FLT3-F691L confers kinase-independent resistance via Grb2-dependent MAPK signaling in AML

Abstract

Abstract FMS-like tyrosine kinase 3 (FLT3) mutations occur in ~20–30% of acute myeloid leukemia (AML) cases and have been effectively targeted with FLT3 inhibitors. However, acquired resistance remains a major limitation of FLT3-targeted therapy. A hot-spot mutation in FLT3, FLT3-F691L (phenylalanine to leucine), emerges in ~12% of patients treated with the FDA-approved FLT3 inhibitor, gilteritinib. Although considered as a gatekeeper mutation that interferes with inhibitor binding, we uncovered a neomorphic, FLT3-kinase-independent mechanism by which FLT3-F691L mediated resistance to gilteritinib and other type I FLT3 inhibitors. To investigate the mechanistic basis of FLT3-F691L resistance to FLT3 inhibitors, we evaluated type I FLT3 inhibitors (gilteritinib and NCGC-1481) in isogenic FLT3-ITD and FLT3-F691L AML models (MOLM14 and CD34⁺ MLL-AF9 cells) in vitro and in vivo. Despite comparable suppression of canonical FLT3-STAT5 signaling in both FLT3-ITD and FLT3-F691L cells, FLT3-F691L AML cells retained leukemic cell viability and function. Moreover, the degree of FLT3-STAT5 inhibition did not correlate with anti-leukemic activity in FLT3-F691L cells, in contrast to the expected functional decline in FLT3-ITD cells. Biochemical assays confirmed that FLT3-F691L did not significantly impair gilteritinib or NCGC-1481 inhibitor binding affinity. Moreover, gene expression profiling revealed that gilteritinib downregulated mitogenic signaling gene signatures in FLT3-ITD AML, but these signatures were preserved in FLT3-F691L AML. We further found that gilteritinib suppressed MAPK signaling (e.g., pRAF, pMEK) in FLT3-ITD AML, but FLT3-F691L AML retained MAPK activity. These findings align with clinical observations where gilteritinib-refractory or relapsed FLT3-ITD patients often acquire either FLT3-F691L or RAS/MAPK mutations, but not both, suggesting parallel mechanisms converging on MAPK signaling. Based on our observation that gilteritinib-treated FLT3-F691L AML retained MAPK signaling despite FLT3 kinase inhibition, we hypothesized a kinase-independent role for FLT3-F691L. To test this, we knocked down FLT3 in FLT3-ITD or FLT3-F691L AML cells using shRNAs. Knockdown of FLT3 in both FLT3-ITD and FLT3-F691L cells suppressed cell viability, leukemic potential, and MAPK (RAF/MEK) signaling in vitro. To further investigate a FLT3 kinase-independent requirement of FLT3-F691L, we employed a FLT3-targeting PROTAC derived from gilteritinib (“CRBN(FLT3)-8”) to degrade FLT3. Unlike gilteritinib or a negative control PROTAC, CRBN(FLT3)-8 robustly suppressed viability, clonogenicity, and RAF/MEK signaling in FLT3-F691L AML, indicating that FLT3-F691L retains essential non-kinase functions. To uncover pathways mediating the FLT3-F691L kinase-independent function, we performed a genome-wide CRISPR-Cas9 dropout screen in isogenic FLT3-ITD and FLT3-F691L AML cells treated with gilteritinib. Grb2, an adaptor protein that links receptor tyrosine kinases to Ras activation via SOS, emerged as a top dependency unique to gilteritinib-treated FLT3-F691L cells. While shRNA-mediated Grb2 knockdown alone had minimal effect on FLT3-F691L AML, co-treatment of shGrb2 cells with gilteritinib significantly impaired viability, clonogenic potential, and RAF/MEK activation. These findings suggest that FLT3-F691L utilizes Grb2 to sustain MAPK signaling despite FLT3 kinase inhibition with gilteritinib. In summary, we report that FLT3-F691L is not exclusively a classical gatekeeper mutation that blocks inhibitor binding; rather, it rewires FLT3 signaling to bypass canonical kinase activity and sustain leukemogenesis via Grb2-mediated Ras/MAPK activation. This neomorphic function offers a mechanistic explanation for gilteritinib resistance and highlights new therapeutic strategies. Our findings support the development of FLT3 degraders or combined targeting of FLT3 and Ras/MAPK pathways to overcome FLT3-F691L-driven resistance in relapsed/refractory AML. Importantly, this work suggests that other recurrent hot-spot mutations in oncogenic kinases may similarly confer kinase-independent functions in cancer and should be systematically investigated as potential drivers of therapeutic resistance.