Initiation of apoptosis in the developing avian outflow tract myocardium



Apoptosis occurs within the cardiac outflow tract (OFT) myocardium during normal development of chick hearts. This peak of apoptosis occurs at stage 30–31 and coincides with dramatic remodeling of the OFT, suggesting that apoptosis occurs to allow proper alignment of the great vessels over their respective ventricles. The signals that initiate apoptosis in this setting are unknown. The aim of this study was to characterize the cells undergoing apoptosis in the cardiac OFT myocardium and the cells that may influence this process. Two cell populations that may initiate apoptosis of the cardiomyocytes are the cardiac neural crest (CNC) cells and epicardial cells. We examined stage 30–31 chick embryos that had undergone removal of the CNC cells or had delayed epicardial growth for alterations of apoptosis. Removal of the CNC cells did not reduce the levels or pattern of apoptosis in the OFT myocardium. In contrast, impeding the growth of the epicardium over the OFT resulted in a 57% reduction in apoptotic cells in the OFT myocardium. Analysis of the apoptotic cells within the OFT myocardium showed that as many as 92% of them expressed cardiomyocyte markers. In the quail, the endothelial marker QH1 identified a component from the epicardium, endothelial cells, in regions where apoptosis is elevated in the OFT myocardium. These results suggest that a component from the epicardium, possibly endothelial cells, is required for the initiation of apoptosis in OFT cardiomyocytes. © 2002 Wiley-Liss, Inc.


Considerable remodeling occurs in the OFT during the transformation of the developing heart from a tubular embryonic structure to the four-chambered mature heart (Goor et al., 1972; de la Cruz et al., 1977; Thompson et al., 1987). This remodeling coincides with high levels of apoptosis in the OFT myocardium (Pexieder, 1973; Hurle and Ojeda, 1979; Watanabe et al., 1998). When levels of apoptosis were altered in chick embryos by using locally applied inhibitors or enhancers of apoptosis (Watanabe et al., 2000, 2001) defects in OFT alignment and orientation resulted. This finding suggests that apoptosis is a critical process required for normal developmental remodeling of the heart, and that defects in the initiation and regulation of apoptosis in the OFT may explain a subset of human congenital conotruncal heart defects. The goal of the present study was to examine the cell types undergoing apoptosis in OFT myocardium and identify those cell types that may be required for this process to proceed.

In the chick embryo, a subset of neural crest cells has been identified that migrate into the developing heart and contributes to normal cardiac development. These cells originate from the area between the mid-otic placode and somite 3 and migrate to the heart after stage 10 (Kirby et al., 1983; Poelmann et al., 1998). Within the OFT, one group of these cells, termed CNC cells, enters the OFT at stage 22 (Waldo et al., 1999) where they will contribute to the mesenchyme of the aorticopulmonic septum, which separates the aortic and pulmonic OFTs. CNC cells have been reported to undergo apoptosis themselves (Poelmann et al., 1998). Their absence results in cardiac defects that resemble those observed in the human condition of persistent truncus arteriosus in which the OFT septum fails to develop (Kirby et al., 1983, 1985).

Simultaneously, another population of CNC cells migrates through the epicardium of the OFT myocardium and gathers at the base of the OFT where they will become the anterior parasympathetic plexus of the heart (Kirby and Stewart, 1983; Verberne et al., 1998). This group of migrating neuroblasts is found at the epicardial surface of the OFT myocardium at stage 24, approximately the same time as the onset of apoptosis of the OFT cardiomyocytes (Kirby et al., 1980). Therefore, the CNC cells may be responsible for initiation of apoptosis in the OFT cardiomyocytes, as suggested by their presence on both endocardial and epicardial surfaces of the OFT myocardium at the time and place cardiomyocyte apoptosis occurs.

The epicardium develops from a grape-like cluster of serosal cells that lies between the liver and the dorsal heart (Manasek, 1969a; Ho and Shimada, 1978; Hiruma and Hirakow, 1989). These cells contact the dorsal area of the heart near the atrioventricular sulcus by stage 17 and completely cover the heart by stage 27 (Hiruma and Hirakow, 1989; Viragh et al., 1993, Manner, 1999). They reach the proximal OFT myocardium at approximately stage 21, before the onset of OFT cardiomyocyte apoptosis. By virtue of the timing and proximity, cells of the epicardium as well as CNC cells are potential initiators of OFT cardiomyocyte apoptosis.

The term “epicardium” is usually used to refer to the outermost single squamous epithelial layer of mesothelial cells and the underlying connective tissue layer adjacent to the myocardium of the mature heart. The newly formed epicardium differs markedly in size and complexity from the mature epicardium, but nonetheless also has a single layer of mesothelial cells and an underlying connective tissue layer (Hiruma and Hirakow, 1989). “Epicardial cells” will be used to refer to any cell within the epicardium. These cells include (1) the mesothelial cells; (2) those mesenchymal cells within the connective tissue that may include cells delaminating from the mesothelial cell layer (termed EPDCs; Gittenberger-de Groot et al., 1998), which include the precursors of the smooth muscle cells and fibroblasts; and (3) other cells within the connective tissue layer of the epicardium including angioblasts, endothelial cells and neural crest cell derivatives. Some of these epicardial cells, (smooth muscle cells, fibroblasts, endothelial cells and/or their precursors) are known to invade the myocardium. We will refer to these as “cells from the epicardium.”

In this study, we tested whether the absence of CNC or the epicardium would result in a reduction of apoptosis in the OFT myocardium. Apoptosis was assessed by using the terminal deoxynucleotidyl transferase–mediated dUTP nick end-labeling (TUNEL) assay at stages when apoptosis of OFT myocardial cells normally peak. A quantitative analysis of apoptotic OFT cells in the normal chick heart was performed to determine what proportion of cells undergoing apoptosis in the OFT myocardium is cardiomyocytes.


Cardiomyocyte Apoptosis in the Chicken Embryo OFT Occurred in the Absence of CNC

OFT myocardial apoptosis in the developing chicken embryo peaks at stages 30–31, when ventricular septation is complete (Watanabe et al., 1998). Therefore, these stages were chosen for observation to maximize the likelihood of detecting inhibitory effects on OFT myocardial apoptosis in the experimental embryos. A total of 14 embryos between the HH stages of 29 and 32 was evaluated. Five sham-operated embryos were compared with three cardiac crest-ablated embryos and six CNC and nodose placode-ablated embryos (termed “doubly ablated”).

Because OFT septation was not complete at the time the embryos were harvested for this analysis, the presence of OFT septation defects could not be assessed. However, it is known that embryonic chick hearts operated by this method undergo abnormal cardiogenesis, with 90% resulting in failure of septation of the OFT (Kirby, 1999) resembling persistent truncus arteriosus. Based on this experience of the laboratory that provided the embryos for our study, we estimated that 90% of the experimental embryos lacked sufficient CNC for normal septation.

The ablation of CNC did not reduce cardiomyocyte apoptosis in the OFT. Embryos that had undergone bilateral CNC ablation alone or in combination with ablation of the nodose placode displayed a diffuse level and pattern of cardiomyocyte apoptosis similar to sham operated controls of the same stage (Fig. 1). The number of TUNEL-positive cells in a set volume of OFT myocardium in control embryos (n = 3) was 19.3 ± 7 (standard deviation), the number in the same volume of OFT myocardium in experimental embryos (n = 5) was 18.7 ± 15 (P = 0.914), with no statistical difference between the two groups. A subset of the doubly ablated embryos (2/6) had a qualitatively larger area of TUNEL-positive myocardium than that in either normal embryos or in the CNC-ablated groups (data not shown). The presence of large areas of apoptosis in the CNC-ablated animals supports our findings that the majority of cells that undergo apoptosis in the OFT myocardium are not CNC cells (because these were not present), but rather cardiomyocytes. Interestingly, all ablated embryos continued to show prominent apoptosis within the cushion mesenchyme of the OFT. This finding suggests that cell types other than CNC cells are undergoing apoptosis within these cushions.

Figure 1.

Apoptosis within the cardiac outflow tract (OFT) myocardium is not reduced in the absence of cardiac neural crest (CNC) cells nor in the doubly ablated embryos. A: Low-magnification view of a sagittal section through a normal stage 30 embryonic chick heart stained with hematoxylin and eosin. Arrows in this panel delineate the proximal and distal OFT borders. B: The area enclosed in the box in A is enlarged to show the cushion mesenchyme (CM) and myocardium (My) of the OFT. The CNC-ablated embryos (C,D), the sham-operated control embryos (E,F), and the double-ablated embryos (G,H) were sagittally sectioned and immunostained with MF20 and anti-titin (C,E,G) to distinguish the myocardium (bright green regions). All exhibited generally similar levels and patterns of terminal deoxynucleotidyl transferase–mediated dUTP nick end-labeling (TUNEL) -stained cells (D,F,H) in the OFT myocardium, which were distributed in the diffuse (small black arrows) and focal (larger black arrows) patterns. Note the apoptotic foci of the myocardium that were negative or reduced in staining for MF20 and anti-titin immunostaining (white arrows in C, E, and G). These foci correspond to areas that were TUNEL-positive except the distal focus (white arrow with asterisk in G). The apparent difference in the specific site of the TUNEL-positive foci between the experimental and control embryos shown here is not significant, because this location varies even within control embryos. Scale bar = 500 μ in A, 250 μ in D (applies to B–H).

A subset of embryos from all three groups had apoptotic foci as observed in normal unoperated embryos (Fig. 1). Two of four sham-operated embryos, two of three CNC-ablated embryos, and three of six doubly ablated embryos exhibited foci. The foci were seen only during HH stages 30 and 31, as in normal embryos, and were located at the same sites as in normal embryos (detailed below). The location of the foci in both abnormal and normal embryos varied in position from embryo to embryo as will be described in more detail in a later section of the Results.

Absence of Epicardium Significantly Reduces Apoptosis

Delay of epicardial growth resulted in reduced apoptosis of the chick OFT myocardium. Stage 30–31 chick embryos that underwent implantation of egg shell membrane between the heart and proepicardial organ were examined for TUNEL reactivity. Six treated embryos with delayed epicardial growth similar to that seen in studies by Manner (1993) were analyzed along with three sham-operated controls. Initially, the surgical procedure led to a high incidence of heart and chest wall malformations thought to be due to the large size of the egg shell membrane used. These malformed embryos were excluded from the analysis. Experimental embryos that did not develop epicardial delay as assessed histologically were also excluded.

The hearts of the experimental embryos included in this analysis appeared similar to control hearts in size and external morphology when examined by stereomicroscopy as whole-mounts. In contrast, there was a dramatic reduction of TUNEL staining in sections of OFT myocardium in experimental embryos (Fig. 2). In regions lacking epicardial coverage, there were few or no TUNEL-positive cells. OFT myocardium at sites covered by epicardium contained higher numbers of TUNEL-positive cells. Control embryos had 101.3 ± 44.5 (SD) cells in a specified volume of the OFT myocardium (see Experimental Procedures section), whereas experimental embryos had 43.2 ± 15.1 cells in the same volume of myocardium (P = 0.0315). One of the six embryos in the epicardial delayed group developed foci of intense apoptotic activity, but within the myocardium that was covered by epicardium.

Figure 2.

The number of terminal deoxynucleotidyl transferase–mediated dUTP nick end-labeling (TUNEL) -positive cells was reduced in the cardiac outflow tract (OFT) myocardium of stage 30 embryos with delay of epicardium formation. Sagittal sections through the base of the OFT of sham-operated embryos (A–C) were compared with those of embryos in which epicardial growth over the OFT myocardium was delayed (D–F). The TUNEL assay (A,D) was used to detect the apoptotic cells in diffuse (small black arrows in A and D) and focal (large black arrow in A) arrangements. The OFT myocardium was delineated by immunofluorescent staining of the same sections by using cardiomyocytes markers (green staining in B, red staining in E). In sham-operated embryos, a thick and well-developed epicardium completely covered the OFT myocardium as seen in section B and by hematoxylin and eosin staining of an adjacent section (C). In the experimental embryos, the epicardial covering was absent over large regions of the OFT myocardium (below the asterisk in D,E,F). A–F: The ventricle is situated below or at the bottom of the photomicrograph. CM, cushion mesenchyme; My, myocardium; Epi, epicardium. The magnification for A–F is 120×.

By qualitative assessment, there was no difference between experimental and control embryos in the thickness of the OFT myocardium. However, there was an obvious difference in the thickness of the ventricular myocardium. Quantitative analysis of histologic sections revealed that the experimental embryos had significantly thinner ventricular myocardial walls compared with control embryos with both ventricles affected. The compact zone of the mural wall, measured near the anteroapical region of the right ventricle, underwent a 54.5% reduction in thickness in experimental embryos, whereas the trabeculated myocardium decreased by 53%, both statistically significant reductions (see Table 1). This is consistent with studies of the development of the mouse heart in which the ventricular wall was thinner when epicardial development was abnormal (Kwee et al., 1995; Wu et al., 1999), when the endothelial component of the epicardium was abnormal (Tevosian et al., 2000; Svensson et al., 2000), and in epicardially ablated quail embryos (Gittenberger-de Groot et al., 2000).

Table 1. Comparison of Mean Apoptotic Cell Number and Thickness of Ventricular Myocardium (±SD) in Stage 30 Epicardially Delayed Embryos Compared with Control Embryosa
 Control (n = 3)Experimental (n = 5) (%)P value
  • a

    Apoptotic cells were counted within a 300 μ × 150 μ × 125 μ region. The same volume of cardiac outflow tract (OFT) myocardium was assayed for control and experimental animals.

Mean number of apoptotic cells/area of OFT myocardium101.3 ± 4543.2 ± 15 (43)0.0315
Mean trabecular thickness (μ) of mural ventricular myocardial wall283 ± 76134 ± 38 (47)0.0092
Mean compact zone thickness (μ) of mural ventricular myocardial wall96.7 ± 3544.0 ± 13 (46)0.0203

Coronary Endothelium Colocalized With Apoptotic Activity in the OFT Myocardium

In view of the fact that the presence of the epicardium over the OFT myocardium appeared to be necessary for apoptosis, we set about determining the epicardial component responsible for initiating this event. The coronary vasculature develops within the epicardium (Mikawa and Gourdie, 1996; Dettman et al., 1998; Gittenberger-de Groot et al., 1998; Manner, 1999) and forms a particularly dense plexus at the base of the OFT at the time apoptosis is initiated in this area (Waldo et al., 1990). We first detected an association between the presence of endothelial cells and the occurrence of apoptosis in the OFT myocardium by using conventionally stained, resin-embedded, semi-thin sections of stage 30 chicken embryos (not shown). Observation of the circumference of transversely sectioned OFT indicated that blood vessels were preferentially present in the epicardium or myocardium adjacent to intensely methylene blue–stained cells or cell fragments. This technique limited us to the identification of endothelial cells in vessels with distinct lumens. Younger vessels without a lumen or endothelial cells at the leading edge of vessels would not be detected. Thus, we turned to the quail system where there is an endothelial marker QH1 that can be used to identify endothelial cells from early in their differentiation. The same or adjacent sections of quail embryos were stained with the TUNEL technique and the immunofluorescent endothelial antibody marker QH1.

At stage 30–31, the peak of apoptosis in the OFT myocardium, QH1-positive cells were abundant where the most TUNEL-positive cells were detected in the OFT myocardium (Fig. 3). At this stage, coronary vessels surround the base of the OFT within the epicardium where they form a capillary plexus (Bogers et al., 1989; Waldo et al., 1990).

Figure 3.

Endothelial markers stain terminal deoxynucleotidyl transferase–mediated dUTP nick end-labeling (TUNEL) -positive regions of the cardiac outflow tract (OFT). Sagittal sections through the normal stage 30 quail embryos were stained with QH1 (A,C,E) to compare with TUNEL-positive cells (B,D,F). The concentration of QH1-positive cells at the proximal OFT(A, white arrows) overlapped with the concentration of TUNEL-positive cells (B, black arrows). A well-developed QH1-positive vessel (white arrows in C) in the epicardium was continuous with a vessel (white in arrow in E) in an adjacent section that appeared to enter the OFT myocardium (white arrows in E) at a level where there were many TUNEL-positive cells within the myocardium (black arrows in D, F). CM, cushion mesenchyme; My, myocardium; Epi, epicardium. Scale bar = 200 μ in B (applies to A–F).

Location of Apoptotic Cells Within the OFT Myocardium

Our previous findings (Watanabe et al., 1998) suggested that apoptosis within the OFT myocardium occurred in two patterns—scattered throughout the OFT myocardium and dense foci—which were confirmed by the present study. The scattered apoptotic cells were spread throughout the width of the OFT myocardial wall from the endocardial to the epicardial side and, at the base of the OFT, around the circumference. The dense foci of apoptotic activity appeared in discrete locations within the OFT myocardium as follows (ordered by decreasing frequency observed): the distal rim of the ventral OFT myocardium adjacent or downstream to the developing aortic valve leaflets, at the junction of the OFT myocardium with the ventricular myocardium, within the dorsal band of OFT myocardium, or in the interventricular septum. The foci at the distal rim of the OFT myocardium are within the myocardium between the aortic and pulmonic OFTs at the level of the developing semilunar valves (Fig. 4). Regions of dense apoptotic activity were also seen in similar areas in the developing quail heart; however, the borders were not as well defined as observed in the apoptotic foci of the chick embryo heart.

Figure 4.

The role of apoptosis in cardiac outflow tract (OFT) maturation. Diagram of transverse sections of the OFT at the level of the developing valves (A), or just below the valves (B–E). A: Preseptation, the OFT in cross-section is a myocardial tube with a single lumen lined by thick endocardial ridges or cushions forming primitive valves. The future lumen of the pulmonary (PA) and aortic (Ao) outflows are marked. The myocardium is enveloped by epicardium at approximately the time that OFT septation begins (Hamburger and Hamilton [HH] stage 25). B,C: Apoptosis may help sculpt the OFT myocardium adjacent to the septum. D,E:Apoptosis may also have a role during connection of the coronary arteries to the lumen of the aorta. The sub-aortic OFT is initially striated myocardium but is replaced by fibrous tissue by the time the coronary arteries mature in E (HH stage 37). AP, aorticopulmonary.

In all cases, the foci were only observed in stage 30 or 31 embryos, and in only approximately 50% of the embryos examined. That only a subset of embryos of these stages have these apoptotic foci suggests that the cells are undergoing apoptosis synchronously during a narrow window of time, which may last only several hours.

Majority of Apoptotic Cells in the OFT Myocardium Express Cardiomyocyte Markers

One explanation of the reduction in OFT apoptosis in the absence of the epicardium is that the cells migrating into the myocardium from the epicardium undergo apoptosis; thus, the levels of apoptosis in this tissue are reduced in their absence. However, previous data have indicated that many of these cells are cardiomyocytes (Hurle and Ojeda, 1979; Watanabe et al., 1998). Therefore, we quantitatively assessed the presence of apoptotic cardiomyocytes by using two methods (1) standard TEM, and (2) double labeling with the apoptosis marker Annexin V and cardiomyocyte markers. We analyzed the proximal OFT myocardium of three stage 30–31 embryos by standard TEM (Fig. 5). TEM allows definitive identification of apoptotic cells and cardiomyocytes by morphologic criteria. The healthy cardiomyocytes had well-organized, aligned myofibrils. Apoptotic cells at various stages of degeneration were observed and were identified by their condensed and darkly stained cytoplasm and nuclei, and large clear vacuoles. They were found in extracellular pockets as well as engulfed by healthy cardiomyocytes. A total of 85 apoptotic cells was examined by using ultrastructural criteria. Approximately 65% (55 of 85) of these contained myofilament bundles, which identified them as cardiomyocytes. The main limitation of this assay is that only a subset of cardiomyocytes in the early stages of apoptosis could have been identified by myofilament structure because of the rapid disintegration of cytoplasmic components during apoptosis. Thus the 35% of unidentified cells could very well contain a large number of apoptotic cardiomyocytes not identifiable by myofilament content, resulting in a misleadingly low number for the percentage of apoptotic cardiomyocytes. To overcome this limitation, we used another technique that allowed us to identify cells in earlier stages of apoptosis (van den Eijnde et al., 1997a,b) when the cardiomyocyte markers might still be intact.

Figure 5.

Ultrastructure of apoptotic cells in the cardiac outflow tract (OFT) myocardium as revealed by standard transmission electron microscopy. Apoptotic cells were identified by ultrastructural criteria in the proximal (A–D) and distal (E) OFT myocardium of stage 30 chicken embryos. Apoptotic cells and fragments had darkly stained cytoplasm and nuclei, multiple clear vacuoles, irregular blebbing of the cell surface, rounded cell shapes, and no cell–cell junctions. Cardiomyocytes were conservatively identified as those cells with myofilaments in their cytoplasm (B, arrowhead in C). D: Myofilaments were not detected in the apoptotic cell. This cell was not counted as a cardiomyocyte, despite the high content of mitochondria, a characteristic of cardiomyocytes in the proximal OFT myocardium. The apoptotic cells in the proximal region were surrounded by or engulfed by healthy cardiomyocytes containing bundles of striated myofilaments. Cells within the apoptotic foci at the distal rim of the OFT myocardium (E) had features of apoptotic cells but could not be definitively identified as cardiomyocytes by myofilament content. The surrounding apparently healthy cardiomyocytes in the distal region have sparse and disorganized bundles of myofilaments (E) compared with those that surrounded apoptotic cells in the proximal OFT (A,B). Scale bars = 1 μ in A (applies to A,E), in C (applies to B,C), in D.

In our previous study, many apoptotic cells of the OFT myocardium were identified as cardiomyocytes by their coexpression of Annexin V and cardiomyocyte markers (Watanabe et al., 1998). In this study, we used the same assay to obtain quantitative data. Among the Annexin V-positive cells in the OFT myocardium, we found that a subset had a ring-like staining pattern similar to that found for molecules in the plasma membrane of intact cells (Fig. 6). The rest appeared to have Annexin V staining throughout the cell. The latter likely represents apoptotic cells or apoptotic bodies late in the process of apoptosis when the integrity of the membranes has been lost, with subsequent staining of intracellular membranes (van den Eijnde et al., 1997a). By assessing the identity of only those apoptotic cells with the ring-like pattern of Annexin V staining, we would only be considering cells in the early stages of apoptosis when they would be more likely to have retained cardiomyocyte markers if they were cardiomyocytes. The percentages of these Annexin V-ringed cells having immunoreactivity to MF20 and anti-titin within the cytoplasm were 91% (53 of 58 cells) for the first embryo, and 92% (59 of 64) for the other. The rest of the Annexin V-ringed cells had no MF20 or anti-titin immunoreactivity in the cytoplasm.

Figure 6.

Confocal images of cells labeled with both Annexin V (red) and immunostained with markers for cardiomyocyte cytoskeletal elements (MF-20 and anti-titin, green). A,B: Cells counted as apoptotic cardiomyocytes (marked with asterisks) had distinct rims of Annexin V staining surrounding green-labeled cardiomyocyte cytoskeletal elements. Annexin V-positive cells or cell fragments that did not have staining in the rim pattern (indicated by white arrowheads) were not included in the total Annexin V-positive count, because their cytoplasmic compartment could not be clearly distinguished. The Annexin V-ringed cell (asterisk) in A is 14 μm in length, and the one in B (asterisk) is 17 μm in length. The images have a z section of 0.8 μm.

The results of the Annexin V and the TEM assays are in agreement that the bulk of the apoptotic cells in the OFT myocardium at the peak of apoptosis are cardiomyocytes. The rest of the apoptotic cells may be cardiomyocytes too far along the apoptotic process to contain identifying elements to classify them as cardiomyocytes. A small percentage of the apoptotic cells may include CNC cells, fibroblasts, smooth muscle cells, or endothelial cells, all of which are known to be in the same vicinity at this time in development.


The aim of this work was to characterize the cells undergoing apoptosis in the OFT myocardium and the cells that may influence this process. The data presented here show that a large proportion of cells that are dying during the peak of apoptosis in the OFT myocardium are cardiomyocytes. These data also support a role for the epicardium but not the CNC cells in initiating apoptosis in the OFT myocardium. One important component from the epicardium, the cardiac vasculature, was associated with the apoptotic cells of the OFT myocardium.

CNC Cells Do Not Initiate Apoptosis of OFT Cardiomyocytes

Our present data strongly suggest that CNC cells are not required for high levels of apoptosis in the OFT myocardium. We observed no reduction in levels of apoptosis within the OFT myocardium or changes in the general pattern and timing of apoptosis either in the absence of CNC cells alone or in embryos that had both the nodose placode as well as the CNC cells removed.

Nodose placode cells are known to regenerate from surrounding ectoderm and may replace the CNC cells that form the cardiac ganglia. The presence of regenerated ganglia or their precursors may explain the absence of effect on the levels of apoptosis within the OFT myocardium of doubly ablated embryos. This scenario appears to be unlikely when we consider the consequences of nodose placode ablation on their normal fate, to become components of the nodose ganglion (Harrison et al., 1995). Removal of a unilateral nodose placode resulted in 54% of embryos with absent nodose ganglia and normal sized nodose ganglion in only 8% of the operated embryos. In those with nodose ganglia, a portion of the replacement cells came from CNC. Therefore it would seem unlikely that embryos lacking both placodes and CNC cells would be able to regenerate the cardiac ganglia.

Epicardium Is Required for the Normal Level of Apoptosis in OFT Myocardium

The other set of cells that is present at the right time and place to initiate apoptosis in the OFT myocardium are the cells entering from the epicardium. When the ensheathment of the OFT myocardium by the epicardial serosa was delayed, apoptosis of OFT cardiomyocytes was greatly reduced. The apoptotic cells detected in the OFT myocardium in experimental embryos were in regions where the epicardium managed to grow, at the proximal-most and distal-most regions. These results indicate that a component of the epicardium is initiating apoptosis in the OFT myocardium.

An alternative explanation for the reduction in the number of apoptotic cells in the myocardium in the absence of epicardium is that the cells from the epicardium themselves represent the bulk of cells that undergo apoptosis within the OFT myocardium. However, we found that the majority of apoptotic cells (as many as 92%) in this region of the OFT myocardium at the peak of apoptosis express cardiomyocyte markers. Therefore, although a small fraction of the apoptotic cells in the OFT myocardium may be cells from the epicardium (8%), they do not represent a large enough fraction to account for the large reduction (57%) in the number of apoptotic cells in the absence of the epicardium. Furthermore, in a quail-chick chimera study, apoptosis was not seen among epicardially derived quail cells that migrated into chick cardiac tissue (Gittenberger-de Groot et al., 1998). These findings are consistent with the hypothesis that an apoptosis signal from the epicardium or from the cells migrating into the myocardium from the epicardium is failing to be delivered in experimental embryos.

Our results suggest that the epicardial coverage is important in achieving the normal level of apoptosis in the OFT myocardium. Although we have evidence that the population of cardiac neural crest cells that are the precursors of the cardiac ganglia are not likely to be involved, we did not identify the specific component from the epicardium that might be responsible for apoptosis in the OFT myocardium. Candidates include factors secreted by the epicardium and components on or secreted by cells migrating in from the epicardium.

Presence of QH1-Positive Endothelial Cells Is Associated With High Levels of Apoptosis

Cells that invade the myocardium from the epicardium become endothelial cells, smooth muscle cells, pericytes, or intramyocardial fibroblasts (Mikawa and Gourdie, 1996; Dettman et al., 1998; Gittenberger-de Groot et al., 1998; Manner, 1999). Endothelial cells express the quail-specific marker QH1 at early stages in their differentiation, before they achieve their distinct vascular structure (Pardanaud et al., 1987; Peault et al., 1988). This marker labeled endothelial cells intermingled with TUNEL-positive cells within the proximal OFT at the peak of apoptosis, when we found the majority of apoptotic cells are cardiomyocytes. An association between the absence of endothelial cell differentiation markers and OFT defects has been found in FOG-2–disrupted mouse embryos (Tevosian et al., 2000; Svensson et al., 2000). These embryos had mesothelial cells of the epicardium ensheathing the myocardium; however, none of the endothelial markers assayed were expressed within the epicardium. OFT abnormalities included pulmonic stenosis and elongated left ventricular OFTs, both of which might be expected if the level of apoptosis in this region were reduced in the absence of endothelial signaling. Indirect evidence from our studies and those of other investigators support a relationship between endothelial cell invasion and cardiomyocyte apoptosis.

The other cell types from the epicardium such as smooth muscle cells, pericytes, and fibroblasts remain potential candidates for the initiation or regulation of apoptosis. Further studies will be required to test which cell type or types from the epicardium initiates apoptosis and whether these cells directly or indirectly initiate or regulate apoptosis in this myocardial cell population.

The coronary vascular bed has been shown to be dramatically reduced by CNC ablation (Hood and Rosenquist, 1992; Hyer et al., 1999). If ingrowing coronary vessels initiate apoptosis as we suggest, then CNC ablation should have reduced apoptosis in our specimens indirectly by decreasing the coronary vasculature. This apparent contradiction to our results can be explained by the report that CNC ablation only affects the developing coronary vessels at the apex but not those in the area analyzed in this study, at the base of the ventricle (Hyer et al., 1999).

Previous studies suggest that apoptosis may be involved in the formation of the coronary artery connections. During cardiogenesis, an extensive capillary plexus forms in the epicardium around the base of the OFT (Waldo et al., 1990) followed by the appearance of multiple connections from it to the aortic sinuses and the eventual persistence of only two, the roots of the mature coronary arteries. The presence of CNC-derived parasympathetic ganglia closely associated with the two remaining coronary arteries (Waldo et al., 1994) suggested that CNCs may be important for the survival of the definitive coronary arteries. The corollary would be that those without the support of the parasympathetic ganglia or precursors would die. This hypothesis has not been tested. Apoptotic cells have been detected by TEM (Aikawa and Kawano, 1982) and by both TEM and TUNEL (Velkey and Bernanke, 2001) at sites where the coronary artery connects to the aortic lumen. The location of the degenerating cells suggested to one group that endothelial cells were undergoing cell death (Aikawa and Kawano, 1982). However the cell types undergoing apoptosis in this region and stage have not been identified, so the possibility remains that cardiomyocyte apoptosis may be involved in coronary artery formation.

One potential mechanism by which ingrowing endothelial cells may cause apoptosis in the cardiomyocyte population of the OFT is by activating the death ligand pathway. Endothelial cells are known to use the Fas pathway to initiate apoptosis in other cells (Sata and Walsh, 1998). Elements of the Fas/Fas-ligand death pathway have been found in both human umbilical vein endothelial cells and adult human coronary endothelial cells (Sata and Walsh, 1998; Sata et al., 2000; Romeo et al., 2000). Fas and other components of the Fas-mediated apoptotic pathway appear to be present and are able to be activated within OFT cardiomyocytes because virally targeted expression of Fas-ligand on these cells overstimulate the loss of OFT tissue (Sallee et al., 1999). However, it is not known whether Fas-ligand or similarly acting ligands are present on endothelial cells at this early stage of coronary development and whether this is the pathway used to initiate apoptosis in the developing cardiomyocytes. Identification of molecules whose expression are regulated at this site may reveal the pathways that lead to the peak of apoptosis in the OFT myocardium.

Sculpting of the OFT Myocardium by Apoptosis

We previously proposed that the high level of apoptosis among OFT cardiomyocytes is important for the general shortening of the whole structure (Watanabe et al., 1998). The results from our current study suggest that apoptosis may also be important for sculpting specific regions of the OFT myocardium (Fig. 4). As the OFT transforms from a single tube to two outflow structures, invaginations develop at the points where the two tracts cross. The sites where invaginations eventually occur are where foci of apoptotic cells were seen at stage 30–31. This location was identified by our histologic analyses and that of Icardo (1990). Furthermore, coronary arteries need to penetrate the muscular wall of the aorta if the coronary circulation is to develop normally. The coronary vessels may bore these holes themselves by initiating apoptosis as discussed earlier, or they may preferentially enter the points where apoptosis has occurred. In the rat, the incipient coronary arteries are known to penetrate the myocardium of the OFT. This myocardium eventually disappears such that the root of the mature coronary arteries are no longer surrounded by myocardium (Ya et al., 1998). The position of the foci that we have observed in chicken embryos at two sites between the pulmonary and aortic outflow vessels at the base of the OFT are consistent with this possibility as well.

The structural changes occurring in the developing OFT myocardium can be explained by the pattern of apoptosis that we and others have observed, but it is likely that many other cellular mechanisms, including differential proliferation and cell shape changes, are also involved in this complex process.

In conclusion, we have shown that the majority of cells that undergo apoptotic cell death in the OFT myocardium are cardiomyocytes. Coverage by the epicardium is required for normal levels of apoptosis to occur within the OFT myocardium, and endothelial cells are present intermingled or adjacent to regions of apoptosis in the OFT myocardium. We hypothesize that it may be the developing vasculature that invades and initiates cardiomyocyte cell death in the outflow tract.


Production of CNC-Ablated Chick Embryos

For the neural crest ablation studies, ablated animals and sham control animals were generously provided by Dr. Margaret Kirby, Medical College of Georgia, Augusta. Fertilized Arbor Acre chicken eggs were used (Seaboard Hatchery, Athens, GA). Fertilized white Leghorn chicken embryos (Gallus gallus) were used for the epicardial ablation and cell type studies, obtained from Squire Valleevue Farm, Ohio. Methods used for ablation of the neural crest and nodose placode have been published elsewhere (Leatherbury et al., 1993; see Fig. 7A). Briefly, animals were incubated at 37°C and 97% humidity in a forced-draft incubator. After 30 hr of incubation, the eggs were windowed. The embryos were lightly stained with neutral red-impregnated agar and staged according to Hamburger and Hamilton (1951). At stages 9 or 10, the cardiac neural folds were ablated bilaterally (Fig. 7A) from the mid-otic placode to the posterior limit of somite three with a pulsed nitrogen/dye laser (Laser Science, Inc., Newton, MA; VSL-377/DLM-110; Kirby et al., 1993). Sham-operated embryos were stained with neutral red only. Two groups of experimental animals were produced. One group underwent bilateral ablation of the neural crest tissue from the mid-otic placode to the caudal boundary of somite three. The second group underwent bilateral ablation of the nodose placodes, ectodermal structures lateral to the CNCs, as well as removal of the CNC as described. The eggs were sealed with cellophane tape and reincubated until the desired stage.

Figure 7.

Sites of intervention. A: Diagram of dorsal surface of the early chick embryo (approximately stage 9) showing the sites of ablated tissues, the cardiac neural crest (CNC) cells between somites 1 (s1) and 3 (s3) and the ectodermal nodose placodes lateral to the CNC. B: Diagram of an older chick embryo (approximately stage 15), showing the placement of the egg shell membrane (ESM) between the proepicardial serosa (PS) and the dorsal aspect of the heart (H). Modified from Kirby (1988a) and Manner (1993).

The second group was included because cells from the nodose placode can reconstitute parasympathetic innervation in the absence of the CNC cells (Kirby, 1988a,b). When the CNC cells are ablated, cholinergic neurons derived from the nodose placode were found within the parasympathetic ganglia of the heart. This finding suggests that cells from the nodose placode may alter their normal fate and replace CNC parasympathetic innervation in the absence of the CNC cells. Therefore, we analyzed embryos in which nodose placode as well as CNC cells were ablated to document the effect of eliminating the influence from CNC cells and cells that assume their role.

Production of Epicardially Delayed Embryos

The technique (Fig. 7B) used to delay growth of the epicardial organ over the myocardium was that of Manner (1993). A rectangular piece of egg shell membrane was inserted between the epicardial serosa and the looped heart tube of stage 15 embryos as previously described. Control embryos were prepared by opening a window and tearing the vitelline membrane, but with no insertion of an egg shell membrane. At this stage, the epicardial serosa, derived from mesenchyme dorsal to the heart loop, has not yet made attachments to the myocardium (Manner, 1993; Hiruma and Hirakow, 1989).

In surviving embryos, cells from the epicardial serosa eventually grow around the egg shell membrane to contact the myocardium, but with a 3- to 4-day delay in the time required to encompass the entire OFT. The epicardium normally completes growth around the heart by stage 24 (day 4 of development), reaching the base of the OFT around stage 21. Epicardially delayed embryos complete the epicardial covering of the myocardium by stages 32–34 (day 7.5–8). As apoptosis of OFT cardiomyocytes begins between stages 24 and 26 and peaks between stages 30 and 31, the epicardial delay provides a window of time during which the epicardium will not cover the OFT myocardium until after the peak of apoptosis. The windowed eggs were sealed with Parafilm and returned to the incubator until stage 30–31, at which time they were fixed in 10% formalin, embedded in paraffin, and assayed as described below.

Histologic Detection of TUNEL-Positive Apoptotic Cells in Chick and Quail

The animals were killed and fixed with 10% neutral phosphate buffered formalin (Ted Pella, Inc., Redding, CA) for 1 hr, embedded in paraffin, and sectioned (7 μ). The sections were deparaffinized and processed for detection of apoptosis by using the ApopTag peroxidase kit (Intergen). This process uses the TUNEL assay (terminal deoxynucleotidyl transferase mediated dUTP nick end labeling technique [Gavrieli et al., 1992]), which identifies cells that have undergone DNA fragmentation, one of the later events in apoptosis.

The sections were counterstained with MF20 and anti-titin to aid in the identification of the cardiomyocytes and to delineate the myocardium. The mouse monoclonal antibody MF20 supernatant was obtained from the Developmental Studies Hybridoma Bank maintained by the Department of Pharmacology and Molecular Sciences, Johns Hopkins University School of Medicine, Baltimore, MD, and the Department of Biological Sciences, University of Iowa, Iowa City, IA, under contract N01-HD-6-2915 from the NICHD. The mouse monoclonal anti-titin ascites (clone T11) was obtained from Sigma, St. Louis, MO (Furst et al., 1988).

Sections processed with the TUNEL technique were rinsed with phosphate buffered saline (PBS) and incubated at room temperature for 1 hr in MF20. The MF20 and anti-titin were diluted with 10% normal goat serum to a concentration of 1:500. The sections were washed in PBS and incubated in goat anti-mouse IgG FITC (Alexa Fluor, Molecular Probes or Texas Red, Cappel), 1:250, for 1 hr at room temperature, rinsed, and cover-slipped in Gelvatol solution (Airvol 523 polyvinyl alcohol, Air Products and Chemicals, Inc., Allentown, PA).

Quail sections were counterstained with QH1, a quail endothelial cell marker, also from the Developmental Studies Hybridoma Bank (see source information for MF20 above). The ascites-derived QH1 was diluted with 10% normal goat serum to a concentration of 1:1,000. Sections were incubated overnight in the primary antibody at 4°C, rinsed in PBS, and incubated in goat anti-mouse IgG FITC 1:250 for 1 hr at room temperature, rinsed, and cover-slipped in Gelvatol solution.

A volume of myocardium 300 μ (length) × 150 μ (width) × 125 μ (thickness) was analyzed in sham and experimental embryos. Only sections with normal patterns of TUNEL-positive cells in the body wall were counted (served as an internal positive control). A region adjacent to the presumptive aortic valve was selected in every embryo to standardize the area analyzed. Every other section (7 μ) was used in the calculations to reduce the possibility of counting cells twice. Statistical significance was determined with a two-tailed unpaired t test.

Preparation of Resin-Embedded Tissues for Analysis by Light and Transmission Electron Microscopy

Embryonic chicken and quail hearts were fixed sequentially in 2% glutaraldehyde and 2% OsO4 in 1% cacodylate buffer, en bloc stained with 1% uranyl acetate, dehydrated, and infiltrated by using ethanol and propylene oxide, and embedded in Polybed 812 resin (Polysciences). Semi-thin sections (0.5 μ) were stained with methylene blue. The apoptotic cells were stained an intense blue color compared with the surrounding tissue and could be easily distinguished from the small, more uniformly shaped and sized blue intracellular granules. Thin sections were poststained with uranyl acetate and lead citrate. Observations and photographs were taken by using the JEOL TEM.

Histologic Detection of Annexin V-Positive Apoptotic Cardiomyocytes in the Chick Embryo OFT

Chick embryos at stages 30–31 were stained for apoptotic cells by using a Ca2+-dependent phosphatidylserine binding protein, Annexin V, to bind to phosphatidylserine exposed on the plasma membrane of apoptotic cells (van den Eijnde et al., 1997a,b). Solutions (<1 μl/embryo) containing Annexin V conjugated to biotin (ApoAlert Annexin V, Clontech Laboratories, Inc., Palo Alto, CA) were injected into the vitelline vein of chicken embryos. After 1 hr, the hearts were dissected, fixed in a graded series of ethanol, and prepared for frozen sectioning (Watanabe et al., 1998). Sections of specimens injected with Annexin V-biotin were stained by using streptavidin-Texas Red (Vector, Burlingame, CA) and simultaneously immunofluorescently stained for the striated muscle markers, an antibody to the isoform of the myosin heavy chain (MF20) or anti-titin (Sigma Immunochemicals, T9030; St. Louis, MO).

In this study, 14-μ cryostat sections from four sagittal levels of the OFT myocardium were observed by confocal microscopy and counted from two stage 31 chick embryos. Cells that had a ring pattern of Annexin V staining along the plasma membrane as well as MF20/anti-titin staining were labeled apoptotic cardiomyocytes (Fig. 6). Cells with only plasma membrane staining of Annexin V were counted as apoptotic cells that were not cardiomyocytes. Cells intensely stained throughout with Annexin V were excluded from this analysis, because it was not possible to determine whether MF20 staining was present or not in the cytoplasm. Annexin V-positive cells within the endocardial cushions served as an internal negative control and expressed no MF20 or anti-titin immunostaining. Costaining with MF20 and anti-titin enhanced the intensity of the immunostaining. Observations were made on a laser confocal microscope (Zeiss LSM 410; Zeiss 40× Plan Neofluor, 0.90 NA, oil immersion). Z sections of less than 1 μ (0.8–0.9) were used to identify Annexin V-positive cells that were coimmunostained with MF20 and anti-titin.


The authors thank Margaret Kirby, Donna Kumiski, Harriet Stadt, and Karen Waldo for their helpful assistance and advice, and José Perez-Pomares for helpful discussions. F.R. received funding from the AHA Ohio Valley Fellowship, and M.W. and S.A.F. received funding from the NIH and AHA.