Participation of Ebf2⁺cells and their progeny in leukemic niche formation during AML progression. (A) Experimental setup for analyzing Ebf2⁺ cells in the AML BM. Normal and MLL-AF9–expressing lineage (LIN) ⁻KIT⁺ cells were generated from FVB/N mouse BM LIN⁻KIT⁺ cells and transplanted to nonirradiated Ebf2-Egfp reporter FVB/N mice. Ebf2⁺ cells were detected by their GFP expression. BM stromal cells in control and AML mice were collected when the mice developed AML. BM stromal cell subsets were analyzed and sorted by FACS for subsequent functional assays and gene expression analysis. (B) Representative FACS profiles showing abnormal expansion of Ebf2⁺ MSPCs in AML mice. The gating strategy used for phenotypic analysis and sorting of Ebf2⁺ cells and MSCs and MPCs within the Ebf2⁺ cell fraction in the BM of control and AML mice. (C) Increased frequencies and numbers of Ebf2⁺ MSPCs within total CD45⁻TER119⁻CD31⁻ stromal cells in AML mice. Data are from 2 independent experiments. (D) Reduced proportion of MSCs and increased proportion of MPCs within Ebf2⁺ cells in AML mice. Data are from 2 independent experiments. (E) Frequency of CFU-Fs in sorted Ebf2⁺ cells from control and AML BM. Each dot in the left panel represents mean values of duplicates or triplicate measurement. Horizontal bars in panels C-E show median values. Data are from 3 independent sorting experiments. (F) qPCR analysis of Cxcl12, Kitl, and Il6 expression in Ebf2⁺ cells from AML and control mice. Data are from 2 to 4 independent experiments. (G) Experimental setup. Eight- to 10-week old Ebf2-Egfp × Ebf2-Cre^ER × Rosa26-^loxpStop^loxp-Tomato mice were first injected with TAM to activate expression of tomato in Ebf2⁺ cells and their progeny. One month later, mice were transplanted either with normal or MLL-AF9–expressing LIN⁻KIT⁺ cells without irradiation. The BM stromal cells in the control and AML mice were collected and analyzed by FACS at the onset of AML. (H) FACS profile showing gating of tomato-expressing stromal cells. Tomato⁺ cells were first gated within CD45⁻TER119⁻CD31⁻7AAD⁻ live cells. Cells were analyzed for CD44⁻ and CD44⁺ cell fraction, and CD44⁻Tomato⁺ cells were then further subdivided into SCA1⁺PDGFRA/CD140A⁺ MSCs (PaS), SCA1⁻CD140A⁺ MPCs, and SCA1⁻CD140A/PDGFRA⁻ mature stromal cells. (I) Frequencies of the total tomato-expressing cells and tomato⁺ MSPCs within the BM stromal cell compartment (CD45⁻TER119⁻CD31⁻) in control and AML mice. (J) Frequencies of the Tomato⁻ MSCs and MPCs within the BM stromal cell compartment (CD45⁻TER119⁻CD31⁻) in control and AML mice. (K) FACS plots showing the distribution of MSCs and MPCs within Tomato⁻ and Tomato⁺ cells in AML mice. SCA1⁺PDGFRA/CD140A⁺ MSCs (PaS) and SCA1⁻CD140A⁺ MPCs were first gated within stromal cells (CD45⁻TER119⁻CD31⁻). Cells were further analyzed for their expression of Tomato and GFP. Gates were made based nontransgenic mice and fluorescence-minus-one control. (L) Dot plots represent the gating strategy used for the expression of Tomato and GFP. Differences were determined using an unpaired parametric Student t test (C [right],D-F,I [left],L), unpaired parametric Student t test with Welch’s correction (F), or Mann-Whitney U test (C [left],I). Horizontal bars represent mean values in panels C-F,I-K. Data are from 2 independent experiments.