The phosphoproteomic landscape of 16 Acute Myeloid Leukemia (AML) cell lines

Acute myeloid leukemia (AML) is a clonal hematopoietic malignancy, characterized by expansion of immature leukemic blasts in the bone marrow. In AML, specific tyrosine kinases have been implicated in leukemogenesis, and are associated with poor treatment outcome. However, targeted therapy using kinase inhibitors (KIs) has had limited success, and may be improved by proper patient selection. We performed phosphotyrosine (pY) based, label-free phosphoproteomics to identify hyperphosphorylated, active kinases in AML cell lines as targets and predictive biomarkers to select KIs for treatment. We identified 3605 class I phosphorylation sites in 16 AML cell lines (EOL-1, KG-1a, MM6, KG-1, ME-1, NB-4, Kasumi-3, MV4-11, THP-1, HEL, HL-60, Kasumi-1, Kasumi-6, ML-2, OCI-AML3, MOLM-13) that exhibited large variation in the number and level of phosphopeptides per cell line (241-2764). Ranking analyses successfully pinpointed the hyperactive kinases PDGFRA, FGFR1, KIT, and FLT3 in eight cell lines with a corresponding kinase mutation. Additionally, we identified unexpected drivers in two more cell lines (PDGFRA in Kasumi-3 and FLT3 in MM6) which proved sensitive to specific kinase inhibitors. Six cell lines without a clear receptor tyrosine kinase (RTK) driver showed evidence of MAPK1/3 activation, consistent with the presence of activating RAS mutations. Our data show the potential of pY phosphoproteomics to identify key drivers in AML cells, and the predictive value of the phosphoproteome profiles in TKi selection for targeted treatment.