William Weiss, MD, PhD

Our laboratory is broadly interested in developing and characterizing mouse models that faithfully recapitulate the biology and genetics of human tumors of the nervous system, and using observations in the mouse to inform the biology, genetics, and therapy of human tumors. Specifically we are:

1) Identifying genes, pathways, and small RNA molecules that contribute to tumorigenesis.
2) Studying tumor initiating cells to understand their contribution to malignant progression.
3) Evaluating new targets, therapies, and mechanistic rationales for combining targeted agents.

Targeted expression of EGFR causes oligodendroglioma in transgenic mice. Gliomas are the leading cause of cancer death in children, and the most common primary adult brain tumor. Aberrant signaling through EGFR features prominently in this disease. We generated a mouse model for oligodendroglioma, by over-expressing activated EGFR under control of the S100 beta promoter. These animals and animals developed by others and driven by a GFAP-Ras transgene, are being used to explore cell of origin for glioma and to develop preclinical therapeutic strategies. We demonstrated that Ras drives proliferation in rare cancer-stem-like cells in the subventricular zone, blocks neuronal differentiation, promotes aberrant glial differentiation resulting in glioma that is polyclonal in nature. In contrast, tumors in the v-erbB model arise from abundant oligodendroglial progenitors in the white matter, not from rare cells in the subventricular zone, and do not show blockade of neuronal differentiation. Since EGFR signals in-part through PI3-kinase, we surveyed in glioma cell lines a collection of PI3K inhibitors synthesized by Kevan Shokat’s lab, demonstrating that concurrent inhibition of PI3K alpha and mTOR was required to inhibit proliferation, and that Akt was a dispensable proliferative intermediate in this pathway. Although dual inhibition of PI3K and of mTOR arrested growth of tumors in-vivo, these agents were inefficient in driving tumor cell death. We have further analyzed signaling nodes in the EGFR-PI3K-MAPK pathways, and have identified three novel targets (cdk1, cdk2 and autophagy/lysosomal trafficking), inhibition of which converts cytostatic PI3K inhibitors into cytotoxic agents. We have submitted an IND for a clinical trial in relapsed glioma, combining inhibitors of PI3K/mTOR with inhibitors of autophagy.

Targeted expression of MYCN to generate models of neuroblastoma and medulloblastoma in transgenic mice. Neuroblastoma is the third most common tumor of childhood. The proto-oncogene MYCN is amplified in one third of children with neuroblastoma and correlates with incurable disease. We generated transgenic mice that over-express MYCN in the neural crest and develop neuroblastoma. Genome-wide screens revealed localized gains and losses well aligned with comparable changes in human tumors. Further genetic screening is utilizing Sleeping Beauty transposons and modifier genetics of strain-specific penetrance to identify specific genes that cooperate with MYCN. Murine tumors are highly proliferative and displayed a complex vasculature, expressing numerous angiogenic markers in common with human disease. Murine neuroblastoma tumors mutant at p53 are resistant to chemotherapy, and are now being developed as a model for relapsed drug-resistant neuroblastoma. We also showed that PI3K inhibition destablilizes Mycn protein, leading to decreased tumor burden in-vivo. We have identified a number of tool and clinical inhibitors of PI3K highly effective against Mycn and well tolerated in-vivo, are dissecting mechanisms through which these drugs block Mycn, and lead to involution of tumors, and are negotiating to initiate clinical trials of these agents in neuroblastoma.

We recently re-derived the TH-MYCN animals using the Tet system, allowing us to regulate transgene expression and to image tumor-associated firefly luciferase expression in-vivo. Intriguingly, targeted expression of Mycn to the brains of transgenic mice led to luciferase and Mycn-positive medulloblastoma, a common pediatric brain tumor also driven by MYCN, affording a new opportunity to study activation and inhibition of Mycn in both neuroblastoma and medulloblastoma. Although a minority of human medulloblastoma is driven by aberrant Sonic hedgehog (Shh) signaling, Shh is central to existing mouse models of this disease. Shh signaling is active in a minority of tumors in our MYCN-driven model, suggesting that we can model Shh-dependent and independent disease. This model demonstrates either classic and large cell/anaplastic pathologies, the latter associated with poor survival in children with medulloblastoma. We are characterizing cells of origin and preclinical therapeutics directed against Mycn protein, and are initiating forward genetic screens to identify metastases genes in this model.