ABSTRACT: Enhancer of Zeste Homolog 2 (EZH2) functions as the catalytic component of the Polycomb Repressive Complex 2, a chromatin modifying complex, which mediates methylation of lysine 27 on histone 3 (H3K27me3), a repressive chromatin mark. In cancer, genetic alterations in EZH2 encompass amplifications, deletions and point mutations, and of particular interest are the recurring mutations in the SET domain at tyrosine 641 (Y641). Hematologic malignancies exhibit a prominent association with EZH2Y641 mutations, primarily observed in B cell malignancies such as germinal center diffuse large B cell lymphoma (DLBCL) and follicular lymphoma (FL) with a mutation frequency around 20%. Leukemias and solid tumors like melanoma, prostate, head and neck, sarcomas, and others demonstrate substantially lower EZH2Y641 mutation frequencies. Because the effect of Ezh2Y641F mutations depend on the existing patterns of H3K27 mono-, di-, and tri-methylation, we hypothesized that the oncogenic potential of Ezh2Y641F mutations depends on the developmental stage and cell of origin during which mutations occur. Towards that end, we employed a conditional and inducible mouse model of Ezh2Y641F, crossed with different tissue-specific Cre drivers. We found that heterozygous expression of Ezh2Y641F early in embryonic development, at birth, or in adult hematopoietic stem/progenitor cells resulted in diminished overall survival, which was primarily attributed to hematopoietic defects. Specifically, expression of Ezh2Y641F at the hematopoietic stem cell (HSC) stage did not induce B cell or myeloid transformation, but instead led to bone marrow failure due to both cell intrinsic and cell extrinsic factors. Analysis of gene expression profiles in pro- and pre pro-B cells isolated from the CD19-Cre and Mx1-Cre crosses demonstrated a significant overlap in differentially expressed genes and pathways such as IFN signaling and IL2-Stat5 signaling. Expression of Ezh2Y641F at different times during hematopoietic development also resulted in expression of different pathways in early B cell progenitors, such as E2F targets, heme metabolism and GTPase activity driven by GBP proteins. Forced expression of Gbp2/3 in several lymphoma/leukemia cell lines revealed subtle changes in cell survival, suggesting a possible mechanism for how expression of Ezh2Y641F could have a different effect on hematopoietic transformation depending on when it is expressed. In summary, our study underscores the critical importance of the timing of the Ezh2Y641F mutation in determining its oncogenic activity. The varied temporal contexts result in distinct downstream changes that govern the ultimate oncogenic impact of EZH2 mutations.
