Mammalian target of rapamycin inhibitors, temsirolimus and torin 1, attenuate stemness-associated properties and expression of mesenchymal markers promoted by phorbol-myristate-acetate and oncostatin-M in glioblastoma cells

Abstract

The phosphatidylinositol 3-kinase/Akt/mammalian target of rapamycin signaling pathway is crucial for tumor survival, proliferation, and progression, making it an attractive target for therapeutic intervention. In glioblastoma, activated mammalian target of rapamycin promotes invasive phenotype and correlates with poor patient survival. A wide range of mammalian target of rapamycin inhibitors are currently being evaluated for cytotoxicity and anti-proliferative activity in various tumor types but are not explored sufficiently for controlling tumor invasion and recurrence. We recently reported that mammalian target of rapamycin inhibitors—rapamycin, temsirolimus, torin 1, and PP242—suppressed invasion and migration promoted by tumor necrosis factor-alpha and phorbol-myristate-acetate in glioblastoma cells. As aggressive invasion and migration of tumors are associated with mesenchymal and stem-like cell properties, this study aimed to examine the effect of mammalian target of rapamycin inhibitors on these features in glioblastoma cells. We demonstrate that temsirolimus and torin 1 effectively reduced the constitutive as well as phorbol-myristate-acetate/oncostatin-M-induced expression of mesenchymal markers (fibronectin, vimentin, and YKL40) and neural stem cell markers (Sox2, Oct4, nestin, and mushashi1). The inhibitors significantly abrogated the neurosphere-forming capacity induced by phorbol-myristate-acetate and oncostatin-M. Furthermore, we demonstrate that the drugs dephosphorylated signal transducer and activator transcription factor 3, a major regulator of mesenchymal and neural stem cell markers implicating the role of signal transducer and activator transcription factor 3 in the inhibitory action of these drugs. The findings demonstrate the potential of mammalian target of rapamycin inhibitors as “stemness-inhibiting drugs” and a promising therapeutic approach to target glioma stem cells.

Introduction

Glioblastoma (GBM) is the most lethal and common central nervous system (CNS) tumor in adults. Despite the advancements in multi-modality treatment, the overall prognosis of GBM patients remains dismal with median survival period of less than 15 months.¹ The standard line of treatment involves surgical removal of tumor followed by radiation and chemotherapy with temozolomide (TMZ) as the first choice of drug.² However, in most cases, treatment with TMZ results in intrinsic or acquired resistance and failure of therapy³ with only 11% of the patients being free from tumor progression after 2 years of treatment.⁴ The problem is further complicated as TMZ enhances production of inflammatory mediators such as CXCL2, CXCL3, and interleukin-8 (IL-8).⁵ The other factors that contribute to resistance in GBM include (a) high degree of heterogeneity of the tumor and (b) the microenvironment comprising of tumor-associated parenchymal cells such as vascular cells, microglia, and peripheral immune cells, and (c) cancer stem cells (CSC) that actively promote tumor progression.⁶ The CSC or glioma initiating cells/glioma stem cells (GIC/GSC) constitute a small population of a tumor, retain stem-like cell features including self-renewal capacity that is crucial for tumorigenicity,^7,8 and confer resistance to radiotherapy and chemotherapy, which leads to recurrence and cancer relapse.^9–11 Therefore, GSCs represent a promising target for anti-GBM therapies.

Phosphatidylinositol 3-kinase (PI3K)/Akt/mammalian target of rapamycin (mTOR) signaling pathway plays a crucial role in survival, proliferation, and progression of tumors in many cancers, including GBM.^12–14 The pathway is negatively regulated by tumor suppressor-phosphatase and tensin homologue (PTEN).¹⁵ In malignant gliomas, PTEN is expressed at a low level that is caused by mutation or loss of heterozygosity and results in highly activated PI3K/Akt/mTOR pathway.^16–19 The epidermal growth factor receptor (EGFR) is frequently overexpressed in GBM and contributes to activation of PI3K/Akt/mTOR pathway. Activation of mTOR facilitates invasive phenotype and tumor dissemination throughout the whole brain^20,21 and is associated with poor patient survival.²² The mTOR protein kinase functions by formation of two distinct macromolecular complexes that regulate different functions of the mTOR network. The mTOR complex 1 (mTORC1) consists of mTOR, raptor, and mLST8 and controls cell growth, progression, and metabolism by regulating ribosomal synthesis, translation mechanism, and autophagy. It acts through effectors such as ribosomal protein S6 kinase beta-1 (S6K1) and eukaryotic initiation factor 4E-binding protein 1 (4EBP1). The mTOR complex 2 (mTORC2) comprises of mTOR, rictor, Sin-1, and mLST8 and regulates the actin cytoskeletal functioning (RhoA, Rac1) through protein kinase C alpha (PKC-α) and pro-survival protein kinase B (Akt/PKB).²³

The mTORC1 inhibitors, everolimus and temsirolimus (TEM), are approved drugs for the treatment of GBM. The results with these drugs have supported the development of other mTOR kinase inhibitors (TORKinibs) such as torin 1 (TOR) and PP242. These inhibitors are potent, selective small molecule inhibitors that bind to adenosine triphosphate (ATP)-binding site of mTOR molecule and efficiently inhibit mTORC1 as well as mTORC2 complexes.^24,25 Many mTOR inhibitors have been extensively evaluated for their effect on controlling tumor growth and sustenance but are not explored sufficiently for limiting tumor invasion and recurrence. We recently reported that mTOR inhibitors—rapamycin (RAP), TEM, TOR, and PP242—suppressed invasion and migration induced by tumor necrosis factor-alpha (TNFα) and tumor promoter (phorbol-myristate-acetate (PMA)) in human GBM cells.²⁶ Given that invasion and migration are characteristics associated with mesenchymal phenotype and stemness-like features in GBM, the aim of this study was to investigate the potential of mTOR inhibitors in affecting the mesenchymal markers and stemness. We demonstrate that TEM and TOR inhibit mesenchymal markers and stem cell–like traits induced by PMA and oncostatin-M (OSM). Based on our findings, we propose that mTOR inhibitors—TEM and TOR—have the potential to be explored as “stemness-inhibiting drugs” and as a part of therapeutic approach in the treatment of GBM.

Materials and methods

Cell lines

The Human Glioblastoma Cell Line (LN-18) was procured from American Type Culture Collection (ATCC). Dulbecco’s modified eagle’s medium (DMEM) with 4 mM l-glutamine, 1.5 g/L sodium bicarbonate, and 4.5 g/L glucose supplemented with 5% heat-inactivated fetal calf serum was used to maintain the cells in a humidified incubator at 37°C with 5% CO₂. Cells were dislodged using trypsin (0.125%)–ethylenediaminetetraacetic acid (EDTA; 0.02%) solution.

The GBM tumor samples were obtained from neurosurgeries performed at Sasoon Hospital, D. Y. Patil Medical College, Hospital & Research Centre, and Inamdar Hospital, Pune. Informed consent was obtained from patients for procurement of tumor samples in accordance with the protocol approved by the Institutional Ethics Committee of National Centre for Cell Science (NCCS). Cell Cultures were obtained by processing tissues using Accutase (Himedia) and Zymefree (Himedia) to obtain adherent cell cultures. Cell cultures up to 30 passages were used in the study. Cells were maintained and subcultured in DMEM with 4 mM l-glutamine, 1.5 g/L sodium bicarbonate, and 4.5 g/L glucose, supplemented with 10% heat-inactivated fetal calf serum in a humidified incubator at 37°C with 5% CO₂.

Treatment

GBM cells were seeded in complete medium and incubated for 24 h. Cells were treated with RAP (10 µM; Calbiochem) or its analog TEM (5 µM; Santa Cruz Biotechnology), which effectively inhibit mTORC1 and, on prolonged exposure, also inhibit mTORC2. TOR (100 nM) and PP242 (100 nM) were purchased from Tocris Bioscience. These compounds are small molecule ATP mimetic site-binding inhibitors of both the mTOR complexes. OSM (50 ng/mL; R&D systems) and phorbol 12-myristate 13-acetate (PMA, 100 ng/mL; Sigma-Aldrich) were used for promoting aggressive phenotypes in cells.

Western blotting

LN-18 and G1 cells were harvested and lysed using radioimmunoprecipitation assay (RIPA) lysis buffer (120 mM NaCl, 1.0% Triton X-100, 20 mM Tris–HCl, pH 7.5, 100% glycerol, 2 mM EDTA, and protease inhibitor cocktail; Roche). Protein concentration in the lysates was estimated using Bradford method (Bio-Rad). Total protein (35 µg) of each sample was electrophoresed on 10% sodium dodecyl sulfate (SDS) polyacrylamide gels at a constant voltage of 65 V and electro-blotted onto polyvinylidene difluoride (PVDF) membrane (Millipore) using Bio-Rad mini-blot module (120 mA per gel, 3 h, 4°C). After blocking with 5% bovine serum albumin (BSA) in Tris-buffered saline with Tween 20 (TBS-T) for 1 h at room temperature, the blots were probed with specific primary antibodies for 2 h at room temperature or overnight at 4°C. Phospho-STAT3 (Y705), phospho-STAT3 (S727), and total signal transducer and activator transcription factor 3 (STAT3) were from Cell Signaling Technology and used at 1:2000 dilution. The other antibodies used in the study were vimentin (1:1000; Thermo Fisher Scientific), fibronectin (1:1000; Sigma-Aldrich), YKL40 (1:1000; Santa Cruz Biotechnology), Oct4 (1:1000; Chemicon), and Sox2 (1:1000; R&D systems). The blots were probed with horseradish peroxidase (HRP)-labeled anti-rabbit or anti-mouse or anti-goat secondary antibodies (1:8000; Bio-Rad) for 1 h. The bands were visualized by chemiluminescence using Super Signal West Femto Maximum Sensitivity Substrate (Pierce), and images were acquired on Amersham Imager 600 instrument (GE Healthcare). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH, 1:10,000; Sigma-Aldrich) was used as loading control. For analysis of relative intensities of protein, densitometry was performed using the ImageJ® software.

Immunofluorescence

LN-18 and G1 cells (5 × 10⁴ cells) grown on coverslips (Thermo Esco) in complete medium for 24 h were treated with TEM and TOR alone or in combination with OSM or PMA for 24 h. Cells were washed twice using cold 1× phosphate-buffered saline (PBS) and fixed in 3.7% paraformaldehyde for 10 min at room temperature. Cells were then permeabilized for 10 min with freshly prepared 0.2% Triton X-100 in PBS, blocked using 1% BSA in PBS for 1 h. The cells were incubated with optimal dilutions (1:200) of primary antibodies—vimentin (Thermo Fisher Scientific), fibronectin (Sigma), YKL40 (Santa Cruz Biotechnology), nestin (Chemicon), Sox2 (R&D systems), and musashi1 (R&D systems) for 2 h—followed by Cy3 conjugated anti-rabbit or anti-mouse or anti-goat secondary antibodies (1:250; Chemicon) for 1 h. 1× PBS was used to wash the cells thrice between each step of incubation. Nuclear staining was done using 4′,6-diamidino-2-phenylindole (DAPI; 0.5 µg/mL) for 30 min, and coverslips were mounted using mounting media. Images were acquired using LEICA confocal microscope (Laser Scanning Microscope-510). Negative controls included cells processed similarly during all steps except for the absence of primary antibodies.

Neurosphere assay

G1 cells (3 × 10⁵) were pre-treated with inhibitors followed by PMA/OSM for 6 h, and 1 × 10³ cells were seeded onto low attachment six-well plates (Corning) in serum-free neurosphere assay medium. The neurosphere assay medium was prepared from a combination of DMEM (4 mM l-glutamine, 1.5 g/L sodium bicarbonate, and 4.5 g/L glucose) and Ham F-12 (HF12) in 1:1 ratio and supplemented with B27 supplement (1× concentration; Gibco), basic fibroblast growth factor (bFGF; 20 ng/mL), and epidermal growth factor (EGF; 20 ng/mL; Peprotech). Cultures were incubated in a humidified incubator with 5% CO₂ at 37°C for 12 days with change of media containing respective treatment after every 3 days. The images were acquired using phase-contrast microscope (Nikon).

Statistical analysis

Quantitative data were represented as mean ± standard error of the mean (SEM) for various groups. The statistical significance between groups was analyzed using an unpaired Student’s t-test to obtain p value. p < 0.05 was considered significant.

Results

mTOR inhibitors downregulate the expression of mesenchymal markers

Initial experiments were conducted to assess the potential of mTOR inhibitors to modulate mesenchymal markers expressed in human GBM cell line—LN-18. The cells were treated with mTORC1 inhibitors—RAP and TEM—and TORKinibs—PP242 and TOR—and the cell lysates were analyzed by western blotting for mesenchymal markers. As depicted in Figure 1(a), LN-18 cells expressed fibronectin abundantly compared to vimentin and YKL40. PMA is a tumor promoter and enhances aggressive features in tumor cells through PKC activation.^27,28 The effect of the mTOR inhibitors was examined in GBM cells stimulated with PMA. The expression of fibronectin and vimentin was robustly elevated by PMA by 41% and 24%, respectively, while YKL40 was increased by 18%. The mTOR inhibitors were effective in decreasing the level of mesenchymal proteins induced by PMA in the order of PP242 > TOR > TEM > RAP. For immunostaining experiments, TOR and TEM were used in the presence of PMA. We observed immunoreactivity to the three mesenchymal proteins by immunofluorescence staining and found higher intensity of staining in cells treated with PMA. The inhibitors decreased the basal as well as PMA-induced fluorescence intensity of the mesenchymal markers in GBM cells (Figure 1(b)).

Figure 1.

mTOR inhibitors downregulate the expression of mesenchymal markers. (a) LN-18 cells were treated with rapamycin (RAP; 10 µM), temsirolimus (TEM; 5 µM), torin 1 (TOR; 100 nM), and PP242 (100 nM) for 24 h in combination with PMA (100 ng/mL) for 24 h before termination of time point, and immunoblotting was performed for mesenchymal markers—fibronectin, vimentin, and YKL40. GAPDH was used as loading control. Data are representative images of three independent experiments. Images represent fold change in mesenchymal markers—fibronectin, vimentin, and YKL40—with respect to GAPDH and are relative to untreated samples through densitometry analysis by ImageJ® and (b) LN-18 cells seeded on coverslips were exposed to TEM (5 µM) and TOR (100 nM) for 24 h alone and in combination with PMA (100 ng/mL) for 24 h before termination of time point and were processed for immunofluorescence staining of mesenchymal markers—fibronectin, vimentin, and YKL40. The merged images depict staining for the markers and nuclear staining as blue (DAPI). Representative images of two independent experiments. Scale: 20 µm.

[Figure omitted. See PDF]

mTOR inhibitors modulate STAT3 activity in LN-18 cells

The STAT3 is a member of the STAT family of proteins and mediates cytokine signaling and nuclear transcription. STAT3 is hyperactivated in malignant gliomas and plays a crucial role in driving these tumor cells toward mesenchymal phenotype.²⁹ OSM exerts its biological activities through the Janus kinase (JAK)/STAT3 pathway that is initiated by binding to its heterodimer receptor, gp130, and leukemia inhibitory factor (LIFR)/oncostatin M receptor (OSMR), and followed by phosphorylation of STAT3 at the Tyr-705 residue and its dimerization and translocation to nucleus.³⁰ We recently reported that OSMR is upregulated in GBM cells, and the expression correlates negatively with patient survival. We also identified the role of STAT3 in inducing mesenchymal markers in glioma cells stimulated with OSM.³¹ On these lines, we speculated whether the effect of mTOR inhibitors in regulating the mesenchymal markers might involve modulation of STAT3 activity. We observed that LN-18 cells treated with TEM, TOR, and PP242 exhibited a significant drop (>50%) in phosphorylated STAT3 (tyrosine 705) while the total STAT3 level was unaltered (Figure 2(a)). PMA enhanced the expression of pSTAT3 (Y705), and the effect was reversed by the inhibitors in the order TOR > PP242 > TEM. RAP affected the basal and PMA-induced pSTAT3 marginally (Figure 2(b)).

Figure 2.

mTOR inhibitors modulate STAT3 activity in LN-18 cells. LN-18 cells were treated with rapamycin (RAP; 10 µM), temsirolimus (TEM; 5 µM), torin 1 (TOR; 100 nM), and PP242 (100 nM); (a) inhibitors alone and (b) in combination with PMA (100 ng/mL) for 24 h before termination of time point, and immunoblotting was performed for phospho-STAT3 (Y705) and total STAT3. GAPDH was used as loading control. Data are representative images of three independent experiments. Images show fold change of treated phospho-STAT3 (Y705) with respect to total STAT3 and relative to untreated samples by densitometry through ImageJ analysis.

[Figure omitted. See PDF]

TEM and TOR modulate PMA-induced-STAT3 activity and vimentin in GBM cell line—G1

Since LN-18 cell line is an established cell line, we chose to conduct experiments with cells in lower passages of GBM cell line (G1) that was developed in our laboratory. The characteristic features of the cell line have been described earlier.³² G1 cells were treated with TEM and TOR in the presence of PMA, and the expression of pSTAT (S727), pSTAT3 (Y705), and vimentin was assessed. We found that G1 cells expressed detectable levels of both phosphorylated forms, pSTAT (S727) and pSTAT3 (Y705). The inhibitors reduced the constitutive expression of the phosphorylated STAT3 forms with the effect being more profound for pSTAT (S727). In context with mTOR pathway, a recent study reported a correlation between a reduction in Ser727 pSTAT3 and decreased mTOR-ser2448.^33,34 As depicted in Figure 3(a), exposure of cells to PMA resulted in greater increase of phosphorylated STAT3 Y705 (36%) compared to STAT3 S727 (17%). This observation is not surprising since stimulation with cytokine results in activation of STAT3 with phosphorylation of Y705 that is important for cooperation with S727 for transcriptional activity.³³ The expression of vimentin in response to PMA in G1 cells was not as robust as in LN-18 cells. The difference in the two cell lines may be related to the basal level of vimentin in the two cell lines. LN-18 cell line has gone through hundreds of passages and exhibits low vimentin expression that was elevated on stimulation with PMA, functioning as tumor promoter. In contrast, G1 cells were in lower passages and retained the aggressive characteristics with high basal level of mesenchymal marker such as vimentin and hence were not responsive to PMA to the same extent. Importantly, despite this disparity, TEM and TOR reduced the vimentin expression in both the cell lines (Figure 3(b)). Immunostaining revealed that PMA-treated G1 cells stained intensely for vimentin, and the inhibitors caused a significant decrease in the fluorescence (Figure 3(c)). These results clearly indicated that the mTOR inhibitors effectively reduced STAT3 activity and decreased the expression of mesenchymal markers in GBM cells.

Figure 3.

TEM and TOR modulate PMA-induced STAT3 activity and vimentin in G1 cells. G1 cells were treated with temsirolimus (TEM; 5 µM) and torin 1 (TOR; 100 nM) alone and in combination with PMA (100 ng/mL) for 24 h before termination of time point. Immunoblotting was performed for (a) phospho-STAT3 (Y705), phospho-STAT3 (S727), and total STAT3 (b) mesenchymal marker vimentin. GAPDH was used as loading control. Data are representative images of two independent experiments. Images present fold change of treated phospho-STAT3 (Y705) and phospho-STAT3 (S727) relative to untreated samples normalized to total STAT3 and vimentin with respect to GAPDH by densitometry analysis using ImageJ software, and (c) immunofluorescence staining of mesenchymal marker—vimentin. The merged images depict staining for vimentin and nuclear staining as blue (DAPI). Representative images of two independent experiments. Scale: 20 µm.

[Figure omitted. See PDF]

TEM and TOR attenuate PMA-induced neural stem cell markers in G1 cells

GSC/GIC contribute to chemo- and radio-resistance that is followed by tumor recurrence in malignant gliomas. Recent studies suggest that GBM tumors with mesenchymal phenotype that are highly invasive and vascular also express stem cell progenitor marker—CD44.³⁵ Another study reported that primary GBM tumor cells express molecular signatures of mesenchymal (MES) and neural stem cells, and differentiation toward either of the lineages could be a decisive factor for treatment regime.³⁶ It was therefore intriguing to investigate the impact of mTOR inhibitors on “stemness-related traits” in GBM cells. The GBM cell line (G1) was ideal for these experiments as the cells exhibit neural stem cell markers—Oct4, Sox2, nestin, and musashi1—and also display resistance to a panel of DNA alkylating drugs.³² Western blotting analysis revealed that PMA upregulated the expression of Sox2 (33%) and Oct4 (21%) compared to control cells. Interestingly, while the basal level of both these proteins was reduced effectively by TOR but not TEM, both the inhibitors reduced PMA-induced expression of Oct4 and Sox2 significantly (Figure 4(a)). We also observed by immunostaining that basal as well as PMA-induced Sox2 was attenuated by the inhibitors. Similar results were seen for neural stem cell markers—nestin and musashi1 (Figure 4(b)).

Figure 4.

TEM and TOR attenuate PMA-induced neural stem cell markers in G1 cells. G1 cells were treated with temsirolimus (TEM; 5 µM) and torin 1 (TOR; 100 nM) alone and in combination with PMA (100 ng/mL) for 24 h before termination of time point. (a) Immunoblotting was performed for stem cell markers Oct4 and Sox2. GAPDH was used as loading control. Data are representative images of two independent experiments. Images represent fold change in stem cell markers Oct4 and Sox2 with respect to GAPDH and relative to untreated samples through densitometry analysis by ImageJ software and (b) immunofluorescence staining of stem cell markers: Sox2, musashi1, and nestin. The merged images depict staining of markers and nuclear staining as blue (DAPI). Representative images of two independent experiments. Scale: 20 µm.

[Figure omitted. See PDF]

TEM and TOR attenuate OSM-induced STAT3 activity and neural stem cell markers in G1 cells

In our previous study, we demonstrated that OSM enhanced the mesenchymal phenotype and stemness markers in G1 cells via STAT3 signaling.³¹ In order to gain better understanding of the effect of mTOR inhibitors on mesenchymal phenotype and stemness-associated molecules, we determined the expression of mesenchymal markers and stemness-related proteins in G1 cells stimulated with OSM. As shown in Figure 5(a), OSM highly activated STAT3 with increase in pSTAT3 (S727) and pSTAT3 (Y705) levels by 134% and 211%, respectively, and the effect was significantly inhibited by TEM and TOR. The expression of vimentin was reduced on exposure to OSM in the presence of mTOR inhibitors as detected by western blotting and immunostaining (Figure 5(b)–(c)).

Figure 5.

TEM and TOR attenuate OSM-induced STAT3 activity and vimentin in G1 cells. G1 cells were treated with temsirolimus (TEM; 5 µM) and torin 1 (TOR; 100 nM) alone and in combination with OSM (50 ng/mL) for 24 h before termination of time point. Immunoblotting was performed for (a) phospho-STAT3 (Y705), phospho-STAT3 (S727), and total STAT3; and (b) mesenchymal marker vimentin. GAPDH was used as loading control. Data are representative images of two independent experiments. Images show fold change for phospho-STAT3 (Y705), phospho-STAT3 (S727), and vimentin for treated in comparison to total STAT3 and GAPDH, and relative to untreated samples by densitometry through ImageJ analysis and (c) immunofluorescence staining of mesenchymal marker vimentin. The merged images depict staining of markers and nuclear staining as blue (DAPI). Representative images of two independent experiments. Scale: 20 µm.

[Figure omitted. See PDF]

STAT3 has an important role in neuronal survival and determination of neural stem cell fate during development of nervous system.³⁷ Since mTOR inhibitors dephosphorylated pSTAT3 in G1 cells, it was of interest to examine whether this effect translated to modulation of stemness-like features in G1 cells. As presented in Figure 6(a), western blotting data demonstrated that OSM enhanced Sox2 and Oct4 levels by 31% and 52%, respectively, compared to untreated cells, and TEM and TOR reversed the effect significantly. The results for Sox2 were confirmed by immunostaining. In addition, the expression of neural stem cell markers—musashi1 and nestin—was upregulated by OSM, and the mTOR inhibitors attenuated the expression of both these proteins in G1 cells (Figure 6(b)). These results suggest that TEM and TOR are efficient inhibitors of stemness-related markers in GBM cells.

Figure 6.

TEM and TOR attenuate OSM-induced neural stem cell markers in G1 cells. G1 cells were treated with temsirolimus (TEM; 5 µM) and torin 1 (TOR; 100 nM) alone and in combination with OSM (50 ng/mL) for 24 h before termination of time point. (a) Immunoblotting was performed for stem cell markers Oct4 and Sox2. GAPDH was used as loading control. Data are representative images of two independent experiments. Images show fold change in stem cell markers, Sox2 and Oct4, for treated in comparison to GAPDH, and relative to untreated samples by densitometry through ImageJ analysis and (b) immunofluorescence staining of stem cell markers: Sox2, musashi1, and nestin. The merged images depict staining of markers and nuclear staining as blue (DAPI). Representative images of two independent experiments. Scale: 20 µm.

[Figure omitted. See PDF]

mTOR inhibitors reverse the neurosphere-forming capacity promoted by PMA and OSM

To further confirm the effect of mTOR inhibitors as effective “stemness-inhibiting drugs,” functional assay for stemness was performed using the neurosphere formation assay. The assay essentially measures the self-renewal capacity of neural stem cells. As depicted in Figure 7, G1 cells exhibited the potential to form neurospheres that was promoted by OSM and PMA. Interestingly, the neurospheres formed in response to PMA were larger while OSM-treated cells formed higher number of neurospheres compared to untreated controls. Importantly, TEM and TOR significantly reduced the neurosphere-forming potential of G1 cells induced by PMA as well as OSM. Collectively, these results suggest that regardless of the stimulators, TEM and TOR were effective in reducing the stemness-associated traits in GBM cells.

Figure 7.

mTOR inhibitors reverse the neurosphere formation promoted by PMA and OSM. G1 cells treated with PMA (100 ng/mL) or OSM (50 ng/mL) alone or in combination with inhibitors—temsirolimus (TEM; 5 µM) or torin 1 (TOR; 100 nM)—for 24 h were used to perform neurosphere formation assay. Representative images of neurosphere formation in G1 cells (magnification 20×). Images were analyzed using ImageJ software. Graphs represent neurosphere number and neurosphere size (area in µM²) ± SEM of two similar experiments performed in duplicates. *p < 0.05 untreated versus PMA/OSM treated.

[Figure omitted. See PDF]

Discussion

High-grade glioma tumors contain GSCs that are characterized by aggressive invasion, mesenchymal phenotype, and stem cell–like features, thus making GSCs an attractive target for designing better and efficient strategies for improved anti-GBM therapy. While mTOR inhibitors are being clinically evaluated and are in various stages of trials,^38,39 their application to target CSCs has not been sufficiently explored. This is the first report to demonstrate that mTOR inhibitors—TEM and TOR—repress stem-like cell properties in GBM cells. The results are significant as the inhibitors were effective even when these features and mesenchymal properties were enhanced in cells by PMA and OSM. Collectively, the data reinforce the potential of these drugs to target cancer cells that are characterized by chemoresistance and recurrence. Our findings also implicate the involvement of STAT3 in the action of these drugs as “stemness-inhibiting drugs.”

The epithelial mesenchymal transition (EMT) is a key step in tumor progression related to invasion and metastasis. Studies by Mani et al.⁴⁰ provided evidence for a link between EMT and formation of stem cells. The investigators demonstrated that cells that had undergone EMT shared many traits of stem cells derived from normal or neoplastic populations. The mesenchymal and neural subtypes of GBM express CD133-stem cell marker.^41,42 PMA is a specific agonist of PKC enzymes and is reported to be a potent inducer of EMT-like features in malignant prostate cancer cells.⁴³ PMA induces transition of epithelial breast cancer cell line (MCF-7) to a mesenchymal phenotype and expression of surface markers that are characteristics of CSC population.⁴⁴ On these lines, we addressed whether mTOR inhibitors could modulate mesenchymal markers as well as stem cell–like properties. For this purpose, we performed experiments with GBM cells that exhibit properties of neural stem cells. PMA elevated the expression of stem cell markers and the potential to form neurospheres. This is not surprising considering the role of various PKC isoforms in maintenance and differentiation of neural differentiation.⁴⁵ Loss of function of PKCζ is reported to be sufficient to support neuronal differentiation in proliferating neural stem cells by mechanism involving inhibition of neuronal fate determinant TRIM32.⁴⁶ We have previously demonstrated that mTOR inhibitors, RAP, TEM, TOR, and PP242, reduced PKC activity in GBM cells.²⁶ In this context, we show here that TEM and TOR markedly reduced the expression of stem cell markers, sex-determining region Y-Box (Sox2), octomer-binding transcription factor (Oct4), nestin, and musashi1, and also decreased the capacity of G1 cells to form neurospheres. In addition, we demonstrated that OSM-induced expression of mesenchymal characters and stem-like features were attenuated by the mTOR inhibitors.

Various signaling pathways including WNT, Notch, STAT3, and Sonic hedgehog involved in stem cell regulation and development are activated in GSCs.⁴⁷ Sox2 and Oct4 are members of the panel of “Yamanaka Transcription factors” that is capable of reprogramming of differentiated stem cells⁴⁸ and play a key role in self-renewal, maintenance, and differentiation of embryonic stem cells and tumor cells.^49,50 Sox2 is a master regulator transcription factor, and its expression and regulation is reported in different cancers.⁵¹ Oct4 is a partner of Sox2 and is also regulated by Sox2. Both these transcription factors positively correlate with pathological grades of gliomas.⁵² The transcription factor, STAT3, promotes cancer migration and invasion by promoting pro-oncogenic inflammatory pathways.^53,54 STAT3 is frequently activated in malignant gliomas and is associated with poor prognosis.^55,56 We focused on investigating the role of STAT3 for further experiments for two reasons; first, PTEN/mTOR signaling through STAT3 is important for maintenance of cancer stem-like cells and is therefore a relevant pathway in development of molecules to target CSC for improved cancer treatment,⁵⁶ and second, our earlier study revealed that OSM induced stem-like cell traits via STAT3-mediated signaling in GBM cells.³¹ To this end, our data demonstrated that mTOR inhibitors—TEM and TOR—attenuated stemness-associated markers—Sox2, Oct4, nestin, and musashi1—and also functionally reduced neurosphere formation induced by PMA and OSM in G1 cells. Furthermore, the inhibitors dephosphorylated STAT3 (Y705), thus implicating the involvement of STAT3 in the process. Foshay et al.⁵⁷ demonstrated that STAT3 is important in the production of neural precursor cells and regulates Sox2 promoter leading to Sox2 expression and resulting in the expression of neural stem cell marker—nestin. Nestin is an intermediate filament and is important for cytoskeletal organization, signaling organogenesis, and metabolism.⁵⁸ It is a prognostic marker of malignant gliomas.⁵⁹ It is noteworthy that regardless of the mechanism of action of TEM (an mTORC1 inhibitor) and the TOR (an inhibitor of C1 and C2 complexes), both the inhibitors dephosphorylated STAT3. Although we have no direct evidence, it is tempting to speculate that downregulation of nestin and Oct4 by mTOR inhibitors might be the result of Sox2 regulation mediated by STAT3. The finding that TEM and TOR reduced activation of STAT3 assumes importance as inhibition of STAT3 is reported to overcome TMZ resistance in GBM by downregulating O⁶-methylguanin-DNA-methyltransferase (MGMT) expression.⁶⁰

Although a number of compounds are being evaluated for their potential to inhibit stemness in in vitro and in vivo studies in tumor cells, they have a long way to go to reach the goal as approved drugs. In this scenario, the findings from this study lead us to conclude that mTOR inhibitors, particularly TEM (as it is an FDA (Food and Drug Administration)-approved drug), are promising therapeutic drugs to target GSCs and pave the way for conducting clinical trials for improved treatment of GBM patients. However, further in-depth studies are needed to decipher the precise mechanism of action of TEM and TOR in limiting stemness in GBM. Considering the challenges posed by CSC and the role of PI3K/Akt/mTOR/STAT3 in conferring drug resistance across a spectrum of tumor types, it is important and relevant to view these mTOR inhibitors from a new perspective as “stemness-inhibiting drugs” than just cytotoxic or anti-proliferative drugs. Furthermore, these drugs could also be explored for combination therapy with other drugs that target bulk of tumor.

The authors thank Mr Jagadish CK Mangu for his contribution in preparation of manuscript.

Author contribution

G.C. and K.N. performed the biological experiments. P.S. coordinated the project. P.S. and G.C. analyzed the data and wrote the main manuscript. D.R. and A.C. are neurosurgeons who provided the tumor samples.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The work was funded by intramural funding of the National Centre for Cell Science (NCCS), Pune, India. Goparaju Chandrika and Kumar Natesh are senior research fellows funded by Council of Scientific and Industrial Research (CSIR), New Delhi, India.