We have previously identified a novel function for the adhesion molecule E-selectin – awakening otherwise dormant hematopoietic stem cells (HSC) and inducing lineage commitment (Winkler et al., Nat Med 2012). Now we show that therapeutic blockade of E-selectin in vivo specifically augments the mobilisation of HSC with highest self-renewal potential following G-CSF administration, and markedly improves subsequent engraftment and reconstitution in mice. From these data we hypothesise that vascular E-selectin acts as a gatekeeper influencing activation of transmigrating HSC.

Firstly we found that administration of mobilizing doses of G-CSF increased the level of E-selectin expressed on the surface of bone marrow (BM) endothelial cells. To determine whether E-selectin influenced HSC mobilisation, we then compared G-CSF-mediated mobilisation in wildtype and E-selectin knock-out mice. We found that although absence of E-selectin did not significantly alter the number of phenotypic HSC or colony-forming cells mobilized into the blood following G-CSF, the absence of E-selectin in mobilised mice did increase the subsequent engraftment and reconstitution potential of mobilised blood analysed by competitive repopulation transplant assays. Next we investigated whether this beneficial effect could similarly be achieved by transient E-selectin blockade using therapeutic doses of GMI-1271, a small synthetic mimetic that specifically blocks the binding of E-selectin to its receptors. Wild-type mice were administered human G-CSF alone (125µg/kg subcutaneously twice daily for 3 days) ± GMI-1271 injections (20 mg/kg BID). The number of mobilized HSC were quantified by rigorous limiting-dilution transplantation of 0.3, 1, 5 or 20 µL mobilised blood in competition with 200,000 congenic BM cells into lethally-irradiated congenic recipients, to enable quantification of long-term reconstituting cells per mL of mobilised blood by Poisson’s distribution. We found the mobilized blood of donor mice injected with both GMI-1271 and G-CSF, showed faster engraftment and 25-fold increased reconstitution potential over blood from G-CSF alone injected controls (p<0.0001). That is after 3 days of G-CSF injections, blood from G-CSF plus GMI-1271 injected mice contained 476 reconstitution units (RU) /mL compared to 20 RU/mL in blood from mice mobilised with G-CSF alone. (One reconstitution unit is the same reconstitution as 100,000 normal bone marrow cells; 95% CI was 200 - 500 RU/mL compared to 11 – 36 RU/mL in blood from G-CSF plus GMI-1271 treated mice, compared to G-CSF alone injected mice respectively). Surprisingly, this dramatic boost to reconstitution potential with GMI-1271 co-administration was not reflected in the numbers of phenotypic HSPC, or number of colony-forming cells mobilised per mL of blood. Together these findings suggest that transient interactions between HSC extravasating across the BM endothelium during G-CSF administration, with E-selectin expressed on the BM vasculature, may inadvertently compromise the reconstitution potential of up to 96% of harvested peripheral blood HSC, indicating an unexpected disadvantage with current HSC harvesting procedures.

These data are consistent with the role of E-selectin in the bone marrow (to awaken otherwise dormant HSC) that we have previously reported, and also point the way forward to a simple remedy (administration of E-selectin antagonist together with G-CSF) during HSC mobilisation to improve short- and long-term engraftment thus accelerate recovery in transplant recipients. In conclusion we propose the novel concept of E-selectin as a vascular ‘gate-keeper’ dampening the potential of migratory HSC.


Winkler:FibroGen Inc.: Research Funding. Magnani:GlycoMimetics Inc.: Employment, Equity Ownership, Membership on an entity's Board of Directors or advisory committees.

Author notes


Asterisk with author names denotes non-ASH members.

This icon denotes a clinically relevant abstract