Dynamic genetic and nongenetic RAS-pathway activation drives resistance to FLT3 and BCL2 inhibitor therapy.

Bulk sequencing of relapsed tumors reveals mutations associated with resistance to cancer therapy but is insufficient to fully assess all causes of relapse. Because of inherent tumor heterogeneity, on-treatment tumor evolution may select genetically distinct clones or shifts in malignant transcriptional states not resolvable by bulk sequencing. We performed multiomic single-cell DNA/protein and RNA/protein profiling of a clinical trial cohort of patients with acute myeloid leukemia (AML) treated on the phase 1b clinical trial of the BCL2 inhibitor venetoclax and the FLT3 inhibitor gilteritinib (Ven/Gilt) to characterize immunophenotypic, transcriptional, and genetic clonal evolution driving resistance. We found that although Ven/Gilt effectively eliminated FLT3-mutant clones, resistance was associated with RAS activation through multiple mechanisms, including selection for RAS-mutant clones, nonmutational upregulation of RAS transcriptional programs, and a shift to RAS-associated monocytic AML differentiation. Using an in vitro model of monocytic differentiation associated with nonmutational RAS transcriptional activation, we demonstrated that RAS-pathway inhibition resensitized to Ven/Gilt. These data illustrate that convergent resistance pathways in patients can be activated via diverse genetic and nongenetic mechanisms. These results underscore that RAS signaling is central to FLT3 and BCL2 inhibitor resistance, is tightly coupled with AML monocytic differentiation, and highlights the potential of RAS-pathway inhibition as a viable clinical strategy to combat resistance. This trial was registered at www.ClinicalTrials.gov as NCT03625505.