Revisiting the lineage contribution of hematopoietic stem and progenitor cells

During vertebrate ontogeny, developmental hematopoiesis takes place in discrete anatomical sites with highly dynamic characteristics. There are sequential waves of hematopoiesis, each producing different lineages that increase in their complexity and diversity of blood lineage potential (Costa et al., 2012; Dzierzak and Bigas, 2018). The earliest two waves of hematopoiesis mainly produce primitive hematopoietic cells, lineage-restricted progenitor cells and multipotent progenitor cells (MPPs) in a hematopoietic stem cell (HSC)-independent manner (Dignum et al., 2021; Frame et al., 2013; Inlay et al., 2014; Palis et al., 1999; Patel et al., 2022; Soares-da-Silva et al., 2021). In mice, the primitive hematopoietic cells, such as erythrocytes, megakaryocytes and myeloid cells, are found in the yolk sac blood islands derived from extra-embryonic mesoderm at embryonic day (E) 7.5 and primarily arise from unipotent hematopoietic progenitor cells (HPCs) (Iturri et al., 2021; Palis, 2016; Soares-da-Silva et al., 2021). Erythrocytes transport oxygen to support quickly expanding embryos (Palis, 2008), whereas megakaryocytes and myeloid cells are important for tissue remodeling and homeostasis (Fig. 1A) (Dzierzak and Bigas, 2018; Tober et al., 2007). Shortly after the primitive wave, at around E8.25, the pro-definitive wave of hematopoiesis occurs: erythro-myeloid progenitors (EMPs), which display both the erythroid and myeloid differentiation abilities, directly derive from yolk sac hemogenic endothelium (HE) (Kasaai et al., 2017). Growing evidence has shown that primitive erythrocytes and EMP-derived erythrocytes are maintained throughout gestation and that EMPs contribute significantly to erythrocyte output until birth (Fraser et al., 2007; McGrath et al., 2015; Soares-da-Silva et al., 2021). EMP-derived macrophages are the common origin of tissue-resident macrophages in the liver (Kupffer cells), epidermis (Langerhans cells), brain (microglia) and lung (alveolar macrophages) from embryonic stages to adulthood (Gomez Perdiguero et al., 2015). In addition, other HSC-independent lineage-restricted progenitor cells have been identified during embryonic development in vertebrates. For example, in zebrafish, lymphoid-erythroid progenitor cells and lymphoid-myeloid progenitor cells, which arise from aortic HE, maintain major developmental hematopoiesis (Ulloa et al., 2021). In mice, an immune-related hematopoietic wave is also detected in the yolk sac, which can produce lymphoid-myeloid progenitor cells and B-1a cells (Boiers et al., 2013; Hadland and Yoshimoto, 2018; Inlay et al., 2014; Yoshimoto et al., 2011). Furthermore, MPPs have also been identified before, or concurrently with, HSC generation; e.g. recently it has been shown that the earliest MPPs emerge at E9-E10 from the para-aortic splanchnopleura/aorta-gonad-mesonephros (AGM) and are derived from Cxcr4-negative, progenitor-restricted HE in mice (Dignum et al., 2021). Similarly, a population of murine embryonic MPPs (eMPPs), which are labeled by the Flt3 (FMS-like tyrosine kinase 3)-Cre reporter line, are generated through endothelial-to-hematopoietic transition (EHT) in the yolk sac and AGM at E10.5, and these eMPPs substantially contribute to hematopoietic maintenance from embryonic stages to young adulthood (Fig. 1B) (Patel et al., 2022). Together, non-HSC derived HPCs and mature cells are necessary and sufficient to sustain hematopoiesis until birth or even young adult stage (Chen et al., 2011; Palis, 2016).

HSC activity is detected in the third wave of definitive hematopoiesis, having the important abilities of self-renewal and multipotent differentiation for long-term hematopoietic maintenance. Live imaging and lineage tracing analyses in vertebrates have shown that embryonic HSCs arise from HE through EHT (Bertrand et al., 2010; Boisset et al., 2010; Kissa and Herbomel, 2010; Zovein et al., 2008). At least two HSC-competent intermediates have been defined as the precursors of HSC during EHT (Table 1): the type I pre-HSCs (CD144⁺CD45⁻CD41^lowCD43⁺) and type II pre-HSCs (CD144⁺CD45⁺CD41^lowCD43^high). These HSC precursors will then mature into HSCs, which involves the simultaneous acquisition of hematopoietic potential and loss of the endothelial potential (Rybtsov et al., 2011; Taoudi et al., 2008; Zhou et al., 2016). In mice, HSCs and their immature precursors are observed within the intra-aortic hematopoietic clusters (IAHCs) of the para-aortic splanchnopleura, AGM region, vitelline arteries and umbilical arteries between E10.5 and E11 (Fig. 1C) (Costa et al., 2012; de Bruijn et al., 2000; Godin et al., 1995; Gordon-Keylock et al., 2013; Mikkola and Orkin, 2006; Nishikawa et al., 2001). From E11.5, HSC precursors and HSCs are released into circulating blood and gradually migrate into the fetal liver, where HSC precursors undergo maturation and HSCs differentiate into HPCs, including MPPs and lineage-restricted progenitor cells, and downstream lineages (Fig. 1D) (Ganuza et al., 2022; Kieusseian et al., 2012; Mikkola and Orkin, 2006; Rybtsov et al., 2016). In the fetal liver, the pool of HSCs expands simultaneously with their differentiation from E12.5 to E16.5 (Ema and Nakauchi, 2000; Swain et al., 2014). Subsequently, HSCs migrate into the bone marrow for maintenance throughout the adulthood from E17.5 onwards. Bone marrow HSCs are classified into two subpopulations: long-term (LT) HSCs and short-term (ST) HSCs. Compared with LT-HSCs, ST-HSCs exhibit a gradual decline in their self-renewal and reconstitution abilities (Morrison and Weissman, 1994; Yang et al., 2005).

Hematopoiesis has long been considered to be sustained by HSCs based on their self-renewal capability (i.e. the generation of HSCs by symmetric division) and multipotency (i.e. the generation of mature hematopoietic lineage cells by asymmetric division or differentiation) (Eaves, 2015; Seita and Weissman, 2010; Weissman and Shizuru, 2008). These functional properties are exemplified by the ability of HSCs to reconstitute the entire hematopoietic system upon transplantation into lethally irradiated recipients, and HSC transplantation is a well-accepted therapy choice for the treatment of malignant and non-malignant hematopoietic diseases (Balassa et al., 2019; Barriga et al., 2012). Studies have shown that the two functional properties of HSCs are executed by different cell states: quiescent HSCs (i.e. LT-HSCs) serve as a reserve of HSC pool, while active HSCs (i.e. ST-HSCs) stepwise differentiate to progenitor cells and mature blood lineages (Arai, 2016; Haas et al., 2018; Wilson et al., 2008). However, it remains unknown whether other means to coordinate these two functional properties exist, especially at embryonic stages, when only a small number of HSCs are generated. Previous studies estimated that the number of cells in the IAHCs of AGM region peaks at E10.5 (∼700 cells per embryo). At E11.5, there are ∼550 HSCs/HPCs generated in the IAHCs per embryo and, among them, one or two functional HSCs become validated through transplantation assay (Dzierzak and Bigas, 2018; Ganuza et al., 2017; Kumaravelu et al., 2002; Medvinsky and Dzierzak, 1996; Muller et al., 1994; Yokomizo and Dzierzak, 2010).

The emergence of HSCs and HPCs (especially embryo-derived MPPs) overlaps spatio-temporally. Therefore, it is important to determine the phenotypic parameters with the specific signatures of HSCs/HPCs. The development of integrated approaches for characterizing HSC molecular signatures and lineage tracing of HSC fates in vivo (Bowling et al., 2020; Weinreb et al., 2020) has transformed our perception of HSC biology and also suggests that the unperturbed blood system is predominately sustained by HPCs. Here, we review the classical and recent viewpoints about the developmental origins, molecular signatures and functional properties of HPCs and HSCs.

Studies on HSC biology began in the mid-20th century. In 1945, studies of fraternal twin cattle with a fused placental circulation during the embryonic period showed that the fraternal twins share the same blood cell types throughout their lives. Importantly, these embryonic blood cells are capable of establishing hematopoietic system and of contributing to the adult blood system in their co-twin hosts (Owen, 1945). This observation indicates that blood circulation contains long-lived blood progenitor cells. In 1961, by performing bone marrow cell transplantation experiments in sub-lethally irradiated mice, Till and McCulloch demonstrated that the transplanted cells are able to produce different types of colonies (granulocytes, macrophages, erythrocytes and megakaryocytes) via cell division and differentiation in the spleen of recipients (Siminovitch et al., 1963; Till and McCulloch, 1961; Wu et al., 1967, 1968). Based on these crucial findings, Till and McCulloch proposed the core concept of HSC (i.e. a hematopoietic cell with clonality). Thus, later studies mainly focus on the purification of HSCs from murine bone marrow. Based on different surface marker combinations (Table 1), HSCs and various HSC-derived HPCs can be sorted to evaluate their performances after transplantation. Given that some HSC-derived HPCs (e.g. MPPs) are able to maintain hematopoiesis in the recipients for a long period of time, it is necessary to perform serial transplantations for the evaluation of multilineage reconstitution and self-renewal abilities, respectively (Giebel and Bruns, 2008; Morrison and Weissman, 1994; Na Nakorn et al., 2002; Yang et al., 2005). Specifically, HSCs are characterized by long-term (>8 weeks after primary transplantation) multilineage differentiation potential and self-renewal capability (after secondary transplantation), while HSC-derived HPCs are characterized by declined (<8 weeks after primary transplantation) multilineage differentiation potential and loss of self-renewal capability (after secondary transplantation). Notably, these findings show that the hematopoietic colonies observed in the spleens (≤2 weeks after primary transplantation) in Till and McCulloch's experiments mainly come from HSC-derived HPCs (Magli et al., 1982; Na Nakorn et al., 2002). Taken together, these findings established the concept of HSCs, i.e. a rare population with multipotent differentiation potential and self-renewal ability for lifelong hematopoiesis (Morrison and Weissman, 1994; Spangrude et al., 1988), and demonstrated the biological function and purification scheme of HSCs.

To illustrate the lineage relationship between HSCs and their progenies, the hematopoietic hierarchy has been proposed as a tree-like model (Akashi et al., 2000; Reya et al., 2001). In this model, HSCs sitting at the top can differentiate into all downstream lineage cells, including MPPs, lineage-restricted progenitor cells, and mature myeloid, lymphoid and erythroid-megakaryocyte lineages (Eaves, 2015; Kondo et al., 1997; Morrison et al., 1997; Pietras et al., 2015) (Fig. 2A). Extensive studies have revised the hematopoietic hierarchy: intermediate-term (IT) HSCs have been identified and sit between LT-HSCs and ST-HSCs (Benveniste et al., 2010); MPPs have been classified into MPP1, MPP2, MPP3 and MPP4 (Pietras et al., 2015; Wilson et al., 2008); and lineage-biased progenitor cells have been found and directly generated from HSCs (Yamamoto et al., 2013). Nonetheless, HSCs are still considered a major source to produce all blood lineage cells during embryogenesis and in adulthood (Zhang et al., 2018).

The hematopoietic hierarchy mainly illustrates the HSC functional behaviors under stress conditions, including transplantation, injury or cell culture in vitro (Zhao and Baltimore, 2015). In perturbed hematopoiesis, HSCs likely execute their lineage potential, but this might not reflect their native contributions under normal conditions. For example, the initial observations of multipotent lineage outcomes from HSCs in the bone marrow are obtained from the serial transplantation experiments (Eaves, 2015; Weissman and Shizuru, 2008). However, later studies revealed that HSCs in steady-state preferentially adopt a fate towards particular lineages (i.e. the megakaryocyte lineage) but not all blood lineages (Carrelha et al., 2018; Rodriguez-Fraticelli et al., 2018). Therefore, it is essential to explore and compare the HSC function under different conditions, which will help to achieve a better understanding of normal hematopoiesis and hematopoiesis under stress.

Owing to the development of new lineage tracing methods, it is now possible to study the native properties of HSCs under physiological conditions. Based on the introduction of random mutations into genomic DNA sequences, DNA barcoding has been developed for lineage tracing via DNA sequencing (Gerrits et al., 2010). DNA barcodes, mediated by viral infection system (Lu et al., 2011), Polylox recombination system (Pei et al., 2017) or CRISPR/Cas9-mediated genome editing system (McKenna et al., 2016), would represent different types of genetic scars, which consist of a mass of random sequence tags. These barcodes can be identified through sequencing and are continuously heritable from founder cells to their progenies during cell division and cell differentiation. Thus, the founder cells and their progenies share the identical and unique DNA barcodes, while different lineages can be clearly distinguished. Among these approaches, the viral genetic barcoding system delivers a library of random nucleotides into target cells, which are transplanted into recipients for lineage tracing. This system provides diversity of labeling and is commonly applied to evaluate the stem cell clonal dynamics (e.g. HSC or cancer stem cell) (Gerrits et al., 2010; Lu et al., 2011; Nguyen et al., 2014). Although the viral genetic barcoding system can identify the lineage relationship, it cannot provide transcriptomic information. Therefore, the molecular signatures for cell identities/states are not precise; meanwhile, this system cannot achieve cell labeling at a given timepoint and in a variety of tissues. To overcome the limitations, the Polylox barcoding system has been developed. In this system, multiple loxP sites with alternating orientations form an intact Polylox barcode and Cre recombination events happen between any two loxP sites. The DNA fragment between two loxP sites will be excised when the two loxP sites are the same orientation, whereas the DNA fragment will be inverted when the two loxP sites are the opposite orientation. Thus, the random recombination of Polylox fragments can generate millions of distinct barcodes. Based on cell-type-specific Cre-loxP inducible recombination system, it is widely used to study cell fate decisions during cell division, cell differentiation or tissue regeneration at the given time point in the target cell types. However, limited or incomplete Cre enzyme activity when confronted with a variety of loxP sites will reduce barcode diversity. In addition, this system allows only a single round of cell labeling and thus it cannot construct multi-layer phylogenetic trees (Chen et al., 2022; Kester and van Oudenaarden, 2018; Pei et al., 2017). Another recently developed barcoding system is the CRISPR/Cas9-mediated genome editing system. Targeting specific endogenous genome loci by guide RNA-Cas9 complex can introduce the insertion-deletion mutations, which act as the traceable elements from founder cells to their descendants. The persistent activity of Cas9 results in the cumulation of edits within barcodes over time, which allows the construction of a multi-layer phylogenetic tree from early embryonic stages to late adult stages. Importantly, several lineage recording technologies based on CRISPR/Cas9 editing and single-cell transcriptomics simultaneously measure the cell type characterization and lineage relationship, which facilitates the understanding of the developmental branches, lineage histories and gene expression cascades in multicellular organisms (McKenna et al., 2016; Spanjaard et al., 2018). Using these systems, HSCs at the top position in the hematopoietic hierarchy are capable of differentiating into all blood lineages (Lu et al., 2011; Pei et al., 2017). However, the extent of the lineage contribution of HSCs to the entire hematopoietic system is not yet clear.

Following the identification of HSC- and HPC-specific markers (Box 1; Table 1), it is possible to evaluate the hematopoietic contribution of HSCs and HPCs. Given that Tie2 (tyrosine-protein kinase receptor) is expressed in embryonic and adult HSCs, the Tie2^MCM/+Rosa^YFP-mediated inducible mouse model has also been used to trace the in vivo behaviors of LT-HSCs and their differentiated progenies, including ST-HSCs, MPPs and lineage-restricted progenitor cells. This finding shows that the properties of HSCs in native hematopoiesis are different from that of transplanted HSCs (Busch et al., 2015). Through the combination of transplantation and limiting dilution analysis, ∼30% LT-HSCs [characterized by the lineage cocktail (Lin)⁻Kit⁺Sca-1⁺CD150⁺CD48⁻ marker combination] are active and contribute to overall hematopoiesis. However, in the steady state, the differentiation frequency of LT-HSCs (an average of 1/110 of LT-HSCs differentiate to ST-HSCs per day) is lower than that of ST-HSCs (an average of 1/22 of ST-HSCs per day differentiate into four types of MPPs) and MPPs (an average of 1/46 of MPPs per day differentiate into common lymphoid progenitors and 1 MPP per day differentiates into four common myeloid progenitors). These data indicate that adult hematopoiesis is largely sustained by ST-HSCs and MPPs, not LT-HSCs. Furthermore, in another study, a transposon tagging-based lineage tracing model has been used to study the blood system in the native state in mouse bone marrow (Rodriguez-Fraticelli et al., 2018). In the physiological state, LT-HSCs (defined as Lin⁻Kit⁺Sca-1⁺Flt3⁻CD150⁺CD48⁻) stay in a slow cell cycle state, and exclusively contribute to the megakaryocyte lineage through a direct differentiation pathway. Whereas ST-HSCs and MPP compartments (MPP1: Lin⁻Kit⁺Sca-1⁺Flt3⁻CD150⁻CD48⁻ MPP2: Lin⁻Kit⁺Sca-1⁺Flt3⁻CD150⁺CD48⁺; MPP3: Lin⁻Kit⁺Sca-1⁺Flt3⁻CD150⁻CD48⁺; MPP4: Lin⁻Kit⁺Sca-1⁺Flt3⁺CD48⁺) mainly contribute to lymphoid, erythroid and myeloid lineage cells. Other studies have also reported LT-HSC as a significant cell source for megakaryocyte lineage without going through intermediate progenitor cell stages (Carrelha et al., 2018; Sanjuan-Pla et al., 2013). Taken together, in their native state, HPCs largely contribute to adult hematopoiesis in the bone marrow.

During development, rapid growth of embryos requires expansion of hematopoietic cell pools and establishment of hematopoietic hierarchy, which is thought to be attributed to the self-renewal and multilineage differentiation abilities of HSCs. However, using innovative methodology, several recent studies have shown that embryonic hematopoiesis is sustained by HPCs in an HSC-independent manner.

Owing to the temporally distinct expression patterns of zebrafish draculin gene, it is possible to distinguish HSCs from embryonic progenitor cells and show that the differentiation timepoint of terminal blood cells from HSCs occurs later than that from embryonic progenitor cells, including lymphoid-myeloid progenitor cells and lymphoid-erythroid progenitor cells (Ulloa et al., 2021). Importantly, this study suggests that embryonic HSCs and HPCs have different lineage differentiation capabilities and kinetics, and that embryonic and early larval hematopoiesis is mainly sustained by the HPC pool and not by HSCs. Recent studies in mice have also confirmed this view: transposon-based cell fate mapping and in situ barcoding has been used to assess the developmental origins of embryonic and adult blood system (Patel et al., 2022). After Dox induction at E9.5, labelled hematopoietic cells were observed to derive not only from HSC clones, but also from eMPPs. During embryonic development, eMPP clones were predominant and contributed to about 60% of mature lineage cells, whereas 9% of mature lineage cells were derived from HSC clones. To investigate the potential biases and lineage output changes with time, the clonal contribution of HSCs and eMPPs from both young and old mice was measured, revealing that (1) eMPPs are the predominant source of lymphoid lineage cells in the young adult; (2) the lymphoid lineage contribution of eMPPs decreases during aging; and (3) HSCs are the main source of myeloid lineage output and become more prominent with age. This study indicates that eMPPs predominantly contribute to hematopoiesis from embryos to young adult stage. Furthermore, the hematopoietic contribution during embryonic development has also been analyzed using an in vivo genetic tracing mouse model (Yokomizo et al., 2022). Using Hlf (hepatic leukemia factor) as a marker of HSC and HPC precursors (pre-HSPCs) in embryonic IAHCs, it has been shown that the Hlf⁺ pre-HSPCs independently mature into Lin⁻Kit⁺Sca-1⁺CD150⁺CD48⁻ HSCs, Lin⁻Kit⁺Sca-1⁺CD150⁻CD48⁻ ST-HSCs and Lin⁻Kit⁺Sca-1⁺CD150⁻CD48⁺ MPPs in the fetal liver. Thus, hematopoietic hierarchy in the fetal liver is not established via stepwise hierarchical differentiation of LT-HSCs, but instead is formed mainly through maturation of Hlf⁺ pre-HSPCs and differentiation of ST-HSCs/HPCs. Furthermore, the knockdown of HSC-specific gene Evi1 (a marker of hematopoietic precursors in the AGM with the potential to preferentially generate HSCs) leads to the significant decrease in HSC number, but does not affect the number and function of HPCs. These data indicate that fetal HPCs mainly contribute to the generation of terminal blood cells during embryonic development (Fig. 2B).

Taken together, these findings show that, under native conditions, HPCs largely contribute to both embryonic and adult hematopoiesis. By contrast, HSCs can reprogram to produce megakaryocyte lineage, despite maintaining the multilineage differentiation potential (Busch et al., 2015; Carrelha et al., 2018; Patel et al., 2022; Rodriguez-Fraticelli et al., 2018; Yokomizo et al., 2022). Importantly, these studies demonstrate that the self-renewal and multilineage differentiation properties of HSCs are likely not executed simultaneously by the same cell type in unperturbed hematopoiesis. The new findings on stem cell-independent lineage contribution have greatly updated our understanding of both developmental and adult hematopoiesis (Fig. 3).

The classical hematopoietic hierarchy model including HSCs, multiple progenitor cells and terminal blood cells was originally proposed to illustrate the HSC functional outputs, such as multipotent differentiation ability, mainly reflecting the lineage potential of HSCs (Seita and Weissman, 2010; Weissman, 2002). Recently, extensive studies using in vivo lineage-tracing systems have revealed the native fates of HSCs and HPCs during embryogenesis and in adulthood. Based on different experimental strategies, we can conclude that HSCs show different functional properties under stress and native conditions. Under stress conditions (e.g. post-transplantation or post-chemotherapy), the re-activated multipotency of HSCs is beneficial for reconstitution of impaired blood systems; however, under native conditions, HSCs are in a quiescent state for reducing replication-induced DNA damage and oxidative stress, and minimally contribute to differentiation outputs or biasedly produce specific blood lineages (e.g. megakaryocyte lineage). More importantly, HPCs, including HSCs derived and/or non-HSCs derived, are a major contributor to embryonic and life-long hematopoiesis (Fig. 3). Together, these findings suggest that it is the time to revisit the potential values of specialized HPCs in translational medicine. Furthermore, these advances have also revised the classical hematopoietic hierarchy model in embryonic and adult stages (Fig. 2).

The blood system is maintained by both HSC-dependent and -independent pathways; HPCs, as a HSC-independent pathway, predominantly maintain hematopoiesis from embryonic to young adult stages. It will be interesting to investigate the underlying biological significance and molecular mechanisms in the future. Given that the onset of stem cell differentiation is tightly associated with cell cycle entry (Cockburn et al., 2022; Pauklin and Vallier, 2014), we reason that, under native conditions, the difference in cell fates between HSCs and HPCs, is partially attributed to the difference in cell cycle states. During nascent HSC generation in mouse embryos, HSC precursors are proliferating with increased transcriptional, biosynthetic and translational activities. When these precursors further undergo maturation, HSCs exhibit a slow-cycling state (Batsivari et al., 2017). In the fetal liver, a large number of HPCs emerge, which are derived from extra-embryonic and intra-embryonic hematopoietic sites. A recent study reveals that fetal liver HPCs show more active cell division and a more robust differentiation ability than HSCs (Ganuza et al., 2022). In the bone marrow, LT-HSCs prefer to remain in a quiescent state. Throughout adult life, HSCs maintain their long-term repopulating ability until the fifth division and undergo self-renewal with progressively lengthening periods (Bernitz et al., 2016). The limited cell division number of HSCs is insufficient to support their ‘proposed’ major contribution to hematopoiesis. In comparison, owing to the high differentiation frequency and active cell cycle state, HPCs can make a major contribution to both embryonic and adult hematopoiesis (Busch et al., 2015; Rodriguez-Fraticelli et al., 2018).

Given that HSCs and HPCs emerge concurrently, it is crucial to distinguish precisely between HSCs and HPCs via molecular signatures, functional outputs and spatial locations in the future studies. In the traditional lineage-tracing strategies, cell type is identified based on signature marker combinations, which will potentially ignore or confuse some rare but important cell types. Single-cell transcriptomics technologies can provide the precise molecular signatures of individual cells and are required for accurate cell type identification. Recently, the integrated strategies of single-cell transcriptomics and lineage tracing technologies have been reported, such as LINNAEUS (lineage tracing by nuclease-activated editing of ubiquitous sequences), scGESTALT (single-cell genome editing of synthetic target arrays for lineage tracing) and ScarTrace (Alemany et al., 2018; Raj et al., 2018; Spanjaard et al., 2018). These combinations will help to simultaneously identify cell types and record cell genetic lineage relationship, which can provide a detailed insight of cell fate decision. Besides, single-cell transcriptomic analyses are widely used to construct unbiased cell-cell interaction networks, which can help to identify the important microenvironmental cells and potential regulatory function.

Although these issues still need to be addressed, these updated views provide a more comprehensive understanding of HSC function and hematopoietic hierarchy, and provide useful insights into further studies on both basic research and clinical applications. For example, the new finding that the proportion of MPPs with enhanced lymphoid production decreases during aging may explain the diminished immune protection in the older people, in which the most recognized changes for the immune system are the involution of the immune tissues and decreased output of lymphocytes (Nikolich-Zugich, 2018; Weiskopf et al., 2009). Thus, application of these lymphoid-primed MPPs will provide new clues and cellular source for the treatment of the elderly immune system. Besides, given that HSC generation in vitro is difficult to achieve and HPCs show more powerful functions in vivo than previously thought, these HPCs appear to be the target cells for in vitro induction and also ideal cellular sources for clinical applications to treat severe hematological diseases.

We thank our lab members for critical reading of this manuscript.