PI3K Isoform Dependence in Adult Hematopoiesis and Myeloid Leukemia

? DESCRIPTION (provided by applicant): Acute myeloid leukemia (AML) is a devastating disease that is diagnosed in about 19,000 people per year, with a 5-year survival of only 24%. The mainstay of treatment is multi-agent chemotherapy and bone marrow transplantation, both associated with high morbidity and mortality, largely due to toxic effects on hematopoietic stem cells (HSC). To enable the design of more specific treatments for AML, it is critical to identify therapeutic targets in leukemia-initiating cells (LICs), since this cell population is important in relapse and resistance. PI3 kinase (PI3K) is a lipid and protein kinase that transduces growth factor and chemokine signals, leading to phosphorylation and activation of the Ser/Thr kinase Akt, which regulates metabolism, the cell cycle, apoptosis, and protein synthesis. Pathologic phosphorylation of Akt is frequently reported in AML patient samples, and inhibition of the PI3K/Akt pathway has shown efficacy in AML cell lines, patient samples, and mouse leukemia models. This pathway is an attractive therapeutic target for leukemia and other malignancies. However, the role of PI3K signaling in normal adult HSCs is unclear. This is an important consideration for the potential toxicity of PI3K/Akt inhibitors in the clinic. In hematopoietic cels, genes encoding four different isoforms of the catalytic subunit of PI3K (p110 ??????? and ?) are expressed. These isoforms have unique functions in normal and cancer cells, but may substitute for each other in some contexts. Two classes of PI3K inhibitors are currently in clinical development: pan-PI3K inhibitors, which target all four PI3K isoforms, and isoform-selective inhibitors, which favor only one or two isoforms. To determine which of the PI3K isoforms are essential for hematopoiesis, and which are more important in leukemic cells, we have generated a series of mouse knockouts to study the roles of each isoform individually in adult HSCs. We recently found that RAS- mutated myeloid leukemias are dependent on the p110alpha isoform of PI3K, and that pharmacologic inhibition of p110alpha is effective in treatment of murine models of RAS-mutated leukemia. However, normal HSC functions are maintained in the absence of p110alpha, making it a safe therapeutic target. It is still not known whether PI3K is essential for adult HSC function, as other PI3K isoforms may compensate for p110alpha in HSCs. Furthermore, the role of PI3K in LICs is unclear. We have now generated novel compound knockout mouse models, which will enable the inducible conditional deletion of multiple PI3K isoforms in adult HSCs. In Specific Aim 1, we will further characterize the redundant roles of the PI3K isoforms in HSC function, self-renewal, proliferation, and differentiation. In Specific Aim 2, we will genetically ablate PI3K isoforms in mouse models of AML, in order to determine the roles of PI3K in LIC function. At the moment, criteria for the selection of AML patients that are most likely to respond to isoform-selective PI3K inhibitor treatment have not been established. To address this issue, in Specific Aim 3, we will delineate the molecular determinants for PI3K dependence and PI3K isoform specificity in AML.