Direct Conversion of Adult Endothelial Cells into Immunecompetent Long-Term Engraftable Clinically Scalable Hematopoietic Stem Cells: Pathway to Therapeutic Translation

The molecular pathways and microenvironmental cues that choreograph the conversion of endothelial cells (ECs) into true engraftable hematopoietic stem cells (HSCs) remain undefined. This is due to lack of models to recreate the short-lived transition from EC to hemogenic cells and to HSCs. Extending on our previous work (Sandler V. et al, Reprogramming of human endothelium into hematopoietic cells requires vascular niche induction. Nature, 511:312-8. 2014), we have developed a novel, sequential, clinically-translatable in vitro model of the adult EC to hematopoietic transition (EHT). This model uses precise, conditional, on-off expression of transcription factors (FosB, Gfi1, Runx1, and Spi1 - FGRS) and an inductive vascular niche to reprogram adult mouse ECs into true HSCs (rEC-HSCs) with high efficiency. During the induction phase (days 0-8), FGRS are conditionally expressed in adult non-lymphatic ECs isolated from Runx1-IRES-GFP reporter mice and co-cultured with the supportive vascular niche cells. During the specification phase (days 8-20), FGRS-transduced VEcad⁺Runx1^-CD45^- ECs activate expression of endogenous Runx1, initiating the hematopoietic program and silencing EC fate. The VEcad⁺Runx1⁺CD45⁺ cells, then complete specification and full commitment to VEcad^-Runx1⁺CD45⁺ hematopoietic stem and progenitor cells (rEC-HSPCs). Specified rEC-HSPCs are then expanded (days 20-28) on the vascular niche generating a large number of hematopoietic cells and, at this point, expression of exogenous FGRS is turned off. Transplantation of rEC-HSPCs (CD45.2) into lethally irradiated (CD45.1) recipient mice reconstitute both short-term and long-term hematopoiesis, and are capable of engrafting secondary and tertiary recipients (rEC-HSCs). Once engrafted, rEC-HSPCs give rise to functional myeloid and lymphoid cells with full complement of polarized T cell subsets. rEC-HSC-derived immune cells undergo T-cell receptor (TCR) rearrangement and reconstitute adaptive immune function in Rag1^-/- mice.

To prove the stem cell potential of rEC-HSCs, we performed clonal analyses on the VEcad⁺Runx1⁺CD45⁺ cells (days 8-20), in which single cells were plated in coculture with vascular niche. Notably, 7 out of 386 CD45.2⁺ clones gave rise to expanding colonies that were capable of 4 months primary multilineage engraftment into lethally irradiated CD45.1⁺ recipients. In addition, limiting dilution transplantation of isochronic VEcad^-Runx1⁺CD45⁺ cells indicated that 1 out of 538 rEC-HSPCs are repopulating rEC-HSCs with primary and secondary engraftment potential. Thus, based on both clonal and limiting dilution transplantation studies, we demonstrate that rEC-HSCs represent true repopulating prototypical HSCs. In addition, since during the expansion phase there are >200 fold expansion of the SLAM-coded KLS population, these data support the notion that during the 20-28 days expansion phase there is robust proliferation of rEC-HSCs. Serial analyses of the primary, secondary and tertiary engrafted rEC-HSCs, showed no evidence of leukemias. Molecular analyses indicated that both in vitro and in vivo expanded rEC-HSCs had complete erasure of vascular signature and stable expression of hematopoietic profile.

Moreover, employing Runx1-IRES-GFP reporter mice enabled deconvolution of stage-specific pathways involved in generation of engraftable rEC-HSCs. Inhibition of TGFβ signaling along with activation of BMP and CXCL12 pathways reinforced the induction phase. Active Notch and CXCL12 signaling throughout the specification and expansion phases supported self-renewal of transplantable rEC-HSCs. This stepwise reprogramming approach provides an interrogatable in vitro platform to elucidate the critical pathways involved in the transition of ECs into hemogenic-like and ultimately hematopoietic cells in vitro. Specifically, these data unequivocally support that our approach lead to generation of large number HSCs, enabling therapeutic application. Thus, our reprogramming platform that does not require transition through a pluripotent state, will facilitate devising strategies to reprogram ECs into abundant autologous long-term repopulating HSCs amenable to genetic modification for treatment of inherited and acquired hematological disorders.