SEARCH

SEARCH BY CITATION

Keywords:

  • lymphangiogenesis;
  • Lyve-1;
  • prox-1;
  • mouse embryonic heart;
  • heart development;
  • lymphatic vessels;
  • collecting lymphatics

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

We aimed to study the spatiotemporal pattern of lymphatic system formation in the embryonic and early postnatal mouse hearts. The first sign of the development of lymphatics are Lyve-1–positive cells located on the subepicardial area. Strands of Lyve-1–positive cells occur first along the atrioventricular sulcus of the diaphragmatic surface and then along the great arteries. Lumenized tubules appear, arranged in rows or in a lattice. They are more conspicuous in dorsal atrioventricular junction, along the major venous and coronary artery branches and at the base of the aorta and the pulmonary trunk extending toward the heart apex. At later stages, some segments of the lymphatic vessels are partially surrounded by smooth muscle cells. Possible mechanisms of lymphangiogenesis are: addition of Lyve-1–positive cells to the existing tubules, elongation of the lymphatic lattice, sprouting and coalescence of tubules. We discuss the existence of various subpopulations of endothelial cells among the Lyve-1–positive cells. Developmental Dynamics 237:2973–2986, 2008. © 2008 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Studies of the lymphatic system have been hampered for a long time by the lack of specific markers of lymphatic vessels. Despite this, many studies have been conducted on the distribution of lymphatic vessels in adult organs using India ink, Evans blue, or resin injections and by anatomical dissections. Hearts of many species have been explored relative to the distribution of major lymphatic trunks and the lymphatic capillary plexuses (Eliškova and Eliška,1974,1976; Gołąb,1977; Böger and Hort,1977; Eliška and Eliškova,1980; Esperança Pina and Sampaio Tavares,1980; Shimada et al.,1990; Miller et al.,1996; Sacchi et al.,1999). However, little is known about the origin and spatiotemporal formation and maturation of the lymphatic system in the embryonic heart. Primordia of the lymphatic vessels in mammals and birds are lymph sacs, which arise from embryonic veins (Sabin,1909). According to Sabin (1902,1904) and Lewis (1905), lymph sacs descend from the venous system and the remaining lymphatics of the surrounding tissues originate by sprouting off lymphatic endothelial cells (LECs) from lymph sacs. This is called the centrifugal theory.

Another theory (centripetal) claims that LECs arise de novo from undifferentiated mesenchymal cells or circulating precursor cells (Wilting et al.,2006) as has been suggested by Huntington and McClure and Kampmeier (Huntington,1908; Huntington and McClure,1910; Kampmeier,1912). According to them, initial lymph sacs arise in the mesenchyme, independent of the veins and secondarily establish venous connections. Circulating precursors of LECs in mammals may also differentiate from hematopoietic-tissue-derived endothelial progenitors which bear Syk- and Slp-76 antigens (Sebzda et al.,2006), from transdifferentiating leukocytes (Buttler et al.,2006) or from transdifferentiating macrophages (Maruyama et al.,2005). Another kind of differentiation in adults has been reported to occur from bone marrow-derived cells which acquire LEC antigens during lymphangiogenesis, also in pathologic conditions (Salven et al.,2003; Religa et al.,2005; Kerjaschki et al.,2006).

Lymphatic vessels in chicken embryos are of dual origin: they arise from veins and from scattered mesodermal precursor cells (Wilting et al.,2006). On the contrary, in zebrafish the LECs of the main thoracic duct vessel arise from endothelial cells of primitive veins, as it has been presented in time-lapse imaging studies (Yaniv et al.,2006). The venous endothelial cells from the parachordal vessel migrate, assume mesenchymal phenotype, proliferate, and subsequently settle in the area where the thoracic duct is formed just ventrally from the dorsal aorta. These cells form disconnected vascular segments which extend rostrally and caudally and coalesce forming a single vessel. Thus, in this case, the LECs of the thoracic duct derive from venous endothelium, by an intermediate stage of migratory mesenchymal cells, but not from cells sprouting directly from veins.

We do not know, however, if the entire lymphatic system in this species is of venous origin or whether it has other contributing sources.

Markers for LECs were detected in recent years. Among these is prospero-related transcription factor Prox-1, which demarcates nuclei of LECs (Wilting et al.,1999; Wigle and Oliver,1999). Lyve-1, the hyaluronian receptor, on the other hand, is a highly specific marker of LECs in many tissues and organs (Banerji et al.,1999). The function of Prox-1 is indispensable for lymphatic vessel development in vivo. Inactivation of Prox1 in mice results in a defective sprouting of endothelial cells (ECs) from the cardinal vein, loss of lymphatic-marker expression, and lack of the lymphatic vasculature, whereas blood vessels seem to be unaffected (Wigle and Oliver,1999). On the contrary, Lyve-1–deficient mice appear normal, without any apparent lymphatic malformations (Hong et al.,2004). Currently there is much review literature devoted to the development of the lymphatic system and diversification of endothelial cell precursors into venous, arterial and lymphatic lines (Wilting et al.,1999; Jeltsch et al,2003; Hong et al.,2004; Harvey and Oliver,2004; Alitalo et al.,2005; Liersch and Detmar,2007; Yamashita,2007).

Sabin's centrifugal theory has been recently supported by studies demonstrating that highly specific markers of LECs, prox-1 and Lyve-1, are already expressed in early anlagen of lymphatics, lymph sac, as well as in the nearby veins of murine embryos before the lymph sacs are present (Wigle and Oliver,1999; Wigle et al.,2002; Oliver,2004). Thus, at least one means of lymphatic system morphogenesis involves the acquisition of the LEC identity and separation from the blood vessel network (Liersch and Detmar,2007).

Derivation of lymphatic vessels in the avian heart has been recently studied by Wilting et al. (2007). Using quail–chicken chimeras, where the proepicardial organ from quail has been grafted to chicken, the authors found that the proepicardium does not deliver precursors for LECs. Proepicardium is a transient organ, present during heart development, which forms on the ventral surface of both venous sinuses (Ho and Shimada,1978; Virágh et al.,1993; Männer et al.,2001) extending to the pericardial cavity and attaching to the naked heart surface. The proepicardium gives rise to the epicardial cover of the heart and is also a source of endothelial, smooth muscle cells and fibroblasts of coronary vasculature as well as interstitial fibroblasts of the developing heart (Mikawa and Fishman,1992; Mikawa and Gourdie,1996; Dettman et al.,1998; Gittenberger-de Groot et al.,1998; Pérez-Pomares et al.,2002). Regardless of the recent progress in studies on lymphatic system development and the origin of lymphatic vessels, our knowledge on the spatiotemporal pattern of lymphatic vessel formation in the embryonic heart is still scarce. There is also a scarcity of data concerning lymphangiogenic patterns during embryonic heart development, although we can presume that the formation of lymphatic vessels may proceed by similar cellular mechanisms as angiogenesis of blood vessels.

The aim of this work was to study morphologic patterns of lymphatic vessel formation, spatiotemporal location of their primordia, the ways of their elongation, maturation, and remodeling in embryonic and early postnatal mouse hearts. We performed the study based on detection of Lyve-1 antigen on endothelial cells according to the literature reports indicating that this antigen is specific for LECs (Banerji et al.,1999; Prevo et al.,2001; Jackson,2003). Because collecting lymphatics and precollector vessels containing smooth muscle cells in their wall have been reported to be present in the human and canine hearts (Kline,1969; Sacchi et al.,1999), we also wanted to study the inception and differentiation of collecting lymphatics in the mouse heart and the occurrence of mural cells in their walls by the detection of smooth muscle actin (SMA) antigen.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

We have studied frozen sections and whole-mount stained hearts of inbred mouse embryos (Balb/c) between embryonic days 11–18 and postnatal days (PD) 4–28. Initially (in 11 days post coitum [dpc] hearts) a few scattered cells faintly labeled with Lyve-1 antibodies were found within the trabecular myocardium and the subepicardium. Cells bearing Lyve-1 antigen did not form any aggregates. On the venous pole of the heart they were present within the proximal part of the left cranial caval vein, in the proximal part of the right cranial caval vein, in the pulmonary vein, in the endocardium near the entrance of the pulmonary vein to the left atrium. Thus, endothelial cells of the early systemic veins bore Lyve-1 antigen. There were also some scattered Lyve-1–positive cells on the dorsal surface of the heart. Only a few Lyve-1–positive cells were found in the entire outflow tract on the arterial pole of the heart (Fig. 1). Liver sinusoids also expressed the Lyve-1 antigen (Fig. 1a). In 12 dpc hearts, scattered cells bearing Lyve-1 antigen were more numerous as compared to those in 11 dpc hearts and were mostly located in the subepicardial area on the diaphragmatic surface of the heart.

thumbnail image

Figure 1. An 11.5 days post coitum (dpc) embryo cut transversely at the level of the heart; selected pictures from serially cut sections; Lyve-1–positive cells are brown; sections were counterstained with hematoxylin. a: Small magnification of a section taken at the level of the left cranial caval vein (marked with a star) and the ventricles; the area in the rectangle is enlarged in b, note that hepatic sinusoids are also Lyve-1–positive; a small star indicates the left anterior cardinal vein which exhibits the Lyve-1 antigen; liver sinusoids exhibit the Lyve-1 antigen. b: Arrows point to Lyve-1–positive cells located on the dorsal surface of the heart and to a population of endothelial cells building partially the lumen of the left cranial caval vein. c: A small magnification of the fetus cut at the level of the entrance point of the pulmonary vein to the left atrium; this area is enlarged in d. d: Endothelial cells of the pulmonary vein are Lyve-1–positive; Lyve-1–positive cells are seen in the endocardial endothelial cells of the left atrium near the opening of the pulmonary vein (arrow); Lyve-1–positive cells are also seen forming a lumen of a small vascular tubule on the surface of the right atrium (small arrow). e: A section through the heart at the level of the opening of the right cranial caval vein (marked with a star) to the right atrium. f: The right anterior cardinal vein with Lyve-1–positive cells forming the lumen of its proximal part; the distal part opens to the right atrium as the right cranial caval vein; faint Lyve-1–positive cells are located on the surface of the right venous valve (arrow). LB, lung buds; RV, right ventricle, LV, left ventricle; LA, left atrium; RA, right atrium; L, liver; OT, outflow tract. Scale bars = 500 μm in a,c,e, 200 μm in b,d,f.

Download figure to PowerPoint

A total of 13 dpc whole-mount–stained hearts were covered with scattered cells expressing the Lyve-1 antigen. These cells were located predominantly on the diaphragmatic surface of the heart and were more numerous close to the base of the heart as compared to the heart apex and as compared to the sternal surface (Fig. 2a–c). This gradient of Lyve-1–positive cells was less pronounced in 13.5 dpc hearts, when these cells were regularly seeded within almost equal distances from each other, covering the entire surface of the heart (Fig. 3a). Groups of cells appeared also deeper within the myocardium and these seemed to derive from epicardially positioned cells sprouting toward myocardium (Fig. 3c). Diaphragmatic surface of the atrioventricular sulcus and coronary sinus was the area of advanced development of Lyve-1–positive tubules (Fig. 3b,d). Few penetrations of cells bearing the Lyve-1 antigen occurred within the atrial wall and the wall of the coronary sinus: these penetrations occurred either from the endocardial (Fig. 3e) or from the epicardial surface (Fig. 3f). This area was also invaded by prox-1–positive cells, however, the latter were less numerous as compared to Lyve-1–positive cells (data not shown).

thumbnail image

Figure 2. a: Whole-mount–stained 13 days post coitum (dpc) hearts; a few Lyve-1–positive cells cover the sternal surface of the heart. b: Cells are more numerous close to the base of the heart as visualized on the left lateral view. c: Diaphragmatic surface is covered with numerous cells. Scale bars = 1 mm.

Download figure to PowerPoint

thumbnail image

Figure 3. A 13.5 days post coitum (dpc) heart, transverse sections; selected serial sections labeled with Lyve-1 antibodies and counterstained with hematoxylin. a: Scattered Lyve-1–positive cells and cellular clusters on the surface of both ventricles (brown). b: A few cellular aggregates are visible deeper in the myocardial wall on the diaphragmatic surface of the heart (arrows). c,e,f: Cellular penetrations across the myocardial wall are visible in some areas (c, arrow), and in the wall of the coronary sinus (e,f, arrows). d: Single, short tubules appear on the diaphragmatic surface of the heart on the surface of coronary sinus (arrow). RV, right ventricle; LV, left ventricle; CS, coronary sinus; RA, right atrium. Scale bars = 500 μm in a, 200 μm in b–f.

Download figure to PowerPoint

In 14.5 dpc hearts, lumenizing tubules, which resembled vessels, were again more conspicuous on the diaphragmatic surface of ventricles and atria with a predominance at the atrioventricular sulcus of the heart (Fig. 4a). Some of the cells or short tubules tended to be arranged in rows (Fig. 4b). Collecting vessels ran cranially on the surface of the left atrium. The primordial lymphatics on the sternal surface of the heart were less developed and consisted of scattered Lyve-1–positive cells. On the roots of the great arteries a net-like structure containing lumen was beginning to form (Fig. 4c). The arterial pole of the heart was invaded by Lyve-1–positive cells which were located between the aorta and the pulmonary trunk. However, these vascular tubes or cellular strands were discontinuous and did not descend below the level of the semilunar valves (Fig. 5a–d).

thumbnail image

Figure 4. ac: Whole-mount–stained 14.5 dpc hearts. Diaphragmatic (a,b) and sternal (c) surface. Lyve-1–positive scattered cells appear on the sternal and diaphragmatic surface of the heart. Lymphatic tubules resembling vessels emerged on the diaphragmatic surface of the left and the right atrium and the coronary sinus; discontinuous tubules arranged in a row appear on the left ventricle of one heart (arrow in a); some tubules are formed (white arrow in b); other tubules seem to coalesce from Lyve-1–positive cells arranged in a row (black arrow in b); the first clusters of cell aggregates in the form of a lattice are seen on the roots of the aorta (arrow in c). Scale bars = 1 mm in a,c, 50 μm in b.

Download figure to PowerPoint

thumbnail image

Figure 5. A 14.5 days post coitum (dpc) heart. Selected photos taken from serially cut transverse sections at the level of great arteries cut above the semilunal valves, immunolabeled with anti-Lyve-1 (brown) and counterstained with hematoxylin; ad: Lyve-1–positive cellular clusters penetrating between the aorta and the pulmonary trunk (marked with black arrows, a,b), and second Lyve-1–positive vesicle on the surface of the aorta (white arrow) are discontinuous while approaching the semilunar valves (c,d). d: At the roots of great vessels, an exuberant accumulation of Lyve-1–positive cells is evident. a, aorta; pt, pulmonary trunk; pa, pulmonary artery. Scale bars = 200 μm.

Download figure to PowerPoint

In 15.5 dpc hearts, lymphatic capillary vessels with lumens of irregular shapes were more developed within the crux of the heart (Fig. 6a,b) extending toward the apex within the subsinusoidal interventricular sulcus and cranially on the atrial surface. Lumenized tubules appeared on the left side of the conus along the left conal vein (Fig. 6c, black arrow). Scattered Lyve-1–positive cells were regularly observed between tubules, within the openings of the lymphatic capillary plexus, and on the remaining surface of the heart. The cellular shapes were either fusiform or formed arborizations reminiscent of nerve-cell outgrowths (Fig. 6b). Single cells or groups of cells seemed to form processes toward the existing tubules (Fig. 6b, black arrow). The arterial pole of the heart was entirely enveloped by lymphatic tubules positioned around the great arteries and between them (Fig. 7a–f). These tubules formed a plexus at the roots of the great arteries which was visible in whole-mount–stained hearts (not shown).

thumbnail image

Figure 6. Whole-mount–stained 15.5 days post coitum (dpc) hearts. ac: Lyve-1–positive scattered clusters of cells cover the entire heart surface; tubules are located on the diaphragmatic (a,b) and the sternal surface (c, arrows) of the heart; in panel a major lymphatic tubules are visible: running on the left atrium, within the atrioventricular sulcus, on the left ventricle, and within the interventricular sulcus; a collecting lymphatic vessel runs cranially on the surface of the coronary sinus (a). Panel b indicates scattered Lyve-1–positive cells among the forming tubules; some cells formed ramifications like nerve cell outgrowths; the arrow in b indicates a ramified cell approaching adjacent tubule; c: The atrial appendages have been removed to better demonstrate lymphatics (arrow) on the sternal surface. Scale bars = 1 mm in a,c, 50 μm in b.

Download figure to PowerPoint

thumbnail image

Figure 7. a–f: A 15.5 days post coitum (dpc) heart; selected pictures taken from serially cut transverse sections through the roots of the great arteries (a–c,e–f) and at the level of the semilunar valves (d). a,b: Cranially a small amount of Lyve-1–positive clusters and tubules along the great arteries (marked with arrows) are demonstrated. c,d: At the level of the semilunar valves, the lymphatic plexus is more developed. e,f: Paraffin sections of a 15.5 dpc heart from corresponding area stained with anti–prox-1; lymphatic vessels around the aorta and the pulmonary trunk labeled with arrows. f: A higher magnification of the vessel between the aorta and the pulmonary trunk shown in e. Scale bars = 200 μm in a–c, 500 μm in d, 100 μm in e,f.

Download figure to PowerPoint

Advanced development of Lyve-1–positive tubules was found in 16.5 dpc hearts. They formed a capillary plexus reminiscent of a net-like structure with separate tubules extending toward the apex (Fig. 8a, white arrow). At higher magnification these tubules seemed to be arranged in discontinuous strands (Fig. 8b, arrows). On the sternal surface of the heart, an accumulation of tubules running on the left side of the conus was more developed, as compared to that at earlier stages. These tubules seemed to form a lattice or a net-like structure with the major trunk discernible (Fig. 8c, arrow). A cell seemed to sprout from one lattice toward another and join the walls of the adjacent set of tubules forming “a bridge” (see also Fig. 11a). There were numerous large-diameter tubules (but with collapsed lumina) forming sprouts within the atrioventricular sulcus on the atria and in the coronary sinus (not shown). A lattice of an irregularly shaped lymphatic capillary plexus at the roots of the great arteries was seen to extend over the conus of the heart (Fig. 8d, arrow).

thumbnail image

Figure 8. a–d: Whole-mount–stained hearts on 16.5 days post coitum (dpc). Lyve-1–positive tubules located on the diaphragmatic (a,b) and sternal surface (c,d) of the heart. a: The tubule on the surface of the left atrium (marked with the large black arrow) running rostrally is presumably a collecting vessel, similar to a tubule running on the surface of the coronary sinus (small black arrow). Lymphatic vessels located on the atrioventricular sulcus form a pattern of arborization; they have large diameters and some of them extended toward the apex into discontinuous tubules (white arrow). b: A higher magnification of tubules indicated by a white arrow in a. c: The arrow points to tubules (or capillary lymphatic plexus) which assemble along the left conal vein. d: A discontinuous lattice of lymphatic capillary plexus is shown on the surface of the pulmonary trunk (pt) and the aorta (a). This plexus spreads on the heart conus (marked with a black arrow). Scale bars = 1 mm in a,c, 0.5 mm in d, 100 μm in b.

Download figure to PowerPoint

thumbnail image

Figure 11. a–c: Whole-mount–stained hearts at the 16.5 days post coitum (dpc; a) and 17.5 dpc (b,c) stage of development. a: Lyve-1–positive cells scattered among the forming tubules; one cell forms a “bridge” between assembling tubules (black arrow). b: A cellular “bridge” between two adjacent existing tubules, marked with a black arrow. c: Lyve-1–positive cells scattered among the lymphatic lattice and within the openings of the lattice (a star) and assuming different shapes (left); lymphangiogenic sprouting from existing tubule (white arrow); a cell seems to sprout toward the existing tubule (black arrow).

Download figure to PowerPoint

On serial sections of 16.5 dpc hearts, Lyve-1–positive cells were observed on the cardiac surface and within the myocardium of both ventricles (Fig. 9a, black arrows). Some of the scattered Lyve-1–positive cells assembled in vesicles or tubules (Fig. 9a,c,d, white arrows). Some tubules or lymphatic vessels had a distended lumen (Fig. 9b,d). Lymphatic endothelial cells that did not seem to form a lumen were also found on the surface of the coronary sinus and in the adjacent myocardium (Fig. 9a,c, black arrows). At the roots of the great vessels lymphatics formed a lattice (Fig. 9e,f).

thumbnail image

Figure 9. a–d: Transverse sections of the mouse heart on 16.5 days post coitum (dpc) stained with anti–Lyve-1 (brown) and anti–smooth muscle actin (navy blue). a: Lyve-1–positive cell aggregates on the surface of the ventricle; some of them form a lumen (upper white arrow); a delicate lattice of Lyve-1–positive cells is shown within the myocardium (black arrows). b: A cross-section of the heart with Lyve-1–positive cellular clusters forming lumen on the surface of the heart. c: A higher magnification of coronary sinus area covered with lumenized lymphatic tubules (white arrows) and nonlumenized Lyve-1–positive aggregates (black arrows). d: The atrioventricular sulcus with a capillary lymphatic vessel with a dilated lumen, marked with a white arrow; coronary artery endothelial cells do not stain with anti–Lyve-1 (black arrow in d) but are surrounded by mural cells (tunica media) exhibiting α-SMA. e,f: A paraffin section (e, small magnification; f, larger magnification) of the arterial pole of the 16.5 dpc heart stained with anti–prox-1; lymphatics were marked with black arrows; the white arrow indicates the lymphatic vessel corresponding to the one shown in d), marked with a white arrow. cs, coronary sinus; a, aorta; p, pulmonary trunk. Scale bars = 100 μm.

Download figure to PowerPoint

In 17.5 dpc hearts, more coalesced Lyve-1–positive tubules and lymphatic vessels were found on the subepicardial surface of the heart. They formed a lattice of branching, continuous tubules around major venous branches: the cardiac caudal vein (Fig, 10a), the right cardiac vein (Fig. 10b), the left conal vein (Fig. 10c,f), the left cardiac vein (Fig. 8e). The lymphatic plexus was equally developed on the diaphragmatic surface of the heart as on the sternal surface of the heart (Fig. 10a,c). This plexus covered major parts of the left ventricle, the right ventricle, the area near the apex, and the conus of the heart. Richly branching tubules were also found around the coronary sinus (the proximal part of the left cranial caval vein; Fig. 10e). By the end of the fetal life, the lymphatic plexus of the heart consisted of two developed drainage areas: one collecting lymph from the left ventricle and the other collecting lymph from the right ventricle. The border between these two systems was visible near the heart apex (Fig. 10g). There were also two major collecting vessels: one on the sternal surface of the heart running along the left conal vein and the other on the diaphragmatic surface of the heart running rostrally on the surface of the left atrium (on the surface of the coronary sinus).

thumbnail image

Figure 10. Whole-mount–stained 17.5 days post coitum (dpc) hearts. Lyve-1–positive tubules and lymphatic vessels are located along the major branches of cardiac veins visible as translucent “ghosts” on the following panels. a: The cardiac caudal vein, marked with a black arrow. b: The right cardiac vein marked with a black arrow. c,f: The left conal vein. e: The left cardiac vein. d: The lymphatic tubules form a lattice around the aorta (a), and the pulmonary trunk (pt); the tubules arborize on the surface of the right (b) and the left (c,f) ventricle. g: The border of two collecting lymphatic systems: draining lymph from the left ventricle (LV) and the right ventricle (RV) lies near the heart apex at the interventricular sulcus. Scale bar = 500 μm.

Download figure to PowerPoint

In whole-mount–stained specimens of the 16 dpc, 17 dpc, and 18 dpc hearts some lymphangiogenic patterns could be discerned at higher magnifications. There were single scattered Lyve-1–positive cells located within the openings of the lymphatic lattice (Fig. 11c, a black star); some of them assumed different forms of cellular shapes: fusiform, spindle-like, star-like with long processes similar to nerve cell outgrowths (see Fig. 5b); they protruded toward the forming tubules (Fig. 11a, black arrow) and existing tubules (Fig. 11b, black arrow) and seemed to come in contact with them forming a “cellular bridge”. Ramifying cells were also positioned adjacent to one end of an existing tubule but without coming in contact with it (Fig. 11c, black arrow). There were also signs of sprouting at the edges of the existing tubules (Fig. 11c, white arrow).

At later stages of development, tubules, lymphatic vessels, and scattered Lyve-1–positive cells appeared around coronary artery branches. The lymphatic system was better developed close to the base of the heart and on diaphragmatic surface of the heart as compared to the apex and the sternal surface (Fig. 12a–h). Lymphatics were also developed on the conus and at the roots of the aorta and the pulmonary trunk (Fig. 12e). The Lyve-1–positive vessels, partly surrounded by smooth muscle cells, were located within the aorto-pulmonary sulcus and on the roots of the pulmonary trunk (Fig. 12h, white arrow). All remaining lymphatic vessels were devoid of cells containing the α-SMA antigen during embryonic and early postnatal period. Moreover, we observed that some endocardial endothelial cells were also stained with anti–Lyve-1 antibodies. These cells were located in both ventricles, auricles, within the coronary sinus, and in the left atrium (in the dorsal wall of the left atrium) (Fig. 12a,f,g). Some Lyve-1–positive cells in the subsinusoidal interventricular sulcus assembled in clusters without yet forming a lumen (Fig. 12d).

thumbnail image

Figure 12. Selected sections from a serially cut 17.5 days post coitum (dpc) heart labeled with anti–smooth muscle actin (SMA; navy blue) and anti-Lyve-1 (brown). a: Lyve-1–positive cells mark some Endocardial endothelial cells within the myocardium (marked with an arrow). b,c: Some accumulations of Lyve-1–positive cells above the mitral valve (arrow in b) and in the atrioventricular sulcus (arrow in c).d: Scattered Lyve-1–positive cells in the interventricular sulcus around the major caudal vein. e: A lattice of lymphatic vessels at the roots of the aorta and the pulmonary trunk (black arrows). f: An exuberant lattice of lymphatic vessels around the base of the coronary sinus. g: A population of endocardial endothelial cells exhibiting Lyve-1 antigen within the left atrium (marked with an arrow). h: SMA-containing cells partially covering a lymphatic vessel located between the pulmonary trunk and the aorta (marked with an arrow). CS, coronary sinus; LA, left atrium; A, aorta; PT, the pulmonary trunk. Scale bar = 100 μm.

Download figure to PowerPoint

On postnatal day 1, the lumenized vessels traversing along the interventricular sulcuses (either subsinusoidal or paraconal) were well developed. A net-like capillary plexus extended further toward the apex and covered both ventricles and atria. Single scattered Lyve-1–positive cells were continuously present in the heart, predominantly within the subepicardium.

In postnatal day (PD) 4 hearts, lymphatics traversed the outer layer of the myocardial wall and were found around the major branches of cardiac veins and coronary arteries (Fig. 13c,d). At the truncus arteriosus (between the pulmonary trunk and the aorta), lymphatic vessels consisted of endothelial cells forming large lumenized vessels, partly surrounded by smooth muscle cells (Fig. 13a,b). In some sections of these lymphatic vessels, smooth muscle cells were located at a distance from the endothelium; in others, smooth muscle cells formed focal accumulations (Fig. 13a, white arrows). Thus, the mural cells around lymphatics were present only in certain segments of the vessels. The area of the heart apex was covered with continuous lymphatic tubules.

thumbnail image

Figure 13. Selected sections of a postnatal day 4 heart double-stained with anti–Lyve-1 (brown) and anti–smooth muscle actin (SMA; navy blue). a,b: Scattered groups of SMA-containing cells around lymphatic vessels between the aorta and the pulmonary trunk are visible (marked with white arrows), some smooth muscle cells are at a distance from lymphatic endothelium, lower arrow in (a). c,d: Lymphatic vessels and tubules are located around the major branches of coronary arteries (marked with black arrows).

Download figure to PowerPoint

During further stages of postnatal development (PD 8, 13, and 28) the number of lymphatic vessels increased, and apart from vessels of larger diameter there were thin-walled lymphatic vessels traversing the outer layer of the myocardial wall and having an irregularly beaded appearance (Fig. 14). In PD-28 hearts single scattered lyve-1–positive cells seemed to be no longer present.

thumbnail image

Figure 14. A postnatal-day-28 heart; whole mount staining with Lyve-1 antibodies. a: Diaphragmatic surface of the heart with exuberant lattice of lymphatic vessels. bf: Various patterns of a mature lymphatic plexus consisting of large lymphatic capillaries and tiny-diameter lymphatics exhibiting a “knotted” (“beaded”) appearance (e, arrows). Scale bars = 1 mm in a, 100 μm in b–d, 200 μm in e,f.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Derivation of Lymphatic Endothelial Cells

Recent studies on murine embryos have shown that the first cells committed to lymph endothelial lineage are located in the anterior cardinal vein in the jugular region (Wigle and Oliver,1999; Wigle et al.,2002). Cells located on the dorsolateral part of the vein sprout into the dorsolateral mesoderm, where they form the jugular lymph sacs. In 14.5 dpc embryos the cells of the jugular lymphatic sacs sprout toward different organs. According to Sabin (1904) lymphatic vessels develop from specific parts of venous system by sprouting and forming primitive lymph sacs which further develop into lymphatics. Another theory, by Huntington and McClure (1908) and Kampmeier (1912), suggests that lymphatic vessels develop by confluence of mesenchymal cells, and only the lymph sacs might be of venous origin. The third theory claims that lymphatics derive from migrating venous endothelial cells, exhibiting an intermediate mesenchymal phenotype, which settle in areas where lymph vessels are formed (Yaniv,2006). Little is known about derivation of lymphatic endothelial cells within developing heart. A current study by Wilting et al. indicates that lymphatic endothelial cells do not derive from the proepicardium and at the same time emphasizes that lymphatic endothelial cells are of extra cardiac source (Wilting et al.,2007).

The objective of our study was not to determine the derivation of endothelial lymphatic cells. However, we could perform certain observations based on a purely morphological approach regarding the points of Lyve-1–positive cells' entrance into the developing heart. First, in the earliest stages studied, Lyve-1–positive endothelial cells mark, partially or entirely, the systemic embryonic veins entering the venous pole of the heart. Lyve-1–positive cells have been also expressed in liver sinusoids. Second, the first sprouts from Lyve-1–positive cells were observed across the wall of the coronary sinus (the left cranial caval vein; Webb et al.,1998; Ciszek et al.,2007) and atria. Third: at later stages of development the arterial pole of the heart (along the pulmonary trunk and the aorta) has been invaded by Lyve-1–positive cells.

The first two observations would support the venous origin of cardiac LECs according to the centrifugal theory. Based on our observation that EC of systemic veins entering the venous pole of the heart of the early embryo exhibit, partially or entirely, Lyve-1 antigen we can not exclude a possibility that developing cardiac veins could also bear Lyve-1 antigen. The first strands of cardiac venous system occur in this area of the heart (Vrancken-Peeters et al.,1997a,b). On the other hand, we observed also in this area some prox-1–positive cells; however, they were less numerous as compared the Lyve-1-bearing cells. We thus suggest that at least some penetrating Lyve-1–positive vessels or strands in 13 dpc hearts would be the primordia of lymphatics.

Another issue comes from an observation by us and other authors (Carreira et al.,2001; Jackson,2003) that apart from being present in LEC, the Lyve-1 antigen is expressed in fetal liver and adult liver sinusoids. Based on this finding, we can hypothesize that liver primordium would be another possible source of Lyve-1–positive cells in the embryonic heart: they may enter the heart by dorsal mesocardium. This supposition has been created by analogy to Poelmann's concept on the possible origin of coronary endothelial cells from liver primordium (Poelmann et al.,1993).

In addition, we observed some scattered Lyve-1–positive cells located on the dorsal and sternal surfaces of the heart; they seemed not to exhibit any continuity with Lyve-1–positive endothelial cells of systemic vessels. This observation would be consistent with the second theory of de novo formation of lymphangioblasts from mesenchyme (Huntington and McClure, 1908; Kampmeier,1912). According to this theory a derivation of LECs would be from mesenchymal cells scattered on the diaphragmatic surface of the heart which enter the heart by means of the dorsal mesocardium. Our morphological approach allowed us to study only the direct derivation of LECs from veins by demonstrating direct connections by means of sprouts from Lyve-1–positive cell aggregates. Interestingly, we observed single, scattered Lyve-1–positive cells continuously occurring throughout prenatal and early postnatal heart development. That would indicate continuous recruitment of these cells in situ from mesenchyme by acquiring the Lyve-1 antigen. We can only speculate that these mesenchymal cells which acquire the Lyve-1 antigen may derive from remote venous endothelium.

Our morphological study did not allow us to demonstrate that venous endothelial cells may detach themselves from systemic veins, assume mesenchymal phenotype, and migrate to other areas of the developing heart, as has been detected by the time-lapse technique in zebrafish embryos during thoracic duct formation (Yaniv et al.,2006). The venous origin of LECs has been recently supported by lineage tracing studies in mouse embryos (Srinivasan et al.,2007). The authors demonstrate that all body lymphatics derive from venous endothelial cells by sprouting, proliferation, and migration. They validated Sabin's theory that lymph sacs are progenitors of LECs throughout the entire bodies of mammalian embryos. They also showed that lymphatic vessels did not arise from hematopoietic-derived cells, as had been previously suggested (Abtahian et al.,2003; Sebzda et al.,2006).

Results by Klika et al. (1972) and Rychter et al. (1971), who observed lymph vessels growing toward the heart of the chicken embryo in the grooves between the large vessels of the heart, have been in accordance with a recent observation by Wilting et al. (2007). Our results support this finding because we observed Lyve-1–positive tubes along the great arteries in 14.5 dpc hearts. We have demonstrated, however, that a lymphatic tube which occurred at the area of the great arteries was discontinuous at this stage of heart development and reached confluence later. In addition, we observed a rich lymphatic plexus (both Lyve-1– and prox-1–positive) which develops prenatally at the roots of the aorta and the pulmonary trunk, which is in line with the observation demonstrated by Wilting et al. (2007) in chicken hearts. Thus the arterial pole of the heart could be an additional entry point of lymphatics to the developing heart.

Formation and Patterning of Lymphatic Vessels

Scattered Lyve-1–positive cells are continuously present on the surface of the heart during prenatal and early postnatal life. Initially there is a gradient of Lyve-1–positive cells with predominance on the diaphragmatic surface and at the base of the heart. This developmental gradient from the base to the apex has been previously demonstrated in chicken hearts by Rychter et al. (1971), Klika et al. (1972), and recently by Wilting et al. (2007) who has also observed it in mouse hearts. Additionally, we observed a gradient of lymphatic development between the sternal and diaphragmatic surfaces of the heart. Compared with the lymphatics on the diaphragmatic surface of the heart, the lymphatics on the sternal surface are much delayed in development; however, as we mentioned previously, the diaphragmatic surface may be colonized by venous as well as lymphatic capillaries which bear the Lyve-1 antigen. This gradient is maintained up to the stage of 16 dpc mouse hearts. Further development proceeds by an extension of the lattice formed by lymphatic capillaries around the major cardiac venous trunks. The presence of a lymphatic lattice along major vein branches, the latter being surrounded by an extensive capillary network (Ratajska et al.,2003), explains the basic function of lymphatics: collecting the interstitial tissue fluid, which is a filtrate from blood capillaries. A similar pattern of lymphatic vessel location along the cardiac veins has been described in postnatal VEGFR3+/Lac-Z stained hearts (Mäkinen et al.,2001). Subsequently, development of the lymphatic system occurs along the route of major branches of coronary arteries.

By the end of the fetal life and postnatally the whole heart is almost completely covered with the lymphatic plexus and this plexus traverses also into the outer layer of the myocardial wall. In later stages of postnatal development (PD 28) lymphatics almost reach their mature phenotype and they resemble the lymphatic vessels described by B¨oger and Hort (1977) and by Shimada et al. (1990). Lymphatics have larger diameter, as compared to that of blood capillaries, and their walls are of irregular contour. Some of the lymphatics located deeper within the myocardial wall have a beaded or knotted appearance also observed by Saban et al. in postnatal hearts and in other organs (Saban et al.,2004). Of interest, in 1-, 4-, and 13-PD hearts single scattered Lyve-1–positive cells among the lymphatic capillaries still seem to be present, which indicates that there is a continuous source of cells which acquire lymphatic phenotype and are not yet integrated with lymphatic vessels. This may indicate that lymphatic system development is not completed in early postnatal life.

A plexus of lymphatics has been observed at the roots of the aorta and the pulmonary trunk. This plexus can be found in 14 dpc hearts forming initially a small focus of lymphatic capillary network; and by the end of the fetal life (17.5–18 dpc) it develops extensively into a dense mesh which is connected with the conal lymphatic plexus. In further stages of postnatal development this aorto-pulmonary lymphatic plexus diminishes markedly in size indicating that it either stops growing further while the heart enlarges during development or it involutes at later stages. The significance of this plexus around the roots of the great arteries during the prenatal life in unknown.

There are many data concerning the presence of collecting lymphatics or major lymphatic trunks within the heart (Eliškova and Eliška,1974; Leak et al.,1978; Esperança Pina and Sampaio Tavares,1980; Wilting et al.,2007). Reports state that collecting lymphatics in adult human hearts have been found at the border between the myocardium and the epicardium and that precollectors have been found around the large coronary branches (Sacchi et al.,1999). A similar pattern of collecting vessel location has been described in the macaca monkey and other mammals (Miller et al.,1996): there are two major branches of collecting vessels—one running upward (cranially) along the anterior interventricular septum and the other running on the right side of the conus and subsequently between the aorta and the pulmonary trunk and further in the mediastinum opening at the venous angle. A similar route of major lymphatic trunks (collecting vessels) along the great arteries has been described in the chicken heart (Klika et al.,1972, Wilting et al.,2007). In the mouse heart there are two major collecting lymphatics: one running along the left conal vein and upward (cranially) along the great arteries and the second one on the diaphragmatic surface of the heart (on the atria and coronary sinus), collecting lymph from two or three minor vessels.

To identify whether the walls of collecting lymphatic vessels of the embryonic mouse hearts contain mural cells (bearing SMA antigen) we performed double labeling (anti-Lyve-1 and anti-SMA). Our data indicate that lymphatic vessels of the embryonic and early postnatal hearts do not contain smooth muscle cells (tunica media), except some segments of vessels positioned between the aorta and the pulmonary trunk, which are partially surrounded by smooth muscle cells. In our study, only some segments of these vessels contained smooth muscle cells positioned at a distance from the lymphatic endothelium, similar to the observations of Sacchi et al. (1999). The absence of mural cells in collecting lymphatics of the embryonic mouse heart is consistent with previous findings by Landsberger and Heym (1974), who did not find smooth muscle cells in the wall of lymphatic vessels of the human fetal heart. Thus, differentiation of lymphatic vessels into smooth-muscle–covered lymphatic trunks and ducts occurs presumably at later stages of development. There might be also differences in the structure of collecting lymphatics between species: in murine hearts collecting vessels might be devoid of smooth muscle cells and further (downstream) sections of collecting lymphatics with a smooth-muscle-cell mantle may occur in other organs besides the heart (Adams and Alitalo,2007), whereas in adult human hearts the collecting lymphatics contain smooth muscle cells. Of interest, there is a possibility of the occurrence of ectopic SMA-containing cells around the lymphatic capillaries in certain pathologic states (Adams and Alitalo,2007).

Lymphangiogenic Patterns

Our morphological observations indicate that several patterns of lymphangiogenesis are present during heart development. The presence of separate cells bearing the Lyve-1 antigen that are arranged in a row would suggest that from this orientation a tubule is easily formed by the coalescence of these cells and lumen formation. This mechanism of lymphatic vessel formation would proceed by vasculogenesis (in situ formation of primordial vessels) similar to that of growing coronary blood vessels (Rongish et al.,1994; Kattan et al.,2004). Simple sprouting of existing tubules observed in our study resembles the angiogenic sprouting described previously by Risau (1997).

Elongation of lymphatic tubules might occur by incorporation of mesenchymal cells between growing segments of Lyve-1–positive tubules or by coalescence of existing tubules (Lyve-1–positive) by their extension (elongation) and proliferation of endothelial cells. These latter steps proceed also by mechanisms similar to blood vessel angiogenesis.

Another of our observations is that Lyve-1–positive cells become ramified, subsequently seem to approach the existing lymphatic tubules and come into direct contact with them. Klika et al. (1972) suggested a theory on the formation of lymphatic capillary bed in the heart. Mesenchymal cells would simply circumscribe the lumen of the lymph capillary and form an integral part of its wall. In other words, according to Klika et al. (1972) major developmental mechanism of lymphangiogenesis occurs by incorporation of mesenchymal cells into coronary lymphatics. We do not know at what stage preceding lumen formation these mesenchymal cells described by Klika et al. become committed to being LECs (bearing Lyve-1 antigen). Because Lyve-1 is the earliest marker of lymphatic endothelial cells (Sleeman et al.,2001; Wigle et al.,2002; Buttler et al.,2006), we can speculate that these mesenchymal cells are already visualized by our method. Thus, our observations would confirm that one of the lymphangiogenic mechanisms is an addition of separate cells committed to lymphatic lineage to existing tubules.

Lymphangiogenesis may also start from single cells that make arborizations similar to nerve cell outgrowths. Such ramified cells are found in the openings of the existing lymphatic lattice. The cellular processes seem to approach the adjacent cellular wall and coalesce forming a branched lymphatic lattice. Such cells were observed on the surface of 12–17 dpc hearts and at later stages of postnatal development. Apparently the presence of single Lyve-1–positive cells on the surface of the heart on postnatal days suggests that lymphatic vessel formation is not completed during the ontogenic life and continues after birth.

In summary, lymphangiogenesis in developing heart proceeds by the same mechanisms as blood vessel angiogenesis: by in situ formation of tubes and vessels, vessel elongation, coalescence of existing tubules, branching of existing tubules by making additional connections within openings of the lymphatic lattice. Addition of single mesenchymal cells which acquire the EC antigen to a forming tubule or to an existing tubule is characteristic for lymphatic vessel formation and occurs also during blood vessel formation (Wilting et al.,1995).

Limitations of the Study

The use of Lyve-1 antibodies as a marker of lymphatic endothelium of the heart is based on literature reports dealing with this marker's specificity (Jackson,2003). Specificity of the Lyve-1 marker is, however, not definite. It is known, for example, that Lyve-1 is expressed in some activated macrophages and in the sinusoidal endothelium of the liver and spleen, where high-molecular-weight hyaluronian is absorbed and degraded (Jackson et al.,2001; Jackson,2003). Lyve-1 antibodies have been demonstrated to label some developing endothelial cells of the venous system (umbilical vein endothelial cells) in other organs like liver sinusoids, spleen, placental syncytiotrophoblast (Sleeman et al.,2001). In our study, Lyve-1 has been expressed in endothelial cells of embryonic veins entering the venous pole of the heart. The expression of Lyve-1 by certain embryonic venous endothelial cells may indicate that either lymphatic vessels derive from venous endothelium (Kato et al.,2006) or venous and lymphatic endothelium share some common antigenic determinants (Lyve-1, or else?). Another unexplained issue is labeling of endothelial cells of the endocardium (for example in atria and some segments of ventricles) with Lyve-1 antibodies, which we observed in our study. This may indicate that different populations of endothelial cells which bear the Lyve-1 antigen are present in the mouse embryonic heart. Further studies are necessary to characterize antigenic expression in various endothelial cell populations of the embryonic heart. These studies would be useful in elucidation of the issue of possible source, differentiation, and interrelationship between various endothelial cell compartments: venous, arterial, blood capillary, lymphatic, and endocardial endothelium.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Animals

Balb/c mouse embryos from embryonic day (ED) 11 to 18 and PD 1–28 were studied. All procedures were approved by Medical University of Warsaw II Local Bioethics Committee. Embryos were removed from the uteri of cadaveric dams (killed by an overdose of ketmine-xylazine) and the hearts were dissected. Subsequently hearts were either frozen immediately in 2-methylobuthane cooled in an acetone–dry ice mixture or fixed in 4% paraformaldehyde in PBS. Paraformaldehyde-fixed hearts were treated according to the whole-mount method, or processed for routine paraffin-embedded specimens.

Immunohistochemistry on Serial Tissue Sections

Frozen sections were serially cut, fixed in cold acetone, and stained with Lyve-1 antibodies (R&D). The secondary antibody was biotin-conjugated mouse anti-rat IgG (Pharmingen). Subsequently, sections were incubated with Streptavidin-peroxidase complex (DAKO). Diaminobenzidine (DAB) was used as chromogen (Sigma). For double-staining the sections were treated with anti-Lyve-1 antibodies (developed with DAB) and with anti–α-smooth muscle actin (SMA, DAKO) antibodies (developed with alkaline phosphatase substrate (NBT/BCIP, DAKO). Paraffin sections were deparaffinized, rinsed with phosphate buffered saline (PBS), treated in a microwave oven 2 × 5 min in citric acid buffer (to retrieve antigenic determinants), and incubated overnight with anti-prox-1 antibodies (Angiobio), then, after being rinsed with PBS, the sections were treated with anti-rabbit IgG biotin conjugate (DAKO), rinsed again and treated with streptavidin–HRP complex (DAKO). The color reaction was developed with DAB (Sigma).

Whole-mount Immunohistochemistry

Hearts fixed in paraformaldehyde were treated according to Ma et al. (1998): in brief, the hearts were rinsed 5 times for 1 hr with PBS-MT (phosphate buffered saline containing 2% non-fat milk and 0.1% Triton-X, incubated overnight with anti-Lyve-1 antibodies diluted with PBS-MT, subsequently rinsed with PBS-MT as above, incubated overnight with biotin-conjugated mouse anti-rat IgG antibodies. The rinse steps were repeated and incubation with Streptavidin–peroxidase complex was performed overnight. After rinsing with PBS-MT the reaction was developed with DAB substrate with metal enhancement (Sigma). The hearts were rinsed with PBS-T, fixed with 50% methanol in PBS and processed with the use of increasing concentrations of glycerol (50%–70%–100%).

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

We thank Anna Podbielska for her excellent technical help. We also thank Prof. O. Olszowska, M. Sztyber, MS, and W. Szypuła, MS, from Department of Biology and Pharmaceutical Botanics, the Medical University of Warsaw, for the use of their stereomicroscope and their generous help in photography of whole-mount labeled hearts. Authors thank two anonymous reviewers whose pertinent comments helped to improve this manuscript. M.J. was a recipient of a “Minigrant” funded by Postgraduate School of Molecular Medicine at the Medical University of Warsaw.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES