Abstract
Introduction: KIT mutations are detected in approximately 20-30% of core binding factor (CBF, t(8;21) and inv(16)/t(16;16)) AML and have been associated with outcome and response to tyrosine kinase inhibitors (TKI). Mutations tend to localize to 2 regions- the receptor's kinase activation loop in exon 17 (E17) and a region of the extracellular domain in exon 8 (E8). The prognostic significance of the 2 mutations has varied in previous series. In this analysis, the functional significance of E8) and E17 KIT mutations in vitro and their response to TKIs were evaluated. We hypothesized that activating variants might be associated with inferior clinical outcome and be more susceptible to KIT-targeted agents.
Methods: COG AAML0531 enrolled 1,070 patients, of which 800 underwent targeted next generation sequencing of the entire KIT gene coding sequence to identify somatic KIT mutations. To evaluate the functional impact of the most common KIT mutations, E8 (D419G, D419indel) and E17 (D816V, N822K) mutations were engineered and transduced into HEK293 cells for evaluation of KIT phosphorylation (pKIT) by immunoblotting and into Ba/F3 cells for cytokine independent proliferation studies. Mutations that resulted in pKIT were evaluated for response to the TKIs dasatinib and crenolanib. We further correlated our mutational and functional analysis to clinical outcome data from COG AAML0531. We analyzed outcomes by treatment arms (no-gemtuzumab ozogamicin (no-GO) vs. GO), focusing on the GO arm given the potentially favorable impact of GO on CBF AML response.
Results: A total of 247 CBF patients were enrolled, with 218 undergoing screening for somatic KIT mutations. Mutations were detected in 55 (25%) patients; 27 (49%) in E8, 26 (47%) in E17, and 2 (4%) involved both exons. Among the 109 patients with CBF AML treated on no-GO arm, KIT mutations were detected in 29/109 (27%); 14 (48%) in E8, 13 (45%) in E17, and 2 (7%) involved both exons. E8 mutations were a combination of missense mutations or 3-9 bp indels that maintained reading frame, while E17 variants were primarily missense mutations involving the D816 or N822 codons, while. Cells transduced with E8 mutations showed baseline receptor phosphorylation and no IL-3 independent proliferation. In contrast, E17 mutations showed autonomous receptor phosphorylation and IL-3 independent proliferation, with D816V showing more robust cellular proliferation and higher levels of pKIT (Fig1A).
We further evaluated the cytotoxic effect of 2 TKIs in engineered cells with activating KIT mutations and KIT wild-type (WT). Neither parental line nor cells expressing KIT -WT had any response to the two TKIs with IC50s > 10,000μM, however cells transduced with the E17 variants D816V or N822K were exquisitely sensitive to the TKIs. Following exposure to dasatinib and crenolanib, N822K+ cells had an IC50 of 3.8nM and 28.3nM respectively, with D816V+ cells demonstrating similar IC50s of 13nM and 14.3nM respectively (Fig 1B).
Given the differential activating potential of E8 vs. E17 variants, we evaluated the clinical implications of the mutations in the 2 regions among CBF patients. On the no-GO arm, E8 KIT mutations lacked clinical significance, with an overall survival (OS) of 81±20% vs. 80±9% (p=0.91) for patients with and without E8 mutations, with a corresponding disease-free survival (DFS) of 57±26% vs. 62±11% (p=0.75) and relapse risk (RR) of 43±28% vs. 36±11% (p=0.59) respectively. In contrast, activating E17 mutations were associated with adverse outcome with an OS of 56±27% vs. 85±7% for those with and without E17 mutations (p=0.02; Fig 1C), with a corresponding DFS 33±27% vs. 65±11% (p=0.05) and a RR of 67±30% vs. 32±11% (p=0.03) respectively. Interestingly, when we evaluated outcomes on the GO arm, patients with E17 mutations had an OS of 77±23% vs. 83±8% for KIT-WT (p=0.62), suggesting that GO may abrogate the adverse prognostic impact of E17 mutations.
Conclusions: Somatic E17 KIT mutations lead to autonomous receptor phosphorylation and cellular proliferation and are associated with inferior clinical outcome when treated with conventional chemotherapy alone. This is in contrast to E8 mutations, which lack functional significance. KIT E17 mutations are uniquely sensitive to the cytotoxic effects of KIT inhibition. Consideration of TKI therapy for at least this subset of pediatric KIT + CBF AML is prudent given their inferior prognosis and this preclinical correlate.
No relevant conflicts of interest to declare.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal