SUMMARY Cerebral arteriovenous malformations (AVMs) are the most common vascular malformations and the leading cause of hemorrhagic strokes. Past studies have demonstrated an important role of endothelial cells (ECs) in cerebral AVMs, and shown that the maintenance of endothelial integrity by bone morphogenetic protein (BMP) and Notch signaling is critical for cerebral vascular formation. However, it is unclear how disturbed crosstalk between BMP and Notch signaling affects EC differentiation at transcriptional regulatory level causing cerebral AVMs. In this proposal, we aim to unearth that the crosstalk between BMP and Notch signaling induces histone deacetylase 2 (HDAC2) to shift the transcriptional landscape of ECs toward ill-fated differentiation causing cerebral AVMs. We will also define if HDAC2 inhibition prevents this ill-fated cell shift and improves cerebral AVMs. In preliminary study, using a new mouse model, we find a striking shift of ECs to mesenchymal-like cells in cerebral AVMs and show that these mesenchymal-like cells cause arteriovenous shunts. Utilizing single-cell RNA sequencing and connectivity Map, we identify HDAC inhibition to prevent ECs from mesenchymal cell differentiation and significantly reduce cerebral AVMs. In human and mouse cerebral AVMs, we find a specific HDAC2 induction. We show that HDAC2 induction alters specific histone modifications, which are responsible for the shift of ECs to mesenchymal cell differentiation. Endothelial-specific deletion of HDAC2 prevents this ill- fated cell shift and reduces cerebral AVMs. Similar results of dysregulated HDAC2 with its downstream effects are also found in cerebral AVMs of hereditary hemorrhagic telangiectasia type 1 (HHT1) and type 2 (HHT2), but not in juvenile polyposis/HHT. Furthermore, we find that HDAC2 is specifically induced in cerebral AVMs by excess BMP through delta-like protein 3 (Dll3) and Notch1 signaling. We uncover that lack of matrix Gla protein (MGP) allows BMP-8b to elevate staphylococcal nuclease domain-containing protein 1 (SND1), which is required for Notch signaling to induce HDAC2 in cerebral AVMs. We hypothesize that HDAC2 induction, downstream of excess BMP and Notch signaling, alters specific histone modifications to shift ECs toward ill-fated differentiation causing cerebral AVMs. In specific Aim 1, we will determine how HDAC2 is dysregulated by the crosstalk between BMP and Notch signaling and shifts endothelial differentiation in cerebral AVMs. In specific Aim 2, we will determine the contribution of HDAC2 induction to human cerebral AVMs. In specific Aim 3, we will determine if limiting HDAC2 improves cerebral AVMs. If successful, the obtained information will provide new insight into the mechanism of AVMs, and HDAC2 inhibition may emerge as a novel therapeutic approach for cerebral AVMs.