Gene signatures associated with mouse postnatal hindbrain neural stem cells and medulloblastoma cancer stem cells identify novel molecular mediators and predict human medulloblastoma molecular classification

Medulloblastoma arises from mutations occurring in stem/progenitor cells located in restricted hindbrain territories. Here we report that the mouse postnatal ventricular zone lining the IV ventricle also harbors bona fide stem cells that, remarkably, share the same molecular profile with cerebellar white matter-derived neural stem cells (NSC). To identify novel molecular mediators involved in medulloblastomagenesis, we compared these distinct postnatal hindbrain-derived NSC populations, which are potentially tumor initiating, with murine compound Ptch/p53 mutant medulloblastoma cancer stem cells (CSC) that faithfully phenocopy the different variants of human medulloblastoma in vivo. Transcriptome analysis of both hindbrain NSCs and medulloblastoma CSCs resulted in the generation of well-defined gene signatures, each reminiscent of a specific human medulloblastoma molecular subclass. Most interestingly, medulloblastoma CSCs upregulated developmentally related genes, such as Ebfs, that were shown to be highly expressed in human medulloblastomas and play a pivotal role in experimental medulloblastomagenesis. These data indicate that gene expression analysis of medulloblastoma CSCs holds great promise not only for understanding functional differences between distinct CSC populations but also for identifying meaningful signatures that might stratify medulloblastoma patients beyond histopathologic staging. SIGNIFICANCE: The functional and molecular comparison between the cell progenitor lineages from which medulloblastoma is thought to arise and medulloblastoma CSCs might lead to the identification of novel, potentially relevant mediators of medulloblastomagenesis. Our findings provide a rationale for the exploitation of mouse CSCs as a valuable preclinical model for human medulloblastoma, both for the definition of CSC-associated gene signatures with predictive mean and for the identification of therapeutically targetable genes.