Spatiotemporal analysis of programmed cell death during mouse cardiac septation
Article first published online: 12 MAR 2004
Copyright © 2004 Wiley-Liss, Inc.
The Anatomical Record Part A: Discoveries in Molecular, Cellular, and Evolutionary Biology
Volume 277A, Issue 2, pages 355–369, April 2004
How to Cite
Sharma, P. R., Anderson, R. H., Copp, A. J. and Henderson, D. J. (2004), Spatiotemporal analysis of programmed cell death during mouse cardiac septation. Anat. Rec., 277A: 355–369. doi: 10.1002/ar.a.20006
- Issue published online: 12 MAR 2004
- Article first published online: 12 MAR 2004
- Manuscript Accepted: 17 NOV 2003
- Manuscript Received: 31 JAN 2003
- British Heart Foundation. Grant Numbers: FS/98010, BS/99001
- University College London Medical School
- Wellcome Trust. Grant Number: 051690
- cardiac development;
- programmed cell death;
Cell death is thought to play an important role in mammalian cardiogenesis, although a precise map of its distribution during the crucial period of cardiac septation has so far been lacking. In this study, the spatiotemporal distribution of programmed cell death (PCD) during mouse cardiac septation is described between embryonic days 10.5 and 13.5. Two types of foci of cell death can be demonstrated in the developing heart. Those with high-intensity, with a PCD index greater than 1%, are clearly visible on individual TUNEL-assayed sections. Low-intensity foci, with a PCD index of less than 1%, become visible only following summation of data. High-intensity foci occur exclusively within the endocardial cushions of the outflow tract and atrioventricular region, appearing at the 52–54 somite stage (late E11.5), concomitant with the formation of the central mesenchymal mass. Low-intensity foci are present throughout the period of cardiac development from E10.5 to E13.5 and are frequently localized to regions of septation, such as the muscular ventricular septum and the mesenchymal cap of the primary atrial septum. Expression of Fas and FasL corresponds to these low-intensity foci, but not those with high-intensity, suggesting that activation of this death receptor may be specifically involved in molecular control of the low-intensity foci. Anat Rec Part A 277A:355–369, 2004. © 2004 Wiley-Liss, Inc.
The normal development of many mammalian organ systems has been shown to depend on the removal of cells by appropriate programmed cell death (PCD) (Jacobson et al.,1997; Vaux and Korsmeyer,1999). While its exact function in the developing heart remains unresolved, a role for PCD during mammalian cardiac development is suggested by targeted mutations of the caspase 8, FADD, and FLIP genes, all of which cause cardiac defects (Varfolomeev et al.,1998; Yeh et al.,1998,2000). Although these studies show that PCD is important for mammalian cardiac development, they do not reveal its function. In order to begin to address this, it is essential to have an accurate spatiotemporal map of the dying cells within the heart.
The development of the avian and mammalian heart from a simple unseptated tube to a complex four-chambered dual-outflow system is dependent on the interactions between a number of different cell types, which originate both within and from outside the primitive heart tube. Cardiac septation results from the coordinated growth, fusion, and subsequent remodeling of the endocardial cushions that form in the atrioventricular canal and the common outflow tract. These cushions, although arising from interactions between the intrinsic myocardium and endocardium, are invaded by extracardiac populations of cells, including neural crest cells (Kirby et al.,1983; Jiang et al.,2000) and epicardially derived cells (Gittenberger-de Groot et al.,1998; Perez-Pomares et al.,2002), which influence significantly their subsequent development. High levels of PCD have been described in the endocardial cushions of chick, rat, and mouse embryos, although the exact distribution and onset of this PCD is unclear and varies between species (Pexieder,1975; Hurle and Ojeda,1979; Poelmann et al.,1998; Ya et al.,1998; Zhao and Rivkees,2000; Abdelwahid et al.,2001; Cheng et al.,2002; Keyes and Sanders,2002). Although several studies have suggested that there are also low levels of PCD elsewhere in the developing heart, for example in the myocardium of the ventricles (Abdelwahid et al.,1999) and outflow tract (Rothenberg et al.,2002), they do not describe comprehensively the spatiotemporal nature of this apoptosis during the period of cardiac septation. Evidence from a number of animal models (Lakkis and Epstein,1998; Bartram et al.,2001; Watanabe et al.,2001) nonetheless indicates that the levels and timing of this apoptosis is critical to the normal cardiac development.
Relatively low levels of PCD have been shown to be important in morphogenesis. Camp and Martin (1996) suggested that the death of between 1% and 3% of the cells in the developing rat mesonephros may be responsible, over the period of a few days, for loss of half of all the cells in the nephrogenic part of that structure. Moreover, it is well established that low-frequency PCD removes half of all motor neurons that develop during embryogenesis (Oppenheim,1991). Low-intensity PCD may be important, therefore, if it is spatially defined and persistent.
The aim of the present study was not only to confirm the detailed location of high-intensity foci of PCD, as identified in other studies of the developing heart, but also to examine the distribution of consistent low-intensity PCD foci. Since a single section assayed by the TdT-mediated dUTP nick end labeling (TUNEL) method represents only a snapshot of the distribution of cells at a relatively late stage of the apoptosis, low-intensity PCD foci may be very difficult to identify, particularly if they affect only a small cardiac region. For this reason, we developed a templating technique in which data were summated from several comparable sections of similarly staged hearts. The templating method has allowed us to identify several regions of persistent low-intensity PCD during cardiac septation, only one of which was visible on individual sections.
MATERIALS AND METHODS
Mice and Embryos
CD1 random-bred mice (Charles River, U.K.) were maintained according to the Animals (Scientific Procedures) Act 1986 of the U.K. government and were exposed to a 12-hr light-dark cycle with constant access to food and water. Pregnant females were killed by cervical dislocation and the embryos were dissected free from the uterus and prepared for immunohistochemistry as described previously (Henderson and Copp,1998). In order to visualize changes in the distribution of PCD during the rapid process of cardiac septation, 31–60 somite embryos [embryonic days (E) 10.5–12.5] were divided into groups, each spanning three somite stages and therefore representing around 6 hr of gestation. A final group spanning the period from late E12.5 to early E13.5, when somitogenesis is complete, was also included.
TUNEL Assay, Immunohistochemistry, and Electron Microscopy
The peroxidase-TUNEL assay was performed according to the manufacturer's protocol (Intergen) on 8 μm transversely sectioned paraffin-embedded embryos except that the proteinase K treatment varied from 6 to 9 min, depending on the gestational age of the embryos. Negative controls involved omission of the TdT enzyme, and positive staining was never seen. Immunohistochemistry with the antibodies to cleaved caspase 3 (Cell Signalling Technology), Fas (pc-69; Oncogene), and FasL (sc-834; Santa Cruz Biotechnology) was performed as recommended by the suppliers. Macrophages were identified using the monoclonal anti-F4/80 antibody, which was a gift from Dr. Paul Martin of London and Professor Saimon Gordon of Oxford (Camp and Martin,1996). Negative controls for immunohistochemistry involved omission of the primary antibody, and positive staining was never seen. Transmission electron microscopy was performed as described previously (Ybot-Gonzalez and Copp,1999) by Mr. Brian Young and Professor David Landon at the Institute of Neurology (London, U.K.).
Summation of Data (Templating)
Individual TUNEL-assayed sections frequently showed only a few isolated apoptotic cells. In order to determine whether PCD occurred reproducibly in particular anatomical locations, we devised a method to summate our data. To create the templates, 16 equidistant representative transverse sections were taken through the heart from each somite group and stained with methyl green. Images of the sections were captured digitally and printed. TUNEL-assayed sections were compared to these templates and matched to a specific image according to developmental stage and morphology. Every apoptotic cell corresponding to a particular template was marked on a transparency, which was overlaid on the printed template. In general, there were 5–8 sections corresponding to each template per embryo, and apoptotic cells from these were all marked on the same transparency. Different colored dots represent apoptotic cells from different individual sections and were used to aid the templating process. The size of the colored dots is not of any significance. This procedure was repeated on a minimum of three different embryos per somite group. Images of the data on the transparencies were then scanned into Adobe Photoshop 4.0. The data from all three embryos were then overlaid onto the image of the appropriate template (Fig. 1).
Quantitation of PCD
The total number of apoptotic cells plotted on a given template was divided by the total number of cells in the template (counted using the KS 300 program; Imaging Associates, Thame, U.K.) multiplied by the number of sections analyzed for that template. This gave an estimate of the proportion of cells that were apoptotic, expressed as a percentage (i.e., the PCD index). This quantitative analysis was performed on data averaged from embryos spanning a range of six somites, that is, incorporating two of the previous somite groups, such as the groups of 31–33 somites and 34–36 somites, and so on. Statistical analysis was performed using the Sigma Stat program (Jandel Scientific) and graphs were drawn using Sigma Plot (SPSS, Jandel Scientific).
Every region of the heart showed some apoptotic cell death throughout the gestational period examined (Table 1). In the ventricles, the atria, and most of the outflow tract, however, trends in PCD could be discerned only subsequent to summation of data, whereas from early E12.5 to E13.5, localization of PCD within the endocardial cushions was apparent on individual TUNEL sections.
|Structure||Developmental Stage (Somites)|
|Atria||Adjacent to venous inlet||L||L||L||L|
|Right venous valve||L||L||L||L|
|Dorsal part of left atrium||L|
|Primary atrial septum||L||L||L||H||H||H||H|
|AV canal||Endocardial cushions||L||L||L||L||H||H||H||H|
|Ventricles||Junction of right ventricle and outflow tract||L||L||L||L||L||L||L|
|Junction of left ventricle and aorta||—||—||—||—||—||—||—||L||L|
|Junction of atria and ventricles||L||L||L|
|Outflow tract||Aortopulmonary septum||L||L||L||L||L||—||—||—||—||—||—|
|Walls of aortic sac||L||L||L||L||L||—||—||—||—||—||—|
|Walls of great vessels||—||—||—||—||—||L||L||L||L||L||L|
|Ventral outflow tract walls||L||L||L||L||L||L||L||L|
|Lateral outflow tract walls||L||L||L||L|
High Levels of PCD Occur in Atrioventricular Endocardial Cushions
From 31 to 48 somites (E10.5 to early E11.5), very little PCD is apparent within the atrioventricular cushions, even using the templating technique (Fig. 2a and b). By 49–51 somites (mid E11.5), however, foci of cell death can be detected on individual TUNEL sections. These are seen close to the junctions of the primary atrial septum with the superior and inferior atrioventricular cushions (Fig. 2c and d). At early E12.5 (52–54 somites), PCD appears to occur more intensely in the atrioventricular cushions than at preceding stages (compare Fig. 2e with d). Although apoptosis occurs diffusely within the superior atrioventricular cushion (Fig. 2e), the most intense focus is seen, particularly at the 55 and 60 somite stage, in the region where the mesenchymal cap of the primary atrial septum fuses with the cushions (Fig. 2f). At E13.5, PCD in the atrioventricular cushions is less extensive than seen previously, being largely limited to regions where the primary atrial septum has not achieved continuity with the inferior and superior endocardial cushions. In addition, PCD is now detectable at the sites of formation of the atrioventricular valvar leaflets at the edges of the atrioventricular cushions (Fig. 2g).
Quantitatively, PCD in the atrioventricular cushions can be seen to increase sharply from 43–48 to 55–60 somites, with a decline at E13.5. At its peak, PCD affects 4% of the cells within the atrioventricular cushions (Fig. 3a).
PCD Is Particularly Intense in Endocardial Cushions of Outflow Tract
Summation of data by templating does not detect significant PCD during formation of the cushions within the outflow tract, at E10.5 and early E11.5 (large arrow in Fig. 4a). By 46–48 somites (mid E11.5), PCD has increased (Fig. 4b), although dying cells are difficult to visualize on individual TUNEL-assayed sections (Fig. 4g). By the 49–51 somite stage, PCD has increased markedly in the proximal outflow tract cushions (Fig. 4d). From this stage onward, dying cells are clearly visible without the aid of data summation.
At the 49–51 somite stage, bright-field microscopy reveals PCD to be diffuse, and not clearly localized to any particular part of the outflow tract cushions (Fig. 4h). High-intensity PCD in the proximal outflow tract cushions appears at 52–54 somites (Fig. 4i) and is maintained at 55–60 somites (Fig. 4j), when the distal portions of the outflow tract cushions have already fused to divide the distal outflow tract. The appearance of this intense PCD is concurrent with that in the atrioventricular cushions, although unlike the atrioventricular region, PCD in the outflow tract does not appear to be associated with sites of fusion of cushions but is seen more proximally within the unfused cushions. By E13.5, when cardiac septation is complete, PCD in the proximal outflow tract cushions has diminished in intensity, although it can still be seen in a central location within the region of fused proximal endocardial cushions (Fig. 4k).
The quantitative increase in PCD observed in the outflow tract cushions, from 43–48 to 55–60 somites, closely resembles the increase observed in the atrioventricular cushions (Fig. 3a). At its peak, at the 55–60 somite stage, PCD affects approximately 6% of the cells in the outflow cushions.
Low-Intensity Foci of PCD Characterize Cardiac Regions Undergoing Septation
Foci of PCD are not detectable in the atria at E10.5 (Fig. 5a), but from 40 somites onward are present mainly within right atrial structures, including the septum spurium and the right venous valve (Fig. 5b, d, f, and g). Although PCD is an order of magnitude less abundant than in the endocardial cushions, the PCD index in the right atrium is consistently greater than in the left atrium (Fig. 3b). At 43–48 somites, a relatively intense focus of cell death affecting 1% of all cells can be detected in the mesenchymal cap of the primary atrial septum (Fig. 3b), just prior to its fusion with the atrioventricular cushions (Fig. 5c). This focus of cell death is seen ventrally within the mesenchymal cap of the atrial septum and not adjacent to the developing oval foramen.
In the ventricles, the most prominent focus of apoptosis is within the muscular ventricular septum, which, even at the earliest stage of its development (31–33 somites), is revealed by templating to have more PCD than the adjacent ventricles (Fig. 6a). By 34–36 somites, the growing septum possesses an obvious focus of PCD (Fig. 6b), and by early E11.5 (40–42 somites), this is continuous with newly formed foci in the nontrabeculated ventricular wall of the ventral part of both ventricles (Fig. 6c). The foci of PCD in the ventricular septum and adjacent ventricles continue to be seen through E11.5–E13.5 (Fig. 6d–g), with the focus in the left ventricle becoming more extensive laterally compared with the right ventricular PCD focus (Fig. 6f). The ventricular septum exhibits PCD that is progressively localized to its central and ventral aspects (Fig. 6g).
Quantitatively, PCD increases gradually in all ventricular structures from E10.5 to E13.5 (Fig. 3c). Apoptotic cells comprise up to 0.6% of ventricular septal cells, which is approximately 10-fold lower than in the endocardial cushions (Fig. 3a). The ventricular walls exhibit a generally lower PCD index, and the index is lower in the right than in the left ventricle.
Outflow Tract Structures
At early E10.5 in the 31–33 somite group, a focus of PCD is detectable in the aortopulmonary septum, dividing the dorsal wall of the aortic sac (data not shown). By the 40–42 somite stage, this focus can still be seen (white arrow in Fig. 4a), but PCD foci also occur in the lateral walls of the aortic sac at the site of inward migration from the neural crest (arrowheads in Fig. 4a). Levels of PCD are higher at this site than in the remainder of the outflow tract. At 46–48 somites, the distal outflow tract is dividing to form the intrapericardial components of the aorta and pulmonary trunk (Fig. 4c). At this stage, PCD foci can be seen at the point of separation between the medial walls of the arterial trunks (arrowhead in Fig. 4c). This focus is maintained until 55–57 somites (arrowhead in Fig. 4e).
PCD in Endocardial Cushions Can Be Detected by Transmission Electron Microscopy and Caspase 3 Immunohistochemistry
In order to confirm that TUNEL-positive nuclei found in the heart are dying by apoptosis, three embryonic hearts were examined by transmission electron microscopy (TEM) at the 55–60 somite stage, when high-intensity PCD foci are visible within the outflow tract cushions (Fig. 3a). Cells with the characteristic TEM morphology of apoptosis were frequently observed in the proximal outflow tract cushions (Fig. 7a–d), whereas such cells were only rarely seen in other regions of the outflow tract. In particular, apoptotic bodies (Fig. 7b and c) and edging of chromatin against the nuclear membrane (Fig. 7d) were characteristic findings. No obviously necrotic cells were seen.
Immunohistochemistry for activated caspase 3 was performed on sections of 55–57 somite embryos in which the central mesenchymal mass is present. In these sections, caspase 3-positive cells were found in both high- and low-intensity PCD foci in the same distribution as that of TUNEL-positive cells (compare Fig. 7e and g with f and h), confirming the involvement of caspase activation in the cardiac cell death observed.
Fas and Fas Ligand Expression Correlates With Low- But Not High-Intensity Foci of PCD
Receptor-mediated PCD has been implicated in cardiovascular development (Varfolomeev et al.,1998; Yeh et al.,1998,2000). Because of this, we attempted to correlate the expression of Fas and Fas ligand (FasL) with sites of PCD in the septating mouse heart. These patterns of expression, to the best of our knowledge, have not been described in the developing mouse heart. Adult thymus, a known site of Fas and FasL expression, was therefore used as a control for positive and negative staining, confirming specificity (Fig. 8e–h).
Immunohistochemistry on cryosections from E10.5–E13.5 hearts showed that expression of Fas was comparable at all the stages examined, being found throughout the myocardium, endocardium, and epicardium, but was absent from the endocardial cushion tissue and its derivatives (Fig. 8a and b). Specifically, no Fas expression was found in the proximal outlet septum, where the highest-intensity focus of PCD is found in the developing heart (asterisk in Fig. 8b). In contrast, FasL was found in isolated cells in discrete regions of the heart, including the septum spurium, the myocardium of the right ventricular outflow tract, the muscular ventricular septum, and the base of the ventricles (Fig. 8c and d). All of these areas corresponded with regions where templating had revealed low-intensity foci of cell death. Expression was not seen at sites of high-intensity PCD foci, such as the proximal outlet septum, or in areas where PCD was not observed.
Macrophages Do Not Localize to Foci of PCD During Cardiac Septation
Anti-F4/80 immunostaining, which labels all macrophages (Austyn and Gordon,1981), was performed between E10.5 and E13.5 to look for the presence of macrophages during cardiac septation. Only very occasional F4/80-positive cells were found in the developing heart, although they were abundant in fetal liver (Fig. 9g). At no point did macrophage staining correlate with either high- or low-intensity PCD foci, either in the muscular ventricular septum (Fig. 9a and b), atrioventricular region (Fig. 9c and d), or the proximal outflow tract cushions (Fig. 9e and f). Moreover, TEM of the outflow tract cushions identified apoptotic debris within the cell membranes of resident cardiac cells, rather than within cells having the characteristics of macrophages (Fig. 7b–d).
In this study, we mapped the distribution of PCD during cardiac septation (see Table 1 and Fig. 10 for summary). High-intensity foci of PCD, with a PCD index of greater than 1%, were identified from the 49–54 somite stage (E12.5) in the endocardial cushions of both the atrioventricular and outflow tract regions and in the mesenchymal cap of the primary atrial septum. This latter tissue has a composition similar to endocardial cushion tissue and may well arise by the same mechanism (Arrechedera et al.,1987; Henderson and Copp,1998).
These are structures whose cellular composition alters dramatically during the period of septation, perhaps reflecting a role for PCD in removing those cell types no longer needed for subsequent development. In contrast, low-intensity foci, with a PCD index of less than 1%, are present throughout the heart at all stages, although most foci increase in prominence with gestational age and correlate with areas of morphological remodeling. This disparity between foci of high- and low-intensity suggest there may be differences in the molecular regulation of PCD. Significantly, low-intensity foci correlate with expression of Fas/FasL, suggesting that death receptor activation may be involved in their regulation.
Need for Templating
The aim of our study was to identify consistent low-intensity foci of cell death as well as the high-intensity foci of PCD that others have identified in studies of the developing heart (Pexieder,1975; Hurle and Ojeda,1979; Poelmann et al.,1998; Ya et al.,1998; Zhao and Rivkees,2000; Abdelwahid et al.,2001; Cheng et al.,2002; Keyes and Sanders,2002). We developed the templating technique in order to summate data for consistent spatially regulated low-intensity PCD. The hearts at all stages were divided into 16 equidistant rostrocaudal segments, and representative views were taken of these segments, providing us with our templates. Visual analysis of these segments, which represent a three-dimensional slice of heart tissue, showed minimal morphological variation within the sections matched to each template. Apoptotic cells for all the sections that corresponded to that template were then marked individually on the template, allowing us to build up a map of all the apoptotic cells in a particular slice of the heart. By examining these together, we showed their location throughout a particular structure. As the heart is considerably larger at E13.5 than at E10.5, the thickness of tissue corresponding to a particular template, and the number of sections corresponding to a single template at E13.5, is considerably larger than at E10.5. For example, at E13.5, each template corresponded to a tissue slice of approximately 60 μm, whereas at E10.5, a single template corresponded to a 40 μm slice of tissue. It is for this reason that comparisons have not been made between numbers of apoptotic cells on templates of different ages. However, the PCD index was derived from the ratio of cells undergoing apoptosis to total cells and was not therefore biased by these developmental changes. This allowed comparisons to be made between hearts of different gestational ages in this analysis. This templating technique, by summating data from different but very similar sections of the heart, and moreover from several embryos at once, thus permits the identification of very low level foci of cell death, but only when they are consistent from section to section within a specific tissue and from embryo to embryo. Indeed, we were able to identify several foci, for example in the systemic venous tributaries and the right venous valve, which had not previously, to the best of our knowledge, been identified. However, as with all methods of data summation, subtle differences in PCD localization to fine intracardiac structure may be missed. Although many studies have been published on cell death in the developing heart, the mouse heart studies have not been as thorough or as quantitative as in other species. Considering the potential importance of cell death in the morphogenesis of many tissues and structures in the heart, we suggest that it is crucial that a reference be provided from which to compare aberrant patterns of programmed cell death. In our opinion, our study provides this reference. Although it might be argued that the anatomical features in individual hearts can differ significantly, or that projecting data collected from sections up to 64 μm apart onto a single section is not appropriate, we believe that the advantages gained from the templating approach significantly outweigh these potential caveats.
Association of High-Intensity Foci of PCD With Changes in Cellular Composition of Septal Structures
The majority of previous studies have reported high levels of programmed cell death in the developing endocardial cushions (Pexieder,1975; Hurle and Ojeda,1979; Poelmann et al.,1998; Ya et al.,1998; Zhao and Rivkees,2000; Abdelwahid et al.,2001; Cheng et al.,2002; Keyes and Sanders,2002). Not unreasonably, these changes have been correlated with septation and remodeling of the cushion tissue. In the embryonic chick heart, the levels of PCD peak first in the outflow tract cushions, then later in the atrioventricular cushions (Cheng et al.,2002; Keyes and Sanders,2002), and have been suggested to follow waves of septation (Cheng et al.,2002). In our study, high-intensity foci of PCD were found simultaneously in the cushions of the atrioventricular canal and outflow tract, albeit in similar patterns to those seen in chick, perhaps reflecting subtle differences in timing of septation between mouse and chick. In the atrioventricular cushions, PCD occurred principally at the points of fusion between the primary atrial septum and the superior atrioventricular cushion, and between the superior and the inferior atrioventricular cushions. It is possible that this might represent removal of endocardial cells lying between apposing structures and thereby allowing cushion fusion. This localized PCD is similar to that described in the atrioventricular cushions in the developing chick heart (Cheng et al.,2002), where PCD levels peaked during the period of fusion of the cushions. There is evidence for a role for PCD in this type of process in other developing systems. For instance, cell death appears to be required for the closure of the neural tube of the developing chick embryo (Weil et al.,1997), suggesting that PCD is involved in epithelial rearrangement or cell fusion. It is possible that the cell death in the fusion seams may in some way induce fusion, perhaps by releasing a factor that renders the adjacent cells more adhesive. The mechanism of programmed cell death, however, is not generally held to involve release of cell contents to the surrounding environment, thus making this hypothesis unlikely. An alternative explanation is that PCD removes another cell type from the atrioventricular cushions. The atrioventricular cushions are known to be invaded by epicardially derived cells, at least in the chick (Gittenberger-de Groot et al.,1998; Perez-Pomares et al.,2002), which are essential for the normal morphological development of the atrioventricular valves (Perez-Pomares et al.,2002), and it may be that their timely removal is an important part of this process. In support of this hypothesis, these epicardially derived cells are found principally in the subendocardial layers (Gittenberger-de Groot et al.,1998; Perez-Pomares et al.,2002) and thus are close to the fusion seams where we found the majority of dying cells.
In contrast, PCD in the outflow tract cushions occurs on either side of the prospective line of fusion at E12.5 and persists in the center of the fused septum of the proximal outflow tract at E13.5. These observations suggest a functional difference between PCD in the outflow tract and in the atrioventricular cushions. Perhaps the most important difference between the outflow tract and atrioventricular cushions in the developing mouse heart is the migration of neural crest cells into the outflow tract, but not the atrioventricular junctions. The high levels of PCD in the outflow tract cushions may be connected with colonization by cardiac neural crest cells, as has been suggested for the chick (Poelmann et al.,1998). Moreover, the timing of the high-intensity PCD in the outflow tract closely matches the stage at which lineage tracing studies have demonstrated a dramatic reduction in numbers of cells derived from the neural crest (Jiang et al.,2000). Hence, removal of cells derived from the neural crest in the outflow tract is likely to occur by apoptosis. High-intensity foci of PCD may therefore reflect a requirement for rapid removal of cell types from the endocardial cushions that are no longer required for further development. Significantly, the areas of apoptosis involve predominantly the cushions of the proximal outflow tract. Here, the major morphological change is the conversion of the mesenchymal cushion tissue into the muscular embryonic outlet septum (Van den Hoff et al.,1999) and later into the free-standing subpulmonary infundibulum. Recent work has suggested that rather than becoming redundant, the mesenchymal cushion cells are recruited into the muscular lineage as the septum is muscularized (Kruithof et al.,2003). It is unlikely, then, that these cells die in large numbers. It has been suggested that apoptotic neural crest cells might be involved in inducing the muscularization of the outflow tract cushions (Poelmann et al.,1998), perhaps by releasing a factor such as TGF-β2, which then induces the muscularization of the septum. As discussed above, the extracellular release of factors is not normally associated with the apoptotic process; nevertheless, elevated levels of TGF-β2 do appear to enhance apoptosis and to be involved in outflow tract septation and remodeling (Bartram et al.,2001; Kubalak et al.,2002). Epicardial cells have been shown to be essential for septation of the outflow tract and muscularization of the outflow tract cushions, since these processes fail in chick embryos where the process of epicardial outgrowth is inhibited (Gittenberger-de Groot et al.,2000). However, epicardial cells are not thought to enter the outflow tract cushions (Manner et al.,1999), and this therefore is unlikely to be the identity of the dying cells in this tissue. The separation of the aortic and pulmonary roots in the area of the proximal outflow tract just proximal to the prominent “dog leg” bend may provide an alternative explanation. The aorta and the pulmonary trunks both arise from the common trunk and, following septation of the outflow tract, become completely separate from one another. It is conceivable that the apoptotic events may be involved in the conversion of these initially septal areas to regions of extracardiac fibroadipose tissue. These ideas are not mutually exclusive, and there may be PCD occurring within several of these cell populations. Lineage tracing experiments will be crucial in determining the identity of the apoptotic cells in the complex environment of the outflow tract cushions.
Low-Intensity PCD Foci Are Associated With Regions of Morphological Remodeling
The heart undergoes extensive remodeling during septation, processes associated with low-intensity foci of cell death. For example, in the atria, the systemic venous tributaries are incorporated solely into the right atrium, while the inferior commissure of the right venous valve contributes to the developing central mesenchymal mass via the vestibular spine (Webb et al.,1998). It is striking that we observed a consistently higher level of PCD in the right than in the left atrial structures, since relatively little remodeling occurs in the left atrium. The separation of the great arteries from each other is also associated with low-intensity foci of cell death. Rothenberg et al. (2002) have described foci of cell death in the chick heart associated with shortening of the outflow tract and have shown that this apoptosis is essential for normal formation of the ventriculoarterial connections (Watanabe et al.,2001). In our study of the mouse heart, we also saw PCD in the outflow tract myocardium during the period in which the outflow tract shortens, suggesting that this mechanism may well be conserved across species.
A further example concerns the peripheral regions of the atrioventricular cushions, which undergo reshaping to form the atrioventricular valvar leaflets at E13.5. At this stage, we found PCD mainly in the peripheral regions of the atrioventricular canal, regions that disappear during the sculpting of the valvar leaflets. Similar findings were reported by Cheng et al. (2002), who also noted PCD in the atrioventricular myocardium of the embryonic chick heart, correlating with regions of formation of conduction tissues. These authors also described high-level PCD in the crest of the ventricular septum at the site of the His bundle and in the ventricular trabeculations. In our study, some PCD was found in both of these locations, although the patterns were not striking. This perhaps reflects a species difference between mouse and chick, the more so because the atrioventricular bundle in the chick is buried within the muscular ventricular septum. Striking foci of PCD were found, however, in the base of the ventricular septum and the adjoining regions of the right and left ventricles. These regions correspond with the ramifications of the fascicular conduction system within the ventricular apices. This pattern was described in the seminal study of Pexieder (1975), although it has not been noted in more recent studies. The septum forms initially at around E10.5, as a flimsy structure in the base of the ventricles. By E13.5, it has become more solid, comprised of compact myocardium, and is similar in composition to the adjoining ventricular wall. There is little direct data relating to the embryonic compaction of the myocardium, although the thickening of the wall appears to result from both proliferation and the compaction of the trabeculations (Sedmera et al.,2000) and to be dependent on invasion of the myocardium by cells derived from the epicardium (Gittenberger-de Groot et al.,2000; Perez-Pomares et al.,2002). The dying cells could, as with the fusing atrioventricular cushions, reflect the loss of endocardial cells from the regions of trabecular fusion. Alternatively, the epicardial cells that invade the myocardial wall may be programmed to die by apoptosis once they have fulfilled their purpose.
Receptor-Mediated PCD Is Found in Areas of Low-Intensity Foci
We have shown that Fas is expressed throughout the heart except in the endocardial cushions. Furthermore, FasL is expressed solely in areas that correspond to regions of low-intensity PCD. This provides a possible mechanism for molecular control of the low-intensity, but not the high-intensity, foci. Targeted deletion of several components of the receptor-mediated cell death pathway, including FADD (Varfolomeev et al.,1998), caspase 8 (Yeh et al.,1998), and FLIP (Yeh et al.,2000), has produced animals with cardiac defects that correspond to the areas of low-intensity PCD foci that we have described. In each case, the main defect is in the ventricular myocardium, manifesting as a hypoplastic ventricular wall. In none of these models is there a defect in the endocardial cushions, supporting the idea that death receptor activation is not responsible for the cell death found in the endocardial cushions. Interestingly, Keyes and Sanders (2002), in their study of the regulation of PCD in the endocardial cushions of the chick heart, showed that caspase 9, which is a mediator of the mitochondrial-initiated pathway of PCD, in other words, not the receptor-mediated pathway, is expressed in the endocardial cushions. Inhibition of this caspase was able to inhibit PCD in cultures of the endocardial cushions. This implies that the mitochondrial-initiated pathway for cell death is active in the high-intensity foci of PCD in the endocardial cushions and supports the idea that different mechanisms regulate the PCD within the high- and low-intensity foci.
The growth factors BMP2 and BMP4 have been associated with PCD in the atrioventricular and outflow tract cushions, respectively (Zhao and Rivkees,2000; Abdelwahid et al.,2001). In culture experiments, BMP4 has been shown to induce PCD in both the endocardial cushions and the ventricular myocardium. In contrast, FGF2 has been shown to inhibit PCD in these tissues. Interestingly, these factors have also been shown to induce, in the case of BMP4, and inhibit, in the case of FGF2, PCD in other developmental systems (Macias et al.,1996; Zou and Niswander,1996), supporting the idea that they may also be involved in regulation of PCD in the developing heart. Certainly, there is a great deal of scope for regulation of PCD by growth factors in specific cardiac tissues.
Clearance of PCD During Cardiac Septation Is by Local Cells, Not by Macrophages
We were unable to demonstrate F4/80-positive macrophages in the vicinity of foci of PCD in the developing heart, and our TEM studies provided evidence that local cardiac cells phagocytose apoptotic debris, at least in the outflow tract cushions. This finding is somewhat surprising since in other developmental systems where PCD occurs at high levels, such as in the developing interdigital webs and the developing kidneys (Austyn and Gordon,1981; Hopkinson-Woolley et al.,1994; Camp and Martin,1996), apoptotic cells are cleared by professional macrophages. Moreover, other reports of PCD in the developing heart have suggested that macrophages are responsible for clearance of apoptotic cells (Pexieder,1975; Ya et al.,1998; Abdelwahid et al.,1999), although in only one study was a specific immunohistochemical marker used to identify macrophages (Ya et al.,1998). Wood et al. (2000) demonstrated that, even in regions of intense PCD such as the interdigital webs, neighboring mesenchymal cells can phagocytose apoptotic cells in the experimental absence of professional macrophages. Hence, our findings of apoptotic bodies apparently undergoing phagocytosis within cardiac mesenchymal cells may indicate that specialized macrophages are at most only partially responsible for clearance of PCD in the developing heart.
We thank Bill Chaudhry for critical reading of the article.
- 1999. Apoptosis in the pattern formation of the ventricular wall during mouse heart organogenesis. Anat Rec 256: 208–217. , , , , , , .
- 2001. Overlapping and differential localization of Bmp-2, Bmp-4, Msx-2 and apoptosis in the developing endocardial cushion and adjacent tissues of the developing mouse heart. Cell Tissue Res 305: 67–78. , , , .
- 1987. Origin of the mesenchymal tissue in the septum primum: a structural and ultrastructural study. J Mol Cell Cardiol 19: 641–651. , , , .
- 1981. F4/80, a monoclonal antibody directed specifically against the mouse macrophage. Eur J Immunol 11: 805–815. , .
- 2001. Double-outlet right ventricle and overriding tricuspid valve reflects disturbances of looping, myocardialization, endocardial cushion differentiation, and apoptosis in TGFb2-knockout mice. Circulation 103: 2745–2752. , , , , , , , , .
- 1996. The role of macrophages in clearing programmed cell death in the developing kidney. Anat Embryol 194: 341–348. , .
- 2002. Spatiotemporal and tissue specific distribution of apoptosis in the developing chick heart. Dev Dyn 223: 119–133. , , , .
- 1998. Epicardium-derived cells contribute a novel population to the myocardial wall and the atrioventricular cushions. Circ Res 82: 1043–1052. , , , , .
- 2000. Epicardial outgrowth inhibition leads to compensatory mesothelial outflow tract collar and abnormal cardiac septation and coronary formation. Circ Res 87: 969–971. , , , , .
- 1998. Versican expression is associated with chamber specification, septation and valvulogenesis in the developing mouse heart. Circ Res 83: 523–532. , .
- 1994. Macrophage recruitment during limb development and wound healing in the embryonic and foetal mouse. J Cell Sci 107: 1159–1167. , , , .
- 1979. Cell death during the development of the truncus and conus of the chick embryo heart. J Anat 129: 427–439. , .
- 1997. Programmed cell death in animal development. Cell 88: 347–354. , , .
- 2000. Fate of the mammalian cardiac neural crest. Development 127: 1607–1616. , , , , .
- 2002. Regulation of apoptosis in the endocardial cushions of the developing chick heart. Am J Physiol Cell Physiol 282: C1348–1360. , .
- 1983. Neural crest cells contribute to normal aorto-pulmonary septation. Science 220: 1059–1061. , , .
- 2003. Recruitment of intra- and extracardiac cells into the myocardial lineage during mouse development. Anat Rec 271: 303–314. , , , .
- 2002. Elevated transforming growth factor μ2 enhances apoptosis and contributes to abnormal outflow tract and aortic sac development in retinoic X receptor α knockout embryos. Development 129: 733–746. , , , .
- 1998. Neurofibromin modulation of ras activity is required for normal endocardial-mesenchymal interaction in the developing heart. Development 125: 4359–4367. , .
- 1996. In vivo inhibition of programmed cell death by local administration of FGF-2 and FGF-4 in the interdigital areas of the embryonic chick leg bud. Anat Embryol 193: 533–541. , , , .
- 1999. Does the subepicardial mesenchyme contribute myocardioblasts to the myocardium of the chick embryo heart? a quail-chick chimera study tracing the fate of the epicardial primordium. Anat Rec 255: 212–226. .
- 1991. Cell death during development of the nervous system. Annu Rev Neurosci 14: 453–501. .
- 2002. Experimental studies on the spatiotemporal expression of WT-1 and RALDH2 in the embryonic avian heart: a model for the regulation of myocardial and valvuloseptal development by epicardially derived cells (EPDCs). Dev Biol 247: 307–326. , , , , , , .
- 1975. Cell death in the morphogenesis and teratogenesis of the heart. Adv Anat Embryol Cell Biol 51: 3–99. .
- 1998. Neural crest cells in outflow tract septation of the embryonic chicken heart: differentiation and apoptosis. Dev Dyn 212: 373–384. , , .
- 2002. Initiation of apoptosis in the developing avian outflow tract myocardium. Dev Dynam 223: 469–482. , , , .
- 2000. Developmental patterning of the myocardium. Anat Rec 258: 319–337. , , , , .
- 1999. Myocardialization of the cardiac outflow tract. Dev Biol 212: 477–490. , , , , , , .
- 1998. Targeted disruption of the mouse Caspase 8 gene ablates cell death induction by the TNF receptors, Fas/Apo1, and DR3 and is lethal prenatally. Immunity 9: 267–276. , , , , , , , , , , , , , , , , , .
- 1999. Cell death in development. Cell 96: 245–254. , .
- 2001. Apoptosis is required for the proper formation of the ventriculo-arterial connections. Dev Biol. 240: 274–288. , , .
- 1998. Formation of the atrioventricular septal structures in the normal mouse. Circ Res 82: 645–656. , , .
- 1997. Is programmed cell death required for neural tube closure. Curr Biol 7: 281–284. , , .
- 2000. Mesenchymal cells engulf and clear apoptotic footplate cells in macrophageless PU.1 null mouse embryos. Development 127: 5245–5252. , , , , , , .
- 1998. Normal development of the outflow tract in the rat. Circ Res 82: 464–472. , , , , , , .
- 1999. Bending of the neural plate during mouse spinal neurulation is independent of actin microfilaments. Dev Dyn 215: 273–283. , .
- 1998. FADD: essential for embryo development and signaling from some, but not all, inducers of apoptosis. Science 279: 1954–1958. , , , , , , , , , , , , , .
- 2000. Requirement for Casper (c-FLIP) in regulation of death receptor-induced apoptosis and embryonic development. Immunity 12: 633–642. , , , , , , , , , , .
- 2000. Programmed cell death in the developing heart: regulation by BMP4 and FGF2. Dev Dyn 217: 388–400. , .
- 1996. Requirement for BMP signalling in interdigital apoptosis and scale formation. Science 272: 738–741. , .