Cell of Origin and Cancer Stem Cell Phenotype in Medulloblastomas

The goal of this project is to test our hypothesis that cellular context in which initiating oncogenic event occurs may have a dominant role over specific oncogene function in determining the molecular phenotype of each tumor. To test this hypothesis, we originally proposed to transform neural stem cells (NSCs) and neural progenitor cells (NPCs) in vivo by expressing an activated form of Notch1 (N1ICD) or deleting XRCC2. However, a reviewer suggested that we replace Xrcc2 deletion with a more clinically relevant oncogene, and we chose to make new medulloblastoma models with oncogenic PIK3CA (PIK3CAH1047R). However, this change caused major delay in our progress since we have not been able to generate any tumors with PIK3CAH1047R expression. We successfully intercrossed PIK3CAH1047R (a frequent mutant allele of PIK3CA observed in human cancer) to Sox2CreER, Atoh1-creER, andp53-/- strains to generate PIK3CAH1047R;Sox2-creER;p53-/- and PIK3CAH1047R;Atoh1-CreER;p53-/- mice. We did not observe any medulloblastomas from these crosses before my laboratory moved to Houston Methodist Research Institute in Sept 2016. We had to restart the crosses here and while we could observe megacephaly in PIK3CAH2017R transgene-expressing brains, we did not observe any medulloblastomas. As a backup, we started generating a YAP-induced medulloblastoma models, and we successfully collected 5 samples from YAPS5A;Atoh-CreER medulloblastomas. When YAPS5A is expressed in NSC in developing brain, it caused major developmental defects and the embryos died in utero. We are now crossing YAPS5A to hGFAPCreER mice to activate the transgene in postnatal NSCs. We will be able to perform the final tumor comparison analyses once we have these tumors.