EVI1 exerts distinct roles in AML via ERG and cyclin D1 promoting a chemoresistant and immune-suppressive environment

Abstract

Abstract Aberrant expression of ecotropic viral integration site-1 (EVI1+) is associated with very poor outcomes in acute myeloid leukemia (AML), mechanisms of which are only partially understood. Using the green fluorescent protein reporter system to monitor EVI1 promoter activity, we demonstrated that Evi1high KMT2A-MLLT1–transformed AML cells possess distinct features from Evi1low cells: the potential for aggressive disease independent of stem cell activity and resistance to cytotoxic chemotherapy, along with the consistent gene expression profiles. RNA sequencing and chromatin immunoprecipitation sequencing in EVI1-transformed AML cells and normal hematopoietic cells combined with functional screening by cell proliferation–related short hairpin RNAs revealed that the erythroblast transformation–specific transcription factor ERG (E26 transformation-specific [ETS]-related gene) and cyclin D1 were downstream targets and therapeutic vulnerabilities of EVI1+ AML. Silencing Erg in murine EVI1+ AML models severely impaired cell proliferation, chemoresistance, and leukemogenic capacity. Cyclin D1 is also requisite for efficient EVI1-AML development, associated with gene expression profiles related to chemokine production and interferon signature, and T- and natural killer–cell exhaustion phenotype, depending on the interferon gamma (IFN-γ)/STAT1 pathway but not on CDK4/CDK6. Inhibiting the IFN-γ/STAT1 pathway alleviated immune exhaustion and impaired EVI1-AML development. Overexpression of EVI1 and cyclin D1 was associated with IFN-γ signature and increased expression of chemokines, with increased exhaustion molecules in T cells also in human AML data sets. These data collectively suggest that ERG and cyclin D1 play pivotal roles in the biology of EVI1+ AML, where ERG contributes to aggressive disease nature and chemoresistance, and cyclin D1 leads to IFN-γ signature and exhausted T-cell phenotypes, which could potentially be targeted.