Fetal liver hematopoiesis revisited: a precast hierarchy

Abstract

Late fetal liver hematopoiesis was thought to primarily rely on hematopoietic stem cells (HSCs). Using new genetic-tracing tools, a study shows that EVI1-positive HSCs mainly undergo expansion in the fetal liver, while differentiated blood cell production depends on HSC-independent intermediate hematopoietic progenitors.

Hematopoietic stem cells (HSCs) originate in the major arteries of the developing embryo¹ and subsequently colonize the fetal liver, where they undergo a marked expansion². In mice, HSCs settle to the bone marrow around birth, where they will maintain their lifelong residency at the top of the hematopoietic hierarchy, while ensuring the continuous production of differentiated hematopoietic cells³. Although differences in progenitor composition between fetal liver and adult bone marrow had been identified⁴ and linked in part to fetal liver colonization by yolk-sac-born erythro-myeloid progenitors (EMPs)⁵, it was historically assumed that late fetal liver hematopoiesis relies mainly on HSCs. However, this assumption was never convincingly demonstrated. One of the reasons for this was that much of our previous knowledge relied on data from transplantation assays, which, although undoubtedly essential for determining lineage potential, offer little information about the in vivo fate of cells in a physiological, unperturbed setting.

Recent data have questioned the extent of the HSC-derived contribution to embryonic and fetal hematopoiesis. Initial evidence came from a study showing that progenitors other than HSCs are necessary — and sufficient — to sustain embryonic development until birth⁶. Another study, using the zebrafish as a model, confirmed these observations and demonstrated that HSC differentiation activity during fetal hematopoiesis is limited⁷. However, the identity of progenitor cells responsible for prenatal hematopoiesis remained an open question. Lineage tracing and sophisticated clonal analysis suggested that embryonic-derived, HSC-independent multipotent progenitors make a long-lasting contribution to multiple lineages, seen until postnatal life^8,9. Several fundamental questions still remained to be addressed: in particular, how is the hematopoietic hierarchy in the fetal liver established, and what are the processes driving HSC generation?

In a study in Nature, Yokomizo et al.¹⁰ now report the generation of a transgenic mouse line that directs Cre recombination to hematopoietic clusters in the major embryonic and yolk sac arteries, while sparing EMPs (Hlf^creERT2 mice). By combining this with a reporter line, and following induction by administration of tamoxifen, they were able to genetically label hematopoietic stem and progenitor cells (HSPCs) to follow their fate. The thorough evaluation of labeling dynamics in the fetal liver revealed that traced cells initially included not only HSCs, but also other hematopoietic progenitors that appeared at a slightly earlier time. This led the authors to identify an HLF⁺KIT⁺ precursor population of HPSCs (pre-HSPCs) that is generated in an HSC-independent way and forms a hierarchical-like structure in the fetal liver (Fig. 1).

Fig. 1 |. Schematic of fetal hematopoietic system generation and maintenance.

HLF⁺ EVI1^lo hematopoietic stem and progenitor cells (HSPCs) originate from major extra- and intraembryonic arteries and colonize the fetal liver, where they form a pre-constructed hematopoietic hierarchy preferentially undergoing differentiation. HLF⁺ cells with the highest levels of EVI1 expression (EVI1^hi) are predominantly generated in intraembryonic arteries and contribute self-renewing definitive hematopoietic stem cells (HSCs). DA, dorsal aorta; UA, umbilical artery; VA, vitelline artery.

Single-cell and bulk transcriptomic profiling of the HLF⁺KIT⁺ subsets at pre-liver stages (which include HSC precursors) revealed phenotypic heterogeneity that was especially evident when comparing intraembryonic and extraembryonic regions. This observation is in line with recent evidence suggesting that cell fates might already be determined at the hemogenic endothelium stage¹¹. This prompted the authors to look for putative factors that specify HSC identity, and they reasoned that these would be preferentially expressed by HSPCs in the embryo proper. EVI1 was identified as one such factor, based on gene expression data and analysis of Evi1^GFP reporter mice.

Yokomizo et al.¹⁰ observed that HSCs express the highest levels of EVI1 among all fetal hematopoietic cells. Thus, they asked whether EVI1 was not simply a marker but also had a functional role in HSC specification. To this aim, they analyzed mouse embryos lacking one Evi1 allele (Evi1^+/- mice), resulting in depletion of EVI1^hi cells, and found that the mice exhibited a marked reduction in HSCs, although other progenitors were unaffected. Moreover, the authors exploited differential expression of Evi1 in a new lineage-tracing strategy (using Evi1^creERT2 mice), which, accordingly, preferentially labeled HSCs. Using this line, they examined the relative dynamics of HSCs and their immediate downstream progeny, called short-term HSCs, in the fetal liver. These experiments showed that HSCs make little contribution to downstream, more lineage-committed progenitors up until the end of gestation, reinforcing the idea that HSCs minimally contribute to fetal hematopoiesis (Fig. 1).

The data by Yokomizo et al.¹⁰ indicate that HSCs do not establish a downstream hematopoietic hierarchy in the fetal liver, as was previously thought. Instead, a ‘precast’ fetal hematopoietic hierarchy is likely to be already imprinted in precursors seeding the fetal liver. This is in agreement with the heterogeneity seen already within intra-aortic hematopoietic clusters, with only 1 in 30 cells representing pre-HSCs¹². The reasons such a system has evolved in vertebrates make for an interesting topic of speculation. It is possible that it represents an efficient way for the embryo to quickly respond to its rapidly evolving needs, which require differentiated hematopoietic cells within a short time frame to ensure oxygen distribution, host immunity and other trophic roles, for example, via tissue-resident immune cells. At the same time, this strategy can ensure sufficient HSC expansion in preparation for their use in adult life. As other studies have shown that HSCs in adults exhibit a limited lympho-myelo-erythroid contribution during steady state^13,14, long-term HSCs may in fact represent a ‘reserve’ population, able to sense insults and activate in case of damage. This traces an interesting parallel with other tissue stem cells that are normally quiescent, such as muscle satellite cells.

Finally, ectopic expression of EVI1 in endothelial cells or in pre-HSPCs specified most progenitors generated by the hemogenic endothelium to a transplantable HSC fate, confirming that progenitor fate determination relies on the expression level of EVI1 in pre-HSPCs. The identification of EVI1 as an HSC-specifying factor has important translational implications. It is possible that directing EVI1 expression in induced pluripotent stem cell-derived hemogenic endothelial cells might provide an efficient strategy to improve HSC production in vitro. To this aim, given the well-established involvement of EVI1 dysregulation in hematological malignancies, it will be essential to unveil the detailed mechanisms that regulate EVI1 expression and function. Hence, several remaining questions still need to be addressed, among them: what are the factors upstream of EVI1 that direct hematopoietic precursors to an HSC fate? At what stage, exactly, does HSC specification take place? To tackle these issues, it will be key to devise new methods for prospective identification of HSC-fated cells in situ.