Abstract

Background

Activating mutations in FLT3 occur in ∼30% of adult acute myeloid leukemia (AML), primarily consisting of internal tandem duplication (ITD) mutations (∼25%) and point mutations in the tyrosine kinase domain (∼5%), commonly at the activation loop residue D835. Secondary D835 mutants in FLT3-ITD are frequently associated with acquired clinical resistance to effective FLT3 TKIs. While molecular docking studies suggest that D835 mutations confer resistance by favoring an active “DFG-in” kinase conformation that is unfavorable to the binding of type II FLT3 TKIs, which bind to a “DFG-out” inactive conformation, specific D835 mutations associated with acquired clinical resistance to AC220 (Smith et al., Nature, 2012) appear to differ from those in sorafenib-treated patients (Man et al., Blood, 2011). For example, the D835H mutation is associated with clinical resistance to sorafenib but not AC220. In addition, in vitro saturation mutagenesis screens of FLT3-ITD have identified various secondary D835 mutations as causes of resistance to different TKIs (Smith et al., Blood, 2013; Smith et al., Nature, 2012). Importantly, currently employed clinical assays only detect the presence or absence of D835 mutations and do not reveal their identity. We hypothesized that D835 mutations differ in their relative resistance to type II FLT3 inhibitors in vitro, and we therefore sought to profile the activity of several clinically-active FLT3 TKIs against all clinically-detected D835 mutations to date.

Results

We created BaF3 cells expressing FLT3-ITD with all clinically detected somatic mutations observed at the FLT3 D835 residue to date: D835A/E/F/G/H/I/N/V/Y/del, and assessed their sensitivity in vitro to the clinically active FLT3 TKIs AC220, sorafenib, and ponatinib (all type II inhibitors; FLT3-ITD IC50 0.53, 2.81, 2.29 nM respectively), as well as to the type I TKI, crenolanib (FLT3-ITD IC50 19.56 nM), which is capable of binding the “DFG-in” active conformation and has been previously reported to retain activity against D835 mutants (Smith et al., ASH, 2012 Abstract #141). While all FLT3-ITD/D835 mutants conferred a degree of resistance to the type II TKIs in proliferation assays (range 3 to ∼2600-fold compared to FLT3-ITD), some mutants display variable resistance. For example, the FLT3-ITD/D835H mutation conferred ∼26-fold resistance to AC220, in contrast to ∼295-fold resistance to sorafenib and ∼100-fold resistance to ponatinib, consistent with the detection of these mutations in patients with acquired clinical resistance to sorafenib but not AC220. D835F/I/V/Y/del mutations conferred the highest degree of resistance to all type II TKIs (relative resistance range ∼107 fold to ∼2600 fold compared to FLT3-ITD). Substitutions at D835E/N conferred the least resistance (<20 fold) to all inhibitors tested. As expected, all D835 mutations retained sensitivity to the type I TKI crenolanib.

Conclusions

Clinically relevant D835 mutations confer differential resistance to clinically active type II FLT3 TKIs, and may retain sensitivity to some agents at clinically achievable concentrations. The molecular mechanisms of resistance to type II TKIs conferred by mutations at D835, which represent the most common secondary kinase domain mutations identified in FLT3 TKI-treated AML patients to date, therefore appear to involve more than simply favoring an active kinase conformation. The in vitro sensitivity of D835 mutations to select FLT3 TKIs may guide the subsequent therapy choice for patients who relapse on TKI therapy with a D835 mutation, as it has for CML patients treated with BCR-ABL TKIs. To this end, clinical assays that specify the identity of D835 mutations are necessary. Further studies will be required to elucidate the molecular mechanisms of variable resistance conferred by select FLT3 D835 mutations.

Disclosures:

Off Label Use: Unapproved drugs: AC220, crenolanib. Off-label use in AML: sorafenib, ponatinib. Shah:Ariad Pharmaceuticals: Consultancy, Research Funding; Ambit Biosciences: Research Funding.

Author notes

*

Asterisk with author names denotes non-ASH members.