�鶹��

Pediatric hindbrain tumors such as posterior fossa ependymoma group A (PFA-EPN) are among the most aggressive childhood brain cancers. Unlike most cancers, they arise with strikingly stable genomes but show catastrophic failure of the epigenetic systems that preserve neural cell identity. In healthy development, early neural stem cells permanently silence germline-specific genes—programs normally active only in sperm and egg precursors. 

In PFA-EPN, a germline gene EZHIP (CXORF67) is a central driver: almost universally expressed in PFA-EPN, it mimics the oncohistone H3K27M, inhibits PRC2, and erases the repressive H3K27me3 mark required for neural lineage stability. Yet, the molecular triggers of EZHIP activation in brain cells, and how this leads to disease through chromatin reorganization, remain unknown.This project will engineer tailored, physiologically relevant models of early hindbrain development using human induced pluripotent stem cells (iPSCs) and fetal-derived neural stem cells patterned to hindbrain identity. We will integrate cutting-edge genomics (Region Capture Micro-C, single-CpG methylation mapping) with high-density CRISPRi perturbations, single-cell RNA sequencing, and high-content imaging. 

This approach will directly link 3D genome architecture, DNA methylation, and transcriptional bursting to chromatin compaction, nuclear topology, and lineage fidelity.By systematically mapping and functionally testing the “regulatory gate-breaks” that permit germline programs to invade neural identity, we aim to define shared mechanisms across germline-driven neural pathologies and PRC2-disruptive cancers. The project has the potential to open new therapeutic avenues for PFA-EPN and related pediatric brain tumors that currently lack targeted, less toxic treatments.