Phosphorylated DEK sustains leukemia stem cells by enabling PBX3-driven transcriptional reprogramming

Leukemia stem cells (LSCs) drive acute myeloid leukemia (AML) initiation, relapse, and chemoresistance, yet the core post-translational events sustaining LSC maintenance remain poorly defined. Here, through phosphoproteomic profiling of normal hematopoietic stem and progenitor cells (HSPCs) versus LSC-enriched populations, we identify DEK phosphorylation as a critical modification during leukemogenesis. Functional studies in MLL-AF9- and HOXA9/MEIS1-driven AML mouse models, as well as patient-derived xenografts (PDXs), demonstrate that DEK deficiency impairs LSC maintenance and AML progression. Moreover, DEK deletion enhances LSC chemosensitivity to the standard-of-care combination of azacitidine and venetoclax (Aza/Ven), whereas DEK overexpression confers robust chemoresistance. Mechanistically, DEK recruits the transcription factor GABPA to upregulate the transcriptional cofactor PBX3, a key oncogenic driver in AML, thereby sustaining the leukemogenic transcriptional program. This DEK-GABPA interaction strictly depends on DEK phosphorylation at Ser301/303/306/307 (the 4S sites), which stabilizes the conformation of the DEK-GABPA complex. We identify casein kinase 2 (CK2) as the upstream kinase that directly phosphorylates DEK-4S sites. Importantly, blockade of DEK phosphorylation via 4S site mutagenesis or treatment with the clinical-stage CK2 inhibitor CX-4945 selectively depletes LSCs while sparing normal HSPCs. Furthermore, combining CX-4945 with venetoclax promotes LSC apoptosis and represses the PBX3-driven leukemogenic transcriptional program, exhibiting synergistic anti-AML effects both in vitro and in vivo. Collectively, our findings uncover a previously unrecognized phosphorylation event (DEK-4S phosphorylation) that sustains LSCs and establish the CK2-DEK axis as a promising LSC-specific therapeutic strategy for AML.