UQCRB inhibitors, 1c (A1893) and 1f (A1938), were synthesized and characterized as described previously (15). The compounds were dissolved in dimethyl sulfoxide (DMSO) at a concentration of 100 mM as a stock solution and then further diluted with culture media for appropriate working doses. The negative control groups were treated with equal volumes of DMSO. Gelatin and laminin were purchased from Sigma-Aldrich (St. Louis, MO, USA), and Matrigel^® was obtained from BD Biosciences (San Jose, CA, USA). Anti-CD133 and anti-HIF-1α antibodies were purchased from Miltenyi Biotec GmbH (Bergisch Gladbach, Germany) and BD Biosciences, respectively. Anti-Nanog, anti-Sox2, anti-Oct4, anti-phospho-Met, anti-Met, anti-phospho-Stat3, anti-Stat3, anti-phospho-Akt, anti-Akt, anti-phospho-Erk1/2, anti-Erk1/2, anti-VEGF and anti-β-actin antibodies were obtained from Cell Signaling Technology (Danvers, MA, USA).

Human glioblastoma cell lines, U87MG and U373MG, were obtained from the Korean Cell Line Bank (KCLB). The cells were cultured in minimum essential medium (Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (Gibco) and 1% penicillin-streptomycin-amphotericin B (Lonza, Walkersville, MD, USA), and maintained at 37°C in a humidified 5% CO₂ incubator. To enrich glioblastoma stem-like cells (GSCs), the cells grown in serum-based media were suspended in Dulbecco's modified Eagle's medium/Nutrient Mixture F-12 (Gibco) containing 1X B-27 serum-free supplement (Gibco), 5 µg/ml heparin (Sigma-Aldrich), 2 mM L-glutamine (Gibco), 20 ng/ml epidermal growth factor (EGF; Gibco), 20 ng/ml basic fibroblast growth factor (bFGF; Koma Biotech, Seoul, Korea) and 1% penicillin/streptomycin (Gibco). The serum-free media with EGF and bFGF were added to the cells twice a week. Neurospheres were passaged every 7 days by dissociating with Accutase (Millipore, Temecula, CA, USA).

Cell growth was evaluated by WST-1 assay, a water-soluble tetrazolium salt method. Neurosphere cells were dissociated with Accutase and seeded into 96-well culture plate at a density of 3×10³ cells/well using the serum-free media with EGF and bFGF. After 7 days of exposure to UQCRB inhibitors, 10 µl WST-1 reagent solution (Dogen, Korea) was added to each well, and the cells were incubated for additional 3 h at 37°C. The absorbance was measured at a wavelength of 450 nm using a microplate reader (Thermo Scientific, Vantaa, Finland).

Cell viability was evaluated by trypan blue exclusion assay. Dissociated neurosphere cells were seeded at a density of 1×10⁵ cells/well in 12-well culture plate using the serum-free media with EGF and bFGF. UQCRB inhibitors were added to each well and the cells were incubated for up to 7 days. The cells were stained with trypan blue and counted using a hemocytometer. Cell viability was calculated as the number of viable cells divided by the total number of cells.

Neurosphere cells were dissociated with Accutase, and seeded into 96-well culture plate at a density of 50 cells per well using the serum-free media with EGF and bFGF. The cells were treated with UQCRB inhibitors and cultured for 1–2 weeks. The number of neurospheres in each well was counted under an optical microscope (Olympus, Center Valley, PA, USA). Data were presented as the percentage of sphere-forming cells relative to DMSO-treated control.

The ibidi culture inserts (ibidi GmbH, Martinsried, Germany), which consist of two chambers separated by a 500-µm divider, were used for the migration assay. The inserts were placed into a laminin-coated 24-well culture plate using sterile tweezers. Neurosphere cells were dissociated with Accutase, and single-cell suspensions were prepared at a density of 5×10⁵ cells/ml, of which 70 µl was transferred to each chamber. After cell attachment for 24 h, the culture inserts were removed by using sterile tweezers. The cells were washed with phosphate-buffered saline (PBS) to remove detached cells and incubated with the GSC culture media in the absence or presence of UQCRB inhibitors for up to 48 h. The perimeter of the central cell-free zone was confirmed under an optical microscope (Olympus).

Cell invasion was assayed using a Transwell^® chamber system with polycarbonate filter inserts with a pore size of 8.0 µm (Corning Costar, Acton, MA, USA). The lower side of the filter was coated with 10 µl gelatin (1 mg/ml) and the upper side was coated with 10 µl Matrigel (3 mg/ml). The GSCs (2×10⁵) were placed in the upper chamber of the filter, and UQCRB inhibitors were added to the lower chamber filled with the serum-free media containing EGF and bFGF. The chamber was incubated at 37°C for 48 h, and the cells were subsequently fixed with methanol and stained with hematoxylin/eosin. The total number of cells that invaded the lower chamber of the filter was counted using an optical microscope (Olympus).

Cell lysates were separated by 10% SDS-PAGE electrophoresis and the separated proteins were transferred to polyvinylidene difluoride membranes (Millipore, Billerica, MA, USA) using standard electroblotting procedures. The blots were blocked and immunolabeled with primary antibodies against CD133, Nanog, Sox2, Oct4, phospho-Met (Tyr1234/1235), Met, phospho-Stat3 (Tyr705), Stat3, phospho-Akt (Ser473), Akt, phospho-Erk1/2 (Thr202/Tyr204), Erk1/2, HIF-1α, VEGF and β-actin overnight at 4°C. Immunolabeling was detected with an enhanced chemiluminescence (ECL) kit (Bio-Rad, Hercules, CA, USA), according to the manufacturer's instructions.

Intracellular ROS and mitochondrial ROS levels were detected with 2′,7′-dichlorodihydrofluorescein diacetate (H₂DCFDA) and MitoSOX™ Red mitochondrial superoxide indicator (Molecular Probes, Eugene, OR, USA), respectively. For the assays, the GSCs seeded at a density of 5×10⁴ cells/well in 96-black well culture plate were treated with UQCRB inhibitors for 48 h. After incubation with H₂DCFDA (10 µM) or MitoSOX Red (5 µM) for 10 min, the fluorescence intensity was detected using a multimode microplate reader (Thermo Scientific) at the excitation/emission wavelengths of 495/529 nm for intracellular ROS and 510/580 nm for mitochondrial ROS, respectively. The fluorescent images were also obtained using an Optinity KI-2000F fluorescence microscope (Korea Lab Tech, Seong Nam, Korea).

The mitochondrial membrane potential was detected using the fluorescent, lipophilic dye, JC-1 (5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazol-carbocyanine iodide, Sigma-Aldrich). At hyperpolarized membrane potentials, this dye forms a red fluorescent J-aggregate, whereas at depolarized membrane potentials, this dye remains in its green fluorescent monomeric form. The GSCs were seeded in 24-black well culture plate at a density of 5×10⁴ cells/well and treated with UQCRB inhibitors for 48 h. The cells were incubated with JC-1 (5 µM) for 20 min and the images were obtained using an Optinity KI-2000F fluorescence microscope.

Human UQCRB-specific siRNA (siUQCRB) was constructed as described previously (13). The sense and antisense sequences of this siRNA were 5′-GGG UUA AUG CGA GAU GAU ACA AUA U-3′ and 5′-AUA UUG UAU CAU CUC GCA UUA ACC C-3′, respectively. For depletion of UQCRB mRNA, U87MG GSCs seeded in laminin-coated culture plate with the GSC culture media were transfected with either scrambled negative siRNA or UQCRB siRNA using Lipofectamine 3000 transfection reagent (Invitrogen, Grand Island, NY, USA) according to the manufacturer's instructions. Interference of UQCRB mRNA was validated through reverse transcriptase-polymerase chain reaction (RT-PCR) analysis using specific primers for UQCRB (sense, 5′-ATGGCTGGTAAGCAGGCC-3′; anti-sense, 5′-CTTCTTTGCCCATTCTTC-3′).

The results are expressed as the mean ± standard error (SE). Student's t-test was used to determine statistical significance between the control and test groups. A P-value of <0.05 was considered statistically significant.

All established glioma cell lines, like most other cancer cell lines, are grown in media containing serum, whereas GSCs are grown as neurospheres in serum-free media, since serum causes irreversible differentiation of primary tumor cells (16,17). Accordingly, in this study, GSCs were propagated under an optimal condition using serum-free media supplemented with basic FGF and EGF (18). In order to determine the therapeutic effect of mitochondrial UQCRB inhibitors against GSCs, two synthetic small molecules that are known to target UQCRB, 1c (A1893) and 1f (A1938), were analyzed in the present study (15) (Fig. 1).

We first examined the effect of UQCRB inhibitors on the proliferation of GSCs derived from U87MG and U373MG cells using a water-soluble tetrazolium salt method. The UQCRB inhibitors suppressed the proliferation of both GSCs in a dose-dependent manner (Fig. 2A). To further evaluate whether the GSC growth inhibition by UQCRB inhibitors was due to cytotoxic effects, a viability assay was performed using the trypan blue exclusion method. The viability of GSCs exceeded 80% relative to DMSO-treated control cells, even after treatment with 40 µM of UQCRB inhibitors for 7 days (Fig. 2B). Therefore, the UQCRB inhibitors can inhibit the proliferation of U87MG and U373MG GSCs at subtoxic concentrations.

Sphere-forming assays have been widely used to evaluate the stemness of cells, their capacity for self-renewal and differentiation, both of which are instrumental in cancer cell formation (6,19). As shown in Fig. 3, the neurosphere formation ability of both GSCs was markedly suppressed by treatment with the UQCRB inhibitors. These results indicate that UQCRB inhibitors can reduce the self-renewal capacity of GSCs.

GSCs have been implicated in accelerating tumor metastasis (20,21). We thus assessed whether UQCRB inhibitors have an effect on the migration and invasion of GSCs. We first confirmed their effect on the migration of both GSCs in 48 h after treatment when compared to control conditions using wound healing assay. The UQCRB inhibitors led to significant reduction of cell migration in U87MG and U373MG GSCs (Fig. 4). Next, invasion assay was performed by employing the Matrigel-coated transwell chamber system. The UQCRB inhibitors distinctly decreased the invasion of U87MG and U373MG GSCs when compared with controls (Fig. 5). These data demonstrate that UQCRB inhibitors suppress the metastatic capability of GSCs.

Recent studies have shown that c-Met signaling induces glioma malignancy by increasing GSC population through upregulation of stemness supporting transcription factors such as Sox2, Klf4, c-Myc, Oct4 and Nanog (22,23). We thus investigated the effect of UQCRB inhibitors on the c-Met signaling cascade in U87MG and U373MG GSCs. Treatment with the UQCRB inhibitors effectively inhibited the phosphorylation of c-Met and its downstream signal transduction effectors including STAT3, Akt and ERK1/2, without affecting the total protein levels, in either GSC (Fig. 6). Recent reports have revealed that hypoxia-inducible factors (HIFs) stimulate specific signaling pathways and transcription factors that control cancer stem cell self-renewal and multipotency and are highly expressed in GSCs (24,25). The UQCRB inhibitors significantly reduced the protein levels of HIF-1α as well as its transcriptional target gene, VEGF in GSCs (Fig. 6). They also decreased the expression of GBM stemness markers such as CD133, Nanog and Sox2, which are known to be regulated by c-Met and HIFs (Fig. 6). Taken together, these data suggest that the suppressive effect of UQCRB inhibitors on the proliferation, stemness, migration and invasiveness of GSCs might be partly associated with the downregulation of c-Met signaling and HIF activity.

Recent studies have indicated that ROS contribute to cancer stem cell-like properties and play pivotal roles in tumorigenesis, metastasis and resistance to radiation and chemotherapy (26–28). Since UQCRB functions as a mitochondrial ROS mediator (12,13), we examined whether UQCRB inhibitors affect intracellular ROS levels in U87MG and U373MG GSCs using a fluorescent probe for ROS measurement, 2′,7′-dichlorofluorescein. The UQCRB inhibitors reduced the intracellular ROS levels in both GSCs (Fig. 7A). Moreover, mitochondrial ROS levels in GSCs were measured using a MitoSOX™ Red mitochondrial superoxide indicator. The UQCRB inhibitors noticeably inhibited the mitochondrial ROS generation in both GSCs (Fig. 7B).

Next, to investigate whether UQCRB inhibitors cause mitochondrial dysfunction in GSCs, the mitochondrial membrane potential was measured using the JC-1 fluorescent marker. Treatment with the UQCRB inhibitors led to a decrease in red fluorescence and an increase in green fluorescence in U87MG and U373MG GSCs, indicating that they cause a loss of the mitochondrial membrane potential in the cells (Fig. 8). These data demonstrate that UQCRB inhibitors may modulate the mitochondrial function in GSCs, especially at mitochondrial ROS generation.

To elucidate the role of UQCRB in the maintenance of GSCs, we performed UQCRB depletion experiments in U87MG GSCs. The cells were transfected with either UQCRB siRNA (siUQCRB) or scrambled negative siRNA, and knockdown of the UQCRB gene was confirmed through RT-PCR analysis (Fig. 9A). We first investigated the effect of UQCRB knockdown on the cancer stem cell-like phenotypes of GBM cells. As shown in Fig. 9B, UQCRB knockdown significantly inhibited the growth of U87MG GSCs. In addition, the neurosphere formation and migration abilities of the GSCs were noteworthily suppressed by UQCRB knockdown (Fig. 9C and D). The results suggest that mitochondrial UQCRB positively regulates the cancer stem cell-like properties of GBM cells.

We next investigated the effect of UQCRB knockdown on the c-Met signaling, HIF and stemness markers in U87MG GSCs. UQCRB depletion reduced the activation of c-Met and its downstream signal transduction mediators, STAT3, Akt and ERK1/2 (Fig. 10). UQCRB knockdown also decreased the expression of HIF-1α and VEGF. The down-regulation of c-Met signaling and HIF activity consequently led to a reduction in the expression of GBM stemness markers such as CD133, Oct4 and Sox2 (Fig. 10).

Furthermore, we assessed whether the inhibition of the GSC characteristics by UQCRB depletion is associated with the regulation of mitochondrial function. UQCRB knockdown significantly suppressed the mitochondrial ROS generation and diminished the mitochondrial membrane potential in U87MG GSCs (Fig. 11). These results strongly demonstrate that UQCRB might function as a crucial mediator of GSC maintenance through the modulation of mitochondrial ROS generation in GSCs.