Common Features of Megakaryocytes and Hematopoietic Stem Cells: What's the Connection?

Summary

Megakaryocytes are rare polyploid bone marrow cells whose function is to produce blood platelets. Since the purification and cloning of the major megakaryocyte cytokine, thrombopoietin, in 1994, considerable progress has been made in understanding the biology of megakaryocyte development. Remarkably, these advances have revealed a number of key features of megakaryocytes that are shared with hematopoietic stem cells (HSCs), such as common surface receptors, lineage-specific transcription factors, and specialized signaling pathways. Why there should be such a close connection between these two cell types remains unclear. In this Prospects article, we summarize the data supporting these shared features and speculate on possible teleological bases. In particular, we focus on common links involving developmental hierarchy, endothelial cells, and bone marrow niche interactions. This discussion highlights new data showing close ontologic relationship between HSCs and specialized “hemogenic” endothelial cells during development, and functional overlap between megakaryocytes/platelets and endothelial cells. Overall, these findings may be of relevance in the development of techniques for HSC ex vivo culture and/or possible generation of HSCs via somatic cell reprogramming.

The hematopoietic system provides a constant supply of blood cells of diverse function. This is a highly active process, with nearly a trillion new blood cells produced every day in an adult human. The ability to sustain this activity throughout the lifetime of an individual is dependent on hematopoietic stem cells (HSCs). These rare, specialized cells are characterized by self-renewal capacity, high proliferative potential, and the ability to differentiate into at least ten distinct blood cell types. Since the identification of HSCs by Till and McCullough in 1961, considerable knowledge has been gained about their biology. In particular, key genes have been identified that are required for their ontogeny, maintenance, and function.

Platelets are anucleate blood elements that play essential roles in hemostasis. Resting platelets circulate as small discoid structures. Upon activation, they undergo dramatic changes that include flattening, spreading, membrane surface alterations, and release of vasoactive and thrombogenic substances. In addition to acting as the “first line of defense” at the time of vascular injury, they serve to repair microscopic vascular damage that occurs on a daily basis. Platelets are also thought to play roles in wound healing and angiogenesis through the delivery of stored growth factors, such as PDGF, bFGF, VEGF, platelet factor-4, TGF-β, thrombospondin-1 to sites of injury.

Platelets are derived directly from megakaryocytes (Mks), which are rare, large, polyploid cells that reside predominantly within the bone marrow. During terminal stages of maturation, megakaryocytes produce long cytoplasmic extensions called proplatelets, which then release platelets from their tips. Because of the extreme rarity of megakaryocytes, accounting for <0.05% of nucleated bone marrow cells, an understanding of the molecular biology of megakaryocyte development has lagged behind many of the other blood lineages. However, the purification and cloning of the major cytokine for megakaryopoiesis, thrombopoietin, in 1994, has enabled efficient culturing of these cells and more detailed analysis of their biology. As more has been learned about the key factors involved in megakaryocyte development, and as studies of HSCs have progressed, one of the surprises has been that these two cell types share a remarkable number of key cell-specific factors. Why this should be remains unknown. In this Prospect article, we first review the common features of megakaryocytes and HSCs, and in the second part, offer possible explanations and future research directions.

Shared Features of Megakaryocytes and Hematopoietic Stem Cells

Cell Surface Receptors

Thrombopoietin (TPO) Receptor (c-mpl)

One of the first noted similarities between Mks and HSCs was the finding that both express the thrombopoietin (TPO) receptor (c-mpl) on their cell surface. TPO receptor expression is functionally important in HSCs activity, as bone marrow transplant studies show that HSCs from c-mpl^−/− mice have a strong disadvantage in long-term repopulating activity compared to those from wild type mice, even when injected at ten times greater number [Kimura et al., 1998]. Likewise, knock out of TPO itself reduces HSC function [Murone et al., 1998]. In humans, germline inactivating mutations of the TPO receptor gene cause severe congenital amegakaryocytic thrombocytopenia (CAMT). Children with this disorder are at high risk of developing complete bone marrow failure, typically within the first decade of life [van den Oudenrijn et al., 2000]. This finding is consistent with a role for TPO signaling in human HSC maintenance.

CXCR4

The receptor CXCR4 is a seven membrane spanning G-coupled receptor that binds the chemokine SDF-1. It plays an important role in HSC bone marrow retention [Mohle et al., 1998]. In fact, CXCR4 antagonists have recently been developed for use as mobilization agents in the collection of peripheral blood HSCs from stem cell transplant donors. CXCR4 is also expressed on Mks, and plays a critical role in homing of Mks to the vascular sinusoidal space, where terminal Mk maturation and proplatelet formation occurs [Wang et al., 1998].

CD150

Morrison and colleagues recently identified the SLAM receptor molecule CD150 as a surface marker that enriches for cells with HSC activity when used in conjunction with previously characterized markers [Kiel et al., 2005]. CD150 is also expressed abundantly on committed megakaryocyte progenitors (MkPs) [Pronk et al., 2007] and on mature primary murine fetal liver Mks (H. Huang and A. Cantor, unpublished observation). In fact, protocols incorporating CD150 for the immunophenotypic isolation of murine HSCs and MkPs are now widely used.

CD41 (GPIIb)

The integrins CD41 (GPIIb) and CD61 (GPIIIa) form a heterodimeric complex, which is present at high copy number on the surface of platelets. The GPIIb/IIIa complex has been extensively studied with respect to its function as a receptor for fibrinogen during platelet aggregation. CD41 expression was initially thought to be unique to the megakaryocyte lineage. However, it is now clear that CD41 is also expressed at lower levels on early multipotential hematopoietic progenitors cells, and, in fact, is one of the earliest markers of hematopoietic development [Mikkola et al., 2003a]. Some studies suggest that it is also contained on populations of cells with HSC activity [Debili et al., 2001].

Other Receptors

Several other cell surface markers have also been reported to be shared between Mks and HSCs, including c-kit, CD34, CD105 (Endoglin), CD31 (PECAM-1), JAM-A, Tie-2, and KDR (VEGF receptor 2).

Transcription Factors

Runx-1

Runx-1 (formerly known as AML-1, Cbf-α2, and PEBPA2B) is a hematopoietic-expressed member of the runt family of transcription factors. These factors share a highly conserved 128 amino acid domain that was first identified in the drosophila runt gene. This region mediates sequence-specific DNA binding, as well as protein-protein interactions with the cofactor molecule Cbf-β. Runx1^−/− mice die between embryonic day 12.5 to 13.5 (e12.5–13.5) from central nervous system hemorrhage and failure of all definitive hematopoiesis [Wang et al., 1996]. The latter is due to impaired emergence of the first definitive HSCs from aorto-gonadal-mesonephros (AGM) region during embryogenesis [North et al., 1999]. In adult mice, Runx-1 haploinsufficiency causes a decrease in the number of HSCs (long-term repopulating cells) [Sun and Downing, 2004]. Conditional knock out studies using the Mx1-Cre system, which allows inducible deletion of targeted alleles within the bone marrow, also show a requirement for Runx1 during megakaryocyte differentiation in adult animals [Growney et al., 2005; Ichikawa et al., 2004]. Runx-1 deficient megakaryocytes have hypolobulated nuclei, underdeveloped cytoplasm, low DNA ploidy and enhanced replating activity in semisolid media culture assays. Haploinsufficiency of Cbf-β also perturbs megakaryopoiesis in mice [Talebian et al., 2007]. In humans, heterozygous Runx-1 germline mutations cause Familial Platelet Disorder with Propensity to Develop AML (FPD/AML), a rare autosomal dominant disorder characterized by thrombocytopenia, platelet dysfunction, and markedly elevated risk for myelodysplastic syndrome (MDS) and leukemia (median incidence ~35%) [Song et al., 1999].

GATA-2

GATA family transcription factors play essential roles during development and in the maintenance of certain adult tissues. There are six known family members in mammals, three of which (GATA-1, -2, and -3) are expressed in hematopoietic tissues. In general, GATA-1 and GATA-2 play reciprocal roles during hematopoiesis, with GATA-2 required for early multipotent stages, and GATA-1 required for terminal maturation of certain lineages. Complete loss of GATA-2 in mice causes a marked defect in HSCs and multipotential progenitor cell expansion resulting in early embryonic lethality [Tsai et al., 1994]. Haploinsufficiency of GATA-2 in adult mice results in abnormal HSC homeostasis [Rodrigues et al., 2005]. GATA factors also play crucial roles in megakaryopoiesis. GATA-1 is required for terminal megakaryocyte maturation and is mutated in X-linked macrothrombocytopenia and acute megakaryoblastic leukemia in children with Down syndrome [Cantor, 2005]. GATA-2 plays an overlapping functional role with GATA-1 during early stages of megakaryopoiesis, both of which require interaction with the cofactor Friend of GATA-1 (FOG-1) [Chang et al., 2002].

Evi-1

The ecotropic viral integration site-1 (Evi-1) is an oncogenic transcription factor in murine and human myeloid leukemia. Evi-1^−/− embryos have markedly reduced numbers of phenotypic HSCs [Goyama et al., 2008]. Functional studies using the Mx1-Cre conditional knock out system show defective HSC self-renewing activity and repopulating capacity. Interestingly, the mice also develop marked thrombocytopenia, but no significant alteration in total white blood cell counts or hemoglobin levels [Goyama et al., 2008]. The mice also show selective delayed recovery of platelets following treatment with the cytotoxic agent 5-fluorouracil (5-FU). Collectively, these findings suggest lineage-specific functional roles for Evi-1 in both HSC and megakaryocyte development/maintenance. Some of the hematopoietic effects of Evi-1 may be mediated through GATA-2 [Yuasa et al., 2005] and/or Runx-1 [Senyuk et al., 2007].

Ets Family Transcription Factors

Ets family transcription factors shared a conserved DNA binding motif, and often integrate cell signaling events with transcriptional activity. At least 28 different Ets family members have been identified and play diverse developmental roles. Several members have been shown to be involved in HSC ontogeny/maintenance and megakaryopoiesis. TEL/ETV6 is frequently rearranged in human leukemias of myeloid or lymphoid origins. TEL^−/− mice are embryonic lethal due to a yolk sac angiogenic defect. Conditional knock out studies using the Mx1-Cre system in adult mice show a dramatic loss of HSCs and lineage-selective impairment of megakaryocyte maturation [Hock et al., 2004].

Fli-1 was first identified as a gene product activated by Friend viral complex insertion in murine erythroleukemia cells [Ben-David et al., 1991]. Knock-out of the Fli-1 gene in mice results in dysmegakaryopoiesis and impaired vascular integrity, leading to death at day e11.5–12.5 from hemorrhage [Hart et al., 2000; Spyropoulos et al., 2000]. Fetal livers from the knockout animals have a marked decrease in total colony forming cells. In humans, heterozygous loss of the Fli-1 gene causes a congenital macrothrombocytopenia associated with Jacobsen or Paris-Trousseau syndrome [Hart et al., 2000; Raslova et al., 2004].

SCL/TAL1

SCL/TAL1 is a basic helix-loop-helix transcription factor that binds E-Box elements as a heterodimer with E-protein transcription factors. SCL^−/− mice die during early embryogenesis with a “bloodless” phenotype [Porcher et al., 1996]. Loss of SCL during adult hematopoiesis results in impaired megakaryopoiesis and erythropoiesis, but normal myelopoiesis and lymphopoiesis [Mikkola et al., 2003b].

SCL/Runx1/GATA-2/Ets Factor Transcriptional Network

Recent work from several groups indicates that the transcription factors discussed above act in a cross- and auto- regulatory transcriptional network (Fig. 1). de Bruijn and colleagues identified a key enhancer element of the Runx-1 gene that is controlled by GATA-2/Lmo2/Ldb-1 /SCL/Ets factors during HSC ontogeny [Nottingham et al., 2007]. Gottgens and colleagues found that key enhancers of the GATA-2, Fli-1, SCL, and Lmo2 genes are occupied by all four of these factors in vivo, and that these factors autoregulate themselves through a recursively wired gene regulatory subcircuit [Gottgens et al., 2002; Landry et al., 2009; Pimanda et al., 2007].

Fig. 1.

Transcription factor network involved in HSC ontogeny. Diagram indicating key transcription factors and their cross- and auto-regulatory circuits. Core factors are indicated in orange, and associated factors are shown in blue. Possible involvement of signaling through Notch and BMP4 pathways are shown in purple. Most of these same factors are specifically involved in megakaryocyte development.

In addition to these cross-regulatory loops, many of these factors physically interact with one another. GATA-1 forms a pentameric complex with SCL/E47/LMo2/Ldb1, and directly interacts with Fli-1 and Runx-1 [Eisbacher et al., 2003; Elagib et al., 2003; Wadman et al., 1997]. GATA-2 also interacts with Runx-1 (H. Huang and A. Cantor, unpublished observation). We recently demonstrated that Runx-1 and Fli-1 physically interact directly themselves, resulting in enhanced transcriptional activity at megakaryocyte-specific genes (H. Huang, A. Cantor, submitted). The combination of direct physical interaction and extensive cross/auto-regulatory relationships between small groups of key transcription factors is an emerging theme of gene regulatory networks [Kim et al., 2008]. Gottgens and colleagues have proposed that GATA-2/SCL/LMo2/Ets (and Runx-1) factors comprise a gene regulatory “kernel” involved in the specification of HSCs. This same “kernel” appears to be active in megakaryocyte development.

HOX-related Genes

In addition to the network factors described above, several other transcription factors have been shown to play key roles in both megakaryocyte and HSC development; notably, several homeobox genes and related proteins. Meis1 and Pbx1 are important cofactors for HOX gene function. Meis1 is expressed at high levels in the AGM mesenchyme, aortic endothelium, intra-aortic clusters, and HSCs [Azcoitia et al., 2005; Hisa et al., 2004]. Meis1-deficient embryos die between e11.5 to 14.5 of gestation from internal hemorrhage, liver hypoplasia and anemia [Azcoitia et al., 2005; Hisa et al., 2004]. The AGM region and fetal liver have markedly reduced colony forming potential. In addition, there is a complete agenesis of the megakaryocyte lineages and patterning defects of the vasculature.

Pbx1 is expressed in hematopoietic progenitors during murine embryonic development. Its absence results in severe anemia and embryonic lethality at e15 to 16 [DiMartino et al., 2001]. Pbx1^−/− embryos display impaired HSC function and reduced numbers of common myeloid progenitors and megakaryocyte- and erythrocyte-committed progenitors.

HOXA9 is expressed in CD34⁺ bone marrow cells and is inactivated as cells leave the CD34⁺ compartment. Loss of HOXA9 in mice leads to a defect in committed progenitors of the myeloid/erythroid and pre-B cell lineages [Lawrence et al., 1997]. Competitive repopulation assays show that the major defect in Hoxa9^−/− animals is a dramatic impairment of their HSCs to repopulate lethally irradiated recipients after bone marrow transplantation, suggesting a key role in early HSCs function [Argiropoulos and Humphries, 2007]. There are no reports about whether HOXA9 is expressed in Mks. However, Kaushansky and colleagues showed that TPO induces HOXA9 nuclear transport in immature hematopoietic cells [Kirito et al., 2004].

Rbm15 (OTT)

Rbm15 (also called OTT for “one: twenty-two”) was first identified as a fusion partner of MKL1 (also call MAL) in acute megakaryocytic leukemias harboring the recurrent chromosomal translocation t(1,22)(p13;q13). This translocation produces a fusion molecule that includes essentially all of the coding regions of both molecules. Rbm15 and MKL1, individually, have subsequently been shown to play important roles during normal megakaryopoiesis. Rbm15 conditional knock out mice display enhanced megakaryopoiesis in both spleen and bone marrow, compared to wild type mice (~5 and 2-fold, respectively)[Raffel et al., 2007]. Interestingly, these mice also have about a 50% increase in the size of the bone marrow Lin⁻Kit⁺Sca-1⁺ (LKS) population of cells, which is highly enriched in HSCs.

Signaling Pathways

Notch Signaling

Studies of drosophila germ cells have established a critical role for Notch signaling pathways in maintaining stem cell pluripotency. Although somewhat controversial, Notch signaling has also been shown to play roles in HSC development, and this appears to require the transcription factor Runx-1 [Burns et al., 2005; Maillard et al., 2008]. Recently, Gilliland and his colleagues showed that Notch signaling also positively regulates megakaryopoiesis both in vitro and in vivo [Mercher et al., 2008]. Activated Notch signaling was shown to markedly increase specification of megakaryocytes from HSCs in a co-culture system, and this effect could be abrogated by inhibition of Notch signaling either with gamma-secretase inhibitors or by expression of the dominant-negative Mastermind-like 1. Notch signaling has also been implicated in modulating GATA-2 expression [Robert-Moreno et al., 2005]. Thus, Notch may play an important role in stimulating the hematopoietic stem cell and megakaryocyte “gene regulatory kernel” (Fig. 2).

Fig. 2.

Hierarchal relationship among hematopoietic cell populations. Schematic diagram depicting the classical pathways of hematopoietic development. A potential direct relationship between ST-HSCs and MEPs is indicated by a dashed line. LT-HSCs, long-term repopulating hematopoietic stem cells; ST-HSCs, short-term repopulating hematopoietic cells; MPP, multipotent progenitor cells; CMP, common myeloid progenitor; CLP, common lymphoid progenitor; MEP, megakaryocyte-erythroid progenitor; GMP, granulocyte-macrophage progenitor; B, B-cell; T, T-cell; NK, natural killer cell. Adapted from [Adolfsson et al., 2005].

Role of Prostaglandins

Prostaglandins are lipid metabolites that play key roles in vascular function and platelet activity. Recently, a chemical screening study in zebrafish showed that agents that enhance prostaglandin E2 synthesis markedly enhance HSC expansion, and those that block prostaglandin synthesis decrease stem cell numbers [North et al., 2007]. Follow-up studies in mice show that PGE2 caused amplification of multipotent progenitors and increased the frequency of long-term repopulating HSCs present in murine bone marrow. Prostaglandins also play important roles in platelet function and are synthesized in megakaryocytic cell lines. In fact, aspirin's mode of action in inhibiting platelet function is via irreversible inactivation of the key prostaglandin metabolic enzyme cyclooxygenase.

Collectively, the findings highlighted above point towards a broad array of cell-specific factors that are shared between megakaryocytes and HSCs. These are summarized in Table 1.

Table 1.

Common features of murine megakaryocytes, HSCs, endothelial cells and hemangioblasts. N, not reported. LT-HSC, long term repopulating hematopoietic stem cells; ST-HSCs, short term repopulating hematopoietic cells.

	Megakaryocyte	HSC	Endothelial cell	Hemangioblast
Receptors
c-mpl	+	+	+	+
CXCR4	+	+	+	+
CD 150	+	+ (LT-HSC)	N	N
CD41	+	+	−	−
CD34	+	+ (ST-HSC)	+	+
CD 105	+	+	+	+
ACE/CD143	N	+	+	+
VE-Cadherin	N	+	+	+
CD31/PECAM-1	+	+	+	+
Tie-2	+	+	+	+
KDR	+	+	+	+
c-kit	+	+	+	+
Transcription Factors
RUNX-1	+	+	+	+
FLI-1	+	+	+	+
GATA2	+	+	+	+
SCL/TAL1	+	+	+	+
TEL/ETV6	+	+	+	N
EVI-1	+	+	N	N
HOXA9	N	+	+	N
MEIS1	+	+	+	N
PBX1	+	+	+	N
Other markers
vWF	+	−	+	N
P-Selectin	+	−	+	N
Signaling
Prostaglandin	+	+	+	N
Notch	+	+/−	+	N
Residence
BM vascular sinusoid	+	+	+	N

Possible Origins of HSC and Megakaryocytes Commonalities

Why do HSCs and megakaryocytes share such a large set of cell-specific factors? Possible explanations include: (1) close hierarchical developmental relationship; (2) common functional requirements; and/or (3) shared microenvironmental interactions. In the following section, we consider these possibilities in greater detail.

Close Hierarchal Relationship

The ability to purify subpopulations of cells based on cell surface marker expression, and then test their lineage potential in vitro and in vivo has provided powerful means to understand the hierarchal relationships between different hematopoietic lineages. The earliest application of this technology led to a model in which long-term repopulating HSCs give rise to short-term repopulating HSCs, and followed by multipotential progenitor cells (MPPs) (Fig. 2). These MPPs make the first lineage commitment choice by differentiating into either common lymphoid (CLPs) or common myeloid progenitors (CMPs). CMPs then give rise to bipotent megakaryocyte-erythroid progenitors (MEPs), and granulocyte-macrophage progenitors (GMPs). CLPs differentiate into T, B and NK lineages. A relatively recent provocative study suggested that MEPs may also arise directly from ST-HSCs (Lin⁻Sca-1⁺c-kit⁺CD34⁺Flt3⁻ cells) in mice [Adolfsson et al., 2005]. However, these new findings have been challenged in a subsequent study that took into account differences in the kinetics of distinct progenitor populations and used more sensitive methods [Forsberg et al., 2006]. This area remains controversial, but raises the possibility that megakaryocytes and erythroid cells are closely related to HSCs developmentally.

The Endothelial Cell Link

The developmental origins of hematopoietic and endothelial lineages are closely linked. Some of the earliest evidence for this comes from studies of in vitro differentiated mouse embryonic stem (ES) cells, which showed that a blast-colony forming cell (BL-CFC) could be isolated that gives rise to colonies with both endothelial and hematopoietic components [Choi et al., 1998]. The common precursor cell was termed a “hemangioblast”. Subsequent studies have provided in vivo evidence for the presence of hemangioblasts.

During embryogenesis, HSCs first arise from the ventral wall of the dorsal aorta during a brief developmental window. They then seed the fetal liver, which is the main hematopoietic organ during embryonic development. Emergence of HSCs coincides with the appearance of intra-aortic clusters of hematopoietic cells that are closely associated with specialized endothelial cells lining the ventral aspect of the dorsal aorta. Two models have been proposed for the origins of these cells (Fig. 3A). In the first model, bipotent mesodermal precursors cells in the subaortic mesenchyme migrate toward the lumen of the aorta and independently produce endothelial cells that line the vessel, and HSCs that transmigrate around the endothelial cells to contribute to the hematopoietic aortic cell clusters. This would be akin to the bipotential hemangioblast concept. In the second model, the mesodermal precursor cells give rise exclusively to specialized “hemogenic endothelial cells”, and these cells then give rise to HSCs (a so-called “linear” model). Three recent papers, including one involving single cell continuous imaging analysis, provide strong evidence for the latter [Chen et al., 2009; Eilken et al., 2009; Lancrin et al., 2009].

Fig 3.

Relationship between HSCs, endothelial cells, and megakaryocytes. (A) Models of HSC emergence from the aorto-gonadal-mesonephros region during embryogenesis. Left panel, “Bipotential” model. In this model, bipotential mesodermal precursor cells give rise independently to endothelial cells lining the ventral aspect of the dorsal aorta and cells that migrate around the endothelial cells to form clusters of cells containing HSCs in the lumen of the aorta. Right panel, “Linear” model. In this model, mesodermal precursor cells give rise to specialized “hemogenic” endothelial cells lining the anterior aspect of the dorsal aorta. These cells then give rise directly to the HSC containing clusters in the dorsal aorta lumen. (B) Co-localization of megakaryocytes and HSCs at bone marrow vascular sinusoids. Diagram depicting vascular sinusoids in adult bone marrow. Mature megakaryocytes are shown in apposition to the bone marrow side of a single cell layer of endothelial cells lining the sinusoid. During megakaryocyte maturation, long cytoplasmic proplatelet extensions form and protrude into the vascular space through fenestrations in the endothelial cell layer. Fragments then are shed into the circulation where they are further processed into platelets. Hypothetical interactions between Mks and HSCs at the vascular sinusoidal niche are indicated by the green bi-directional arrow. Hypothetical low-level de novo generation of HSCs from vascular sinusoidal endothelial cells is indicated by the dashed arrow.

The Weissman laboratory recently developed a novel experimental model to study to generation of the hematopoietic bone marrow niche [Chan et al., 2009]. This involves injection of fetal bone-derived cells under the renal capsule, followed by examination for establishment of a bone marrow cavity capable of supporting HSCs. Cell fractionation studies show that the cell population responsible for establishing the hematopoietic microenvironment expresses the endothelial marker CD105 (Endoglin), and that chondrocyte neovascularization is required for establishing the niche. In combination with the previously noted findings, it seems clear that hematopoietic origins have a close ancestral relationship with endothelial cells, at least during embryonic development.

There is considerable functional overlap and interplay between megakaryocytes and endothelial cells. The ultimate function of platelets is to repair disrupted endothelium and “plug” up minute holes. This occurs via adhesion to exposed subendothelium structures, activation, aggregation, cell flattening, and activation of angiogenesis. Both platelets and endothelial cells utilize prostaglandin signaling pathways, and modulate hemostasis and thrombosis. Both megakaryocytes and endothelial cells synthesize and secrete von Willebrand factor (vWF), which is involved in linking platelets to exposed basement membrane, and P-selectin, which acts as a key adhesion molecule during hemostasis. Activated platelets also secrete a large number of vasoactive and angiogenic modulatory factors.

In further support of this idea, many of the commonly expressed “lineage-specific” factors of megakaryocytes and HSCs are also present in endothelial cells and/or “hemangioblasts”, many of which have been shown to be functionally important (Table I). These include the surface receptors c-mpl, CXCR4, CD34, CD105, CD31 (PECAM), JAM-A, Tie-2, KDR, and c-kit; and the transcription factors Runx-1, Fli-1, GATA-2, SCL/TAL1, TEL/ETV6, HOXA9, Meis1, and Pbx. In fact, TPO alone supports BL-CFC formation and gives rise to secondary hematopoietic colonies and endothelial cells [Perlingeiro et al., 2003]. Thus, there appears to be close functional and compositional overlap between megakaryocytes, HSCs and endothelial cells.

Common Bone Marrow Localization

HSCs are thought to reside in specialized niches within the bone marrow. Classically, the HSC niche has been defined as a site in close proximity to the endosteum (the “osteoblastic” niche). However, recent work suggests that HSCs also reside in proximity to vascular sinusoids within the bone marrow (the “vascular niche”) [Kiel et al., 2005]. In this study, about 60% of CD150⁺CD48⁻CD41⁻Lin⁻ HSCs were localized in bone marrow sinusoidal endothelium, and 14% were associated with endosteum [Kiel et al., 2005]. In addition, HSCs maintenance is supported by bone marrow endothelial sinusoids, whereas N-cadherin -mediated homophilic adhesion with osteoblasts is dispensable [Kiel et al., 2009]. This area remains somewhat controversial, and it is possible that there is overlap in osteoblastic and vascular niches as some of the vascular sinusoids may be relatively close to the endosteum. Interestingly, mature Mks also reside predominantly at bone marrow sinusoids, juxtaposed to vascular endothelial cells (Fig. 3B). This has most recently been documented in the context of living mice using in vivo imaging techniques [Junt et al., 2007]. This study also demonstrated the extension of long proplatelet processes from the megakaryocytes through fenestrations in the endothelial cell layer and into the vascular space, where proplatelet fragments (and presumably platelets) were shed into the circulation. The homing of Mks to the vascular sinusoid is mediated in large part by the chemoattractive effects of SDF-1 on the CXCR4 receptor, a system shared with HSCs [Avecilla et al., 2004]. Thus, the functional and gene expression similarities of megakaryocytes and HSCs may reflect a common requirement for homing to and interaction with the specialized vascular sinusoidal niche.

Conclusions and Future Prospects

From the work described above, it is clear that megakaryocytes and HSCs share a remarkable number of specialized factors, and this overlap extends to endothelial cells and hemangioblasts. In addition, close physical association of these cell types occurs in the adult bone marrow. It is not known at this time whether vascular sinusoidal megakaryocytes make direct cell-cell contacts with HSCs. However, it is possible that megakaryocytes themselves may contribute to the vascular HSC niche. If so, they could perhaps be utilized for ex vivo HSC culture or expansion.

The recent demonstration of HSC derivation directly from “hemogenic” endothelium during embryogenesis raises the provocative idea that low-level de novo generation of HSCs might occur from vascular sinusoidal endothelial cells at bone marrow sinusoids in adult mice. Monvoisin et al. generated mice expressing tamoxifen-inducible Cre under the control of endothelial-specific vascular endothelial cadherin (VE-Cadherin) gene promoter [Monvoisin et al., 2006]. When these mice were interbred with reporter mice containing the β-galactosidase gene knocked into the ubiquitously expressed ROSA26 locus and preceded by a loxP-flanked stopper cassette, a small but detectable number of hematopoietic bone marrow cells (0.3 −/+ 0.1 %) became labeled after Cre was activated by tamoxifen injection in adult mice. Although there are a number of alternative explanations for this finding, one intriguing possibility is that HSCs may be capable of being generated de novo from vascular sinusoidal endothelial cells at very low levels in adult animals. If present, it is possible that such a system could be more robust following injury and regeneration.

Cellular reprogramming from somatic cells has become a reality with the remarkable demonstration that a small set of key transcription factors can generate pluripotent stem cells when over expressed in adult murine and human fibroblasts. Given the significant overlap in key transcription factors expressed in Mks and HSCs, cultured Mks may represent an efficient source of somatic cells for reprogramming into HSCs.

The discovery of shared key factors between Mks and HSCs suggests that any new findings in Mk biology may have parallels in HSC biology, and vice versa. Thus, one field may help to inform the other. On the other hand, Mks and HSCs do have unique properties. Therefore, comparison of their differentially expressed factors may represent a somewhat simplified system to elucidate the molecular basis of their specialized activities, such as stem cell self-renewal/pluripotency and megakaryocyte endomitosis. While all of the ideas discussed above are highly speculative, they are testable and may have implications for regenerative medicine, stem cell transplantation, and some bone marrow failure disorders.