EZH2 Mutations Are Drivers of Clonal Hematopoiesis and Leukemic Transformation in a Mouse Model of Primary Myelofibrosis

Primary Myelofibrosis (PMF) is a myeloproliferative neoplasm characterized by aberrant myeloid differentiation, associated with disruption of the bone marrow niche with subsequent fibrosis development and a high risk of leukemic transformation. The phenotypical complexity observed in PMF likely reflects the heterogeneous mutation profile of the neoplastic stem cells driving the disease. In our former work, we identified a CD133+ hematopoietic stem / progenitor cell (HSPC) population from patient peripheral blood that can drive major PMF morbidity parameters in a xenotransplantation mouse model. Mutational analysis of the JAK2 locus at the single cell level within the CD133+ population showed highly variable levels of cells with a JAK2^+/+, JAK2^V617F/+, or JAK2^V617F/V617F genotype, indicating that clonality is unlikely driven by JAK2 mutations. In two of these patient samples, and in a third patient sample with CALR-fs* mutations, we identified a high load of missense mutations in EZH2 (45 to 95%), suggesting they may be critical for the clonal expansion of the neoplastic stem cell compartment.

EZH2 mutations are found in circa 7% of PMF patients and are correlated with poor prognosis. EZH2 is a critical enzymatic subunit of the Polycomb Repressor Complex 2, which initiates gene repression of select genes through its intrinsic activity for methylating lysine-27 of histone H3 (H3K27). To date, the exact contribution of EZH2 mutations to PMF evolution or AML transition has not been clarified. CD133+ HSPC carrying EZH2 mutations either with JAK2 or CALR mutations were transplanted into immunodeficient NOD-scid-gamma (NSG) mice. Mice engrafted with patient samples carrying either EZH2-Y633C and JAK2-V617F or EZH2-Y733* and CALR-fs* mutations showed a strikingly similar phenotype, including high human cell engraftment (10-20%), skewed myelopoiesis, dysplastic human megakaryocytes, splenomegaly, anemia, and fibrosis in either the BM or spleen. In the case of xenotransplanted mice receiving CD133+ cells with a low JAK2 burden and EZH2-D265H mutations, we observed the highest engraftment in our mouse model (62-95%) and in one case AML transition with >50% CD133+ human blasts in murine bone marrow. Notably, AML arose from a CD133+ EZH2^D265H/+ cell that lacked JAK2^V617Fmutation.

We thus conclude that EZH2 mutations confer to CD133+ neoplastic stem cells a predisposition to clonal aberrant hematopoiesis; whereas acquisition of JAK2^V617F or CALR mutations likely leads to the observed myeloproliferation and disruption of megakaryocytic and erythroid regulation . Moreover, our results demonstrate that epigenetic mutations (like EZH2^D265) and not JAK2^V617F are critical for AML transition. Our data underscore the importance of post-transcriptional modifiers of histones in altering the epigenetic landscape of neoplastic stem cells, whose clonal growth sustains aberrant myelopoiesis and expansion of pre-leukemic clones.