MR-spektroskopische Metabolitenanalyse von Biopsatproben für die Tumordiagnose

Abstract

Gliomas are the most common primary brain tumours among adults and reveal an altered metabolism which can be analyzed by NMR spectroscopy. The tumour delineation is still a challenging task because gliomas are invasively growing tumours which tend to relapse after neurosurgical resection due to (unrecognized) remaining tumour cells. Therefore, the aim of this study was to characterize the regional metabolic profiles of different tumour sections of malignant gliomas. For this purpose specimens of active, viable tumour core, tumour margin and more distant tumour margin were extracted with an optimized dual phase extraction procedure and hydrophilic as well as lipophilic tissue extracts were recorded by NMR spectroscopy. Different pattern recognition methods (partial least squares regression, linear discriminant analysis and self organizing maps) were used to recognize relevant metabolites for an intratumoral discrimination and to visualize the high dimensional data. Furthermore, the metabolic profiles of the different glioma types were compared. The tumours revealed characteristic metabolic profiles which reflected typical neoplastic alterations, i.e. glycine and phosphocholine were markers of malignancy and elevated in tumour core extracts of glioblastomas. Statistical analysis showed that NAA, N-Acetylgalactosamine and creatine were relevant metabolites for an intratumoral discrimination and decreased from more distant tumour margin towards the tumour core because of a displacement of neuronal structures as well as an altered glycoconjugate and energy metabolism in the tumour cells. Furthermore, the latter caused elevated levels of alanine and branched chain amino acids in the tumour core extracts. Among lipids dolichol(phosphate), triacylglycerides and cholesteryl esters were relevant components which were preferentially detected in the tumour core. Also polyunsaturated fatty acids (i.e. arachidonic and linoleic acid) were elevated in the tumor core. In addition, the ratio of galactosyl cerebrosides to sphingomyelins represented a good metabolic discriminator. In order to get further insights in tumour lipid metabolism a solid phase extraction for the separation of neural lipids was developed. Lipid extracts of the three different tumour regions were separated in free fatty acids, neutral and acidic phospholipids, cholesteryl esters, triacylglycerides, cholesterol, di- and monoacylglycerides. Results showed a correlation of polyunsaturated fatty acids with polar lipids in the tumour core confirming a higher demand of membrane fluidity and disordered signaling pathways. In conclusion, hydrophilic as well as lipophilic extracts of the different tumour regions showed distinct alterations in their metabolic profiles which enable an intratumoral discrimination. Metabolic differences are caused by a variable fraction of neoplastic cells in the biopsies, which decreases from tumour core over tumour margin to the more distant tumour margin, and further towards healthy cells with a tumour affected metabolism.