- Top of page
- MATERIALS AND METHODS
- LITERATURE CITED
Morris (J. Anat., 1976;121:47–64) proposed that the subepicardial mesenchyme might represent a continuing source of myocardioblasts during embryonic and fetal development. Recent studies have shown that the epicardium and subepicardial mesenchyme, and the coronary vasculature are all derived from a region of the pericardial wall, called the proepicardial serosa. In avian embryos, the cells from the proepicardial serosa colonize the heart via a secondary tissue bridge formed by attachment of proepicardial villi to the heart. In the present study, Morris's hypothesis was tested by tracing the fate of the proepicardial serosa. This was achieved by constructing quail-chick chimeras. The proepicardial serosa was transplanted from HH16/17 quail embryos to HH16/17 chick embryos (ED3). A new transplantation technique facilitated an orthotopic attachment of the quail proepicardial villi to the chicken heart, and prevented the attachment of the chicken proepicardial villi to the heart. The fate of the grafted quail cells was traced in chimeras from ED4 to ED18 with immunohistochemistry, using quail-specific antibodies (QCPN, QH-1). From ED4 onward, the transplant was connected to the dorsal heart wall via its proepicardial villi. Starting from the point of attachment of the quail proepicardial villi to the heart, the originally naked myocardium became almost completely covered by quail-derived epicardium, and quail mesenchymal cells populated the subepicardial, myocardial, and subendocardial layers including the av-endocardial cushions. Quail cells formed the endothelial and smooth muscles cells of the coronary vessels, and the perivascular and intramyocardial fibroblasts. Quail myocardial cells were never found in the subepicardial, myocardial, and subendocardial layers. This suggests that the subepicardial mesenchyme normally does not contribute a substantial number of myocardioblasts to the developing avian heart. The new transplantation technique presented facilitates the production of chimeric hearts in which the derivatives of the proepicardial serosa are almost completely of donor origin. Thistechnique might be useful for future studies analyzing the role of certain genes in cardiac development by the creation of somatic transgenics. Anat Rec 255:212–226, 1999. © 1999 Wiley-Liss, Inc.
The early embryonic heart tube consists of two distinct cell layers, an inner endothelial tube in contact with the blood (primitive endocardium), and an outer contracting epithelium in contact with the pericardial fluid (primitive myocardium).
Up to the late 1960s, it was thought that these two epithelia gave rise to the majority of structures found in the mature heart. The primitive myocardium was thought to provide not only the definitive myocardium but also the epicardium, the mesenchyme and coronary vessels of the subepicardial layer, and the intramyocardial blood vessels subjacent to the subepicardium (e.g., Kölliker, 1879; Streeter, 1945; DeHaan, 1965). The primitive endocardium was thought to provide the definitive endocardium, the mesenchyme of the subendocardial layer, and the endothelium of the intramyocardial blood vessels subjacent to the subendocardium (e.g., Grant, 1926; Voboril and Schiebler, 1969; Rychterová, 1977). The only structures that were thought to be derived from extracardiac sources were the peripheral nervous system of the heart and the endothelium of the main stems of the coronary vessels. The latter were thought to arise from endothelial sprouts growing out from the aorta (coronary arteries) and the sinus venosus (coronary veins) into the subepicardial layer where they united with the forerunner of the distal portions of the definitive coronary circulation, a primitive vascular plexus formed by the local mesenchyme (Grant, 1926; Goldsmith and Butler, 1937; Ogden, 1968; Conte and Pelligrini, 1984).
During the past 3 decades, these views on the origin of the cellular elements of the mature heart have changed remarkably. The initial stimulus for these changes came from Manasek's study of myocardial development in the early chick embryo (Manasek, 1968). He reported that the primitive myocardium consisted of myocardioblast only and, therefore, could not be the source of the epicardium, the subepicardial mesenchyme, and the subepicardial blood vessels.
Since then numerous studies have been performed to clarify the origin of the cells constituting these tissues. The results suggest that the epicardium (Shimada and Ho, 1980; Virágh and Challice, 1981; Komiyama et al., 1987; Hiruma and Hirakow, 1989; Hirakow, 1992; Männer, 1992, 1993), the mesenchymal cells of the subepicardial layer (Van den Eijnde et al., 1995; Pérez-Pomares et al., 1997), the endothelium and smooth muscle cells of the coronary vasculature (Langford et al., 1990; Mikawa and Fishman, 1992; Poelmann et al., 1993; Mikawa and Gourdie, 1996; Dettman et al., 1998), the perivascular and intermyocardial fibroblasts (Mikawa and Fishman, 1992; Dettman et al., 1998), and some cells in the atrioventricular valves (Gittenberger-de Groot et al., 1998) are all derived from a primarily extracardiac source, namely from the coelomic wall covering the sinus venosus with its adjacent portion of the septum transversum (mammals) or liver anlage (chick/quail). In the following, this region of the coelomic wall will be called the “proepicardial serosa.”
During the formation of the epicardium, the epithelium of the proepicardial serosa is characterized by the formation of villous protrusions. In avian embryos, these villi attach to the opposite surface of the heart, establishing a secondary dorsal mesocardium (Männer, 1992). Descriptive and experimental studies suggest that this tissue bridge normally serves as the main route for the transfer of extracardiac cells from the proepicardial serosa to the developing heart. Descriptive studies have shown that the formation of the epicardium and the subsequent formation of the subepicardial mesenchyme and subepicardial vascular plexus normally proceed from the point of attachment of the secondary dorsal mesocardium (Männer, 1992). Experimental prevention of the attachment of the proepicardial villi to the developing chick heart is followed by a severe delay in the formation of the epicardium and by extreme underdevelopment of the subepicardial mesenchyme (Männer, 1993).
Despite this recent progress of our knowledge on structural development of the embryonic heart we found that, in the past, little attention was focused on the origin of clinically important myocardial structures found in the subepicardial layer. These are subepicardial pathways of accessory atrioventricular conductance (Langberg et al., 1993; Milstein et al., 1997), and myocardial bridges over the subepicardial branches of the coronary arteries (Angelini et al., 1983; Yamaguchi et al., 1996). Morris (1976) reported on evidence that subepicardial mesenchymal cells could differentiate into myocardioblasts. He suggested that the subepicardial mesenchyme might represent a continuing source of myocardioblasts during embryonic and fetal development. Based on this hypothesis and on the fact that the subepicardial mesenchyme is derived from the proepicardial serosa (Van den Eijnde et al., 1995; Pérez-Pomares et al., 1997), one might speculate that some myocardial cells of the mature heart, especially those found in the subepicardial layer, could possibly be derived from the primarily “extracardiac” mesenchyme of the proepicardial serosa. Indeed, this idea seems to be supported by the occurrence of myocardial cells in cultures of proepicardial serosa (Langford et al., 1990) and in secondary mesocardia subjected to mechanical stress (Männer, 1993), as well as by some electron microscopic findings in mouse embryos suggesting the transformation of mesenchymal cells from the region of the proepicardial serosa into myocardial cells of the ventrocaudal wall of the sinus venosus. (Virágh and Challice, 1973). Furthermore, there is evidence for the possibility that even fibroblasts from adult hearts can differentiate into myocardial cells (Eghbali et al., 1991).
The aim of the present study was to establish whether the subepicardial mesenchyme normally contributes myocardial precursor cells to the developing chick heart. To answer this question, the fate of the extracardiac primordium of the subepicardial mesenchymal cells was traced by the quail-chick chimera technique. Pieces of the coelomic wall carrying the proepicardial serosa were transplanted from quail embryos to the pericardial cavity of chick embryos. Thereby, a new transplantation technique was used facilitating the production of chicken hearts almost completely covered by epicardium and subepicardial mesenchyme derived from the quail-donor. The fate of the grafted quail cells was traced with immunohistochemistry, using quail-specific antinuclear and antiendothelial antibodies.
- Top of page
- MATERIALS AND METHODS
- LITERATURE CITED
The present fate mapping study was performed to establish whether the subepicardial mesenchyme represents a continuing source of myocardioblasts during the embryonic and fetal periods (Morris, 1976). For this purpose, quail-chick chimeras were produced by grafting the extracardiac source of the epicardium and subepicardial mesenchyme—called the proepicardial serosa—from quail- to chick embryos. The quail-chick chimera technique was used for “labeling” the proepicardial serosa since it facilitates a long-lasting “labeling” of the entire population of transplanted cells, by quail-specific cell markers (Le Douarin et al., 1996), and, additionally, excludes the possibility of unintentional labeling of the original myocardium. Furthermore, the analysis of quail-chick chimeras can relatively simply be performed by use of the quail-specific monoclonal antibody QCPN, which has recently been proven to facilitate immunohistochemical staining of the quail cell nuclei in embryos from early to late stages of development (Huang et al., 1997; Gittenberger-de Groot et al., 1998; Bertossi et al., 1998; Burns and Le Douarin, 1998).
Classical in vivo labeling techniques with the injection of vital dyes into the area of interest do not lead to a long-lasting cell labeling, since the dyes dilute by successive cell divisions. Recent labeling techniques using replication incompetent retroviruses encoding β-galactosidase have the advantage of permanent labeling of the infected cells (Mikawa et al., 1992a,1992b; Hyer and Mikawa, 1997). It seems, however, that this technique does not facilitate labeling of the entire cell population of the proepicardial serosa. After in vitro inoculation of virus, only 5–10% of the cells in cultured explants of proepicardial villi expressed β-galactosidase (Mikawa and Gourdie, 1996). Moreover, in cases of in ovo labeling with replication incompetent retroviruses, it might be impossible to rule out that the infection of a given cell has resulted from potential leakage of virus from the injection site (Mikawa and Gourdie, 1996).
Quail-chick chimeras have previously been constructed for tracing the fate of the proepicardial serosa (Poelmann et al., 1993; Vrancken Peeters, 1997; Dettman et al., 1998; Gittenberger-de Groot et al., 1998; Pérez-Pomares et al., 1998). In these studies, the proepicardial serosa of the donor was loosely placed into the pericardial cavity of the host, and the normal transfer of the cells from the host proepicardial serosa to the heart was not prevented. Thus, the hearts produced by this transplantation technique received only a limited contribution of cells from the transplant to the epicardium, the subepicardial mesenchyme and the coronary vessels. Although the analysis of such hearts has provided some important new insights into the structural development of the embryonic heart, they are of limited use for a study trying to prove Morris's (1976) hypothesis. For such a study it would be more convenient to have chicken hearts almost completely covered by quail epicardium and quail subepicardial mesenchyme.
There are some practical problems in constructing such hearts. Surgical removal of the proepicardial serosa causes lethal bleeding of the embryo, since the removed tissue constitutes a part of the ventral wall of the sinus venosus. Thus, the proepicardial serosa cannot be removed from the chick-host, and the transplant from the quail-donor cannot be implanted into the coelomic wall at the site of the proepicardial serosa of the chick-host. The practical problems to be resolved when constructing chicken hearts almost completely covered by quail epicardium and quail subepicardial mesenchyme are (1) to find a method to prevent or minimize the transfer of cells from the proepicardial serosa of the chick-host to the heart, and (2) to find a method to anchor the quail-transplant in a position within the chicken pericardial cavity enabling the transfer of cells from the quail proepicardial serosa to the chick embryo heart by a pathway close to the normal one.
In the present study, these problems were resolved by use of a new transplantation technique. This transplantion technique is a modification of an experimental technique previously used to prevent the normal transfer of cells from the proepicardial serosa to the heart (Männer 1993). In this previous experiment, the attachment of the villi of the proepicardial serosa to the heart was blocked by the placement of a small piece of the egg shell membrane between the villi and the heart. In the present study, the piece of the egg shell membrane was mounted with a graft of quail proepicardial serosa prior to its implantation into the pericardial cavity of a chick embryo (Fig. 1A–C). This membrane then served not only as a barrier for blocking the attachment of the chicken proepicardial villi to the heart, but simultaneously served as a carrier for the quail-transplant enabling orthotopic attachment of its proepicardial villi to the dorsal wall of the chick embryo heart (Fig. 1D).
This new transplantation technique facilitated the production of chicken hearts almost completely covered by quail epicardium and quail subepicardial mesenchyme (Fig. 1C,E,F). These hearts additionally showed an extensive invasion of quail mesenchymal cells into the myocardial and subendocardial layers. Quail cells constituted all the structures previously found to be of proepicardial origin, but quail myocardial cells were never found in the subepicardial, myocardial, or subendocardial layers nor in the cardiac conduction system (Fig. 6A,C). This suggests that the subepicardial mesenchyme normally does not contribute a substantial number of myocardioblasts to the developing avian heart. However, since Morris (1976) performed his studies on human and rat embryos it cannot be excluded that the subepicardial mesenchyme contributes some myocardial cells to the mammalian heart. It is a challenge for the future to find an experimental approach to test Morris's (1976) hypothesis in mammalian embryos.
If subepicardial mesenchymal cells do not differentiate into myocardioblasts then the question remains of how myocardial strands found within the connective tissue of the subepicardial layer are being formed. These myocardial strands form subepicardial pathways of accessory atrioventricular conductance (Langberg et al., 1993; Milstein et al., 1997), myocardial bridges over the subepicardial branches of the coronary arteries (Angelini et al., 1983; Yamaguchi et al., 1996), and myocardial strands in secondary mesocardia (Lechleuthner and Dominiak, 1990; Männer, 1993).
Subepicardial Pathways of Atrioventricular Conductance
Accessory pathways of atrioventricular conductance are usually localized deep in the atrioventricular groove close to the fibrous annulus, but sometimes they are also found subepicardially traversing the atrioventricular groove distant from the fibrous annulus (Langberg et al., 1993; Milstein et al., 1997). In the past, accessory atrioventricular muscle bundles were commonly regarded as abnormally persisting remnants of the embryonic atrioventricular junctional myocardium (Truex et al., 1960). Recent data from human embryos, however, indicate that the former area of the embryonic atrioventricular junctional myocardium lies deep in the atrioventricular groove of the mature heart, and that there are no myocardial bundles present in the subepicardial mesenchyme of the embryonic atrioventricular groove (Wessels et al., 1996). Therefore, it has been concluded that subepicardial atrioventricular muscle bundles are secondary structures, “acquired” during fetal and/or postnatal life (Wessels et al., 1996). Considering the data from the present study it is unlikely that these myocardial structures are acquired by differentiation of subepicardial mesenchymal cells into myocardiocytes. Recent data from quails and humans have demonstrated that the proximal part of the tunica media of the subepicardial coronary veins is formed by atrial myocytes, and that the formation of the coronary venous media is a relatively late process in cardiac development (Vrancken Peeters, 1997; Vrancken Peeters et al., 1997b). In the present study, we observed small subepicardial veins traversing the atrioventricular groove of hearts from older chick embryos/fetuses (Fig. 8B,C). The fact that the media of these veins sometimes consisted almost completely of host cells, continuous with the atrial and ventricular myocardium suggests that subepicardial pathways of atrioventricular conductance might be formed by myocardial cells migrating from the atrial myocardium towards the ventricular myocardium along the endothelial tube of subepicardial coronary veins. Another source might be ventricular myocardial cells migrating into the subepicardial layer in close association to coronary arteries and veins (Fig. 9B,C). Both possibilities would correspond to clinical observations suggesting relations between accessory pathways of atrioventricular conductance and some morphological features of the coronary sinus and the cardiac veins (Sealy, 1994; Schumacher et al., 1995).
Myocardial Bridges Over the Subepicardial Branches of the Coronary Arteries
Myocardial bridges over the subepicardial branches of the coronary arteries are a frequent anatomical finding in human beings and other mammals (for reviews see Angelini et al., 1983; Yamaguchi et al., 1996). Clinical studies have extensively focused on this subject since myocardial bridging is suspected of exerting an ischemic effect on the myocardium. In the past, myocardial bridges were regarded as congenital and this idea seems to be supported by a study on human embryos where myocardial bridges were observed in 30% of the embryos studied (Vrancken Peeters, 1997). Recent angiographic findings, however, suggest that myocardial bridging might also be acquired postnatally (Vongpatanasin et al., 1997). Unfortunately, there is no information available as to how myocardial bridges, whose ultrastructure deviates somewhat from that of the common myocardium (Yamaguchi et al., 1995), are formed during embryonic or postembryonic development. The data from the present study suggest that it is unlikely that these myocardial structures arise from the diffferentiation of subepicardial mesenchymal cells into myocardial cells. It remains to be established whether myocardial bridges are primary structures that became “isolated” from the myocardial wall by an intramyocardial formation of segments of the subepicardial coronary arteries, or whether they are secondary structures, formed by the migration of myocytes over primarily subepicardial segments of the coronary arteries. The occurrence of subepicardial strands of ventricular myocardium close to subepicardial branches of coronary arteries suggests the latter possibility (Fig. 9B,C).
Myocardial Strands in Secondary Mesocardia
Ligament-like tissue bridges traversing the free pericardial cavity between the pericardial wall and the ventricular part of the heart are a normal and frequent finding in non-mammalian vertebrates (Grant and Regnier, 1926). They may contain blood vessels as well as some myocardial strands (Lechleuthner and Dominiak, 1990). Embryological studies have shown that these tissue bridges are secondary structures, formed during the embryonic period by the attachment of villous protrusions of the pericardial epithelium to the developing ventricles (Greil, 1903; Männer, 1992). Actually, the best known example for these “secondary mesocardia” is the secondary dorsal mesocardium of avian embryos, which is formed by the villi of the proepicardial serosa. In a previous experimental study on chick embryos it was found that myocardial strands were formed within the primarily mesenchymal core of a secondary mesocardium if this tissue bridge was subjected to mechanical stress (Männer, 1993). The origin of these newly formed myocardial strands could not be established and, therefore, it was suspected that they arose either from myocardial differentiation of local mesenchymal cells or from the immigration of myocardial cells from the myocardial wall. The data from the present study indicate not only that it is unlikely that myocardial strands of secondary mesocardia arise from local mesenchymal cells, the data, furthermore, give conclusive evidence that these structures are formed by the emigration of myocardial cells from the myocardial wall into the mesenchymal core of a secondary mesocardium (Fig. 9A).
Taken together it can be stated that the data from the present study suggest that myocardial strands found in the subepicardial layer of the mature heart do not arise from myocardial differentiation of local mesenchymal cells but rather from myocardial cells emigrating from the myocardial layer into the subepicardial connective tissue. It remains to be established what factors might trigger these migratory processes.
Further Implications for the Understanding of the Histogenesis of the Heart
Although the present study was conducted primarily to show whether the subepicardial mesenchyme contributes myocardioblasts to the developing heart, it has also some further implications for the understanding of the structural development of the heart. The new transplantation technique presented facilitates the production of chimeric hearts in which the structures derived from the cells of the proepicardial serosa are almost completely derived from the implanted donor tissue. This transplantation technique might be useful not only for conventional fate mapping studies but also for studies analyzing the role of certain genes in cardiac histogenesis by the creation of somatic transgenics (Hyer and Mikawa, 1997).
With respect to the fate of the proepicardial serosa, the present study has confirmed not only the results of previous investigations, it has also added some new findings that might be helpful in elucidating the structural development of the outflow tract of the embryonic heart. Up to now, it was thought that the myocardial portion of the embryonic outflow tract became completely ensheathed by proepicardium-derived epicardium, whereas the epicardium of the mesenchymal portion of the embryonic outflow tract was thought to be derived from the pericardial epithelium at the junction between the outflow tract and the dorsal wall of the pericardial cavity (Noden et al., 1995). The distal boundary of the proepicardium-derived outflow tract epicardium, therefore, was thought to correspond to the distal boundary of the original outflow tract myocardium (Noden et al., 1995), which, in turn, was considered to be a stable anatomical landmark indicating the level of the future arterial orifices (Bartelings and Gittenberger-de Groot, 1988). The present results, however, show that only the proximal portion of the original outflow tract myocardium becomes ensheathed by proepicardium-derived epicardium. The distal portion of the original outflow myocardium becomes ensheathed by an epicardial layer of different origin continuous with the serosal epithelium at the junction between the outflow tract and the dorsal wall of the pericardial cavity (Fig. 2C). This finding, which matches completely observations made in a previous experimental study on chick embryos (see Fig. 8B in Männer, 1993), suggests that the distal portion of the original outflow tract myocardium becomes ensheathed by primarily “pericardial” epithelium reaching the outflow tract via its junction with the dorsal wall of the pericardial cavity. The present findings, furthermore, show that the topographical relationship between the outflow tract myocardium and the two different outflow tract epicardia do not remain stable but change during a period spanning the time between completed formation of the epicardium on ED6 and completed formation of the semilunar valves on ED8. During this period, the distal boundary of the outflow tract myocardium shows a proximal shift with respect to the distal boundary of the proepicardium-derived outflow tract epicardium, so that the two boundaries finally meet at the level of the orifices of the semilunar valves (Figs. 2C,E,F, and 7A). This finding casts doubt on the view that the distal boundary of the embryonic outflow tract myocardium is a stable anatomical landmark (Bartelings and Gittenberger-de Groot, 1988; Noden et al., 1995). It merely seems to be in accord with other findings on chick and rat embryos suggesting a regression of the distal portion of the outflow tract myocardium before development of the semilunar valves, possibly by transdifferentiation (Argüello et al., 1978; Yu et al., 1998) and/or apoptosis (Watanabe et al. 1998) of its myocardial cells. In vitro studies have shown that the phenotype of myocardial cells can be modified by cell-cell interactions with epicardial cells (Eid et al., 1992). It would be interesting to check whether the regression of the distal outflow tract myocardium might be related to the origin of its epicardial covering.