Single-Cell Pharmacodynamic Monitoring of Ribosomal Protein S6 Phosphorylation During AC220 Therapy Demonstrates in Vivo FLT3 Inhibition in Circulating Leukemic Blasts and Accurately Detects the Development of AC220 Resistance.

Quizartinib (AC220) is a potent, highly selective inhibitor of FLT3, KIT, and FMS tyrosine kinases with promising clinical activity in AML patients, particularly those with FLT3 internal tandem duplication (FLT3-ITD) mutations. However, limited biochemical FLT3 inhibition in leukemic blasts in vivo by AC220 has been previously described (Perl, et al ASH 2012 #3502). As a biomarker for FLT3 inhibition, we assessed monitoring of phosphorylated ribosomal protein S6 (pS6) at serines 235/236 by flow cytometry. S6 is a downstream target of the PI3 kinase/mammalian target of rapamycin (mTOR) pathway, is constitutively phosphorylated in nearly all FLT3-ITD+ samples, and its phosphorylation shows dynamic changes in response to FLT3 ligand (FL) or FLT3 inhibitors in vitro. We hypothesized that S6 phosphorylation would be a biomarker for FLT3 inhibition and here provide final analysis of AC220 phamacodynamic monitoring at our center.

Serial peripheral blood samples were collected during a phase 2 AC220 clinical trial (Cortes, et al. ASH 2011, #2576). Blood was aliquoted within four hours of collection and a subset was exposed to signaling inhibitors (ex vivo AC220, rapamycin × 30 min.) or activators (phorbol ester/PMA or FL × 10 min.) to establish dynamic controls. Following incubation, samples were formaldehyde-fixed and red cells lysed with the permeabilizing agent triton X-100. Samples were stored at −20C in glycerol-containing medium. After collecting all time points, samples were simultaneously thawed, denatured with ice-cold methanol, and stained for flow cytometry. Blasts were identified by using CD45 and side scatter and confirmed by expression of multiple surface markers (CD33, CD34, CD117, HLA-DR, etc.). Constitutive S6 phosphorylation was defined by comparing unstained (fluorescence minus one, FMO) and PMA and/or cytokine-stimulated cells. Biochemical sensitivity to AC220 was defined as a reduction in the percentage of positive phospho-S6 blasts by >50% as compared to baseline.

21 subjects provided whole blood samples (15 FLT3-ITD+, 6 FLT3-WT), 15 had evaluable peripheral blasts (>100 blasts/microliter) for p-S6 monitoring. Only one subject with FLT3-WT had sufficient circulating blasts for analysis. By contrast, 13/15 FLT3-ITD+ subjects' blasts were evaluable. As previously described by our group and others using flow cytometry, pS6 is heterogeneous in primary AML samples and, at basal state, frequently only demonstrable in a subset of blasts. The mean percentage of blasts demonstrating S6 phosphorylation (pS6+) prior to AC220 therapy was 15% (median 4%, range <0.5–97%). To determine if AC220 inhibits FLT3 kinase in vivo, we serially monitored pS6 in subjects prior to and following their initial dose of AC220. 10/13 subjects in which AC220 reduced or eliminated tumor burden in blood and/or marrow were biochemically sensitive to AC220 by flow. The mean percentage of pS6+ blasts 2 hours following a single dose of AC220 was 3% (median 1%, range <0.5–11%). Samples evaluated at predicted drug peak and trough concentration gave similar reductions in pS6, suggesting the drug inhibits signaling throughout the entire dosing interval. Importantly, all subjects with biochemically-sensitive leukemia cleared their peripheral blasts by day 29. In contrast, in the only subject with de novo clinical resistance to AC220, no reduction of S6 phosphorylation was observed. Re-activation of S6 phosphorylation occurred in 5/5 subjects who relapsed during AC220 and for whom samples were available. 4/5 subjects had treatment-emergent FLT3 mutations (e.g. FLT3-D835) in addition to a demonstrable FLT3-ITD; the fifth showed clonal selection for a FLT3-WT clone. S6 phosphorylation in these relapse samples was not potently suppressed by additional ex vivo AC220 in 3/4, suggesting that the cells developed primary resistance to the drug and that drug failure was not secondary to pharmacokinetic changes over time.

We demonstrate the feasibility and utility of flow cytometry for phospho-S6 to monitor the biochemical efficacy of FLT3 inhibitors. The potential of these methods to predict clinical response/resistance paired with the rapid turnaround time of flow cytometry suggests potential future application of this technology in the screening evaluation of patients being considered for therapy with novel signal transduction inhibitors.

Perl:Astellas Pharmaceuticals: Consultancy. Carroll:GlaxoSmithKlein: Research Funding.