Developmental Origin, Injury and Epigenomic Regulation of NF1-Associated Peripheral Nerve Sheath Tumors

Project Abstract Individuals with neurofibromatosis type 1 (NF1) have an approximately 160-fold increased risk of developing malignant peripheral nerve sheath tumor (MPNST). As a leading cause of death for NF1 patients. MPNST has no effective therapy and thus there is an urgent need for new therapies. The dramatically increased risk of

developing MPNSTs is caused by the presence of plexiform neurofibromas (PNFs), the major benign precursor lesion for NF1-MPNST. It has been proposed that PNFs are congenital lesions, arising from the early stages of nerve development when neural-crest stem cells differentiate into Schwann cell (SC) lineages, which give rise

to either myelinating or nonmyelinating SCs (mSCs or nmSCs). In the normal nerve, unmyelinated axons are sorted and ensheathed by nmSCs into individual pockets, forming Remak bundles. Whereas no defect in SC precursors or mSCs was observed, Nf1 loss (Nf1-/-) during early nerve development induced a pocket defect in

Remak bundles, characterized by abnormally sorted unmyelinated axons. These abnormal Remak pockets progress to a stage with axonal degeneration and abnormal proliferation of dissociated SCs, leading to the formation of PNFs. Axonal degeneration may contribute to PNF formation by inducing a nerve injury

environment, a concept supported by the observation that Nf1 loss in mature SCs is not sufficient to induce PNFs unless an injury to the nerve occurs. Malignant transformation of PNFs to MPNSTs requires at least two additional genetic alterations: sequential inactivation of CDKN2A, and then either SUZ12 or EED - two

essential components of Polycomb Repressive Complex 2 (PRC2). PRC2 catalyzes histone modification H3K27me3 to repress gene expression throughout the genome. Loss of PRC2 specifically observed in MPNSTs, but not in benign tumors, suggests that PRC2-mediated H3K27me3 normally represses expression of the oncogenic drivers responsible for malignant transformation of PNFs to MPNSTs. However, Eed/PRC2 is

dispensable during normal mouse SC development and myelination. Further, loss of the Eed/PRC2 tumor suppressor unexpectedly inhibits proliferation of injury-induced reprogrammed SCs, accompanied by derepression of Cdkn2a expression. Here, we propose to test two related hypotheses: (1) the developmental

Nf1-/- Remak pocket defect and its associated axonal degeneration (nerve injury) drive Nf1-/- SCs to form PNFs and (2) nerve injury response induces an epigenomic switch, rendering reprogrammed PNFs or SCs susceptible to malignant transformation by sequential loss of CDKN2A and PRC2. We will determine the role

of the developmental Remak pocket defect in NF1-MPNST formation (Aim 1), investigate tumor suppressive mechanisms in injury-induced reprogramed SCs (Aim 2), and develop therapeutic strategies based on the injury-induced epigenomic switch in reprogrammed SCs (Aim 3). We will identify injury-induced and PRC2-

repressed oncogenic drivers for MPNST formation via epigenomic approaches, develop synergistic therapies using a high-throughput drug repurposing screen, and test them GEM- and patient-derived preclinical models.