A Role for Thrombopoietin in Hemangioblast Development



Vascular endothelial growth factor (VEGF) and stem cell factor (SCF) act as growth factors for the hemangioblast, an embryonic progenitor of the hematopoietic and endothelial lineages. Because thrombopoietin (TPO) and its receptor, c-Mpl, regulate primitive hematopoietic populations, including bone marrow hematopoietic stem cells, we investigated whether TPO acts on the hemangioblasts that derive from differentiation of embryonic stem cells in vitro. Reverse transcriptase polymerase chain reaction analysis detected expression of c-Mpl beginning on day 3 of embryoid body differentiation when the hemangioblast first arises. In assays of the hemangioblast colony-forming cell (BL-CFC), TPO alone supported BL-CFC formation and nearly doubled the number of BL-CFC when added together with VEGF and SCF. When replated under the appropriate conditions, TPO-stimulated BL-CFC gave rise to secondary hematopoietic colonies, as well as endothelial cells, confirming their nature as hemangioblasts. Addition of a neutralizing anti-VEGF antibody did not block TPO enhancement of BL-CFC formation, suggesting that TPO acts independently of VEGF. These results establish that Mpl signaling plays a role in the earliest stages of hematopoietic development and that TPO represents a third growth factor influencing hemangioblast formation.


Thrombopoietin (TPO) is the primary physiological regulator of platelet production. Although it was initially thought to be a lineage-dominant factor primarily affecting megakaryopoiesis [13], numerous in vitro studies have established that TPO plays an important role in the maintenance and expansion of hematopoietic stem cells (HSCs) [410]. Analysis of bone marrow (BM) and blood of mice lacking either TPO [11, 12] or its receptor, c-Mpl [11, 13, 14], revealed deficiencies in the progenitor pool and stem cell compartment in addition to profound thrombocytopenia. More recently, Kimura et al. demonstrated that c-mpl−/− mice were severely deficient in spleen colony-forming cells (CFC-S) as well as long-term repopulating HSCs, and that engraftment ability was extinguished on serial transfer into secondary and tertiary hosts, suggesting that the absence of c-Mpl signaling might affect the self-renewal properties of these cells [15].

In normal mice, c-Mpl is expressed in megakaryocytes, platelets, and primitive hematopoietic cells of fetal liver and BM [16, 17]. c-mpl expression can be found in a highly purified population of dormant HSCs [17], and it has been demonstrated that 70% of highly enriched populations of murine LinloSca+c-kit+ and human CD34+CD38 express c-mpl [18]. The same authors, using competitive repopulation assays, showed that the c-Mpl+ fraction contained all of the repopulating capacity in the murine system and almost all of the engraftment potential in a mouse-human xenograft model [18]. All these findings support the concept that c-Mpl plays an important role in the early stages of hematopoiesis. Interestingly, low levels of expression have also been detected in endothelial cells derived from the umbilical cord [16, 19] and more abundantly in liver endothelial cells [20].

The close association between endothelial cells and primitive hematopoietic precursors in the embryonic yolk sac suggests that the endothelium plays a critical role in early hematopoietic development, providing the microenvironment required for stem cell proliferation and differentiation. Endothelial and blood precursors also express several common critical regulatory genes and antigenic markers [2125]. These observations have led to the hypothesis that a common progenitor, the hemangioblast, gives rise to both endothelial and hematopoietic cells [26, 27]. Recently, a cell with properties of the hemangioblast has been identified during in vitro differentiation of embryonic stem (ES) cells into embryoid bodies (EBs) [28]. This precursor, referred to as the blast colony-forming cell (BL-CFC) forms in response to vascular endothelial growth factor (VEGF) and stem cell factor (SCF), and represents a transient population of cells that stand at the juncture of the endothelial and hematopoietic lineages [29].

Based on the role of TPO as a growth factor for early hematopoiesis, we hypothesized that TPO might have a similar effect on the hemangioblast. In this study, we demonstrated that TPO and its receptor are expressed at the yolk sac stage of hematopoiesis in the embryo, and that TPO stimulates hemangioblast formation during differentiation of ES cells in vitro. These data suggest a role for TPO at the earliest stages of hematopoietic development.

Materials and Methods

Growth and Differentiation of ES Cells

The 129/Sv-derived ES cell line CCE [30] was maintained on gelatinized flasks in Dulbecco's modified Eagle's medium (Sigma, St. Louis, MO; http://www.sigmaaldrich.com) supplemented with 1,000 U/ml leukemia inhibitory factor (Amgen, Thousand Oaks, CA; http://www.amgen.com), 15% knock-out serum replacement (GIBCO; Carlsbad, CA; http://www.lifetech.com), 0.1 mM nonessential amino acids (Sigma), and 0.1 mM of β-mercaptoethanol (Sigma) in the absence of feeders for up to 15 passages. For differentiation cultures, the cells were dissociated with trypsin (0.25%; GIBCO)/EDTA (1 mM; Sigma) to form a single-cell suspension. Cells were washed three times with phosphate-buffered saline (PBS) and EBs were generated by plating 104 cells/ml in Iscove's modified Dulbecco's medium (IMDM; Sigma) with 15% fetal calf serum (FCS; Sigma), 50 μg/ml ascorbic acid (Sigma), 200 μg/ml iron-saturated transferrin (Sigma), 4.5 × 10−4 M monothioglycerol (MTG; Sigma), and 0.9% methylcellulose (M3120, StemCell Technologies; Vancouver, B.C.; http://www.stemcell.com) in 35-mm petri dishes (StemCell Technologies).

Infection of ES cells with MSCVF36VmplGFP

We used a fusion protein composed of the c-Mpl receptor-signaling domain linked to a modified FK506-binding protein dimerization domain (F36V). In this system, activation of c-Mpl occurs through dimerization, which is controlled by adding a chemical inducer of dimerization (CID) [31, 32]. MSCVF36VmplGFP retrovirus (generously provided by Anthony Blau, University of Washington, Seattle, WA) was transfected into 293T cells using the Ca2+ phosphate precipitation method [33]. Virus containing supernatant was collected 48 hours after transfection and filtered through a 0.45-mm filter before use. The cell line CCE was spin-infected with high-titer viral supernatant in ES cell medium at 1,300 g for 90 minutes on a Beckman centrifuge (GH-3.8 rotor; Beckman Coulter; Fullerton, CA; http://www.beckman.com). Supernatant was removed the next day, and cells were passaged for 1 month, with the brightest 10% of GFP+ cells selected for subculture by fluorescence-activated cell sorting (FACS) each week. After 1 month, the brightest 0.1% of cells were FACS deposited into single wells of a 96-well dish preplated with mitotically inactivated mouse embryonic fibroblasts. Wells with single ES cell colonies were harvested and expanded into clonal cell lines. These clones were tested individually for activity of the FKBP fusion protein by generating EBs, disrupting and plating in the presence and absence of CID (AP20187, generously provided by ARIAD; Cambridge, MA; http://www.ariad.com/regulationkits). The clone CCE.j gave the best induction and was used for the blast colony-forming cell (BL-CFC) experiments.

BL-CFC Assay

EBs were collected at 4 days post differentiation, washed in PBS, and treated with 0.25% trypsin (GIBCO) for 3 minutes at 37°C. EBs were disrupted to single cells by repeated passage through a 23-G needle and plated at 5 × 104 cells in 1 ml of methylcellulose medium (M3120) with 10% FCS, 50 μg/ml ascorbic acid, 200 μg/ml iron-saturated transferrin, and 4.5 × 10−4 M MTG, in the presence and absence of TPO (25 ng/ml; Peprotech; Rocky Hill, NJ; http://www.peprotech.com), and in the presence or absence of the following factor conditions: VEGF, 5 ng/ml (Peprotech); SCF, 100 ng/ml (Peprotech); and VEGF/SCF. In some experiments, an anti-VEGF blocking antibody was added to the methylcellulose at an adequate concentration to neutralize all of the effects of VEGF supplementation (0.3 μg/ml; R&D Systems; Minneapolis, MN; http://www.rndsystems.com). Cultures were maintained in a humidified incubator at 37°C in an environment of 5% CO2 in air. After 5 days, developing BL-CFC were counted and picked for replating studies.

Generation of Hematopoietic Cells

BL-CFC were plucked and replated into secondary methylcellulose media containing interleukin-3 (IL-3), IL-6, erythropoietin (Epo), and SCF (M3434; StemCell). Cultures were maintained as described above and secondary hematopoietic colonies were scored at 7–10 days of growth.

Generation of Endothelial Cells and Their Characterization

Individual BL-CFC were picked and transferred to matrigel-coated microtiter wells (Collaborative Research; Bedford, MA) containing IMDM with 10% FCS, 10% horse serum (GIBCO), VEGF (5 ng/ml), insulin growth factor-1 (IGF-1, 10 ng/ml; Peprotech), Epo (2 U/ml; R&D), basic fibroblast growth factor (bFGF, 10 ng/ml; Peprotech), IL-11 (50 ng/ml; R&D), SCF (100 ng/ml), endothelial cell growth supplement (ECGS, 100 μg/ml; Collaborative Research), L-glutamine (2 mM), and 4.5 × 10−4 MTG [29]. After 3–4 days in culture, nonadherent cells were removed and adherent cells were cultured for an additional 1–2 weeks in IMDM with 10% FCS, 10% horse serum, VEGF (5 ng/ml), IGF-1 (10 ng/ml), bFGF (10 ng/ml), ECGS (100 μg/ml), L-glutamine (2 mM), and 4.5 × 10−4 MTG. When confluent, cells were harvested by trypsinization and analyzed for flk-1 and tie-2 expression.

Cell Staining

BL-CFCs were plucked onto prewashed glass cover slips that were coated with a thin layer of matrigel and cultured in 12-well dishes in medium containing both hematopoietic and endothelial cytokines, as described above. Four to 7 days following the initiation of the cultures, the nonadherent hematopoietic cells were removed and the adherent cells were cultured for an additional 1–2 weeks in medium containing only endothelial growth factors (see above). For fluorescence analysis, adherent cells were initially cultured in the presence of 10 μg/ml of Dil-Ac-LDL (Biomedical Technologies; Stoughton, MA; http://www.btiinc.com) at 37°C for 2 hours. Following this incubation, the cells were washed three times and fixed for 10 minutes in PBS containing 3% paraformaldehyde and 3% sucrose. The fixed cells were washed two to three times and incubated with fluorescein isothiocyanate-mouse anti-CD31 (Pharmingen; San Diego, CA; http://www.bdbiosciences.com/pharmingen) for 1 hour. Following this staining, cells were washed again (five times) and the coverslip with the cells was mounted onto a slide for analysis [29]. The images were acquired with a digital confocal microscope using the Zeiss LSM-510 software program (Oberkochen, Germany).

Mouse Dissection

Mature 129SvEv females were caged with breeding males. The day of vaginal plug observation was considered as day 0.5 postcoitum. Pregnant dams were sacrificed by CO2 asphyxiation and embryos isolated at E8.25 for yolk sac (YS) and E13.5 for fetal liver (FL) dissections. Adult BM cells were obtained by flushing femurs from adult mice.

Reverse Transcriptase Polymerase Chain Reaction (RT-PCR) Analysis

Gene expression patterns in EBs, YS, FL, and BM cells, as well as BL-CFC and their derived cells, were determined using the global amplification strategy of Brady et al. [34]. Total RNA was isolated using RNA STAT-60 reagent (Tel-Test “B,” Inc.; Friendswood, TX) as recommended by the manufacturer. First strand cDNA was produced using Superscript II reverse transcriptase (GIBCO). One microgram of total RNA was hybridized with 140 ng random hexamers in first strand reverse transcriptase buffer (50 mM Tris, pH 8.3, 75 mM KCl, 3 mM MgCl2, 10 mM dithiothreitol, 500 μm deoxyribose nucleotide triphosphates) followed by addition of 200 U Superscript II reverse transcriptase and incubation at 42°C for 50 minutes. Five percent of first strand reaction was used for each ensuing PCR reaction. Primer sequences were as follows: A) β-actin forward, GTGGGGCGCCCCAGGCACCA; β-actin reverse, CTCCTTAATGTCACGCACGATTTC; B) c-mpl forward, CCTACTGCTGCTAAAGTGGCAAT; c-mpl reverse, CAATAGCTTAGTGGTAGGTAGGA; C) TPO forward, TCTGTCCAGCCCCGTAGGTC; TPO reverse, GTTCCATCCACAGGTCCGTG; D) flk-1 forward, CACCTGGCACTCTCCACCTTC; flk-1 reverse, GATTT CATCCCACTACCGAAAG; E) tie-2 forward, ATGGAC TCTTTAGCCGGCTTA; and tie-2 reverse, CCTTATAGC CTGTCCTCGAA.


Although c-Mpl expression has been detected in FL, BM [16, 17], and peripheral blood from embryos (E12) [35], its expression at earlier stages of embryonic development has not been explored. We investigated the expression of TPO and the c-Mpl receptor in YS cells by RT-PCR, using FL and BM as positive controls. Both TPO and c-Mpl expression was observed in precirculation YS cells (day 8.25) (Fig. 1), confirming our suspicion that c-Mpl signaling might play a role in early hematopoietic development.

Figure Figure 1..

Detection of c-Mpl and TPO mRNA by RT-PCR.mRNA was isolated from EB cells at different time points (left side), as well as day 8.25 YS, day 13.5 FL, and adult BM cells (middle). In another experiment, mRNA was isolated from individual BL-CFC that had grown for 5 days in the presence of TPO alone, VEGF/SCF/TPO, or VEGF/SCF (right side) and evaluated for c-Mpl expression. Actin mRNA was used as an internal control for fidelity of RNA amplification.

The close association between endothelial cells and primitive hematopoietic precursors during embryogenesis led to the hypothesis of the hemangioblast as a common precursor. Such a cell has not been cultured from embryos, but can be measured as a blast colony forming cell (BL-CFC) that arises as the first hematopoietic elements during in vitro differentiation of ES cells [28, 29]. We performed RT-PCR analyses for c-Mpl in ES cells that were differentiated as EBs various times. c-Mpl expression was first detected by day 3 of EB development (Fig. 1), the day BL-CFC first arises [28], and was still present by day 12 of EB differentiation. We evaluated TPO mRNA and, to our surprise, found expression in ES cells, which continued or decreased with differentiation (Fig. 1). Taken together, our results establish that TPO and its receptor, c-Mpl, are present at the earliest stages of hematopoietic development in the YS and developing EB.

ES cells had been differentiating into EBs for 4 days when we detected the maximal number of BL-CFC. The EBs were disrupted and plated in medium for detection of BL-CFC [28] with or without supplemental TPO, VEGF, and SCF (Fig. 2). When added alone, TPO supported BL-CFC formation (Fig. 3A). Moreover, TPO stimulated an approximately twofold increase in the number of BL-CFC in cultures supplemented by VEGF and/or SCF (Fig. 3A). In order to evaluate if BL-CFC grown under different conditions express the receptor for TPO, mRNA was obtained from single colonies, which were plucked in duplicate for each culture condition (TPO alone, VEGF/SCF/TPO, and VEGF/SCF). We detected c-Mpl expression in all samples examined (Fig. 1, right side), except for one TPO-stimulated colony probably due to insufficient mRNA (lane 1). The finding that BL-CFC grown in the absence of TPO (lanes 5 and 6) express c-Mpl suggests that the hemangioblast has an intrinsic capacity to respond to TPO.

Figure Figure 2..

Scheme for characterization of hematopoietic and endothelial cells from BL-CFC.EBs were differentiated for 4 days, disrupted, and plated in methylcellulose blast medium (BL-MCM) supplemented with no growth factors (base), VEGF/SCF, VEGF, or SCF, with or without TPO, as indicated. After 5 days in culture, BL-CFC were scored and plucked for replating studies. Single colonies were dispersed and replated for secondary hematopoietic CFCs in methylcellulose containing a cocktail of growth factors for myeloid colonies (M3434, StemCell) or plated in matrigel-coated microtiter wells with endothelial growth factors.

Figure Figure 3..

TPO/c-Mpl signaling promotes the formation of BL-CFC.Number of BL-CFC arising under the following conditions: A) EB cells were plated in methylcellulose containing VEGF/SCF, VEGF, SCF, or neither, ± TPO; B) EB cells were plated in methylcellulose containing VEGF or TPO, in the presence or absence of anti-VEGF blocking antibody; C) CCE.j EB cells (F36Vmpl infected) were plated in methylcellulose in the presence and absence of CID. *p < 0.05; **p < 0.01

It has been reported that TPO can stimulate VEGF secretion [36]. To assess if the stimulatory effect of TPO on BL-CFC formation was due to a direct effect of this growth factor or due to induction of VEGF secretion, we added an anti-VEGF blocking antibody to medium containing TPO or VEGF alone (used as a control). A saturating amount of anti-VEGF antibody sufficient to reverse VEGF-stimulated BL-CFC formation to background levels had no effect on BL-CFC stimulated by TPO (Fig. 3B). These results support the conclusion that TPO can play a direct role in hemangioblast differentiation, proliferation, or survival.

We also questioned whether conditional regulation of c-Mpl receptor signaling would enhance hemangioblast formation. For this purpose, ES cells were infected with a retrovirus carrying the chemical dimerizer-dependent domain of FKBP (F36V) fused to the cytoplasmic domain of c-Mpl [31], allowing for signals to be delivered in a regulated manner. Infected ES cells were cultured for several weeks and FACS sorted to isolate cells with the highest levels of green fluorescent protein expression. Subclones were derived and tested individually for activity of the FKBP fusion protein. Clone CCE.j gave the best induction and was used for the BL-CFC experiments. CCE.j cells grown in the presence of CID yielded an increased number of BL-CFC when compared with cells grown in the absence of CID (Fig. 3C). The number of BL-CFC generated by untreated CCE.j clonal cell line was lower than that of the parental ES cell line (Fig. 3A). Since we have observed a similar outcome with other clones (data not shown), we believe that overexpression of the construct may be somewhat deleterious to blast colony formation. Nevertheless, the increase in the number of BL-CFC in the presence of CID demonstrates that c-Mpl signaling can augment BL-CFC formation.

In order to confirm that the TPO-responsive precursors correspond to the hemangioblast, we proceeded with further hematopoietic and endothelial characterization (Fig. 4). We compared the hematopoietic replating potential of BL-CFC colonies grown under different conditions, including VEGF/SCF, VEGF/SCF/TPO, TPO alone, and base media with no growth factor added. Single colonies were picked and dispersed into methylcellulose with a cocktail of hematopoietic cytokines. We observed no hematopoietic replating potential for the infrequent BL-CFC that appeared in unsupplemented base media. However, BL-CFC grown under the three combinations of growth factors yielded similar frequencies of secondary hematopoietic colonies after replating (range, 58.8%–65.4%; Table 1). The average number of secondary CFCs per replated BL-CFC was low in BL-CFC stimulated by TPO alone, with no obvious bias in the types of colonies produced (Fig. 4A). The overall number of secondary hematopoietic colonies was enhanced by the addition of VEGF and SCF (Table 1), in agreement with a previous report that demonstrated that SCF potentiates the hematopoietic replating of VEGF-stimulated BL-CFC [28]. These data indicate that TPO-stimulated BL-CFC have hematopoietic potential that is further augmented by VEGF and SCF.

Figure Figure 4..

Characterization of hematopoietic and endothelial potential of BL-CFC.Individual BL-CFC were plucked, disrupted, and replated under the following conditions: A) Methylcellulose medium supplemented with hematopoietic growth factors (IL-3, IL-6, SCF, and Epo); B) Matrigel-covered coverslips for Dil-Ac-LDL uptake and CD31 staining of adherent cells populations: (I) Morphology of NIH-3T3 (negative control), D4T (endothelial cell line/positive control), and TPO-stimulated blast colony-derived adherent cells under phase contrast, (II) Analysis for their ability to take up Dil-Ac-LDL (red), and (III) Staining for the presence of CD31 (green); C) Eight BL-CFC replated on matrigel-covered plates were analyzed for flk-1 and tie-2 expression by RT-PCR.

Table Table 1.. Analyses of BL-CFC for hematopoietic and endothelial potential
  1. a

    p < 0.05

  2. b

    *Replating is lower than the other two conditions. ND = not determined.

Condition% BL-CFC that yield secondary CFCsAverage n secondary CFCs per BL-CFC% BL-CFC that yield adherent endothelial-like cellsAc-LDL+ cells derived from BL-CFC
TPO (n = 26)65.43.1 ± 2.1*56.33/3
VEGF, SCF (n = 14)64.312.8 ± 1266.62/2
VEGF, SCF, TPO (n = 17)58.88.7 ± 8.354.51/1
Base (n = 10)0NDNDND

Individual BL-CFC were also replated to determine their potential to form adherent endothelial cells (Fig. 2). BL-CFC grown under all tested combinations of growth factors yielded comparable frequencies of adherent endothelial-like cells (range, 54.5%–66.6%; Table 1). Endothelial cells express the platelet endothelial cell adhesion molecule (PECAM-1, CD31) [37] and take up acetylated low-density lipoprotein (Ac-LDL) [38]. Adherent cells derived from BL-CFC grown in the presence of VEGF/SCF (n = 2), VEGF/SCF/TPO (n = 1), and TPO alone (n = 3) were analyzed by immunofluorescence for the presence of CD31 and for their capacity to take up Ac-LDL. The fibroblastic cell line, NIH3T3, and endothelial cell line, D4T [39], served as negative and positive controls, respectively (Fig. 4B). All colonies gave rise to cells that were positive for both parameters (Table 1). In Fig. 4B, a representative example of adherent cells derived from a TPO-stimulated BL-CFC is shown. These cells show Dil-Ac-LDL uptake (red) as well as high levels of CD31 expression (green). That adherent cells derived from blast colonies express CD31 and take up Ac-LDL strongly suggests that they are from the endothelial lineage.

By RT-PCR, we analyzed the expression of the VEGF receptor, flk-1, as well as the endothelial marker, tie-2 [40] in adherent cells derived from TPO-stimulated BL-CFC. All clones of adherent cells expressed detectable levels of either or both flk-1 and tie-2 (Fig. 4C). Our replating efficiency of 60% under either hematopoietic or endothelial conditions argues that a minimum of 20% of TPO-stimulated colonies have bipotentiality. However, given that the maximum replating efficiency reported for this assay is 80%, as described by Choi et al. [29], we believe that the majority of TPO-stimulated colonies are bipotential. Thus, we conclude that the TPO-stimulated precursors from EBs are hemangioblasts.


Given the extensive literature implicating TPO in HSC function [5, 9, 18], as well as the presence of its receptor, c-Mpl, on primitive hematopoietic and endothelial cells [16, 17, 19, 20], we investigated whether TPO signaling might play a role in the earliest stages of hematopoietic development, when mesoderm first becomes committed to hematopoietic cell fate. The hemangioblast, a common precursor of the hematopoietic and endothelial lineages, is believed to represent the first hematopoietic precursor cell formed in the YS. This cell has never been identified from mouse embryos but has been characterized as BL-CFC in differentiating cultures of mouse ES cells, which recapitulate the developmental program of YS hematopoiesis [28]. In this study, we demonstrate that the c-Mpl receptor is expressed in the murine embryonic YS and in EBs at the time of hemangioblast formation, consistent with its role in the earliest stages of hematopoietic development. When added to differentiating cultures of ES cells, TPO augments BL-CFC formation, as does activation of a conditionally regulated chimeric c-Mpl receptor. The TPO-Mpl-stimulated BL-CFC show hematopoietic and endothelial cell potential during serial culture, verifying their identity as hemangioblasts. Our data thus establish a role for TPO in the earliest stages of embryonic hematopoietic development, and implicates TPO along with VEGF and SCF as growth factors required for optimal culture of primitive hematopoietic cells during differentiation of ES cells in vitro.

The spatial association between embryonic hematopoietic precursors and angioblasts in the YS blood islands, as well as the common expression of Flk-1 [22, 25], PECAM [41], and CD34 [24] in hematopoietic and endothelial cells, prompted the hypothesis of a common precursor, the hemangioblast. Using a model system based on the in vitro differentiation of ES cells into EBs, Keller and colleagues identified the BL-CFC, a single cell that forms in response to VEGF, and when replated, yields endothelial and hematopoietic progeny [28, 29]. This is the most striking evidence to date for the existence of the hemangioblast. Corroborating the idea that the VEGF/Flk-1 interaction is fundamental for hemangioblast development, Nishikawa and colleagues demonstrated that Flk-1+ cells purified from differentiated ES cells show both hematopoietic and endothelial potential [42].

While VEGF, acting through its receptor FLK-1, stimulates hemangioblast formation, this signaling pathway is not unique in specifying hemangioblast development. Mice deficient in Flk-1 have dramatic defects in the development of hematopoietic and endothelial lineages, but Flk-1−/− embryos retain hematopoietic activity [43], and VEGF-deficient embryos display both hematopoietic and endothelial cells [44, 45]. Flk-1−/− ES cells generate BL-CFC, though at greatly reduced numbers [46], and there is no difference in hematopoietic and endothelial marker expression of Flk-1+/+, Flk-1+/−, and Flk-1−/− EB cells [46]. These data argue that VEGF is not the sole regulator of hemangioblast development and that other factors must be involved. Previous evidence that TPO-Mpl signaling plays a role in blast colony formation has been suggested by data that Mpl may act as a substitute for the tyrosine kinase receptor, Flk-1 [47]. Enforced expression of c-Mpl in Mpl-deficient ES cells resulted in a TPO-dependent response for BL-CFC, which are normally strictly dependent on VEGF. During the preparation of this manuscript, the same authors demonstrated a synergistic effect of TPO (pegylated recombinant human megakaryocyte growth and development factor) on the VEGF-dependent blast cell colony in wild-type c-mpl+/+ ES cells [48], suggesting that c-Mpl is naturally expressed in the BL-CFC. While careful characterization of the BL-CFC identified in their system was not performed, their data corroborate the results we describe here.

In our studies, BL-CFC stimulated by VEGF/SCF or TPO (alone or in combination) presented a comparable frequency of BL-CFC that yielded secondary hematopoietic colonies. However, when TPO was incorporated into the culture medium, the number of secondary CFCs was reduced. Such a result may be due to an enhancement of self-renewal at the expense of differentiation, a conclusion that is consistent with data from long-term cultures that have revealed TPO to be a self-renewal factor involved in the amplification of long-term culture-initiating cells [9, 49]. Furthermore, Yagi et al. have demonstrated by transplantation studies in lethally irradiated mice that 2-month-old long-term BM cultures stimulated by TPO resulted in superior lymphoid/ myeloid reconstitution when compared with long-term cultures maintained without TPO [9]. These data provide evidence that TPO is an important mediator of self-replication in vitro and in vivo. In support of these findings, it has been recently demonstrated by introducing chimeric cytokine receptors into murine BM cells that c-Mpl signaling can promote self-renewal of multipotent hematopoietic progenitor cells, whereas signals provided by G-CSF and Flt-3 do not [50]. Our results are consistent with a role for TPO in maintaining the self-renewal potential of the hemangioblast at the expense of differentiation to CFCs that can be scored by methylcellulose assay.

Although several studies have demonstrated the growth-promoting effects of TPO on early hematopoietic progenitors [5, 6, 9, 10], as well as in the growth and activation [19, 20] of endothelial cells, we have characterized a novel role for TPO signaling at the level of the hemangioblast, the bipotential precursor of these two lineages. In agreement with our results in the blast assay, we detected c-Mpl expression during EB differentiation as early as day 3, the first day when BL-CFC form. Moreover, we have detected expression of the TPO receptor in BL-CFC stimulated by VEGF and SCF, as well as in precirculation YS cells (day 8.25). The BL-CFC has not been identified in the embryo, but we anticipate that such a cell should be detectable from the precirculation YS. Our data establish a role for TPO and its receptor, c-Mpl, at early stages of hematopoiesis and implicate TPO as a potential physiological regulator of the common precursor of hematopoietic and endothelial cells.

Since hematopoietic and endothelial development are not extinguished by targeted inactivation of either c-Mpl or TPO genes [12], there must be redundancy of cytokine receptor pathways acting on the hemangioblast. In this regard, in addition to VEGF/Flk-1 [43], some reports have suggested that bone morphogenetic protein-4 [51, 52] and transforming growth factor β-1 [53] also play a role at early stages of development. Thus, the hemangioblast and the HSC are regulated by overlapping sets of signaling molecules. Our results establish a model in vitro system for further studies to define the unique contributions of TPO on the earliest stages of hematopoietic commitment.


This work was supported by grants from the National Science Foundation-MIT Biotechnology Process Engineering Center, the Burroughs Wellcome Fund, the National Institutes of Health (CA76418, CA86991, and DK59279), the Alberta Heritage Foundation for Medical Research, and the Canadian Institutes for Health Research. G.Q.D. is the Birnbaum Scholar of the Leukemia and Lymphoma Society of America. We thank C. Anthony Blau (University of Washington) for the F36Vmpl construct and ARIAD Pharmaceuticals for AP20187. This work was conducted utilizing the W.M. Keck Foundation Biological Imaging Facility at the Whitehead Institute.