de novo generation of human haematopoietic stem cells.

Development and maintenance of the haematopoietic system relies on a scant number of self-renewing haematopoietic stem cells (HSCs) residing in the adult bone marrow and representing the top of a complex cellular hierarchy. Transplantation of HSCs, harvested from either bone marrow, mobilized peripheral blood or umbilical cord blood (UCB), has become the standard of care for numerous hereditary and malignant blood diseases. However, the limited availability of optimally human leukocyte antigen (HLA)- matched donor HSCs remains a challenge, especially for individuals of non-Caucasian background or mixed ethnicity. While the immunologic naïveté of UCB enables transplantation despite antigen mismatch, the relatively low HSC dose slows engraftment and raises the threat of graft failure. In vitro expansion of UCB HSCs has been vigorously investigated, but despite substantial progress, current protocols are not yet clinically approved. Consequently, and because of considerable interest in illuminating fundamental aspects of blood development, de novo generation of HSCs from non haematopoietic sources has become a major objective for the field — a ‘holy grail’ — with wide-ranging implications for HSC biology and transplantation medicine. Here we are summarizing our latest efforts and progress towards de novo generation of bona fide haematopoietic stem cell.

HSCs have been shown to arise from specialized endothelial subpopulations with hemogenic potential. In animal models, the transition from endothelial to hematopoietic identity has been directly observed both using live and explanted embryonic tissues; however, analogous studies using human embryos are handicapped by technical and ethical obstacles. Human embryonic stem cells (hESCs) provide an in vitro platform for studying the initial events involved in the differentiation of hematopoietic progenitor cells (HPCs). But a major impediment to the isolation, expansion, and study of hESC-derived HPCs is that their putative cells of origin, hemogenic endothelium, exist only ephemerally, and the early stages of hematopoietic ontogeny in this context have not been described.

A clear demarcation of the ontogeny of hematopoietic cells arising from hESCs would enable the integration of in vivo and in vitro developmental studies and accelerate efforts to generate therapeutically useful cell types. To these ends, we generated a transgenic hESC line that separately identifies emergent endothelial and hematopoietic cells during differentiation and used it to directly observe the spectrum of phenotypic progression from hemogenic endothelium to multipotent HPCs and their derivatives. Live imaging of hemogenic ECs during endothelial to hematopoietic transition (EHT) and subsequent differentiation identified phenotypic milestones during hemato-endothelial specification and revealed a temporal bias in lineage potential that correlates with discreet waves of hemogenesis noted in mouse and human fetal tissues.

Despite recapitulating key developmental branching points and acquiring phenotypical attributes of haematopoietc stem cells, our efforts did not yield engraftable HSCs1. Therefore, to generate such cells, we decided to re-ignite “hemogenic memory” in adult endothelial cells by overexpressing key transcription factors.

Direct conversion of cellular identities through transcription factor (TF)-mediated reprogramming represents an alternative strategy to directed differentiation. This approach allows the cell fate of interest to be obtained through the expression of key cell-fate determining TFs.  To circumvent transition through a destabilizing pluripotency state, attempts have been made to reprogram non-haematopoietic cell types into HSCs but these efforts produced haematopoietic progenitor-like cells with poor engraftment potential.

The inability to generate HSCs could be explained by lack of proper environmental cues to self-renew reprogrammed HSCs. We successfully reprogrammed human umbilical vein endothelial cells (HUVEC) to engraftable HSC-like cells through direct conversion by constitutive expression of FOSB, GFI1, RUNX1, SPI1 (FGRS). Propagation of these cells onto a vascular-niche-like environment substantially enhanced reprogramming efficiency, emphasizing the importance of inductive cues from the physiological micro-environment in the orchestration of haematopoietic specification. The converted cells acquired colony-forming potential and were successfully engrafted in immuno- deficient mice after primary and secondary transplantation, producing long-term myeloid and B lymphoid progeny. However, due to constitutive expression of SPI1 and poor recipient lymphoid competence, we could not generate and mature T cell progeny.

Using a conditional expression of FGRS along with congenic transplantation assay, we were able to prove that transient expression of these four transcription factors in adult ECs along with co-culture with proper vascular microenvironment was sufficient to generate HSCs with transcriptome and long-term self-renewal capacity similar to those of adult haematopoietic stem cells,  clonal engraftment and serial primary and secondary multi-lineage reconstitution, and restoring antigen-dependent adaptive immune function3.

Conversion of adult mECs to HSCs appears to be efficient. We showed that a single clone of runx1+ FGRS-EC yield 598 ± 463 bona fide haematopoietic stem cells. Considering we are obtaining 70,000 runx1+ FGRS-EC per conversion, we could theoretically have achieved a clinically translatable dose of 24,000 bona fide HSCs.