a Representative Flow cytometry gating strategy for healthy lineage− CD34+ CD38− cells (n = 2) compared to patient CD34+ CD38− AML blasts (n = 2) on left, and their E-selectin-binding potential after labell with fluorescent recombinant human E-selectin–IgM on right, b Left panel: Gating strategy for Murine BM HSPC (GFP− Lineage− CD11b− KIT+) and MLL-AF9 AML blasts (GFP+ Lineageneg CD11b+ KIT+) with E-selectin-IgM binding potential (Right panel). Graph shows difference in median florescence intensity (MFI) of E-selectin binding in HSPC from healthy (non-leukemic) hosts, and between non-malignant HSPC and AML blasts within same host. Each linked pair of dots are data from same mouse (n = 5/group, two-tailed paired t test p = 0.0004). c Bolus administration of E-selectin antagonist GMI-1271 in vivo selectively mobilizes a small proportion of AML blasts into the peripheral blood. Left panel; time-course of relative change in GFP+ KIT+ AML blasts circulating in peripheral blood post GMI-1271 (red lines) or saline vehicle injection (gray lines). Each line represents data from an individual mouse (n = 5/group). Expressed as fold above starting (zero time) baseline. Statistics: two-tailed t test; 4 h p = 0.0153; 7 h p = 0.0230; 10 h p = 0.0063; 27 h p = 0.0153. Right panel; HSPC (GFPneg Lineageneg CD11b− KIT+) or AML blasts (GFP+Lineageneg CD11b+ KIT+) in peripheral blood. Shown is fold-change at 10 h post injection of GMI-1271 (GMI, red circles) or saline vehicle (gray triangles) above baseline. Each dot represents data from an individual mouse. Mean ± S.D. (n = 5 mice/group). One way-ANOVA multiple comparisons with post hoc Bonferroni correction. Source data are provided as a Source Data file.