BCOR-PRC1.1复合物的表观遗传重塑决定AML对IDH抑制剂的应答与耐药

Background: IDH1/2 mutations occur in about 20% of AML and generate 2-hydroxyglutarate, which blocks TET-mediated hydroxymethylation, induces CpG island hypermethylation, and enforces a differentiation arrest. Although mutant-specific IDH inhibitors induce remissions, relapse is common. Our prior study identified recurrent acquisition of BCOR Loss of Function mutations at resistance. Because BCOR/PRC1.1 recruitment is regulated by CpG methylation, we hypothesized that IDH-driven hypermethylation disrupts BCOR binding, that IDH inhibitor mediated demethylation restores PRC1.1 occupancy, and that relapse associated BCOR loss prevents this restoration. Methods: Using isogenic TF-1 IDH2-R140Q models, we performed EvoC 6-base sequencing, BCOR ChIP-seq, and RNA-seq across WT, IDH2 mutant, and enasidenib treated cells. Published CRISPR-Cas9 screening data from IDH1 mutant AML (Liu et al. Cancer Research 2022) were reanalyzed to define major key genes in IDH-inhibitor-induced differentiation. Results: IDH2 mutation induced coordinated promoter hypermethylation and 5hmC loss. EvoC sequencing showed that about 60% of DMRs were hypermethylated and enriched upstream of transcription start sites. 5hmC was markedly reduced and fully restored by enasidenib, consistent with impaired TET2 activity. Notably, 87% of hypermethylated DMRs with 5hmC loss underwent demethylation after 28 days of treatment, coinciding with normalized 2HG. BCOR occupancy decreased at promoter-proximal regions in IDH2 mutant cells and was restored by enasidenib.Integration of methylation, BCOR ChIP-seq, and RNA-seq showed that most BCOR-loss promoters were hypermethylated, transcriptionally activated, and normalized with treatment. Restored BCOR targets were enriched for Wnt, PI3K/Akt/mTOR, RTK, and MAPK pathways. Promoters of JUNB and WNT11 showed concordant hypermethylation, reduced BCOR binding, and increased expression in IDH2 mutant cells, all reversed by enasidenib.CRISPR-Cas9 data identified Bcor as a top gene required for differentiation upon IDH inhibition, exceeding canonical regulators such as Cebpa and Cebpb. Bcor targets ( Junb and Wnt11 ) also ranked as essential effectors in the differentiation arm, matching our multi-omics findings. Conclusions: IDH2 mutation establishes an epigenetic blockade marked by promoter hypermethylation, 5hmC loss, and loss of BCOR-PRC1.1 occupancy, enabling aberrant activation of stemness programs. Enasidenib reverses these defects by restoring 5hmC, promoter methylation balance, BCOR binding, and PRC1.1-mediated repression. These data support a model in which IDH-inhibitor-induced differentiation requires intact BCOR function and repression of BCOR targets such as JUNB and WNT11, defining how IDH-mutant epigenetic dysregulation and BCOR loss converge to limit therapeutic efficacy.