Medulloblastoma (MB) is a cancer arising within the cerebellum, and is the most common malignant brain tumor in children. Current care which relies on radiotherapy and chemotherapy, only cures 60% of patients, and has serious long-term neurocognitive sequelae in survivors, hence the unmet medical need to develop more efficacious, less toxic therapies for this disease. Malignant brain tumors including the sonic hedgehog (SHH) subgroup of MB are driven by tumor propagating cells (cancer stem cells; CSCs) expressing the cell surface glycoprotein CD15. A knowledge gap exist in that the molecular mechanisms controlling CSC and non CSC phenotypes are poorly delineated. The central hypothesis for this proposal is that the PI3K/PTEN signaling axis exerts control over the CSC phenotype in SHH MB and hence represents a target for therapeutic exploitation with PI3K inhibitors. Importantly, work from our laboratory has demonstrated that PTEN expression is suppressed in the CSC compartment of the Smo transgenic (SHH pathway) mouse model of MB. The clinical relevance of the PTEN deficient Smo transgene mouse model is highlighted by our finding that PTEN expression is lost in human MB, and the observation that PTEN loss correlates with poor survival in MB patients. Our grant proposal, guided by our preliminary data proposes 3 specific aims: 1) Determine if genetic alterations in PTEN or specific isoforms of PI3K are necessary and/or sufficient for SHH driven CSC phenotype and tumorigenesis 2) Utilize high throughput kinome library screens and genomic profiling of mouse and human MB CSCs and non CSCs to identify a molecular signature and synthetic lethality effects which will predict sensitivity or resistance to a PI-3K inhibitor. 3) Elucidate the epigenetic mechanism by which PTEN is regulated in the CSC compartment in human SHH PDX models and the SmoTg GEMM model; Focused on PTEN specific miRNA regulatory networks. We will use the Math1creER recombinase to generate the conditional deletion of PTEN, p110 and/or p110 in normal cerebellar granule neuronal precursor cells (GNPs) or in the CSC compartment and determine the effects on the development of MB and CSC phenotype. In Aim 2, we will expand upon our recent report demonstrating the activity of the PI3K inhibitor, BKM120 against CSCs to include synthetic lethality screens with other targeted agents and conventional chemotherapeutics. An innovative component of our application is the recent discovery of a novel miRNA network which epigenetically regulates PTEN in the CSC compartment which is associated with the CSC phenotype. This observation will be explored in our GEMMs , MB PDX models and in human MB tumor samples to determine the role of these miRNAs in the regulation of PTEN, PI3K and the CSC phenotypes including tumor initiating activity. The proposed work is significant in that it will: 1) establish a role for PTEN, p110 and p110 in CSC biology in MB 2) identify novel synthetic lethalities in combin...