Glioblastomas (GBMs), the most frequent and aggressive primary brain tumors, are characterized by increased proliferation, resistance to chemotherapy and radiotherapy and invasion into the surrounding normal brain tissue.1, 2 Current treatments include surgery, radiation therapy and chemotherapy.3, 4 Unfortunately, the prognosis of patients with GBMs remains extremely poor and has not changed significantly during the past several years.5, 6 Therefore, novel therapeutic approaches are needed to improve the poor prognosis of these patients.
Recently, a small subpopulation of CD133+ cancer stem cells has been identified in specimens of GBM.7, 8 These glioma stem cells (GSCs) express additional stem cell markers, exhibit self-renewal and differentiation to glial and neuronal lineages, and can initiate xenograft tumors.9, 10 Cancer stem cells in various tumors, including the GSCs, have been implicated in the enhanced radio-resistance and in the repopulation of tumors following these treatments.10, 11 Thus, delineating the molecular mechanisms underlying the increased resistance of these cells to anticancer therapies is of utmost importance.
Autophagy is a cellular pathway involved in protein and organelle degradation.12, 13 This process is regulated by a series of autophagy-related genes (ATGs) and a number of signaling molecules such as mTOR, AKT, and class I and class III phosphatidylinositol 3-kinase.14, 15 Autophagy is frequently activated in tumor cells following anticancer therapies such as chemotherapeutic drugs16, 17 or γ-irradiation18 and can either contribute to cell death or represent a mechanism of resistance to these treatments.19, 20 Treatment of glioma cells with temozolomide, etoposide and γ-radiation induces autophagy in glioma cells and autophagy inhibitors increase the sensitivity of glioma cells to γ-radiation and temozolomide, suggesting that autophagy acts as a protective mechanism in glioma cells in response to these treatments. Although the induction and role of autophagy has been described in various cancer cells, its role in GSCs has not been yet reported.
In this study we examined the induction of autophagy by γ-radiation and its role in the radioresistance of CD133+ GSCs. We found that γ-radiation induced autophagy in GSCs and that the CD133+ cells exhibited a larger degree of autophagy compared with the CD133− cells. Similarly, the CD133+ cells expressed higher levels of LC3, ATG5 and ATG12 and inhibition of autophagy sensitized these cells to γ-radiation. These results suggest that targeting autophagy may abrogate the resistance of GSCs to γ-radiation and can lead to the development of novel therapeutic approaches for the treatment of GBMs.
Material and methods
Anti-beclin1, ATG5, ATG7, ATG12 and actin were obtained from Santa Cruz (Santa-Cruz, CA). Anti-LC3 antibody was obtained from Cell Signaling Technology (Beverly, MA).
GSCs and enrichment of CD133+ cells
GBM specimens were dissociated in 0.05% Trypsin/EDTA for 4 hr at RT followed by incubation in DMEM/F-12 medium containing 0.7 mg/ml ovomucoid. The tissue was then triturated mechanically with fire-narrowed Pasteur pipette and filtrated through a 40-mm mesh. Cells were density centrifuged in Lympholyte-M to remove red blood cells and were then maintained in neurosphere medium supplemented with 20 ng/ml EFG and 20 ng/ml FGFb. Spheroids were examined for self-renewal, expression of astrocytic, oligodendrocytic and neuronal markers upon plating on poly-D-ornithine in serum-containing medium and for the expression of CD44, Bmi-1, CD133, Musashi-1, Sox2 and Nestin. For enrichment of CD133+ cells, neurospheres were dissociated into single-cell suspension and the cells were then centrifuged for 10 min at 300g. The isolation was performed according to the MACS CD133 kit manual (Miltenyi Biotech, Auburn, CA). Briefly, cells were resuspended in PBS solution (pH7.2, 0.5% BSA, 2 mM EDTA) per 108 cells. Then blocking reagent FcR (100 μl/108 cells) (Miltenyi Biotech) and CD133/1 Biotin antibody were added and mixed at 4°C for 10 min. Cells were washed in 10–20× volume with PBS solution three times. The pellet was resuspended in antibiotin microbeads (100 μl/108 cells) and mixed at 4°C for 10 min. Cells were washed three times in PBS solution and then added to a prewashed magnetic separation (MS) column on the magnetic holder. The column was washed three times and cells were collected as the negative fraction. The positive fraction was run through another fresh MS column and washed three times, collected and analyzed by FACS or placed in a 12-well dish with NM media (i.e., DMEM/F12, 1×N2 supplement, 50 μg/ml BSA, 25 μg/ml Gentamicin, Pen/Strep, 20 ng/ml EGF and 20 ng/ml FGFb).
CD133+ cells were transfected with SureSilencing ATG5, Beclin1 or control shRNA plasmids (SuperArray, Frederick, MD). Prior to transfection, the GSC spheroids were mechanically dissociated to smaller spheroids, and transfection was performed by electroporation using the Nucleofector device program A027 and the mouse NSC Nucleofector kit (Amaxa Biosystems, Gaithersburg, MD).
Detection of acidic vesicular organelles using acridine orange staining
For the detection of vacuoles, cells were stained with acridine orange at a final concentration of 1 μg/ml for 15 min, removed from the plate with trypsin-EDTA, and collected in phenol red-free growth medium. Green (510–530 nm) and red (>650 nm) fluorescence emission from cells illuminated with a blue (488 nm) excitation light was measured with a FACSCalibur from Becton Dickinson using CellQuest software.
WST-1 assay was performed according to the manufacturer's instructions. Briefly, cells transfected with control shRNA, ATG5 or beclin 1 shRNA plasmids (Superarray) were plated in a 96-well plate at a concentration of 25,0000 cells/ml. Cells were then irradiated with 5 Gy and incubated for 24 and 48 hr thereafter. The WST-1 assay (Roche Diagnostics, Mannheim, Germany) was used to assess the number of viable cells and the relative absorbance was measured at 595 nm.
Neurosphere formation assay
For the ability of CD133+ cells to form secondary neurospheres, disaggregated cells were irradiated in the presence of autophagy inhibitors or in cells in which ATG5 and beclin1 were silenced for 48 hr. Cells were plated in 24-well plates at a density of 100 cells/well through limiting dilution and the number of neurospheres/well was determined 7 days thereafter for eight different wells. Spheres that contained more than 20 cells were scored. Results are presented as % of maximal neurospheres formed in control untreated cells.
Preparation of cell homogenates and immunoblot analysis
Cells lysates (30 μg protein) were resolved by SDS-PAGE and transferred to nitrocellulose membranes. Following incubation with the primary antibody, specific reactive bands were detected using a goat anti-rabbit or goat anti-mouse IgG conjugated to horseradish peroxidase (BioRad, Hercules, CA) and the immunoreactive bands were visualized by the ECL Western blotting detection kit (Amersham, Arlington Heights, IL).
The results are presented as the mean value ± SE. Data were analyzed using analysis of variance and a Student's t test. For real-time PCR data, Student's t test (with correction for data sets with unequal variances) was done using Prism 4 (GraphPad Software, Inc., San Diego, CA).
γ-Radiation induces autophagy in CD133+ GSCs
In these studies, we employed two preparations of CD133+ cells, which were generated from two different GBM specimens (HF2355 and HF2359). The HF2355 cells were grown as spheroids, expressed CD133 and Sox2 (Fig. S1A) and differentiated to neuronal, astrocytic and oligodendrocytic lineages upon plating on poly-D-ornithine coated plates in serum-containing medium (Fig. S1B). Moreover, the CD133+ cells exhibited self-renewal and on plating of 1 cell per well a small percentage (12–15%) of the cells formed spheres within 40 days; further minimal dilution assays confirmed that these cells have the potential to grow infinitely. In addition, these cells generated tumors that recapitulated the tumors of origin when injected intracranially (Fig. S1C). Similar results were obtained for the HF2359 cells (data not shown).
γ-radiation has been shown to induce autophagy in glioma cells21; however, the induction of autophagy in GSCs has not been yet reported. We therefore examined the effect of γ-radiation on the induction of autophagy in the HF2355 and HF2359 CD133+ cells. The cancer cell neurospheres were γ-irradiated (5 Gy) and analyzed following 48 hr in culture for the expression of LC3I and LC3-II using Western blot analysis. As demonstrated in Figure 1a, γ-radiation increased the expression of LC3-II in the two cell preparations that were examined.
Since CD133+ cells are more resistant than CD133− cells to radiation and chemotherapeutic drugs,11 we compared the induction of autophagy by γ-radiation in these two cell subpopulations. As presented in Figures 1b and 1c, the expression of LC3-II was significantly increased in the HF2359 CD133+ cells following 24 and 48 hr (Fig. 1b) and in the HF2355 CD133+ cells after 48 hr of treatment (Fig. 1c), whereas no significant increase was observed in the CD133− cells of both HF2359 and HF2355 GSCs (Figs. 1b and 1c).
Autophagy is characterized by sequestering cytoplasmic proteins into the lytic component—a process which leads to the formation of acidic vesicular organelles (AVO).16 Similar to the results that were obtained with the LC3-II expression, we found that γ-radiation induced a significantly larger increase in the development of AVO in the CD133+ cells as compared with the CD133− cells in both the HF2359 (Fig. 1d) and HF2355 (Fig. 1e) cells. Immunofluorescence staining of the CD133+ cells with anti-CD133 antibody and FACS analysis demonstrated that around 62% of the HF2359 cells and 55% of the HF2355 cells expressed the CD133 protein in these experiments.
Differential expression of LC3, ATG5 and ATG12 in CD133+ and CD133− cells
We then examined the expression of various autophagy-related proteins in the CD133+ and CD133− cells. Using Western blot analysis we found that in both the HF2355 and the HF2359 cell lines, the CD133+ cells expressed higher levels of ATG5, ATG12 and LC3 compared with the CD133− cells, whereas no differences were observed in the expression of beclin 1 and ATG7 (Fig. 2).
Inhibition of autophagy by bafilomycin A1 enhances the cytotoxic effect of γ-radiation in the CD133+ cells
To examine the role of autophagy in the cytotoxic effect of γ-radiation on the CD133+ cells, we employed the autophagy inhibitors, 3-methyladenine (3-MA), a phosphatidylinositol 3-phosphate kinase inhibitor22 and bafilomycin A1, a specific vacuolar type H+-ATPase inhibitor, which inhibits autophagy at a late stage by inhibiting fusion between the autophagosome and lysosome.23 γ-irradiation of the HF2359 CD133+ cells induced only a small decrease in cell viability. In contrast, combined treatment of the cells with bafilomycin A1 and γ-radiation significantly decreased the viability of the cells, beyond the effect that was induced by either γ-radiation or bafilomycin alone (Fig. 3a). The other autophagy inhibitor that was used in this study, 3-MA, had a smaller effect on the viability of the γ-irradiated cells (Fig. 3a). Similar results were obtained using neurosphere formation assay. Irradiation of the CD133+ cells only slightly decreased the ability of the CD133+ cells to generate neurospheres. Treatment with the autophagy inhibitors alone did not interfere with neurosphere formation; however, both inhibitors exerted a strong inhibitory effect on neurosphere formation in γ-irradiated cells (Fig. 3b). In addition, these cells generated significantly smaller neurospheres.
The CD133− cells exhibited higher sensitivity to γ-radiation compared with the CD133+ cells. However, treatment of the cells with either 3-MA or bafilomycin A1 only marginally increased the sensitivity of these cells to the cytotoxic effect of γ-radiation (Fig. 3c).
Silencing of beclin1 and ATG5 increases the sensitivity of Cd133+ cells to γ-radiation
To further determine the role of autophagy in the cytotoxic effect of γ-radiation, we silenced the expression of ATG5, which is required for autophagy at the stage of autophagosome-precursor synthesis and beclin1, which participates in the autophagosome formation, in the HF2359 CD133+ cells. For these experiments we transfected the CD133+ cells with shRNAs targeting beclin1 and ATG5 and using electroporation. The neurospheres were gently dissociated and electroporated using the Nucleofector Amaxa. In preliminary experiments, we observed around 70% transfection efficiency in the cells transfected with GFP using this approach (Fig. 4a). As presented in Figure 4b, transfection of the HF 2359 CD133+ cells with either beclin 1 or ATG5 shRNAs significantly decreased the expression of these proteins as compared with cells transfected with the control shRNA.
We then examined the response of the silenced cells to γ-radiation. The CD133+ cells transfected with control, ATG5 or beclin1 shRNAs were incubated for three days with neurosphere medium and then γ-irradiated (5 Gy). The viability of the cells was determined following 48 hr using the WST-1 assay. Silencing of ATG5 or beclin1 did not affect the viability of the cells. Similarly, γ-irradiation of the cells induced only a small decrease in cell viability. In contrast, silencing of beclin1 and ATG5 decreased significantly the viability of the CD133+ cells following γ-radiation as compared with the effect observed in control shRNA-transfected cells (Fig. 4c). In addition, silencing of beclin 1 or ATG5 significantly decreased the ability of the cells to generate neurospheres in the γ-irradiated cells (Fig. 4d).
In this study we examined the induction of autophagy by γ- radiation in CD133+ cells and the role of autophagy in the radioresistance of these cells. Recent studies have implicated the small subpopulation of CD133+ GSCs in the repopulation of tumors and in the radioresistance of gliomas.7, 8, 11 One of the mechanisms that mediate the resistance of glioma cells to different antitumor therapies are the induction of autophagy16, 21, 23, 24; however, the role of this process in GSCs has not been described. We found that similar to its effect on glioma cells, γ-radiation also induced autophagy in CD133+ cells.
γ-radiation induced a larger degree of autophagy in the CD133+ cells as compared with the CD133− cells. Similarly, the CD133+ and CD133− cells expressed different levels of the autophagy-related proteins, LC3, ATG5 and ATG12, whereas only minor differences were observed in the expression of beclin1. LC3 is a homologue of Apg8 and essential for the formation of the autophagosome,25 whereas ATG5 and ATG12 are part of an ubiquitin-like conjugation system, which is associated with a protein/phospholipids complex that localizes to the autophagosomal membranes.26, 27 The importance of the increased expression of LC3, ATG5 and ATG12 in the CD133+ cells is not completely understood; however, it may contribute to the enhanced ability of these cells to undergo autophagy in response to γ-radiation and other stimuli. The majority of the reports regarding increased expression of different autophagy-related proteins are in response to treatments that induce autophagy such as γ-radiation.18, 28 In addition, there are a number of studies that reported altered beclin1 expression in different tumors suggesting a role of this protein in cell death and survival depending on the cellular context.29, 30 Similarly, the expression of LC3 has been recently reported to be highly expressed in gastrointestinal cancers.31
Inhibition of autophagy by bafilomycin A1 significantly enhanced the sensitivity of the CD133+ cells to γ-radiation, whereas a more moderate effect was observed with the other autophagy inhibitor 3-MA. Similarly, silencing of ATG5 and beclin1 rendered the CD133+ cells more sensitive to the cytotoxic effects of γ-radiation. In contrast, the CD133− cells exhibited higher sensitivity to γ-radiation; however, inhibition of autophagy only marginally affected their radiosensitivity. These results suggest that in the CD133+ cells autophagy plays a protective role against the effect of γ-radiation and therefore may contribute to the radioresistance of these cells. Silencing of different autophagy-related proteins has recently been shown to increase the radiosensitivity of various cancer cells.18 Similarly, inhibition of autophagy increased the sensitivity of glioma cells to γ-radiation,22 temozolomide32 and arsenic trioxide.24 Interestingly, only bafilomycin A1, which acts after the association of LC3 with the autophagosome membrane, enhanced the cytotoxicity of temozolomide.32
Autophagy may protect cancer cells against radiation damage by decreasing cytoplasmic acidification, by providing catabolites that are required for the repair processes and by removing toxic molecules.19, 33 The increased radioresistance of CD133+ cells compared with CD133− cells has been recently attributed to the preferential activation of the DNA damage checkpoint response and an increase in DNA repair capacity.11 Thus, irradiation of CD133+ cells increased the phosphorylation of proteins such as Chk1, Chk2 and ATM. The association of the increased checkpoint activation and the induction of autophagy by γ-radiation in the CD133+ cells is currently not understood; however, a recent report suggested that the activation of ATM following γ-radiation can lead to the phosphorylation of LKB-1 and to inhibition of the mTOR pathway via an ATM/LKB-1/AMPK pathway, which may contribute to the induction of autophagy.34 Studies of the role of ATM in the induction of autophagy in irradiated CD133+ cells are currently in progress.
In summary, the results of this study demonstrate that CD133+ cells express higher levels of the autophagy-related proteins, ATG5, ATG12 and LC3. Moreover, γ-radiation induces a larger degree of autophagy in these cells as compared with CD133− cells and inhibition of this process significantly radiosensitizes the CD133+ GSCs. Thus, the induction of autophagy in these cells may contribute to their radioresistance and autophagy inhibitors could enhance the sensitivity of CD133+ GSCs to γ-radiation.