Abstract
Myeloid malignancies, including myeloproliferative neoplasms (MPN), myelodysplastic syndromes, and acute myeloid leukemia (AML), are predominantly sporadic diseases caused by the progressive accumulation of somatic mutations. However, rare inherited predispositions have been identified and account for familial clustering. We previously identified a germline tandem duplication at 14q32 (CNV) involving ATG2B and GSKIP in a very large family from West Indies (>60 cases). This CNV predisposes heterozygous carriers to a broad spectrum of myeloid hematologic malignancies with a nearly complete penetrance. The CNV is associated with early-onset clonal hematopoiesis, frequently involving TET2 or JAK2V617F mutations, and follows distinct evolutionary trajectories: either progressing from essential thrombocythemia to myelofibrosis and AML or transforming directly into AML. While ATG2B is a key regulator of autophagy and mitochondrial homeostasis, and GSKIP modulates GSK3β and PKA signaling pathways that intersect with metabolism, proliferation, and stem cell maintenance, the precise mechanisms by which their overexpression drives hematopoietic transformation remain unclear.
To dissect the functional impact of the CNV on hematopoietic stem cells (HSCs), we combined in vitro and in vivo approaches. We overexpressed ATG2B/GSKIP in human hematopoietic cell lines and employed induced pluripotent stem cells (iPSCs) derived from CNV carriers. In parallel, we generated a knock-in (KI) mouse model mimicking the CNV by trans-allelic targeted meiotic recombination and assessed its effects on hematopoiesis, both alone and in combination with Jak2V617F or Tet2 knockout (KO).
Overexpression of ATG2B/GSKIP in hematopoietic cell lines led to mildly reduced proliferation. Transmission electron microscopy revealed marked mitochondrial abnormalities, including increased mitochondrial size and number, highly developed endoplasmic reticulum, and aberrant lipid droplets. Consistently, ATG2B/GSKIP overexpression increased mitochondrial mass and impaired mitochondrial fission, as indicated by elevated TOM20 levels and increased [S637] DRP1 phosphorylation. Seahorse assays demonstrated a 2-fold reduction in basal and maximal respiration, glycolytic reserve, and mitochondrial ATP production, while glycolytic ATP production was preserved. LC-MS/MS metabolomics revealed a selective and profound effect on α-ketoglutarate (αKG) levels, which led to activation of TET family demethylases, suggesting a link between metabolic reprogramming and epigenetic dysregulation. Notably, similar mitochondrial defects and impaired megakaryocyte differentiation were observed in megakaryocytes derived from human iPSCs carrying the CNV.
To assess the in vivo impact of the 14q32 CNV on hematopoiesis, we analyzed KI mice harboring the CNV, which showed mild leukocytosis without overt hematological abnormalities. Lin- progenitors displayed mitochondrial abnormalities and impaired HSC cell cycle entry, as assessed by Ki67/DAPI staining. Transcriptomic analysis of Lin- progenitors confirmed overexpression of Atg2b and Gskip and revealed signatures of mitochondrial dysfunction,along with suppression of key proliferative pathways including cell cycle, translation, RAS, and JAK/STAT signaling, hallmarks of metabolically repressed HSC state. In both competitive and serial bone marrow transplantation assays, CNV HSCs showed reduced repopulation capacity and impaired fitness.
We next examined cooperation between the CNV and MPN drivers. The CNV significantly altered the phenotypes of Jak2V617F and Tet2 KO models: polycythemia was attenuated in CNV× Jak2V617F mice, while platelet and leukocyte counts increased, and thrombocytopenia was alleviated in CNV×Tet2 KO animals. Notably, despite intrinsic fitness defects, CNV×Jak2V617F HSCs displayed enhanced long-term clonal expansion in competitive transplantation with CNV hematopoietic cells only, demonstrating that Jak2V617F confers a strong clonal advantage to HSC in a CNV context.
Overall, our findings indicate that the 14q32 CNV disrupts HSC homeostasis through mitochondrial dysfunction, inhibition of JAK/STAT signaling, and αKG-driven epigenetic remodeling. The ensuing HSC dysfunction may create a permissive state for leukemogenesis by promoting the selection and expansion of clones with both signaling (JAK2V617F, CALR, MPL, RAS) and epigenetic (TET2 and IDH1/2) mutations.
