BMX kinase mediates resistance to gilteritinib. (A) Differential gene expression for gilteritinib-unresponsive patients. Red indicates genes with adjusted P < .1 and fold change ≥1.5; see also supplemental Table 5. (B) UMAP representation of myeloblast cells stratified by gilteritinib response (rows) and time point relative to starting therapy (columns). BMX expression is shown as the Log₁₀-transformed and color scheme corresponds to scale shown. (C) MV4-11 and (D) MOLM-13 grown under normoxic (−) or hypoxic (+) conditions for 24 hours, followed by treatment with gilteritinib (+) alone or in combination with BMX-IN-1 (1.5 and 2.5 μM, respectively), CHMFL-BMX-078 (3 and 5 μM, respectively) or remibrutinib (3 and 5 μM, respectively) for 48 hours. Cell viability was assessed by MTT (n = 4-6). Representative data of three independent experiments. (E) MV4-11 was cultured in hypoxia (H) or normoxia (N) for 24 hours. Inhibition by gilteritinib alone or in combination with BMX-IN-1 (1.5 μM) or CHMFL-BMX-078 (3 μM) was determined after 48 hours by MTT (n = 6). Representative data of two independent experiments. (F) MV4-11 wild-type and BMX CRISPR KO was cultured in hypoxia (H) or normoxia (N) for 24 hours, treated with increasing concentration of gilteritinib for 48 hours and assessed by MTT (n = 6). Representative data of two independent experiments. (G) Inhibition of cell growth of human primary FLT3-mutated AML samples treated with gilteritinib (100 nM) alone, CHMFL-BMX-078 (1.5 μM) alone, or the combination. Samples were grown in hypoxia for 24 hours, treated for 48 hours, and viability was determined by CellTiter-Glo assay (n = 3). (H) MV4-11 wild-type and BMX CRISPR knockout cells were cultured in hypoxia for 24 hours and treated with 5 nM gilteritinib for 1 hour. Western blots were carried. Representative data from three independent experiments shown. For statistical analysis, **P ≤ .01, ***P ≤ .001, ****P ≤ .0001, as determined by two-tailed, unpaired Student t test.