Combined inhibition of Notch and FLT3 produces synergistic cytotoxic effects in FLT3/ITD+ acute myeloid leukemia

Internal tandem duplication (ITD) mutations of FMS-like tyrosine kinase-3 (FLT3) are the most frequent genetic alterations in acute myeloid leukemia (AML) and predict a poor prognosis. FLT3 tyrosine kinase inhibitors (TKIs) provide short-term clinical responses, but the long-term prognosis of FLT3/ITD+ AML patients remains poor. Notch signaling is important in numerous types of tumors. However, the role of Notch signaling in FLT3/ITD+ AML remains to be elucidated. In the current study, we found that Notch signaling was activated upon FLT3-TKI treatment in FLT3/ITD+ cell lines and primary cells. As Notch signaling can be blocked by γ-secretase inhibitors (GSIs), we examined the combinatorial antitumor efficacy of FLT3-TKIs and GSIs against FLT3/ITD+ AML and explored the underlying molecular mechanisms. As a result, we observed synergistic cytotoxic effects, and the treatment preferentially reduced cell proliferation and induced apoptosis in FLT3/ITD+ AML cell lines and in primary AML cells. Furthermore, the combination of FLT3-TKI and GSI eradicated leukemic cells and prolonged survival in an FLT3/ITD+ patient-derived xenograft AML model. Mechanistically, differential expression analysis suggested that CXCR3 may be partially responsible for the observed synergy, possibly through ERK signaling. Our findings suggest that combined therapies of FLT3-TKIs with GSI may be exploited as a potential therapeutic strategy to treat FLT3/ITD+ AML.

supplemented with protease inhibitors (ServeBio). The lysates were quantified by the Bradford assay, and equal amounts of lysates were separated by SDS-PAGE and transferred to PVDF membranes (GE Healthcare, NJ, USA). Proteins were detected by immunoblotting using ChemiDoc XRS+ imaging System (Bio-Rad) following incubation with primary antibodies and appropriate secondary antibodies. Images were obtained and analyzed using Image Lab Software (Bio-Rad). The antibodies used are listed in Supplementary Table 3.

Flow cytometry analysis
Flow cytometry analysis was performed using BD FACSCalibur machine (BD Biosciences) or Beckman Coulter FC500 flow cytometer (Beckman Coulter). To measure the CD34+ cell proportion of human peripheral blood samples, mononuclear cells were collected after isolation and the designated drug treatments. Cells were then washed two times with PBS and stained with anti-CD34 antibody (Biolegend, CA, USA) for 30 minutes at 4℃ . After washing, the samples were subjected to flow cytometry analysis.
To quantify leukemic engraftment in the orthotopic AML xenograft model, the PB and spleen cells were isolated after red cell lysis, labeled with anti-human CD45 antibody (Biolegend) and analyzed by flow cytometry. Data were analyzed using Cytoexpert software (Beckman Coulter) and FlowJo software (version 10.0.7r2; TreeStar, USA).

Generation of FLT3/ITD knock-in Cell Lines
FLT3/ITD knock-in cell lines were generated according to the reported method [1]. To generate short guide RNA (sgRNA) constructs, sequences targeting the 14th exon of FLT3 were loaded at crispr.mit.edu. The BbsI restriction site (CACC) was added to the selected oligos. The oligomers were then annealed and cloned into SKM-1 cells (1 × 10 6 cells per well in 1 mL culture medium in a 12-well plate) were co-transfected with 1 µg sgRNA plasmid and 2 µg donor plasmid using Lonza 4D nucleofector (Lonza, Basel, Switzerland) according to a previously reported method [2]. GFP+ cells were sorted by FACS after 2 weeks of BSD drug selection and seeded in a 96-well plate, with a single cell per well. Subsequently, the cells were cultured, expanded and genotyped. For genotyping, PCR reactions were performed using KOD Fx Polymerase (Toyobo) following the manufacturer's instructions. The PCR products were gel-extracted, and the insertion was confirmed by DNA sequencing. FLT3-ITD-Loxp-BSD-Loxp knock-in cells were transfected with 1 µg pBS505 EF1 alpha EGFP Cre (Addgene #11955), and GFP+ cells were sorted by FACS after 72 h and seeded at a single cell per well. After cell expansion, genotyping was performed again, and SKM-1 FLT3-ITD knock-in cells were finally identified.
The primers used for genotyping are listed in Supplementary Table 2. Conversely, the FLT3/ITD knock-in clones were validated by q-PCR and immunoblotting.

Cell transfection
For Hes1 knockdown, equal amount of non-silencing (NS) or Hes1-targeted (Hes1_1, Hes1_2 and Hes1_3) siRNAs were transfected using Lonza 4D nucleofector (Lonza, Basel, Switzerland) according to the manufacturer's protocol. AC220 was added 24h after transduction. Lentivirus expressing pCDH vector and DNMAML were kindly provided by Hudan Liu (Medical Research Institute, Wuhan University, Wuhan, China) and lentiviral transduction was performed as described before [3]. For transfection, 2x10 5 exponentially growing cells were seeded in 0.5mL culture medium with additional 4μg/mL polybrene (Genechem). Lentivirus were added at an MOI of 25. Three days after transduction, MOLM13 cells were selected in culture medium containing 1ug/mL puromycin for 1 week.

RNA-seq analysis and Gene Set Enrichment Analysis (GSEA)
Principal component analysis (PCA) was performed with R package gmodels The expression of ICN1, ICN2, ICN3 and ICN4 in Molm13 and MV4-11 cells were measured by western blotting following treatment with indicating concentrations of DAPT for 12 h. GAPDH was used as a loading control. Images are representative of three independent experiments.

Supplemental Figure 3. Generation and verification of CRISPR/Cas9-mediated FLT3/ITD mutation knock-in in SKM-1 cells (A)
Schematic overview of the strategy used to knock-in the FLT3/ITD mutation in FLT3 exon 14. (B) Quantitative RT-PCR analysis of FLT3 transcript levels in SKM-1 wild type cells (FLT3/WT) and three FLT3/ITD knock-in clones relative to GAPDH.
Error bars indicate the average fold change relate to FLT3/WT cells of three independent experiments ± SD (**P < 0.01; ***, P < 0.001). (C) Representative image of immunoblot analysis of the signaling mediators downstream of FLT3 in SKM-1 wild type and FLT3/ITD+ clones. The cells were pre-starved of growth factors for 12 h and subsequently exposed to FLT3 ligand (0.1ng/ml) for 10 minutes before harvesting.  (n=4) were treated with 0~100 µM DAPT for 48 hours and apoptosis was measured by Annexin V binding. Error bars indicate the average ± SD (ns. not significant).

Supplemental Figure 5. Treatment with sorafenib and DAPT alone or in combination is well tolerated
The mouse weight for each cohort, measured once a week, is shown (n=12). Error bars indicate the average ± SD.

Supplemental Figure 6. Summary of differentially expressed genes (DEGs)
The numbers of differentially expressed genes (DEGs) identified by the gene expression level are presented.

Supplemental Figure 7. Expression of downstream of FLT3
The expression of phospho-STAT5 (pSTAT5), total STAT5, phospho-AKT (pAKT) and total AKT in Molm13 and MV4-11 cells was measured by western blotting following treatment with AC220 (2.5 nM) and/or DAPT (2.5 µM) for 12 h. GAPDH was used as a loading control. Images are representative of three independent experiments.

Supplemental Figure 8. AMG487 had little effect on peripheral blood mononuclear cells from healthy donors
Peripheral blood mononuclear cells obtained from healthy donors(n=4) were treated with 0~25 µM AMG487 for 48 hours and apoptosis was measured by Annexin V binding. Error bars indicate the average ± SD (ns. not significant).
Supplemental Figure 9. Representative raw plots for apoptosis assay.