PPARγ regulated CIDEA affects pro-apoptotic responses in glioblastoma

Abstract

Refractoriness of glioblastoma multiforme (GBM) to current treatment paradigms has necessitated identification of new targets to better the existing therapeutic strategies. One such target is peroxisome proliferator-activated receptor gamma (PPARγ) - a transcription factor involved in regulation of lipid metabolism and inflammation. Expression of PPARγ, a known regulator of cell death-inducing DFFA-like effector (CIDEA), is modulated by hypoxia inducible factor (HIF-1α). While the involvement of CIDEA in lipid metabolism is known, its role in malignancies remains largely unknown. An elevated PPARγ and low CIDEA level was observed in GBM tumors as compared with surrounding non-neoplastic tissue. As reciprocal relation exists between PPAR and HIF-1α: and as HIF-1α is a key component in glioma progression, their role in regulating CIDEA expression in glioblastoma was investigated. Although HIF-1α inhibition had no effect on CIDEA expression, pharmacological inhibition of PPARγ elevated CIDEA levels. PPARγ mediated upregulation of CIDEA was accompanied by decreased recruitment of NFκB and SP1 to their predicted binding sites on CIDEA promoter. Ectopic expression of CIDEA triggered apoptosis, activated JNK, decreased HIF-1α activation and increased PPARγ levels in glioma cells. While CIDEA overexpression induced actin cytoskeletal disruption, cell cycle arrest, release of pro-inflammatory cytokine IL-6 in a JNK-dependent manner; CIDEA mediated apoptotic cell death, decreased STAT3 phosphorylation and increased p53 acetylation was JNK independent. This study highlights for the first time the existence of (i) PPARγ-CIDEA regulatory loop in glioma and (ii) novel function of CIDEA as regulator of glioma cell survival.

Figure 1

PPARγ and CIDEA expression in glioblastoma. (a) CIDEA expression is significantly low in GBM tumors from different regions of brain as indicated by Gene Expression Omnibus (GEO) database (data set record no. GDS4470). The significance is calculated by two-tailed Mann–Whitney U-test. (b) CIDEA mRNA expression is significantly low in GBM as compared with normal brain tissue. Figure is presented of TCGA expression profiles as retrieved from Oncomine database. (c) Significantly enriched pathways from co-expressed genes of CIDEA. Co-expression data were collected from Oncomine with >80% correlation coefficient in GBM and analyzed using DAVID Bioinformatics Resource 6.7. (d) Western blot analysis demonstrating levels of CIDEA and PPARγ in GBM tumor as compared with surrounding non-neoplastic tissue. The figure shows blots from four independent tumor samples. Blot was re-probed for GAPDH to establish equal loading.

Figure 2

Inter-regulatory relationship between PPARγ and CIDEA. (a) Western blot showing effect of PPARγ and/or HIF-1α inhibition on CIDEA protein expression in glioma cell lines. Blots were re-probed for GAPDH to establish equivalent loading. (b) Real-time PCR indicating elevated CIDEA mRNA expression on PPARγ inhibition. Graph represents fold change of CIDEA total mRNA expression. 18s rRNA was used as control. (c) Ectopic expression of CIDEA elevates PPARγ levels in glioma cell lines. Inset showing heightened CIDEA on transfection with overexpression (OE) construct. Blots were re-probed for GAPDH to established equal loading. (a and c) Representative blot from three independent experiments with identical results. (d) CIDEA negatively regulates HIF-1α transcriptional activation. The graphs represent fold change in HIF-1α luciferase reporter activity over control in cells transfected with CIDEA overexpression construct. Values (b and d) represent the means±S.E.M. from three independent experiments. * denotes significant change from control (P<0.05). LW6 and T007 are HIF-1α and PPARγ inhibitors, respectively.

Figure 3

Inhibition of PPARγ reduces NFκB and SP1 binding on CIDEA promoter. (a) PPARγ inhibitior T007 has no effect on NFκB and SP1 expression. Blot is representative of two independent experiments. Blots were re-probed for GAPDH to establish equal loading. (b) ChIP-qPCR assays demonstrating decreased binding of NFκB to its cognate sites on CIDEA promoter. DNA isolated from control and PPARγ inhibitor treated A172 glioma cells pre and post immunoprecipitation with anti-NFκB antibody, was amplified using specific primer sets. Binding affinity of NFκB was found to be low at three different putative binding sites on CIDEA promoter on inhibition of PPARγ. (c) PPARγ inhibition decreases SP1 binding to its cognate site on CIDEA promoter at −26 to +120 position, as indicated by ChIP-qPCR assay. Graph (b and c) represents fold change as calculated from Ct values of two independent experiments for a single site.

Figure 4

CIDEA overexpression induces cell death. (a) PPARγ inhibition reduces glioma cell viability. The graph represents the viable cells, fold change over control, observed when glioma cells were treated with 50 μM of T0070907 for 40 h, as determined by MTS assay. (b) CIDEA overexpression reduces cell viability in glioma cell lines, as determined by MTS. (c) CIDEA overexpression induces phospho-JNK expression in glioma cells. Blot is representative of three independent experiments. Blots were re-probed for β-tubulin to establish equal loading control. (d) CIDEA overexpression induces JNK co-localization into the mitochondria. Representative image showing mitochondria stained with mitotracker green, JNK with Alexa Fluor 594-tagged secondary antibody (red) and nucleus with DAPI (blue). (e) CIDEA-mediated glioma cell death is JNK independent. Graph represents the viable cells, fold change over control, observed when CIDEA overexpressing cells were treated with 10μM JNK inhibitor for 40 h, as measured by MTS assay. The graph (a, b, e) represents the means±S.E.M. from three independent experiments. * Significant decrease from control (P≤0.05). (f) Overexpression of CIDEA induces JNK-mediated G2/M phase cell cycle arrest in A172 glioma cell line. Left panel shows population of cells arrested at G2/M phase, the right panel indicates the fold change in cell population. Graph is representative data of two independent experiments.

Figure 5

Overexpression of CIDEA increases p53 acetylation in a JNK-independent manner. CIDEA overexpression affects acetylated p53 and total p53 levels in the nucleus (a) and the cytosol (b) of glioma cells in a JNK-independent manner. A representative blot is shown from three independent experiments with identical results. Blots were re-probed with C23 (for nuclear extract) or GAPDH (for cytosolic extract) as loading control.

Figure 6

JNK regulates CIDEA mediated actin cytoskeletal disruption. CIDEA overexpression alters the levels of phospho-VASP (a) and phospho-Cofilin (b) in glioma cell lines in a JNK dependent manner. A representative blot is shown from two independent experiments with identical results. Blots were re-probed with β-tubulin to establish loading control. (c) CIDEA overexpression affects actin cytoskeleton architecture in a JNK-dependent manner. Confocal imaging depicts that association between Cofilin (green, Alexa Fluor 488) and actin (red, rhodamine-phalloidin) observed in CIDEA overexpressing cells is reverted by JNK inhibition. Representative images from two independent experiments are shown for indicated conditions.

Figure 7

CIDEA induces IL-6 levels but abrogates STAT3 activation in glioma cells. (a) Ectopic expression of CIDEA elevates IL-6 secretion in a JNK-dependent manner as revealed by Cytokine Bead Array. Graph represents representative data from two independent experiments in three cell lines. (b) CIDEA overexpression abrogates cytosolic and nuclear pSTAT3 (Y705) levels in a JNK independent manner. Blots were re-probed with C23 (for nuclear extract) or β-tubulin (for cytosolic extract) to establish equal loading. Blots shown are representative of three independent experiments. (c) Proposed mechanism of regulation of CIDEA and its role in glioma cell survival.