|Home | About | Journals | Submit | Contact Us | Français|
Recent studies have begun to elucidate how the endothelial lineage is specified from the nascent mesoderm [1, 2]; however, the molecular mechanisms which regulate this process remain largely unknown. We hypothesized that Notch signaling might play an important role in specifying endothelial progenitors from the mesoderm, since this pathway acts as a bipotential cell fate switch on equipotent progenitor populations in other settings [3, 4]. We found that zebrafish embryos with decreased levels of Notch signaling exhibited a significant increase in the number of endothelial cells, while embryos with increased levels of Notch signaling displayed a reduced number of endothelial cells. Interestingly, there was a concomitant gain of endothelial cells and loss of erythrocytes in embryos with decreased Notch activity, without a significant effect on cell proliferation or apoptosis. Lineage tracing analyses indicate that the ectopic endothelial cells in embryos with decreased Notch activity originate from mesodermal cells that normally produce erythrocytes. Taken together, these data suggest that Notch signaling negatively regulates the number of endothelial cells by limiting the number of endothelial progenitors within the mesoderm, possibly functioning as a cell fate switch between the endothelial and the hematopoietic lineages.
In order to determine the function of Notch signaling during the specification of the endothelial lineage, we first evaluated the effects of decreased Notch activity on endothelial cells at early stages of zebrafish development. Tg(kdrl:EGFP)s843  embryos were treated with 100μM DAPT (in 5% DMSO), a known chemical antagonist of Notch signaling , or 5% DMSO (vehicle alone) from 6 to 18 hours post-fertilization (hpf), and the number of endothelial nuclei within a single transverse section of each embryo was quantified. At the 10th somite level, an average of 6.66 (s=2.42) endothelial nuclei was present in embryos treated with DAPT (n=44), compared to an average of 4.03 (s=1.27) endothelial nuclei in control embryos treated with DMSO (n=30), representing a 65% increase (ANOVA, p<10−6) (Fig. 1). Analysis at the 5th somite level yielded similar results (ANOVA, p<10−5; Suppl. Fig. 1). The observed increase in the number of GFP-positive endothelial cells in DAPT-treated embryos does not appear to be a transgene-dependent phenomenon since similar increases were observed in other transgenic lines, such as Tg(fli1a:nGFP)y7  and Tg(kdrl:Ras-mCherry)s896 , that have labeled endothelial cells (data not shown). In addition, a significant increase in the number of endothelial cells was also observed in homozygous mindbomb (mibta56b) mutant embryos, which lack an E3 ubiquitin ligase required for the activation of Notch by Delta  (Suppl. Fig. 2).
We utilized fluorescence activated cell sorting (FACS) to better quantify the effect of DAPT treatment on the entire endothelial population. To minimize interference from the GFP expression in the pharyngeal endoderm of the Tg(kdrl:EGFP)s843 embryos , casanovas4 (cass4) mutant embryos  were used. Previous studies have shown that the cass4 mutant embryos completely lack endoderm, with little effect on hematovascular development . We found that DAPT-treated 18 hpf embryos contained 55% more GFP-positive endothelial cells compared to DMSO-treated embryos (Suppl. Fig. 3). Therefore, Notch signaling appears to negatively regulate the number of endothelial cells throughout the entire embryo. Consistent with our findings, all known notch genes and two delta genes are expressed in the zebrafish gastrula (Suppl. Fig. 4).
Previous studies have demonstrated that Notch signaling is required at multiple time points to regulate various aspects of vascular development [11–15]. To define the time frame in which Notch signaling modulates the number of endothelial cells, embryos were treated with DAPT during distinct stages of development. Embryos treated from 6 to 10 hpf and from 14 to 18 hpf did not exhibit obvious changes in the number of endothelial cells. In contrast, embryos incubated with DAPT from 10 to 14 hpf displayed a similar increase in the number of endothelial cells as those continuously treated from 6 to 18 hpf (Suppl. Fig. 5 and data not shown). Interestingly, decreased Notch activity at this stage did not inhibit the differentiation of arterial endothelial cells (as determined by EphrinB2a expression), suggesting that the function of Notch signaling in limiting the number of the endothelial cells can be separated from the previously reported function in promoting arterial endothelial cell fate.
In order to test whether increased Notch activity might cause the opposite effect on the number of endothelial cells, Tg(hsp70l:Gal4)kca4;Tg(UAS:myc-notch:intra)kca3  embryos were heat-shocked at shield stage (6 hpf) to induce the expression of the intracellular domain of Notch (NICD). Embryos expressing the NICD (detected by presence of the myc epitope) exhibited a significant decrease in the number of endothelial cells compared to non-heat-shocked control siblings (Fig. 1C). An average of 4.42 (s=1.09) endothelial nuclei were present in the control embryos (n=31), while an average of 3.14 (s=0.80) endothelial nuclei were found in the embryos expressing NICD (n=28), representing a 29% loss in endothelial cell number (ANOVA, p<10−5; Fig. 1D). Therefore, increased Notch activity leads to a reduction in the number of endothelial cells. Taken together, our results show that Notch signaling negatively regulates the number of the endothelial cells during early zebrafish development.
To determine whether the observed gain of endothelial cells in embryos with decreased Notch activity was caused by changes in cell proliferation or cell death, we measured these events in DAPT- and DMSO-treated embryos at 12, 14, and 16 hpf. Neither BrdU incorporation nor staining with the mitotic marker phospho-HistoneH3 showed any discernable differences in cell proliferation within the endothelial cells between DAPT- and DMSO-treated embryos (Suppl. Fig. 6). In addition, Acridine Orange and anti-Caspase3/8 staining did not reveal any differences in cell death between the two groups (data not shown), suggesting that these processes are not likely to be the major cause for the increased number of endothelial cells in embryos with decreased Notch activity.
We then speculated that the excess endothelial cells might have resulted from the expansion of the endothelial progenitor population within the mesoderm, potentially at the expense of other related lineages. To test this hypothesis, we investigated the effect of modulated Notch signaling on the hematopoietic lineage, which develops in close spatiotemporal proximity to the endothelial lineage during embryogenesis [2, 17]. In DAPT-treated embryos, the expression level of hematopoietic markers, including gata1 , ikaros , and draculin , drastically decreased in comparison to DMSO-treated embryos (Fig. 2A and data not shown). The reduced expression of these in situ markers appears to be specific to the hematopoietic lineage, since the non-hematopoietic markers krox20  and otx1  (Fig. 2A) showed equivalent levels of expression.
To confirm that the reduced expression of hematopoietic lineage markers detected by in situ hybridization was the result of a decreased number of hematopoietic cells, we used Tg(gata1:EGFP)la781  embryos to count the number of GFP-positive erythrocytes in DAPT-and DMSO-treated embryos. Since DAPT-treated embryos exhibited a drastic increase in the number of endothelial cells in their posterior half where the majority of hematopoietic cells are erythrocytes, counting the number of erythrocytes serves as a simple surrogate measure to identify any discernable changes in the number of hematopoietic cells. Compared to an average of 8.07 (s=2.38) erythrocyte nuclei per section in DMSO-treated embryos (n=27), embryos treated with DAPT (n=19) had an average of 4.89 (s=2.15) erythrocyte nuclei (Fig. 2B), approximately a 39% decrease (ANOVA, p<10−4; Fig. 2C). Conversely, embryos over-expressing the NICD exhibited a significant increase in the number of erythrocytes compared to control siblings (ANOVA, p<10−6; Fig. 2D–E).
Unlike the hematopoietic lineage, no discernable changes in the number of somitic (detected with S58 and F59 antibodies ) or pronephric duct (Tg(pax2a:GFP)e1 ) cells were detected in embryos treated with DAPT, suggesting that the contribution of progenitors giving rise to other non-axial mesodermal lineages was negligible (Suppl. Fig. 7 and data not shown).
To further characterize the effects of perturbed Notch signaling on the endothelial and hematopoietic lineages, we treated Tg(kdrl:Ras-mCherry)s896;Tg(gata1: EGFP)la781 embryos with DAPT. Interestingly, the embryos exhibited a concomitant loss of erythrocytes and gain of endothelial cells (Fig. 3A), suggesting that Notch signaling might regulate the number of the endothelial and hematopoietic cells during the same developmental stages. To investigate the possibility of a transformation of hematopoietic progenitors into endothelial progenitors at early stages in embryos with compromised Notch signaling, homozygous bloodless (bds) mutant embryos were treated with DAPT. The bds mutation obliterates all primitive hematopoietic cells via lineage-specific apoptosis, without impacting the endothelial lineage . Thus, if hematopoietic progenitors within the mesoderm are transformed to endothelial progenitors as a result of decreased Notch activity, there should be no discernable change in the number of endothelial cells in the bds mutant embryos.
In DAPT-treated bds mutant embryos (n=20), an average of 6.05 (s=1.32) endothelial nuclei were present per section, while DMSO-treated bds mutant embryos (n=22) and DAPT-treated phenotypic wild-type siblings (n=24) contained an average of 5.95 (s=1.13) and 8.04 (s=2.36) endothelial nuclei, respectively (Fig. 3B). Thus, we found no significant difference in the number of endothelial cells in bds mutant embryos treated with DAPT (ANOVA, p<0.80; Fig. 3C). These data suggest that decreased Notch activity causes a transformation of hematopoietic progenitors to endothelial progenitors, and that Notch signaling might function as a cell fate switch between the two lineages.
To further test the function of Notch signaling as a cell fate switch between the endothelial and hematopoietic lineages, lineage tracing experiments were performed . Single-cell stage embryos were injected with DMNB-caged fluorescein conjugated dextran. At shield stage, approximately ten cells within the ventral margin of the zebrafish gastrula, which predominantly generate the primitive hematopoietic lineage, were uncaged with UV irradiation. Subsequently, embryos were treated with DAPT or DMSO (Fig. 4A). If there is a direct transformation of hematopoietic progenitors to endothelial progenitors as a result of decreased Notch activity, the increased contribution to the endothelial lineage from the ventral margin should be detected. As the cells from the ventral marginal zone rarely contribute to the endothelial lineage , an average of 0.16 (s=0.37) endothelial cells per focal plane contained uncaged fluorescein in DMSO-treated embryos (n=18) (Fig. 4B and C). In contrast, in DAPT-treated embryos (n=13), an average of 1.61 (s=1.01) endothelial cells per focal plane showed co-localization with uncaged fluorescein (Fig. 4B and C). These data support a model in which hematopoietic progenitors within the ventral marginal zone of the zebrafish gastrula are transformed to endothelial progenitors upon the inhibition of Notch signaling, which eventually leads to an increased number of endothelial cells (Fig. 4D).
Considering the existence of common progenitors for the endothelial and hematopoietic lineages in zebrafish , it is tempting to speculate that Notch signaling might function as a bipotential cell fate switch for the two lineages by regulating the hemangioblasts, resulting in descendents preferentially adopting an endothelial fate. However, the observed magnitude of the increase in endothelial cells after inhibition of Notch signaling suggests that while such an event is possible, it probably does not account for the full effect.
Our experiments provide evidence that Notch signaling negatively regulates the number of endothelial cells by modulating the specification of endothelial progenitors. This novel function of Notch signaling can be separated from a previously known function in promoting the differentiation of arterial endothelial cells at later stages of development. In addition, our data indicate that the effect of Notch signaling on endothelial cell number is the result of the direct transformation of hematopoietic progenitors into endothelial progenitors. Taken together, our results suggest that Notch signaling functions as a cell fate switch between the endothelial and the hematopoietic lineages within the nascent mesoderm.
Zebrafish (Danio rerio) embryos were obtained from the UNC-CH Zebrafish Aquaculture Core Facility and raised as previously described . The following mutant and transgenic lines were used: bdsa75 , cass4 , mibta56b , Tg(gata1:EGFP)la781 , Tg(hsp70l:Gal4)kca4 , Tg(fli1a:nGFP)y7 , Tg(kdrl:EGFP)s843 , Tg(kdrl:Ras-mCherry)s896 , Tg(pax2a:GFP)e1 , and Tg(UAS:myc-notch:intra)kca3 . Homozygous mibta56a embryos were identified by PCR, and homozygous bdsa75 embryos were identified based on the lack of ikaros expression by RT-PCR.
Embryos were treated with 100μM DAPT (CalBiochem) as previously described. Embryos were treated from 6 to 18 hours post-fertilization (hpf), unless noted otherwise, and subsequently fixed in 4% paraformaldehyde (PFA) overnight prior to processing. To induce ectopic Notch signaling, embryos from crosses between Tg(hsp70l:Gal4)kca4/+ and Tg(UAS:myc-notch:intra)kca3/+;Tg(kdrl:EGFP)s843 individuals were raised at 28°C until shield stage (6 hpf), and subjected to treatment at 40°C for 30 minutes. Bright-field microscopy and myoD  in situ hybridization were used to verify that the treated embryos exhibited no developmentally delays compared to the control embryos (Suppl. Fig. 8).
Immunohistochemistry and whole mount in situ hybridization were performed as previously described . The antibodies and riboprobes used can be found in Supplemental Data. Polymerized actin was visualized by phalloidin (Molecular Probes) at 1:300 and nuclei were visualized with TOPRO (Molecular Probes) at 1:10,000. For immunohistochemistry, processed samples were mounted in Vectashield (Vector Laboratories) and the images were acquired using a Zeiss 510 Meta confocal microscope (Michael Hooker Microscopy Facility, UNC-CH) or a Zeiss LSM5 Pascal confocal microscope. For whole mount in situ hybridization, embryos were mounted in benzylbenzoate:benzyl alcohol and documented with Zeiss Axiocam.
For BrdU incorporation, homozygous Tg(kdrl:EGFP)s843 embryos were treated with DAPT or DMSO from 10 to 18 hpf and also incubated in 10mM BrdU in 5% DMSO from 14 hpf to 16 hpf. At 18 hpf, embryos were fixed in 4% PFA overnight and then incubated in rat monoclonal anti-BrdU followed by anti-rat Alexa Fluor 568.
Lineage tracing was performed as previously described . Briefly, Tg(kdrl:EGFP)s843 embryos were injected with 4.6nl of 0.2% DNMB-caged, biotinylated, lysine fixable fluorescein dextran (Molecular probes) in 0.2M KCl and allowed to develop until shield stage (6 hpf). The cells located in the first three vegetal tiers of the ventral margin were activated using pulses from a 360 nm laser focused through a 20× objective. Embryos were then treated with DAPT or DMSO and fixed at 18 hpf. To detect activated fluorescein and distinguish it from GFP, embryos were incubated in biotinylated rabbit IgG anti-FITC antibody followed by Alexa Fluor 568 Streptavidin (Molecular Probes).
The data were analyzed with one-way ANOVA for the effect of different treatments on the number of cells of interest. A p-value smaller than 0.01 was regarded as significant.
The authors would like to thank the members of the Jin lab for helpful discussions and Drs. Cheol-Hee Kim and Hae-Chul Park for in situ probes. We also thank Drs. Victoria Bautch, Frank Conlon, Wiebke Herzog, Mark Majesky, Cam Patterson, and John Rawls for discussion and critical reading of the manuscript. This study was supported by an NSF Pre-doctoral Fellowship to K. M. V, Agency for Science, Technology and Research, Singapore (A*STAR) to Y.-J. J., the NHLBI (HL54737) and Packard foundation to D.Y.R.S, and a Basil O’Connor Scholar Award from the March of Dimes Foundation, the NHLBI (HL090960), and the University of North Carolina at Chapel Hill Research Fund to S.-W. J.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.