Responsiveness of FLT3-ITD+ AML to Kinase Inhibition Correlates with High Allele Ratio and Relapsed Disease but Not to Inhibition of Canonical Targets

Fms-like tyrosine kinase 3 (FLT3) is a receptor tyrosine kinase and is the most commonly mutated gene in acute myeloid leukemia (AML). Second generation FLT3 inhibitors such as quizartinib (AC220) are clinically active in relapsed FLT3-ITD+ patients. However, not all patients respond and to date primary resistance has not been characterized. A previous study proposed that AML cells from patients with relapsed FLT3 mutant AML and samples with high allelic burden for FLT3 ITD are more sensitive to FLT3 inhibitor cytotoxicity (Pratz KW, et al Blood 2010). We performed studies to test this hypothesis and determine if cells with homodimeric FLT3 ITD are more heavily dependent on FLT3 ITD for growth and survival than cells expressing heterodimeric FLT3 WT:FLT3 ITD. 16 primary AML samples that contain FLT3 ITD mutations were incubated in increasing concentrations of the second generation FLT3 inhibitor crenolanib and assayed for survival in short term liquid culture assays. Only 6 samples demonstrated greater than 30% inhibition of survival in this culture system whereas 10 samples showed little or no cytotoxic response. Consistent with previous results (Pratz, et. al, Blood 2010), responding samples tended to be from relapsed patients and to have higher FLT3 ITD allelic ratios. We then analyzed FLT3 expression and phosphorylation levels as well as inhibition of and crenolanib inhibition of FLT3 phosphorylation, as well as the canonical downstream signal transduction pathways STAT5, ras/MAPK, and PI3K/AKT/mTOR in 13 FLT3-ITD+ primary AML samples. For 11/13 samples, crenolanib strongly inhibited phosphorylation of FLT3 kinase. However, neither FLT3 protein expression nor baseline phosphorylation level correlated with cytotoxic response in liquid culture assays. Crenolanib inhibited phosphorylation of STAT5, ribosomal S6 and ERK to varying degrees and inhibition of none of these pathways consistently correlated with cytotoxicity. Overall, these results are consistent with the hypothesis that FLT3 ITD mutant AML with high allele burden or relapsed samples are more addicted to FLT3 ITD. To further examine this topic, we studied AML cell lines that express only wild type FLT3 (THP1), both FLT3-ITD and WT FLT3 (MOLM14) or FLT3-ITD but not WT-FLT3 (Mv4;11, TF1-ITD). Growth of all three cell lines expressing FLT3-ITD but not THP1 cells was inhibited by crenolanib. Crenolanib inhibited tyrosine phosphorylation and activation of downstream signaling pathways in all three FLT3-ITD+ cell lines. Importantly, crenolanib was active at 10 nM in TF1-ITD cells but higher concentrations were required to inhibit signaling in Mv4;11 cells. This suggests that allele ratio alone does not determine sensitivity to FLT3 inhibitors. Interestingly, FLT3 ligand (FL) impairs inhibition of FLT3 by kinase inhibitor and, again, mutant cell lines are similarly responsive to FL in the presence of kinase inhibitor. In summary, these data demonstrate that AML cells have variable dependence on mutant FLT3 for survival. Samples with high allele ratio and relapsed samples behave in a manner consistent with oncogene addiction and are likely to show cytotoxicity to FLT3 inhibition. However, the basis for oncogene addiction is unclear and does not depend on activation of the canonical signal transduction pathways known to be downstream of FLT3. Interestingly, a recent unbiased screen of phosphorylated proteins in FLT3-ITD+ AML to predict clinical response to AC220 did not identify tyrosine phosphorylation of canonical FLT3 targets, but correlated clinical response with serine phosphorylation on EEPD1-S160, BCL11A-S630, and RANBP3-S333 (Schaab C, et al. Leukemia 2014). Analysis of the effects of FLT3 inhibitors upon these proteins in FLT3 mutant primary samples and cell lines is ongoing.