Expression of PGC1α in glioblastoma multiforme patients
- Authors:
- Published online on: April 3, 2017 https://doi.org/10.3892/ol.2017.5972
- Pages: 4055-4076
-
Copyright: © Cho et al. This is an open access article distributed under the terms of Creative Commons Attribution License.
Abstract
Introduction
Peroxisome proliferator-activated receptor γ coactivator 1α (PGC1α) regulates metabolism (1,2), mitochondrial biogenesis and energy homeostasis (3,4). A number of studies have reported PGC1α as a central regulator of thermogenesis, mitochondrial biogenesis and adaptation to fasting in brown adipose tissue, skeletal muscle, cardiac muscle and the liver (1,5). By contrast, PGC1α in the central nervous system is less associated with energy state or thermogenesis (6). PGC1α expression in the central nervous system is high in the embryonic and early postnatal stages, but is decreased during maturation. PGC1α is expressed mostly by γ-aminobutyric acid-ergic neurons; however, a low level of PGC1α is also expressed in glia in the mature brain (7). There is a significant association between PGC1α and the metabolism of reactive oxygen species. PGC1α-null mice are considerably more sensitive to the neurodegenerative effects of the oxidative stressors 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine and kainic acid, which suggests that PGC1α has a role in cellular antioxidant defense (8).
Numerous clinical studies have reported a significant association between PGC1α and a number of types of cancer. In breast, colon and ovarian cancer (9–12), a significant decrease in PGC1α expression accelerated the ‘Warburg effect’, which allows cancer cells to switch from mitochondrial to glycolytic metabolism to meet the metabolic requirements of proliferation (13). By contrast, increased PGC1α expression is present in melanoma, with a corresponding decrease in patient survival (14). The role of PGC1α in a number of cancer types remains unclear and warrants further studies.
Glioblastoma multiforme (GBM) is the most prevalent and invasive type of brain tumor. It aggressively infiltrates and spreads to the surrounding brain tissue via extensive microvascular proliferation. Numerous necrotic areas surrounded by palisading tumor cells are often observed (15). Although novel therapeutic strategies and improved clinical diagnostics have been introduced, GBM remains one of the most fatal diseases (16). An extensive amount of research has been performed to determine the mechanisms of unlimited proliferation in GBM, as well as its robust resistance to existing drugs and therapies (17,18) In the present study, the expression of PGC1α in normal cortical tissues and GBM tissues was compared. The results of the present study indicate that PGC1α may be a novel biomarker for GBM, as well as a novel target for future GBM therapy development.
Materials and methods
Patient samples
All experiments were performed in accordance with approved guidelines of Chungnam National University Hospital (CNUH; Daejeon, Republic of Korea). The Institutional Review Board of the CNUH approved the experimental protocols and all patients provided written informed consent prior to surgery. A total of 49 patients undergoing tumor resection surgeries at the Department of Neurosurgery, CNUH were enrolled, and pathological diagnoses were confirmed by the Department of Pathology, CNUH via immunohistochemistry. First-time GBM diagnosis was used as the selection criterion, resulting in 26 patient samples that were included in the present study (Table I). The mean age of the patients was 58 years (range, 35 to 74 years). Normal brain tissue samples were obtained from cadavers or from autopsies of surrounding normal brain tissues of consenting GBM patients that underwent surgery (approval no. CNUH 2013-11-006).
Tissue microarray and immunostaining
Tissue microarrays (TMA) were used to perform the comparative histological analysis of normal brain and GBM tissues. The paraffin-embedded sample tissues were de-paraffinized and rehydrated in a graded alcohol series. Tissues were retrieved using 0.01 M citrate buffer (pH 6.0) and heated in a microwave vacuum histoprocessor (RHS-1; Milestone Medical, Bergamo, Italy) at a controlled temperature of 121°C for 15 min. Following washing with phosphate-buffered saline (pH 7.4), tissue sections were incubated with anti-PGC1α antibody (1:200; Santa Cruz Biotechnology, Inc., Dallas, TX, USA; #SC13067) overnight in a humidity chamber at 4°C. Immunohistochemical staining of the tissue sections was performed using avidin-biotin peroxidase complex as previously described (19,20). Additional TMA samples of normal cortex and GBM tissues were obtained from US Biomax, Inc. (Rockville, MD, USA).
All immunostaining was performed with antibodies that detected the N-terminal epitope of PGC1α (1:200; Santa Cruz Biotechnology, Inc.; #sc-13067). For immunofluorescence analysis, PGC1α and COX4 (1:200; Cell Signaling Technology, Inc., Danvers, MA, USA; #4D11-B3-E8) were used as above but with either a Cy3-conjugated antibody (1:500; anti-rabbit; GE Healthcare Life Sciences Chalfont, UK; #PA43004) or a Cy2-conjugated secondary antibody (1:200; anti-mouse; GE Healthcare Life Sciences; #PA42002). Cell nuclei were visualized with DAPI, and double-stained sections were visualized using an Axiophot microscope (Carl Zeiss AG, Oberkochen, Germany).
Bioinformatics
The mRNA expression of 18,988 probes from 38 GBM cell lines was analyzed using the publicly available Broad-Novartis Cancer Cell Line Encyclopedia (CCLE) database (https://portals.broadinstitute.org/ccle/home) (21). The level of PGC1α mRNA expression among the 38 GBM cell lines was determined using CCLE. The mRNA expression data was normalized using the RankNormalize module in GenePattern (http://www.broadinstitute.org/cancer/software/genepattern). Gene Neighbors and Class Neighbors modules in GenePattern (http://www.broadinstitute.org/cancer/software/genepattern) were used to select genes that were closely associated with PGC1α (22). Hierarchical clustering was performed using complete linkage and Pearson rank-correlation distance with software provided by GenePattern (HierarchicalClustering; version 6). The colors in the heat-maps show the relative gene expression compared to the mean expression, with red being higher and blue lower. From the 18,988 gene set, 100 genes that were most correlated with PGC1α were selected for classification by Gene Ontology Enrichment Analysis (GO terms) using Database for Annotation, Visualization and Integrated Discovery (DAVID; http://david.abcc.ncifcrf.gov) (23). Differentially expressed genes (DEGs) were classified according to GO terms based on their biological process, molecular function or cellular component. DAVID provided an overview of extensive pathways (www.biocarta.com) in which various genes interacted, as well as the number of DEGs per pathway with a P-value representing gene enrichment. Gene enrichment score with P<0.05 represents a strong association rather than random chance (23). For genes with unknown biological processes, GeneMANIA database (http://www.genemania.org) was used to predict their function (24).
Statistical analysis
ImageJ software (version 1.47; National Institutes of Health, Bethesda, MD, USA) was used to quantify the optical density (pixels/mm2) or the intensity of images. The results from immunohistochemical staining were analyzed by a paired t-test between two groups. Data were presented as the mean ± standard error. Statistical analyses were performed using the Prism 5.0 software (GraphPad Prism Software, Inc., La Jolla, CA, USA). P<0.05 was considered to indicate a statistically significant difference. Data transformation (log conversion) selection and statistical analyses were performed with either the Microsoft Excel 11.0 (Microsoft Corporation, Redmond, WA, USA) or Prism 5.0 software.
Results
PGC1α is highly and variably expressed in GBM patients
To determine the association between PGC1α and GBM, levels of PGC1α protein in GBM and control (normal cortex) tissues were compared using publicly available TMAs from US Biomax, Inc. (Fig. 1). PGC1α was weakly detectable in the nuclei of cortical tissues in the control, whereas it was highly and sporadically expressed throughout the GBM tissues. Furthermore, PGC1α was mostly expressed within the cytoplasm with pale nucleic density (Fig. 1A). Bright-field immunohistochemical analysis of TMA images using a densitometer revealed that PGC1α expression varied between tumor samples (Fig. 1B).
For additional validation, PGC1α mRNA levels were determined in GBM cell lines (n=38) using the Broad-Novartis CCLE database (21). Comparative analysis of PGC1α expression in GBM and five other types of cancer, including liver, ovarian, endometrial, breast and prostate carcinoma revealed that although there were variations in PGC1α mRNA expression between the GBM cell lines (Fig. 1D), the level of expression was increased in GBM compared to other cancer cell lines (Fig. 1C). Overall, these data demonstrate that PGC1α expression was increased in a subpopulation of GBM cells.
PGC1α is localized to the mitochondria in GBM
As a transcriptional coactivator, PGC1α is reported to be localized in the nuclei of the normal cortex (25). However, immunofluorescence analysis demonstrated localization of PGC1α in the perinuclear or cytoplasmic areas of GBM tissues (Fig. 2A). To confirm the subcellular localization of PGC1α, double staining with anti-PGC1α and anti-COX4 (a mitochondrial marker) antibodies was employed. There was a certain level of colocalization of PGC1α and COX4, thereby indicating that PGC1α was expressed in the mitochondria in GBM in addition to the perinuclear or cytoplasmic areas (Fig. 2B).
Gene Neighbors of PGC1α
Bioinformatics analysis of PGC1α-associated genes was performed. PGC1α mRNA expression levels detected in the GBM cell lines (n=38; Table II) ranged from 3.71 (log2) to 8.83 (log2), which corresponds to a fold-change of 2.38. A total of 100 genes that were strongly correlated with PGC1α were selected using Gene Neighbors (Fig. 3A) and classified using DAVID (23). Genes with significant differences (P<0.05) were classified into two groups based on GO terms: Biological process and cellular components (Tables III and IV). Genes highly expressed in GBM cell lines were largely associated with the generation of metabolite precursors and energy (e.g., the hexose or monosaccharide metabolic processes), oxidation reduction (e.g., mitochondrial electron transport, nicotinamide adenine dinucleotide to ubiquinone and the oxidoreduction coenzyme metabolic process), energy derivation by the oxidation of organic compounds [e.g., acetyl-CoA metabolic and catabolic processes, oxidative phosphorylation, tricarboxylic acid (TCA) cycle, aerobic respiration and glycolysis, and coenzyme metabolic and catabolic processes (e.g., cofactor catabolic process) (Fig. 3B). Notably, highly expressed genes were associated with the mitochondria (e.g., mitochondrial membrane, mitochondrial matrix and mitochondrial respiratory chain), organelle membranes (e.g., organelle inner membrane) and the cellular envelope (Fig. 3C). This observation is in agreement with the finding that PGC1α is localized in the mitochondria in GBM as previously described.
Table III.List of Gene Neighbors of peroxisome proliferator-activated receptor γ coactivator 1α differentially expressed in glioblastoma multiforme cells. |
Table IV.Annotated summary of Gene Neighbors of peroxisome proliferator-activated receptor γ coactivator 1α. |
PGC1α expression is highly correlated with mitochondrial function in GBM
Two-way hierarchical clustering of targeted gene sets was performed between five GBM cell lines with the highest (LNZ308, LN464, DBTRG05MG, LN235 and SNU626) and lowest levels (LN229, KNS60, SF172, SNU466 and KS1) of PGC1α expression. The expression of TCA cycle-(P<0.0001), oxidative phosphorylation (OXPHOS)-(P<0.0001) and lipogenesis-associated genes (P<0.01) was significantly increased in the PGC1α-upregulated cells compared with the PGC1α-downregulated cells (Fig. 4A-C). Furthermore, the expression of antioxidant-associated genes was significantly increased in the PGC1α-upregulated cell lines compared with the PGC1α-downregulated cell lines (Fig. 4D; P<0.0001). Taken together, the data in Figs. 3 and 4 suggest that metabolic and mitochondrial genes were highly expressed in parallel with PGC1α. Notably, genes associated with mitochondrial functions, including TCA cycle, OXPHOS, lipogenesis and antioxidant genes, were highly expressed in cells with high PGC1α levels (Fig. 4), which corroborates the results from a recent study (26) and the colocalization data as previously described in the present study.
Class Neighbors of PGC1α up- and downregulated GBM cell lines
Bioinformatics analysis using Class Neighbors yielded two classes of GBM cell lines. Class A contained the ten most PGC1α-upregulated GBM cell lines, and class B contained the ten most PGC1α-downregulated GBM cell lines (Fig. 5A). Out of a total of 18,988 probe sets, 100 genes that were most strongly correlated with classes A and B and most highly expressed were selected. DAVID analysis classified these genes into three groups based on GO terms: i) Biological process, ii) molecular function and iii) cellular components (Fig. 5B and C; Tables V–VIII). GeneMANIA database analysis resulted in the identification of 52 genes with previously unknown biological interactions with PGC1α, including necdin (NDN).
Table V.List of class A genes highly expressed in peroxisome proliferator-activated receptor γ coactivator 1α-upregulated glioblastoma multiforme cells. |
Table VIII.Annotated summary of class B of peroxisome proliferator-activated receptor γ, coactivator 1α. |
In addition, when genes were analyzed according to cell signaling pathway (BioCarta database), 3 signaling pathways in class A and 5 in class B were identified as statistically significant (P<0.05; Table IX). The results of the present study demonstrate that class A genes play roles in signaling pathways associated with metabolic and mitochondrial electron transport and that class B genes are involved in signaling pathways associated with differentiation and immune function.
Discussion
The objective of the present study was to investigate the association between aberrant expression of PGC1α and GBM, and the role PGC1α may have in patient survival. Protein level data demonstrated that PGC1α expression was increased in a subpopulation of tumor cells, although there were variations between different GBM cell lines and patients. PGC1α localization was identified to differ between GBM tissues and the normal cortex (Fig. 2). These results corroborated with a previous study that detected a brain-specific isoform of PGC1α in the cytoplasm rather than the nucleus (27). It was also reported that the PGC1α isoform becomes localized in the mitochondria via phosphatase and tensin homolog-induced putative kinase 1 and voltage-dependent anion channel (28).
This present study also demonstrated that PGC1α was expressed in the mitochondria of GBM cells. Based on these corroborating results, it is predicted that PGC1α-mediated mitochondrial biogenesis and respiration is increased in GBM cells.
To investigate the role PGC1α has in GBM cells, several bioinformatics analyses were performed. The analyses demonstrated that metabolic and mitochondrial genes were highly correlated with PGC1α in a number of GBM cell lines. Class Neighbors analysis classified PGC1α-expressing GBM cell lines into two groups: Class A and B. Class A contained genes associated with development, neurogenesis, cell structure and motility. Class B contained genes associated with immunity, oncogenesis and signaling, including intracellular, T cell-mediated, ligand-mediated and-calcium mediated pathways. Class A genes are involved in mitochondrial and metabolic pathways, whilst class B genes are involved in differentiation and immune pathways. These data reinforce the hypothesis that PGC1α may have an important role in regulating mitochondrial and metabolic signaling pathways in the GBM microenvironment.
A notable result was the association of NDN with PGC1α. NDN is reported to function as a tumor suppressor in GBM (29) and controls the proliferation of white adipose progenitor cells (30). NDN interacts with PGC1α via nicotinamide adenine dinucleotide dependent protein deacetylase (Sirt-1) and two transcription factors, E2F1 and P53, suggesting that interactions with these cell cycle regulating factors are key to its function (31). Therefore, it is hypothesized that PGC1α enhances antioxidant capacity in GBM by interacting with NDN and Sirt1, leading to delayed progression of necrosis and ultimately increasing overall patient survival. Future studies that elucidate the molecular interactions of PGC1α are required to derive improved insights into the diagnosis, prognosis and treatment of GBM.
Acknowledgements
This work was financially supported by the Chungnam National University Hospital Research Fund in 2012 (SH Kim) and the Basic Science Research Program through the National Research Foundation of Korea, which was funded by the Ministry of Science, ICT and Future Planning (grant no. 2013R1A1A1A05006966) and the Ministry of Education, Science & Technology of South Korea (grant nos. 2012R1A1A2004714 and 2012M3A9B6055302).
References
Finck BN and Kelly DP: PGC-1 coactivators: Inducible regulators of energy metabolism in health and disease. J Clin Invest. 116:615–622. 2006. View Article : Google Scholar : PubMed/NCBI | |
Knutti D and Kralli A: PGC-1, a versatile coactivator. Trends Endocrinol Metab. 12:360–365. 2001. View Article : Google Scholar : PubMed/NCBI | |
Scarpulla RC: Metabolic control of mitochondrial biogenesis through the PGC-1 family regulatory network. Biochim Biophys Acta. 1813:1269–1278. 2011. View Article : Google Scholar : PubMed/NCBI | |
Puigserver P and Spiegelman BM: Peroxisome proliferator-activated receptor-gamma coactivator 1 alpha (PGC-1 alpha): Transcriptional coactivator and metabolic regulator. Endocr Rev. 24:78–90. 2003. View Article : Google Scholar : PubMed/NCBI | |
Lin J, Handschin C and Spiegelman BM: Metabolic control through the PGC-1 family of transcription coactivators. Cell Metab. 1:361–370. 2005. View Article : Google Scholar : PubMed/NCBI | |
Tritos NA, Mastaitis JW, Kokkotou EG, Puigserver P, Spiegelman BM and Maratos-Flier E: Characterization of the peroxisome proliferator activated receptor coactivator 1 alpha (PGC 1alpha) expression in the murine brain. Brain Res. 961:255–260. 2003. View Article : Google Scholar : PubMed/NCBI | |
Cowell RM, Blake KR and Russell JW: Localization of the transcriptional coactivator PGC-1alpha to GABAergic neurons during maturation of the rat brain. J Comp Neurol. 502:1–18. 2007. View Article : Google Scholar : PubMed/NCBI | |
St-Pierre J, Drori S, Uldry M, Silvaggi JM, Rhee J, Jäger S, Handschin C, Zheng K, Lin J, Yang W, et al: Suppression of reactive oxygen species and neurodegeneration by the PGC-1 transcriptional coactivators. Cell. 127:397–408. 2006. View Article : Google Scholar : PubMed/NCBI | |
Watkins G, Douglas-Jones A, Mansel RE and Jiang WG: The localisation and reduction of nuclear staining of PPARgamma and PGC-1 in human breast cancer. Oncol Rep. 12:483–488. 2004.PubMed/NCBI | |
Feilchenfeldt J, Brüundler MA, Soravia C, Tötsch M and Meier CA: Peroxisome proliferator-activated receptors (PPARs) and associated transcription factors in colon cancer: Reduced expression of PPARgamma-coactivator 1 (PGC-1). Cancer lett. 203:25–33. 2004. View Article : Google Scholar : PubMed/NCBI | |
Jiang WG, Douglas-Jones A and Mansel RE: Expression of peroxisome-proliferator activated receptor-gamma (PPARgamma) and the PPARgamma co-activator, PGC-1, in human breast cancer correlates with clinical outcomes. Int J Cancer. 106:752–757. 2003. View Article : Google Scholar : PubMed/NCBI | |
Zhang Y, Ba Y, Liu C, Sun G, Ding L, Gao S, Hao J, Yu Z, Zhang J, Zen K, et al: PGC-1alpha induces apoptosis in human epithelial ovarian cancer cells through a PPARgamma-dependent pathway. Cell Res. 17:363–373. 2007. View Article : Google Scholar : PubMed/NCBI | |
Heiden MG Vander, Cantley LC and Thompson CB: Understanding the Warburg effect: The metabolic requirements of cell proliferation. Science. 324:1029–1033. 2009. View Article : Google Scholar : PubMed/NCBI | |
Vazquez F, Lim JH, Chim H, Bhalla K, Girnun G, Pierce K, Clish CB, Granter SR, Widlund HR, Spiegelman BM and Puigserver P: PGC1α expression defines a subset of human melanoma tumors with increased mitochondrial capacity and resistance to oxidative stress. Cancer cell. 23:287–301. 2013. View Article : Google Scholar : PubMed/NCBI | |
Plate KH and Risau W: Angiogenesis in malignant gliomas. Glia. 15:339–347. 1995. View Article : Google Scholar : PubMed/NCBI | |
Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, Belanger K, Brandes AA, Marosi C, Bogdahn U, et al: Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med. 352:987–996. 2005. View Article : Google Scholar : PubMed/NCBI | |
Cancer Genome Atlas Research Network, . Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature. 455:1061–1068. 2008. View Article : Google Scholar : PubMed/NCBI | |
Verhaak RG, Hoadley KA, Purdom E, Wang V, Qi Y, Wilkerson MD, Miller CR, Ding L, Golub T, Mesirov JP, et al: Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell. 17:98–110. 2010. View Article : Google Scholar : PubMed/NCBI | |
Lee YS, Kang JW, Lee YH and Kim DW: ID4 mediates proliferation of astrocytes after excitotoxic damage in the mouse hippocampus. Anat Cell Biol. 44:128–134. 2011. View Article : Google Scholar : PubMed/NCBI | |
Yi MH, Kim S, Zhang E, Kang JW, Park JB, Lee YH, Chung CK, Kim YM and Kim DW: IQGAP1 expression in spared CA1 neurons after an excitotoxic lesion in the mouse hippocampus. Cell Mol Neurobiol. 33:1003–1012. 2013. View Article : Google Scholar : PubMed/NCBI | |
Barretina J, Caponigro G, Stransky N, Venkatesan K, Margolin AA, Kim S, Wilson CJ, Lehár J, Kryukov GV, Sonkin D, et al: The cancer cell line encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature. 483:603–607. 2012. View Article : Google Scholar : PubMed/NCBI | |
Golub TR, Slonim DK, Tamayo P, Huard C, Gaasenbeek M, Mesirov JP, Coller H, Loh ML, Downing JR, Caligiuri MA, et al: Molecular classification of cancer: Class discovery and class prediction by gene expression monitoring. Science. 286:531–537. 1999. View Article : Google Scholar : PubMed/NCBI | |
Huang DW, Sherman BT, Tan Q, Collins JR, Alvord WG, Roayaei J, Stephens R, Baseler MW, Lane HC and Lempicki RA: The DAVID gene functional classification tool: A novel biological module-centric algorithm to functionally analyze large gene lists. Genome Biol. 8:R1832007. View Article : Google Scholar : PubMed/NCBI | |
Vlasblom J, Zuberi K, Rodriguez H, Arnold R, Gagarinova A, Deineko V, Kumar A, Leung E, Rizzolo K, Samanfar B, et al: Novel function discovery with GeneMANIA: A new integrated resource for gene function prediction in Escherichia coli. Bioinformatics. 31:306–310. 2015. View Article : Google Scholar : PubMed/NCBI | |
Meng H, Liang HL and Wong-Riley M: Quantitative immuno-electron microscopic analysis of depolarization-induced expression of PGC-1alpha in cultured rat visual cortical neurons. Brain Res. 1175:10–16. 2007. View Article : Google Scholar : PubMed/NCBI | |
Deighton RF, Le Bihan T, Martin SF, Gerth AM, McCulloch M, Edgar JM, Kerr LE, Whittle IR and McCulloch J: Interactions among mitochondrial proteins altered in glioblastoma. J Neurooncol. 118:247–256. 2014. View Article : Google Scholar : PubMed/NCBI | |
Soyal SM, Felder TK, Auer S, Hahne P, Oberkofler H, Witting A, Paulmichl M, Landwehrmeyer GB, Weydt P and Patsch W; European Huntington Disease Network, : A greatly extended PPARGC1A genomic locus encodes several new brain-specific isoforms and influences Huntington disease age of onset. Hum Mol Genet. 21:3461–3473. 2012. View Article : Google Scholar : PubMed/NCBI | |
Choi J, Batchu VV, Schubert M, Castellani RJ and Russell JW: A novel PGC-1α isoform in brain localizes to mitochondria and associates with PINK1 and VDAC. Biochem Biophys Res Commun. 435:671–677. 2013. View Article : Google Scholar : PubMed/NCBI | |
Chapman EJ and Knowles MA: Necdin: A multi functional protein with potential tumor suppressor role? Mol Carcinog. 48:975–981. 2009. View Article : Google Scholar : PubMed/NCBI | |
Fujiwara K, Hasegawa K, Ohkumo T, Miyoshi H, Tseng YH and Yoshikawa K: Necdin controls proliferation of white adipocyte progenitor cells. PloS One. 7:e309482012. View Article : Google Scholar : PubMed/NCBI | |
Niinobe M, Koyama K and Yoshikawa K: Cellular and subcellular localization of necdin in fetal and adult mouse brain. Dev Neurosci. 22:310–319. 2000. View Article : Google Scholar : PubMed/NCBI |