Cancer cells upregulate anabolic processes to maintain high rates of cellular turnover. Limiting the supply of macromolecular precursors by targeting enzymes involved in biosynthesis is a promising strategy in cancer therapy. Several tumors excessively metabolize glutamine to generate precursors for nonessential amino acids, nucleotides, and lipids, in a process called glutaminolysis. Here we show that pharmacological inhibition of glutaminase (GLS) eradicates glioblastoma stem-like cells (GSCs), a small cell subpopulation in glioblastoma (GBM) responsible for therapy resistance and tumor recurrence. Treatment with small molecule inhibitors compound 968 and CB839 effectively diminished cell growth and in vitro clonogenicity of GSC neurosphere cultures. However, our pharmaco-metabolic studies revealed that only CB839 inhibited GLS enzymatic activity thereby limiting the influx of glutamine derivates into the TCA cycle. Nevertheless, the effects of both inhibitors were highly GLS specific, since treatment sensitivity markedly correlated with GLS protein expression. Strikingly, we found GLS overexpressed in in vitro GSC models as compared with neural stem cells (NSC). Moreover, our study demonstrates the usefulness of in vitro pharmacometabolomics to score target specificity of compounds thereby refining drug development and risk assessment.

The Notch signaling network determines stemness in various tissues and targeting signaling activity in malignant brain cancers by gamma-secretase inhibitors (GSI) has shown promising preclinical success. However, the clinical translation remains challenging due to severe toxicity side effects and emergence of therapy resistance. Better anti-Notch directed therapies, specifically directed against the tumor promoting Notch receptor 1 signaling framework, and biomarkers predicting response to such therapy are of highest clinical need. We assessed multiple patient datasets to probe the clinical relevance Notch1 activation and possible differential distribution amongst molecular subtypes in brain cancers. We functionally assessed the biological effects of the first-in-human tested blocking antibody against Notch1 receptor (brontictuzumab, BRON) in a collection of glioma stem-like cell (GSC) models and compared its effects to genetic Notch1 inhibition as well as classical pharmacological Notch inhibitor treatment using gamma-secretase inhibitor MRK003. We also assess effects on Wingless (WNT) stem cell signaling activation, which includes the interrogation of genetic WNT inhibition models. Our computed transcriptional Notch pathway activation score is upregulated in neural stem cells, as compared to astrocytes; as well as in GSCs, as compared to differentiated glioblastoma cells. Moreover, the Notch signature is clinical predictive in our glioblastoma patient discovery and validation cohort. Notch signature is significantly increased in tumors with mutant IDH1 genome and tumors without 1p and 19q co-deletion. In GSCs with elevated Notch1 expression, BRON treatment blocks transcription of Notch pathway target genes Hes1/Hey1, significantly reduced the amount of cleaved Notch1 receptor protein and caused significantly impairment of cellular invasion. Benchmarking this phenotype to those observed with genetic Notch1 inhibition in corresponding cell models did result in higher reduction of cell invasion under chemotherapy. BRON treatment caused signs of upregulation of Wingless (WNT) stem cell signaling activity, and vice versa, blockage of WNT signaling caused induction of Notch target gene expression in our models. We extend the list of evidences that elevated Notch signal expression is a biomarker signature declaring stem cell prevalence and useful for predicting negative clinical course in glioblastoma. By using functional assays, we validated a first in man tested Notch1 receptor specific antibody as a promising drug candidate in the context of neuro oncology and propose biomarker panel to predict resistance and therapy success of this treatment option. We note that the observed phenotype seems only in part due to Notch1 blockage and the drug candidate leads to activation of off target signals. Further studies addressing a possible emergence of therapy resistance due to WNT activation need to be conducted. We further validated our 3D disease modeling technology to be of benefit for drug development projects.

Glioblastoma (GBM) is the most aggressive malignant primary brain tumour in adulthood. Despite strong research efforts current treatment options have a limited impact on glioma stem-like cells (GSCs) which contribute to GBM formation, progression and chemoresistance. Invasive growth of GSCs is in part associated with epithelial–mesenchymal-like transition (EMT), a mechanism associated with CD73 in several cancers. Here, we show that CD73 regulates the EMT activator SNAIL1 and further investigate the role of enzymatic and non-enzymatic CD73 activity in GBM progression. Reduction of CD73 protein resulted in significant suppression of GSC viability, proliferation and clonogenicity, whereas CD73 enzymatic activity exhibited negative effects only on GSC invasion involving impaired downstream adenosine (ADO) signalling. Furthermore, application of phosphodiesterase inhibitor pentoxifylline, a potent immunomodulator, effectively inhibited ZEB1 and CD73 expression and significantly decreased viability, clonogenicity, and invasion of GSC in vitro cultures. Given the involvement of adenosine and A3 adenosine receptor in GSC invasion, we investigated the effect of the pharmacological inhibition of A3AR on GSC maintenance. Direct A3AR inhibition promoted apoptotic cell death and impaired the clonogenicity of GSC cultures. Taken together, our data indicate that CD73 is an exciting novel target in GBM therapy. Moreover, pharmacological interference, resulting in disturbed ADO signalling, provides new opportunities to innovate GBM therapy.