View Program - 63rd Annual Drosophila Research Conference

Central nervous system (CNS) tumors are the leading cause of solid tumor death in children. Among the deadliest and most common pediatric brain tumors are diffuse intrinsic pontine glioma (DIPG) and posterior fossa ependymoma type A (PFA). Two key molecular events drive nearly all cases of these two cancers: a lysine-to-methionine mutation at residue 27 on histone H3 (H3 K27M) in DIPG or elevated expression of the previously uncharacterized protein EZHIP in PFA. These tumors share striking similarities, including a near-complete loss of histone H3 trimethylation at lysine 27 (H3K27me3), a mark that contributes to transcriptionally silent chromatin. We have previously shown that H3 K27M and EZHIP are potent inhibitors of the H3K27 histone methyltransferase Polycomb repressive complex 2 (PRC2) in mammalian cell culture. Despite this, low levels of residual H3K27me3 remain at sites of initial PRC2 recruitment. These data leave open the possibility that oncogenesis is the result of transcriptional repression maintained at specific genes, aberrant activation of other genes that lose the repressive H3K27me3 mark, or a combination of both. To begin to address this question, we have modeled these cancers by expressing both oncoproteins in Drosophila melanogaster, the organism in which PRC2 was originally identified. Tissue-specific expression of either EZHIP or H3 K27M causes detrimental phenotypes and the loss of H3K27me3. We have leveraged these Drosophila models to screen 438 conserved, chromatin-regulatory genes for suppression or enhancement of the phenotypes. Our RNAi screen has identified over 50 genes whose knockdown modifies the phenotype caused by H3 K27M in the wing. Remarkably, these suppressors restore normal development despite the continued loss of H3K27me3. Shared features of these suppressors suggest that restoring normal development requires a precise balance between the repressive H3K27me3 and marks of active chromatin at gene regulatory elements. Ongoing studies will determine whether these emerging mechanisms can be leveraged to prevent or even reverse the phenotypes of these potent oncoproteins.

Imbalances in active and repressive chromatin states underlie phenotypes caused by the oncoproteins H3 K27M and EZHIP