Immunolocalization of the Vascular Endothelial Growth Factor ...

Histochemical Journal 30, 627±634 (1998)

Immunolocalization of the vascular endothelial growth factor receptor-2 in the subepicardial mesenchyme of hamster embryos: identi®cation of the coronary vessel precursors J O S E M A R I A P E R E Z - P O M A R E S , D A V I D M A C I A S , L I N A G A R C I A - G A R R I D O and  N M UN  PULI Ä OZ-CHA RAMO Department of Animal Biology, Faculty of Science, University of MaÂlaga, 29071 MaÂlaga, Spain Received 8 December 1997 and in revised form 21 April 1998

Summary The earliest evidence of the development of the cardiac vessels in mammals is the emergence of subepicardial blood islands, which are thought to originate from mesenchymal progenitors. In order to identify these progenitor cells, we have studied the immunohistochemical localization in the heart of Syrian hamster embryos of the type 2 vascular endothelial growth factor receptor, the earliest molecule known to be expressed in the vasculogenic cell lineage. Only a few immunoreactive subepicardial mesenchymal cells were present by 10 days post coitum. By 11 days post coitum, the subepicardial mesenchymal cells became abundant at the dorsal part of the ventricle, the atrioventricular and the conoventricular grooves. About 20% of cells were labelled with the antibody. Immunoreactive cells were isolated or formed pairs, short cords, rounded clusters or ring-like structures at the subepicardium or, occasionally, within the ventricular myocardium. Other labelled cells were simultaneously cytokeratin immunoreactive. By 12 days post coitum, most immunoreactive mesenchymal cells have been replaced by a capillary network. We propose that an active process of vascular differentiation occurs between 10 and 12 days post coitum in the subepicardium of this species, and it might be a suitable model for the study of vasculogenetic mechanisms. # 1998 Chapman & Hall

Introduction The current theory of vertebrate embryo vascularization states that vasculogenesis (i.e. de novo assembly of vessels from mesenchymal progenitors) predominates in the endodermal territories and closely associated mesoderm. However, in ectodermal derivatives and their associated mesoderm, angiogenesis (i.e. growth of already differentiated vessels) is the main mechanism of vascularization (Risau & Flamme, 1995). It is uncertain to what extent the cardiac vascularization ®ts with this model. Although the endoderm plays an essential role in early cardiogenesis, the differentiation of the primitive cardiac capillary plexus occurs when the heart has moved away from the endoderm. However, morphological and experimental evidence, obtained mainly from avian embryos, suggests that the vascular plexus of  To whom correspondence should be sent.

0018±2214 # 1998 Chapman & Hall

the heart forms through self-assembly of mesenchymal precursors, i.e. through vasculogenesis (Viragh & Challice, 1981; Tokuyasu, 1985; Icardo et al., 1990; Bolender et al., 1990; Viragh et al., 1990; Mikawa & Fischman, 1992; MunÄoz-ChaÂpuli et al., 1996). The existence of an endodermal induction on cardiac vasculogenesis could be maintained under the assumption that the cardiac angioblasts migrate into the heart from the liver region (Poelmann et al., 1993; Tomanek, 1996). In the mammalian embryo, the earliest evidence of cardiac vascularization is the emergence of blood island-like structures, mainly located at the intercameral grooves (Hirakow, 1983, 1992; Hutchins et al., 1988). The forerunners of these structures have not yet been identi®ed, and no information is available about their features, abundance or localization. For these reasons, we aimed to identify the putative cardiac vessel progenitors by using the

628 earliest known marker of the cells committed in the vascular and haematopoietic lineage, the type 2 receptor of the vascular endothelial growth factor (VEGFR-2), a transmembrane tyrosine kinase receptor also known as foetal liver kinase-1 in mouse and KDR in humans (Millauer et al., 1993; Quinn et al., 1993; Yamaguchi et al., 1993; Shalaby et al., 1995; Risau & Flamme, 1995). The ®rst aim of this paper was to localize, through immunohistochemistry, the VEGFR-2 in the subepicardial mesenchymal cells of embryos of Syrian hamster, as well as to describe the earliest stages of cardiac vasculogenesis in this species. Our second objective was to check if collocalization of VEGFR-2 and cytokeratin occurs in the subepicardial mesenchymal cells, as we have shown elsewhere that most subepicardial mesenchymal cells in the hamster and chick embryos are characterized by a transient cytokeratin immunoreactivity, and have suggested that it might be caused by the origin of these cells through delamination from the epicardial mesothelium (PeÂrez-Pomares et al., 1997). Materials and methods The investigation conformed with the Guide for the Care and Use of Laboratory Animals published by the US National Institute of Health (NIH publication no. 8523, revised 1985). The hamster sample studied consisted of 27 embryos of Syrian hamster (Mesocricetus auratus), the animal model currently used in our studies on cardiac development. Mature male and female hamsters were mated, and the end of the coitus was considered as day 0 of the embryonic development. Pregnant females were killed by chloroform overdose, and the embryos were obtained by laparotomy and uterotomy, freed from foetal membranes and ®xed. The embryos were collected at stages 10 (5), 10.3 (2), 10.5 (1), 11 (15) and 12 (4) days post coitum. These stages were selected according to our previous experience on the early stages of cardiac vascularization in this species (PeÂrez-Pomares et al., 1997). The embryos were ®xed in 40:40:20 by volume methanol±acetone±distilled water for 8±12 h. After ®xation, the embryos were dehydrated in an ethanolic series ®nishing in n-butanol, and paraf®n embedded. Serial sections (5 ìm and 10 ìm) were obtained with a Leitz microtome and collected on slides coated with poly-Llysine. The sections were dewaxed in xylene, hydrated in an ethanolic series and washed in Tris±phosphate-buffered saline (TPBS), pH 7.8. The further treatment of the slides varied according to the technique used, as described below.

Peroxidase immunohistochemistry Endogenous peroxidase activity was quenched by incubation for 30 min with 3% hydrogen peroxide in TPBS, and endogenous biotin was blocked with the avidin±biotin

PEÂREZ-POMARES et al. blocking kit (Vector). After washing with TPBS, nonspeci®c binding sites were saturated for 30 min with 16% sheep serum, 1% bovine serum albumin in TPBS (blocking buffer). The slides were then incubated overnight at 48C in the primary antibody diluted in blocking buffer containing 0.8% sodium dodecyl sulphate and 0.2% Tween-20. This detergent treatment, although detrimental to the quality of the tissue, was mandatory for performing the immunolabelling. Control slides were incubated in diluent only or with rabbit serum diluted 1:200. After incubation, the slides were washed in TPBS containing 0.4% sodium dodecyl sulphate and 0.1% Tween-20, incubated for 1 h at room temperature in biotin-conjugated anti-rabbit goat IgG (Sigma) diluted 1:100 in blocking buffer, washed again and incubated for 1 h in avidin±peroxidase complex (Sigma) diluted 1:150 in TPBS. Peroxidase activity was developed with Sigma Fast 3,39-diaminobenzidine tablets according to the instructions of the supplier. The polyclonal anti-VEGFR-2 antibody (a gift from Dr A. Schuh, University of Toronto) was raised in a rabbit immunized against the mouse VEGFR-2 carboxyl tail obtained in bacteria as a recombinant fusion protein. Total IgG was isolated from the antiserum and immunoaf®nity puri®ed. The IgG obtained was used at a concentration of 6 ìg mlÿ1 . This antibody has been used in immunohistochemistry to show the existence of differentiated endothelial cells unable to form normal vessels in VEGFR-1de®cient mice (Fong et al., 1995). The antibody recognizes a single band of about 200 kDa in Western blots of mouse embryonic extracts, but no band is detected in extracts of VEGFR-2 ÿ=ÿ embryos or VEGFR-2 ÿ=ÿ embryonic stem cells differentiated in vitro, although they produce VEGFR-1 (A. Schuh, personal communication).

Fluorescence immunohistochemistry and double labelling For immuno¯uorescence, the hamster sections were blocked as described above, incubated overnight in antiVEGFR-2 diluted at 12 ìg mlÿ1 in blocking buffer containing 0.8% sodium dodecyl sulphate and 0.2% Tween-20, washed, incubated for 1 h in biotin-conjugated goat antirabbit IgG (Sigma) diluted 1:100 in blocking buffer, washed and incubated in extravidin±¯uorescein conjugate (Sigma) diluted 1:100 in TPBS. After washing, the sections were mounted in 1:1 v=v buffer±glycerol containing 0.2% 1,4diazabicyclo[2.2.2] octane. For double labelling, the sections stained as described above were blocked again in blocking buffer with 0.1% Triton X-100 and incubated for 2 h with monoclonal antipan cytokeratin (C2562; Sigma, UK), a mix of six monoclonal antibodies that stains most types of keratin in the epithelial cells of several vertebrates, but shows no cross-reaction with non-epithelial normal human tissues. It was used at 1:60 dilution. The washed sections were then incubated for 45 min in rhodamine-conjugated goat anti-mouse IgG (Sigma) diluted 1:50. Controls were incubated only with one primary antibody and then with both secondary antibodies, in order to detect any crossreaction between the secondary and the primary antibodies. The sections were observed in a Nikon Microphot FXA

Immunohistochemical identi®cation of coronary vessel precursors equipped with epi¯uorescence and in a laser confocal microscope Leica TCS-NT, using ®lters speci®c for the ¯uorescein and rhodamine ¯uorochromes.

Results The anti-VEGFR-2 antibody speci®cally labelled all the endothelial cells of the hamster embryos studied, including the endocardial cells. Many large and rounded liver cells, probably haematopoietic, were also labelled, as well as a number of mesenchymal cells mainly located in the splanchnic mesoderm of the gut and the peripheral areas of the pulmonary primordia (Fig. 1). The intensity of the labelling was variable, from faint to well-marked, but distinctive owing to the frequent presence of a perinuclear spot. In the heart of the Syrian hamster embryos,

Fig. 1. Immunostaining of the endothelial cells and the putative endothelial progenitors with the anti-vascular endothelial growth factor receptor-2 polyclonal antibody. Hamster embryo of 11 days post coitum. Transverse section. The antibody speci®cally stains the endothelial cells as well as a number of non-endothelial cells from the liver (LI), which are presumably haematopoietic, and from the cortical areas (arrowheads) of the pulmonary primordia (P). Note the lack of stained cells in the avascular vertebral chondri®cation (VT) and the invasion of the spinal cord (SC) by vessels arising from the perineural vascular plexus (PNP). Scale bar ˆ 100 ìm.

629

VEGFR-2-immunoreactive subepicardial mesenchymal cells were observed for the ®rst time by 10±10.3 days post coitum, about 12 h after the earliest subepicardial mesenchymal cells had appeared. The immunoreactive cells were mainly located at the posterior part of the sinoatrial and atrioventricular grooves. The VEGFR-2-immunoreactive subepicardial mesenchymal cells were isolated or formed open or closed rings (Fig. 2a). Subepicardial blood-containing vessels were scarce at these embryonic stages and restricted to the posterior part of the atrioventricular groove. Mesenchymal cells were abundant at the subepicardium of the 11 days post-coitum embryos, particularly at the atrioventricular groove, dorsal part of the ventricle and proximal out¯ow tract. Approximately 20% of the mesenchymal cells were VEGFR-2 immunoreactive. These cells were very often associated in pairs, short cords, rounded clusters or ring-like structures (Figs 2 and 3), which were isolated from differentiated vessels or from other similar cell assemblages, as demonstrated by the study of serial sections and confocal series. The antibody also labelled the endothelium of the bloodcontaining vessels, which were mainly located at the subepicardium of the atrioventricular groove (Fig. 2b, c and e). Immunoreactive cells were observed relatively frequently in apparent contact with the outer surface of these differentiated vessels (Fig. 2a, b and e). Other immunoreactive mesenchymal cells were observed in the intramyocardial domains of the ventricle, mainly at the proximity of the atrioventricular groove, where the myocardium seemed to have loosened (Fig. 2c±e). VEGFR-2immunoreactive subepicardial mesenchymal cells were not detected in the atrium, except in its ventralmost area close to the atrioventricular groove (Fig. 2e). The controls revealed a non-speci®c background staining in the myocardium and, sometimes, in the outer surface of the primitive epicardium, but not in the subepicardial mesenchymal cells (Fig. 2f). By 12 days post coitum, the number of immunoreactive mesenchymal cells had decreased markedly. However, a plexus of subepicardial blood-containing vessels was present in the areas where the immunoreactive mesenchymal cells were most abundant, i.e. the atrioventricular, conoventricular and interventricular grooves as well as the dorsal part of the ventricle. A few subepicardial mesenchymal cells were simultaneously cytokeratin and VEGFR-2 immunoreactive in 11 days post-coitum embryos (Figs 3 and 4). The cytokeratin staining of the double-labelled cells was usually weaker than that of the epicardial cells or the VEGFR-2ÿ subepicardial mesenchymal cells, but it was clearly stronger than the background staining. Some double-labelled cells were

630 associated in clusters or rings (Figs 3 and 4). Others were large, rounded and isolated, similar to the haematopoietic cells of the liver, which were also double labelled (Fig. 4). The contact of the doublelabelled cells with prolongations of cytokeratin‡ =VEGFR-2ÿ cells was frequent (Fig. 3). Collocalization of these antigens was not obtained by 12 days post coitum, probably because of the marked decrease in the number of both cytokeratin- and VEGFR-2-immunoreactive subepicardial mesenchymal cells. The control sections showed that the double labelling cannot be attributed to crossreactions of the secondary with the primary antibodies. Discussion We have carried out a study aimed at identifying the forerunners of the primary cardiac vascular plexus in a mammalian species. Although the coronary vessels were regarded for a long time as sprouts of the aortic or cardiac endothelia, recent reports have provided evidence that the development of this plexus occurs through vasculogenesis (i.e. commitment of mesodermal cells in the endothelial lineage; Risau et al., 1988; Noden, 1991; Poole & Cof®n, 1991; Risau, 1991; Risau & Flamme, 1995). This evidence is mainly morphological (Viragh & Challice, 1981; Tokuyasu, 1985; Icardo et al., 1990; MunÄoz-ChaÂpuli et al., 1996), but also experimental. Mikawa and Fischman (1992) and Mikawa and Gourdie (1996) have demonstrated, through retroviral labelling and direct cell lineage

PEÂREZ-POMARES et al. analysis, that the coronary arteries of the chick embryo form by coalescence of discontinuous cell clones (i.e. by in situ vasculogenesis). Knowledge about the morphogenesis of the coronary vessels may help us to understand the basic mechanisms underlying the development of a number of coronary vascular anomalies of clinical relevance (Poelmann et al., 1993). On the other hand, the existence of a process of vasculogenesis in a welldemarcated embryonic environment such as the subepicardium might provide a relevant model for the study of the basic mechanisms of endothelial differentiation. As stated above, the evidence for a primary vasculogenic process in the mammalian heart is largely morphological, as markers of cells with vasculogenic potential have not become available until recently. We have shown that, in the hamster embryo, a wave of mesenchymal cells expressing the tyrosine kinase VEGFR-2 appears between 10 and 11 days post coitum at the subepicardium of the atrioventricular and conoventricular grooves as well as in the dorsal part of the ventricle. Apparently, in only 24 h, these immunoreactive cells have coalesced into a network of bloodcontaining vessels, which extends around the atrioventricular and conoventricular grooves and throughout the dorsal surface of the ventricle. The VEGFR-2-immunoreactive subepicardial mesenchymal cells were frequently associated in cords, rings and clusters, which could be regarded as precursors of differentiated vessels or, in some cases, immature

Fig. 2. Immunostaining of the heart of Syrian hamster embryos with the anti-vascular endothelial growth factor receptor2 antibody. (a) Embryo 10 days post coitum. Frontal section. Subepicardium of the most posterior area of the atrioventricular groove, the cardiac region where the earliest blood vessels can be seen. The antibody labels the endothelial cells of two blood vessels (BV), a ring-like structure (large arrow) and an isolated cell (small arrow) in contact with the atrial myocardium (M). Other immunoreactive mesenchymal cells are in contact with the outer surface of a blood vessel (arrowheads). EP, epicardial mesothelium. Scale bar ˆ 12:5 ìm. (b) Embryo 11 days post coitum. Transverse section. Subepicardium of the atrioventricular groove. The immunostaining reveals a connection between a blood vessel (star) and a tube-like structure (arrow). The endocardium (EN) is also immunoreactive. A, atrium; EP, epicardial mesothelium; V, ventricle. Scale bar ˆ 12 ìm. (c) Embryo 11 days post coitum. Transverse section. Atrioventricular groove. This ®gure shows different arrangements of the immunoreactive cells such as isolated (arrowheads), forming cords (large arrow) or clusters (little arrow), either in the subepicardium (SE) or within the ventricular myocardium (V). Two blood-containing vessels with labelled endothelium are present in the subepicardium (stars). EP, epicardial mesothelium; MC, unlabelled mesenchymal cells. Scale bar ˆ 10 ìm. (d) Embryo 11 days post coitum. Transverse section. Atrioventricular groove. A group of labelled rounded cells is arranged in a compact cluster (arrow). An isolated VEGFR-2-immunoreactive cell (arrowhead) can also be seen inside the ventricular myocardium (V). Most subepicardial mesenchymal cells (MC) are unlabelled. A, atrium; EP, epicardial mesothelium. Scale bar ˆ 9 ìm. (e) Embryo 11 days post coitum. Transverse section. Two large vessels (stars) have developed in this embryo at the level of the atrioventricular groove. Isolated stained cells can be seen in the subepicardium of the groove and within the ventricular myocardium (arrow and little arrowhead respectively). The presence of stained cells in the subepicardium of the atrium (A), such as that shown by a large arrowhead, is rare and restricted to the areas closest to the atrioventricular groove. EP, epicardial mesothelium; EN, endocardium. Scale bar ˆ 20 ìm. (f) Embryo 11 days post coitum. Atrioventricular groove. Control section incubated with rabbit serum diluted 1:200 instead of the primary antibody. The blood vessel (star) and the mesenchymal cells (MC) are not stained. The staining of the myocardium (M) is nonspeci®c. The epicardial mesothelium, negative in this section, occasionally showed a non-speci®c staining of its apical surface. Scale bar ˆ 30 ìm.

Immunohistochemical identi®cation of coronary vessel precursors

631

632

PEÂREZ-POMARES et al.

Fig. 3. Embryo 11 days post coitum. Transverse section. Subepicardium of the atrioventricular groove. Laser confocal microscopy. (a) VEGFR-2 immunoreactivity. A ring-like structure (asterisk), formed by several cells and probably a precursor of a capillary vessel, is conspicuously labelled, as well as some isolated cells (white arrows). Other labelled cells are forming a cord (yellow arrows). (b) Cytokeratin immunoreactivity. Note the strong labelling of the epicardial mesothelium (EP) and most mesenchymal cells (MC). The cells shown by the white arrows are double labelled. The capillary-like structure shows a faint cytokeratin immunoreactivity, although greater than the background staining. The contact between a cytokeratin-immunoreactive mesenchymal cell and this structure is highlighted by an arrowhead. M, ventricular myocardium. Scale bar ˆ 10 ìm. Fig. 4. Embryo 11 days post coitum. Transverse section. Subepicardium of the dorsal part of the ventricle. Laser confocal microscopy. (a) VEGFR-2 immunoreactivity. The staining reveals ring-like structures (white and yellow arrows) formed by several cells, located in the subepicardium as well as within the ventricular myocardium (M). Many cells in the liver (L), probably haemopoietic cells, are also VEGFR-2 immunoreactive. (b) Cytokeratin immunoreactivity. The subepicardial ring-like structure (white arrow) is clearly double labelled, as well as the liver cell indicated by an arrowhead and, to a lesser degree, the presumptive haemopoietic cells. Prolongations of cytokeratin-immunoreactive mesenchymal cells are in contact with the double-labelled, subepicardial structure. EP, epicardial mesothelium; SM, splanchnic mesothelium. Scale bar ˆ 10 ìm.

angiocysts. Other isolated immunoreactive cells were found within the ventricular myocardium, an observation that could be related to the intramyocardial arrangement of the adult coronary arteries in this species. This ®nding suggests that the intramyocardial vascular plexus might originate not only by vascular sprouting and ingrowth from the

subepicardial network, but also through migration of subepicardial angioblasts. The transient wave of VEGFR-2-immunoreactive subepicardial mesenchymal cells has been recorded in the hamster embryo by 10±11 days post coitum. In the mouse embryo, whose developmental stages are similar to those of the hamster between 9 and 12

Immunohistochemical identi®cation of coronary vessel precursors days post coitum after our own observations, vascular endothelial growth factor (VEGF) production has been demonstrated in the myocardium, with a maximum expression by 10.5 days post coitum, i.e. at the onset of the development of the subepicardial capillary plexus (Dumont et al., 1995). Furthermore, exogenous VEGF stimulated the appearance of cord-like structures among the mesenchymal cells migrating from rat embryo hearts explanted on collagen gels (Ratajska et al., 1995). These observations, together with our report of VEGFR-2-immunoreactive subepicardial mesenchymal cells, suggest that the signalling system VEGF=VEGFR-2 can play a crucial role in the early development of the cardiac vessels. On the other hand, the existence of this signalling system may explain why a primary vasculogenic process occurs in the heart without a close mesoderm±endoderm interaction. This interaction is a key factor according to some current theories of embryonic vasculogenesis (Flamme, 1989; Pardanaud et al., 1989), and it seems to be mediated by VEGF secreted by the endoderm and the VEGFR-2 located at the mesodermal cells (Breier et al., 1995; Palis et al., 1995). By 10±11 days post coitum, the myocardium, through its secretion of vascular endothelial growth factor, might be playing the role of the endoderm as the inducer of the responsive mesodermal cells that express the type 2 receptor of this growth factor. It would also explain the normal vascularization of the embryonic heart from intrinsic progenitors, even when it is grafted in oculo, far away from any endodermal tissue (Rongish et al., 1994). Some of the VEGFR-2-immunoreactive subepicardial cells were grouped in compact clusters of rounded cells, which might be interpreted as immature angiocysts. This is an interesting observation, as most morphological descriptions of the development of the coronary system in mammalian embryos have reported the presence of subepicardial blood island-like vessels (Hirakow, 1983, 1992; Hutchins et al., 1988). In this regard, the similarity of some large and rounded VEGFR-2-immunoreactive subepicardial mesenchymal cells to the haematopoietic stem cells located in the liver at the same embryonic stages should be noted. We have reported recently that most subepicardial mesenchymal cells, as well as some capillarylike structures of the subepicardium of chick and hamster embryos, are transiently labelled with anticytokeratin antibodies. We have proposed that this feature may be caused by the persistence of remains of a cytokeratin cytoskeleton indicative of a mesothelial origin (PeÂrez-Pomares et al., 1997). The collocalization of VEGFR-2and cytokeratin immunoreactivity in some subepicardial mesenchymal cells, reported here, allows the possibility that mesothe-

633

lial-derived cells can acquire a vasculogenic potential in the embryonic heart, a hypothesis that has been proposed recently (Markwald et al., 1996; MunÄoz-ChaÂpuli et al., 1996). However, a detailed cell lineage study of the subepicardial mesenchymal cells would be needed to address this point. Acknowledgements The authors sincerely thank Dr Schuh (University of Toronto) for his kind gift of the anti-VEGFR-2 antibody and JesuÂs SantamarõÂa for his help with laser confocal microscopy. This work was supported by grant PB95-0475 (Ministerio de EducacioÂn y Cultura, Spain). J. M. PeÂrez-Pomares is the recipient of a fellowship from the Ministerio de EducacioÂn y Cultura, Spain. References (1990) Coronary vessel vasculogenesis. In: Embryonic Origins of Defective Heart Development (edited by B O C K M A N , D . E . & K I R B Y , M . L . ). Ann. NY Acad. Sci. 588, 340±4. B R E I E R , G . , C L A U S S , M . & R I S A U , W . (1995) Coordinate expression of VEGF-receptor 1 (¯t-1) and its ligand suggests a paracrine regulation of murine vascular development. Dev. Dyn. 204, 228±39. BOLENDER, D., OLSON, L.M. & MARKWALD, R.R.

DUMONT, D.J., FONG, G.H., PURI, M.C., GRADWOHL, G.,

(1995) Vascularization of the mouse embryo: a study of ¯k-1, tek, tie and vascular endothelial growth factor expression during development. Dev. Dyn. 203, 80±92. F L A M M E , I . (1989) Is extraembryonic angiogenesis in the chick embryo controlled by the endoderm? A morphological study. Anat. Embryol. 180, 259±72. ALITALO, K. & BREITMAN, M.L.

FONG, G.H., ROSSANT, J., GERTSENSTEIN, M. & BREITMAN, M . (1995) Role of the Flt-1 receptor tyrosine kinase in regulating the assembly of vascular endothelium. Nature 376, 66±70. H I R A K O W , R . (1983) Development of the cardiac blood vessels in staged human embryos. Acta Anat. 115, 220±30. H I R A K O W , R . (1992) Epicardial formation in staged human embryos. Acta Anat. Nippon 67, 616±22. HUTCHINS, G.M., KESSLER-HANNA, A. & MOORE, G.W.

(1988) Development of the coronary arteries in the embryonic human heart. Circulation 77, 1250±7.

 NDEZ-TERA  N, M.A. & OJEDA, J.L. ICARDO, J.M., FERNA

(1990) Late heart embryology. The making of an organ. In: Handbook of Human Growth and Developmental Biology, Vol. 3 (edited by M E I S A M I , E . & T I M I R A S , P . S . ), pp. 25±49. Boca Raton: CRC Press.

MARKWALD, R.R., EISENBERG, C., EISENBERG, L., TRUSK,

(1996) Epithelial±mesenchymal transformations in early avian heart development. Acta Anat. 156, 173±86. M I K A W A , T . & F I S C H M A N , D . A . (1992) Retroviral analysis of cardiac morphogenesis: discontinuous formation of coronary vessels. Proc. Natl. Acad. Sci. USA 89, T., & SUGI, Y.

PEÂREZ-POMARES et al.

634 9504±8.

(1996) Pericardial mesoderm generates a population of smooth muscle cells migrating into the heart along with ingrowth of the pericardial organ. Dev. Biol. 174, 221±32.

MIKAWA, T. & GOURDIE, R.G.

MILLAUER, B., WIZIGMANN-VOOS, S., SCHNURICH, H., M A R T ÂI N E Z , R . , M O L L E R , N . P . H . , R I S A U , W . , U L L R I C H ,

(1993) High af®nity VEGF binding and developmental expression suggest Flk-1 as a major regulation of vasculogenesis and angiogenesis. Cell 72, 835±46.

A.

 P U L I , R . , M A C I A S , D . , R A M O S , C . , G A L L E G O , Ä OZ-CHA MUN A . & A N D R E S , V . (1996) Development of the sub-

epicardial mesenchyme and the early cardiac vessels in the dog®sh (Scyliorhinus canicula). J. Exp. Zool. 275, 95±111. N O D E N , D . M . (1991) Development of craniofacial blood vessels. In: The Development of the Vascular System (edited by F E I N B E R G , R . N . , S H E R E R , G . K . & A U E R B A C H , R . ), pp. 1±24. Basle: Karger. P A L I S , J . , M C G R A T H , K . E . & K I N G S L E Y , P . D . (1995) Initiation of hematopoiesis and vasculogenesis in murine yolk sac explants. Blood 86, 156±63.

Natl. Acad. Sci. U.S.A. 90, 7533±7. RATAJSKA, A., TORRY, R.J., KITTEN, G.T., KOLKER, S.J. &

(1995) Modulation of cell migration and vessel formation by vascular endothelial growth factor and basic ®broblast growth factor in cultured embryonic heart. Dev. Dyn. 203, 399±407. R I S A U , W . (1991) Vasculogenesis, angiogenesis and endothelial cell differentiation during embryonic development. In: The Development of the Vascular System (edited by F E I N B E R G , R . N . , S H E R E R , G . K . & A U E R B A C H , R . ), pp. 58±68. Basle: Karger. R I S A U , W . & F L A M M E , I . (1995) Vasculogenesis. Annu. Rev. Cell. Dev. Biol. 11, 73±91. TOMANEK, R.J.

RISAU, W., SARIOLA, H., ZERWES, H.G., SASSE, J., EKBLOM,

(1988) Vasculogenesis and angiogenesis in embryonic-stem-cellderived embryoid bodies. Development 102, 471±8.

P., KEMLER, R. & DOETSCHMAN, T.

RONGISH, B.J., TORRY, R.J., TUCKER, D.C. & TOMANEK, R.J.

(1994) Neovascularization of embryonic rat hearts cultured in oculo closely mimics in utero coronary vessel development. J. Vasc. Res. 31, 205±15.

SHALABY, F., ROSSANT, J., YAMAGUCHI, T.P., GERTSEN-

P A R D A N A U D , L . , Y A S S I N E , F . & D I E T E R L E N - L I EÁ V R E , F .

STEIN, M., WU, X.F., BREITMAN, M.L. & SCHUH, A.C.

(1989) Relationship between vasculogenesis, angiogenesis and haemopoiesis during avian ontogeny. Development 105, 473±85.

(1995) Failure of blood-island formation and vasculogenesis in ¯k-1-de®cient mice. Nature 376, 62±6. T O K U Y A S U , K . T . (1985) Development of myocardial circulation. In: Cardiac Morphogenesis (edited by F E R R A N S , V . J . , R O S E N Q U I S T , G . & W E I N S T E I N , C . ), pp. 226±37. New York: Elsevier. T O M A N E K , R . J . (1996) Formation of the coronary vasculature: a brief review. Cardiovasc. Res. 31, E46±E51. V I R A G H , S . & C H A L L I C E , C . E . (1981) The origin of the epicardium and the embryonic myocardial circulation in the mouse. Anat. Record 201, 157±68.

P EÂ R E Z - P O M A R E S , J . M . , M A C IÂ A S , D . , G A R C IÂ A - G A R R I D O , L . Â P U L I , R . (1997) Contribution of the Ä OZ-CHA & MUN

epicardium to the subepicardial mesenchyme in hamster and chick embryos. Dev. Dyn. 210, 96±105.

POELMANN, R.E., GITTENBERGER-DE GROOT, A.C., MENÈ K E N K A M P , R . & H O G E R S , B . (1993) TINK, M.M.T., BO

Development of the cardiac coronary vascular endothelium, studied with antiendothelial antibodies, in chicken-quail chimeras. Circ. Res. 73, 559±68. P O O L E , T . J . & C O F F I N , J . D . (1991) Morphogenetic mechanisms in avian vascular development. In: The Development of the Vascular System (edited by F E I N B E R G , R . N . , S H E R E R , G . K . & A U E R B A C H , R . ), pp. 25±36. Basle: Karger.

QUINN, T.P., PETERS, K.G., DE VRIES, C., FERRARA, N. &

(1993) Fetal liver kinase 1 is a receptor for vascular endothelial growth factor and is selectively expressed in vascular endothelium. Proc.

WILLIAMS,

L.T.

VIRAGH, S., KALMAN, F., GITTENBERGER-DE GROOT, A.C.,

(1990) Angiogenesis and hematopoiesis in the epicardium of the vertebrate embryo heart. In: Embryonic Origins of Defective Heart Development (edited by B O C K M A N , D . E . & K I R B Y , M . L . ). Ann. NY Acad. Sci. 588, 455±8. POELMANN, R.E. & MOORMAN, A.F.M.

YAMAGUCHI, T.P., DUMONT, D.J., CONLON, R.A., BREIT-

(1993) ¯k-1, an ¯t-related receptor tyrosine kinase is an early marker for endothelial cell precursors. Development 118, 489±98.

MAN, M.L. & ROSSANT, J.