Acute myeloid leukemia (AML) is a grievous illness with unmet medical need. Recently developed targeted therapies have been approved, but only for subsets of patients and the depth and duration of these treatments is not yet established. Bispecific T cell Engaging (BiTE®) antibody constructs engage the body’s endogenous T cells to target tumor cells. To evaluate FLT3 as a target for BiTE® therapy, target expression profiling was performed in primary AML patient samples and normal tissues. FLT3 antigen positivity on the AML bulk population was found in 78% of patients, irrespective of FLT3 mutational status. FLT3 antigen positivity on the AML leukemia stem cell (LSC) population was found in 79% of patients. FLT3 RNA was detected in cerebellum; however, the majority of cerebellar FLT3 transcripts was truncated and cell surface protein was not detected. Two FLT3 BiTE® antibody constructs, one with a half-life-extending (HLE) Fc domain and one without (canonical), induced potent T cell-dependent cellular cytotoxicity (TDCC) against AML cell lines vitro and provided a survival advantage in an AML xenograft model. In nonhuman primate studies, FLT3 BiTE® molecules exhibited reproducible pharmacokinetic profiles and decreased FLT3+ cells in the blood and bone marrow. Nine out of 14 primary AML patient samples responded to ex vivo treatment with FLT3 BiTE®, which trended with FLT3 expression and E:T ratio. Combination with tyrosine kinase inhibitors or PD-1 blockade increased FLT3 BiTE® -mediated cytotoxicity. The data here support the clinical development of FLT3 HLE BiTE® for AML.

HUMAN CEREBELLUM FLT3 PROTEIN ANALYSIS (MASS SPECTROMETRY) IN-GEL DIGESTION AND LC-MS/MS PARAMETERS: Coomassie-stained gel bands from the FLT3 cerebellum immunoprecipitation and the EOL-1 positive control cells were excised, cubed, and subjected to in-gel tryptic digestion using the UCSF in-gel digestion protocol (http://msf.ucsf.edu/protocols.html). Peptides were extracted and analyzed by reverse-phase liquid chromatography followed by tandem mass spectrometry using a NanoAcquity UPLC (Waters, Milford, MA) in-line with a Thermo Orbtrap Velos mass spectrometer (San Jose, CA) with one acquisition per gel band with the control sample run last to prevent carry-over. The MS method consisted of a high-resolution (30,000) full MS scan (400-2000m/z), followed by low resolution collision-induced dissociation (CID) of the top 10 most abundant precursor ions. BIOINFORMATIC DATABASE SEARCHES: Mass spectral data sets were initially searched using Proteome Discoverer Version 1.4.1.14 (Thermo, San Jose, CA) against the SwissProt (23 Sept. 2013) human protein database with a total of 455,465 sequences. Search parameters were as follows: a precursor tolerance of 10 ppm, with tryptic enzyme constraints, a static modification of carboxyamidomethylation for cysteine, and a dynamic modification of oxidation for methionine residues. The data were further evaluated using the in-house program MassAnalyzer to help identify peptides from FLT3 in each band. EXTRACTION OF MS1 ION INTENSITY INFORMATION TO SUPPORT FLT3 PEPTIDE IDENTIFICATION: To generate a spectral library, mass spectral data files were searched individually using PEAKS Studio Version 7.5, build 20150615 (Waterloo, ON, CAN) against the canonical full-length sequence of human FLT3 (Uniprot entry P36888-1; FLT3_HUMAN). Search parameters in PEAKS were as follows: trypsin digestion with two missed cleavages, variable modifications of methionine oxidation, deamidation of asparagine, conversion of glutamine and glutamic acid to pyroglutamic acid, and N-terminal protein acetylation. Carbamidomethylation of cysteines was set as a fixed modification. Spectral libraries of identified peptides were generated using BiblioSpec for creating and searching tandem MS peptide libraries in Skyline as previously described. Raw data files for each associated gel band were directly imported into Skyline using the ProteoWizard data access library. Following import, ion intensity chromatograms were displayed for the top three isotopic peaks and each peptide was inspected manually for proper peak integration based on theoretical isotopic ratios and retention time of the MS/MS spectra. Ion intensity measurements were then compared across individual acquisitions. All peptide quantitation was carried out using the most abundant isotopic peak area.

FLT3 was immunoprecipitated from human cerebellum and EOL-1 cells. Samples were separated by SDS-PAGE and bands were cut out corresponding to western blot analysis of FLT3.