Leukemic mutation FLT3-ITD is retained in dendritic cells and disrupts their homeostasis leading to expanded Th17 frequency

Abstract

Dendritic cells (DC) are mediators between innate and adaptive immune responses to pathogens and tumors. DC development is determined by signaling through the receptor tyrosine kinase Fms-like tyrosine kinase 3 (FLT3) in bone marrow myeloid progenitors. Recently the naming conventions for DC phenotypes have been updated to distinguish between "Conventional" DCs (cDCs) and plasmacytoid DCs (pDCs). Activating mutations of FLT3, including Internal Tandem Duplication (FLT3-ITD), are associated with poor prognosis for acute myeloid leukemia (AML) patients. Having a shared myeloid lineage it can be difficult to distinguish bone fide DCs from AML tumor cells. To date, there is little information on the effects of FLT3-ITD in DC biology. To further elucidate this relationship we utilized CITE-seq technology in combination with flow cytometry and multiplex immunoassays to measure changes to DCs in human and mouse tissues. We examined the cDC phenotype and frequency in bone marrow aspirates from patients with AML to understand the changes to cDCs associated with FLT3-ITD. When compared to healthy donor (HD) we found that a subset of FLT3-ITD+ AML patient samples have overrepresented populations of cDCs and disrupted phenotypes. Using a mouse model of FLT3-ITD+ AML, we found that cDCs were increased in percentage and number compared to control wild-type (WT) mice. Single cell RNA-seq identified FLT3-ITD+ cDCs as skewed towards a cDC2 T-bet^- phenotype, previously shown to promote Th17 T cells. We assessed the phenotypes of CD4⁺ T cells in the AML mice and found significant enrichment of both Treg and Th17 CD4⁺ T cells in the bone marrow and spleen compartments. Ex vivo stimulation of CD4⁺ T cells also showed increased Th17 phenotype in AML mice. Moreover, co-culture of AML mouse-derived DCs and naïve OT-II cells preferentially skewed T cells into a Th17 phenotype. Together, our data suggests that FLT3-ITD+ leukemia-associated cDCs polarize CD4⁺ T cells into Th17 subsets, a population that has been shown to be negatively associated with survival in solid tumor contexts. This illustrates the complex tumor microenvironment of AML and highlights the need for further investigation into the effects of FLT3-ITD mutations on DC phenotypes and their downstream effects on Th polarization.

Keywords: AML; FLT3; T helper (T) 17 cells; Treg; dendritic cells; single cell.

Figure 1

Human Bone Marrow Phenotyping Reveals Disruption of cDC Populations. (A) Representative dot plots showing human bone marrow cDCs. cDCs defined as Lin(CD3e, CD19, CD56, CD14, AXL, CD123)-CD45⁺CD11c⁺HLA-DR⁺. (B) Summary graph for (A) Each symbol is one human sample. HD n=9, FLT3-WT AML n=12, FLT3-ITD⁺ AML n=10. (C) Representative dot plots showing cDC1 and cDC2 subsets of the cDCs identified in (A). cDC1 defined as CD1c-XCR1⁺ and cDC2 defined as XCR1-CD1c⁺. (D–F) Summary graphs for (C) Each symbol is one human sample. HD n=9, FLT3-WT AML n=12, FLT3-ITD+ AML n=10. (G) Representative dot plots showing CLEC10A expression of cDC2s. (H, I) Summary dot plots of CLEC10a+ and CLEC10a- cDC2s as a frequency of total cDCs (A). * = P < 0.05; ** = P < 0.01; *** = P < 0.001; **** = P < 0.0001; ns = P > 0.05.

Figure 2

AML mice have increased frequencies of myeloid cells and cDCs. (A) Comparison of dissected spleens from healthy WT mice and diseased AML mice. Ruler for scale (cm). (B) Summary graph of spleen wet weight for WT and AML mice. Each symbol is one mouse. WT n=18 AML n=17. (C) RBC lysed single-cell suspensions of WT and AML mice spleens were stained with monoclonal antibodies for known immune cell markers. Summary graph for frequency of CD19-CD3e-CD11b⁺ splenocytes. Each symbol is one mouse. WT n=17 AML n=16. (D) RBC lysed single-cell suspensions of WT and AML PBMCs were stained with monoclonal antibodies for known immune cell markers. Summary graph for frequency of CD19-CD3e-CD11b⁺ PBMCs. Each symbol is one mouse. WT n=4 AML n=10. (E) Survival curves of reference C57BL/6J mice from Jackson Laboratories (B6) and our mouse model AML mice. (F) Representative dot plots showing mouse splenic cDCs. cDCs defined as Lin(Ly6C, Ly6G, F4/80, CD3e, CD19, NK1.1)- CD45⁺CD11c^hiMHC-II^hi. (G) Summary graph of cDC cell counts per spleen of WT and AML mice. Each symbol is one mouse. WT n=5 AML n=6. (H) Summary graph of cDC cell counts per bone marrow of WT and AML mice. Each symbol is one mouse. WT n=5 AML n=6. (I) Summary graph of (F). Each symbol is one mouse. WT bone marrow n=10 WT spleens n=23. AML bone marrow n=13 AML spleens n=16. (J) Representative histograms of phospho-FLT3 (pFLT3) staining in cDCs as measured by flow cytometry. RBC lysed single-cell suspensions of WT and AML PBMCs were stained with monoclonal antibodies for known immune cell markers at the surface and were stained for pFLT3 in the cytoplasm. (K). Summary graph showing geometric mean fluorescent intensity (gMFI) pFLT3 staining of blood circulating cDCs. Each symbol is one mouse. WT n=4 AML n=11. * = P < 0.05; ** = P < 0.01; **** = P < 0.0001.

Figure 3

Single cell RNA-seq analysis of AML and WT mouse spleens identify changes in DC populations. (A) UMAP plot of DC subsets identified from WT and AML mice after magnetic bead enrichment for DCs. (B) Mean expression of various DC gene markers across annotated DC subtypes. (C) Mean abundance of protein markers across annotated DC subtypes. (D) DC subtype proportions across samples AML (n=4) and WT (n=4). (E) Violin plots of DC subtype proportions compared between WT and AML groups. Differences in means were determined using Student’s t-test.

Figure 4

AML mouse blood exhibits a Th17 inflammatory phenotype. (A) Representative dot plots for gating on mouse blood CD4⁺ T cells by flow cytometry. Top row is WT and bottom row is AML. (B) Transcription factor expression of CD4⁺ T cells. Top row is summary bar charts of transcription factor expression in CD4⁺ T cells from (A). Each symbol is an individual mouse. WT (n=6) and AML (n=6). Bottom row is representative histograms of each transcription factor. (C) Blood serum cytokines in WT (n=21) and AML (24) mice. Each symbol is an individual mouse. Summary bar charts for selected inflammatory cytokines measured using BioLegend LegendPlex multiplex assay Mouse Inflammation Panel. (D) Summary bar chart for bone marrow IL-17A+ CD4⁺ T cells after stimulation with Cell Activation Cocktail for six hours. WT n=6 and AML n=6. (E) Summary bar chart for spleen IL-17A+ CD4⁺ T cells after stimulation with Cell Activation Cocktail for six hours. WT n=6 and AML n=6. (F) Summary bar charts for Luminex measurement of human IL-10 cytokine detected in HD and AML patient samples of mixed peripheral blood (PB) and bone marrow (BM). HD (n=6) and FLT3-ITD+ AML (n=50) and FLT3-WT AML (n=251). (G) Summary bar charts for Luminex measurement of human IL-17 cytokine detected in HD and AML patient samples of mixed peripheral blood (PB) and bone marrow (BM). HD (n=6) and FLT3-ITD+ AML (n=50) and FLT3-WT AML (n=251). * = P < 0.05; ** = P < 0.01.

Figure 5

AML mice retain more OT-II cells and AML DCs promote IL-17A production. (A) Experimental outline of OT-II adoptive cell transfer. CD45.1+ AML (n=8) and WT (n=8) mice are injected intravenously with 200k naïve CD45.2+ OT-II cells on Day 0. 24 hours later mice were injected intraperitonealy with 200 µg whole OVA protein. 10 days later mice were sacrificed and spleens were harvested for flow cytometry analysis of remaining CD45.2+ OT-II cells. (B) Representative dot plots of spleen resident CD45.2+ OT-II cells by flow cytometry. Frequency of live ingle cells noted in the plots. (C) Summary bar charts of OT-II cell counts (left panel) and frequency of live single cells (right panel). Each symbol is an individual mouse and summary data collected from two experiments. AML (n=8) and WT (n=8). (D) Experimental outline of in vitro co-cultures of OT-II cells and cDCs. Naïve OT-II cells were magnetically sorted from spleen tissue using BioLegend MojoSort Mouse CD4 Naïve T Cell Isolation Kit. After magnetic sorting samples were verified for purity by flow cytometry. cDCs were magnetically isolated using BioLegend MojoSort Mouse CD11c+ beads. After magnetic sorting samples were verified by flow cytometry. OT-II cells were plated at 50k cells per well and cDCs were plated at 20k cells per well for five days in Th17 polarizing media using CellXVivo Mouse Th17 Cell Differentiation Kit CDK017 from R&D Systems. (E) Representative dot plots of IL-17A+ OT-II cells by flow cytometry. Frequency of CD4⁺ cells noted in the plots. (F) Summary bar charts of IL-17A+ CD4⁺ T cells from (E). Each symbol is an individual sample and summary data collected from four experiments. AML (n=8) and WT (n=7). * = P < 0.05; ** = P < 0.01; *** = P < 0.001.