J Mol
Cell
Cardiol
23,
1411-1422
Distribution
Cynthia
(1991)
ofthe Neural Cell Adhesion During Heart Development L. Burroqhsl,
Micbiko
Watanabe*
Molecule
and Dennis
(NCAM)
E. Morsel*
lDepartment of Anatomy, Medical College of Ohio, PO Box 10008, Toledo, OH43699, USA 2Department of Pediatrics, D ivision of Pediatric Cardiology, Case Western Reserve University, 2074 Abington Road, Cleveland, OH 44106, USA (Received 9 January 1990, accepted in revisedform 30 July 1991) C. L. BURROUGHS, M. WATANABE AND D. E. MORSE. Distribution of the Neural Cell Adhesion Molecule (NCAM) During Heart Development.Jou~lofMofcculnrand Cellular Cardiology (1991) 23, 1431-1422. The neural cell adhesion molecule, NCAM, was localized in the embryonic chick heart from Hamburger-Hamilton stage 14 up to hatching and in the adult heart. A monoclonal antibody directed to NCAM was used with the indirect antibody technique to stain frozen sections with immunoperoxidase. The myocardium showed immunoreactivity at stages 15 and 21, with little to no staining of epicardium, endocardium or atrioventricular endocardial cushion tissue. At stage 22, additional immunoreactivity was found in the endocardium ofboth the atrial septum and the atrial and ventricular surfaces of the atrioventricular cushions. Endocardial-derived mesenchymal cells within the cushions were also immunostained for NCAM. A gradient of NCAM staining was evident in the ventricular wall by stage 16. The staining intensity in the myocardium subjacent to the epicardium was less than found near the ventricular lumen. Biochemical analyses revealed that the embryonic heart expresses polysialylated NCAM. Upon desialylation with the endoneuraminidase Endo-N, the predominant heart NCAM has an apparent molecular weight of 155 to 160 kDa, which is distinct in size from the predominant forms found in embryonic chick nervous system (180, 140 and 120 kDa). NCAM expression is regionally regulated in the heart. The pattern of its expression is consistent with our hypothesis that it is involved in (1) differentiation of the atrial and ventricular walls, (2) fusion of the atrial septum with the endocardial cushions, (3) fusion of the endocardial cushions, and (4) formation and remodeling of ventricular trabeculae. KEY WORDS: Endocardial
NCAM; Cushions.
Myocyte;
Heart;
Embryo;
Chick;
Introduction The embryonic vertebrate heart efficiently and effectively propels blood to other rapidly proliferating systems while undergoing major restructuring to form septa, valves, trabeculae, an intrinsic impulse generating system and a coronary circulation. The embryonic cardiomyocytes must be sufficiently adherent to each other to tolerate the forces imposed by the rhythmic contractions of the heart yet suitably plastic
in
their
unions
to
permit
tissue
re-
modeling. Several proteins are associated with cell-cell adhesion during morphogenesis. This study examines the expression of one of these molecules, the neural cell adhesion molecule (NCAM), during normal cardiogenesis in the embryo and adult chicken. ‘To
whom
correspondence
0022-2828/91/121411
should + 12 $03.00/O
Endocardium;
Trabeculae;
Septation;
Immunostain;
NCAM serves as a homophilic ligand in the formation of cell-cell bonds (Rutishauser et al., 1982). It is a cell surface glycoprotein expressed in a variety of early embryonic and adult tissues including neurons, glia and skeletal muscle myoblasts and myotubes. NCAM may influence such basic developmental processes as cell migration, segregation of cells into discrete areas within a tissue, axonal guidance and the formation of synapses (for reviews see Edelman, 1986; Rutishauser and Jessel, 1988). Regulation of the function of this molecule can be achieved by controlling its quantity (for reviews see Rutishauser, 1984, 1986) and polysialic acid 1986; Edelman, content (Rutishauser et al., 1988). NCAM has been detected immunohistologically in the developing heart of various
be addressed. 0 1991 Academic
Press
Limited
1412
C. L. Burroughs
species. NCAM was observed at the end of gastrulation in the presumptive cardiac region in the chick. By Hamburger-Hamilton stage 23 (Hamburger and Hamilton, 1951), it was restricted to the myocardium and absent in the epicardium and endocardium (Thiery et al., 1982; Hoffman et al., 1985). Expression of NCAM was studied in rat hearts from 12-day embryos to post-natal and adult stages cells (Wharton et al., 1989). Myocardial stained for NCAM, while the endo- and epicardium did not. In addition, it was noted that cardiac nerves and ganglia and the endocardial cushions were NCAM positive. The immunostaining intensity declined during development and became restricted in the adult to conduction tissue and valve cusps. The study of human heart (Gordon et al., 1990) revealed that the sinoatrial and atrioventricular nodes and the valvular cusps were stained more prominently for NCAM than surrounding tissues. Furthermore NCAM, though typically down-regulated in the adult, was expressed in high amounts in cardiac transplants and in certain disease states. The aim of this investigation was to document the presence and form of NCAM in specific regions of the chick heart. This was accomplished in the embryo and the adult by correlation of known morphological events with changes in the distribution of NCAM using immunohistochemical staining of frozen sections and immunoblot analysis.
Materials
and Methods
Fixation and preparation of specimens White Leghorn chick embryos were harvested from Hamburger-Hamilton (H-H) stages 12 and 14 to 31 inclusive. Thereafter stages 36, 39, 40, 43 and 46 were utilized. Adult chicken hearts were also collected. Whole embryos up to H-H stage 27 were immersed in fixative at 4OC for 1 to 3 h. The isolated hearts from older embryos and adults were either immediately immersion-fixed or perfused, and subsequently immersed in fixative for 1 to 3 h at 4OC. Tissue was fixed in periodate-lysine-paraformaldehyde PW fixative (McClean and Nakane, 1974) and was infiltrated in changes of 7 % and 15 % sucrose in phosphate buffer each at 4’C for 3 to 4 h.
et al.
This was followed by a third change in a 25% sucrose/l0 % glycerin solution for 3 to 4 h at 4X. Hearts were then frozen in Tissue-Tek OCT compound in a dry ice/acetone slurry. Frozen sections, 5 to 15 pm thick were cut with an IEC cryostat, collected on gelatin-coated slides, air-dried for 30min, and immunostained. Hearts were sectioned in the coronal, sagittal or transverse plane.
Immunohistochemistry In order to reduce background staining, sections were incubated for 15 min in 50 mM NH&l in Hank’s balanced salt solution (HBSS) followed by 30 min in 5 to 10% normal goat serum in HBSS (NGS/HBSS). All antibodies were diluted in 5 to 10% NGSIHBSS. A mouse monoclonal antibody 5E (1: 100) was utilized as the first antibody. Incubation was in humidified containers for either 2 h at room temperature or overnight at 4°C. Slides were then washed in four changes of phosphate buffered saline (PBS) for 5 to 10 min each and the second antibody was applied for 2 h at room temperature. The second antibody was peroxidase-conjugated goat IgG (1: 100). Sections were again washed in four changes of PBS for 5 to 10min each, washed in one change of Tris buffer (pH 7.5), and then incubated in diaminobenzidine (DAB) reaction mixture (60 mg DAB/300 ml Tris buffer, pH 7.5, with 0.03 % H202) for 10 to 30 min. The slides were rinsed in distilled water and coverslipped. The specificity of this procedure was tested by alternate negative control sections incubated with pre-immune serum, NGS/HBSS or monoclonal antibody no. 12 IgG in place of the first antibody. Monoclonal antibody no. 12 against Drosophila vitellogenin (Watanabe et al., 1982) was used as a negative control to an irrelevant protein. It is of the same subtype IgGi as the primary antibody (Frelinger and Rutishauser, 1986; Watanabe et al., 1986). Pre-treatment to eliminate endogenous peroxidase activity of red blood cells was not routinely performed. Reactivity in these morphologically distinct cells served as a positive control for the DAB reaction. Tissues known to be positive for NCAM (neural tube, somites) were included in immunostained samples as positive controls. Cardiac tissue examined in this part of the study included the
NCAM developing atria1 septum, atrioventricular canals.
outflow
tract
in Heart and
Immunoadsorption Immunoadsorption (Tosney et al., 1986) was used to concentrate NCAM and to separate it from more abundant heart proteins such as myosin. One volume of tissue was pelleted by centrifugation (13 000 g, 5 min), solubilized in 10~01 of Nonidet-P40 buffer (0.5% NP-40, 0.8% NaCl, 0.02% KCl, 0.02% KH2P04, 0.015 % Na,HPO,, 200 KIU Aprotinin) by sonification and mixed overnight with Sepharose beads (BioRad, Affi-Gel 10) coupled to anti-NCAM monoclonal antibody 5E, which is directed to an extracellular region on the polypeptide chain of NCAM (Frelinger and Rutishauser, 1986; Watanabe et al., 1986). When using heart tissue from animals of stage 36 or older, the homogenate was pre-adsorbed with control beads in order to remove the bulk of myosin which nonspecifically binds. After the beads were incubated with the tissue extracts, they were washed five times in the Nonidet-P40 buffer, and prepared for Western blot analysis as described below. The combination of immunoadsorption and Western blot analysis allowed us to concentrate NCAM for optimal detection and to purify the NCAM away from abundant proteins such as myosin which would distort the protein bands of the gel (Tosney et al., 1986).
Glycosidase treatments Endoneuraminidase N (Endo-N), the soluble virion-free enzyme from bacteriophage KIF, was 200- to 300-fold purified from KIFinduced lysate of Escherichia coli (Vimr et al.. 1984). This enzyme has no detectable proteolytic or exoneuraminidase activity. It was used at a concentration of c. 25 U/gg of tissue. This concentration removes the bulk of the polysialic acids from brain tissue which has a higher content of heavily polysialyiated NCAM than does cardiac tissue. Endo-N digestion was carried out in phosphate buffer, pH 7.4, on ice for 15 to 60min. Western blot analysis Protein
preparations
were
solubilized
by
1413
Development
boiling for 2 to 3min in two-times concentrated Laemmli (1970) sample buffer (0.0625MTris buffer at pH6.8, 2% SDS, 5 % 2-mercaptoethanol, 10 % glycerol, 0.001% bromophenol biue). The proteins were separated on 7% polyacrylamide gels by SDS-PAGE and transferred to nitrocellulose filters (Schleicher and Schuell, 0.45pm) by electrophoresis (Hoefer, TE 42 Transphor with TE 51 power supply set at 0.5A for 3 to 12 h). The NCAM on the nitrocellulose was detected using the indirect immunostaining procedure (Towbin et al., 1979; Johnson et al., 1984). The filters were incubated with rabbit antisera or monoclonal antibodies directed embryonic against chick brain NCAM, washed, incubated in antibody conjugated with peroxidase and directed to IgGs of the appropriate species, and washed. Bound peroxidase was detected using the chromogenic substrate 4-chloro- 1-naphthol (Hawkes et al., 1982). Pre-stained proteins (Rainbow Markers, Amersham) were used as molecular weight standards. As a control for non-specific binding of polypeptides, beads with their active sites blocked by treatment with 1 M ethanolamine hydrochloride or coupled to control monoclonal antibody no. 12 (see earlier section) directed to an irrelevant antigen were incubated with solubilized samples. For positive control, extracts of embryonic brain and optic chiasm were incubated with the antibody-derivatized beads. Embryonic brain was used as a source of tissue to demonstrate the SDS-PAGE profile of NCAM with high polysialic acid levels. Optic chiasm from the hatched chick or adult was used as a source to demonstrate the location of neuronal NCAM polypeptides at 180, 140 and 120 kDa.
Results Myocardium and epicardium NCAM was detected in the atrial, ventricular and septal myocardium at all embryonic stages of this study [Fig. l(a),(b)]. The adult ventricular and atrial myocardia were also positive for NCAM [Figs. l(c)-(e), 5(f)]. There was a decrease in NCAM immunostaining intensity within the myocardium with development. In the ventricle a gradient of immunostaining intensity was evident in that trabeculae carnae
1414
C. L. Burroughs
et al.
FIGURE 1. The myocardium stained with immunoperoxidase for NCAM in the embryonic [(a),(b)] and adult [(c),(d)] heart. (a) Transverse section through the stage 14 chicken embryo, with the myocardial tissue (M) of the tubular heart, the somites (So) and the neural tube (NT) stained for NCAM, and epithelium of the stomach (St) negative for NCAM. (b)Frontal section at stage 27; atria1 (A) and ventricular (LV and RV) myocardia, including trabeculae, are NCAM-positive. The adult ventricular myocytes were also NCAM-positive along cell surfaces as shown in sections parallel(c) and perpendicular(d) to the longitudinal axis of the myofibers. A blood vessel (BV) and surrounding connective tissue clearly detected with phase optics(e) were negative for NCAM (d). In (d) the DAB deposition in red blood cells (rbc) is due to endogenous peroxidase and was evident in control sections (not shown). Bars: (a) 1 mm; (b)0.5 mm; (c)-(e), 0.1 mm.
stained with a greater intensity than mural myocardium. This became more dramatic at progressive stages of development (Fig. 2). The epicardium was difficult to distinguish at early stages due to the flattened nature of this mesothelium, which is closely applied to the myocardium. In perikaryal regions where these cells were discernible, they were NCAMnegative [Fig. 3(a)]. At more advanced stages and in the adult, all components of the thickened epicardium remained NCAM-negative IFig. 3(b)-(d)].
Endocasdium Endocardial cells were characterized by a modulation of NCAM expression. Its presence coincided with morphological changes involved in restructuring the interior of the heart. At early stages the staining of the endocardium varied [Fig. 4(a),(c),(d)]. The endocardium of the conotruncal ridges of the outflow tract was NCAM-positive [Fig. 4(e)]. Cells of the ridge matrix were also positive. A similar pattern of NCAM expression was
FIGURE 2. A gradient of NCAM immunostaining intensity was apparent in the ventricular heads) was more intensely stained than the mural myocardium of stages 16 [(a),(a’)], 21 [(b),(b’)], and (d) correspond to location of arrows in (a’).(b’),(r’) and (a’), respectively. V. ventricular
wall ofthe embryonic heart. 23 [(c).(c’)] and 40 [(d),(d’)]. chambers. Bars = 0.1 mm.
Myocardium of the trabeculae (arrowThe locations of arrows in (a),(b),(c)
C. L. Burroughs
1416
et al.
(a)
FIGURE 3. NCAM-negative epicardial cells (arrowheads and EP) were flattened positive myocardial cells (My) at stages 14 (a) and 27 (b) and in the adult ventricular chamber; AS, atrial septum. Bars: (a) 50/.~m; (b)-(d) 100~111.
observed during atrioventricular (AV) cushion formation [Fig. 4(f)]. An unstained endocardium characterized the entire atrial septum prior to stage 22 [Fig. 5(a)]. Subsequently these cells at the free margin of the septum became NCAM-positive [Fig. 5(b)]. Cells within the expanded subendocardial space (SES) were also immunostained. The endocardium of the atria1 surfaces of the AV cushions was reactive for NCAM at this stage. The staining pattern of the septal and cushion components remained unchanged after fusion. Remnants of the SES persisted for several stages and separated atria1 myocytes from cushion mesenchymal cells [Fig. 5(c)]. In regions of the septum which became thin and ruptured to form foramina secunda [Fig. 5(d),(e)] the endocardium remained unreactive.
closely applied to the NCAM(c) and atrial (d) regions. A, atrial and
Biochemical analysis The major polypeptide form from stage 29 cardiac tissue had apparent molecular weights of 155 to 160 kDa after Endo-N treatment [Fig. 6(a), lane + 1, although the 180 kDa form of NCAM was also detectable. This compares with 180, 140 and 120 kDa polypeptides from chicken optic nerve [Fig. 6(a), lane on]. Before Endo-N treatment the NCAM in stage 29 cardiac tissue was heterogeneously polysialylated and migrated as a smear from 155 to >230 kDa. Distinct bands of presumably desialylated protein with relative mobilities of 155 to 160 kDa and 180 kDa were detectable within the smear [Fig. 6(a), lane - 1. The forms of NCAM expressed in cardiac tissue from stages 19 to 39 were qualitatively similar in polysialylation [Fig. 6(b)]. The moderate sialylation of embryonic heart
NCAM
in Heart
Development
1417
FIGURE 4. (a) and (c) Intensely stained myocardium in stage 15 embryos. In sagittal section the neural tube (NT) and somites (S) were also NCAM-positive. (b)Control section did not show the presence of NCAM. (c)The acellular cardiac jelly ( l ) was NCAM-negative. Endocardial cells at this stage varied in their degree of immunostaining from dark (large arrowhead) to light (small arrowhead). Under phase optics (d) the darker-staining cells were hypertrophic and/or had processes that extended into the cardiac jelly. The developing truncal cushions (TC) of the outflow tract (OT) at stage 24 (e) and endocardial cushions (ECC) of the atrioventricular canal at stage 22 (f) also had NCAM-positive endocardial (arrowheads) and cushion matrix cells. A, atrium; V, ventricle. Bars: (a),(b),(e) and (f)0.5mm; (c) and (d) 100gm.
NCAMs was less than that found in embryonic brain NCAM [Fig. 6(b), lane br]. The 155 to 160 kDa cardiac specific form of NCAM was the most abundant form detected from stage 19 up to stage 35. The 180 kDa NCAM polypeptide was evident in heavily loaded lanes probed with the sensitive iodinated second antibody method as early as stages 19 and 24
(not shown). The 140 kDa NCAM polypeptide became detectable in stage 31 cardiac tissue and became as abundant as the 155 to 160 kDa form by stage 36. A faint band at 120 kDa was visible in sialylated NCAM from stages 38 and 39. It was indistinguishable in Endo-N treated samples.
1418
C. L. Burroughs
et al.
FIGURE 5. (a) and (b) In frontal sections at stages 21 and 22, the myocardium (My) of the atria1 septum was ~OSICIVC for NCAM. At stage 21 the myocardium ofthe atria1 septum was separated from the NCAM-negative endocardiunr (En) by the subendocardial space. (b) The endocardium at the leading edge of the elongated atria1 septum (arrowhead) became immunoreactive at stage 22. (c)The stage 25 atria1 septum had fused with the endocardial cushions, but the subrndocardial space (star) of the atrial septum could still be distinguished from the surrounding endocardial cushion tissue. (d)Conspicuous, NCAM-negative, rounded cells (arrowhead) were found on the surface of the atria1 septum at stage 25. (e) Secondary perforations (open arrow) and lightly stained regions of attenuated atria1 tissues (arrowhead) were cvidcnt at stage 28. (f) Myocytes of the adult atria1 septum continued to express NCAM on their surfaces. Red blood cells stained due to endogenous peroxidase activity (arrowheads). Bars: (a),(b).(c),(e) and (f)O. 1 mm; (d)0.05 mm.
Discussion Cardiocytes express NCAM prior to formation of the tubular heart through adulthood (Thiery et al., 1982; Wharton et al., 1989; Gordon et al., 1990). At the light microscopic level, NCAM appears to be evenly distributed over the entire surface of embryonic myocardial cells as is the case for immature skeletal myocytes. Embryonic cardiocytes progressively increase their area of apposed plasma membranes to form ionically and mechanically coupled complexes known collectively as intercalated discs (Manasek, 1968). The gap juncadhesion tion and calcium-dependent molecule components of the intercalated disc
are affected in their function or maturation by NCAM immunological perturbation (Keane et al., 1988; Rutishauser et al., 1988; Watanabe et al., 1989; Knudsen et al., 1990). Thus, NCAM is potentially involved in regulating the formation, maturation and function of these junctions between myocardial cells in embryonic and adult chickens. Regional variability in NCAM expression in the heart has been observed. The atrioventricular node, sinus node and valvular cusps have been noted to have higher levels of NCAM expression than surrounding tissues of the fetal heart (Gordon et al., 1990). Downregulation of NCAM expression has been
NCAM
in Heart
Development
1419
FIGURE 6. The major NCAM polypeptides from chicken nervous trssue such as the optic nerve (on) had apparent molecular weights of 180 kDa, 140kDa and 120kDa [panel (a), arrows from top to bottom and panel (c)l. (a) After removal of polysialic acid with Endo-N the major NCAM polypeptides from stage 29 cardiac tissue has apparent molecular weights of 155 to 160 kDa (lane + ); before Endo-N treatment, cardiac tissue is polysialylated with hetcrw geneous molecular weight, but protein bands at 155 to 160 kDa and 180 kDa are detectable (lane - ). The forms ofcardiac NCAM before Endo-N treatment (b) and after Endo-N treatment (c) are shown with the exception of stage 34. Heart tissue from this particular stage was not Endo-N-treated to allow side-by-side comparison of polysialylated and desiaiylated forms of heart NCAM. The embryonic stages of tissue source are indicated across the top. In (b) highly polys~nlylatrd embryonic brain NCAM (br) is shown
detected with maturation in rat and human heart (Wharton et al., 1989; Gordon et al., 1990). In this study, we have noted a ventricular gradient of NCAM expression which may reflect the differences in maturation between the mural myocytes and those of the trabeculae. The elevated immunoreactivity of trabecular myocytes is consistent with the findings of Wharton et al. (1989) in the rat heart prior to 18 days of gestation. In vitro assays have shown that alterations of NCAM concentration can dramatically affect vesicle aggregation (Hoffman and Edelman, 1983) and neurite outgrowth (Doherty et al., 1990). Therefore even subtle differences in NCAM expression, which can be demonstrated at the histological level, may have a profound physiological influence on the behavior of the cardiomyocytes. The regulated pattern of NCAM expression
in skeletal muscle strongly supports a role for NCAM in neuron-to-myocyte interactions. Normal adult skeletal muscle expresses NCAM only around the neuromuscular junction (Covault and Sanes, 1985, 1986; Rieger et al. , 1985; Tosney et al., 1986). During establishment of innervation, during myopathy or after surgical or pharmacological denervation, NCAM is homogeneously expressed at the surface of the muscle fiber. A similarly close correlation of NCAM expression and innervation has not been made for cardiac cells. In the rat, most of the cardiomyocytes lose NCAM immunoreactivity postnatally and only express NCAM in the adult heart in discrete regions around the coronary arteries and the ventricular septum (Wharton et al., 1989). Those cells that are positive retain NCAM reactivity in both intercalated disc regions and on surfaces that do not necessaril)
1420
C. L. Burroughs
correspond to regions of interaction with nerve synapses. This is in contrast to the case of the mature myoneural junctions of skeletal muscle (Covault and Sanes, 1986). We find that chick cardiomyocytes retain NCAM immunoreactivity throughout embryonic, fetal and adult stages with some decrease in the intensity of immunostaining with maturation. The evidence is strong that NCAM is necessary for interactions between differentiating cardiomyocytes and between a subpopulation of mature cardiomyocytes before, during and after cardiac innervation. This trend has been observed in other systems (Thiery et al., 1982; Crossin et al., 1985; Tosney et al., 1986; Rutishauser, 1986; Watanabe et al., 1989). More recently, Gordon et al. (1990) have made correlations that indirectly support NCAM’s role in innervation. The sinus node and atrioventricular node, those areas that receive dense innervation, are more highly NCAMpositive at 15 weeks of gestation in the human. Furthermore, human transplanted hearts, which are denervated and never reinnervated express higher levels of NCAM than the normal adult heart. Evidence exists from in vitro studies that the interaction between brain membrane vesicle and cardiomyocyte are capable of being attenuated by anti-NCAM (Hoffman et al., 1985). In summary, the correlation of NCAM expression with cardiomyocyte innervation is not as striking as for skeletal muscle. The difference in patterns of expression between skeletal and NCAM cardiac muscle may reflect differences in the morphology and physiology of the respective nerve-muscle junctions. The role of NCAM in cardiac innervation is unclear. An apparent discrepancy exists as to whether the endocardium is negative (Thiery et al., 1982; Wharton et al., 1989) or positive (Mjaatvedt and Markwald, 1989) for NCAM. This is possibly due to the variability in staining of the endocardium with developmental stage and anatomical region. We find low levels of NCAM on endocardial cell surfaces beginning at stage 15. More intensely positive cells are present at specific regions directly involved in the internal remodeling of the heart. These regions include the facing surfaces of the endocardial cushions, the atrial surface of the endocardial cushions, the free edge of the atria1 septum and the endocardial
et al. ridges of the outflow tract. The endocardium of each of these regions is imminently invohcd in fusion as part of the septation of the heart. The spatiotemporal expression of NCAM on the free edge of the atria1 septum correlates directly with numerous endocardial cell modifications which occur just prior to fusion with the endocardial cushions (Arrechedera et al., 1987). The suggestion is that NCAM may have a role in directing and establishing the adherence of two opposing endothelial surfaces. In the case of the cushions of the atrioventricular canal and outflow tract, NCAM may be related to the differentiation of mesenchymal cells from the endocardium. These mesenchymal cells separate from the endocardium to populate the cushion matrix (Bernanke and Markwald, 1982; Markwald et al., 1977). Our observations support those of Mjaatvedt and Markwald (1989) that a downregulation of NCAM expression exists as mesenchymal cells migrate away from the endocardial surface. We deduce from SDS-PAGE and immunoblot profiles that heart tissue from chicken embryo stages 19 to 39 expresses polysialylated NCAM. It appears as a diffuse streak, characteristic of heterogeneously glycosylated NCAM, which focuses to a broad band of apparent molecular weight 155 to 160 kDa when desialylated by treatment with Endo-N, an endoneuraminidase specific for polysialic acid of NCAM. Since there is evidence that the polysialic acid is a general attenuator of cell-cell interactions (Rutishauser et al., 1988), the implication is that embryonic cardiocyte cell-cell interactions may be regulated by not only NCAM, but also by its polysialic acid. Developmental changes in the size of NCAM polypeptides expressed were detected in Endo-N-treated samples. At later stages NCAM polypeptide bands became detectable at 140 and 180 kDa in addition to the 155 to 160 kDa band. Since we analyzed all heart tissue together, these NCAM proteins may reflect components contributed by nervous tissue as well as by the myocardial and endocardial cells. Sequence analysis has shown that some of the skeletal muscle specific forms of NCAM resemble the 120 kDa NCAM isolated from neural tissue (Dickson et al., 1987). These lack a transmembrane domain and are bound to the plasma membrane by a phosphatidyl-
NCAM
in Heart
inositol linkage (Moore et al., 1987). In addition, these skeletal muscle NCAMs have an extracellular stretch of 37 amino acids near the membrane which is not found in the predominant nervous system NCAMs. A cDNA clone from an embryonic chick heart library was found to have a 93 base-pair insert similar, but not identical, to the skeletal-specific domain identified in human skeletal myocytes (Prediger et al., 1988). Sequence analysis of an NCAM genomic clone allowed identification of four exons in this region. The significance of the multiple NCAM polypeptide forms is not known; however, it has been postulated that the lack of a transmembrane domain may affect the stability of NCAM within the membrane (Gennarini et al., 1984). The differences in size of cytoplasmic domains may have profound effects on interactions of NCAM with cytoplasmic components such as cytoskeletal elements. Differences in the extracellular domain adjacent to the membrane may have consequences for how the NCAM interacts laterally with other molecules within the plasma membrane or how flexible it is as a hinged molecule (Hall and Rutishauser, 1987; and Becker et al., 1989). The regulation stabilization of the special mechanical and electrotonic interactions between cardiocytes may require specialized NCAM molecules.
Development
1421
In summary, this paper presents evidence that NCAM is associated with cell-cell interactions in the embryonic and adult chicken heart. The most obvious possibility is that NCAM functions in myocyte-myocyte interactions since this molecule is present on cardiomyocytes from the stages of tubular heart to adult. A significant observation is that NCAM is present on the surfaces of endocardial cells in regions where differentiation to mesenchymal cells and contact during tissue fusion occur. This presents the intriguing possibility that NCAM may not only function in the modulation of cell-cell interactions, but also in tissue-tissue interactions. Any impairment in such a basic process as tissue fusion could have major ramifications for morphogenesis of the heart. Resulting heart defects could include septal defects, endocardial cushion defects or aberrant truncal septation. Acknowledgements The authors thank Dr Giora Ben-Shachar for allowing us to photograph his histology slides of the ventricle, and Dr Urs Rutishauser and laboratory members for generosity with prepared reagents and advice. This work was supported in part by grants from American Heart Association, Northeast Ohio Affiliate and NIH R29 HL38172 OIAI.
References ARRECHEDERA H, ALVAREZ M, STRAUSS M, AYESTA C (1987) Origin of mesenchymal tissue in the septum primum: a structural and ultrastructural study. J Mel Cell Cardiol 19: 641-651. BECKER JW, ERICKSON HP, HOFFMAN S, CUNNINGHAM BA (1989) Topology ofcell adhesion molecules. Proc Nat1 Arad Sci USA 86: 1088-1092. BERNANKE DH, MARKWALD RR (1982) Migratory behavior of cardiac cushion tissue cells in collagen lattice culture system. Dev Biol 91: 235-245. COVAULT J, SANES JR (1985) Neural cell adhesion molecule (N-CAM) accumulates in denervated and paralyzed skeletal muscle. Proc Nat1 Acad Sci USA 82: 4544-4548. COVAULT J, SANES JR (1986) Distribution of N-CAM in synaptic and extrasynaptic portions of developing and adult skeletal muscle. J Mol Cell Cardiol 102: 716-730. CROSSIN KL, CHUONC CM, EDELMAN GM (1985) Expression sequences of cell adhesion molecule. Proc Nat1 Acad SCI USA 82: 6942-6946. DICKSON G, GOWER HJ, BARTON CH, PRENTICE HM, ELSOM VL, MOORE SE, Cox RD. QUINN C, PUTT W, WALSH FS (1987) Human muscle neural cell adhesion molecule (N-CAM): identification of a muscle-specific sequence in the extracellular domain. Cell 50: 1119-1130. DOHERTY P, FRUNS M, SEATON P, DICKSON G, BARTON CH, SEARS TA, WALSH FS (1990) A threshold effect of the major isoforms of NCAM of neurite outgrowth. Nature 343: 464-466. EDELMAN GM (1986) Cell adhesion molecules in the regulation of animal form and tissue pattern. Annu Rev Cell Biol2: 81-116. FRELINGER AL, RUTISHAUSER U (1986) Topography of N-CAM structural determinants. II. Placement of monoclonal antibodv eoitooes. 1 Mol Cell Cardiol 103: 1729-1737. GENNA~IN; G,* D&OSTINI-BAZIN H, GORIDIS C (1984) Studies on the transmembrane disposition of the neural cell adhesion molecule N-CAM. The use of liposome-inserted radioiodinated N-CAM to study its transbilayer orientation. EurJ Biochem 142: 65-73.
1422
C. L. Burroughs
et
al.
L, WHARTON J, MOORE SE, WALSH FS, Moscoso JC. PENKETH R, WALLWORK J. TA~L.OR KM, YA(:OUH MH, POLAK JM (1990) Myocardial localization and isoforms of neural cell adhesion molecule (N-CAM) in the developing and transplanted human heart. J Clin Invest 86: 1293-1300. HALL AK, RUTISHAUSER U (1987) Visualization of neural cell adhesion molecule by electron microscopy. J Cell Biol 104: 1579-1586. HAMBURGER V, HAMILTON HL (1951) A series of normal stages in the development of the chick embryo. J Morph01 88: 49-92. HAWKES R, NIDAY E, GORDON J (1982) A dot-immunobinding assay for monoclonal and other antibodies. Anal Biochem 119: 142-147. HOFFMAN S, EDELMAN GM (1983) Kinetics of homophilic binding by embryonic and adult forms of the neural cell adhesion molecule. Proc Nat1 Acad Sci USA 80: 5762-5766. HOFFMAN S, GRUMET M, EDELMAN GM (1985) Cell adhesion molecules in the histogenesis of nerve and muscle. In: Cardiac Morphogmesis, edited by VJ Ferrans, G Rosenquist, C Weinstein. New York, Else&r Science, pp 36-43. JOHNSON DA, GAUTSCH JW, SPORTSMAN JR, ELDER JH (1984) Improved technique utilizing nonfat dry milk for analysis of proteins and nucleic acids transferred to nitrocellulose. Gene Anal Tech 1: 3-8. KEANE RW, MEHTA PP, ROSE B, HONIC LS, LOEWENSTEIN WR, RUTISHAUSER U (1988) Neural differentiation, NCAM-mediated adhesion and gap junctional communication in neuroectoderm. A study in vitro. J Cell Biol 106: 1307-1319. KNUDSEN KA, MCELWEE SA, MYERS L (1990) A role for the neural cell adhesion molecule, NCAM, in myoblast interaction during myogenesis. Dev Biol 138: 159-168. LAEMMLI UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680-685. MCCLEAN IW, NAKANE PK (1974) Periodate-lysine-paraformaldehyde fixative. A new fixative for immunoelectron microscopy. J Histochem Cytochem 22: 1077-1083. MANASEK F (1968) Histogenesis of the embryonic myocardium. Am J Cardiol 25(2): 149-168. MARKWALD RR, FITZHARRIS TP, MANASEK F (1977) Structural development of endocardial cushions. Am J Anat 148: 85-120. MJAATVEDT CH, MARKWALD RR (1989) Induction of an epithelial-mesenchymal transition by an in ~ivo adheron-like complex. Dev Biol 136: 118-128. MOORE SE, THOMPSON J, KIRKNESS V, DICKSON JE, WALSH FS (1987) Skeletal muscle neural cell adhesion molecule (N-CAM): changes in protein and mRNA species during myogenesis of muscle cell lines. J Cell Biol 105: 1377-1386. PREDICER EA, HOFFMAN S, EDELMAN GM, CUNNINGHAM BA (1988) Four exons encode a 93 base-pair insert in three neural cell adhesion molecule mRNAs specific for chicken heart and skeletal muscle. Proc Nat1 Acad Sci USA 85: 9616-9620. RIECER F, GRUMET M, EDELMAN GM (1985) N-CAM at the vertebrate neuromuscular junction. J Cell Biol 101: 285-293. RUTISHAUSER U (1984) Developmental biology of a neural cell adhesion molecule. Nature 310: 549-554. RUTISHAUSER U (1986) Differential cell adhesion through spatial and temporal variations of NCAM. Trends Neurosci g(8): 374-377. RUTISHAUSER U, ACHESON A, HALL AK, MANN DM, SUNSHINEJ (1988) The neural cell adhesion molecule (NCAM) as a regulator of cell-cell interactions. Science 240: 53-57. RUTISHAUSER U, HOFFMAN S, EDELMAN GM (1982) Binding properties of a cell adhesion molecule from neural tissue. Proc Nat1 Acad Sci USA 79: 685-689. RUTISHAUSER U, JESSEL TM (1988) Cell adhesion molecules in vertebrate neural development. Physiol Rev 68: 819-857. THIERY J-P, DUBAND J-L, RUTISHAUSER U, EDELMAN GM (1982) Cell adhesion molecules in early chicken embryogenesis. Proc Nat1 Acad Sci USA 79: 6737-6741, TOSNEY KW, WATANABE M, LANDMESSER L, RUTISHAUSER U (1986) The distribution of NCAM in the chick hindlimb during axon outgrowth and synaptogenesis. Dev Biol 114: 437-452. TOWBIN H, STAEHELIN T, GORDON J (1979) Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some application. Proc Nat1 Acad Sci USA 76: 4350-4354. VIMR ER, MCCOY RD, VOLLCER HF, WILKISON NC, TROY FA (1984) Use of prokaryotic-derived probes to identify poly(sialic acid) in neonatal neuronal membranes. Proc Nat1 Acad Sci USA 81: 1971-1975. WATANABE M, FRELINGER AL, RUTISHAUSER U (1986) Topography of N-CAM structural and functional determinants. I. Classification of monoclonal antibody epitopes. J Cell Biol 103: 1721-1727. WATANABE M, KOBAYASHI H, RUTISHAUSER U, KATAR M, ALCALA J, MAISEL H (1989) NCAM in the differentiation of embryonic lens tissue. Dev Biol 135: 414-423. WATANABE M, PALTER KB, MAHOWALD AP (1982) Monoclonal antibodies recognize different components of the Dmsophila embryos. J Cell Biol95(2): 148a. WHARTON J, GORDON L, WALSH FS, FLANICAN TP, MOORE SE, POLAK JM (1989) Neural cell adhesion molecule (NCAM) expression during cardiac development in the rat. Brain Res 483: 170-176. GORDON
Cell
Cardiol
23,
1411-1422
Distribution
Cynthia
(1991)
ofthe Neural Cell Adhesion During Heart Development L. Burroqhsl,
Micbiko
Watanabe*
Molecule
and Dennis
(NCAM)
E. Morsel*
lDepartment of Anatomy, Medical College of Ohio, PO Box 10008, Toledo, OH43699, USA 2Department of Pediatrics, D ivision of Pediatric Cardiology, Case Western Reserve University, 2074 Abington Road, Cleveland, OH 44106, USA (Received 9 January 1990, accepted in revisedform 30 July 1991) C. L. BURROUGHS, M. WATANABE AND D. E. MORSE. Distribution of the Neural Cell Adhesion Molecule (NCAM) During Heart Development.Jou~lofMofcculnrand Cellular Cardiology (1991) 23, 1431-1422. The neural cell adhesion molecule, NCAM, was localized in the embryonic chick heart from Hamburger-Hamilton stage 14 up to hatching and in the adult heart. A monoclonal antibody directed to NCAM was used with the indirect antibody technique to stain frozen sections with immunoperoxidase. The myocardium showed immunoreactivity at stages 15 and 21, with little to no staining of epicardium, endocardium or atrioventricular endocardial cushion tissue. At stage 22, additional immunoreactivity was found in the endocardium ofboth the atrial septum and the atrial and ventricular surfaces of the atrioventricular cushions. Endocardial-derived mesenchymal cells within the cushions were also immunostained for NCAM. A gradient of NCAM staining was evident in the ventricular wall by stage 16. The staining intensity in the myocardium subjacent to the epicardium was less than found near the ventricular lumen. Biochemical analyses revealed that the embryonic heart expresses polysialylated NCAM. Upon desialylation with the endoneuraminidase Endo-N, the predominant heart NCAM has an apparent molecular weight of 155 to 160 kDa, which is distinct in size from the predominant forms found in embryonic chick nervous system (180, 140 and 120 kDa). NCAM expression is regionally regulated in the heart. The pattern of its expression is consistent with our hypothesis that it is involved in (1) differentiation of the atrial and ventricular walls, (2) fusion of the atrial septum with the endocardial cushions, (3) fusion of the endocardial cushions, and (4) formation and remodeling of ventricular trabeculae. KEY WORDS: Endocardial
NCAM; Cushions.
Myocyte;
Heart;
Embryo;
Chick;
Introduction The embryonic vertebrate heart efficiently and effectively propels blood to other rapidly proliferating systems while undergoing major restructuring to form septa, valves, trabeculae, an intrinsic impulse generating system and a coronary circulation. The embryonic cardiomyocytes must be sufficiently adherent to each other to tolerate the forces imposed by the rhythmic contractions of the heart yet suitably plastic
in
their
unions
to
permit
tissue
re-
modeling. Several proteins are associated with cell-cell adhesion during morphogenesis. This study examines the expression of one of these molecules, the neural cell adhesion molecule (NCAM), during normal cardiogenesis in the embryo and adult chicken. ‘To
whom
correspondence
0022-2828/91/121411
should + 12 $03.00/O
Endocardium;
Trabeculae;
Septation;
Immunostain;
NCAM serves as a homophilic ligand in the formation of cell-cell bonds (Rutishauser et al., 1982). It is a cell surface glycoprotein expressed in a variety of early embryonic and adult tissues including neurons, glia and skeletal muscle myoblasts and myotubes. NCAM may influence such basic developmental processes as cell migration, segregation of cells into discrete areas within a tissue, axonal guidance and the formation of synapses (for reviews see Edelman, 1986; Rutishauser and Jessel, 1988). Regulation of the function of this molecule can be achieved by controlling its quantity (for reviews see Rutishauser, 1984, 1986) and polysialic acid 1986; Edelman, content (Rutishauser et al., 1988). NCAM has been detected immunohistologically in the developing heart of various
be addressed. 0 1991 Academic
Press
Limited
1412
C. L. Burroughs
species. NCAM was observed at the end of gastrulation in the presumptive cardiac region in the chick. By Hamburger-Hamilton stage 23 (Hamburger and Hamilton, 1951), it was restricted to the myocardium and absent in the epicardium and endocardium (Thiery et al., 1982; Hoffman et al., 1985). Expression of NCAM was studied in rat hearts from 12-day embryos to post-natal and adult stages cells (Wharton et al., 1989). Myocardial stained for NCAM, while the endo- and epicardium did not. In addition, it was noted that cardiac nerves and ganglia and the endocardial cushions were NCAM positive. The immunostaining intensity declined during development and became restricted in the adult to conduction tissue and valve cusps. The study of human heart (Gordon et al., 1990) revealed that the sinoatrial and atrioventricular nodes and the valvular cusps were stained more prominently for NCAM than surrounding tissues. Furthermore NCAM, though typically down-regulated in the adult, was expressed in high amounts in cardiac transplants and in certain disease states. The aim of this investigation was to document the presence and form of NCAM in specific regions of the chick heart. This was accomplished in the embryo and the adult by correlation of known morphological events with changes in the distribution of NCAM using immunohistochemical staining of frozen sections and immunoblot analysis.
Materials
and Methods
Fixation and preparation of specimens White Leghorn chick embryos were harvested from Hamburger-Hamilton (H-H) stages 12 and 14 to 31 inclusive. Thereafter stages 36, 39, 40, 43 and 46 were utilized. Adult chicken hearts were also collected. Whole embryos up to H-H stage 27 were immersed in fixative at 4OC for 1 to 3 h. The isolated hearts from older embryos and adults were either immediately immersion-fixed or perfused, and subsequently immersed in fixative for 1 to 3 h at 4OC. Tissue was fixed in periodate-lysine-paraformaldehyde PW fixative (McClean and Nakane, 1974) and was infiltrated in changes of 7 % and 15 % sucrose in phosphate buffer each at 4’C for 3 to 4 h.
et al.
This was followed by a third change in a 25% sucrose/l0 % glycerin solution for 3 to 4 h at 4X. Hearts were then frozen in Tissue-Tek OCT compound in a dry ice/acetone slurry. Frozen sections, 5 to 15 pm thick were cut with an IEC cryostat, collected on gelatin-coated slides, air-dried for 30min, and immunostained. Hearts were sectioned in the coronal, sagittal or transverse plane.
Immunohistochemistry In order to reduce background staining, sections were incubated for 15 min in 50 mM NH&l in Hank’s balanced salt solution (HBSS) followed by 30 min in 5 to 10% normal goat serum in HBSS (NGS/HBSS). All antibodies were diluted in 5 to 10% NGSIHBSS. A mouse monoclonal antibody 5E (1: 100) was utilized as the first antibody. Incubation was in humidified containers for either 2 h at room temperature or overnight at 4°C. Slides were then washed in four changes of phosphate buffered saline (PBS) for 5 to 10 min each and the second antibody was applied for 2 h at room temperature. The second antibody was peroxidase-conjugated goat IgG (1: 100). Sections were again washed in four changes of PBS for 5 to 10min each, washed in one change of Tris buffer (pH 7.5), and then incubated in diaminobenzidine (DAB) reaction mixture (60 mg DAB/300 ml Tris buffer, pH 7.5, with 0.03 % H202) for 10 to 30 min. The slides were rinsed in distilled water and coverslipped. The specificity of this procedure was tested by alternate negative control sections incubated with pre-immune serum, NGS/HBSS or monoclonal antibody no. 12 IgG in place of the first antibody. Monoclonal antibody no. 12 against Drosophila vitellogenin (Watanabe et al., 1982) was used as a negative control to an irrelevant protein. It is of the same subtype IgGi as the primary antibody (Frelinger and Rutishauser, 1986; Watanabe et al., 1986). Pre-treatment to eliminate endogenous peroxidase activity of red blood cells was not routinely performed. Reactivity in these morphologically distinct cells served as a positive control for the DAB reaction. Tissues known to be positive for NCAM (neural tube, somites) were included in immunostained samples as positive controls. Cardiac tissue examined in this part of the study included the
NCAM developing atria1 septum, atrioventricular canals.
outflow
tract
in Heart and
Immunoadsorption Immunoadsorption (Tosney et al., 1986) was used to concentrate NCAM and to separate it from more abundant heart proteins such as myosin. One volume of tissue was pelleted by centrifugation (13 000 g, 5 min), solubilized in 10~01 of Nonidet-P40 buffer (0.5% NP-40, 0.8% NaCl, 0.02% KCl, 0.02% KH2P04, 0.015 % Na,HPO,, 200 KIU Aprotinin) by sonification and mixed overnight with Sepharose beads (BioRad, Affi-Gel 10) coupled to anti-NCAM monoclonal antibody 5E, which is directed to an extracellular region on the polypeptide chain of NCAM (Frelinger and Rutishauser, 1986; Watanabe et al., 1986). When using heart tissue from animals of stage 36 or older, the homogenate was pre-adsorbed with control beads in order to remove the bulk of myosin which nonspecifically binds. After the beads were incubated with the tissue extracts, they were washed five times in the Nonidet-P40 buffer, and prepared for Western blot analysis as described below. The combination of immunoadsorption and Western blot analysis allowed us to concentrate NCAM for optimal detection and to purify the NCAM away from abundant proteins such as myosin which would distort the protein bands of the gel (Tosney et al., 1986).
Glycosidase treatments Endoneuraminidase N (Endo-N), the soluble virion-free enzyme from bacteriophage KIF, was 200- to 300-fold purified from KIFinduced lysate of Escherichia coli (Vimr et al.. 1984). This enzyme has no detectable proteolytic or exoneuraminidase activity. It was used at a concentration of c. 25 U/gg of tissue. This concentration removes the bulk of the polysialic acids from brain tissue which has a higher content of heavily polysialyiated NCAM than does cardiac tissue. Endo-N digestion was carried out in phosphate buffer, pH 7.4, on ice for 15 to 60min. Western blot analysis Protein
preparations
were
solubilized
by
1413
Development
boiling for 2 to 3min in two-times concentrated Laemmli (1970) sample buffer (0.0625MTris buffer at pH6.8, 2% SDS, 5 % 2-mercaptoethanol, 10 % glycerol, 0.001% bromophenol biue). The proteins were separated on 7% polyacrylamide gels by SDS-PAGE and transferred to nitrocellulose filters (Schleicher and Schuell, 0.45pm) by electrophoresis (Hoefer, TE 42 Transphor with TE 51 power supply set at 0.5A for 3 to 12 h). The NCAM on the nitrocellulose was detected using the indirect immunostaining procedure (Towbin et al., 1979; Johnson et al., 1984). The filters were incubated with rabbit antisera or monoclonal antibodies directed embryonic against chick brain NCAM, washed, incubated in antibody conjugated with peroxidase and directed to IgGs of the appropriate species, and washed. Bound peroxidase was detected using the chromogenic substrate 4-chloro- 1-naphthol (Hawkes et al., 1982). Pre-stained proteins (Rainbow Markers, Amersham) were used as molecular weight standards. As a control for non-specific binding of polypeptides, beads with their active sites blocked by treatment with 1 M ethanolamine hydrochloride or coupled to control monoclonal antibody no. 12 (see earlier section) directed to an irrelevant antigen were incubated with solubilized samples. For positive control, extracts of embryonic brain and optic chiasm were incubated with the antibody-derivatized beads. Embryonic brain was used as a source of tissue to demonstrate the SDS-PAGE profile of NCAM with high polysialic acid levels. Optic chiasm from the hatched chick or adult was used as a source to demonstrate the location of neuronal NCAM polypeptides at 180, 140 and 120 kDa.
Results Myocardium and epicardium NCAM was detected in the atrial, ventricular and septal myocardium at all embryonic stages of this study [Fig. l(a),(b)]. The adult ventricular and atrial myocardia were also positive for NCAM [Figs. l(c)-(e), 5(f)]. There was a decrease in NCAM immunostaining intensity within the myocardium with development. In the ventricle a gradient of immunostaining intensity was evident in that trabeculae carnae
1414
C. L. Burroughs
et al.
FIGURE 1. The myocardium stained with immunoperoxidase for NCAM in the embryonic [(a),(b)] and adult [(c),(d)] heart. (a) Transverse section through the stage 14 chicken embryo, with the myocardial tissue (M) of the tubular heart, the somites (So) and the neural tube (NT) stained for NCAM, and epithelium of the stomach (St) negative for NCAM. (b)Frontal section at stage 27; atria1 (A) and ventricular (LV and RV) myocardia, including trabeculae, are NCAM-positive. The adult ventricular myocytes were also NCAM-positive along cell surfaces as shown in sections parallel(c) and perpendicular(d) to the longitudinal axis of the myofibers. A blood vessel (BV) and surrounding connective tissue clearly detected with phase optics(e) were negative for NCAM (d). In (d) the DAB deposition in red blood cells (rbc) is due to endogenous peroxidase and was evident in control sections (not shown). Bars: (a) 1 mm; (b)0.5 mm; (c)-(e), 0.1 mm.
stained with a greater intensity than mural myocardium. This became more dramatic at progressive stages of development (Fig. 2). The epicardium was difficult to distinguish at early stages due to the flattened nature of this mesothelium, which is closely applied to the myocardium. In perikaryal regions where these cells were discernible, they were NCAMnegative [Fig. 3(a)]. At more advanced stages and in the adult, all components of the thickened epicardium remained NCAM-negative IFig. 3(b)-(d)].
Endocasdium Endocardial cells were characterized by a modulation of NCAM expression. Its presence coincided with morphological changes involved in restructuring the interior of the heart. At early stages the staining of the endocardium varied [Fig. 4(a),(c),(d)]. The endocardium of the conotruncal ridges of the outflow tract was NCAM-positive [Fig. 4(e)]. Cells of the ridge matrix were also positive. A similar pattern of NCAM expression was
FIGURE 2. A gradient of NCAM immunostaining intensity was apparent in the ventricular heads) was more intensely stained than the mural myocardium of stages 16 [(a),(a’)], 21 [(b),(b’)], and (d) correspond to location of arrows in (a’).(b’),(r’) and (a’), respectively. V. ventricular
wall ofthe embryonic heart. 23 [(c).(c’)] and 40 [(d),(d’)]. chambers. Bars = 0.1 mm.
Myocardium of the trabeculae (arrowThe locations of arrows in (a),(b),(c)
C. L. Burroughs
1416
et al.
(a)
FIGURE 3. NCAM-negative epicardial cells (arrowheads and EP) were flattened positive myocardial cells (My) at stages 14 (a) and 27 (b) and in the adult ventricular chamber; AS, atrial septum. Bars: (a) 50/.~m; (b)-(d) 100~111.
observed during atrioventricular (AV) cushion formation [Fig. 4(f)]. An unstained endocardium characterized the entire atrial septum prior to stage 22 [Fig. 5(a)]. Subsequently these cells at the free margin of the septum became NCAM-positive [Fig. 5(b)]. Cells within the expanded subendocardial space (SES) were also immunostained. The endocardium of the atria1 surfaces of the AV cushions was reactive for NCAM at this stage. The staining pattern of the septal and cushion components remained unchanged after fusion. Remnants of the SES persisted for several stages and separated atria1 myocytes from cushion mesenchymal cells [Fig. 5(c)]. In regions of the septum which became thin and ruptured to form foramina secunda [Fig. 5(d),(e)] the endocardium remained unreactive.
closely applied to the NCAM(c) and atrial (d) regions. A, atrial and
Biochemical analysis The major polypeptide form from stage 29 cardiac tissue had apparent molecular weights of 155 to 160 kDa after Endo-N treatment [Fig. 6(a), lane + 1, although the 180 kDa form of NCAM was also detectable. This compares with 180, 140 and 120 kDa polypeptides from chicken optic nerve [Fig. 6(a), lane on]. Before Endo-N treatment the NCAM in stage 29 cardiac tissue was heterogeneously polysialylated and migrated as a smear from 155 to >230 kDa. Distinct bands of presumably desialylated protein with relative mobilities of 155 to 160 kDa and 180 kDa were detectable within the smear [Fig. 6(a), lane - 1. The forms of NCAM expressed in cardiac tissue from stages 19 to 39 were qualitatively similar in polysialylation [Fig. 6(b)]. The moderate sialylation of embryonic heart
NCAM
in Heart
Development
1417
FIGURE 4. (a) and (c) Intensely stained myocardium in stage 15 embryos. In sagittal section the neural tube (NT) and somites (S) were also NCAM-positive. (b)Control section did not show the presence of NCAM. (c)The acellular cardiac jelly ( l ) was NCAM-negative. Endocardial cells at this stage varied in their degree of immunostaining from dark (large arrowhead) to light (small arrowhead). Under phase optics (d) the darker-staining cells were hypertrophic and/or had processes that extended into the cardiac jelly. The developing truncal cushions (TC) of the outflow tract (OT) at stage 24 (e) and endocardial cushions (ECC) of the atrioventricular canal at stage 22 (f) also had NCAM-positive endocardial (arrowheads) and cushion matrix cells. A, atrium; V, ventricle. Bars: (a),(b),(e) and (f)0.5mm; (c) and (d) 100gm.
NCAMs was less than that found in embryonic brain NCAM [Fig. 6(b), lane br]. The 155 to 160 kDa cardiac specific form of NCAM was the most abundant form detected from stage 19 up to stage 35. The 180 kDa NCAM polypeptide was evident in heavily loaded lanes probed with the sensitive iodinated second antibody method as early as stages 19 and 24
(not shown). The 140 kDa NCAM polypeptide became detectable in stage 31 cardiac tissue and became as abundant as the 155 to 160 kDa form by stage 36. A faint band at 120 kDa was visible in sialylated NCAM from stages 38 and 39. It was indistinguishable in Endo-N treated samples.
1418
C. L. Burroughs
et al.
FIGURE 5. (a) and (b) In frontal sections at stages 21 and 22, the myocardium (My) of the atria1 septum was ~OSICIVC for NCAM. At stage 21 the myocardium ofthe atria1 septum was separated from the NCAM-negative endocardiunr (En) by the subendocardial space. (b) The endocardium at the leading edge of the elongated atria1 septum (arrowhead) became immunoreactive at stage 22. (c)The stage 25 atria1 septum had fused with the endocardial cushions, but the subrndocardial space (star) of the atrial septum could still be distinguished from the surrounding endocardial cushion tissue. (d)Conspicuous, NCAM-negative, rounded cells (arrowhead) were found on the surface of the atria1 septum at stage 25. (e) Secondary perforations (open arrow) and lightly stained regions of attenuated atria1 tissues (arrowhead) were cvidcnt at stage 28. (f) Myocytes of the adult atria1 septum continued to express NCAM on their surfaces. Red blood cells stained due to endogenous peroxidase activity (arrowheads). Bars: (a),(b).(c),(e) and (f)O. 1 mm; (d)0.05 mm.
Discussion Cardiocytes express NCAM prior to formation of the tubular heart through adulthood (Thiery et al., 1982; Wharton et al., 1989; Gordon et al., 1990). At the light microscopic level, NCAM appears to be evenly distributed over the entire surface of embryonic myocardial cells as is the case for immature skeletal myocytes. Embryonic cardiocytes progressively increase their area of apposed plasma membranes to form ionically and mechanically coupled complexes known collectively as intercalated discs (Manasek, 1968). The gap juncadhesion tion and calcium-dependent molecule components of the intercalated disc
are affected in their function or maturation by NCAM immunological perturbation (Keane et al., 1988; Rutishauser et al., 1988; Watanabe et al., 1989; Knudsen et al., 1990). Thus, NCAM is potentially involved in regulating the formation, maturation and function of these junctions between myocardial cells in embryonic and adult chickens. Regional variability in NCAM expression in the heart has been observed. The atrioventricular node, sinus node and valvular cusps have been noted to have higher levels of NCAM expression than surrounding tissues of the fetal heart (Gordon et al., 1990). Downregulation of NCAM expression has been
NCAM
in Heart
Development
1419
FIGURE 6. The major NCAM polypeptides from chicken nervous trssue such as the optic nerve (on) had apparent molecular weights of 180 kDa, 140kDa and 120kDa [panel (a), arrows from top to bottom and panel (c)l. (a) After removal of polysialic acid with Endo-N the major NCAM polypeptides from stage 29 cardiac tissue has apparent molecular weights of 155 to 160 kDa (lane + ); before Endo-N treatment, cardiac tissue is polysialylated with hetcrw geneous molecular weight, but protein bands at 155 to 160 kDa and 180 kDa are detectable (lane - ). The forms ofcardiac NCAM before Endo-N treatment (b) and after Endo-N treatment (c) are shown with the exception of stage 34. Heart tissue from this particular stage was not Endo-N-treated to allow side-by-side comparison of polysialylated and desiaiylated forms of heart NCAM. The embryonic stages of tissue source are indicated across the top. In (b) highly polys~nlylatrd embryonic brain NCAM (br) is shown
detected with maturation in rat and human heart (Wharton et al., 1989; Gordon et al., 1990). In this study, we have noted a ventricular gradient of NCAM expression which may reflect the differences in maturation between the mural myocytes and those of the trabeculae. The elevated immunoreactivity of trabecular myocytes is consistent with the findings of Wharton et al. (1989) in the rat heart prior to 18 days of gestation. In vitro assays have shown that alterations of NCAM concentration can dramatically affect vesicle aggregation (Hoffman and Edelman, 1983) and neurite outgrowth (Doherty et al., 1990). Therefore even subtle differences in NCAM expression, which can be demonstrated at the histological level, may have a profound physiological influence on the behavior of the cardiomyocytes. The regulated pattern of NCAM expression
in skeletal muscle strongly supports a role for NCAM in neuron-to-myocyte interactions. Normal adult skeletal muscle expresses NCAM only around the neuromuscular junction (Covault and Sanes, 1985, 1986; Rieger et al. , 1985; Tosney et al., 1986). During establishment of innervation, during myopathy or after surgical or pharmacological denervation, NCAM is homogeneously expressed at the surface of the muscle fiber. A similarly close correlation of NCAM expression and innervation has not been made for cardiac cells. In the rat, most of the cardiomyocytes lose NCAM immunoreactivity postnatally and only express NCAM in the adult heart in discrete regions around the coronary arteries and the ventricular septum (Wharton et al., 1989). Those cells that are positive retain NCAM reactivity in both intercalated disc regions and on surfaces that do not necessaril)
1420
C. L. Burroughs
correspond to regions of interaction with nerve synapses. This is in contrast to the case of the mature myoneural junctions of skeletal muscle (Covault and Sanes, 1986). We find that chick cardiomyocytes retain NCAM immunoreactivity throughout embryonic, fetal and adult stages with some decrease in the intensity of immunostaining with maturation. The evidence is strong that NCAM is necessary for interactions between differentiating cardiomyocytes and between a subpopulation of mature cardiomyocytes before, during and after cardiac innervation. This trend has been observed in other systems (Thiery et al., 1982; Crossin et al., 1985; Tosney et al., 1986; Rutishauser, 1986; Watanabe et al., 1989). More recently, Gordon et al. (1990) have made correlations that indirectly support NCAM’s role in innervation. The sinus node and atrioventricular node, those areas that receive dense innervation, are more highly NCAMpositive at 15 weeks of gestation in the human. Furthermore, human transplanted hearts, which are denervated and never reinnervated express higher levels of NCAM than the normal adult heart. Evidence exists from in vitro studies that the interaction between brain membrane vesicle and cardiomyocyte are capable of being attenuated by anti-NCAM (Hoffman et al., 1985). In summary, the correlation of NCAM expression with cardiomyocyte innervation is not as striking as for skeletal muscle. The difference in patterns of expression between skeletal and NCAM cardiac muscle may reflect differences in the morphology and physiology of the respective nerve-muscle junctions. The role of NCAM in cardiac innervation is unclear. An apparent discrepancy exists as to whether the endocardium is negative (Thiery et al., 1982; Wharton et al., 1989) or positive (Mjaatvedt and Markwald, 1989) for NCAM. This is possibly due to the variability in staining of the endocardium with developmental stage and anatomical region. We find low levels of NCAM on endocardial cell surfaces beginning at stage 15. More intensely positive cells are present at specific regions directly involved in the internal remodeling of the heart. These regions include the facing surfaces of the endocardial cushions, the atrial surface of the endocardial cushions, the free edge of the atria1 septum and the endocardial
et al. ridges of the outflow tract. The endocardium of each of these regions is imminently invohcd in fusion as part of the septation of the heart. The spatiotemporal expression of NCAM on the free edge of the atria1 septum correlates directly with numerous endocardial cell modifications which occur just prior to fusion with the endocardial cushions (Arrechedera et al., 1987). The suggestion is that NCAM may have a role in directing and establishing the adherence of two opposing endothelial surfaces. In the case of the cushions of the atrioventricular canal and outflow tract, NCAM may be related to the differentiation of mesenchymal cells from the endocardium. These mesenchymal cells separate from the endocardium to populate the cushion matrix (Bernanke and Markwald, 1982; Markwald et al., 1977). Our observations support those of Mjaatvedt and Markwald (1989) that a downregulation of NCAM expression exists as mesenchymal cells migrate away from the endocardial surface. We deduce from SDS-PAGE and immunoblot profiles that heart tissue from chicken embryo stages 19 to 39 expresses polysialylated NCAM. It appears as a diffuse streak, characteristic of heterogeneously glycosylated NCAM, which focuses to a broad band of apparent molecular weight 155 to 160 kDa when desialylated by treatment with Endo-N, an endoneuraminidase specific for polysialic acid of NCAM. Since there is evidence that the polysialic acid is a general attenuator of cell-cell interactions (Rutishauser et al., 1988), the implication is that embryonic cardiocyte cell-cell interactions may be regulated by not only NCAM, but also by its polysialic acid. Developmental changes in the size of NCAM polypeptides expressed were detected in Endo-N-treated samples. At later stages NCAM polypeptide bands became detectable at 140 and 180 kDa in addition to the 155 to 160 kDa band. Since we analyzed all heart tissue together, these NCAM proteins may reflect components contributed by nervous tissue as well as by the myocardial and endocardial cells. Sequence analysis has shown that some of the skeletal muscle specific forms of NCAM resemble the 120 kDa NCAM isolated from neural tissue (Dickson et al., 1987). These lack a transmembrane domain and are bound to the plasma membrane by a phosphatidyl-
NCAM
in Heart
inositol linkage (Moore et al., 1987). In addition, these skeletal muscle NCAMs have an extracellular stretch of 37 amino acids near the membrane which is not found in the predominant nervous system NCAMs. A cDNA clone from an embryonic chick heart library was found to have a 93 base-pair insert similar, but not identical, to the skeletal-specific domain identified in human skeletal myocytes (Prediger et al., 1988). Sequence analysis of an NCAM genomic clone allowed identification of four exons in this region. The significance of the multiple NCAM polypeptide forms is not known; however, it has been postulated that the lack of a transmembrane domain may affect the stability of NCAM within the membrane (Gennarini et al., 1984). The differences in size of cytoplasmic domains may have profound effects on interactions of NCAM with cytoplasmic components such as cytoskeletal elements. Differences in the extracellular domain adjacent to the membrane may have consequences for how the NCAM interacts laterally with other molecules within the plasma membrane or how flexible it is as a hinged molecule (Hall and Rutishauser, 1987; and Becker et al., 1989). The regulation stabilization of the special mechanical and electrotonic interactions between cardiocytes may require specialized NCAM molecules.
Development
1421
In summary, this paper presents evidence that NCAM is associated with cell-cell interactions in the embryonic and adult chicken heart. The most obvious possibility is that NCAM functions in myocyte-myocyte interactions since this molecule is present on cardiomyocytes from the stages of tubular heart to adult. A significant observation is that NCAM is present on the surfaces of endocardial cells in regions where differentiation to mesenchymal cells and contact during tissue fusion occur. This presents the intriguing possibility that NCAM may not only function in the modulation of cell-cell interactions, but also in tissue-tissue interactions. Any impairment in such a basic process as tissue fusion could have major ramifications for morphogenesis of the heart. Resulting heart defects could include septal defects, endocardial cushion defects or aberrant truncal septation. Acknowledgements The authors thank Dr Giora Ben-Shachar for allowing us to photograph his histology slides of the ventricle, and Dr Urs Rutishauser and laboratory members for generosity with prepared reagents and advice. This work was supported in part by grants from American Heart Association, Northeast Ohio Affiliate and NIH R29 HL38172 OIAI.
References ARRECHEDERA H, ALVAREZ M, STRAUSS M, AYESTA C (1987) Origin of mesenchymal tissue in the septum primum: a structural and ultrastructural study. J Mel Cell Cardiol 19: 641-651. BECKER JW, ERICKSON HP, HOFFMAN S, CUNNINGHAM BA (1989) Topology ofcell adhesion molecules. Proc Nat1 Arad Sci USA 86: 1088-1092. BERNANKE DH, MARKWALD RR (1982) Migratory behavior of cardiac cushion tissue cells in collagen lattice culture system. Dev Biol 91: 235-245. COVAULT J, SANES JR (1985) Neural cell adhesion molecule (N-CAM) accumulates in denervated and paralyzed skeletal muscle. Proc Nat1 Acad Sci USA 82: 4544-4548. COVAULT J, SANES JR (1986) Distribution of N-CAM in synaptic and extrasynaptic portions of developing and adult skeletal muscle. J Mol Cell Cardiol 102: 716-730. CROSSIN KL, CHUONC CM, EDELMAN GM (1985) Expression sequences of cell adhesion molecule. Proc Nat1 Acad SCI USA 82: 6942-6946. DICKSON G, GOWER HJ, BARTON CH, PRENTICE HM, ELSOM VL, MOORE SE, Cox RD. QUINN C, PUTT W, WALSH FS (1987) Human muscle neural cell adhesion molecule (N-CAM): identification of a muscle-specific sequence in the extracellular domain. Cell 50: 1119-1130. DOHERTY P, FRUNS M, SEATON P, DICKSON G, BARTON CH, SEARS TA, WALSH FS (1990) A threshold effect of the major isoforms of NCAM of neurite outgrowth. Nature 343: 464-466. EDELMAN GM (1986) Cell adhesion molecules in the regulation of animal form and tissue pattern. Annu Rev Cell Biol2: 81-116. FRELINGER AL, RUTISHAUSER U (1986) Topography of N-CAM structural determinants. II. Placement of monoclonal antibodv eoitooes. 1 Mol Cell Cardiol 103: 1729-1737. GENNA~IN; G,* D&OSTINI-BAZIN H, GORIDIS C (1984) Studies on the transmembrane disposition of the neural cell adhesion molecule N-CAM. The use of liposome-inserted radioiodinated N-CAM to study its transbilayer orientation. EurJ Biochem 142: 65-73.
1422
C. L. Burroughs
et
al.
L, WHARTON J, MOORE SE, WALSH FS, Moscoso JC. PENKETH R, WALLWORK J. TA~L.OR KM, YA(:OUH MH, POLAK JM (1990) Myocardial localization and isoforms of neural cell adhesion molecule (N-CAM) in the developing and transplanted human heart. J Clin Invest 86: 1293-1300. HALL AK, RUTISHAUSER U (1987) Visualization of neural cell adhesion molecule by electron microscopy. J Cell Biol 104: 1579-1586. HAMBURGER V, HAMILTON HL (1951) A series of normal stages in the development of the chick embryo. J Morph01 88: 49-92. HAWKES R, NIDAY E, GORDON J (1982) A dot-immunobinding assay for monoclonal and other antibodies. Anal Biochem 119: 142-147. HOFFMAN S, EDELMAN GM (1983) Kinetics of homophilic binding by embryonic and adult forms of the neural cell adhesion molecule. Proc Nat1 Acad Sci USA 80: 5762-5766. HOFFMAN S, GRUMET M, EDELMAN GM (1985) Cell adhesion molecules in the histogenesis of nerve and muscle. In: Cardiac Morphogmesis, edited by VJ Ferrans, G Rosenquist, C Weinstein. New York, Else&r Science, pp 36-43. JOHNSON DA, GAUTSCH JW, SPORTSMAN JR, ELDER JH (1984) Improved technique utilizing nonfat dry milk for analysis of proteins and nucleic acids transferred to nitrocellulose. Gene Anal Tech 1: 3-8. KEANE RW, MEHTA PP, ROSE B, HONIC LS, LOEWENSTEIN WR, RUTISHAUSER U (1988) Neural differentiation, NCAM-mediated adhesion and gap junctional communication in neuroectoderm. A study in vitro. J Cell Biol 106: 1307-1319. KNUDSEN KA, MCELWEE SA, MYERS L (1990) A role for the neural cell adhesion molecule, NCAM, in myoblast interaction during myogenesis. Dev Biol 138: 159-168. LAEMMLI UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680-685. MCCLEAN IW, NAKANE PK (1974) Periodate-lysine-paraformaldehyde fixative. A new fixative for immunoelectron microscopy. J Histochem Cytochem 22: 1077-1083. MANASEK F (1968) Histogenesis of the embryonic myocardium. Am J Cardiol 25(2): 149-168. MARKWALD RR, FITZHARRIS TP, MANASEK F (1977) Structural development of endocardial cushions. Am J Anat 148: 85-120. MJAATVEDT CH, MARKWALD RR (1989) Induction of an epithelial-mesenchymal transition by an in ~ivo adheron-like complex. Dev Biol 136: 118-128. MOORE SE, THOMPSON J, KIRKNESS V, DICKSON JE, WALSH FS (1987) Skeletal muscle neural cell adhesion molecule (N-CAM): changes in protein and mRNA species during myogenesis of muscle cell lines. J Cell Biol 105: 1377-1386. PREDICER EA, HOFFMAN S, EDELMAN GM, CUNNINGHAM BA (1988) Four exons encode a 93 base-pair insert in three neural cell adhesion molecule mRNAs specific for chicken heart and skeletal muscle. Proc Nat1 Acad Sci USA 85: 9616-9620. RIECER F, GRUMET M, EDELMAN GM (1985) N-CAM at the vertebrate neuromuscular junction. J Cell Biol 101: 285-293. RUTISHAUSER U (1984) Developmental biology of a neural cell adhesion molecule. Nature 310: 549-554. RUTISHAUSER U (1986) Differential cell adhesion through spatial and temporal variations of NCAM. Trends Neurosci g(8): 374-377. RUTISHAUSER U, ACHESON A, HALL AK, MANN DM, SUNSHINEJ (1988) The neural cell adhesion molecule (NCAM) as a regulator of cell-cell interactions. Science 240: 53-57. RUTISHAUSER U, HOFFMAN S, EDELMAN GM (1982) Binding properties of a cell adhesion molecule from neural tissue. Proc Nat1 Acad Sci USA 79: 685-689. RUTISHAUSER U, JESSEL TM (1988) Cell adhesion molecules in vertebrate neural development. Physiol Rev 68: 819-857. THIERY J-P, DUBAND J-L, RUTISHAUSER U, EDELMAN GM (1982) Cell adhesion molecules in early chicken embryogenesis. Proc Nat1 Acad Sci USA 79: 6737-6741, TOSNEY KW, WATANABE M, LANDMESSER L, RUTISHAUSER U (1986) The distribution of NCAM in the chick hindlimb during axon outgrowth and synaptogenesis. Dev Biol 114: 437-452. TOWBIN H, STAEHELIN T, GORDON J (1979) Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some application. Proc Nat1 Acad Sci USA 76: 4350-4354. VIMR ER, MCCOY RD, VOLLCER HF, WILKISON NC, TROY FA (1984) Use of prokaryotic-derived probes to identify poly(sialic acid) in neonatal neuronal membranes. Proc Nat1 Acad Sci USA 81: 1971-1975. WATANABE M, FRELINGER AL, RUTISHAUSER U (1986) Topography of N-CAM structural and functional determinants. I. Classification of monoclonal antibody epitopes. J Cell Biol 103: 1721-1727. WATANABE M, KOBAYASHI H, RUTISHAUSER U, KATAR M, ALCALA J, MAISEL H (1989) NCAM in the differentiation of embryonic lens tissue. Dev Biol 135: 414-423. WATANABE M, PALTER KB, MAHOWALD AP (1982) Monoclonal antibodies recognize different components of the Dmsophila embryos. J Cell Biol95(2): 148a. WHARTON J, GORDON L, WALSH FS, FLANICAN TP, MOORE SE, POLAK JM (1989) Neural cell adhesion molecule (NCAM) expression during cardiac development in the rat. Brain Res 483: 170-176. GORDON
