FLT3 Mutant to Wild Type Allelic Ratio and Clinical Status Are Predictive of Response to FLT3 Inhibitors in AML.

Abstract 1716

Poster Board I-742

AML patients who harbor FLT/ITD mutations have a poor prognosis. Several small molecule kinase inhibitors with activity against FLT3 are in pre-clinical and clinical development. These compounds, derived from several different chemical classes, might be predicted to have very different effects both in vitro and in vivo on FLT3-mutant AML. We therefore compared 5 different inhibitors (Lestaurtinib, AC220, KW-2449, Sorafenib, and Sunitinib) for potency against mutant and wild type FLT3, as well as for cytotoxic effect against a series of 13 primary blast samples obtained from acute myeloid leukemia (AML) patients harboring internal tandem duplication (FLT3/ITD) mutations. Using immunoblot assays of Molm14 (FLT3/ITD, mutant) and SEMK2 (FLT3, wild type) in culture medium containing 10% bovine serum, the IC50 for inhibition of FLT3/ITD autophosphorylation for each of the drugs was as follows: Lestaurtinib 2 nM; AC220 1 nM; KW-2449 10 nM; Sorafenib 3 nM; and Sunitinib 1 nM. For wild type, these values were: Lestaurtinib 10 nM; AC220 5 nM; KW-2449 36 nM; Sorafenib 28 nM; and Sunitinib 2 nM. All 5 drugs, therefore, were more potent against the FLT3/ITD compared to wild type FLT3. Taking plasma protein binding into account, AC220 is predicted to be the most potent of these compounds in vivo (IC50 18.4 nM in plasma), while Lestaurtinib is predicted to be the least potent in vivo (IC50 700 nM in plasma). Using immunoblot assays of FLT3 in FLT3/ITD primary blast samples, we found that complete inhibition of FLT3 autophosphorylation does not always induce cell death, implying that some FLT3/ITD AML is not truly addicted to FLT3 signaling. This phenomenon was confined to patient samples obtained at diagnosis. Relapsed samples and samples with a high mutant-to-wild type allelic ratio were invariably more responsive to cytotoxicity from FLT3 inhibition compared with the samples obtained at diagnosis or those with a low mutant allelic ratio. Using an IL-3 rescue assay as well as published assays of inhibition of other kinases, we ranked the inhibitors according to their relative selectivity for FLT3. The most FLT3-selective agent was AC220, followed by sorafenib, then KW-2449, sunitinib, and finally lestaurtinib. The selectivity for FLT3 of a given inhibitor clearly influenced the cytotoxic response it induced in primary samples, in that lestaurtinib, the least selective (or most multi-targeted) was broadly cytotoxic to virtually all FLT3/ITD samples studied, while the more selective agents were more effective against samples obtained at relapse or samples with high FLT3 mutant-to-wild type ratios. FLT3 mutant AML, therefore, seems to undergo an evolution at relapse to a state of being more dependent on mutant FLT3 signaling. While multi-targeted FLT3 inhibitors may have greater efficacy in vitro, particularly in the diagnostic samples, it remains to be seen how well-tolerated such agents are clinically. Our results have important implications for the potential therapeutic use of FLT3 inhibitors.