BGS-2456 Is a Novel Potent Covalent Inhibitor of FLT3 That Highly Discriminates Against KIT and Is Not Toxic Toward Normal Hematopoiesis in Vitro

Introduction: Class III receptor tyrosine kinases include FLT3, KIT, PDGFRA/B and CSF1R. Activating mutations in FLT3 represent the most common genetic anomalies in AML (30%). While there are numerous examples of KIT TKIs that do not inhibit FLT3 (imatinib, avapritinib, dasatinib), to date, all clinically active FLT3 TKIs (quizartinib, gilteritinib, midostaurin, sorafenib) fail to spare KIT inhibition. Preclinical studies suggest that coinhibition of FLT3 and KIT can result in non-specific myelosuppression, which can limit dose intensification and the ability to safely achieve maintenance dosing. A potent FLT3 TKI that is non-myelosuppressive may enable prolonged treatment in conjunction with chemotherapy and thereby substantially improve clinical outcomes, similar to BCR::ABL1 TKIs in Ph+ ALL (Slayton et al, JCO 2018). FF-10101 is a recently described covalent inhibitor to irreversibly bind FLT3. In vitro and in vivo studies have revealed that FF-10101 exhibits substantial activity against FLT3 and resistant mutants F691L and D835V/Y. However, it does not effectively discriminate against KIT inhibition (Yamara et al, Blood 2018). In clinical experience, a majority of responses were associated with persistent myelosuppression (Levis et al, JCO 2021 39:15_suppl, 7008). Further clinical development of FF-10101 has been abandoned. Here we describe a novel covalent inhibitor (BGS-2456) that potently targets FLT3 with outstanding selectivity against KIT.

Results: To confirm that BGS-2456 is a covalent FLT3 inhibitor, we tested various Cys→Ser substitutions in FLT3-ITD. Of these, only substitution of C695 resulted in loss of activity, suggesting that this residue is critical for covalent bond formation with BGS-2456, as has also been demonstrated for FF-10101 (Yamara et al, Blood 2018). Additionally, the compound without the covalent warhead (BGS-2457) was found to be unimpacted by the C695S substitution and substantially less potent than BGS-2456 (Figure).

To characterize in vitro properties of BGS-2456 and whether it achieves selectively against FLT3 relative to KIT, we performed cell viability assays of the patient-derived FLT-ITD-positive AML Molm14 cell line as well as the gastrointestinal stromal tumor cell line GIST-T1, which contains an activating juxtamembrane mutation in KIT. We compared EC50 ratios with other FLT3 TKIs and found that BGS-2456 is highly potent, with an EC50 value of ~0.2nM. Moreover, BGS-2456 displayed remarkable selectivity toward FLT3, >10-fold greater than all other TKIs tested. Kinome profiling confirmed FLT3 as the kinase that was most highly inhibited by BGS-2456 among the 300 kinases assessed, with BMX being a distant second (Figure).

To assess whether the selectivity of BGS-2456 results in less myelosuppression in vitro, we performed CFU assays of healthy bone marrow. Compared to FF-10101 and gilteritinib, BGS-2456 exhibited the least amount of hematologic toxicity, facilitating in vitro proliferation and differentiation of normal hematopoietic progenitor cells even at 100x EC50 concentration against Molm14 cells.

Finally, we assessed BGS-2456's ability to retain activity against the known clinically problematic drug-resistant FLT3-ITD mutants D835Y and F691L. Encouragingly, BGS-2456 retained activity against both D835Y and F691L mutants in the low nanomolar range.

Conclusion: This is the first description of a potent and exquisitely specific FLT3 inhibitor that spares KIT inhibition and displays no myelosuppression in vitro at >100x EC50 concentration. The potent inhibitory effects of BGS-2456 on both D835Y and F691L mutants support its promise as a best-in-class TKI for the treatment of FLT3-mutant AML. Efforts to molecularly dissect the basis of the high degree of selectivity of BGS-2456 are ongoing.