An RNAi Screen Reveals a Pattern of Addiction to Epigenetic Pathways That Distinguishes MLL-AF9 Leukemia From Normal Hematopoietic Cells

Abstract 65

Epigenetic pathways act to control gene expression in a heritable fashion without altering DNA sequence, typically involving the control of chromatin structure. Several lines of evidence implicate the involvement of epigenetics in the pathogenesis of human cancer, however it remains uncertain to what extent manipulating epigenetic pathways can fully rectify malignant cellular states for therapeutic benefit. To systematically explore this issue, we have developed a novel RNAi screening pipeline that can distinguish the epigenetic requirements for normal and malignant hematopoiesis. This was achieved by first constructing a custom shRNA library targeting all known enzymatic complexes that regulate chromatin structure (1100 shRNAs in total). Next, each shRNA was delivered systematically to cells derived from a mouse model of chemotherapy-resistant acute myeloid leukemia driven by the oncogenes MLL-AF9 and NRAS, or to several non-transformed hematopoietic cell lines of different lineages. shRNAs were scored for their capacity to differentially inhibit growth of leukemic cells without influencing growth of non-leukemic cells. Each of the identified genes was then evaluated in vivo for its influence on normal reconstitution of the hematopoietic system following transplantation of shRNA-infected hematopoietic stem and progenitor cells into lethally irradiated recipient mice. Each gene was likewise suppressed in leukemia cells in vivo using both constitutive and conditional RNAi vectors. The net result of the in vivo testing was identification of 6 genes encoding different regulators of chromatin structure whose suppression provides therapeutic benefit in a mouse model of therapy-resistant AML, without significantly influencing the production of normal blood lineages. In support of the accuracy of our screening protocol, one of the identified genes from the screen encodes the protein Menin, a known MLL-AF9 cofactor essential for disease initiation and shown to be dispensable for steady-state hematopoiesis in knockout mice.

Two of the identified hits in our screen are the genes Eed and Suz12, which encode two subunits of the PRC2 Polycomb complex (Eed and Suz12), which catalyzes histone H3K27 methylation to suppress gene expression. Inhibiting PRC2 function in MLL-AF9 leukemia cells leads to monocytic differentiation, as revealed by FACS, RT-qPCR, and cell morphology analysis. Microarray experiments coupled with chromatin immunoprecipitation identified a core program of myeloid fate determinants that are suppressed by PRC2 via H3K27 methylation. In addition, we observe that MLL-AF9 directly occupies several Polycomb gene promoters to upregulate their expression in leukemic cells. Our findings highlight an unexpected alliance between the MLL-AF9 oncogene (a Trithorax protein) and PRC2 (a Polycomb complex), which act together to block myeloid differentiation in AML. Our findings also highlight the utility of employing RNAi in vivo to identify novel therapeutic targets in otherwise chemotherapy-resistant disease models.