|
Vascular cell adhesion molecule-1 (VCAM-1) is a member of the immunoglobulin superfamily, and as such shares considerable homology with other cell adhesion molecules (CAM's). VCAM-1 is the coreceptor for a4b1 integrin, a heterodimeric receptor that can mediate both cell-cell and cell-extracellular matrix interactions. a4b1 (VLA-4; CD49d/CD29) has known specificity for two ligands: VCAM-1 and fibronectin (FN), while VCAM-1 is only known to bind a4 integrins (Glukhova and Thiery, 1993; reviewed in Hemler, 1990).
The interaction between a4b1 and VCAM-1 is important for the differentiation of hematopoietic cells in the bone marrow (Miyake et al., 1991a), as well as for the alignment and fusion of secondary myoblasts during skeletal muscle development (Rosen et al., 1992). The importance of these interactions suggested that other developmental processes involving heterotypic cell-cell interactions might also require this receptor pair. For this reason, antibodies against VCAM-1 and the a4 integrin subunit were used to immunolabel frozen sections of mouse embryos collected throughout gestation.
Based on previous studies, we expected to see VCAM-1 and a4 on all skeletal muscle, particularly late in development. We also expected to see these proteins on hematopoietic tissues such as the yolk sac, liver, thymus, and spleen. We began our search by looking at these tissues to confirm our ability to detect VCAM-1 and a4 and especially to confirm the expression on skeletal muscle, which had previously been described only in culture (Rosen et al., 1992). Not only were VCAM-1 and a4 present in these tissues, but both were also localized to the extraembryonic membranes, heart, and central nervous system (CNS), and a4 was found on several forms of smooth muscle.
Materials and Methods:
Antibodies. PS/2 (Miyake et al., 1991a; diluted 1:5), R1-2 (Holtzmann et al., 1989a; diluted 1:2), and anti CD49d (Pharmingen; diluted 1:5) are rat anti-mouse monoclonal antibodies to a4. M/K-1 (Miyake et al., 1991a; diluted 1:2) and anti-NCAM-110 (Pharmingen; diluted 1:5) are rat anti-mouse monoclonal antibodies to VCAM-1. Rabbit antisera to laminin (Sanes et al., 1986; diluted 1:3000), FN (Collaborative Research; diluted 1:75), neurofilament (Dahl, 1983; diluted 1:3000), and von Willebrand Factor (vWF) (Dako; diluted 1:20,000) were used where indicated. Fluorescein goat anti-rat (Jackson Immunochemicals; diluted 1:200) and rhodamine donkey anti-rabbit (Jackson Immunochemicals; diluted 1:200) were used as secondary antibodies.
Histology. Timed pregnant and normal adult mice (outbred Swiss-Webster x Balb/c) were obtained from a colony maintained by the laboratory. The presence of vaginal plugs was used to define embryonic day 0 (E0); birth occured early on E19. Whole embryos or various tissues dissected from adult mice were placed in O.C.T. mounting medium (Miles) and quick frozen in liquid nitrogen. Ten µm sections were cut on a cryostat, mounted on SuperFrost Plus slides (Fisher), and stored at - 70°C until needed. Sections were fixed in methanol at -20°C and incubated with a mixture of the indicated monoclonal and polyclonal antibodies, diluted in PBS containing 2% fish gelatin (Sigma), at 4°C overnight. Then, sections were washed with PBS and incubated with a mixture of secondary antibodies diluted in PBS containing 2% fish gelatin for 1 hr at room temperature. After further washing, sections were mounted under Vectashield (Vector Laboratories) and observed with epifluorescent illumination. Similar patterns of immunolabeling were seen with each anti-a4 monoclonal antibody and with each anti-VCAM-1 monoclonal antibody.
Results:
Expression of VCAM-1 and a4 appears to be a property of all differentiating skeletal muscles
Previously, we found that VCAM-1 and a4b1 are expressed in a developmentally regulated fashion in mouse intercostal muscle (Rosen et al., 1992). a4 was evident as the premuscle masses began to fuse into primary myotubes at embryonic day 13 (E13), whereas VCAM-1 was not apparent until at least E15 as secondary myogenesis began to occur. The expression of a4 on premuscle masses in the intercostal region at E13 is shown in Figure 2.1a. As with the intercostal region, a4 was also evident at E13 on premuscle masses in the developing vertebral column and on the diaphragm, tongue and back (Figures 2.1b-d). Additionally, a4 was found on premuscle masses around the eye at E13 (data not shown). As we described previously in the intercostal region, VCAM-1 does not appear on skeletal muscle until at least E15 - VCAM-1 and a4 expression on differentiating muscle in the back is shown in Figures 2.1d and d', respectively. Together, these results demonstrate that expression of VCAM-1 and a4 is not restricted to the intercostal region, and suggests that it is characteristic of all differentiating skeletal muscle.
|
Colocalization of VCAM-1 and a4 on Vascular Smooth Muscle
Expression on vascular smooth muscle. The expression of VCAM-1 and a4 in developing skeletal muscle raised the question as to whether these receptors are expressed on smooth muscle. Figure 2.2a shows that a4 is expressed on the aorta at E15. Expression on the aorta was evident as early as E10 (Figure 2.9c), and it persisted around the adult vessel (data not shown). FN was abundant on both the endothelium and the surrounding smooth muscle layers of the aorta throughout embryogenesis (Figure 2.2b) and in the adult (data not shown), indicating that a4 and its extracellular matrix ligand, FN, colocalize throughout development in aortic vascular smooth muscle. In contrast to skeletal muscle, VCAM-1 was not detected on normal vascular smooth muscle; however we have found that its expression can be activated on vascular smooth muscle cells by inflammatory cytokines (Iademarco, et al., 1998).
|
Expression in the lung. Vascularization in the lung does not begin until around E15 after substantial differentiation, including repeated rounds of airway branching, has already occurred. Blood vessels then form from mesenchyme surrounding the airways. At E15, a4 and FN were found on the smooth muscle layers around newly forming blood vessels, and on the endothelium of these vessels (Figures 2.2c and c'). a4, but not FN, was also found on the surrounding mesenchyme, which gives rise to the endothelium and smooth muscle of the vessels. In contrast to vascular smooth muscle, no a4 was detected around airway smooth muscle. Although VCAM- 1 was not detected on newly forming vessels in the lung, it was evident along with a4 on the mesenchymal cells (Figure 2.2e).
Expression on smooth muscle in other tissues. The presence of a4 and FN on vascular smooth muscle prompted us to investigate the pattern of these proteins on other smooth muscles. Figure 2.3a shows that a4 is expressed in the lamina propria of adult intestinal villi. The lamina propria was delineated by double immunolabeling for laminin (Figure 2.3a'). Some of the a4 in the lamina propria appeared to be on immune cells that reside in this area; however, it was also evident on the smooth muscle layer. This smooth muscle layer was also positive for FN (data not shown). As with vascular smooth muscle, VCAM-1 was not detected on intestinal smooth muscle; however, it was found on mononuclear cells scattered throughout the intestinal villi (Figure 2.3b). VCAM-1 has been detected previously on macrophages, therefore, it is likely these VCAM-1 positive cells in the intestine are resident macrophages. Neither a4 nor VCAM-1 was detected on embryonic intestine.
|
The muscularis layer of the adult stomach was also found to express a4 (Figure 2.3e). However, in marked contrast to other tissues, VCAM-1 was evident on this smooth muscle layer (Figure 2.3f). VCAM-1 was also evident on cells scattered throughout the surrounding adventitia that are likely to be macrophages. While a4 appeared as streaks along the surface of the smooth muscle, VCAM-1 exhibited a punctate pattern on the cell surface. As with the intestine, neither a4 nor VCAM-1 was apparent in embryonic stomach (data not shown). Thus, the appearance of VCAM-1 and a4 on smooth muscle of the adult stomach and the presence of a4 on smooth muscle in the adult intestine is in contrast to the developmentally regulated pattern of a4 on smooth muscle of the trachea and esophagus.
Expression of VCAM-1 and a4 in immune tissues
At E13, a4 was evident on immune cells throughout the liver (Figure 2.4a); however, relatively little a4 was detected in adult liver, which is no longer involved in immune cell production (data not shown). VCAM-1 expression in embryonic (Figure 2.4b) and adult liver (data not shown) was confined to scattered cells, which presumably are macrophages. In the spleen, a4 was expressed on tightly packed lymphoid cells in splenic nodules of the white pulp (Figure 2.4c). The outline of the nodules is evident from FN immunolabeling of sheaths surrounding the nodules (Figure 2.4c'). a4 was also apparent on more loosely packed lymphoid cells in the surrounding red pulp. Immunolabeling for VCAM-1 was sparse in the white pulp (Figures 2.4d and e), but concentrated in the marginal zone between the nodules and red pulp. It is likely that many of the VCAM-1-positive cells are macrophages, which are known to concentrate in the marginal zone; however, some of the labeling may also be on reticular cells.
|
a4-positive cells were detected throughout the thymus, but they were most prominent in the medullar zone, where lymphoblasts and immature lymphocytes are loosely organized - few a4-positive cells were evident in the cortical zone, which consists mainly of small lymphocytes (Figure 2.4f). FN expression appeared to be confined to epithelial reticular cells, their desmosomal processes, and Hassall's corpuscles, which are composed of reticular cells (Figure 2.4f'). VCAM-1 expression was largely confined to the medullar region where intensely immunolabeled cells - probably macrophages - were apparent (Figure 2.4g).
Expression of VCAM-1 and a4 in the developing placenta
After implantation of the embryo, stromal cells in the uterus undergo a decidual reaction - they accumulate glycogen and lipids, and become polyploid. These decidual cells surround maternal blood vessels where they appear to disrupt surrounding basement membrane and interact directly with endothelial cells. Molecular mechanisms that are important for this interaction are unknown. Subsequently, giant trophoblasts followed by spongiotrophoblasts migrate through the labyrinth of the placenta, which constitutes the maternal/embryonic circulatory interface, and into the uterus. __ integrins have been shown previously to be important for trophoblast migration. Eventually, trophoblasts displace the endothelium of maternal vessels, and this is thought to be facilitated by the preceding activity of decidual cells. These early events are followed by formation of the placenta from the fusion of the chorion and allantois, extra-embryonic membranes of the fetus. The chorioallantoic membrane forms the fetal blood vessels of the placenta, and the stalk of the allantois develops into the umbilical cord.
|
a4 was present on mesodermal cells in the chorion of the developing placenta at E8, but was not expressed on trophoblasts (Figure 2.5b). Also, scattered a4-positive decidual cells were seen in the labyrinthine region of the developing placenta (Figure 2.5b). Although surrounded by FN, these cells did not appear to express FN on their surface (data not shown): furthermore, neither a4 nor VCAM-1 was evident on muscle cells of the myometrium (data not shown). Although VCAM-1 was not evident in the maternal placenta, VCAM-1 was strongly expressed on the allantois at E8 (Figure 2.5a). The existence of a4 on the chorion and VCAM-1 on the allantois suggest a role for this receptor pair in the fusion of these tissues to form the chorioallantoic membrane, which matures into the fetal aspect of the placenta. The importance of this interaction has been confirmed in mice lacking either receptor, a large portion of which die around E8.5 from a failure of chorioallantoic fusion and placentation (Gurtner, et al., 1995).
Expression of VCAM-1 and a4 in the yolk sac
As with other blood vessels, a4 was evident on the smooth muscle surrounding large vitelline vessels in the yolk sac at E13 (Figures 2.6a and b). The endothelium of these vessels expressed vWF (Figures 2.6a' and b'). Blood islands of the yolk sac, which are sites of hematopoiesis, were also positive for a4. Here, a4 was found on cells in the lumen of the vessels themselves (Figures 2.6b and c). In contrast to the vitelline vessels the endothelium of these blood islands was largely vWF-negative - only scattered staining of endothelium was apparent (Figures 2.6a' and b'). FN was also evident on blood islands; however, it also extended to areas immediately surrounding these vessels (Figure 2.6c'). A number of the vessels containing yolk sac blood islands appeared to have collapsed; however, vessels that still had relatively large lumens, and thus appeared to be active sites of hematopoiesis, immunolabeled intensely for VCAM-1 (Figure 2.6d; compare the lumens of vessels in Figures 2.6b and c to those in Figure 2.6d). By E13, the hematopoietic function of the yolk sac is diminishing as the liver becomes the principal site of hematopoiesis in the embryo.
|
Expression of VCAM-1 in the developing CNS
Sections of embryonic mouse brain and whole embryonic mouse were immunostained with rat anti-mouse antibodies to VCAM-1. VCAM-1 expression was found on the neuroepithelium in the developing spinal cord (Figure 2.7a), the diencephalon (Figure 2.7b), and the hippocampus (Figure 2.7c). VCAM-1 was also present on the choroid plexus (Figure 2.7d). This expression of VCAM-1 in the CNS maintained a distinct spatiotemporal pattern through late embryogenesis. Figures 2.7b shows that VCAM-1 is expressed in the ventricular zone of embryonic brain. VCAM-1 is concentrated on the surface of cell bodies in this region. However, there is additional immunostaining for VCAM-1 on cellular projections extending from the ventricular region; these projections are indicative of radial glia whose somata also reside in the ventricular zone. Cells in the ventricular zone also expressed nestin (Sheppard et al., 1995) and did not express neurofilament (Figure 2.7b'), a marker for differentiated neurons. These results suggest that VCAM-1 is expressed on progenitor cells and radial glia in the ventricular zone.
|
VCAM-1 was also found in the ventricular zone of the embryonic spinal cord (Figures 2.7a). As in the brain, VCAM-1 appeared to be concentrated on the surface of cell bodies, with lower concentrations on radial projections extending out of the ventricular zone. There was no overlap in immunostaining for VCAM-1 and a4 (Figure 2.7a'), suggesting that VCAM-1 works independent of its co-receptor in the CNS. Cells in the ventricular zone of the spinal cord also immunostained for nestin, and nestin-positive projections were evident extending from the ventricular zone across the spinal cord (Sheppard et al., 1995). As in the brain, we conclude that VCAM-1 is expressed on neural progenitor and radial glial cells in the spinal cord.
Expression of VCAM-1 in the CNS is developmentally regulated, and no immunostaining was evident in the adult mouse CNS (data not shown). Radial glial cells are transient during development and eventually give rise to astrocytes (Hirano and Goldman, 1988). VCAM-1 was not found on cells that were positive for glial fibrillary protein, which is a marker for astrocytes (Sheppard et al., 1995), suggesting that VCAM- 1 expression diminishes as radial glia take on properties of astrocytes. a4b1 integrin was identified on the neural crest derived peripheral nervous system, including the dorsal root ganglia and sensory nerve fibers, although these cells never contact the VCAM-1 positive neuroepithelium (Figure 2.7a').
Expression of VCAM-1 and a4 during cardiogenesis
Murine heart development involves the proliferation and differentiation of cells in the splanchnic mesoderm of the head referred to as the precardial plate (Manasek, 1968). Arising from hematopoietic tissue similar to the blood islands of the yolk-sac, the precardial plate forms two lateral cardiac primordia comprised of the thickened mesodermal cells of the promyocardium overlying the columnar proendocardial cells. By embryonic day 8 (E8), these primordia have been brought together ventral to the foregut and joined to form the primitive heart tube (DeRuiter, et al., 1992). Subsequently, the heart tube begins to fold upon itself, and the dorsal mesocardium extends villi to form a layer of epicardium, which encapsulates the heart and tethers it to the dorsal wall of the pericardial coelom (Manner, 1992). At E10, flexure has juxtaposed the bulbus cordis and common atrial chamber. The heart is divided into four chambers, over the next several days, through the generation and expansion of septa and through the growth of endocardial cushions. By E16, cardiogenesis is essentially complete except for continued growth into adulthood (reviewed in Kaufman, 1992).
VCAM-1 and a4 expression in E8 - E8.5 - Initial fusion of the lateral cardiac primordia occurs at E8, so that by E8.5, the primitive heart tube is completely formed. At E8, FN is evident in the basal lamina lining the endocardium as well as in the surrounding cardiac jelly; FN is also seen around the pericardium (Figure 2.8a). VCAM-1 is strictly localized to the myocardial cells that surround the joined endocardium which is negative (Figure 2.8b). There is no a4 in the heart at this stage (data not shown). At E8.5, FN is still evident in the cardiac jelly, endocardial basal lamina, and on the pericardium; FN is also evident in the basal lamina of the common atrial chamber (Figure 2.8c). VCAM-1 is localized to the myocardium but not the endocardium of the common ventricular chamber, however it appears localized mostly to the ventral portion of the myocadium in the common atrial chamber (Figure 2.8d).
|
VCAM-1 and a4 expression in E9 - E10 - During E9, flexure of the primitive heart tube juxtaposes the bulbus cordis and common atrial chamber. Also at this stage, formation of the epicardium and endocardial cushions begin. FN is evident in the cardiac jelly and endocardium, and it is also seen in the forming villi of the dorsal mesocardium (Figure 2.9a). Expression of VCAM-1 is restricted to the compact layer of myocardial cells in both the atrial and ventricular chambers, although the bulbus cordis is negative (Figure 2.9b). Expression of a4 appears on the villi of the dorsal mesocardium and on cells around the outer perimeter of the myocardium, presumably epicardial precursor cells (Figure 2.9c). Previous work has shown these epicardial precursor cells to originate from the dorsal mesocardial villi (Manner, 1993). Some expression of a4 is also evident on cushion mesenchymal cells in the cardiac jelly of the common atrial chamber (Figure 2.9c).
|
During E10, endocardial cells lining the truncus arteriosus and atrioventricular canal dedifferentiate and migrate into the cardiac jelly to form the mesenchymal layers, from which the endocardial cushions arise. FN is evident underlying the endocardium and throughout the cardiac jelly, particularly in the region of the endocardial cushions (Figure 2.9d). VCAM-1 expression is unchanged from E9 (data not shown). Expression of a4 is now evident on the cushion mesenchymal cells and the fully formed epicardium (Figure 2.9e).
VCAM-1 and a4 expression at E14 - In the week following E9, the heart forms septa bisecting and separating the atrial and ventricular chambers, and dividing the truncus arteriosus into the aorta and pulmonary artery. At E14, FN is evident on the epicardium, myocardium, and endocardium of the ventricular and atrial chambers; FN is especially evident in the area of the endocardial cushion (Figure 2.10a).__VCAM-1 expression changes significantly by E14: it is evident weakly on the compact myocardium but strongly on the developing ventricular septum (Figure 2.10b). The expression of a4 is still detectable on the epicardium and on the endocardial cushion, and can also be seen on the outflow tracts (Figures 2.10c). At this stage myosin is evident in the myocardium, particularly in the ventricular chambers (Figure 2.10d).
|
There is little morphological change in the heart between E16 and birth: VCAM-1 expression continues to diminish until absent in the adult, and expression of a4 is lost in the epicardium and atrioventricular cushion but continues on the aortic arch and pulmonary artery into adulthood (Sheppard, et al., 1994).
Discussion:
The events of vertebrate morphogenesis are becoming more clearly defined in terms of the required movements of cell populations and the remodeling of tissue. As a consequence, the importance of cellular adhesion interactions is now being recognized. Here we describe the spatiotemporal expression of VCAM-1 and a4 integrin in several aspects of murine development, including myogenesis of the skeletal musculature, differentiation of the neuroepithelium of the CNS, adhesion of the placenta, function of hematopoietic tissues and morphogenesis of the heart. Our data suggest that a4 and VCAM-1 act cooperatively and independently in varying ways as cells interact heterotypically and homotypically to form several essential embryonic tissues.
We have examined the patterns of VCAM-1 and a4 integrin expression in embryonic and adult mouse tissues. Prior to beginning this study, VCAM-1 and a4 were found to be expressed in adult bone marrow (Miyake et al., 1991a; Miyake et al., 199b; Ryan et al., 1991). a4 was found on maturing immune cells and on undifferentiated cells, whereas VCAM-1 was present on stromal cells. The interaction between a4 integrin and VCAM-1 mediates attachment of immune cells to stromal cells, and reagents that block this interaction inhibit maturation of the immune cells. We found that VCAM-1 and a4 were expressed at embryonic sites of hematopoiesis such as liver and yolk sac, suggesting that these receptors could have parallel roles in embryonic and adult hematopoiesis.
Previously, we found that VCAM-1 and a4 were expressed in a developmentally regulated pattern in skeletal muscle from the intercostal region (Rosen et al., 1992). Here, we extend this observation by showing that these receptors appear to be expressed in all differentiating skeletal muscle, thus implying that they have a general role in skeletal muscle differentiation.
We found that a4 is also expressed on vascular smooth muscle. However, unlike skeletal muscle, a4 was present on vessels very early during embryogenesis and it persisted into adulthood. Also, in contrast to skeletal muscle, VCAM-1 was not present on normal vascular smooth muscle. Like a4, FN was found on smooth muscle and endothelium in the lung, suggesting that the interaction between a4 integrins and FN could be an early event during vascularization that is important for organization of smooth muscle cells around newly forming vessels. Although VCAM-1 was not evident on vessels in the lung, it was present along with a4 on surrounding mesenchyme, which gives rise to the vascular endothelium and smooth muscle. Therefore, it is conceivable that an interaction between a4 integrins and VCAM-1 on the undifferentiated mesenchyme could be important for early angiogenic events, whereas a subsequent interaction between a4 integrins and FN on newly forming vessels could be important for condensation of smooth muscle layers around the vessels.
The pattern of a4 and VCAM-1 on smooth muscle layers appears to be under intricate tissue-specific control. For example, a4 was not detected on the smooth muscle of airways in the developing lung, demonstrating a difference between vascular and airway smooth muscle. In contrast to the more distal airways, the trachea does express a4 on its smooth muscle layer. But, expression of a4 on the trachea and on the esophagus is transient: a4 was evident at E13 but not at E15. After expression of a4 dissipated on the smooth muscle of the esophagus, both a4 and VCAM-1 appeared distally on smooth muscle of the adult stomach and a4 appeared on the adult intestine. These patterns suggest a proximal to distal gradient of a4 and VCAM-1 in the digestive tract.
The finding of VCAM-1 on smooth muscle in stomach was surprising because VCAM-1 was not evident on smooth muscle in other tissues, and it suggests a unique role for this receptor in the specialized smooth muscle of the stomach. In contrast to skeletal muscle where we have identified promoter elements that control the pattern of VCAM-1 expression (Jesse et al., 1998), nothing is yet known about how the pattern of this receptor is controlled in gastric smooth muscle.
As with smooth muscle, the pattern of VCAM-1 on vascular endothelium was quite restricted: VCAM-1 was apparent on vessels in only two tissues, the kidney and the myometrium. And in these two tissues, expression was only apparent on a subset of vessels. Because VCAM-1 is known to be expressed on endothelium in response to inflammatory cytokines (Osborn et al., 1989; Rice et al., 1989; Iademarco et al., 1992), it is conceivable that the vessels which expressed VCAM-1 were simply undergoing an inflammatory response. Alternatively, such vessels could be located in areas that are rich in cytokines, which are important for local differentiation events.
Cardiogenesis. Cardiac morphogenesis is accomplished through the migration and aggregation of precursor cells to form the three layers of the four-chambered adult heart. Interactions between cells and each other or their environment are needed to drive this extensive tissue remodeling required for cardiomorphogenesis. Adhesion molecules, such as PECAM-1 (Baldwin, et al., 1994a) and b1 integrins (Drake, et al., 1992), and extra-cellular matrix molecules, such as laminin (Davis, et al., 1989) and fibronectin (FN) (Icardo, et al., 1992), have been identified in specific layers of the heart, suggesting a molecular basis for these essential interactions.
After the joining of the two lateral cardiac primordia begins to shape the primitive heart, VCAM-1 is strongly evident throughout the forming myocardium. This expression coincides with the expression of myosin and desmin in the myocardium previously described in rat embryogenesis (Baldwin, et al., 1991). The only known ligand of VCAM- 1, a4 expression appears, at this stage, in the branchial arches which contribute to the aortic arches and truncus arteriosus (Noden, 1991), which do not express VCAM-1. In the absence of a4, the activity of VCAM-1 must involve an unknown interaction if it is to have an effective role in cardiogenesis.
The formation of the epicardium during E9 requires the dynamic remodeling of the dorsal mesocardium, and results in the extension of villi. Subsequently, pre-epicardial cells migrate from the dorsal mesocardium and across the outer surface of the compact myocardium (Manner, 1992). The placement of a shell membrane between the dorsal mesocardium and heart in chick embryos prevents the formation of an epicardium indicating that cellular interactions are required (Manner, 1993). The dorsal mesocardial origin of the epicardial cells is supported by the patterns of labeling we see for a4 during epicardium formation. The expression of b1 integrin (Drake, et al., 1992) and extracellular matrix molecules (Gallagher, et al., 1993) in the forming dorsal mesocardium and epicardium in chick embryos has similarly been demonstrated. These observations suggest a role for adhesive cell-matrix interactions governing the formation of mesocardial villi and adhesive cell-cell interactions later governing formation of the epicardium. Furthermore, when VCAM-1 expression or a4 integrin subunit expression was eliminated in mice, the heart did not form an epicardium (Baldwin and Buck, 1994; Kwee, et al., 1995; Yang, et al., 1995). Here we suggest, that as the dorsal mesocardium forms a4-positive villi, pre-epicardial cells from these villi utilize the interaction of these counter-receptors to migrate onto and upon the VCAM-1 positive compact myocardium to surround the heart to form the epicardium.
The morphological division, or septation, of the cardiac chambers and tracts is commonly achieved by the expansion and subsequent fusion of two cell masses. The interventricular septum for example, results from the expansion of the muscular portion of the ventricular septum and its fusion with the expanding atrioventricular cushion. Between E10 and E16, localization of VCAM-1 expression becomes increasingly restricted to the muscular ventricular septum. The importance of VCAM-1 in the genesis of the muscular ventricular septum is confirmed by VCAM-1 knock-out mice, which die by E12 with an underdeveloped, or in some cases no, ventricular septum (Kwee et al., 1995). The atrioventricular cushion develops from the dedifferentiation of endocardial cells. These cells express a4 as they migrate into the cardiac jelly which is composed mainly of collagen types I and IV, laminin, and FN (Little, et al., 1989). It has been demonstrated in vitro, that endocardial cushion cells require multiple components for migration into an extracellular matrix (Markwald, et al., 1984; Davis, et al., 1989). Formation of the atrioventricular septum can be prevented by injection of anti-FN antibodies into developing chicks in vivo (Icardo, et al., 1992), a result reminiscent of the congenital heart defect persistent atrioventricular canal disorder (Rogers and Edwards, 1948). Although a4 and VCAM-1 are expressed in the atrioventricular cushion and muscular ventricular septum respectively, the endocardial lining around each of these portions of the developing interventricular septum prevents the interaction of these co-receptors until fusion is complete.
Our studies indicate that VCAM-1 and the a4 integrins have important roles in the morphogenesis of the heart. The spatiotemporal patterns of expression implicate these adhesion molecules in the formation of two of the principal layers of the heart, specifically the myocardium and the epicardium, while a similar member of the immunoglobulin superfamily, PECAM-1, have been localized spatiotemporally to the third principal layer, the endocardium (Baldwin, et al., 1994). Expression of VCAM-1 and a4 has also been implicated in the formation of the major anatomical divisions of the heart, specifically the septa. Although classically considered as a receptor ligand pair, the expression of these molecules during cardiogenesis, as elsewhere during embryogenesis, indicates that both can function in the absence of the other.
It will be interesting to determine how expression of VCAM-1 is controlled during development. For example, is there a common mechanism such that some promoter elements that are important for expression in embryonic skeletal muscle are also important for expression in other developing and adult muscular tissues, specifically the heart, or are there multiple tissue-specific mechanisms controlling the pattern of this receptor? The studies presented here, examining the expression of VCAM-1 and a4 in the mouse, provide the foundation necessary to address what controls the pattern of expression of these receptors in developing tissues, but also support the present model describing roles for these receptors during tissue differentiation.
Chapter 1 | Chapter 2 | Chapter 3 | Chapter 4 | References
