Pilocytic astrocytomas (PAs) and disseminated diffuse leptomeningeal glioneuronal tumors (DLGTs) are two types of brain cancer common among pediatric patients. PAs are low grade gliomas, generally presenting as non-infiltrating tumor masses, but their anatomical location can have profound consequences, with symptoms ranging from pressure headaches, cranial nerve defects, ataxia, loss of visual acuity, diabetes, and precocious puberty. Surgery is the treatment of choice for these patients, although radical resection is not always possible, In these cases, adjuvant radiation and/or chemotherapy is often administered with acute and long-term toxicities that can be debilitating in young patients. In fact, although the majority of patients have a good prognosis in terms of long-term survival following surgical resection, approximately 50% of patients suffer from morbidity due to recurrence or therapy-related side effects, making PA a disease with an unmet need for better therapeutic options. DLGTs, although less common than PAs, are even more challenging clinically due to diffuse leptomenigial infiltration which precludes surgical intervention, leading to higher mortality rates approaching 80%. By far the most common genetic event?observed in nearly 70% of cases of both PA and DLTG?is a recurrent tandem duplication on chromosome 7. As a consequence of this rearrangement, the N-terminal portion of KIAA1549 becomes fused to the C-terminal portion of BRAF, which includes the kinase domain. Loss of BRAF?s N-terminal regulatory domain in turn, results in constitutive dimerization and downstream signaling in a RAS-independent manner. To generate an accurate mouse model of human cancer driven by complex chromosomal rearrangements, our laboratory has recently developed a novel CRISPR-based approach to induce specific chromosomal rearrangements in vitro and, more importantly, in vivo . The CRISPR-Cas9 system is ideally suited for in vivo genome editing because it only requires co-expression of Cas9 and an appropriately designed RNA molecule (sgRNA) to guide the bacterial endonuclease Cas9 to the desired cut site. The method we have developed is based on the simultaneous expression of Cas9 and 2 sgRNAs designed to cleave at the desired breakpoints. As a proof of concept, we have demonstrated the feasibility of this strategy by generating novel mouse models of EML4-ALK fusion driven lung adenocarcinomas and BCAN-NTRK1 fusion-driven brain cancer. Encouraged by these successes, we propose to use in vivo chromosomal engineering to model the KIAA1549:BRAF rearrangement in the mouse brain (Aim 1). We have already obtained a large body of preliminary data that demonstrate the feasibility of this approach. Ex vivo generation of the KIAA1549:BRAF fusion in adult neural stem cells produces tumors upon orthotopic injection that have characteristics of human DLTG. Efforts to faithfully recapitulate PA pathology will also be explored using novel in vivo CRISPR-Cas9 modeling at early developmental stages. We will use our models of pediatric glioma to investigate the molecular mechanisms through which KIAA159-BRAF promotes tumor initiation and progression (Aim 2). Finally, we will use this model to directly test the therapeutic potential of a novel BRAF inhibitor (Aim 3).
A subset of tumors of the central nervous system that are found in children and adults is driven by a genetic event in which chromosomes are rearranged in cells that initiate these tumors. This rearrangement produces a fusion protein not found in normal cells that can drive the rapid growth and expansion of cells which harbor this new protein. Chromosome rearrangements are difficult to model in mice but we have devised a protocol using a new gene editing technique (called CRISPR) that precisely models one such fusion event and recapitulates in mice the disease seen in humans. In this proposal we will use this mouse model to understand the mechanistic basis of this tumor type and develop novel therapeutic strategies based on rational drug design.