Article
The omics landscape of tumour metabolism in glioblastoma multiforme
Search Medline for
Authors
-
Dieter Henrik Heiland
- Department of Neurosurgery, Medical Center – University of Freiburg, Faculty of Medicine, University of Freiburg, Freiburg, Germany, Freiburg, Deutschland
-
Annette Gäbelein
- Department of Neurosurgery, Medical Center – University of Freiburg, Faculty of Medicine, University of Freiburg, Freiburg, Deutschland
-
Jakob Wörner
- Universitätsklinikum Freiburg, Institut für Physikalische Chemie, Freiburg, Deutschland
-
Pamela Franco Jimenez
- Department of Neurosurgery, Medical Center - University of Freiburg, Faculty of Medicine, University of Freiburg, Germany, Freiburg, Deutschland
-
Sabrina Heynckes
- Department of Neurosurgery, Medical Center – University of Freiburg, Faculty of Medicine, University of Freiburg, Freiburg, Deutschland
-
Dietmar Pfeifer
- Universitätsklinikum Freiburg, Klinik für Hämatologie, Onkologie und Stammzelltransplantation, Freiburg, Deutschland
-
Stefan Weber
- Universitätsklinikum Freiburg, Institut für Physikalische Chemie, Freiburg, Deutschland
-
Irina Mader
- Klinik für Neuroradiologie, Klinik für Neuroradiologie, Bad Krozingen, Deutschland
-
Oliver Schnell
- Department of Neurosurgery, Medical Center – University of Freiburg, Faculty of Medicine, University of Freiburg, Freiburg, Deutschland
| Published: | June 9, 2017 |
|---|
Outline
Text
Objective: Tumourigenesis is driven by the reprogramming of cellular metabolism, which derives directly or indirectly from genetic or epigenetic alterations. The purpose of this study was to analyze the global metabolic/transcriptomic profile of glioblastoma (in-vivo 21 patients and ex-vivo 33 patients) by newly designed bioinformatical tools and to validate altered metabolic pathways in an established glioblastoma cell-model.
Methods: In a fist step, a wide spectrum of genetic and metabolomic “high-throughput” methods were performed to compute an integrative metabolomic/transcriptomic network (H-MRS, 1D-NMR, RNA-sequencing, expression array, next-generation sequencing). Different bioinformatical tools were then combined to compute the network. In the second step, our investigated network was validated in a glioblastoma cell-model, which integrates 6 different stem-like cell lines cultured under normoxia/hypoxia conditions. Identified altered metabolic pathways were validated within the cell-model and analyzed with above described analytic approaches. Additionally, two newly designed bioinformatial tools were investigated to model the dynamic processes of tumor specific metabolic alterations by a Bayesian approach (MetaAnalysis and MetabolicTracking)
Results: First, the metabolomic/transcriptomic network resulted in numbers of tumor specific alterations particularly affected the energy, cell-cycle and neurotransmitter metabolism. However, the metabolite creatine was found to be highly connected within different expression subgroups (proneural vs. mesenchymal) and other biological functions such as invasiveness and cell-cycle. Secondly, the cell-model validation showed a metabolic shift of creatine into energy-associated substrates (glycine, serine) and away from neural-associated metabolic pathways (glutamate, proline, arginine) under hypoxia conditions. This metabolic shift could be rescued by increasing the creatine-source within the cells, which also resulted in a reshaped GBM subgroup pattern on the transcriptional level.
Conclusion: Creatine was identified as a central connector between proneural-, or mesenchymal associated metabolic/transcriptiomic alterations. Therefore, our results highlight the strong coherence between metabolic and genetic alteration and accentuates the relevance of an evaluation of tumor-metabolism in glioblastoma multiforme.
