Definitive erythropoiesis in chicken yolk sac - Wiley Online Library

DEVELOPMENTAL DYNAMICS 237:3332–3341, 2008

PATTERNS & PHENOTYPES

Definitive Erythropoiesis in Chicken Yolk Sac Hiroki Nagai and Guojun Sheng*

The first wave of erythropoiesis in amniotic animals generates all primitive erythrocytes and takes place exclusively in yolk sac mesoderm. It is less clear, however, to what extend and for how long the yolk sac contributes to the second wave of erythropoiesis which gives rise to definitive erythrocytes for later embryonic and adult use. Here, we examine the initiation, duration, and site of definitive erythrocyte formation in chicken yolk sac. We show that the earliest definitive erythrocytes are generated in yolk sac venous vessels surrounding major arteries at embryonic day (E) 4 – 4.5, and that mature definitive erythrocytes enter circulating at E4.5–E5. This takes place at a time when yolk sac vasculature remodels extensively to generate paired arterial/venous vessels. The yolk sac remains the predominant site for definitive erythropoiesis from E5 to E10, and continues to generate definitive erythrocytes at least until E15. Similar to primitive erythropoiesis, definitive erythropoiesis in the yolk sac is accompanied by the expression of transcriptional regulators gata1, scl, and lmo2. Furthermore, our data suggest that one main source of definitive erythropoietic cells is the pre-existing vascular endothelial cells. It remains unclear whether yolk sac derived hematopoietic progenitors that do not undergo erythropoiesis in the yolk sac may take up intraembryonic niches and contribute to erythropoietic stem cell population after hatching. Developmental Dynamics 237:3332–3341, 2008. © 2008 Wiley-Liss, Inc. Key words: chicken; erythropoiesis; yolk sac; hematopoiesis; definitive; hemoglobin; globin; betaA; artery; vein; niche; locked nucleic acid Accepted 15 August 2008

INTRODUCTION It has been well-documented that extraembryonic tissues, including the yolk sac and allantois in birds and mammals and the placenta in mammals, contribute to the generation of definitive erythrocytes during amniote development (Dantschakoff, 1908; Moore and Owen, 1965; Moore and Metcalf, 1970; Dieterlen-Lievre et al., 1976; Weissman et al., 1978; Beaupain et al., 1979; Caprioli et al., 1998; Palis and Yoder, 2001; Gekas et al., 2005; McGrath and Palis, 2005; Mikkola et al., 2005; Ottersbach and Dzierzak, 2005; Zeigler et al., 2006; Corbel et al., 2007; Samokhvalov et al., 2007; Lux et al., 2008; Rhodes et

al., 2008). What remain not well-understood are the precise cellular source and niche of definitive erythropoiesis and the relative contributions of intraembryonic- and extraembryonic-generated definitive erythrocytes in embryonic blood. In birds, chick/quail chimera studies suggest that yolk sac derived cells contribute to the majority of circulating blood cells prior to embryonic day (E) 12 (DieterlenLievre et al., 1976; Beaupain et al., 1979), although intraembryonically generated hematopoietic cells can be detected in the ventral wall of dorsal aorta as early as at E3– 4 (Jaffredo et al., 2005). It is unclear when and where hematopoietic cells generated

from the dorsal aorta initiate terminal differentiation to supply definitive erythrocytes for the developing embryo. Primitive erythrocytes in chickens express mainly rho and epsilon beta globins and definitive erythrocytes mainly betaA beta globin (Nakazawa et al., 2006; Alev et al., 2008). Globin transcript and protein analyses (Simons, 1966; Wilt, 1967; Shimizu, 1972; Bruns and Ingram, 1973; Minie et al., 1992; Mason et al., 1995; Singal and vanWert, 2001) and erythrocyte morphological analysis (Fraser, 1961; Lucas and Jamroz, 1961; Minie et al., 1992) indicate that definitive erythrocytes in chickens appear in circulation at about E5. We

Laboratory for Early Embryogenesis, RIKEN Center for Developmental Biology, Kobe, Hyogo, Japan *Correspondence to: Guojun Sheng, Laboratory for Early Embryogenesis, RIKEN Center for Developmental Biology, Kobe, Hyogo 650-0047, Japan. E-mail: [email protected] DOI 10.1002/dvdy.21746 Published online 10 October 2008 in Wiley InterScience (www.interscience.wiley.com).

© 2008 Wiley-Liss, Inc.

ERYTHROPOIESIS IN CHICKEN YOLK SAC 3333

have recently reported that a burst of definitive erythrocyte generation between E5 and E7 results in the majority of circulating blood cells displaying a definitive profile (Alev et al., 2008). These data suggest that definitive erythrocytes in chickens are being generated at a time when hematopoietic cells are being specified intraembryonically in dorsal aorta. It is unclear, however, how these two events are linked. Using betaA globin as a marker for definitive erythrocytes undergoing terminal differentiation and several hematopoietic transcriptional regulators as markers for hematopoiesis in progress, we analyzed the initiation, duration and site of definitive erythropoiesis in the yolk sac.

RESULTS betaA LNA Probe Is Used to Detect Definitive Erythrocytes We have previously reported that betaA transcripts and protein, although highly expressed in definitive erythrocytes, are also weakly expressed in primitive erythrocytes (Alev et al., 2008). For the current study, the presence of betaA transcripts is used as a marker for definitive erythrocytes as relatively large difference in the levels of its expression in primitive and definitive blood cells makes this distinction unambiguous. In order to clearly distinguish different globin transcripts and to avoid the transient nature of nascent nuclear transcripts detected by intron-based specific probes (Nagai and Sheng, 2007; Alev et al., 2008), we tested the feasibility of generating a locked-nucleic-acid (LNA)based in situ hybridization probe for betaA. LNA probes have been widely used for detecting microRNAs because of its stable annealing with short stretches of RNA sequences (Wienholds et al., 2005; Darnell et al., 2006; Cao et al., 2007; Obernosterer et al., 2007), but so far have not been reported for regular mRNA in situ analysis in chickens. We reasoned that the abundance of globin transcripts in erythrocytes and the highly discriminative nature of short LNA probes may make them ideally suitable for distinguishing different globin tran-

scripts. Two LNA probes for betaA were generated, with one of them giving strong and clean signals and used for all subsequent studies (see the Experimental Procedures section).

Definitive Erythrocytes Are Generated in Venous Vessels Adjacent to Arteries The percentage of betaA protein among all beta globins in circulating blood shows a rapid increase between E5 and E7 (Alev et al., 2008), suggesting that definitive erythrocytes are being abundantly generated during or before this period. We investigated the location of this generation by performing whole-mount in situ hybridization analysis using betaA LNA probe with E5 yolk sac tissue (Fig. 1A,B; see also Experimental Procedures). A small percentage of betaA positive definitive cells could be seen in circulation in major vessels (Fig. 1C, arrowheads), mixed with betaA negative primitive blood cells. Cell clusters with betaA positive signals, however, were observed much more abundantly in association with smaller vessels. These vessels appeared to have slow or no blood flow, and are of venous nature, as most arterial vessels have started to sink into the yolk by this stage (Fig. 1D, black arrows). Larger clusters were detected adjacent to arteries (Fig. 1D, red arrows) and smaller ones in between major vessels (Fig. 1D, red arrowheads). Cells in these clusters have the morphology of less mature blood cells (Fig. 2A-C), similar to differentiating primitive blood cells before the initiation of circulation (Nakazawa et al., 2006). Larger clusters surrounding major arteries were observed to have both intravascular (Fig. 2A, arrowheads) and extravascular locations (Fig. 2A, arrows) in association with venous vessels. Smaller clusters located in between major vessels were often seen to be entirely composed of erythropoietic cells (Fig. 2B). In both large and small vessels, intravascular clusters (Fig. 2C red arrowhead) were often observed with some primitive erythrocytes next to them (Fig. 2C, black arrowhead). Whole-mount view of betaA stained yolk sac tissue (Fig. 1E,F) indicated that these clusters are lined along all major arteries (Fig. 1F, arrows) and

smaller vessels situated between major arteries (Fig. 1F, arrowheads).

Definitive Erythropoiesis in the Yolk Sac Is Accompanied by the Expression of Hematopoietic Transcription Regulators Primitive erythropoiesis in chickens is preceded by the expression of several hematopoietic transcriptional regulators, including gata1, scl, and lmo2 (Nakazawa et al., 2006). Genetic and molecular analyses in mice indicated that these regulators are critical for both primitive (Warren et al., 1994; Shivdasani et al., 1995; Fujiwara et al., 1996; Porcher et al., 1996; Robb et al., 1996; Takahashi et al., 1997) and definitive erythropoiesis (Shivdasani et al., 1995; Porcher et al., 1996; Robb et al., 1996; Yamada et al., 1998; Shimizu et al., 2001). We, therefore, asked whether they are also involved in definitive erythropoiesis in chicken yolk sac. In situ hybridization analysis of E5 yolk sac tissue showed that the expression of all three genes has a similar whole-mount pattern to that of betaA (Fig. 3A,D,G), with dense clusters lined along main arteries (Fig. 3B,E,H, arrows) and smaller clusters spread in between them (Fig. 3B,E,H, arrowheads). Sections of these stained tissues revealed that all three genes, like betaA, are strongly expressed in differentiating definitive erythrocytes in both venous vessels (Fig. 3C,F,I, arrows) adjacent to arteries, and smaller vessels in between (Fig. 3C,F,I, arrowheads).

Generation of Definitive Erythrocyte in the Yolk Sac Starts at E4 –E4.5 and Lasts After E15 We next asked how early the yolk sac definitive erythropoiesis starts and for how long it lasts. Yolk sac tissues from E3 to E15 embryos were dissected out for in situ analysis with betaA LNA probe. The timing of the earliest generation of definitive erythrocytes was assessed using E3-E4.5 yolk sac tissues, with half a day increments and with precise stages judged by embryobased Hamburger and Hamilton criteria. The earliest definitive erythropoi-

3334 NAGAI AND SHENG

Fig. 1. Definitive erythropoiesis in the yolk sac revealed with betaA LNA probe. A,B: Overview of embryonic day (E) 5 embryo showing the embryo, allantois and yolk sac (side view in A and top view in B). C: Some betaA positive definitive erythrocytes (arrowheads) can be seen among primitive erythrocytes in circulation within a yolk sac artery. A single cell-thick layer of large yolk sac endodermal cells can be seen covering the ventral surface of the arterial vessel. D: More abundant betaA-positive cells can be seen around major arteries (black arrows) as clusters associated with venous vessels (red arrows). Smaller clusters can be seen in small vessels in between (red arrowheads). E: Whole-mount view of a piece of betaA stained E5 yolk sac tissue. F: Magnified view of a region in E. Arrows indicate intense staining along arteries. Arrowheads indicate less dense staining in smaller vessels and vascular plexus. Scale bars ⫽ 25 ␮m in C and 100 ␮m in D.

Fig. 2. Intra- and extravascular locations of definitive erythropoietic clusters. Sections of tissue shown in Figure 1E. A: Around the artery, both intravascular (arrowheads) and extravascular (arrows) positive cells are associated with venous vessels. B: A small vessel with all associated cells being betaA positive. C) A vessel with most associated cells being betaA-positive (red arrowhead) and with some betaA-negative primitive blood cells in the center (black arrowhead). Scale bars ⫽ 25 ␮m.

etic clusters were detected at E4 –E4.5 (Fig. 4A–D), with some E4 yolk sac being negative (Fig. 4B) and some positive (Fig. 4C,E) for betaA. Positive E4 yolk sac tissues have betaA clusters surrounding large yolk sac arteries proximal to the developing embryo (Fig. 4C,E, arrows in E), with the most proximal part, however, being nega-

Fig. 3.

Fig. 3. Definitive erythropoietic cells are positive for scl, gata1 and lmo2. A,D,G: Whole-mount view of embryonic day (E) 5 yolk sac tissue stained for scl (A), gata1 (D) or lmo2 (G). B,E,H: Magnified view of a region in A, D and G, respectively. C,F,I: Section view of tissue shown in A, D and G, respectively. Arrows indicate intense staining along arteries. Arrowheads indicate less dense staining in smaller vessels and vascular plexus. Scale bars ⫽ 25 ␮m.


Fig. 4.

tive (Fig. 4C,D, arrowheads). BetaA positive clusters become much more abundant at E4.5 (Fig. 4D,F), and at E5 and afterwards, intensely stained definitive erythropoietic clusters are seen throughout the yolk sac as shown in Figure 1. Sectioning of positive E4 yolk sac revealed that the earliest definitive erythrocytes are generated in venous vessels adjacent to arteries (Fig. 4G-L), either as intravascular clusters (Fig. 4G,H,J,K, arrowheads), extravascular clusters (Fig. 4H,K, arrows) or individual cells (Fig. 4I,L, arrows). In situ analysis with later stage yolk sac tissues showed that extensive erythropoiesis takes place at least until E10 (Fig. 5A,B). Circulating blood at this stage contains predominantly definitive erythrocytes (Alev et al., 2008), although betaA transcript levels in circulating mature definitive erythrocytes are much reduced. High levels of betaA transcripts could be seen in erythropoietic clusters (Fig. 5C,D, arrowheads in D) surrounding arteries (Fig. 5C,D, arrows). Similar to at earlier stages, these clusters are also positive for hematopoietic transcriptional regulators (data not

Fig. 4. Initiation of definitive erythropoiesis in the yolk sac. All panels are stained for betaA transcripts. A,B: Definitive erythropoiesis does not happen in embryonic day (E) 3 (A) or some E4 (B) embryos. C: Some E4 embryos start to initiate definitive erythropoiesis, along main arteries proximal to the embryo, with magnified view shown in E. The most proximal region is negative (arrowheads). D: E4.5 yolk sac has more abundant definitive erythropoietic clusters, with magnified view shown in F. The most proximal region is still negative (arrowheads). E: Magnified view of C. Arrows indicate betaA-positive clusters. F: Magnified view of D. G–L: Sections of C. G: (magnified in J): betaA positive cluster (arrowhead) inside a venous vessel adjacent to an artery. H (magnified in K): A venous vessel associated with some intravascular positive cells (arrowheads) and extravascular positive cluster (arrows). I (magnified in L): Individual betaA-positive cells (arrows) can be seen associated with ventral wall of a venous vessel. Dotted lines in J–L outline the vessel. Scale bars ⫽ 25 ␮m.

Fig. 5.

Fig. 5. Duration of definitive erythropoiesis in the yolk sac. All panels are stained for betaA transcripts. A,B: Embryonic day (E) 10 yolk sac. C,D: Section of E10 yolk sac (C), with magnified view in D. Black arrows: arteries. Red arrowheads: erythropoietic clusters in venous vessels around the artery. E: E15 yolk sac. Arrows indicate erythropoietic clusters. Scale bars ⫽ 100 ␮m in C and 25 ␮m in D.


shown). In E15 yolk sac, a reduced, although still quite prominent, level of erythropoiesis could be detected (Fig. 5E, arrows), accompanied by the expression of hematopoietic regulators (data not shown).

Pre-existing Vascular Endothelium Is a Likely Source for Definitive Erythropoietic Cells To understand the source of cells undergoing definitive erythropoiesis, we investigated more closely the transition period between E2, when embryonic circulation starts, to E5, when definitive erythrocytes are detected in circulation. First, we asked whether enhanced proliferation of perivascular cells may account for the rapid appearance of these clusters. Phosphohistone H3 staining, marking all mitotic cells, indicated that there is no prominent increase in the number of mitotic cells in peri-vascular regions during this period in either wholemount (Fig. 6A,B) or section (data not shown) examination. A similar observation, in addition to the lack of prominent apoptosis during this period, has also been reported previously (le Noble et al., 2004). This suggested that these definitive erythropoietic cells are derived either from blood-borne hematopoietic cells or from pre-existing vascular cells. It has been reported that during this period (E3-E4), hemogenic endothelial cells are detected in the ventral wall of dorsal aorta intraembryonically (Jaffredo et al., 2005; Dieterlen-Lievre et al., 2006). We, therefore, investigated whether a similar process may be taking place in extraembryonic vessels. Like betaA, scl and gata1 were detected in yolk sac erythropoietic clusters surrounding large arteries at E4.5, along main arteries (Fig. 6D for gata1, data not shown). Consistent with the idea that these transcription factors regulate definitive blood cell differentiation, the earliest expression of both scl and gata1 in definitive erythropoietic cells could be seen at E3.5–E4 (Fig. 6C,E-G for gata1 and 6L,M for scl), slightly earlier than betaA at E4-E4.5. scl and gata1, however, were mainly detected in cells that do not have an endothelial morphology; whereas lmo2 could be seen

during this period in forming erythropoietic clusters (Figs. 6O, 7C,D, arrows) and in vascular endothelial cells surrounding them (Figs. 6O, 7C,D, arrowheads), suggesting that definitive erythropoietic cells in the yolk sac may be generated by means of a hemogenic endothelium step, similar to what has been suggested in dorsal aorta in both mice and chickens. Although the appearance of betaA positive definitive erythropoietic clusters at E4 – 4.5 is unambiguous, the expression of lmo2, scl, and gata1 showed a temporal connection to primitive erythrocytes and primitive vascular endothelium. During early primitive erythropoiesis, lmo2 and scl mark hemangioblast cells prior to terminal differentiation (Nakazawa et al., 2006). After terminal differentiation of primitive erythrocytes, scl expression becomes restricted to blood cells, whereas lmo2 is expressed in both blood and endothelial cells (Nakazawa et al., 2006; Weng et al., 2007). At E3, a day after the initiation of circulation, strong lmo2 expression becomes restricted to endothelial cells (Fig. 6J,K) and scl expression remains prominent in all blood cells (Fig. 6H,I). Between E3 and E3.5, scl expression is rapidly lost in primitive blood cells and only to reappear in definitive erythropoietic cells about half a day later (Fig. 6L,M). During the same period, lmo2 expression is lost in arterial vascular endothelial cells (Figs. 6N, 7A,B, notice the lack of arterial vessel staining). This loss is maintained at later stages (Figs. 6O, 7C,D, notice the lack arterial vessel staining). Although at this stage most venous vessel endothelial cells show lmo2 positivity, strong up-regulation could be seen in some endothelial cells in venous vessels situated around arteries (Fig. 7A,B arrows) and in association with small erythropoietic clusters (data not shown).

Yolk Sac Definitive Erythropoiesis Is Linked to Remodeling of Yolk Sac Vasculature The physiological trigger for this drastic change during E3–E4 in scl and lmo2 expression is unclear. During the same period, intraembryonic hemogenic endothelial cells were re-

ported to appear in dorsal aorta. It is also during this period that the yolk sac vasculature undergoes a major remodel from primary (apposing) to secondary (paired) arterial/venous connection (le Noble et al., 2004). We, therefore, investigated spatial and temporal relationships between yolk sac definitive erythropoiesis and vascular remodeling. E3 to E6 embryos were injected in ovo and intracardiacally with fluorescein labeled dextran to visualize all vessels, followed by injection of rhodamine labeled dextran mixed with fixative (see the Experimental Procedures section). Cardiac arrest caused by the fixative was observed within 5-10 sec, allowing visualization of all vessels with fluorescein and main yolk sac arteries with rhodamine. In E3 embryos (Fig. 8A), wellarborized yolk sac arterial and venous trees are readily discernable. The arterial/venous pairing, however, is still rudimentary at this stage and is restricted to very proximal regions (double-headed arrow in merge). In E4 embryos, the arterial/venous pairing extends more peripherally (Fig. 8B, double-headed arrows in merge). This process seems to take place gradually as the embryo develops, and by E6, most large and proximal arterial branches are paired with large venous vessels (Fig. 8C, paired blue and red arrowheads in Fig. 8F–H). Smaller arteries at E6 still lack paired veins with a similar caliber (red arrowheads in Fig. 8H–K), although poorly circulating smaller venous vessels could be often seen situated on top of these arteries (blue arrows in Fig. 8I,K). At E4, when pairing is actively taking place, arteries could often be seen to have clusters of densely packed cells in poorly circulating vessels over them (Fig. 8E, arrows), in agreement with our in situ hybridization observation that smaller venous vessels surrounding main arterial vessels contain numerous erythropoietic clusters.

DISCUSSION A century ago, Dantschakoff reported a detailed histological description of definitive erythropoiesis in chicken yolk sac (Dantschakoff, 1908). This has not been followed up with recently available tools. Our analyses with molecular markers largely confirmed her


observations. In particular, we show that the yolk sac is the major source of cells and niche for definitive erythropoiesis and that this takes place in venous vessels around major arteries. In addition, we report here that this is tightly connected to vascular remodeling in the yolk sac. We propose that there are two phases of vascular remodeling in chicken yolk sac. The first phase, in which yolk sac vascular plexus is remodeled to form well-arborized arterial and venous trees, takes place between E2 and E3, soon after the initiation of embryonic circulation. This phase of vascular remodeling has been reported to be controlled by hemodynamics in the yolk sac (le Noble et al., 2004). Our observations suggest that yolk sac definitive erythropoiesis does not coincide with, although it may be a consequence of, this remodeling. In fact, large venous vessels formed in this phase are negatively correlated with the definitive erythropoiesis. The second phase is initiated after E3, during which many smaller venous vessels adjacent to major arteries fuse to form larger venous vessels lying next to, and subsequently over, arteries. Vast majority of yolk sac definitive erythropoiesis is associated with these fusing Fig. 6.

Fig. 6. Onset of definitive erythropoiesis and its link to primitive blood and vasculature. A,B: Phospho-histone H3 staining shows uniform mitotic pattern, with no enhanced mitosis along major arteries. A: Embryonic day (E) 3 yolk sac. B: E5 yolk sac. C,D: gata1 is robustly and restrictively expressed along major arteries at E4 (stage 23, C) and E4.5 (stage 24, D). E: Magnified view of C. F,G: Sections of C, showing venous vessel-associated expression of gata1 adjacent to arteries. H: scl expression at E3 (stage 18). I: Magnified view of H, showing positive primitive blood cell staining. J: lmo2 expression at E3 (stage 18). K: Magnified view of J, showing positive endothelial cell staining. L: scl expression at E3.5– 4 (stage 22). M: Magnified view of a region in L. N: lmo2 expression at E3–3.5 (stage 20). O: lmo2 expression at E4 (stage 23). Scale bars ⫽ 25 ␮m.

Fig. 7.

Fig. 7. Pre-existing vascular endothelium as a source for definitive erythropoietic cells. A: Section of yolk sac in Fig. 6N. B: Magnified view. Arrows: strong lmo2-positive signals in venous endothelial cells adjacent to the artery. Dotted lines outline basal surface of artery and vein. C,D: Sections of yolk sac shown in Figure 6O. Red arrowheads: positive endothelial cells. Black arrowheads: negative endothelial cells. Arrows: positive clusters. Scale bars ⫽ 25 ␮m.


venous vessels. This phase seems to continue throughout embryonic development. Our observations indicate that yolk sac definitive erythropoiesis takes place both intravascularly (inside venous vessel lumen) and extravascularly (in perivascular space). As no extraordinary cell proliferation has been observed, one likely source of the erythropoietic cells is pre-existing endothelial cells of these venous vessels. Although we cannot exclude a possible contribution from blood-borne erythropoietic cells that are derived either from intraembryonic tissues or from primitive erythropoietic wave in the yolk sac, the scale and rapidity of yolk sac definitive erythropoiesis and the lack of enhanced cell division suggest that pre-existing venous endothelium is one main cellular source. These hemogenic endothelial cells lose endothelial integrity and initiate erythropoiesis either inside the vessel lumen or extravascularly. This notion of hemogenic endothelium, however, should be distinguished from the debated definition of hemogenic endothelial cell as bi-potential blood/endothelial progenitor or stem cell, which after cell division gives rise to either an endothelial cell/hematopoietic cell daughter pair (bi-potential progenitor) or a hemogenic endothelial cell/ hematopoietic cell daughter pair (bipotential stem cell). Our data cannot rule out or distinguish between these possibilities. Another possible main source of cells for definitive erythropoiesis in the yolk sac, which has not been investigated in this work, is differentiated primitive erythrocytes. De-differentiation of chicken primitive erythroid cells into definitive type multipotential hematopoietic progenitors has been reported in culture system (McNagny and Graf, 2003). Differentiation plasticity of hematopoietic cells, with de-differentiation as one possible manifestation of the plasticity, has been increasingly realized (Dzierzak, 2002; Graf, 2002; Orkin and Zon, 2002; Prindull and Fibach, 2007). Our data, although not directly supporting it, are not incompatible with a possible dual source model. We do not know whether the induction of yolk sac erythropoiesis is triggered by hemodynamic changes in these remodeling vessels or by a sud-

den increase of pro-erythropoietic cytokines derived either from developing organs intraembryonically or from associated yolk sac arteries as they sink into the yolk. It is possible that most of the early vessel cells are hemogenic and gradually lose the hemogenic potential. This is supported by our observation that lmo2 is expressed in all vessel cells at E3 and its expression is lost between E3 and 4 in endothelial cells of the vessels that are negatively associated with erythropoietic clusters. This loss takes place in most yolk sac arterial vessels and later on in major venous vessels that are not paired with arteries. In our live embryo observation, the venous vessels that are associated with extensive erythropoiesis, although having slower blood flow due to numerous erythropoietic clusters bulging into the vessel and thus restricting the vessel flow, are nevertheless connected to the general circulation (Fig. 8D–K). It, therefore, remains unclear what are the mechanisms used to ensure that only relatively mature definitive erythrocytes enter the circulation. Occasionally, we have observed circulating scl-positive, but never betaA-positive, clusters inside large vessels intraembryonically, including inside dorsal aorta and heart chambers (data not shown). These clusters are to be distinguished from smaller clusters of hemogenic cells on the ventral wall of dorsal aorta (Jaffredo et al., 2005), which we have also consistently observed to be positive for hematopoietic transcription factors used in our work (Fig. 8L–O, arrows). It is, therefore, possible that a small percentage of erythropoietic cells from yolk sac erythropoietic clusters may enter the circulation before starting terminal differentiation, and subsequently take up embryonic niches for differentiation. We have previously reported that E4 blood contains a rich repertoire of hematopoietic progenitor and stem cells (McIntyre et al., 2008). Given the scale of yolk sac erythropoiesis, and possibly hematopoiesis to generate other blood lineages, the yolk sac may also be the source for these circulating progenitor and stem cells. The nearly identical timing in the formation of hematopoietic cells intraembryonically in dorsal aorta and

extraembryonically in the yolk sac suggests a physiological connection between these two events. The scale of definitive erythropoiesis in the yolk sac indicates that intraembryonically generated erythropoietic cells do not make a significant, if any, contribution to yolk sac definitive erythropoiesis. However, the identical timing also strongly supports the notion that erythropoietic cells attached to the ventral wall of dorsal aorta are generated intraembryonically from aortic endothelial cells. A critical difference between these two events may be in the regulation of terminal differentiation (i.e., from erythropoietic cells to definitive erythrocytes). In chicken yolk sac, most erythropoietic cells undergo terminal differentiation in situ, soon after hematopoietic induction; whereas dorsal aorta derived hematopoietic cells remain dormant for a long time, either in circulation or in paraaortic foci, until proper niches for erythropoiesis (e.g., bone marrow) mature. In mammals, yolk sac derived hematopoietic cells do not undergo definitive erythropoiesis in situ, but rather find a secondary niche in placenta, allantois or fetal liver soon after their generation. It remains possible that there is an intrinsic qualitative difference between these two definitive erythropoietic populations, and that only the dorsal aorta derived one will contribute to true hematopoietic/ erythropoietic stem cell population.

EXPERIMENTAL PROCEDURES Embryology, In Situ Hybridization, and Histology Fertilized eggs were purchased from Shiroyama Farm (Kanagawa, Japan) and incubated to desired stages at 38.5°C. All embryonic day stages correspond to full incubation time (e.g., E3 indicates 72 hr of incubation), and confirmed by Hamburger and Hamilton staging system. All in situ hybridization analyses with yolk sac tissues followed the general in situ protocol (Stern, 1998) with two additional modifications. First, after fixation, fine incisions were made in yolk sac tissues with needles to avoid all possible trappings in subsequent hybridization steps. Second, the concentra-


Fig. 8. Relationship of definitive erythropoiesis with yolk sac vascular remodeling and dorsal aorta. A,B,C: Images of yolk sac vascular trees at embryonic day (E) 3 (A), E4 (B), and E6 (C). Left panels: bright field views. Green: fluorescein dextran visualization of all vessels. Red: rhodamine dextran visualization of arterial trees. Doubleheaded arrows in merged view indicate degree of arterial/venous pairing. D,E: Bright field view (D) and fluorescein visualization (E) of E4 artery (right vessel). Dense cell clusters can be seen situated over artery (arrows in E) within poorly circulating venous vessel. F–K: Fluorescein visualization of E6 vessels. Pairing of artery (red arrowhead) and vein (blue arrowhead) can be seen in proximal vessels (F) and main branches (G,H), whereas more distally located arteries (red arrowhead in J,K) and smaller side branches (red arrowhead in H,I) still lack paired veins. I and K show magnified view of H and J, respectively. Small poorly circulating venous vessels can be seen over unpaired arteries (blue arrows in I and K). Single blue arrow in H shows main venous vessel that is resulted from early yolk sac vascular remodeling and does not get paired with an artery. L–O: Expression of scl (L,N) and lmo2 (M,O) in the ventral wall of dorsal aorta at E3.5– 4. Arrows indicate hematopoietic cell clusters associated with the ventral wall. Scale bars ⫽ 25 ␮m.


tion and duration of Proteinase K treatment were adjusted to increase probe penetration (twice the normal concentration and/or twice the normal treatment time, depending on the age of the yolk sac). The corresponding regions for gata1, scl, and lmo2 probes have been described previously (Nakazawa et al., 2006). After in situ hybridization, yolk sac tissues were photographed whole-mount with Olympus SZX12 microscope (Olympus, Tokyo, Japan) and Olympus DP70 camera, followed by paraffin embedded sectioning. All sections were 10 ␮m thick and were examined with Olympus BX51 microscope. Anti–Ser10-phosphorylated histone H3 antibody was purchased from Upstate Cell Signaling (NY, #06-570).

Labeled Locked-Nucleic-Acid (LNA) Probe and In Situ With LNA Probes Two betaA 5⬘-FAM labeled LNA probes were purchased from Nippon EGT Co. Ltd. (Toyama, Japan), corresponding to 12-31 nucleotides (gTaggTgCucCgTgAuCuTu, with LNA replacements shown in capital letters) and 14-35 nucleotides (gguuguAGgugcucCguGAucu, with LNA replacements shown in capital letters) after the stop codon in the 3⬘ UTR region. The second one (14-35) was used for all studies presented here. The in situ procedure for labeled LNA probe was identical to normal procedure (as mentioned above), except that the hybridization was done at 48°C instead of 68°C. AP coupled anti-fluorescein antibody (Roche, #11061700) was used for FAM epitope detection in the LNA probe.

Imaging of Yolk Sac Vasculature With Fluorescent Dyes For yolk sac vascular imaging, fertilized eggs incubated to desired stages were opened and injected intracardiacally with 0.1– 0.5 ␮l of 10 mg/ml 500 kDa fluorescein-labeled dextran (Molecular Probes, #D-7136). Within 1 min, all intraembryonic and extraembryonic vessels were observed to contain injected dye. Images were taken either with live embryos after this

first injection, or after the second injection. The second injection, with 0.5-1 ␮l of 20-40 mg/ml 3 kDa tetramethylrhodamine-labeled dextran (Molecular Probes, #D-3308) together with 1–2% (final concentration) of paraformaldehyde, was performed after reducing the heart beat rate with a few drops of ice-cold Pannett-Compton solution. This ensured a complete stop of heart beat within 5–10 sec, resulting in the second dye labeling only major yolk sac arteries. Images were taken immediately after the second injection.

ACKNOWLEDGMENTS We thank Dr. I.M. Samokhvalov and Dr. C. Alev for critical discussions, Dr. C. Alev for the translation of Dr. W. Dantschakoff’s work, Dr. W. Weng for help with anti-phospho-H3 staining and for proofreading, and Mr. Y. Sheng for help with figure preparation.

REFERENCES Alev C, McIntyre BA, Nagai H, Shin M, Shinmyozu K, Jakt LM, Sheng G. 2008. BetaA, the major beta globin in definitive red blood cells, is present from the onset of primitive erythropoiesis in chicken. Dev Dyn 237:1193-1197. Beaupain D, Martin C, Dieterlen-Lievre F. 1979. Are developmental hemoglobin changes related to the origin of stem cells and site of erythropoiesis? Blood 53: 212-225. Bruns GA, Ingram VM. 1973. The erythroid cells and haemoglobins of the chick embryo. Philos Trans R Soc Lond B Biol Sci 266:225-305. Cao X, Pfaff SL, Gage FH. 2007. A functional study of miR-124 in the developing neural tube. Genes Dev 21:531-536. Caprioli A, Jaffredo T, Gautier R, Dubourg C, Dieterlen-Lievre F. 1998. Blood-borne seeding by hematopoietic and endothelial precursors from the allantois. Proc Natl Acad Sci U S A 95:1641-1646. Corbel C, Salaun J, Belo-Diabangouaya P, Dieterlen-Lievre F. 2007. Hematopoietic potential of the pre-fusion allantois. Dev Biol 301:478-488. Dantschakoff V. 1908. Untersuchungen u¨ber die Entwickelung des Blutes und Bindegewebes bei den Vo¨geln. Anatomische Hefte 37:471-587. Darnell DK, Kaur S, Stanislaw S, Konieczka JH, Yatskievych TA, Antin PB. 2006. MicroRNA expression during chick embryo development. Dev Dyn 235:31563165. Dieterlen-Lievre F, Beaupain D, Martin C. 1976. Origin of erythropoietic stem cells in avian development: shift from the yolk

sac to an intraembryonic site. Ann Immunol (Paris) 127:857-863. Dieterlen-Lievre F, Pouget C, Bollerot K, Jaffredo T. 2006. Are intra-aortic hemopoietic cells derived from endothelial cells during ontogeny? Trends Cardiovasc Med 16:128-139. Dzierzak E. 2002. Hematopoietic stem cells and their precursors: developmental diversity and lineage relationships. Immunol Rev 187:126-138. Fraser RC. 1961. Hemoglobin formation in the chick embryo. Exp Cell Res 25:418427. Fujiwara Y, Browne CP, Cunniff K, Goff SC, Orkin SH. 1996. Arrested development of embryonic red cell precursors in mouse embryos lacking transcription factor GATA-1. Proc Natl Acad Sci U S A 93:12355-12358. Gekas C, Dieterlen-Lievre F, Orkin SH, Mikkola HK. 2005. The placenta is a niche for hematopoietic stem cells. Dev Cell 8:365-375. Graf T. 2002. Differentiation plasticity of hematopoietic cells. Blood 99:3089-3101. Jaffredo T, Nottingham W, Liddiard K, Bollerot K, Pouget C, de Bruijn M. 2005. From hemangioblast to hematopoietic stem cell: an endothelial connection? Exp Hematol 33:1029-1040. le Noble F, Moyon D, Pardanaud L, Yuan L, Djonov V, Matthijsen R, Breant C, Fleury V, Eichmann A. 2004. Flow regulates arterial-venous differentiation in the chick embryo yolk sac. Development 131:361-375. Lucas AM, Jamroz C. 1961. Atlas of avian hematology. Washington, D.C.: U.S. Dept. of Agriculture. vi, 271 p. pp. Lux CT, Yoshimoto M, McGrath K, Conway SJ, Palis J, Yoder MC. 2008. All primitive and definitive hematopoietic progenitor cells emerging before E10 in the mouse embryo are products of the yolk sac. Blood 111:3435-3438. Mason MM, Lee E, Westphal H, Reitman M. 1995. Expression of the chicken betaglobin gene cluster in mice: correct developmental expression and distributed control. Mol Cell Biol 15:407-414. McGrath KE, Palis J. 2005. Hematopoiesis in the yolk sac: more than meets the eye. Exp Hematol 33:1021-1028. McIntyre BA, Alev C, Tarui H, Jakt LM, Sheng G. 2008. Expression profiling of circulating non-red blood cells in embryonic blood. BMC Dev Biol 8:21. McNagny KM, Graf T. 2003. E26 leukemia virus converts primitive erythroid cells into cycling multilineage progenitors. Blood 101:1103-1110. Mikkola HK, Gekas C, Orkin SH, Dieterlen-Lievre F. 2005. Placenta as a site for hematopoietic stem cell development. Exp Hematol 33:1048-1054. Minie ME, Kimura T, Felsenfeld G. 1992. The developmental switch in embryonic rho-globin expression is correlated with erythroid lineage-specific differences in transcription factor levels. Development 115:1149-1164. Moore MA, Metcalf D. 1970. Ontogeny of the haemopoietic system: yolk sac origin


of in vivo and in vitro colony forming cells in the developing mouse embryo. Br J Haematol 18:279-296. Moore MA, Owen JJ. 1965. Chromosome marker studies on the development of the haemopoietic system in the chick embryo. Nature 208:956 passim. Nagai H, Sheng G. 2007. Cis-cotranscription of two beta globin genes during chicken primitive hematopoiesis. PLoS ONE 2:e703. Nakazawa F, Nagai H, Shin M, Sheng G. 2006. Negative regulation of primitive hematopoiesis by the FGF signaling pathway. Blood 108:3335-3343. Obernosterer G, Martinez J, Alenius M. 2007. Locked nucleic acid-based in situ detection of microRNAs in mouse tissue sections. Nat Protoc 2:1508-1514. Orkin SH, Zon LI. 2002. Hematopoiesis and stem cells: plasticity versus developmental heterogeneity. Nat Immunol 3:323-328. Ottersbach K, Dzierzak E. 2005. The murine placenta contains hematopoietic stem cells within the vascular labyrinth region. Dev Cell 8:377-387. Palis J, Yoder MC. 2001. Yolk-sac hematopoiesis: the first blood cells of mouse and man. Exp Hematol 29:927-936. Porcher C, Swat W, Rockwell K, Fujiwara Y, Alt FW, Orkin SH. 1996. The T cell leukemia oncoprotein SCL/tal-1 is essential for development of all hematopoietic lineages. Cell 86:47-57. Prindull GA, Fibach E. 2007. Are postnatal hemangioblasts generated by dedifferentiation from committed hematopoietic stem cells? Exp Hematol 35:691-701. Rhodes KE, Gekas C, Wang Y, Lux CT, Francis CS, Chan DN, Conway S, Orkin SH, Yoder MC, Mikkola HK. 2008. The emergence of hematopoietic stem cells is

initiated in the placental vasculature in the absence of circulation. Cell Stem Cell 2:252-263. Robb L, Elwood NJ, Elefanty AG, Kontgen F, Li R, Barnett LD, Begley CG. 1996. The scl gene product is required for the generation of all hematopoietic lineages in the adult mouse. Embo J 15:41234129. Samokhvalov IM, Samokhvalova NI, Nishikawa S. 2007. Cell tracing shows the contribution of the yolk sac to adult haematopoiesis. Nature 446:1056-1061. Shimizu K. 1972. Ontogeny of chicken hemoglobin. I. Electrophoretic study of the heterogeneity of hemoglobin in development. Dev Growth Diff 14:43-55. Shimizu R, Takahashi S, Ohneda K, Engel JD, Yamamoto M. 2001. In vivo requirements for GATA-1 functional domains during primitive and definitive erythropoiesis. Embo J 20:5250-5260. Shivdasani RA, Mayer EL, Orkin SH. 1995. Absence of blood formation in mice lacking the T-cell leukaemia oncoprotein tal-1/SCL. Nature 373:432-434. Simons JA. 1966. The ontogeny of the multiple molecular forms of hemoglobin in the developing chick under normal and experimental conditions. J Exp Zool 162: 219-230. Singal R, vanWert JM. 2001. De novo methylation of an embryonic globin gene during normal development is strand specific and spreads from the proximal transcribed region. Blood 98:3441-3446. Stern CD. 1998. Detection of multiple gene products simultaneously by in situ hybridization and immunohistochemistry in whole mounts of avian embryos. Curr Top Dev Biol 36:223-243. Takahashi S, Onodera K, Motohashi H, Suwabe N, Hayashi N, Yanai N, Nabesima

Y, Yamamoto M. 1997. Arrest in primitive erythroid cell development caused by promoter-specific disruption of the GATA-1 gene. J Biol Chem 272:1261112615. Warren AJ, Colledge WH, Carlton MB, Evans MJ, Smith AJ, Rabbitts TH. 1994. The oncogenic cysteine-rich LIM domain protein rbtn2 is essential for erythroid development. Cell 78:45-57. Weissman IL, Papaioannou V, Gardner R. 1978. Fetal hematopoietic origins of the adult hematolymphoid system. In: Clarkson B, Marks PA, Till JE, editors. Differentiation of normal and neoplastic hematopoietic cells. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory. pp 33-47. Weng W, Sukowati EW, Sheng G. 2007. On hemangioblasts in chicken. PLoS ONE 2:e1228. Wienholds E, Kloosterman WP, Miska E, Alvarez-Saavedra E, Berezikov E, de Bruijn E, Horvitz HR, Kauppinen S, Plasterk RH. 2005. MicroRNA expression in zebrafish embryonic development. Science 309:310-311. Wilt FH. 1967. The control of embryonic hemoglobin synthesis. Adv Morphog 6:89-125. Yamada Y, Warren AJ, Dobson C, Forster A, Pannell R, Rabbitts TH. 1998. The T cell leukemia LIM protein Lmo2 is necessary for adult mouse hematopoiesis. Proc Natl Acad Sci U S A 95:3890-3895. Zeigler BM, Sugiyama D, Chen M, Guo Y, Downs KM, Speck NA. 2006. The allantois and chorion, when isolated before circulation or chorio-allantoic fusion, have hematopoietic potential. Development 133:4183-4192.