Abstract The lack of genetically accurate models that recapitulate the genetics of human tumors hampers the development of new treatments for rare tumor types. One strategy to create an accurate model would be to use stem cells from a cancer's tissue of origin and add key oncogenic driver elements in combinations that are seen in tumors. These elements should also drive oncogenic transformation relevant to that particular tumor type. To validate this strategy, we used human neural stem and progenitor cells derived from the cerebellar anlage to develop a model of Group 3 MYC-driven medulloblastoma, a rare and frequently fatal pediatric brain tumor. The cells were transduced with c-MYC, dominant-negative p53, constitutively active AKT and hTERT, key oncogenic driver elements associated with aggressive medulloblastoma and injected into moice as orthotopic xerographs. These cells formed tumors in the brains of mice and resulted in decreased survival to a median of 117 days. These tumors histologically, pathologically, and genetically resembled the MYC-driven Group 3 medulloblastoma tumors. In contrast, human neural stem cells immortalized with SV40 did not form tumors in mice. To facilitate rapid screening of a large number of available drugs, we developed an in silico analysis technique, DiSCoVER (Disease-model Signature vs. Compound-Variety Enriched Response) that uses the expression profile of the human neural stem cell models and existing drug sensitivity databases to identify novel therapeutic targets. The DiSCoVER analysis predicted aggressive medulloblastoma would be sensitive to CDK inhibitors. Among the top hits were the CDK4/6 inhibitor palbociclib. Indeed, treatment of Palbociclib in human neural stem cell medulloblastoma model and high-serum medulloblastoma cell lines led to decreased proliferation by an average of 62% as measured by MTS assay compared to vehicle control (p<0.01). Palbociclib treatment also increased the percentage of apoptotic cells by an average of 150% as measured by cleaved caspase-3 immunofluorescence and cleaved-PARP western blot compared to vehicle control (p< 0.04). Treatment with palbociclib significantly extended the survival of mice with orthotopic medulloblastoma xenografts (p = 0.003). We validated our human neural stem cell model of Group 3 medulloblastoma and show that this model, as well as patient-derived high MYC medulloblastoma models, are sensitive to CDK4/6 inhibition in vitro and in vivo. Our results strongly suggest that Palbociclib can be an effective treatment for poor-prognosis MYC-driven medulloblastomas in carefully selected patients (Group 3/C1). In summary, we have presented here a new method of generating genetically accurate models for rare tumors using region-specific human stem and progenitor cells and a novel analysis technique to find therapeutics to target them. Citation Format: Allison Rose Hanaford, Tenley C. Archer, Antoinette Price, Ulf Kahlert, Jarek Maciaczyk, Nikkhah Guido, Wlliam Kim, Tobias Ehrenberger, Paul Clemons, Vlado Dancik, Brinton Seashore-Ludlow, Vasanthi Viswanathan,, Michelle Stewart, Matthew Rees, Alykahn Shamji, William Hahn, Stuart Schreiber, Ernest Fraenkel, Scott Pomeroy, Jill Mesirov, Pablo Tamayo, Charles G. Eberhart, Eric H. Raabe. DiSCoVERing innovative therapies for rare tumors: Combining genetically accurate disease models with advanced in silico analysis to identify novel therapeutic targets. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 2476.
Read full abstract