Pre-Clinical Activity of Novel Hypoxia-Activated FLT3 Inhibitors in FLT3-Mutated AML

Activating mutations in the Fms-like tyrosine kinase 3 (FLT3) are most frequently observed molecular abnormalities in AML, which lead to constitutive activation of the tyrosine kinase. While several mechanisms of resistance to FLT3 tyrosine kinase inhibitors (TKIs) have been identified, commonly failure is associated with lack of complete and sustained inhibition of FLT3 tyrosine kinase activity. FLT3 inhibition is particularly challenging in blasts within the hypoxic bone marrow (BM) microenvironment, which plays a crucial role in AML survival. Widespread hypoxia has been shown by us and others to be prevalent within the confines of the leukemic BM space, and to limit the efficacy of standard chemotherapy (Benito et al., PLoS One 6(8):e23108, 2011). With the goal of exploiting hypoxia as a physiological target and increasing selectivity to FLT3-mutant AML blasts, we have developed a series of hypoxia-activated prodrugs of FLT3 TKIs (HA-TKIs) based on the known clinical compounds sunitinib, AC220, MLN-518 and crenolanib. Bioreductive trigger conjugation was performed at the final step of the synthesis to provide the prodrugs as previously described (Lu et al., Tetrahedron, 69, p9130-9138, 2013). Under hypoxia, these compounds can undergo enzymatic one-electron reduction and then fragment selectively to release the cell permeable TKI, generating higher local concentrations of TKIs and limiting normal tissue exposures.

Pulse radiolysis was used to determine the electron affinity (E(1)) and rate of fragmentation of each prodrug. The E(1) values ranged from -430 mV to -497 mV, while the rate of fragmentation of the prodrugs following one-electron reduction under hypoxia ranged from 28 s^-1 to 400 s^-1. SN37156 and SN37169 in particular demonstrated E(1) and k_frag parameters ideal for hypoxia-selective cellular metabolism (-449 mV, 85 s^-1; -449 mV, 62 s^-1, respectively).

With the purpose of identifying the HA-TKIs with the most selective and dose-potent in vitro efficacy, we first performed an Oxic/Hypoxic screening of the hypoxia-dependent activity of HA-TKIs against FLT3-mutant AML cell lines. From the 10 compounds originally developed, SN37156 and SN37169 showed hypoxia-dependent apoptosis and growth arrest. SN37156 demonstrated the best activity against Molm13, Molm14 and MV4-11 cells under hypoxia with hypoxic cytotoxicity ratios (oxic/hypoxic IC₅₀ ratios) of 4.9, 11 and 17.6, respectively.

In our secondary screening we compared the activity of SN37156, SN37169 and their parental inhibitors on target modulation. Both HA-TKIs selectively inhibited FLT3 and ERK auto phosphorylation in MV4-11 cell line under hypoxic but not oxic conditions, with similar potency to the parental inhibitor under normal oxygen culture conditions.

We next investigated the effects of both HA-TKIs on inhibition of FLT3 downstream targets, including AKT, ERK, pS6 and STAT5 using time-of-flight mass cytometry (CyTOF). The data demonstrated distinct patterns of signaling inhibition, with a dose-dependent p-ERK and p-S6 inhibition in MV4-11 and Molm-14 cells upon treatment with both HA-TKIs. In turn, both compounds at 100nM inhibited p-AKT and p-STAT5.

Single dose tolerability studies were conducted in tumor-free C57BL6 mice. Prodrugs were administered by IP injection in 20% β-hydroxypropylcyclodextrin solution starting at 17.8 µmol/kg and escalating in 1.33-fold increments until lethality or severe mortality (mean body weight loss of >10%) was observed. The maximal tolerated dose for both, SN37156 and SN37169 was established at 56.1 µmol/kg. To test efficacy of HA-TKIs in vivo, we injected C57BL6 mice with genetically engineered Baf3-ITD/luc/GFP cells. Mice were randomized to treatment groups on day 8 when engraftment was documented by bioluminescent imaging (BLI), and treated with either vehicle or 56.1 µmol/kg of SN37156 or SN37169 I.P. every 3 days. Multiple doses (up to 10 so far; study is ongoing) were well tolerated. BLI demonstrated reduced leukemia burden in both treatment groups compared to control.

In summary, our data indicate that HA-TKIs SN37156 and SN37169 inhibit mutant FLT3 and its selected downstream signaling targets under conditions resembling marrow hypoxia, and induce hypoxia-dependent growth arrest and apoptosis in FLT3-mutant AML cells. Planned studies in FLT3-mutant AML PDX will evaluate in vivo efficacy to eliminate primary human AML cells from hypoxic BM microenvironment.