Progress report for the project "Tumor metabolism - interaction between glioblastoma cells and brain" (November 2016)
The research project focuses on three main aims and work packages. The first aim addresses the question whether glycolyis and glioblastoma cell migration are mechanistically linked. We hypothesize that the glycolysis enzyme glucose-6-phosphate-isomerase (GPI) plays an important role in glioblastoma cell migration and invasion. GPI functions not only as a glycolysis enzyme but is also known as autocrine motility factor (AMF), a factor that can be secreted from cells and can bind to the cell surface receptor AMFR, thereby stimulating the motility of for example melanoma cells in an autocrine fashion. Our data obtained so far show that both GPI/AMF as well as its receptor AMFR are most strongly expressed in the most severely hypoxic regions in glioblastoma tissue. Furthermore, immunostaining of a tissue microarray revealed that high expression of GPI/AMF is associated with a shorter survival of glioblastoma patients. Analyses on a panel of different glioblastoma cell lines showed that the expression of both GPI/AMF as well as AMFR is strongly enhanced by hypoxia and that in particular the secretion of GPI/AMF into the culture medium is increased under hypoxic conditions. Functional experiments demonstrated that GPI/AMF strongly enhances the chemotactic migration of glioblastoma cells, whereas proliferation tends to be decreased, especially at higher concentrations. The GPI/AMF-Inhibitor erythrose-4-phosphate (E4P) inhibited the migration of glioblastoma cells, and by immunodepleting of GPI/AMF from the culture supernatant we demonstrated that glioblastoma cells stimulate their motility in an autocrine fashion. Downregulation of GPI/AMF expression by shRNA transduction resulted in reduced migration whereas proliferation was largly unchanged. Currently, in vivo experiments are outstanding, in which cells with downregulated GPI/AMF expression will be implanted into the brains of immunocompromized mice as xenograft tumors in order to study their growth and invasion. In summary, the results obtained thus far demonstrate that GPI/AMF is a potent pro-migratory factor for glioblastoma cells, which can in part explain the association between glycolysis and migration. Consequently, GPI/AMF is potentially promising target for the inhibition of glioblastoma cell invasion in brain.
The second aim is to analyze the relevance of intra- and intertumoral heterogeneity for glioblastoma cell metabolism. Intratumoral metabolic heterogeneity is for example obvious in glioblastoma tissue, in which the most severly hypoxic regions display very high expression of glycolysis enzymes but comparatively low expression of pentose phosphate pathway enzymes, whereas the inverse expression pattern is observed in highly proliferative tissue areas. To facilitate analyses of the causes and mechanisms of metabolic heterogeneity at the clonal level, we developed a clonal colour tracing system (optical barcoding). The technique allows to track the fate of individual cells in heterogeneous cultures as well as in xenograft tumors in vivo. Moreover, individual cell clones can be re-sorted from tumors after growth in vivo and can then be used for differential analyses of adaptation to metabolic stress, such as hypoxia and glucose starvation. In order to identify intertumoral differences in metabolic regulation in gliomas with different genetic background, transgenic mouse glioma cell lines with alterations of the TP53, PDGF-B, CDKN2a and EGFR genes will be studied in collaboration with Prof. Glass. Currently, the gene expression profiles of these cell lines are being analzyed by RNAseq.
The third aim is to characterize the hypoxic regulation of additional pathways that directly or indirectly branch off glycolysis. For this purpose, quantitative PCR analyses and Western blot analyses of enzymes of the serine synthesis pathway, the folate cycle, the methionine cycle and the purine and pyrimidine synthesis pathways have been performed. Preliminary findings demonstrate that enzymes of the folate cycle are mostly downregulated by hypoxia, whereas enzymes of the serine synthesis pathway are upregulated. Analyses on tumor material revealed that phophoglycerate dehydrogenase, the first enzyme of the serine synthesis pathway is particularly highly expression in IDH1-mutated gliomas. These analyses now have to be extended and to be validated on additional cell lines to generate a more complete picture. In addition, functional analyses on selected enzymes need to be performed.