Progress report for the project "Relevance of Tumor Metabolism for Glioblastoma-Host Interaction" (February 2015).
In the nine months since our previous interim report in May 2014, we continued to work on the three work packages defined in the original project proposal.
In the first work package we investigate whether the assocation between the pentose phosphate pathway (PPP) and proliferation as well as the association between glycolysis and cell migration exist also independently of changes in oxygen levels. We continued to optimize the methodologies for selecting highly proliferative as opposed to highly migratory cell subpopulations. Using a technique in which cells are labeled with PKH67, a fluorescent dye which incorporates into cell membranes and dilutes with every cell division, we were able to enrich highly proliferative populations from different glioblastoma cell lines. Expression analyses revealed that these fast cycling cells displayed a significantly reduced expression of glycolysis enzymes but increased expression of PPP enzymes compared with slowly cycling cells. The selection of highly migratory cell subpopulations from several different glioblastoma cell lines was successfully achieved by using transwell-assays, in which glioblastoma cells migrate through pores of a membrane. In contrast to highly proliferative cells, the highly migratory cells exhibited increased expression of glycolysis enzymes but reduced expression of PPP enzymes. Furthermore, the expresion of tumor stem cell markers was increased in highly migratory cell populations, whereas highly proliferative cells displayed reduced expression of these markers, indicating that slowly cycling cells tend to possess a more pronounced stem-like phenotype.
In the second work package we investigate whether glycolysis and the PPP are of direct, causal relevance to glioblastoma cell migration or proliferation, respectively. To address this question, we performed analyses with enzyme inhibitors, which showed that inhibition of glucose-6-phoshate dehydrogenase [(G6PD), the first and key regulatory enzyme of the PPP] with 6-aminonicotinamide inhibited proliferation while simultaneously stimulating migration. In contrast, enzyme inhibition upstream of both pathways, at the level of hexokinase, using 2-deoxyglucose, inhibited proliferation as well as migration. These findings were reproduced on several glioblastoma stem-like cell lines. In addition, the expression of G6PD and Aldolase C (the most strongly hypoxia-induced enzyme of glycolsis according to our data) was knocked down using shRNA in two glioblastoma cell lines. In vitro analyses showed that downregulation of G6PD consistently entailed increased proliferation, whereas downregulation of Aldolace C caused decreased migration. Moreover, in vivo experiments using a rapidly dividing glioblasoma cell line showed that downregulation of G6PD resulted in prolonged survival in an orthotopic intracranial mouse glioblastoma xenograft model, whereas knockdown of Aldolase C resulted in shortened survival.
In the third work package we assess whether the oxygen concentration-dependent switch between the PPP and glycolysis is a general phenomenon or whether it is specific for glioblastoma cells cultured under neural stem cell conditions. Enzyme expression analyses under hypoxia as opposed to normoxia were performed on normal (untransformed) cells present in the human brain (astrocytes, mesenchymal stem cells, endothelial cells, fibroblasts, mononuclear cells) as well as on various tumor cell types (glioblastoma, liver cancer, and breast cancer). Our results consistently showed that the hypoxia-induced increased expression of glycolysis enzymes, along with reduced expression of PPP enzymes, is a general phenomenon, present at the transcript as well as at the protein level.
In future work, several of the enrichment analyses (part 1) need to be repeated and extended, in particular with optimized time points for cell selection, and experiments with selective stimulation of proliferation or migration have to be performed. Studies testing the sensitivity of cell populations to enzyme inhibitors remain to be conducted. Additional in vivo experiments need to be done, and experiments inhibiting glucose-6-phosphate isomerase (part 2) are still outstanding. Finally, the question whether different enzyme isoforms which are preferentially expressed in different types of cells show the same regulation as those preferentially expressed in brain and/or cancer are to be performed (part 3), and metabolic flux analyses are still pending.