The mTORC1/mTORC2 inhibitor AZD2014 enhances the radiosensitivity of glioblastoma stem-like cells

Abstract

Background

The mammalian target of rapamycin (mTOR) has been suggested as a target for radiosensitization. Given that radiotherapy is a primary treatment modality for glioblastoma (GBM) and that mTOR is often dysregulated in GBM, the goal of this study was to determine the effects of AZD2014, a dual mTORC1/2 inhibitor, on the radiosensitivity of GBM stem-like cells (GSCs).

Methods

mTORC1 and mTORC2 activities were defined by immunoblot analysis. The effects of this mTOR inhibitor on the in vitro radiosensitivity of GSCs were determined using a clonogenic assay. DNA double strand breaks were evaluated according to γH2AX foci. Orthotopic xenografts initiated from GSCs were used to define the in vivo response to AZD2014 and radiation.

Results

Exposure of GSCs to AZD2014 resulted in the inhibition of mTORC1 and 2 activities. Based on clonogenic survival analysis, addition of AZD2014 to culture media 1 hour before irradiation enhanced the radiosensitivity of CD133+ and CD15+ GSC cell lines. Whereas AZD2014 treatment had no effect on the initial level of γH2AX foci, the dispersal of radiation-induced γH2AX foci was significantly delayed. Finally, the combination of AZD2014 and radiation delivered to mice bearing GSC-initiated orthotopic xenografts significantly prolonged survival as compared with the individual treatments.

Conclusions

These data indicate that AZD2014 enhances the radiosensitivity of GSCs both in vitro and under orthotopic in vivo conditions and suggest that this effect involves an inhibition of DNA repair. Moreover, these results suggest that this dual mTORC1/2 inhibitor may be a radiosensitizer applicable to GBM therapy.

Whereas radiotherapy significantly prolongs the survival of patients with glioblastoma (GBM), the median survival rate of patients with GBM remains 12 to 15 months after diagnosis even in combination with surgery and chemotherapy.¹ An approach to improving the effectiveness of GBM therapy is the development of molecularly targeted radiosensitizers, a strategy that requires a thorough understanding of the mechanisms mediating cellular radioresponse. Along these lines, studies have recently shown that radiation selectively regulates mRNA translation, a process that operates independently from transcription.^2,3 With respect to functional consequence, the radiation-induced changes in mRNA translation correlate to changes in the corresponding protein, in contrast to changes in the radiation-induced transcriptome. Because translational control of gene expression is a component of the cellular radioresponse, we recently tested the role of eukaryotic initiation factor 4E (eIF4E), the rate-limiting component in cap-dependent translation initiation, as a determinant of radiosensitivity.⁴ In that study, knockdown of eIF4E was shown to enhance the radiosensitivity of tumor but not normal cell lines, which suggested that strategies targeting eIF4E activity may provide tumor selective radiosensitization.

A critical regulator of eIF4E is the mechanistic target of rapamycin (mTOR), which plays a critical role in regulating mRNA translation and protein synthesis in response to a variety of environmental signals. mTOR is the kinase component of 2 distinct complexes: mTOR complex 1 (mTORC1) and mTOR complex 2.⁵ The major substrates for mTORC1 kinase activity are eIF4E-binding protein 1 (4E-BP1) and the ribosomal protein s6 kinase 1 (S6K1). In the hypophosphorylated state, 4E-BP1 binds to eIF4E preventing its association with eIF4G, the formation of the eIF4F complex, and cap-dependent translation.⁶ However, when 4E-BP1 is phosphorylated by mTORC1, it is released from eIF4E, and the eIF4F cap-complex is assembled.⁶ With respect to regulating eIF4E, the critical substrate of mTORC2 is AKT at s473, which can indirectly lead to enhancement mTORC1 activity.^7,8

mTOR is frequently dysregulated in GBM⁹ and is a major downstream effector of a number of signaling pathways including PI3K/AKT, RAS/MAPK, and RTKs, which have been implicated in gliomagenesis.^10,11 Accordingly, mTOR kinase has been suggested as a target for GBM therapy. Most studies targeting mTOR in GBM^12,13 and cancer in general¹⁴ have focused on the allosteric inhibitor rapamycin and its analogs (rapalogs), which incompletely inhibit mTORC1 output and do not inhibit mTORC2.¹⁵ As single agents, these drugs have shown modest activity with respect to patient outcomes,¹⁶ which has been attributed to their incomplete inhibition of 4E-BP1 phosphorylation, feedback activation of AKT, and/or the lack of mTORC2 inhibition.^15,17 In contrast to the allosteric inhibitors like rapamycin, more recently developed competitive inhibitors of mTOR inhibit mTORC1 output more completely and inhibit mTORC2, which prevents the feedback activation of AKT following S6K inhibition.^7,18–21 We recently showed that for established tumor cell lines, in contrast to rapamycin, the mTORC1/2 inhibition achieved by the competitive inhibitor PP242 enhanced tumor cell radiosensitivity.²² However, PP242 has unfavorable pharmacokinetics in humans²³ and is not considered applicable to GBM therapy. Thus, to investigate the potential of mTOR to serve as a target for GBM radiosensitization, we determined the effects of the competitive inhibitor AZD2014, which has recently entered clinical trials as a single agent,²⁴ on the radiosensitivity of glioblastoma stem-like cells (GSCs) in vitro and GSC-initiated orthotopic xenografts.

Materials and Methods

GSC Culture

In vitro studies were performed using 4 neurosphere-forming cultures isolated from human GBM surgical specimens: GBMJ1 and GBAM1²⁵; NSC23²⁶ (kindly provided by Dr. Frederick Lang, MD Anderson Cancer Center), and 0923.²⁷ Neurospheres were maintained in stem cell medium consisting of DMEM/F-12 (Invitrogen), B27 supplement (1X) (Invitrogen), and human recombinant bFGF and EGF (50 ng/mL each) (R&D Systems ). All cultures were maintained at 37°C in an atmosphere of 5% CO₂/7% O₂.²⁸ CD133+ cells (GBMJ1, GBAM1, and NSC11) or CD15+ cells (0923) were isolated from each neurosphere cultures by FACS²⁵ and used as a source for the described experiments. The CD133+ and CD15+ cell cultures met the criteria for tumor stem-like cells²⁹ including self renewal, differentiation along glial and neuronal pathways, expression of stem cell related genes, and formation of brain tumors when implanted in immunodeficient mice.^25,28,30 For use in an in vitro experiment, CD133+ or CD15+ neurosphere cultures were disaggregated into single cells as described²⁵ and seeded onto poly-L-lysine (Sigma) or poly-L-ornithine/laminin (Sigma)³¹ coated tissue culture dishes in stem cell media. Under these conditions, single-cell glioma stem cells attach and proliferate maintaining their CD133+ or CD15+ expression and stem-like characteristics.²⁵ Monolayer cultures were treated with AZD2014 (Astra-Zeneca) dissolved in dimethyl sulfoxide (DMSO) or vehicle control. Radiation was delivered using a 320 kV X-ray machine (Precision XRay Inc.) at a dose rate of 2.3 Gy/min.

Clonogenic Survival Assay

GSC neurospheres were disaggregated into single cells and plated at clonal density into 6-well plates coated with poly-L-lysine, which results in adherent colony formation.²⁵ Twenty-four hours after seeding, a time sufficient for attachment yet prior to the onset of cell division, the specified treatment was delivered. Colonies were stained 21 days later with 0.5% crystal violet, and the number of colonies containing at least 25 cells was determined. After normalizing for cytotoxicity induced by AZD2014 alone, radiation survival curves were generated. Data presented are the mean ± SE of 3 independent experiments.

Immunoblot Analysis

GSCs were plated into poly-L-ornithine/laminin coated plates and grown to approximately 70% confluency. Cells were lysed in 50 mmol/L Tris-HCl (pH 7.5), 150 mmol/L NaCl, 2 mmol/L EDTA, 2 mmol/L EGTA, 25 mmol/L NaF, 25 mmol/L β-glycerophosphate, 0.2% Triton X-100, 0.3% NP-40, and 0.1 mmol/L sodium orthovanadate, supplemented with 1x phosphatase inhibitor cocktails II and III (Sigma-Aldrich), and 1x HALT protease inhibitor cocktail (Thermo Scientific) for 15 minutes on ice. Protein quantification was done with BCA protein assay (Thermo Scientific). Proteins were diluted in SDS-PAGE and electrophoresed and transferred (Bio-Rad). Membrane was blocked in 5% bovine serum albumin (Fisher), incubated with antibody overnight at 4°C, and incubated with HRP-coupled secondary antibody for 2 hours at room temperature. Bands were visualized with Pierce ECL Western Blotting Substrate (Thermo Scientific). Primary antibodies included anti-4E-BP-1, anti-AKT, anti-phospho-AKT S473, anti-phospho-4E-BP-1 T37/46, anti-phospho-4E-BP-1 S65 (Cell Signaling Technology), anti-phosph-S6K T389, anti-S6K (Epitomics), and anti-β-actin (Sigma-Aldrich). Donkey anti-rabbit and sheep anti-mouse Horseradish Peroxidase conjugated secondary antibodies were used (GE Healthcare).

Immunofluorescent Analysis of γH2AX Foci

GSCs were seeded onto LabTek CC2-treated tissue culture slides (Thermo Fisher) and 24 hours later subjected to the specified treatment. Slides were then fixed with 10% neutral buffered formalin, permeabilized with 0.1% Triton X-100, and blocked with 1% bovine serum albumin in PBS containing 5% goat serum. The slides were incubated with antibody to phospho-H2AX (Millipore) followed by secondary antibody with goat-anti-mouse-Alexa488 (Invitrogen), and mounted with Prolong gold antifade reagent containing DAPI (Invitrogen) to visualize nuclei. Cells were analyzed on a Zeiss upright fluorescent microscope. Data presented are the mean ± SE of 3 independent experiments in which 50 cells were evaluated.

Apoptotic Cell Death

Cells undergoing apoptosis were quantified according to annexin V staining (Annexin V Apoptosis Detection Kit, BD Biosciences). Briefly, cells were resuspended in 1x Annexin V Binding Buffer and incubated with Annexin V-Cy5 antibody in the dark at room temperature for each treatment condition. Hoechst 33258 was added for live/dead discrimination, and samples were analyzed by flow cytometry (Millipore guava EasyCyte flow cytometer).

G₂/M Checkpoint

Activation of the G₂/M cell cycle checkpoint was defined according to mitotic index as described by Xu et al.³² GSC cells were seeded into poly-L-ornithine/laminin coated tissue culture plates and treated with AZD2014 (2 µM) and/or radiation (2 Gy) 24 hours later. Immediately after irradiation, cells were treated with nocodazole (50 ng/mL) (Sigma) to prevent cells from exiting mitosis.³³ Cells were collected at times indicated and stained with anti-phosphorylated histone H3 (Millipore) and analyzed with Guava EasyCyte flow cytometer (Millipore). Mitotic cells were those defined by histone H3 expression with a 4N DNA content.³²

Intracranial Xenografts

Eight-to-10 week old athymic female nude mice (NCr nu/nu; NCI Animal Production Program) were used in these studies. For in vivo studies, CD133+ GBMJ1 cells engineered to express luciferase using the lentivirus LVpFUGW-UbC-ffLuc2-eGFP2, a bimodal expression vector fused with the combination of the bioluminescent protein ffLuc2 and fluorescent protein eGFP2 under the control of the UbC promoter, were used as previously described.³⁴ For orthotopic implantation, mice were anesthetized using with 2% isoflurane in an oxygen/air (40/60%) mixture, and the gas levels were adjusted to maintain normal breathing rate. The head was held in a stereotaxic jig (Stoelting Co.), a central dorsal incision of ∼2 cm was made, and 10⁵ CD133+ cells injected in a total volume of 5 µL at 1.0 mm anterior and 2.0 mm lateral to the bregma to a depth of 3.5 mm at a rate of ∼1 µL/min.³⁰ Bioluminescent imaging (BLI) was performed as described³⁴ beginning at 1 week after implantation. At 12 days postimplantation, consistent BLI was detected in all mice, which were then randomized into 4 treatment groups: control, AZD2014 (50 mg/kg, oral gavage), irradiation (IR), 3 × 12 Gy), and AZD2014 plus IR. Specifically, AZD2014 treatment was followed by IR (12 Gy) for 3 consecutive days. For irradiation (Pantak×Ray), mice were anesthetized using a cocktail of ketamine/xylazine/acepromazine and placed in well-ventilated Plexi glass jigs with shielding for the entire torso of the mouse along with critical normal structures of the head (eg, ears, eyes, neck). Mice were monitored every day until the onset of neurologic symptoms (morbidity). All experiments were performed as approved by the principles and procedures in the NIH Guide for Care and Use of Animals.

Immunofluorescent Histochemistry

At the initial signs of morbidity, mice were euthanized by CO₂ inhalation and perfused with 4% paraformaldehyde in PBS (pH 7.4) via cardiac puncture. Brains were then removed and placed in 10% buffered formalin before embedding in paraffin. The paraffin-embedded brains were cut into 6-μm-thick slices; sections were deparaffinized in xylene and rehydrated in decreasing amounts of alcohol. Sections were boiled in citrate buffer and incubated in 1% bovine serum albumin in PBS containing 10% goat serum. Primary antibodies anti-mouse human nestin (Millipore) and anti-rabbit phosphorylated 4E-BP1 t37/46 (Cell Signaling) were incubated overnight at 4°C followed by secondary antibodies Alexa Fluor 488 antirabbit IgG and Alexa Fluor 555 antimouse IgG, and then mounted with mounting media with DAPI (Vector) to visualize nuclei. Micrographs were generated using a Zeiss confocal microscope.

Statistical analysis

In vitro experiments were repeated 3 times and statistical analysis performed using Student' t test. Data are presented as mean ± SE. For in vivo studies, Kaplan–Meier curves were generated and log-rank values calculated.

Results

To investigate the effects of AZD2014 on the radiosensitivity of GSCs, initial studies focused on GBMJ1 cells. This GSC line is CD133+ and has the in vitro stem-cell like characteristics of continuous self-renewal, expression of stem-cell related genes, and the capacity to partially differentiate along glial and neuronal pathways.^29,35 For analyses of mTOR kinase activity, GBMJ1 neurospheres were disaggregated and grown on poly-l-ornithine/laminin coated tissue culture plates, monolayer conditions under which GSCs maintain their CD133 expression and stem-cell like characteristics.²⁸ Initially, mTORC1 and mTORC2 activities were determined at 1 hour as a function of AZD2014 concentration using p-S6K (t389) and p-4E-BP1 (t37/46 and s65) as readouts for mTORC1 activity and p-AKT (s473) as a marker for mTORC2 activity. As shown in Fig. 1A, 1 µM AZD2014 resulted in a decrease in p-S6K and p-4E-BP1 as well as p-AKT (s473), indicative of a decrease mTORC1 and mTORC2 activities. A somewhat greater inhibition was achieved by 2 µM with no further decrease in mTORC1/2 activities at 4 µM. mTOR kinase activity was then determined as a function of time after addition of 2 µM AZD2014. To determine mTORC1/2 inhibition as a function of exposure time, AZD2014 was added to GBMJ1 cultures and collected at the specified times (Fig. 1B). Inhibition of mTORC1 and mTORC2 was detectable by 1 hour, reaching a maximum decrease by 6 hours, which was then maintained for at least 24 hours. To determine whether radiation influences mTOR activity, GBMJ1 cells were exposed to 2 Gy and collected for immunoblot analysis at times out to 2 hours (Fig. 2). Based on levels of p-S6K, p-4E-BP1 and p-AKT, radiation did not significantly modify mTORC1 or mTORC2 activity.

Fig. 1.

Effect of AZD2014 on mTORC1 and mTORC2 activities in CD133+ GBMJ1 cells. (A) Cells in monolayer culture were exposed to the indicated concentration of AZD2014 for 1 hour and collected for immunoblot analysis. (B) Cells were exposed to AZD2014 (2 µM) for the specified time and collected for analysis. β-actin was used as a loading control; blots are representative of 2 independent experiments.

Fig. 2.

Influence of radiation on mTORC1 and mTORC2 activities. GBMJ1 CD133+ cells were irradiated (2 Gy) and collected at the specified times for immunoblot analysis. β-actin was used as a loading control; blots are representative of 2 independent experiments.

The effect of AZD2014 on the radiosensitivity of GBMJ1 cells was then measured by clonogenic survival analysis. For this study, GBMJ1 CD133+ neurospheres were disaggregated into single cells and seeded in specified numbers onto poly-l-lysine coated tissue culture plates. Under these conditions, GSCs grow as adherent colonies and maintain their CD133 expression.²⁸ After seeding cells were allowed to attach for 24 hours, AZD2014 was then added at a concentration of 2 µM, which induces the maximum mTOR inhibition (Fig. 1), and cultures were irradiated 1 hour later. Twenty-four hours after irradiation, stem cell media was removed and fresh drug-free media was added; cultures were fed with fresh media weekly, and colonies were counted after 21 days. Addition of AZD2014 1 hour prior to irradiation enhanced the radiosensitivity of GBMJ1 cells, resulting in a dose enhancement factor at a surviving fraction of 0.10 (DEF) of 1.35 (Fig. 3A). AZD2014 (2 µM, 25 h) alone reduced surviving fraction of GBMJ1 cells to 0.72 ± 0.05. To determine whether AZD2014-induced radiosensitization was unique to GBMJ1 cells, the same treatment protocol was applied to the CD133+ GSCs NSC23 and GBAM1 (Fig. 3B and C). AZD2014 exposure enhanced the radiosensitivity of NSC23 and GBAM1 cells with DEFs of 1.33 and 1.51, respectively. Treatment of NSC23 and GBAM1 with AZD2014 alone reduced surviving fractions to 0.88 ± 0.02 and 0.85 ± 0.07, respectively. Given that CD133 is not the only marker for isolating GSCs, the study was extended to the GSC line 0923, which has the in vitro and in vivo characteristics of a tumor stem-like cells, but in contrast to the GSCs evaluated above was isolated based on CD15 expression.²⁷ As shown in Fig. 3D, AZD2014 addition 1 hour prior to irradiation enhanced radiosensitivity of 0923 cells with a DEF of 1.33; AZD2014 (2 µM, 25 h) alone reduced the surviving fraction of 0923 cells to 0.77 ± 0.05. These results indicate that this competitive mTOR inhibitor enhances the in vitro radiosensitivity of GSCs, although AZD2014 alone has little effect on survival.

Fig. 3.

Effects of AZD2014 on GSC radiosensitivity. (A) GBMJ1 CD133+ (B) NSC23 CD133+, (C) GBAM1 CD133+, (D) 0927 CD15+. Cells were seeded into poly-L-lysine coated tissue culture plates and allowed to attach overnight with AZD2014 (2 µM) then added 1 hour before irradiation. Twenty-four hours after irradiation, media was removed, and fresh drug-free media was added. Colony-forming efficiency was determined 21 days later, and survival curves were generated after normalizing for cytotoxicity induced from drug alone. Values represent the mean ± SE of 3 independent experiments.

In the initial treatment protocol evaluating the effects of AZD2014 on GSC radiosensitivity (Fig. 3) the mTOR inhibitor was added to the culture media 1 hour before irradiation. To determine whether this was the optimal exposure protocol for radiosensitization as well as to generate insight into the mechanisms involved, AZD2014 (2 µM) was added to GBMJ1 culture media at various times before and after irradiation followed by clonogenic survival analysis (Fig. 4). In each experiment AZD2014 was removed 24 hours after exposure to radiation, and all survival curves were generated after normalizing for cell killing caused by AZD2014 treatment alone. Treatment of GBMJ1 cells with AZD2014 24 hours before irradiation had no significant effect on their radiosensitivity. Addition of AZD2014 24 hours prior to irradiation resulted in the same degree of radiosensitization (DEF of 1.35) as when added 1 hour before irradiation. When the mTOR inhibitor was added 1 hour after irradiation, the radiosensitivity of GBMJ1 was also increased with a DEF of 1.51. These data indicate that the AZD2014-induced radiosensitization also occurs when the drug is added after irradiation, which is somewhat unusual for radiosensitizers.

Fig. 4.

The influence of timing of AZD2014 treatment on GSC radiosensitivity. GBMJ1 CD133+ cells were seeded and allowed to attach overnight. AZD2014 (2 µM) was added to cultures 24 hours before irradiation (24 h Pre-IR), 2 hours before (2 h Pre-IR), 1 hour before (1 h Pre-IR), or 1 hour after (1 h Post-IR) irradiation. Twenty-four hours after irradiation, media was removed, and fresh drug-free media was added. Colony-forming efficiency was determined 21 days later, and survival curves were generated after normalizing for cytotoxicity induced from drug alone. Values represent the mean ± SE of 3 independent experiments.

To begin to address the mechanism of AZD2014-induced radiosensitization, we focused on GBMJ1 and GBAM1 cells. Given that mTOR inhibitors have been shown to induce apoptosis in certain tumor cell lines,^36,37 we determined whether the AZD2014-induced radiosensitization was due to an enhancement of radiation-induced apoptosis. In this study, apoptosis was defined by Annexin V staining at 24 hours after exposure to 4 Gy for cells with and without AZD2014 treatment. As previously shown for GBMJ1 and other GSCs,²⁵ radiation alone did not induce a significant apoptotic response. AZD2014 alone also had no effect on apoptosis; this mTOR inhibitor also had no effect on radiation-induced apoptosis (data not shown). These results indicate that apoptosis is not the mechanism mediating the AZD2014-induced radiosensitization.

Because mTOR can influence the translation of proteins involved in cell cycle progression,³⁸ a potential mechanism of AZD2014-induced radiosensitization is the abrogation of cell cycle checkpoints. Critical to radiosensitivity is the activation of the G2/M checkpoint, the inhibition of which enhances radiation-induced cell death.³⁹ Radiation-induced activation of the G2/M checkpoint results in a decrease in the mitotic index (% cells in mitosis). To quantify the percent of cells in mitosis, flow cytometry is used to identify cells expressing phosphorylated histone H3 with a 4N DNA content.³² In this analysis, to improve the accuracy of detecting cells moving into mitosis, cultures were treated immediately after irradiation with nocodazole (50 ng/mL), which prevents cells from exiting mitosis resulting in a linear accumulation of mitotic cells.³³ As shown in Fig. 5A (left panel), untreated GBMJ1 cells continue to accumulate in mitosis for at least 24 hours after the addition of nocadazole. Radiation (2 Gy) significantly delayed the increase in percent mitotic cells consistent with the activation of the G2/M checkpoint.³² Whereas AZD2014 (2 µM) alone slowed the accumulation of cells in mitosis, it did not affect the initial delay induced by radiation. Similar results were obtained for GBAM1 cells (Fig. 5A, right panel). These data indicate that AZD2014-induced radiosensitization is not the consequence of abrogation of the G2/M checkpoint.

Fig. 5.

Influence of AZD2014 on the G2/M checkpoint and H2AX foci levels in irradiated GBMJ1 and GBAM1 cells. (A) G₂/M checkpoint activation was determined by mitotic index (% cells in mitosis). Left panel: GBMJ1; right panel: GBAM1. AZD2014 (2 µM) was added 1 hour before irradiation (IR) (2 Gy), which was followed by immediate addition of nocodazole (50 ng/mL). Cells were collected at specified time points for cell cycle distribution analysis and determination of phospho-H3 expression. Values represent the mean ± SE of 3 independent experiments. (B) Radiation-induced γH2AX foci formation and dispersal. Left panel: GBMJ1; right panel: GBAM1. AZD2014 (2 µM) was added 1 hour prior to irradiation (2 Gy) with cells collected at specified times. The number γH2AX foci were determined in at least 50 nuclei per treatment condition. Values represent the mean ± SE of 3 independent experiments, *P < .05.

The critical lesion responsible for radiation-induced cell death is the DNA double strand break (DSB). Because γH2AX foci correspond to radiation-induced DSBs and their dispersal correlates with DSB repair,^40–42 the effects of AZD2014 on radiation-induced γH2AX were evaluated (Fig. 5B). In this study AZD2014 (2 µM) was added 1 hour before irradiation (2 Gy), with γH2AX nuclear foci determined at times out to 24 hours. For GBMJ1 cells (Fig. 5B, left panel), no difference in foci levels was detected between control (vehicle) and AZD2014 treated cells at 1 hour after irradiation, suggesting that mTOR inhibition had no effect on the initial levels of radiation-induced DSBs. However, at 6 hours and 24 hours after irradiation, the number of γH2AX foci remaining in the AZD2014 treated cells was significantly greater than in control cells. In GBAM1 cells (Fig. 5B, right panel), no difference in foci levels was detected between control (vehicle) and AZD2014 treated cells at 1 hour or 6 hours after irradiation. However, at 24 hours, the number of radiation-induced γH2AX foci remaining in the AZD2014 treated cells was significantly greater than in control cells. These data suggest that AZD2014-induced GSC radiosensitization involves an inhibition of the repair of radiation-induced DNA DSBs.

To determine whether the enhancement of tumor cell radiosensitivity measured in vitro extends to an orthotopic model, GBMJ1 cells were used to initiate intracerebral xenografts in nude mice, as previously described.³⁰ Initially, the ability of AZD2014 to inhibit mTOR activity in GBMJ1 orthotopic xenografts was tested. At the onset of tumor-induced morbidity, AZD2014 (50 mg/kg) was delivered by oral gavage; brains were collected 2 hours later and subjected to immunofluorescent histochemical analysis. Sections were obtained from nonnecrotic portions of the tumor. Human-specific nestin antibody was used to verify the identity of tumor cells. As shown in Fig. 6, total as well as phosphorylated AKT and 4E-BP1 were clearly detectable in brain tumor xenografts from control mice. Whereas AZD2014 treatment had no apparent effect on the expression of total 4E-BP1 and Akt, in treated mice, there was a significant reduction in the levels of p-4EBP1and p-AKT, indicative of mTORC1 and mTORC2 inhibition, respectively. These data indicate that AZD2014 penetrates the tumor blood-brain barrier at sufficient levels to inhibit mTOR kinase.

Fig. 6.

Effects of AZD2014 on mTOR activity in orthotopic xenografts initiated from CD133+ GBMJ1 cells. At the onset of morbidity (mean, 52 days), mice bearing orthotopic xenografts were exposed to vehicle or AZD2014 (50 mg/kg, oral gavage) and collected 2 hours later for immunohistochemical evaluation: total 4EBP1 (green), p4E-BP1 t37/46 (green), AKT (green), pAKT s473 (green), nestin to identify human tumor cells (red), and nuclei (blue), 40x magnification.

Because of its ability to inhibit mTOR activity in the GBMJ1 orthotopic xenografts, the effect of AZD2014 on the radioresponse of these brain tumors was determined. For this study, GBMJ1 cells were engineered to express β-luciferase, and bioluminescent imaging (BLI) used to establish tumor presence in each mouse and for randomization into the treatment groups.³⁴ Specifically, at 12 days after intracerebral implant when bioluminescence was clearly detectable in all mice indicative of tumor, mice were randomized according to BLI signal into 4 groups: vehicle (control), radiation (12 Gy), AZD2014 (50 mg/kg), and AZD2014 plus radiation. AZD2014 was delivered once a day (50 mg/kg, oral gavage) for 3 days with the tumor locally irradiated (12 Gy) immediately after each drug treatment. Mice were followed until the initial onset of morbidity. As shown in Fig. 7, whereas AZD2014 treatment alone had no effect on mouse survival as compared with control treatment (P = .63), IR treatment alone resulted in a significant increase in survival (P = .03). The survival of mice receiving the combination protocol (AZD2014 + IR) was significantly increased as compared with control (P = .014) and importantly as compared with IR alone (P = .03). For control, AZD2014, IR and AZD2014 + IR treatments the median survival times were 53, 56 (+3), 62 (+9) and 82 (+29) days, respectively, indicating that the combination protocol resulted in a greater than additive increase in survival. Thus, these data are consistent with AZD2014 enhancing the radiosensitivity of GBMJ1 orthotopic xenografts.

Fig. 7.

Influence of AZD2014 on the radioresponse of orthotopic xenografts initiated from CD133+ GBMJ1 cells. At 12 days after orthotopic implant, mice were randomized and treatment initiated as described. Mice were followed until the onset of morbidity. Kaplan–Meier survival curves were generated with log-rank analysis for comparison.

Discussion

In the study presented here, radiation-induced GSC death was defined by clonogenic survival analysis, the gold standard for evaluating intrinsic radiosensitivity. When in EGF/FGF supplemented neural basal medium, which maintains their stem-like properties, GSCs do not attach to standard tissue culture plastic. However, when plates are coated with poly-L-lysine, GSCs grow as adherent colonies and, in contrast to growth in medium containing FBS, preserve their stem-like cell properties including CD133 expression.²⁸ Thus, this method allows for defining radiosensitivity according to clonogenic analysis of the GSC phenotype. Whereas the identification and isolation of GSCs has been primarily based on the stem cell associated protein CD133,²⁹ not all GSCs express CD133⁴³; other markers have been used to isolate GSCs from neurospheres generated from human GBM surgical specimens. Along these lines, Son et al reported that stage-specific embryonic antigen 1 (SSEA-1/CD15) could be used to isolate GSCs that meet the criteria for tumor stem-like cells.²⁷ As shown here, the radiosensitivity of the CD15 expressing GSC line 0923 was similar to that of the 3 CD133+ GSC lines. Whereas AZD2014 treatment alone had little effect on GSC survival, this mTOR inhibitor enhanced the intrinsic radiosensitivity of GSCs expressing either CD133 or CD15. These results suggest a general applicability of AZD2014 as a radiosensitizer of GSCs.

Given the number of mTORC1 and mTORC2 substrates, whether the radiosensitization induced by AZD2014 is initiated via a single downstream event or whether multiple mTOR substrates are involved remains to be determined. However, based on analysis of γH2AX foci induction and dispersion, it appears that AZD2014-mediated radiosensitization is the result of an inhibition of DNA double strand break repair. Furthermore, radiosensitization was induced when AZD2014 was added after irradiation, consistent with an effect on some aspect of the DNA repair process. Although the direct interaction of mTOR or 1 of its substrates with a component of the DNA repair machinery cannot be eliminated, the role of mTOR as a critical regulator of gene translation in response to a variety of stress and environmental signals may provide a mechanistic basis for the inhibition of DSB repair in AZD2014-treated cells. Along these lines, as for other competitive mTOR inhibitors, AZD2014 effectively inhibits the phosphorylation of 4E-BP1 (Fig. 1), which prevents its release of eIF4E and thus reduces the level of eIF4E available for cap-dependent translation.¹⁸ A recent study using microarray analysis of polysome-bound RNA showed that after exposure to another competitive mTOR inhibitor PP242, among the genes whose translation was significantly suppressed were a number coding for DNA repair proteins.²³ Moreover, in our recent study using RIP-Chip analysis, irradiation was found to increase eIF4E binding to over 1 000 unique transcripts, a significant number of which were associated with the functional category of DNA Replication, Recombination and Repair.⁴ Thus, the AZD2014-mediated inhibition of gene translation may play a role in its radiosensitizing actions.

Investigations aimed at developing radiosensitizing agents for GBM have traditionally focused on long-established glioma cell lines. However, the biology of such cell lines, as reflected by genetic abnormalities, gene expression, and orthotopic growth patterns, has little in common with GBM in situ.⁴⁴ With respect to a more biologically accurate model system, data now suggest that GBMs are driven and maintained by a subpopulation of clonogenic cells referred to as glioma stem-like cells (GSCs). In addition to in vitro properties in common with normal neural stem cells, GSCs grown as brain tumor xenografts replicate the invasive growth patterns of GBMs in situ as well as the genotype and gene expression patterns of the GBM from which they originated. Given that GSC initiated orthotopic xenografts simulate GBM biology, it would appear that they should also provide a relevant model system for investigating molecularly targeted radiosensitizers. Accordingly, the potential of AZD2014 as a radiosensitizing agent applicable to GBMs was further evaluated using a GSC-initiated xenograft. As shown, AZD2014 penetrates the blood-brain barrier to effectively inhibit both mTORC1 and mTORC2 activities suggestive of its clinical relevance in the treatment of CNS malignancies. Moreover, the combination of AZD2014 and radiation significantly prolonged the survival of mice bearing a GSC brain tumor xenograft. It should be noted that this prolongation of survival was attained when AZD2014 was delivered for only 3 days. AZD2014 is currently under evaluation in a phase I clinical trial as a single agent;²⁴ the data presented here suggest that this competitive mTOR inhibitor may be an effective radiosensitizing agent applicable to GBM therapy.

Funding

Division of Basic Sciences, National Cancer Institute (Z1A BC011372, Z1A BC011373).

Conflict of interest statement. All authors have seen and agreed with the contents of the manuscript. The authors have no conflicts of interest related to this work and confirm the originality of this study.