J Neurooncol (2015) 121:19–29 DOI 10.1007/s11060-014-1602-3

LABORATORY INVESTIGATION

Tight regulation between cell survival and programmed cell death in GBM stem-like cells by EGFR/GSK3b/PP2A signaling Demirkan B. Gu¨rsel • Matei A. Banu • Nicholas Berry • Roberta Marongiu • Jan-Karl Burkhardt • Keith Kobylarz Michael G. Kaplitt • Shahin Rafii • John A. Boockvar

•

Received: 14 February 2014 / Accepted: 23 August 2014 / Published online: 26 October 2014 Ó Springer Science+Business Media New York 2014

Abstract Malignant gliomas represent one of the most aggressive forms of cancer, displaying high mortality rates and limited treatment options. Specific subpopulations of cells residing in the tumor niche with stem-like characteristics have been postulated to initiate and maintain neoplasticity while resisting conventional therapies. The study presented here aims to define the role of glycogen synthase kinase 3 beta (GSK3b) in patient-derived glioblastoma (GBM) stem-like cell (GSC) proliferation, apoptosis and invasion. To evaluate the potential role of GSK3b in GBM, protein profiles from 68 GBM patients and 20 normal brain samples were analyzed for EGFRmediated PI3kinase/Akt and GSK3b signaling molecules including protein phosphatase 2A (PP2A). To better understand the function of GSK3b in GBM, GSCs were isolated from GBM patient samples. Blocking GSK3b

Electronic supplementary material The online version of this article (doi:10.1007/s11060-014-1602-3) contains supplementary material, which is available to authorized users. D. B. Gu¨rsel  M. A. Banu  N. Berry  J.-K. Burkhardt  J. A. Boockvar (&) Laboratory for Translational Brain Tumor and Stem Cell Research, Department of Neurological Surgery, Weill Cornell Brain Tumor Center, Weill Cornell Medical College, 525 East 68th Street, Box 99, New York, NY 10021, USA e-mail: [email protected]; [email protected] R. Marongiu  M. G. Kaplitt Department of Neurological Surgery, Weill Cornell Brain Tumor Center, Weill Cornell Medical College, New York, NY 10021, USA K. Kobylarz  S. Rafii Department of Genetic Medicine, Howard Hughes Medical Institute, Ansary Stem Cell Institute, Weill Cornell Medical College, New York, NY 10021, USA

phosphorylation at Serine 9 attenuated cell proliferation while concomitantly stimulating apoptosis through activation of Caspase-3 in patient-derived GSCs. Increasing GSK3b protein content resulted in the inhibition of cell proliferation, colony formation and stimulated programmed cell death. Depleting GSK3b in GSCs down regulated PP2A. Furthermore, knocking down PP2A or blocking its activity by okadaic acid inactivated GSK3b by increasing GSK3b phosphorylation at Serine 9. Our data suggests that GSK3b may function as a regulator of apoptosis and tumorigenesis in GSCs. Therapeutic approaches targeting GSK3b in glioblastoma stem-like cells may be a useful addition to our current therapeutic armamentarium. Keywords Glioblastoma  Cancer stem cells  EGFR  Glycogen synthase kinase 3  PP2A  Apoptosis

Introduction The GBM stem-like cell (GSC) hypothesis argues that human gliomas display a cellular hierarchal organization, with a heterogeneous mix of terminally differentiated cells along with a smaller population of progenitor and stem-like cells. Tumor initiating cells are thus proposed to be tumor stem-like cells with highly proliferative and self-renewal characteristics. These GSCs are also capable of differentiating into various lineages, such as astrocytes, oligodendrocytes and primary neurons similar to the normal neural stem cell population (NSCs) [1–5]. Glycogen synthase kinase-3beta (GSK3b) is a cytoplasmic serine/threonine multifunctional protein kinase that was originally identified as the pivotal regulator of glycogen metabolism [6]. Although GSK3b initially

123

20

became an attractive drug target for its role in diabetes [7], Alzheimer’s disease and mood disorders, more recent studies revealed an important role in human cancers [8–11]. Wnt signaling modulates GSK3b phosphorylation at Serine 9, thereby regulating cell growth [12]. Furthermore, activation of ErbB2 down regulates p27 and augments cyclin D1 expression trough PI3kinase/Akt and MAPK signaling, suggesting a potential role of GSK3b in the deregulation of G1-to-S transition of the cell cycle in mammary epithelial cell transformation [13]. GSK3b has been described as an important pro-survival regulator in pancreatic cancer [14] and as a pro-apoptotic factor in colorectal cancer [15]. GSK3b plays an active role in several cellular processes, including cell differentiation [16, 17], ceramide-induced mitochondrial apoptosis [18] and microtubular reorganization and cell cycle control [19]. Little is known about the involvement of GSK3b in glial transformation or about the precise molecular mechanism through which growth factor mediated GSK3b signaling is involved in GSC growth, apoptosis and tumorigenesis. Therefore, the present study aims to investigate the role of GSK3b in GSC survival and apoptosis. We report the regulation of GSK3b activation by several pathways including EGFR-mediated PI3kinase/Akt and protein phosphatase 2A (PP2A) signaling, with direct impact on GSC proliferation, apoptosis and transformation.

Experimental procedures Methods and experimental procedures used for this study have been previously published by our group and are described in detail in the Supplementary section of this manuscript [5, 20].

Results Inhibition of EGFR/PI3Kinase/Akt -mediated signaling modulates GSK3b activity by dephoshorylation of Serine-9 in patient-derived GSCs To determine if the function of GSK3b is dependent on the activation of EGFR and PI3kinase/Akt signaling, LN229 cells overexpressing EGFRwt (Fig. 1a) and GSCs (Fig. 1b–d) were excited exogenously with EGF (20 ng/ml) for 100 , 300 and 1 h. In parallel, to establish if we can reverse the EGF effect, we blocked EGFR tyrosine phosphorylation with AG1478 (10 lM) (Fig. 1b, c) or OSI774 (10 lM) (Fig. 1c). Moreover, we interrupted PI3kinase with LY294002 (20 lM) (Fig. 1b). EGF stimulated the phosphorylation of tyrosine residues at 1,173

123

J Neurooncol (2015) 121:19–29

(Fig. 1a, b), 1,086 (Fig. 1a, c), 1,068 and 1,148 (data not shown). The strongest level of phosphorylation in GSCs was achieved at 30 min after EGF stimulation (Fig. 1b). Furthermore, activation of EGFR at all sites increased the phosphorylation of Akt at Serine 473 and GSK3b at Serine 9 (Fig. 1a–c). However, the EGFR inhibitor OSI774 did not show any significant effect on our patientderived GSCs (Fig. 1c) whereas GSCs treated with the GSK3b inhibitor LiCl increased the phosphorylation of GSK3b at Serine 9 (Fig. 1c). Response to OSI774 seems to be highly patient specific. In addition, using AAV-1, we overexpressed wtPTEN, a known suppressor of PI3kinase, in our patient GSCs and observed that phosphorylation of GSK3b at Serine 9 was under the control of the PI3kinase/Akt pathway (Fig. 1d). Furthermore, AG1478 and LY294002 reversed the effect of EGF stimulation by dephosphorylating EGFR, Akt and GSK3b (Fig. 1b, c). These results suggest that GSCs respond to EGF by phosphorylating EGFR, Akt and GSK3b, while inhibitors of EGFR and PI3kinase can reverse this effect. Active GSK3b stimulates apoptosis via Caspase-3dependent pathway To determine how GSK3b exerts its role in GSCs and to verify if Serine 9 on GSK3b has a functional significance in GSCs, we performed apoptosis and cell proliferation experiments with the following conditions: cells alone, EGFtreated (20 ng/ml), PTEN overexpressing, AG1478-treated (10 lM) and LY294002-treated (20 lM). We found that blocking GSK3b phosphorylation at Serine 9 with either PTEN overexpression, AG1478 or LY294002, resulted in the inhibition of GSC proliferation (Fig. 2a), decreasing Ki67 (Fig. 2b, c) and stimulating apoptosis via Caspase-3 (Fig. 2d, e). LY294002, AG1478 and PTEN overexpression attenuated proliferation by 39 % (p = 0.00024), 18.6 % (p = 0.0029) and 13.2 % (p = 0.05507) respectively compared to untreated cells at day five (Fig. 2a). Higher inhibition values were obtained with LY294002, AG1478 and PTEN, 44.2, 26 and 21.1 % respectively, compared to cells treated with EGF at day five (p = 0.00073) (Fig. 2a). Inhibition of GSK3b phosphorylation at Serine 9 resulted in Caspase-3 increase: AAV-1 PTEN, 12.2 % (SD ± 2.8, p = 0.00065); AG1478, 27.1 % (SD ± 7.4, p = 0.0052) and LY294002, 66.7 % (SD ± 3.6, p = 0.000008) (Fig. 2d, e). Furthermore, treating GSCs with different inhibitors led to Ki67 decrease: AAV-1 PTEN, 54.6 % (SD ± 4.7, p = 0.0005); AG1478, 33.7 % (SD ± 3.4, p = 0.0019) and LY294002, 71.3 % (SD ± 6.6, p = 0.0004) (Fig. 2c, b) and to caspase-3 dependent apoptosis (Fig. 2d, e). To show that the mechanism of cell death is apoptotic, we treated GSCs with Z-VAD-FMK (final concentration: 10 lg/ml), a cell

J Neurooncol (2015) 121:19–29

21

Fig. 1 The regulation of GSK3b via the EGFR/PI3kinase-Akt pathway. Western blots on LN 229 and patient derived GBM stemlike cell lysates after treatment with EGF (E?), EGFR phosphorylation blockers AG1478 and OSI774 and PI3kinase inhibitor LY294002. In panel d, cells were either left untreated (cells column), treated with EGF (E?) or transduced with low and high titer AAVPTEN constructs, and western blots were performed on the protein lysates. Actin is used as a loading control. EGF activates EGFR on tyrosine 1,173 and 1,086, Akt on Serine 473 (a–c). Once

phosphorylated, Akt induces the phosphorylation of GSK3b on Serine 9, inactivating its function (a–d). The effect of EGFRmediated signaling was reversed by interrupting tyrosine and Serine phosphorylation with AG1478 (b, c) and LY294002 (b). An alternative EGFR inhibitor, OSI774, showed no effect. GSCs treated with LiCl, a GSK3b inhibitor, increased phosphorylation at Serine 9 (c). GSCs overexpressing PTEN, tumor suppressor for PI3kinase, showed activated GSK3b with decreased Serine 9 phosphorylation (d)

permeable pan-caspase inhibitor that irreversibly binds to the catalytic site of caspase proteases. Cells treated with the Akt inhibitor LY294002 for 1 and 6 h (final concentration: 20 lM) showed an increase in the positive apoptotic cell index in a time-dependent manner (Fig. 2f). Co-administration of LY294002 with the caspase inhibitor interrupted programmed cell death. Furthermore, CC-3 was blocked, whereas MCL1 expression was increased when cells were co-treated with Z-VAD-FMK and LY294002 (Fig. 2g), suggesting that MCL1 may play an important role in GSK3b induced apoptosis of GSCs. To further demonstrate that

GSK3b plays an important role in GSC apoptosis, we treated cells either with LY294002 or with the GSK3b inhibitor LiCl co-administered with LY294002 (Fig S3). Treating cells with LY294002 stimulated the expression of CC-3. Expression of CC-3 was totally diminished when LiCl was co-administered with LY294002, suggesting that inhibiting GSK3b may interfere with the CC-3 apoptotic pathway. These experiments suggest that the pro-apoptotic activity of GSK3b depends on the phosphorylation status at Serine 9. Dephosphorylation drives GSCs into apoptosis, partly through a Caspase-3 mediated mechanism.

123

22

123

J Neurooncol (2015) 121:19–29

J Neurooncol (2015) 121:19–29 b Fig. 2 Significance of the EGFR/PI3kinase/GSK3b pathway in GSC

proliferation and apoptosis. a Proliferation assay of patient derived stem-like cells under different conditions: untreated (cells), treated with EGF, transduced with AAV-PTEN, treated with AG1478 and treated with LY294002; *p \ 0.01; **p \ 0.001. b, c Representative images of Ki67 immunocytochemistry on patient derived stem-like cells treated with the same conditions as in panel a. Ki67 labeling index is depicted in panel c. d, e Representative immunocytochemistry images and labeling index of cleaved caspase 3 under the same conditions. f Cleaved caspase 3 immunocytochemistry in untreated patient derived stem-like cells, cells treated with LY294002 and LY294002 ? Z-VAD-FMK. g Western blot on cell lysates treated with the same conditions as those in panel f. Actin is used as a loading control. Activated GSK3b attenuated GSC proliferation in all three points, while 20 ng/ml of EGF were sufficient to promote cell growth compared to cells untreated with EGF (a). Inhibition of EGFR and Akt decreased Ki67 levels (b, c) and induced GSC apoptosis by activation of cleaved caspase-3 (d, e) P values were calculated with respect to control (untreated cells) as * = p \ 0.05, ** = p \ 0.001 (a). Cells treated with the Akt inhibitor LY294002 showed a higher apoptotic cell index at 6 h compared to 1 h (F). Co-administration of LY294002 with the caspase inhibitor Z-VAD-FMK interrupted programmed cell death. Furthermore, CC-3 was blocked whereas MCL1 expression was increased when GSCs were co-treated with Z-VAD-FMK and LY294002 (G)

23

demonstrated that increasing the GSK3b protein content resulted in the inhibition of GSC colony formation, irrespective of the EGF feedback loop (Fig. 3c, d). GSK3b regulates early and late apoptosis of GSCs The direct effect of GSK3b signaling on GSC early and late apoptosis was then analyzed by fluorescence-activated cell sorting (FACS) after transient transfection for 48, 72 and 96 h either with empty vector or with a plasmid bearing GSK3b wt (Fig. 4a, b). Small aliquots from each time interval were examined to verify transfection efficiency by western blot and immunocytochemistry with GSK3b and HA antibodies whereas the rest of the samples were fixed in 4 % PFA and subjected to flow cytometry. Results obtained from these experiments indicated that increasing GSK3b wild type stimulated GSC programmed cell death while untransfected or empty vector transfected cells remained unchanged (Fig. 4a, b). These experiments revealed that GSK3b functions as a regulator of apoptosis in GSCs.

Overexpressing wild type GSK3b in GBM stem-like cells inhibits proliferation and colony formation

PP2A promotes GSK3b activity by dephosphorylating GSK3b at serine 9

Given the pro-apoptotic nature of GSK3b, we hypothesized that down regulation of GSK3b in certain patients may promote GSC escape from apoptosis, conferring them a growth advantage. To this end, we transiently overexpressed GSK3b wild type (HA GSK3b wt pcDNA) in GSCs. We verified GSK3b protein levels after 48 h of transfection (Fig. 3a lower panel) using both immunocytochemistry and western blot analysis (data not shown). Following transfection, we counted cells under the following conditions, each in triplicates, over a 5 day period: untransfected, vector only (pcDNA 3.1) and GSK3b wt (HA GSK3b wt pcDNA) (Fig. 3b). We found that the proliferation of GSCs overexpressing GSK3b wt was inhibited by 31.5 % (p = 0.019978) compared to untransfected cells or cells transfected with vector alone (Fig. 3b). To confirm that the phenotypic changes we observed were indeed the result of GSK3b up-regulation rather than transient diluted expression of GSK3b, we monitored the protein expression pattern at day one, three and five (data not shown) and observed cell morphology at day five (Fig. 3a). To substantiate that attenuated GSC proliferation by GSK3b up-regulation also inhibits colony formation in soft agar, we implanted 1.5 9 105 GBM spheres under the following conditions: unstransfected, EGF treated (50 ng/ml), with vector alone-1 (pcDNA3.1), with vector alone-2 (AAV-1 YFP), LiCl (20 mM), AG1478 (10 lM), AAV-PTEN, LY294002 (20 lM) and GSK3b wt (Fig. 3c, d). A 3D soft agar assay noticeably

To explain how GSK3b function is regulated by potential crosstalk with PP2A, we knocked down GSK3b using siGSK3b (Fig. 5a, b) and analyzed PP2A protein expression. Silencing GSK3b resulted in down regulation of PP2A in all three subunits in GSCs (Fig. 5c). In order to determine if PP2A regulates GSK3b activity by modifying its phosphorylation profile at Serine 9, we inhibited PP2A activity with short hairpin RNA (shPP2AB56 b) specific to the B subunit (Fig. 5d). Interrupting PP2A at the B subunit stimulated phosphorylation of GSK3b at Serine 9 while reducing total GSK3b at 72 h post-transduction (Fig. 5d). To confirm this finding we also inhibited PP2A activity with okadaic acid (OA), showing that the inactivation of PP2A B subunit also inhibits the activity of GSK3b by increasing its phosphorylation at Serine 9 (Fig. 5e). This suggests that PP2A promotes GSK3b pro-apoptotic activity by dephosphorylating GSK at Serine 9.

Discussion The tumorigenic characteristics of GBM stem-like cells are believed to be the consequence of genetic alterations, which eventually lead to an aberrant regulation of signaling pathways. Understanding how these factors and effector proteins interact with each other under stress will offer significant insight into their distinct role in the pathogenesis of brain tumors.

123

24

J Neurooncol (2015) 121:19–29

Fig. 3 Inhibition of proliferation and anchorage-independent growth in GSCs overexpressing GSK3b. a Morphology and immunocytochemistry of patient derived stem-like cells transfected with GSK3b pcDNA. b Proliferation assay of cells untransfected or transfected with GSK3b. C. Soft agar colony formation of patient derived stem like cells treated with different conditions. D. Quantitative analysis of colonies at 2 and 3 weeks in soft agar under different conditions. GBM stem-like cells were transfected either with HA GSK3b wt pcDNA or pcDNA as vehicle. Cells were lysed 48 h after transfection to confirm GSK3b protein expression or seeded on cover slips and incubated with HA antibody to demonstrate the expression and the localization of GSK3b with immunostaining (a-lower panel). Cells overexpressing wild type GSK3b had attenuated cell proliferation

while no effect was observed in untransfected cells or GSCs transfected with vector alone (b, p = 0.02). Upper panel a shows the different morphology of cells at day five. To assess the impact of GSK3b on in vitro transformation, GSCs used in proliferation experiments were also utilized for a colony formation assay in soft agar (c) 1 9 105 cells were mixed with 0.3 % agar suspension. The following conditions were used: cells alone, EGF-treated cells, pcDNA3.1 (empty vector for GSK3b wt), AAV-1 YFP (empty vector for wt PTEN), LiCl, AG1478, AAV-PTEN, LY294002 and GSK3b wt. GSK3b wt overexpression inhibited colony formation, while EGF-treated cells, pcDNA3.1, AAV-1 YFP and LiCl GSCs showed an increase not only in number (d) but also in size of formed colonies by week three

To test our hypothesis, we established an in vitro model by generating glioblastoma stem-like cell lines from GBM patients. We corroborated the stemness of the cells with their ability to express stem cell markers (Nestin, Musashi1, Sox-2 and Oct-4), their ability to self-renew, differentiate and their in vivo tumorigenic capacity. To understand the interplay of several proteins with GSK3b, we screened tissues from 68 GBM patients and 20 normal brains with over 30 antibody probes. Our western blot results revealed that 50 % of our screened GBM patients highly expressed the EGFR wt protein and 10 % of these patients were also positive for phosphorylated tyrosine residues. Furthermore, 20 % of the patients in the EGFR wt positive cohort also expressed the mutant form of EGFR (EGFR vIII). Our results are in agreement with Ekstrand et al. [21] which initially suggested an essential role of EGFR in glioma

pathogenesis. Activation of PI3kinase signaling has been implicated in glioma patients with poor prognosis. The average survival of GBM patients with elevated PI3kinase and Akt activity was 11 months while patients with lower levels showed a median survival of 40 months [22]. Our results confirm these previous findings in which mutations or loss of PTEN led to activation of the PI3kinase pathway [23]. We found that approximately 60 % of our patients expressed variable degrees of GSK3b down regulation compared to the normal brain samples. This may suggest a potential role of GSK3b in the etiology of GBM and thereby regulated by EGFR, PI3kinase and PP2A signaling. Our findings regarding down regulation of GSK3b in GBM patients are contrary to the results of Korur et al. [16] and Miyashita et al. [24]. The number of samples examined by both groups (32 and 23 primary gliomas by Korur and

123

J Neurooncol (2015) 121:19–29

25

Fig. 4 The effect of GSK3b on early and late apoptosis in GSCs. a Flow cytometry for apoptosis markers Annexin V and Propidium Iodide of untransfected and GSK3b transfected patient derived stemlike cells. b Quantitative analysis of early and late apoptotic cells by flow cytometry, either for cells transfected with an empty vector or cells transfected with GSK3b pcDNA. GBM stem-like cells were transfected with GSK3b wt pcDNA or pcDNA (empty vector) for 48 h and incubated with FITC Annexin V in a buffer containing

Propidium Iodide (PI) and analyzed by flow cytometry. GSK3b wt overexpressing GSCs (a, b) showed stronger FITC Annexin positive staining than either untransfected or empty vector transfected cells. This data suggests that GSK3b wt overexpression in GBM stem-like cells increases both early and late apoptosis. Similar results were replicated in 5 independent experiments with different patient-derived cell lines

Miyashita respectively) was smaller compared to our study. In addition, Korur et al. reasoned the stronger GSK3b expression could potentially be the result of higher necrosis levels found in GBM tissue with elevated protein degradation. GSK3b levels observed in this study were unlikely to be influenced by protein degradation given that protein profiling on our patient samples was entirely consistent with previous reports for EGFR, PTEN and Akt. If protein degradation directly led to the diminishing GSK3b levels, one would have expected to see similar effects on other proteins as well, which was not the case here. Previous studies have indicated that GSK3b activity was blocked by growth factor stimulation followed by the activation of PI3kinase/Akt signaling in various cell types [25–28]. Pap and Cooper [29] showed that pre-incubation

of PC12 cells with nerve growth factor (NGF) inhibited GSK3b apoptotic activity by 30–40 %. This blockage was reversed by LY294002 and Wortmannin, PI3kinase inhibitors, suggesting that NGF inhibition of GSK3b was driven by PI3kinase/Akt signaling [29]. The pro-apoptotic role of GSK3b in EGFR overexpressing patient-derived GSCs has not been well established. Therefore, in this study, we sought to determine (a) whether GSK3b activity relies at least in part on EGFR-mediated PI3kinase/Akt signaling and (b) the mechanism through which GSK3b functions as an apoptotic regulator in GSCs. Our results demonstrate that EGF binding to its receptors induces its auto-phosphorylation on EGFR tyrosine residues 1,173, 1,086 and 1,068, 1,148. Phosphorylated EGFR activates Akt at Serine 473, which then results in GSK3b phosphorylation at

123

26

J Neurooncol (2015) 121:19–29

Fig. 5 Regulation of GSK3b signaling by PP2A. a Western blot assessing GSK3b protein levels in siRNA treated patient derived stem-like cells. b Immunocytochemistry of nestin and GSK3b in cells treated with various concentrations of GSK3b siRNA. c Western blot on untreated cells and cells treated with GSK3b siRNA. d Western blot on cells transduced with shPP2A lentivirus in order to knock down the PP2A B56 b subunit. e Western blot on cells treated with

various concentrations of okadaic acid (OA) for 30 min and 12 h. Actin is used as a loading control for all western blots in this figure. Dose series for GSK3b siRNA shows effective silencing at 1 lM concentrations (a, b). Silencing GSK3b resulted in down regulation of PP2A (c) in patient-derived GSCs. Knocking down PP2A on the B56 b subunit (d) or treating cells with OA (e) increased the phosphorylation of GSK3b while decreasing total levels of GSK3b

Serine 9. To assess the effect of GSK3b Serine 9 phosphorylation on GSC proliferation and apoptosis, we either stimulated cells with EGF or inhibited the activity of the EGFR and Akt signaling pathway biochemically and genetically. We found that dephosphorylation of GSK3b on Serine 9 attenuated GSC proliferation and induced its proapoptotic activity via Caspase-3 activation. Song et al. [30]

showed that GSK3b is essential for the activation of Caspase-3 in stress induced apoptosis. Recent studies revealed that Bcl2L12 acts as an anti-apoptotic regulator in GBMs by inhibiting post-mitochondrial apoptotic signaling pathway [31–33]. In a recent study, Cho et al. [34] reported that overexpression of Bcl2L12 resulted in suppressed activation of Caspase-3 and Caspase-7. Here we provide

123

J Neurooncol (2015) 121:19–29

evidence that Serine 9 phosphorylation is an important determinant in patient-derived GSC apoptosis, mediated in part by Caspase-3 activation. To date, the role of GSK3b in apoptosis is controversial. Although overexpression of GSK3b can stimulate apoptosis [35, 36], GSK3b knockdown in mouse embryos is lethal, with liver degeneration due to increased apoptotic activity in hepatocytes [37]. Here we show an indirect intrinsic loop where EGF treated GSCs repressed GSK3b apoptosis. Furthermore, this inhibition was prevented by AG1478, LY294002 and AAV-PTEN. To determine the direct effect of GSK3b on GSC apoptosis and proliferation, we overexpressed the wild type GSK3b in GSK3b-deficient GSCs. Our results indicate that GSK3b overexpression increased both early and late apoptosis and impaired cell proliferation. Similarly, Ma and co-workers [38] found that GSK3b is significantly down regulated in human squamous cell carcinoma and basal cell carcinomas. Because GSK3b acts as a negative regulator in several oncogenic pathways, we hypothesized that it might also function as a tumor suppressor during transformation of GSCs. Our results prove that GSK3b overexpression in stem-like cells inhibits colony formation in soft agar. These experiments suggest that GSK3b may indeed function as a regulator of apoptosis and tumorigenesis in GSCs. Serine/Threonine PP2A is believed to function as tumor suppressor. It has been shown that the inhibition of PP2A stimulated cellular transformation [39]. Aside from its tumor suppressor role, PP2A has been shown to exhibit different roles in GBM as well. Lu et al. [40] demonstrated that the inhibition of PP2A with a combination of temozolomide (TZM) and the specific PP2A inhibitor LB-1.2 increased GBM chemosensitivity in a xenograft mouse model. We have shown previously that higher PP2A activity is correlated with poor outcomes in GBM patients and enhanced PP2A activity, regulated by oxygen tension, increases survival of GBM-derived tumor stem-like cells [41]. Activity of PP2A negatively regulates Akt. Overexpression of PP2A at the B subunit (B55a) significantly blocked Akt phosphorylation at Threonine 308 and to a lesser extent at Serine 473. Moreover, PP2A overexpression impaired NIH 3T3 cell proliferation, while silencing PP2A promoted the survival of pro-lymphoid FL5.12 cells [42]. This data led us to investigate the interaction between GSK3b and PP2A in GSCs. We showed an aberrant down regulation of PP2A in GBM patients compared to the normal brain. Previous work has shown that Akt phosphorylation is a substrate for PP2A, which is a negative regulator for GSK3b activity [42]. PP2A regulation on GSK3b could have clinical relevance in GBM patients. Therefore, therapies targeting PP2A, GSK3b or both may increase a patient’s response to chemotherapy. There is no available information supporting a direct association

27

between these two signaling molecules in GBM patients. To determine if there is a direct link between GSK3b and PP2A, we knocked down GSK3b and PP2A in GSCs. Silencing GSK3b down regulated PP2A in all subunits. Knocking down PP2A or blocking the activity with okadaic acid inhibited the activity of GSK3b by Serine 9 phosphorylation. Results showing a direct interaction between GSK3b and PP2A suggest that GSK3b activity may potentially be regulated by PP2A in GSCs. The present study proposes that GSK3b activity is a critical regulator of GSC survival and apoptosis. The balance between the two processes relies on the regulation of GSK3b. Positive EGFR signaling promotes cell proliferation and transformation by (a) the activation of PI3kinase/ Akt signaling and (b) inactivation of GSK3b. PP2A stimulates or/and maintains GSK3b in an active state by suppressing Serine 9 phosphorylation. Inhibition of PP2A increases the phosphorylation of GSK3b, therefore inhibiting its pro-apoptotic activity in human GSCs. We therefore postulate that combined therapies targeting EGFR, GSK3b and PP2A signaling pathways may be more beneficial in our GBM patients. Acknowledgments The authors would like to thank Maria Irina Chiriac for excellent technical assistance in preparing the figures of this manuscript. Disclosure The authors declare that they have no conflict of interest. The authors report no financial or material support concerning the materials or methods used in this study or the findings specified in this paper.

References 1. Bleau AM, Howard BM, Taylor LA, Gursel D, Greenfield JP, Lim Tung HY, Holland EC, Boockvar JA (2008) New strategy for the analysis of phenotypic marker antigens in brain tumorderived neurospheres in mice and humans. Neurosurg Focus 24:E28. doi:10.3171/FOC/2008/24/3-4/E27 2. Pollard SM, Yoshikawa K, Clarke ID, Danovi D, Stricker S, Russell R, Bayani J, Head R, Lee M, Bernstein M, Squire JA, Smith A, Dirks P (2009) Glioma stem cell lines expanded in adherent culture have tumor-specific phenotypes and are suitable for chemical and genetic screens. Cell Stem Cell 4:568–580. doi:10.1016/j.stem.2009.03.014 3. Sanai N, Alvarez-Buylla A, Berger MS (2005) Neural stem cells and the origin of gliomas. N Engl J Med 353:811–822. doi:10. 1056/NEJMra043666 4. Wang JC (2007) Evaluating therapeutic efficacy against cancer stem cells: new challenges posed by a new paradigm. Cell Stem Cell 1:497–501 5. Gursel DB, Shin BJ, Burkhardt JK, Kesavabhotla K, Schlaff CD, Boockvar JA (2011) Glioblastoma stem-like cells-biology and therapeutic implications. Cancers 3:2655–2666. doi:10.3390/ cancers3022655 6. Embi N, Rylatt DB, Cohen P (1980) Glycogen synthase kinase-3 from rabbit skeletal muscle. Separation from cyclic-AMPdependent protein kinase and phosphorylase kinase. Eur J biochem/FEBS 107:519–527

123

28 7. Ring DB, Johnson KW, Henriksen EJ, Nuss JM, Goff D, Kinnick TR, Ma ST, Reeder JW, Samuels I, Slabiak T, Wagman AS, Hammond ME, Harrison SD (2003) Selective glycogen synthase kinase 3 inhibitors potentiate insulin activation of glucose transport and utilization in vitro and in vivo. Diabetes 52:588–595 8. Ougolkov AV, Fernandez-Zapico ME, Bilim VN, Smyrk TC, Chari ST, Billadeau DD (2006) Aberrant nuclear accumulation of glycogen synthase kinase-3beta in human pancreatic cancer: association with kinase activity and tumor dedifferentiation. Clinical cancer research : an official journal of the American Association for Cancer Research 12:5074–5081. doi:10.1158/ 1078-0432.CCR-06-0196 9. Ougolkov AV, Billadeau DD (2006) Targeting GSK-3: a promising approach for cancer therapy? Future Oncol 2:91–100. doi:10.2217/14796694.2.1.91 10. Kang T, Wei Y, Honaker Y, Yamaguchi H, Appella E, Hung MC, Piwnica-Worms H (2008) GSK-3 beta targets Cdc25A for ubiquitin-mediated proteolysis, and GSK-3 beta inactivation correlates with Cdc25A overproduction in human cancers. Cancer Cell 13:36–47. doi:10.1016/j.ccr.2007.12.002 11. Ding Q, Xia W, Liu JC, Yang JY, Lee DF, Xia J, Bartholomeusz G, Li Y, Pan Y, Li Z, Bargou RC, Qin J, Lai CC, Tsai FJ, Tsai CH, Hung MC (2005) Erk associates with and primes GSK-3beta for its inactivation resulting in upregulation of beta-catenin. Mol Cell 19:159–170. doi:10.1016/j.molcel.2005.06.009 12. Shi CS, Huang NN, Harrison K, Han SB, Kehrl JH (2006) The mitogen-activated protein kinase kinase kinase kinase GCKR positively regulates canonical and noncanonical Wnt signaling in B lymphocytes. Mol Cell Biol 26:6511–6521. doi:10.1128/MCB. 00209-06 13. Lenferink AE, Busse D, Flanagan WM, Yakes FM, Arteaga CL (2001) ErbB2/neu kinase modulates cellular p27(Kip1) and cyclin D1 through multiple signaling pathways. Cancer Res 61:6583–6591 14. Ougolkov AV, Fernandez-Zapico ME, Savoy DN, Urrutia RA, Billadeau DD (2005) Glycogen synthase kinase-3beta participates in nuclear factor kappaB-mediated gene transcription and cell survival in pancreatic cancer cells. Cancer Res 65:2076–2081. doi:10.1158/0008-5472.CAN-04-3642 15. Tan J, Zhuang L, Leong HS, Iyer NG, Liu ET, Yu Q (2005) Pharmacologic modulation of glycogen synthase kinase-3beta promotes p53-dependent apoptosis through a direct Bax-mediated mitochondrial pathway in colorectal cancer cells. Cancer Res 65:9012–9020. doi:10.1158/0008-5472.CAN-05-1226 16. Korur S, Huber RM, Sivasankaran B, Petrich M, Morin P Jr, Hemmings BA, Merlo A, Lino MM (2009) GSK3beta regulates differentiation and growth arrest in glioblastoma. PLoS One 4:e7443. doi:10.1371/journal.pone.0007443 17. Li Y, Lu HM, Li G, Yan GM (2010) Glycogen synthase kinase3beta regulates astrocytic differentiation of U87-MG human glioblastoma cells. Acta Pharmacol Sin 31:355–360. doi:10.1038/ aps.2010.10 18. Lin CF, Chen CL, Chiang CW, Jan MS, Huang WC, Lin YS (2007) GSK-3beta acts downstream of PP2A and the PI 3-kinaseAkt pathway, and upstream of caspase-2 in ceramide-induced mitochondrial apoptosis. J Cell Sci 120:2935–2943. doi:10.1242/ jcs.03473 19. Cohen P, Frame S (2001) The renaissance of GSK3. Nat Rev Mol Cell Biol 2:769–776. doi:10.1038/35096075 20. Gursel DB, Beyene RT, Hofstetter C, Greenfield JP, Souweidane MM, Kaplitt M, Arango-Lievano M, Howard B, Boockvar JA (2011) Optimization of glioblastoma multiforme stem cell isolation, transfection, and transduction. J Neurooncol 104:509–522. doi:10.1007/s11060-011-0528-2 21. Ekstrand AJ, James CD, Cavenee WK, Seliger B, Pettersson RF, Collins VP (1991) Genes for epidermal growth factor receptor,

123

J Neurooncol (2015) 121:19–29

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

transforming growth factor alpha, and epidermal growth factor and their expression in human gliomas in vivo. Cancer Res 51:2164–2172 Chakravarti A, Zhai G, Suzuki Y, Sarkesh S, Black PM, Muzikansky A, Loeffler JS (2004) The prognostic significance of phosphatidylinositol 3-kinase pathway activation in human gliomas. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 22:1926–1933. doi:10.1200/ JCO.2004.07.193 Sathornsumetee S, Reardon DA, Desjardins A, Quinn JA, Vredenburgh JJ, Rich JN (2007) Molecularly targeted therapy for malignant glioma. Cancer 110:13–24. doi:10.1002/cncr.22741 Miyashita K, Kawakami K, Nakada M, Mai W, Shakoori A, Fujisawa H, Hayashi Y, Hamada J, Minamoto T (2009) Potential therapeutic effect of glycogen synthase kinase 3beta inhibition against human glioblastoma. Clinical cancer research : an official journal of the American Association for Cancer Research 15:887–897. doi:10.1158/1078-0432.CCR-08-0760 Cross DA, Alessi DR, Cohen P, Andjelkovich M, Hemmings BA (1995) Inhibition of glycogen synthase kinase-3 by insulin mediated by protein kinase B. Nature 378:785–789. doi:10.1038/ 378785a0 Cross DA, Alessi DR, Vandenheede JR, McDowell HE, Hundal HS, Cohen P (1994) The inhibition of glycogen synthase kinase-3 by insulin or insulin-like growth factor 1 in the rat skeletal muscle cell line L6 is blocked by wortmannin, but not by rapamycin: evidence that wortmannin blocks activation of the mitogen-activated protein kinase pathway in L6 cells between Ras and Raf. Biochem J 303(Pt 1):21–26 Hurel SJ, Rochford JJ, Borthwick AC, Wells AM, Vandenheede JR, Turnbull DM, Yeaman SJ (1996) Insulin action in cultured human myoblasts: contribution of different signalling pathways to regulation of glycogen synthesis. Biochem J 320(Pt 3):871–877 Saito Y, Vandenheede JR, Cohen P (1994) The mechanism by which epidermal growth factor inhibits glycogen synthase kinase 3 in A431 cells. Biochem J 303(Pt 1):27–31 Pap M, Cooper GM (1998) Role of glycogen synthase kinase-3 in the phosphatidylinositol 3-Kinase/Akt cell survival pathway. J Biol Chem 273:19929–19932 Song L, De Sarno P, Jope RS (2002) Central role of glycogen synthase kinase-3beta in endoplasmic reticulum stress-induced caspase-3 activation. J Biol Chem 277:44701–44708. doi:10. 1074/jbc.M206047200 Stegh AH, Kim H, Bachoo RM, Forloney KL, Zhang J, Schulze H, Park K, Hannon GJ, Yuan J, Louis DN, DePinho RA, Chin L (2007) Bcl2L12 inhibits post-mitochondrial apoptosis signaling in glioblastoma. Genes Dev 21:98–111. doi:10.1101/gad.1480007 Stegh AH, Kesari S, Mahoney JE, Jenq HT, Forloney KL, Protopopov A, Louis DN, Chin L, DePinho RA (2008) Bcl2L12mediated inhibition of effector caspase-3 and caspase-7 via distinct mechanisms in glioblastoma. Proc Natl Acad Sci USA 105:10703–10708. doi:10.1073/pnas.0712034105 Stegh AH, Chin L, Louis DN, DePinho RA (2008) What drives intense apoptosis resistance and propensity for necrosis in glioblastoma? A role for Bcl2L12 as a multifunctional cell death regulator. Cell Cycle 7:2833–2839 Chou CH, Chou AK, Lin CC, Chen WJ, Wei CC, Yang MC, Hsu CM, Lung FW, Loh JK, Howng SL, Hong YR (2012) GSK3beta regulates Bcl2L12 and Bcl2L12A anti-apoptosis signaling in glioblastoma and is inhibited by LiCl. Cell Cycle 11:532–542. doi:10.4161/cc.11.3.19051 Pap M, Cooper GM (2002) Role of translation initiation factor 2B in control of cell survival by the phosphatidylinositol 3-kinase/ Akt/glycogen synthase kinase 3beta signaling pathway. Mol Cell Biol 22:578–586

J Neurooncol (2015) 121:19–29 36. Bijur GN, De Sarno P, Jope RS (2000) Glycogen synthase kinase3beta facilitates staurosporine- and heat shock-induced apoptosis. Protection by lithium. J Biol Chem 275:7583–7590 37. Hoeflich KP, Luo J, Rubie EA, Tsao MS, Jin O, Woodgett JR (2000) Requirement for glycogen synthase kinase-3beta in cell survival and NF-kappaB activation. Nature 406:86–90. doi:10. 1038/35017574 38. Ma C, Wang J, Gao Y, Gao TW, Chen G, Bower KA, Odetallah M, Ding M, Ke Z, Luo J (2007) The role of glycogen synthase kinase 3beta in the transformation of epidermal cells. Cancer Res 67:7756–7764. doi:10.1158/0008-5472.CAN-06-4665 39. Schonthal AH (2001) Role of serine/threonine protein phosphatase 2A in cancer. Cancer Lett 170:1–13 40. Lu J, Kovach JS, Johnson F, Chiang J, Hodes R, Lonser R, Zhuang Z (2009) Inhibition of serine/threonine phosphatase

29 PP2A enhances cancer chemotherapy by blocking DNA damage induced defense mechanisms. Proc Natl Acad Sci USA 106:11697–11702. doi:10.1073/pnas.0905930106 41. Hofstetter CP, Burkhardt JK, Shin BJ, Gursel DB, Mubita L, Gorrepati R, Brennan C, Holland EC, Boockvar JA (2012) Protein phosphatase 2A mediates dormancy of glioblastoma multiforme-derived tumor stem-like cells during hypoxia. PLoS One 7:e30059. doi:10.1371/journal.pone.0030059 42. Kuo YC, Huang KY, Yang CH, Yang YS, Lee WY, Chiang CW (2008) Regulation of phosphorylation of Thr-308 of Akt, cell proliferation, and survival by the B55alpha regulatory subunit targeting of the protein phosphatase 2A holoenzyme to Akt. J Biol Chem 283:1882–1892. doi:10.1074/jbc.M709585200

123