Nrf2-driven TERT regulates pentose phosphate pathway in glioblastoma

Abstract

Given the involvement of telomerase activation and dysregulated metabolism in glioma progression, the connection between these two critical players was investigated. Pharmacological inhibition of human Telomerase reverse transcriptase (hTERT) by Costunolide induced glioma cell apoptosis in a reactive oxygen species (ROS)-dependent manner. Costunolide induced an ROS-dependent increase in p53 abrogated telomerase activity. Costunolide decreased Nrf2 level; and ectopic Nrf2 expression decreased Costunolide-induced ROS generation. While TERT knock-down abrogated Nrf2 levels, overexpression of Nrf2 increased TERT expression. Inhibition of hTERT either by Costunolide, or by siRNA or dominant-negative hTERT (DN-hTERT) abrogated (i) expression of Glucose-6-phosphate dehydrogenase (G6PD) and Transketolase (TKT) - two major nodes in the pentose phosphate (PPP) pathway; and (ii) phosphorylation of glycogen synthase (GS). hTERT knock-down decreased TKT activity and increased glycogen accumulation. Interestingly, siRNA-mediated knock-down of TKT elevated glycogen accumulation. Coherent with the in vitro findings, Costunolide reduced tumor burden in heterotypic xenograft glioma mouse model. Costunolide-treated tumors exhibited diminished TKT activity, heightened glycogen accumulation, and increased senescence. Importantly, glioblastoma multiforme (GBM) patient tumors bearing TERT promoter mutations (C228T and C250T) known to be associated with increased telomerase activity; exhibited elevated Nrf2 and TKT expression and decreased glycogen accumulation. Taken together, our findings highlight the previously unknown (i) role of telomerase in the regulation of PPP and glycogen accumulation and (ii) the involvement of Nrf2-TERT loop in maintaining oxidative defense responses in glioma cells.

Figure 1

Costunolide induces glioma cell death and decreases telomerase activity in ROS/p53-dependent manner. (a) Costunolide reduces viability of A172 and U87MG glioma cells, but has no effect on p53 mutant T98G cells. Graph shows percent change in viability upon treatment with different doses of Costunolide for 24 h as determined by MTS assay. (b) The increase in DHE fluorescence induced upon Costunolide treatment is abrogated in the presence of ROS inhibitor NAc. (c) Costunolide-induced increase in TUNEL-positive cells is abrogated in the presence of NAc. Graph depicts the percent of TUNEL-positive cells treated with Costunolide, NAc or both. (d) ROS inhibition rescues Costunolide-mediated decrease in telomerase activity. Graph shows fold change in telomerse activity over control in glioma cells treated with different combinations of Costunolide and NAc. (e) Overexpression of SOD-1 rescues Costunolide-mediated decrease in hTERT expression. Following transfection of glioma cells with SOD-1 overexpression construct for 48 h, cells were treated in the presence and absence of Costunolide for an additional 24 h and hTERT levels were determined. Inset demonstrates SOD-1 expression in transfected cells. (f) Costunolide-induced increase in p53 expression is ROS dependent. Western blot images depicting p53 levels in cells treated with Costunolide in the presence or absence of NAc. (g) siRNA-mediated knock-down of p53 prevents Costunolide-mediated decrease in telomerase activity. Graph represents the telomerase activity of glioma cells transfected with either p53 siRNA or scrambled (NS) siRNA and treated with Costunolide. Values are expressed as fold change over control. Inset confirming the transfection efficiency of p53 siRNA. Values in the graph (a–d and g) represent the means±S.E.M. from three independent experiments. *Significant change from control, ^#Significant change from Costunolide-treated cells (P<0.05). Blots (e and f) are representative images of three independent experiments showing similar results. Blots were re-probed for c23 to establish equivalent loading

Figure 2

Existence of Nrf2-TERT regulatory loop in glioma cells. (a) Costunolide decreases Nrf2 levels in an ROS-dependent manner. Western blots showing Nrf2 levels in nuclear extracts of glioma cells treated with Costunolide in the presence and absence of NAc. (b) Transfection with DN-hTERT construct decreases Nrf2 levels in glioma cells, as demonstrated by western blot analysis. Inset confirming hTERT levels upon with transfection with DN-hTERT construct. (c) Western blot analysis demonstrating decreased Nrf2 levels in cells upon siRNA-mediated knock-down of hTERT. Inset confirming the transfection efficiency of hTERT siRNA. (d) Nrf2 overexpression increases TERT expression and (e) telomerase activity. Inset in (d) demonstrates Nrf2 expression in transfected cells. (f) Costunolide-induced ROS generation is abrogated upon Nrf2 overexpression. The graph represents DHE fluorescence intensity in cells transfected with either Nrf2 overexpression construct (OE-Nrf2) or empty vector and treated in the presence or absence of Costunolide. Fluorescence intensity values are expressed as fold change over control. Blots (a–d) are representative images of three independent experiments showing similar results. Blots were re-probed for β-actin or c23 to establish equivalent loading. Values in (e and f) represent the means±S.E.M. of three independent experiments. *Denotes significant change from control or mock-transfected group, ^#depicts significant change from Costunolide-treated cells (P<0.05)

Figure 3

TERT regulates pentose phosphate pathway in glioma cells. (a) Costunolide decreases expression of G6PD and TKT in an ROS-dependent manner. Western blot demonstrating G6PD and TKT levels in cells treated with different combinations of Costunolide and NAc. (b) Costunolide-mediated decrease in TKT activity is rescued upon ROS inhibition. Graph showing TKT activity in cells treated with Costunolide in the presence and absence of NAc. (c) Transfection of glioma cells with DN-hTERT decreases G6PD and TKT levels in glioma cells, as depicted by western blot analysis. (d) Decreased TKT activity in glioma cells transfected with DN-hTERT construct. Values represent fold change in TKT activity in DN-hTERT-transfected cells over mock-transfected control. (e) Western blot demonstrating decreased G6PD and TKT levels upon siRNA-mediated knock-down of hTERT. (f) siRNA-mediated knock-down of TERT decreases TKT activity in glioma cells. Values represent fold change in TKT activity in TERT siRNA-transfected cells over NS-siRNA-transfected control. (g) Western blot analysis depicting G6PD and TKT levels in glioma cells transfected with either Nrf2 overexpression construct (OE-Nrf2) or empty vector and treated in the presence or absence of Costunolide. Blots shown in (a, c, e, and g) are representative images of three independent experiments showing an identical trend. Blots were re-probed for β-actin to establish equivalent loading. Values in (b, d, and f) are means±S.E.M. of three independent experiments. *Denotes significant change control or mock-transfected group, ^#depicts significant change from Costunolide-treated cells (P<0.05)

Figure 4

TERT regulates glycogen accumulation and senescence. (a) Costunolide induces glioma cell senescence in an ROS-dependent manner. The graph represents percentage β-gal-positive cells observed upon treatment with Costunolide in the presence and absence of NAc. (b) Costunolide decreases phospho-glycogen synthase GS(P) levels in an ROS-dependent manner. Western blot depicting GS(P) and GS levels in cells treated with different combinations of Costunolide and NAc. (c) Costunolide-mediated increase in glycogen accumulation is ROS dependent. Graph showing glycogen levels in cells treated with Costunolide in the presence and absence of NAc. (d) Transfection with DN-hTERT diminishes GS(P) levels and (e) increases glycogen accumulation in glioma cells. (f) siRNA-mediated knock-down of TERT decreases phosphorylated GS levels and (g) increases glycogen accumulation. (h) Western blot analysis demonstrating diminished GS(P) levels in glioma cells upon siRNA-mediated knock-down of TKT. (i) siRNA-mediated knock-down of TKT increases glycogen accumulation. (j) Western blot analysis depicting the effect of Nrf2 overexpression on Costunolide-induced changes in GS(P) and GS levels in glioma cells. Blots (b, d, f, h, and j) are representative of three independent experiments showing similar results. Blots were re-probed for β-actin to establish equivalent loading. The values in the graph (a, c, e, g, and i) represent the mean±S.E.M. from three independent experiments. *Significant change from control or non-specific siRNA or mock-transfected cells, ^#significant change from Costunolide-treated cells (P<0.05)

Figure 5

Costunolide inhibits pentose phosphate pathway and increases glycogen accumulation in heterotypic xenograft glioma model. (a) Significant reduction in tumor volume in Costunolide-treated glioma xenografts as compared with untreated groups (n=7). (b) Costunolide-treated tumors show decreased telomerase activity as compared with control groups (n=4). (c) Elevated ROS levels in Costunolide-treated tumors as compared with untreated group, as indicated by increased DHE fluorescence (n=4). (d) Western blot showing decreased hTERT, Nrf2, G6PD, TKT, and GS (P) levels in the total cell lysates prepared from Costunolide-treated tumors as compared with untreated groups. Blot is representative images (n=6). Blot was re-probed for β-actin to establish equivalent loading. (e) Decreased TKT activity and (f) elevated glycogen levels in Costunolide-treated tumors as compared with untreated controls. The values in graph (b, c, e, and f) indicate means±S.E.M. of n=4. Statistical analysis was done using unpaired Student's t-test. Results were considered as significant, when P-value was equal to or less than 0.05. *Denotes significant change from the untreated group. (g) Costunolide increases glycogen accumulation in cells undergoing senescence as demonstrated by increased PAS and β-glactosidase staining. Sections derived from Costunolide-treated U87 xenografts were co-stained for β-glactosidase and glycogen (PAS). Representative image from animals from each group is shown (n=4)

Figure 6

Co-relation between TERT, TKT, and Nrf2 in GBM tumors. (a) Sanger sequencing indicating C228T and (b) C250T mutation in GBM tumor samples. (c) qRT-PCR indicating elevated TKT and (d) Nrf2 mRNA expression levels in C228T- or C250T-mutated tumor samples. Result were analyzed using non-parametric Mann–Whitney test. (e) Immunohistochemistry showing decreased glycogen accumulation in GBM tumor bearing C228T or C250T mutation. MT and WT represent TERT mutant and wild-type GBM tumors, respectively. (f) Proposed model demonstrating the existence of Nrf2-TERT regulatory loop, and its involvement in regulating metabolic adaptation in glioma cells. *denotes significant change from WT