Cdk6 targeting reverses leukemia in germline cebpa Q83X AML mouse model with CMP-origin transformation

Abstract

Abstract CEBPA mutations are frequent drivers of familial AML (FAML) and common in sporadic disease. While bZIP-in-frame mutations (bZIP-if) confer favorable risk, relapse remains common within five years, and patients with N-terminal mutations (N-ter) exhibit inferior survival. Germline CEBPA N-ter mutations carry an 80.36% lifetime AML risk. To model this pathogenesis, we established the first conditional CebpaQ83X knock-in mouse, based on a familial pedigree (4 generations, 11 AML cases) with hereditary N-ter mutation (NM_004364.5: c.247C>T; p.Q83*). This model recapitulated human AML progression, enabling mechanistic investigation of CEBPA -driven leukemogenesis. Unexpectedly, both prenatal and postnatal induction of CebpaQ83X caused significant reductions in neutrophils and monocytes in mice by 6 weeks, while hemoglobin and platelet levels remained normal. Analysis of this pre-leukemic marrow (6 weeks) revealed a myeloid differentiation arrest, characterized by expanded CMPs and LT-HSCs alongside depleted GMPs and mature myeloid cells. Serial colony assays at this stage confirmed enhanced progenitor self-renewal and impaired myeloid commitment. Furthermore, transplantation of 6-week marrow demonstrated multilineage reconstitution failure (myeloid/T/B) with a persistent CMP/GMP imbalance. However, by 4-6 months, 83% of mice developed myeloid expansion, marked by a 2- to 20-fold increase in neutrophil/monocyte-like and GMP-like cells. Within this cohort, 33% spontaneously progressed to lethal AML exhibiting leukocytosis, anemia, thrombocytopenia, and blasts phenotypically matching human FAML. Transplantation of marrow from these 6-month mice provoked rapid AML development within 1.5-2 months, characterized by >90% GMP-like blasts. scRNA-seq uncovered CMP dominance in pre-leukemic and leukemic HSPCs of CebpaQ83X mice. Strikingly, leukemic cells exhibited lineage infidelity—transcriptionally CMP-like despite GMP surface phenotype. Critically, integrated single-cell multi-omics (RNA-seq/targeted-seq) and WES in 27 CEBPA-mutant AML patients confirmed leukemic mutations enriched in CMP-like populations. This resolves the transformation stage controversy, establishing CMP—not GMP—as the origin for CEBPA-mutant AML. Paradoxically, CebpaQ83X LT-HSCs functioned as primed pre-malignant reservoirs—exhibiting hyperproliferation (Ki67/EdU staining, scRNA-seq) and elevated C/ebpα protein despite multilineage reconstitution failure (whole BM/sorted LT-HSC transplants). Multi-omics integration (ATAC-seq/RNA-seq on CMPs; scRNA-seq subclustering) identified Cdk6 as a druggable vulnerability: upregulated with open chromatin, hub-leveraging leukemogenic pathways (cell cycle/transformation). Targeting CDK6 with palbociclib potently suppressed leukemic progenitors and clonogenicity exclusively in CebpaQ83X cells, demonstrating genetic dependency. This study establishes germline CebpaQ83X drives leukemogenesis through a three-stage pathogenesis: hyperproliferative, C/ebpα-hi LT-HSCs serve as pre-malignant reservoirs with impaired differentiation capacity; malignant initiation at CMP stage despite GMP-like surface phenotype—resolving the cellular origin paradox; epigenetic rewiring positions Cdk6 as a proliferation hub in arrested CMPs. Critically, palbociclib suppresses LT-HSC/CMP proliferation and leukemic outgrowth, providing the first mechanistic rationale for CDK6-targeted therapy in CEBPA N-ter mutant AML.