SEARCH

SEARCH BY CITATION

Keywords:

  • heart;
  • epicardium;
  • coronary development;
  • endothelial (precursor) cells;
  • migration

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED

The coronary vasculature develops from mesothelial and endothelial precursor cells (EPCs) derived from the proepicardial organ (PEO), which migrate over the heart to form the epicardium. By epithelial-mesenchymal transition (EMT), the subepicardium and epicardium-derived cells (EPDCs) are formed. EPDCs migrate into the myocardium, where they differentiate into smooth muscle cells and fibroblasts that stabilize the developing coronary vasculature and contribute to myocardial architecture. Complete PEO ablation results in embryonic lethality due to cardiac defects, including a looping disorder with a too wide inner curvature. To investigate the behavior of early coronary contributors, we analyzed normal quail embryos and found lumenized endothelial vessels in the subepicardium already at stage HH19. Furthermore, EPCs had penetrated into the myocardium of the inner curvature. To confirm that the myocardium of the inner curvature is specifically permissive for EPCs and to study early EPDC migration in more detail, chimeric chicken embryos harboring a quail PEO were analyzed. Lateral epicardial outgrowth and EMT were observed throughout, but migration into the myocardium was restricted to the inner curvature between HH19 and 22. The permissive myocardial area expanded to the atrium, atrioventricular canal, and trabeculated ventricle at stage HH23–24. In contrast, outflow tract myocardium was never found to be permissive for EPDCs and EPCs until HH30, not even when the quail PEO was attached directly onto it. We conclude that early coronary formation starts in the inner curvature and hypothesize that the presence of PEO-derived cells is essential for the maturation of the inner curvature and subsequent looping of the heart tube. © 2005 Wiley-Liss, Inc.

Heart development starts with the formation of a simple tube consisting of an inner endocardial layer and an outer myocardial layer, with cardiac jelly in between. With increase in size and functioning of the heart, the development of a coronary circulation becomes necessary. Coronary formation is classically divided in primary formation of the endothelial network, during the process of angiogenesis, and in a secondary stabilization of the vessel wall by smooth muscle cells (SMCs) and pericytes during the process of arteriogenesis. Both processes are initiated by the development of the proepicardial organ (PEO) and the epicardium during the late looping phase of the tubular heart. The PEO can be seen as villous protrusions from the ventral wall of the sinus venosus and consists of mesothelial, mesenchymal, and endothelial (precursor) cells (Poelmann et al., 1993). From the proepicardial villi, cells transverse the pericardial cavity along extracellular matrix bridges (Nahirney et al., 2003) and attach to the dorsal side of the developing ventricle (Männer, 1992), where they start spreading over the myocardial surface of the heart tube to form the epicardium (Vrancken Peeters et al., 1995). By epithelial-mesenchymal transition (EMT) of epicardial cells, the mesenchymal subepicardium is formed. Cells from this layer, the so-called epicardium-derived cells (EPDCs), migrate and differentiate into coronary smooth muscle cells and adventitial fibroblasts to stabilize the developing coronary vessels (Vrancken Peeters et al., 1999) and into interstitial fibroblasts to give rigidity to the myocardium (Gittenberger-de Groot et al., 1998). Furthermore, they migrate to the endocardial cushions of the atrioventricular canal (Gittenberger-de Groot et al., 1998). The importance of proper development of the (sub)epicardium and hence of correct EPDC deposition and differentiation is underlined by an increasing number of studies in which epicardial development was manipulated, either mechanically (Gittenberger-de Groot et al., 2000) or genetically (Kwee et al., 1995; Yang et al., 1995; Moore et al., 1999; Tevosian et al., 2000; Lie-Venema et al., 2003). The spectrum of cardiac defects ranges from severe abnormalities resulting in embryonic lethality, e.g., with very thin myocardium, absence of coronary development, and abberant cardiac looping in quail embryos with a completely ablated PEO (Gittenberger-de Groot et al., 2000), to milder disturbances as diminished coronary SMC deposition and abnormal patterning of coronary arteries in antisense Ets-treated chicken embryos (Lie-Venema et al., 2003).

Whereas the migration patterns of the proepicardium and their derivative mesenchymal EPDCs have been studied quite extensively (Wessels and Perez-Pomares, 2004), detailed studies on the developmental behavior of the proepicardial endothelial (precursor) population during early epicardial and coronary development have begun to emerge only recently (Kattan et al., 2004). We therefore analyzed the initial events of coronary endothelial (precursor) cell migration at the onset of coronary development.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED

Normal and Chimeric Embryos

Normal endothelial (precursor) cell behavior was studied in Japanese quail (Coturnix coturnix japonica) embryos. To study EPDC and endothelial (precursor) cell migration in more detail, chimeras were generated using White Leghorn chicken (Gallus domesticus) embryos as hosts and quail embryos as donors. Embryos were staged according to the criteria of Hamburger and Hamilton (1951).

Quail-chicken chimeras were made basically according to the chimerization technique as described by Poelmann et al. (1993). In brief, the proepicardium of an Hamburger and Hamilton stage 15–18 (HH15–18) quail embryo was isolated together with a tiny piece of liver tissue to provide endothelial precursor cells to the proepicardial transplant. The isolated proepicardium and adjacent liver tissue were transplanted into the pericardial cavity of a HH15–18 chicken host embryo through the naturally occurring hiatus in the body wall that exists until HH18. The transplanted tissue was positioned at different sites along the developing heart tube: adjacent to the sinus venosus and atrial region, in the inner curvature or dorsally of the ventricle (Fig. 1). Chimeras were harvested between stage HH19 and HH30. The number and characteristics of successful chimerizations, verified by a specific quail nuclear immunohistochemical staining with QCPN, are listed in Table 1.

thumbnail image

Figure 1. Schematic drawing of the looped heart tube of an HH17 chicken embryo illustrating the sites to which quail PEOs were transplanted in order to generate the quail-chicken chimeras described in Table 1. PEOs of equivalent age were transplanted to either (1) the inner curvature, (2) a site adjacent to the sinus venosus on top of the endogenous PEO, or (3) a site dorsocaudal of the developing ventricle of HH15–17 chicken embryos. A, atrium; AVC, atrioventricular canal; Li, liver; OT, outflow tract; SV, sinus venosus; V, ventricle.

Download figure to PowerPoint

Table 1. Characteristics of quail-chicken chimeras used in this study*
Harvesting stage (HH)Stage at time of transplantation (HH)Transplant siteAttachment sitea
Quail PEOChicken host
  • *

    Only chicken embryos in which the PEO was grafted successfully are listed. For the site of transplantation, see also Figure 1. Donor PEOs inserted next to the developing ventricle (at position 3 in Fig. 1) never attached correctly.

  • a

    A, atrium; AVC, atrioventricular canal; IC, inner curvature; Li, liver; ND, not determined; OT, outflow tract; SV, sinus venosus; V, ventricle.

191615ICA, AVC, OT
201717ICIC, V
201717ICOT
201715ICA, AVC, OT
201616ICAVC, IC, OT
201516ICOT
201516SVOT
20+1716SVOT
211515ICAVC, IC, OT
211618SVA, OT
211616SVIC, OT
211518SVA
211818SVLi
211718SVOT
211516ICAVC, OT
221718ICIC, V
221717ICIC,
22+1516SVA, AVC
231818SVA, OT
241815SVA, OT
24+1617SVV
24+1517SVIC, V
251716SVOT
251515SVOT
251717ICA
251718ICIC
251717SVAVC, IC, OT, Li,
261816SVA, OT
26+1616SVIC, OT
271717SVV
301615ICND

Histological Procedures

Unless indicated otherwise, tissue processing and immunohistochemistry were as described earlier (Vrancken Peeters et al., 1999). Normal and chimeric embryos were fixed by overnight immersion in 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer (pH 7.4) at 4°C. Serial sections were immunostained with the monoclonal antibodies QCPN (Hybridoma Bank, Baltimore, MD; diluted 1:4), QH1 (Hybridoma Bank; diluted 1:500), and HHF35 (Dakopatts M635, Glostrup, Denmark; diluted 1:500) that recognize specifically quail cell nuclear antigen, quail endothelial (precursor) cells, and muscle actins, respectively. For the QCPN staining, antigen unmasking was applied by incubating the sections for 3 × 4 min in 0.01 M citrate buffer (pH 6.0) at 98°C in a microwave. Sections were examined by light microscopy. An Olympus AX-70 light microscope in combination with an Olympus DP-12 digital camera was used for microphotography.

Definitions

QH1-positive cells.

The QH1 antibody recognizes specifically quail endothelial cells and their precursors. As was defined earlier (Poelmann et al., 1993), only QH1-positive cells that line a lumen are considered to be true endothelial cells. Dispersed single or clustered cells located in either mesenchymal or mesothelial regions of the (sub)epicardium or the myocardium are considered as endothelial precursor cells when they do not line a lumen, but have already a relatively flat endothelial cell-like appearance. Isolated round and relatively large QH1-positive cells are considered to be cells of the hematopoietic lineage (hemangioblasts) that can give rise to several blood cell types and to endothelial precursor cells (Choi, 2002). QH1 also recognizes endocardial cells. Thus, it is impossible to discern PEO-derived endothelial (precursor) cells from endocardial cells originating from the primary heart field by staining quail embryonic sections with QH1.

PEO, epicardium, subepicardium, and EPDCs.

The PEO is defined as the cauliflower-like protrusion that sprouts from the sinus venosus, traverses the pericardial cavity, and attaches to the myocardium of the ventricle. The PEO contains a variety of cell and tissue types, namely, endothelial (precursor) cells, capillaries, epithelium, and mesenchyme. Once the PEO has attached to the myocardial surface, the outer mesothelial layer is defined as the epicardium. The subepicardial mesenchymal cells, originating from the epicardium by EMT, form the subepicardium. EPDCs encompass all cells derived from the epicardium, that is, the subepicardial mesenchymal cells and their derivatives (smooth muscle cells and interstitial fibroblasts).

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED

Early Endothelial (Precursor) Cell Migration in Normal Quail Embryos

To study the earliest events in coronary formation, we analyzed PEO-associated endothelial cell migration in normal quail embryos between HH18 and HH27. As can be seen in Figure 2A and B, endothelial precursor cells were present in the mesothelial epicardium and in the mesenchymal subepicardium already at stage HH17+. The close juxtaposition of the liver primordium and the proepicardium, clearly visible in transverse sections of HH17 and HH21 embryos (Fig. 2A and D), suggests that the liver, which is a hemangioblastic organ at this age (Wong and Cavey, 1993), is the primary source of these proepicardial endothelial precursor cells. Migration of the EPCs through the proepicardial organ results in the formation of lumenized endothelial capillaries already at stage HH19 (Fig. 2C and D). These early capillaries form in the subepicardium that is located dorsally of the primitive atrioventricular canal and ventricle, at the initial attachment site of the PEO. In this lumenized vessel, network erythrocytes were found only sporadically in six quail embryos (HH20–22; Fig. 2D). Blood vessels containing notable amounts of erythrocytes were observed for the first time in the subepicardium covering the dorsal side of the ventricle and atrioventricular canal at HH24.

thumbnail image

Figure 2. Early angiogenesis visualized by QH1 staining in the epicardium of quail embryos at stage HH17 (transverse sections in A and B), HH19 (sagittal section in C), and HH21 (transverse section in D). At HH17+, endothelial (precursor) cells, indicated by arrow in A and B, are already present in the proepicardial organ. Given the close proximity of the liver (Li) and the PEO (A, B, and D), these endothelial cells are most likely derived from the hematopoietic liver tissue. A few endothelial precursor cells have migrated into the myocardium of the inner curvature (IC; encircled in C). Lumenized endothelial vessels are observed from HH19 onward (arrowheads in C and D). An arrow indicates a lumenized vessel filled with erythrocytes in a transverse section of an HH21 quail embryo in D. An endothelial capillary network, probably continuous with the developing liver sinusoids, is present in the subepicardium at the site where the PEO attaches to the myocardium. The QH1-positive cells in the myocardium (asterisks) are endocardial cells lining the trabecular lumina. Graft-derived endothelial precursor cells do not enter the myocardium in this region. Scale bars = 100 μm (A, C, and D); 50 μm (B)

Download figure to PowerPoint

The earliest myocardial invasion by endothelial precursor cells was observed from HH20 onward, although not in the region of this early vessel network, but in the inner curvature (Fig. 3). As was deduced from the position of these endothelial precursor cells in serial sections, they were derived from or connected to the endothelium of the more caudally located subepicardial capillaries. Because the QH1 antibody stains endocardial cells equally well, it appeared to be difficult to discern PEO-associated endothelial precursor cells in the myocardium from the endocardium-lined ends of the trabecular lumina.

thumbnail image

Figure 3. Series of transverse sections of a stage HH20+ quail heart stained with QH1 to demonstrate endothelial (precursor) cells (AC) and with HHF35 to delineate the myocardium (D, consecutive section to B). Endothelial precursor cells invading the myocardium seen only at the site of the inner curvature (arrowheads) are shown in B. Endothelial precursor cells in the subepicardium surrounding the outflow tract and atrioventricular canal do not enter the myocardium (arrows in A and C). A, atrium; AVCu, atrioventricular cushion; OT, outflow tract. Scale bars = 100 μm.

Download figure to PowerPoint

Quail-Chicken Chimeras

To circumvent the difficulty that endocardial cells closely resemble endothelial (precursor) cells in an immunohistochemical QH1 staining, chimeric embryos were generated, in which the PEO of a quail embryo was transplanted to the pericardial cavity of a chicken embryo of equivalent age. In the resulting quail-chicken chimeras, QH1 could react only with the PEO-associated endothelial cells, leaving the chicken-derived endocardial cells unstained. A developmental series of quail-chicken chimeras (HH19–30) was made, in which the donor PEO was placed at several sites along the developing heart tube at HH16–17 (Fig. 1, Table 1), just at the time point when proepicardial cells start to transverse the pericardial cavity to reach the heart. The chicken host PEO was fully preserved. Thus, transplantaton did not interfere with the endogenous outgrowth of the chicken PEO and the heart could develop normally (Poelmann et al., 1993). This could also be deduced from the fact that the embryos had normal heart morphology.

With respect to the site of transplantation, we observed that PEO attachment to the heart and epicardial outgrowth succeeded in 31 out of 56 surviving embryos when the PEO was placed adjacent to the sinus venosus or in the inner curvature of the looping heart. In contrast, PEOs transplanted dorsocaudally of the ventricular myocardium either did not attach in 15 out of 25 cases or attached to the liver in the remaining 10 embryos. This may be indicative for a limited capacity of the ventricular myocardium to attract PEO-derived cells to its surface.

The site and extent of quail epicardial outgrowth in the embryos where the PEO had attached well have been summarized in Table 2. The presence of a subepicardial mesenchymal layer was a prerequisite for the ingrowth of PEO-derived cells into the underlying myocardium. More surprisingly, we observed that only the myocardium of the inner curvature was permissive for endothelial precursor cells and mesenchymal EPDCs between stage HH19 and HH22 (Fig. 4). The myocardium of the atrioventricular canal and that of the outflow tract were not invaded by PEO-derived cells, not even when a multicellular layer of subepicardial mesenchyme was present. The myocardium of the inner curvature is not covered with cushion tissue at its luminal side. However, PEO-derived cell invasion of the myocardium was not directly related to the absence of cushion tissue, since it was also observed in cushion-bearing myocardium of the atrioventricular canal at later stages of development (not shown). From HH23 onward, PEO-derived cells were also observed in the myocardium of the atria (Fig. 5A–C) and in the ventricular myocardium (Fig. 5D and E). The myocardium of the outflow tract was not permissive for EPDCs, at least untill HH26. Only at HH30 were EPDCs observed in the outflow tract myocardium up to its most distal border (Fig. 5H–J).

Table 2. Overview of sites of EMT and PEO-derived cell migration into the myocardium in the quail-chicken chimeras used in this study*
Harvesting stage (HH)SVAAVCVICOT
  • *

    The presence of graft-derived subepicardial mesenchyme (1) and endothelial (precursor) cell and EPDC migration (2) into the myocardium is indicated. Although a multicellular layer of subepicardial mesenchyme could be present at several sites along the heart tube, before stage 23 ingrowth of PEO-derived cells occurred exclusively in the inner curvature. A, atrium; AVC, atrioventricular canal; IC, inner curvature; OT, outflow tract; SV, sinus venosus; V, ventricle

19 11 1,21
20   11,2 
20     1
20 11 1,21
20  1 1,21
20  1 1,21
20     1
20+      
21 1111,21
21 11 1,21
21    1,21
2111    
21   1  
21     1
21  111,21
22   11,2 
22  1 1,2 
23 11,2 1,2 
23 1,2   1
2411,21,21,21,21
24   1,21,21
24  1,21,21,2 
25     1
25     1
25 1,2   1
25 1,2  1,21
25   1,21,21
26  1,21,21,21
26+  1,21,21,21
27  1,21,21,21
30 1,21,21,21,21,2
thumbnail image

Figure 4. Serial sections of a stage HH19 (AF) and HH21 (G and H) quail-chicken chimeric embryos in which a quail PEO (HH16) was transplanted to the inner curvature of a chicken host (HH15). Sections were stained with QH1 for quail endothelial (precursor) cells (A, D, and G), with QCPN for quail-derived cells (B, E, and H), and with HHF35 to demarcate the myocardium (C and F). Note that PEO-derived cells do not migrate into the myocardium of the atrioventricular canal (AVC) and that of the outflow tract (OT) in the upper panel. The lower panel depicts serial sections 35 μm more caudally and shows that only the myocardium of the inner curvature (IC) is permissive for endothelial precursor cells (arrows). QH1 and QCPN staining (arrowheads) mostly colocalize, indicating that the first cells to enter the inner curvature are endothelial precursor cells rather than mesenchymal EPDCs. However, the fact that not only endothelial (precursor) cells invade the myocardium at these early stages of development is shown in G and H, which depict serial sections of the inner curvature of the HH21 chimeric embryo that received a quail PEO graft with the most minimal piece of liver. Arrows and closed arrowheads indicate endothelial (precursor) cells; open arrowheads indicate mesenchymal EPDCs. Scale bars = 100 μm.

Download figure to PowerPoint

thumbnail image

Figure 5. Serial transverse sections of the hearts of stage HH23 (AC), stage HH26 (D and E), and stage HH30 (FJ) quail-chicken chimeras, stained with QH1 (A, F, and H), QCPN (B, D, E, G, and I), and HHF35 (C and J). Arrows indicate the presence of PEO-derived cells in the atrial myocardium at HH23. The outflow tract myocardium is resistant to invasion by endothelial precursor cells and EPDCs. At HH26, the trabecular myocardium contains EPDCs. This is clearly visible in regions with a well-developed quail-derived subepicardium. Serial sections of the right ventricular wall of an HH30 chimera show that endothelial PEO-derived cells invade the compact myocardium (arrows in F), whereas mesenchymal EPDCs reach also the trabecular myocardium and subendocardial space (arrowheads in G). Only at HH30 were quail-derived cells penetrating the outflow tract myocardium observed (H–J). The inset shows that photomicrographs H–J concern the most distal outflow region, at the myocardial border. A, atrium; OT, outflow tract; V, ventricle. Scale bars = 100 μm.

Download figure to PowerPoint

From our set of quail-chicken chimeras, we could not precisely infer whether the endothelial precursor cells and EPDCs reached their intramyocardial locations only via the epicardium on the outside, or also by spreading within the myocardium. Only sporadically quail-derived EPDCs were observed in the myocardium underneath a layer of chicken-derived epicardium, and most quail EPDCs were deposited in the myocardium beneath a quail-derived QCPN-positive epicardial region. Figure 5D and E illustrate how an epicardium rich in quail cells delivers many quail EPDCs to the underlying myocardium, whereas an epicardium with less quail cells covers a myocardium in which less QCPN-positive EPDCs are present. Thus, endothelial precursor cell and EPDC migration from the epicardial to the myocardial layer seems to be a quite local process. Along the heart tube, the endothelial QH1 staining in the myocardium colocalized with the nuclear signal of the QCPN-positive quail cells in a large majority (95%) of the embryos, indicating that the entrance of endothelial precursor cells into the myocardium might precede that of mesenchymal EPDCs. EPDCs penetrating beyond the compact myocardium into the trabecular myocardium and subendocardial space were mostly found underneath those regions of the compact myocardium that were supplied well with endothelial (precursor) cells (Fig. 5F and G).

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED

In the present study, we analyzed the behavior of early coronary contributors both in normal quail embryos and in proepicardial quail-chicken chimeras. Many aspects of the development of the coronary vasculature have been studied earlier by our group and others (Luttun and Carmeliet, 2003). For example, the intimate topographical relationship of liver primordium and the sinus venosus was documented by Virágh et al. (1993). The liver primordium is highly vascularized with developing capillaries, which, together with the perihepatic mesenchyme, enter the wall of the sinus venosus. It has been further observed that the caudal part of the splanchnic vessel plexus becomes part of the liver sinusoids, and that the more cranial part extends in the direction of the proepicardial organ (Vrancken Peeters et al., 1997). Thus, the endocardial cells that make up the capillaries developing in the proepicardial strands can be considered to be continuous with the endothelial cells of the liver sinusoids. Further evidence for a role of liver-derived endothelial (precursor) cells in the development of the coronary vasculature comes from earlier chimera experiments, in which the quail PEO was excised very precisely and transplanted to the pericardial cavity of a chicken host without any liver tissue. In the resulting embryos, quail-derived endothelial cells were never observed (Poelmann et al., 1993). Although the precise origin of the endothelial (precursor) cells in the PEO is still a matter of debate (Kattan et al., 2004), indirect proof of endothelial (precursor) cell migration through the PEO has been provided recently by the finding that larger numbers of QH1-positive cells were observed at the base of the PEO than in its villous tips in HH17 quail embryos (Kattan et al., 2004).

In earlier studies, true vascularization in the subepicardium, defined by the presence of lumenized capillaries, was described not to occur before HH25 (Poelmann et al., 1993; Virágh et al., 1993; Vrancken Peeters et al., 1997). We now show that lumenized endothelial vessels, in tissue sections sometimes observed with an erythrocyte inside, can be found in the subepicardium already at stage HH19. Because some of these were localized close to the developing liver, they may well have been arisen by sprouting from the sinus venosus-liver plexus. Thus, coronary formation at the dorsal side of the developing heart, at the attachment site of the outgrowing PEO, would be initiated by angiogenesis, rather than by vasculogenesis. On the other hand, as was described both in earlier and in more recent publications, large hemangioblastic cells were observed in the epicardium, the subepicardium, and on the naked heart surface (Vrancken Peeters et al., 1997; Kattan et al., 2004). These are indicative for the de novo vasculogenesis that has been shown to be instrumental in coronary development (Mikawa and Fischman, 1992).

In the young quail-chicken chimeras (HH19–23), only in the inner curvature were quail-derived EPDCs able to migrate further than the subepicardium and to find a permissive myocardium to migrate into. PEO transplants located next to the outflow tract or close to the developing ventricle failed to deposit coronary precursor cells into the myocardial layer of the AV canal, the atria, and the ventricles until HH23/24. To our knowledge, this is the first in vivo study demonstrating that the invasion of coronary contributors into the myocardium and thus the onset of myocardial vascularization depend on characteristics of the underlying myocardium. In the avian embryo, the PEO protrudes from the dorsal mesothelium to the dorsal side of the developing common ventricle, from where it begins to ensheath the myocardium in the region of the atrioventricular canal. Instead of entering the myocardium in this region, the PEO-associated endothelial cells travel all the way to the inner curvature to initiate the formation of intramyocardial coronary capillaries. In our quail-chicken chimera experiments, quail endothelial precursor cell and EPDC ingrowth started exclusively in the myocardium of the inner curvature, whereas outflow tract myocardium was refractory to this, even when the quail PEO was attached directly onto it. This also confirms the notion that the myocardial microenvironment is an important determinant for endothelial precursor cell and EPDC migration and coronary vessel formation (Morabito et al., 2001, 2002). As for the nature of the myocardial signals, we can only speculate. A marked anatomical difference between the myocardium of the inner curvature compared with that of the atrioventricular canal and the outflow tract is that there is no endocardial cushion present on the inside. Possibly, angiogenic regulators (inhibitors) secreted by the endocardium or cushion mesenchyme induce the myocardium to be nonpermissive for endothelial cells coming from the other side in the AV canal and outflow tract. Good candidates to function as such antiangiogenic regulators are TGF-β2 and TGF-β3. In vitro analysis showed that TGF-β1, -2, and -3 inhibit the myocardial signals that regulate epicardial EMT (Morabito et al., 2001). In chicken embryos, both TGF-β2 and -3 are expressed in the myocardium and endocardium of the AV canal and outflow tract during cushion formation around stage HH17 (Boyer et al., 1999), just preceding the endothelial (precursor) ingrowth of the inner curvature. In mouse embryos, strong TGF-β2 expression is seen in the outflow tract and atrioventricular cushions and underlying myocardium between embryonic day (ED) 9.5 and ED12.5, comparable to chicken stage HH20–25. At ED11.5–12.5, epicardial TGF-β2 and -3 expression increases, especially in the epicardium covering the ventral side of the developing ventricles (Molin et al., 2003). This might explain why the subepicardium is relatively thin at this location. After ED12.5, TGF-β2 expression remains high in the cushion mesenchyme of the outflow tract and AV canal. On the proangiogenic side, VEGF becomes upregulated in the atrioventricular field of the heart from ED10.5 onward (Dor et al., 2001). VEGF expression stimulates endothelial proliferation and migration (Carmeliet, 2000) and has been shown to activate the endothelial proteolytic machinery (Prager et al., 2004). Thus, the AV region might become prone to invasion by epicardially derived endothelial and mesenchymal cells, whereas the outflow tract myocardium remains inaccessible until the VEGF signaling routes can be fully employed. In quail embryos, increased endothelial VEGF-R2 and VEGF-R3 expression coincides with the ingrowth of coronary capillaries at the base of the truncus arteriosus to form the coronary plexus that precedes the formation of the coronary stems (Tomanek et al., 2002).

The fact that we did not find myocardial endothelial (precursor) cells and EPDCs of quail origin in regions covered by a chicken epicardium indicates that myocardial entry of these cells is indeed a very local process, as was already suggested by earlier findings in older chimeric embryos (Gittenberger-de Groot et al., 1998). The local delivery of coronary components only to the underlying myocardium could also be seen in stage HH35 quail embryos in which epicardial outgrowth was partially inhibited. Only the ventricular regions covered with epicardium acquired a compact myocardium with regularly arranged coronary vessels. Naked compact ventricular myocardium was completely devoid of these vessels, indicating that the local presence of coronary precursor cells is a prerequisite for coronary vessel formation (data not shown).

In the set chimeras that we analyzed, the endothelial (precursor) cells seemed to be the first to enter the myocardium. At later stages, more intramyocardial mesenchymal EPDCs were observed. Only in the ventricular myocardium, where epicardium-derived fibroblasts also contribute to myocardial rigidity, did the EPDCs outnumber the endothelial (precursor) cells from stage HH25/26 onward. However, in myocardial regions where coronary artery stabilization is one of the EPDC functions, many of the proepicardium-derived cells are endothelial cells, which seem to be followed by the mesenchymal EPDCs. It is tempting to speculate that, probably via PDGF-B signaling (Nishishita and Lin, 2004), the endothelial (precursor) cells pave the way for the entrance of the mesenchymal cells that will assist in the formation of stable vessels. Further support for the idea that penetration by endothelial (precursor) cells may facilitate mesenchymal EPDC migration comes from our observation in older chimeras that only in regions where the compact myocardium was supplied with numerous quail endothelial cells did the underlying trabecular myocardium contain many quail-derived EPDCs.

Finally, between HH20 and HH24, the inner curvature is not only the site where coronary formation starts, it is also a region that is very much involved in the last phase of the cardiac looping process. During this phase, the proximal two-thirds of the outflow tract or primitive conus shifts from the right lateral position to its final position, ventral to the right atrium (Männer, 2000). In quail embryos in which proepicardial outgrowth was inhibited, cardiac looping was severely disturbed (Gittenberger-de Groot et al., 2000). Although we did not directly address the issue, the present study might give a first concrete clue for a role of the epicardium in the looping process. We hypothesize that the selective ingrowth of proepicardially derived endothelial cells into the inner curvature is the starting signal for the myocardium to rearrange its fibrils and for the conus to shift leftward.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED

The authors gratefully acknowledge Jan H. Lens and Bas Blankevoort for excellent photographic assistance and digital artwork. They also thank Jaime Martodihardjo for the initial characterization of the chimeric embryos and Daniël G.M. Molin for critical reading of the manuscript.

LITERATURE CITED

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED