FLT3 is genetically essential for ITD-mutated leukemic stem cells but dispensable for human HSCs

Leukemic stem cells (LSCs) fuel relapse in acute myeloid leukemia (AML), but therapies tailored at eradicating LSCs without harming healthy hematopoietic stem cells (HSCs) are lacking. FLT3 is frequently mutated in AML and associated with relapse; but FLT3 targeting has met with limited clinical success. This raises questions of whether more potent inhibitors would increase effectiveness but whether toxicity to HSC would become limiting. Here, we tested the consequence of complete FLT3 ablation using CRISPR/Cas9 FLT3 knock-out (FLT3-KO) in human HSCs and LSCs followed by functional xenograft assays to test their ability to regenerate human hematopoiesis and leukemia, respectively. FLT3-KO in HSCs from human fetal liver, cord blood and adult bone marrow showed no impairment in multilineage hematopoiesis in primary and secondary xenografts. By contrast, FLT3-KO LSCs from 6 of 7 FLT3-ITD mutated AMLs were able to generate short-term engraftment but were completely exhausted by 12 weeks. Thus, FLT3 is essential for LSC long-term propagation. This dependency was unique to FLT3-ITD AML samples as non-FLT3-ITD AML samples generated leukemic grafts upon FLT3-KO. Transcriptomic analysis revealed that FLT3-KO induced downregulation of DNA repair and cell cycle checkpoints, uniquely in FLT3-ITD AML, but not in healthy HSCs or other AMLs. Our research highlights a critical distinction between healthy HSCs and LSCs: whereas healthy hematopoiesis proceeds unperturbed upon FLT3-KO, FLT3-ITD leukemogenesis is impaired through elimination of LSCs. This evidence underscores the necessity for more potent FLT3-targeting and places FLT3 as an ideal therapeutic target to selectively eradicate LSCs, while sparing HSC. We used three primary patient samples of acute myeloid leukemia (AML) with the internal tandem duplication (ITD) in FLT3 – samples ITD 1, ITD 2 and ITD 3; three primary patient samples of AML without the ITD mutation – samples WT 1, WT 2 and WT 3; and three samples of human umbilical cord blood (CB). AML samples were thawed, and the fraction previously determined (PMID: 27926740) to be enriched in leukemic stem cell activity was sorted: CD34+ population was sorted from all samples, except from WT 3 where the CD45+ CD3- population was sorted. CB samples were thawed and the CD34+CD38- population was sorted. AML samples were cultured for 24-h, while CB cells were cultured for 48-h. Subsequently, FLT3-KO and OR2W5-KO were performed using CRISPR/Cas9 technology through electroporation. After electroporation, cells were cultured for 3 additional days to achieve the maximal reduction on FLT3 expression after FLT3-KO in culture. Subsequently, cells were collected and viably sorted using SYTOX Blue Dead Cell Stain (Thermo Fisher Scientific). A minimum number of 13 x103 live cells was sorted per sample (range 13 to 800 x103). One fifth of the sample was collected for total DNA isolation and subsequent determination of KO percentage mediated by CRISPR/Cas9. The rest of the sample was used for total RNA purification and DNase treatment using the RNeasy Micro kit (Qiagen). RNA integrity was measured using the RNA 6000 Pico kit (Agilent) and the average RNA integrity scores were the following: 8 for AML-ITD samples (6.9-10), 10 for AML-WT samples (9.7-10) and 9.95 for CB samples (9.8-10). To generate RNAseq libraries, total RNA was used as input with SMART-Seq V4 Ultra Low Input RNA kit (Clontech). The cDNA profiles were confirmed on the bioanalyzer using the High Sensitivity 14 DNA kit (Agilent) and next-generation sequencing libraries were made using the Nextera XT DNA Library Preparation kit (Illumina). Sequencing was performed on the NovaSeq 6000 system using 2 x 150 bp paired-ended sequencing to obtain ~50 M reads per sample. >>>Submitter states that raw data will be submitted to EGA<<<