The vitamin A derivative retinoic acid has previously been shown to have teratogenic effects on heart development in mammalian embryos. The craniomedial migration of the precardiac mesoderm during the early stages of heart formation is thought to depend on a gradient of extracellular fibronectin associated with the underlying endoderm. Here, the effects of retinoic acid on migration of the precardiac mesoderm have been investigated in the early chick embryo. When applied to the whole embryo in culture, the retinoid inhibits the craniomedial migration of the precardiac mesoderm resulting in a heart tube that is stunted cranially, while normal or enlarged caudally. Similarly, a local application of retinoic acid to the heart-forming area disrupts the formation of the cardiogenic crescent and the subsequent development of a single mid-line heart tube. This effect is analogous to removing a segment of endoderm and mesoderm across the heart-forming area and results In various degrees of cardia bifida. At higher concentrations of retinoic acid and earlier developmental stages, two completely separate hearts are produced, while at lower concentrations and later stages there are partial bifurcations. The controls, in which the identical operation is carried out except that dimethyl sulphoxide (DMSO) is used instead of the retinoid, are almost all normal. We propose that one of the teratogenic effects of retinoic acid on the heart is to disrupt the interaction between precardiac cells and the extracellular matrix thus inhibiting their directed migration on the endodermal substratum.

In the chick embryo, precardiac mesoderm cells are initially present in the epiblast, after which they migrate through the primitive streak and into the lateral plate mesoderm during gastrulation (Rosenquist and DeHaan, 1966; Rosenquist, 1970). By stage 5 (Hamburger and Hamilton, 1951), there are two distinct bilateral heart-forming areas (Rawles, 1943). Soon after this, the precardiac mesoderm undergoes a directed craniomedial migration until the two sides meet in the mid-line, over the head fold (DeHaan, 1963c). The two branches of this cardiogenic crescent then condense and form bilateral heart tubes. With lateral body folding, these two heart tubes are brought closer together until they start to fuse in a craniocaudal direction, forming a single mid-line heart (Manasek, 1968; Stalsberg and DeHaan, 1969). These early stages of heart development are illustrated in Fig. 1.

Fig. 1.

Diagrams showing a ventral view of normal heart development in the chick embryo from Hamburger and Hamilton (1951) stages 5 (19-22 h) to 10 (33-38h). At stage 5 (A), the left and right heart-forming areas (Ihfa and rhfa) are present as large diffuse patches of precardiac mesoderm cells lying on either side of the notochord (n), Hensen’s node (hn) and primitive streak (ps). At stage 6 (B), the precardiac mesoderm begins its craniomedial migration toward the developing head fold (hf). By stage 7 (C), the precardiac cells meet in the mid-line over the head fold to form a single, more condensed cardiogenic crescent (cc), while the overlying endoderm invaginates to form the anterior intestinal portal (aip). At this point, the unfused neural folds (nf) are visible, as are the first pair of somites (s) and the segmental plates (sp). At stage 8 (D), the precardiac mesoderm condenses further and begins to differentiate into bilateral heart tubes (ht). At stage 9 (E), the tubes fuse cranially in the mid-line and the sinus venosus (sv) forms over the anterior intestinal portal. The neural folds have fused along much of their length, so that the neural tube (nt) and forebrain (fb) are visible. At stage 10 (F), the bilateral heart tubes continue to fuse in a craniocaudal direction, forming a single mid-line heart that begins to bend to the right, giving rise to the primitive ventricle (v).

Fig. 1.

Diagrams showing a ventral view of normal heart development in the chick embryo from Hamburger and Hamilton (1951) stages 5 (19-22 h) to 10 (33-38h). At stage 5 (A), the left and right heart-forming areas (Ihfa and rhfa) are present as large diffuse patches of precardiac mesoderm cells lying on either side of the notochord (n), Hensen’s node (hn) and primitive streak (ps). At stage 6 (B), the precardiac mesoderm begins its craniomedial migration toward the developing head fold (hf). By stage 7 (C), the precardiac cells meet in the mid-line over the head fold to form a single, more condensed cardiogenic crescent (cc), while the overlying endoderm invaginates to form the anterior intestinal portal (aip). At this point, the unfused neural folds (nf) are visible, as are the first pair of somites (s) and the segmental plates (sp). At stage 8 (D), the precardiac mesoderm condenses further and begins to differentiate into bilateral heart tubes (ht). At stage 9 (E), the tubes fuse cranially in the mid-line and the sinus venosus (sv) forms over the anterior intestinal portal. The neural folds have fused along much of their length, so that the neural tube (nt) and forebrain (fb) are visible. At stage 10 (F), the bilateral heart tubes continue to fuse in a craniocaudal direction, forming a single mid-line heart that begins to bend to the right, giving rise to the primitive ventricle (v).

Several mechanisms for the directed migration of the precardiac mesoderm have been proposed (DeHaan, 1963a,b,c, 1964), but more recently, Linask and Lash (1986, 1988a,b) have provided evidence that this is related to a gradient of fibronectin in the extracellular matrix between the precardiac mesoderm and underlying endoderm. Using immunofluorescence and scanning electron microscopy, Linask and Lash (1986) found that, at stage 5, there is no fibronectin, or fibrillar extracellular matrix material at the mesoderm-endoderm interface in the heart-forming area. After stage 5, however, with the onset of precardiac cell migration, there is a concurrent increase in the amount of both fibronectin and fibrillar material in this region. They observed that the cells move on the fibrillar mesh and that, during stages 6 to 8, there is a gradient in both the amount of this material and in the concentration of fibronectin, increasing cranially. They suggested that this fibronectin gradient directs the migration of the precardiac mesoderm via an haptotactic mechanism involving differential adhesiveness.

Linask and Lash (1988a) provided further evidence for the involvement of fibronectin by exposing chick embryos of stages 5 to 8 to an antibody to fibronectin. They found that the anti-fibronectin disrupted precardiac cell migration, resulting in hearts that were stunted cranially and others that showed various degrees of bifurcation, indicating a failure in the formation of the cardiogenic crescent. They suggested that precardiac cell migration is prevented by the anti-fibronectin blocking the interaction between the cells and the extracellular fibronectin.

Since the production of craniofacial defects in mouse embryos (Kalter and Warkany, 1961), the teratogenic effects of vitamin A and particularly retinoic acid have been well documented (review. Spom et al. 1984). Shenefelt (1972) has shown, in the hamster, that almost every organ, or tissue system can be affected by retinoic acid if the embryo is treated at the ‘critical period’ of development. There are numerous examples of retinoic acid inhibiting the migration of mesenchymal cells, particularly neural crest cells, both in vitro (Morriss, 1975; Thorogood et al. 1982; Smith-Thomas et al. 1987) and in vivo (Poswillo, 1975; Hassell et al. 1977). In several such studies, it has been suggested that this inhibited migration is related to an alteration in the interactions between the cells and components of the extracellular matrix. This may be one of the mechanisms by which retinoids have their teratogenic effects, such as in the production of craniofacial abnormalities.

The aim of this study was to determine more clearly how retinoic acid affects early heart development and whether this was related to an effect on cell migration. The experiments demonstrate that the retinoid has a very similar effect to that of physically disrupting the precardiac mesoderm and of disturbing the fibronectin gradient in the extracellular matrix, as in the experiments of Linask and Lash (1986, 1988a), to produce stunted and bifurcated hearts. This suggests that retinoic acid may act directly on precardiac cell-extracellular matrix interactions and, in this way, inhibit the migration of these cells.

Hens’ eggs (Ross Brown) were incubated at 37.5°C to obtain embryos of stages 3 to 7 (Hamburger and Hamilton, 1951). Embryos were dissected from the yolk in P and C saline (Pannett and Compton, 1924) and explanted to culture using the technique of New (1955) in which the embryo lies in the culture dish with its endodermal side uppermost. A few embryos had reached stages 7+ and 8− and these were also used. Three types of experiments were done: (1) exposing the whole embryo to medium containing retinoic acid; (2) the removal of a piece of precardiac mesoderm and (3) the implantation of an anion exchange bead, previously soaked in retinoic acid, into the precardiac mesoderm.

(1) Exposing whole embryos to retinoic acid in medium

All-irarw-retinoic acid (Sigma Lot 104F-1035), with a relative molecular mass of 300, was made up in stock solutions of 3 mg ml-1 with dimethyl sulphoxide (DMSO, Sigma) and kept as aliquots at −70°C. The retinoid was thawed immediately before use and diluted with DMSO if a lower concentration was required. Five concentrations of retinoic acid were used: 10−6M, 5×10−6M, 10−5M, 5×105M and 10−4M. A measured amount of the appropriate retinoid solution was added to fresh, sterile tissue culture medium (Wellcome 199 with 10 % fetal calf serum, Gibco and 1 % penicillin streptomycin, Gibco) and, for all concentrations, the proportion of the DMSO in the tissue culture medium was always kept at the minimum (1%) that still allowed the dissolution of the retinoic acid. This was in order to avoid the possibility of DMSO toxicity. There were two control solutions: the DMSO control was 1 % DMSO in medium, while the untreated control consisted of tissue culture medium alone.

Excess fluid was removed from the top of each New culture and 100/11 of one of the prepared solutions was carefully added so that the entire blastoderm surface was covered. Cultures were then incubated at 37.5°C for 24 h. All procedures were carried out in subdued light and microscope illumination was covered with green filters in order to prevent isomerisation of the retinoic acid.

(2) The removal of a piece of precardiac mesoderm

Using a dissecting microscope, a rectangular incision was made through the endoderm and mesoderm layers in the heart-forming area with a tungsten needle, taking care not to cut through the underlying ectoderm. The two layers, which adhere strongly to each other, were then gently peeled away from the ectoderm with the tungsten needle. The piece was always removed from the left heart-forming area (right-hand side, as viewed ventrally in New culture) and care was taken to ensure that the rectangle included both medial and lateral borders of the visible precardiac mesoderm. Two types of experiments were performed: those in which the piece of precardiac mesoderm was removed and then immediately replaced in its original orientation and others where the piece was removed and discarded. The controls were embryos set up in New culture, but not operated on. All specimens were incubated at 37.5°C for 24h.

(3) Retinoid implants in the precardiac mesoderm

AG1-X2 anion exchange beads in formate form (Bio-rad, California) were soaked in retinoic acid and then implanted into the left heart-forming area. These beads allow a controlled release of the retinoid into the tissue over several hours (Eichele et al. 1984). The all-trww-retinoic acid (Sigma Lot 104F-1035) was made up in stock solutions of 10mg ml-1, l.0mg ml-1, 0.1 mg ml-1 and 0.01 mg ml-1 by dissolving it in dimethyl sulphoxide (DMSO, Sigma) and kept as 100/d aliquots at −70 °C. The retinoid solution was thawed immediately before use and a 100/4 drop was placed on a piece of Parafilm in a plastic Falcon dish. Using a dissecting microscope and eyepiece graticule, beads of approximately 100 gm in diameter were picked with fine forceps and placed in the drop of retinoic acid. The forceps were wiped thoroughly with tissue before picking out another bead. The beads were soaked in the retinoid solution for 20 min at room temperature. Each bead was then quickly rinsed in two 100 μl drops of sterile tissue culture medium (Wellcome 199 with 10% fetal calf serum, Gibco and 1 % penicillin streptomycin, Gibco) on Parafilm. They were then transferred to a plastic Falcon dish containing 1 μl of sterile tissue culture medium, in which they were soaked for 20 min at 37.5 °C. A slit was made in the endoderm over the centre of the left-hand heart-forming area of the embryo with a tungsten needle. A bead was then removed from the tissue culture medium and pushed through this slit into the precardiac mesoderm.

There were two types of control: an untreated control and a DMSO control, which underwent the same operation as the retinoid-treated embryos except that the beads were soaked in DMSO alone before being transferred to tissue culture medium and implanted into the precardiac mesoderm. Following treatment, specimens were incubated for 24 h at 37.5°C.

Analysis of specimens

For all three types of experiments, embryos were photographed in whole-mount after the 24 h incubation period. They were then either fixed in buffered formal saline and processed to wax for 8 gm serial histological sections stained with haematoxylin and eosin, or fixed in glutaraldehyde and processed for scanning electron microscopy. The x2 test was used for estimating the statistical significance of the results: a probability of 5 % was used as the level of significance.

(1) Whole embryo exposure to retinoic acid

There was a wide range of heart malformations seen in embryos treated with retinoic acid. At one extreme, was the complete absence of any heart, but just an indication of precardiac mesoderm on either side of the head region, as seen in Fig. 2F. The most frequently occurring anomaly, however, was a heart tube that was stunted cranially, but enlarged caudally. Thus, the bulbus and ventricle could be anything from moderately stunted, as in Fig. 2C and D, to completely absent (Fig. 2E), while the atrium and/or sinus venosus were large and swollen. The scanning electron micrograph of Fig. 3B and coronal section of Fig. 2D show this anomaly more clearly. These hearts were usually associated with an oddly shaped and sized anterior intestinal portal and ‘winged’ cranial somites, both of which can be seen in Fig. 2E. The winged somites, so named because of the wing-like processes that extend laterally and cranially from the somites, appear to be folds of pericardium stretching from the somites to the heart region (see Fig. 2D).

Fig. 2.

Whole embryo exposure to dimethyl sulphoxide (DMSO) or retinoic acid (RA). (A) Normal stage 11 embryo, which had been treated with DMSO at stage 4. It has a well-developed heart (h), which has started to bend to its right (left in the photograph), a normally shaped anterior intestinal portal (aip), and regularly shaped and spaced somites (s). (B) Coronal section of A showing the well-formed heart, sinus venosus (sv), AIP and somites. The pericardial cavity (pc) surrounds the heart and sinus venosus, but is separated from the somites by the pericardium (arrows). (C) Stage 11 embryo, which had been exposed to 10−5 M RA at stage 7 and now has a heart that is stunted cranially, but with a large, expansive sinus venosus and somites which are slightly winged cranially (ws) and continuous with folds (arrows) extending from the heart region. (D) Coronal section of D showing the abnormal heart and sinus venosus and the wing-shaped cranial somites. The long folds seen in C appear to be pockets of pericardium (arrows) which are continuous with these somites. (E) Specimen which had been treated with 10−4M RA at stage 5 (should now be stage 12-13) with a very stunted head, no heart development except for a large sinus venosus and its associated wide, square AIP and prominently winged somites, which appear to be continuous with the sinus venosus. (F) An embryo which had been treated with 10−4 M RA at stage 4 and should now be stage 11 – 12, but has no heart, with only an indication of undeveloped precardiac mesoderm (pern) on either side of the very stunted head. Scale bar=500 μm.

Fig. 2.

Whole embryo exposure to dimethyl sulphoxide (DMSO) or retinoic acid (RA). (A) Normal stage 11 embryo, which had been treated with DMSO at stage 4. It has a well-developed heart (h), which has started to bend to its right (left in the photograph), a normally shaped anterior intestinal portal (aip), and regularly shaped and spaced somites (s). (B) Coronal section of A showing the well-formed heart, sinus venosus (sv), AIP and somites. The pericardial cavity (pc) surrounds the heart and sinus venosus, but is separated from the somites by the pericardium (arrows). (C) Stage 11 embryo, which had been exposed to 10−5 M RA at stage 7 and now has a heart that is stunted cranially, but with a large, expansive sinus venosus and somites which are slightly winged cranially (ws) and continuous with folds (arrows) extending from the heart region. (D) Coronal section of D showing the abnormal heart and sinus venosus and the wing-shaped cranial somites. The long folds seen in C appear to be pockets of pericardium (arrows) which are continuous with these somites. (E) Specimen which had been treated with 10−4M RA at stage 5 (should now be stage 12-13) with a very stunted head, no heart development except for a large sinus venosus and its associated wide, square AIP and prominently winged somites, which appear to be continuous with the sinus venosus. (F) An embryo which had been treated with 10−4 M RA at stage 4 and should now be stage 11 – 12, but has no heart, with only an indication of undeveloped precardiac mesoderm (pern) on either side of the very stunted head. Scale bar=500 μm.

Fig. 3.

Scanning electron micrographs of hearts from DMSO and retinoid-treated embryos. (A) Stage 13 embryo, which had been treated with DMSO at stage 8, displaying a normal heart tube that has lengthened and started to bend to its right. The chambers of the primitive heart are labelled: t, truncus; b, bulbus cordis; v, ventricle; a, atrium; sv, sinus venosus; aip, anterior intestinal portal. The opening in the sinus venosus is due to breakage during processing. (B) Heart of a stage 12 embryo, which had been treated at stage 7 with 104M RA. Caudally, the heart (atrium and sinus venosus) is well developed, but cranially is very stunted. The large heart loop seen in A is replaced by a small compartment (arrow). Scale bar=100 μm.

Fig. 3.

Scanning electron micrographs of hearts from DMSO and retinoid-treated embryos. (A) Stage 13 embryo, which had been treated with DMSO at stage 8, displaying a normal heart tube that has lengthened and started to bend to its right. The chambers of the primitive heart are labelled: t, truncus; b, bulbus cordis; v, ventricle; a, atrium; sv, sinus venosus; aip, anterior intestinal portal. The opening in the sinus venosus is due to breakage during processing. (B) Heart of a stage 12 embryo, which had been treated at stage 7 with 104M RA. Caudally, the heart (atrium and sinus venosus) is well developed, but cranially is very stunted. The large heart loop seen in A is replaced by a small compartment (arrow). Scale bar=100 μm.

The range of heart abnormalities induced by whole embryo exposure to retinoic acid was difficult to separate into distinct groups, but there was a strong indication that both the incidence and severity of malformations increased as the concentration of the retinoid increased and the developmental stage of the embryo at time of treatment decreased. For example, all of the specimens that showed no heart tube formation had been treated with the higher concentrations at stages 3 to 5, while those showing cranially deficient hearts had been treated with lower concentrations, at these stages, or higher concentrations at the later stages. Although they do not indicate severity, Table 1 and Fig. 4 show the relationship between stage, concentration and incidence of anomaly. The difference between each treatment group and its DMSO-treated counterpart was determined to be significant or not using the x2 test (Table 1).

Table 1.

Incidence of heart anomalies for each whole embryo treatment and developmental stage at time of application

Incidence of heart anomalies for each whole embryo treatment and developmental stage at time of application
Incidence of heart anomalies for each whole embryo treatment and developmental stage at time of application
Fig. 4.

Incidence of heart anomalies for each whole embryo treatment and developmental stage, at time of application. As the concentration of retinoic acid increases and developmental stage decreases, the number of embryos with various degrees of stunted hearts increases (data in Table 1).

Fig. 4.

Incidence of heart anomalies for each whole embryo treatment and developmental stage, at time of application. As the concentration of retinoic acid increases and developmental stage decreases, the number of embryos with various degrees of stunted hearts increases (data in Table 1).

(2) Precardiac mesoderm removal

In this series of experiments, a piece of precardiac mesoderm was removed and replaced in its original orientation in 12 embryos, while the piece was removed and not replaced in 28 embryos. There were 10 unoperated controls. Within minutes of the incision in the heart-forming area having been made, the surrounding tissue retracted so that the operated area was almost double in size. At the same time, the extirpated tissue contracted until it was less than half its original size (Fig. 5A). Subsequently, the operated area failed to heal over, in most cases, resulting in a permanent discontinuity in the precardiac mesoderm. Thus, most of the specimens that had a piece of mesoderm removed and immediately replaced did not fully heal and were affected in the same way as those that did not have the explant replaced. Of these 12, 2 developed normal hearts, 8 displayed a full bifurcation of the heart (Fig. 5B), 1 had a partial bifurcation and 1 displayed situs inversus -ie. a left-handed heart (bent to the embryo’s left instead of to the right). Three of the embryos with full cardia bifida had two smaller areas of cardiac tissue on the operated side, one of which was probably derived from the excised piece of the heart-forming area and the other from the precardiac mesoderm caudal to this.

Fig. 5.

Whole-mount photographs of embryos immediately after and 24 h after precardiac mesoderm removal. (A) A rectangle containing the endoderm and mesoderm has been removed from the heart-forming area of a stage 5 embryo and immediately replaced, leaving the underlying ectoderm intact. The operated area has become larger due to the edges of the incision retracting, while the piece cut away has retracted (arrow), hp, head process; pern, precardiac mesoderm; ps, primitive streak. (B) 24 h later, the same embryo has developed two hearts (arrows), of about equal size, which were beating independently. The anterior intestinal portal (arrowhead) has been displaced to the embryo’s right and is associated with the medial border of the right hand heart. (C) A rectangle (arrow) containing the endoderm and mesoderm has been removed from the heart-forming area of a stage 7 embryo. (D) 24 h later, the embryo has developed two hearts (arrows), a large one on its right (left in the photograph) and a smaller one on its left, which were beating independently. The AIP (arrowhead) has been displaced to the embryo’s right, associated with the medial border of the larger heart, and the foregut (fg) appears to be shifted to the right of the embryo. Scale bar=500 μm.

Fig. 5.

Whole-mount photographs of embryos immediately after and 24 h after precardiac mesoderm removal. (A) A rectangle containing the endoderm and mesoderm has been removed from the heart-forming area of a stage 5 embryo and immediately replaced, leaving the underlying ectoderm intact. The operated area has become larger due to the edges of the incision retracting, while the piece cut away has retracted (arrow), hp, head process; pern, precardiac mesoderm; ps, primitive streak. (B) 24 h later, the same embryo has developed two hearts (arrows), of about equal size, which were beating independently. The anterior intestinal portal (arrowhead) has been displaced to the embryo’s right and is associated with the medial border of the right hand heart. (C) A rectangle (arrow) containing the endoderm and mesoderm has been removed from the heart-forming area of a stage 7 embryo. (D) 24 h later, the embryo has developed two hearts (arrows), a large one on its right (left in the photograph) and a smaller one on its left, which were beating independently. The AIP (arrowhead) has been displaced to the embryo’s right, associated with the medial border of the larger heart, and the foregut (fg) appears to be shifted to the right of the embryo. Scale bar=500 μm.

Of the 28 embryos that had the precardiac mesoderm removed and not replaced (Fig. 5C), 1 developed a normal heart, 21 displayed full cardia bifida (Fig. 5D), 4 had a partial bifurcation and 2 displayed situs inversus. The 10 untreated controls were all normal. In most cases of cardia bifida, other tissues appeared to be unaffected. In a small number of cases (3/21), however, double hearts were accompanied by stunted heads and neural tube defects due to the accidental removal of some of the ectoderm underlying the operated area. In some embryos, the mesoderm strongly adhered to the ectoderm and its removal was impossible without disrupting the ectodermal layer. When even a very small area was depleted of all three layers, tension in the blastoderm caused the hole to enlarge. The two related structures which were consistently disrupted when cardia bifida occurred, were the anterior intestinal portal and the foregut. With complete separation of the two hearts, these structures were reduced to a small flap of endoderm which was barely visible, but in cases of partial bifurcation were more prominent and sometimes normal. There were a few embryos in which the anterior intestinal portal and foregut appeared to be associated with the right heart only. An example of this is shown in Fig. 5D in which the embryo has an almost normal-looking heart tube on its right and there appears to be an anterior intestinal portal-like structure associated with the medial border of this tube. In fine with this is a fold of tissue extending to the right of the head, which appears to be a pocket of endoderm connected to the portal, forming an ectopic foregut. Fig. 5B shows a similar anomaly arising from a precardiac mesoderm explant removed and replaced. The embryo’s right-hand heart has formed a sinus venosus, rotated about 90 degrees, with an associated anterior intestinal portal.

In all cases of full cardia bifida, the left and right hearts appeared to beat independently, with no apparent relationship between the two. Often the left side was beating faster than the right side, but the opposite situation was never seen. With partial bifurcation of the heart, the left side beat first, followed closely by the right side.

(3) Retinoid implants in the precardiac mesoderm

171 embryos were used in these experiments of which 150 viable specimens were obtained, including 29 untreated controls and 31 DMSO-treated controls. The embryos treated with retinoic acid displayed heart malformations very similar to those produced by the removal of the precardiac mesoderm, particularly various degrees of cardia bifida. The incidence and severity of these effects depended on the concentration of the retinoid used in the bead implants, while the stage at which the implant was made appeared to affect only the severity of the defect produced. The incidence of the various cardiac anomalies, for each treatment, is shown in Table 2 and Fig. 8.

Table 2.

Incidence of each class of heart morphology for each implant treatment with all stages grouped together

Incidence of each class of heart morphology for each implant treatment with all stages grouped together
Incidence of each class of heart morphology for each implant treatment with all stages grouped together

As in the specimens that had the precardiac mesoderm removed, the two hearts resulting from retinoid implants always beat independently, with the left one usually beating faster. Another similarity was that with full bifurcations, most embryos had only a small, barely visible pocket of endoderm in place of the anterior intestinal portal and foregut (see SEM in Fig. 7A), but in some, either the left or right heart had developed a sinus venosus-like structure with what appeared to be an anterior intestinal portal associated with it (Fig. 6B). One difference between the results of the two experiments was that the two hearts produced by retinoic acid were usually of about equal size, rather than the right one being larger.

Fig. 6.

Whole-mount photographs of embryos immediately after, or 24 h after a bead (b) soaked in l.0mg ml-1 RA has been implanted into the left heart-forming area. (A) Stage 7 embryo immediately after the bead has been implanted. (B) 24 h later, the same embryo has two hearts (arrows) which were beating independently. There is no midline anterior intestinal portal, or foregut, but both left and right hearts appear to have a rudimentary sinus venosus-like structure (arrowheads). There is an indication of winged somites (ws) connected to folds of tissue in the heart region. (C) Stage 12 embryo, which had been treated at stage 8. There are two separate hearts (arrows), which were beating independently, but they share a common sinus venosus and AIP (arrowhead). (D) Stage 12 embryo, which had been treated at stage 6, resulting in two independently beating hearts (arrows) that are slightly connected over the large, square AIP (arrowhead). Scale bar=500 μm.

Fig. 6.

Whole-mount photographs of embryos immediately after, or 24 h after a bead (b) soaked in l.0mg ml-1 RA has been implanted into the left heart-forming area. (A) Stage 7 embryo immediately after the bead has been implanted. (B) 24 h later, the same embryo has two hearts (arrows) which were beating independently. There is no midline anterior intestinal portal, or foregut, but both left and right hearts appear to have a rudimentary sinus venosus-like structure (arrowheads). There is an indication of winged somites (ws) connected to folds of tissue in the heart region. (C) Stage 12 embryo, which had been treated at stage 8. There are two separate hearts (arrows), which were beating independently, but they share a common sinus venosus and AIP (arrowhead). (D) Stage 12 embryo, which had been treated at stage 6, resulting in two independently beating hearts (arrows) that are slightly connected over the large, square AIP (arrowhead). Scale bar=500 μm.

Fig. 7.

Scanning electron micrographs of embryos 24 h after beads soaked in retinoic acid have been implanted into the heart-forming area. (A) Stage 14 embryo, which had been treated with 1.0 mg ml-1 RA at stage 7. There is a complete separation of the two hearts (arrow’s). The rudimentary anterior intestinal portal (arrowhead) was broken up when the endoderm was removed. The bead, which is hidden from view, is in the mid-line dorsal to the AIP. (B) Stage 13 embryo, which had been treated with 0.1 mg ml-1 RA at stage 7. There is a partial cardia bifida with the two hearts (arrows) connected over the AIP (arrowhead) and the bead (b) positioned between them. (C) Stage 13 embryo, which had been treated with 0.1mg ml1 RA at stage 8. There is a less severe division of the heart than in B and there is a common midline sinus venosus (sv) and AIP (arrowhead). The bead is in the midline. Scale bar=100 μ m.

Fig. 7.

Scanning electron micrographs of embryos 24 h after beads soaked in retinoic acid have been implanted into the heart-forming area. (A) Stage 14 embryo, which had been treated with 1.0 mg ml-1 RA at stage 7. There is a complete separation of the two hearts (arrow’s). The rudimentary anterior intestinal portal (arrowhead) was broken up when the endoderm was removed. The bead, which is hidden from view, is in the mid-line dorsal to the AIP. (B) Stage 13 embryo, which had been treated with 0.1 mg ml-1 RA at stage 7. There is a partial cardia bifida with the two hearts (arrows) connected over the AIP (arrowhead) and the bead (b) positioned between them. (C) Stage 13 embryo, which had been treated with 0.1mg ml1 RA at stage 8. There is a less severe division of the heart than in B and there is a common midline sinus venosus (sv) and AIP (arrowhead). The bead is in the midline. Scale bar=100 μ m.

Fig. 8.

Bar, chart showing the incidence of normal, retarded, partially bifurcated and fully bifurcated hearts for each treatment (data in Table 2). The incidence of full and partial cardia bifida increases with the concentration of retinoic acid used.

Fig. 8.

Bar, chart showing the incidence of normal, retarded, partially bifurcated and fully bifurcated hearts for each treatment (data in Table 2). The incidence of full and partial cardia bifida increases with the concentration of retinoic acid used.

With partial bifurcations, there were two separate hearts, which were joined in the mid-line at the sinus venosus and were associated with a normally sized and shaped anterior intestinal portal (Fig. 6C and 7B). The embryo in Fig. 6D shows a greater separation of the two hearts that lie caudal to the very large anterior intestinal portal, but the prominence of the portal suggests that there is a strip of cardiac tissue connecting the two hearts. In all such cases, the two hearts were beating independently. Less severe partial bifurcations were usually associated with lower retinoid concentrations and resulted in large, ‘semi-detached’ hearts which appeared to consist of two distinct areas of cardiac tissue joined in the mid-line (Fig. 7C). There was a variable beating pattern in these hearts as, in most, the left side beat first followed closely by the right side, but in others, the heart appeared to beat as one unit. Other malformations resulting from retinoid implants, were. single hearts which were stunted, or of an anomalous shape and hearts displaying situs inversus, which were slightly stunted and bent to the embryo’s left, rather than to the right. The final position of the bead, 24 h after it was implanted, was not always consistent, but this did not have an effect on the type, or degree of anomaly produced. In most cases, the bead remained associated with heart tissue, but it could also be found buried deep in the head at heart level, or cranial to this (Fig. 6B) and sometimes ended up outside the embryo, to the left of the head (Fig. 6C).

The anomalies described here do not appear to be a result of the bead implant itself, as 24 of the 31 DMSO-treated embryos were normal. Of the remaining 7, only 1 displayed complete cardia bifida, 2 had partial bifurcations, 2 displayed situs inversus and 2 had stunted hearts. The incidence of these abnormalities was not much higher than in the untreated controls. Of the 29 untreated embryos, 1 had a partial bifurcation, 1 displayed situs inversus and 2 had stunted hearts. In previous experiments, it has been observed that untreated embryos in New culture occasionally display such malformations.

Concentration effect

Of the four concentrations of retinoic acid used for soaking the beads before implanting them into the precardiac mesoderm, the highest, 10 mg ml-1, was found to be extremely toxic and so was not used in the analysis of a concentration effect. With the other three, there was a direct relationship between the concentration of retinoic acid used and both the incidence and severity of heart malformations produced (see Table 2 and Fig. 8). Thus, embryos treated with l.Omgml-1 had the highest incidence of full cardia bifida, with a lower incidence of partial bifurcations, even fewer with hearts classified as retarded (stunted, or situs inversus) and very few with norma! hearts. Those treated with 0.1 mg ml-1 had fewer full bifurcations, more partial bifurcations and more hearts classified as retarded and normal than the specimens treated with the higher concentration. Finally, the embryos treated with 0.01 mg ml-1 had a lower incidence of both full and partial cardia bifida and more hearts classified as retarded and normal, than the embryos treated with the other two concentrations.

The X2 test was used for estimating the statistical significance of these results with a probability of 5 %. For each of the three concentrations of retinoic acid, the incidence of partial cardia bifida was significantly greater than that resulting from treatment with DMSO. In the case of full cardia bifida, on the other hand, only the two highest concentrations were significantly different from the DMSO-treated group. Nevertheless, the total number of all three classes of anomalies was significantly greater for each of the three concentrations of retinoic acid than it was for the DMSO controls.

Stage effect

The 35 embryos that were treated with 1.0 mg ml-1 retinoic acid were used for the analysis of a stage effect. The relationship between developmental stage and type of anomaly can be seen in Table 3 and Fig. 9 which show the incidence of each class of malformation resulting from this treatment at each stage. For a given concentration, the stage of development at which the implant was made did not appear to greatly influence the incidence of heart malformations, but did affect the severity, or type of anomaly produced. Thus, embryos that were treated at stages 5 and 6 had a higher incidence of full cardia bifida than those treated at stage 7 or later. The older embryos displayed more partial than full bifurcations and a higher incidence of retarded hearts (stunted, or situs inversus). Again, using the x2 test, the group of embryos treated at stage 5 and that treated at stage 6 both showed a significantly greater incidence of full cardia bifida than the group treated at stage 7+, whilst the older group had a significantly higher number of specimens with partial cardia bifida than either of the other two. When all three classes of anomalies were pooled, none of the treatment groups were found to be significantly different from either of the other two.

Table 3.

Incidence of each class of heart morphology, resulting from treatment with l.Omgml1 retinoic acid, for each stage

Incidence of each class of heart morphology, resulting from treatment with l.Omgml1 retinoic acid, for each stage
Incidence of each class of heart morphology, resulting from treatment with l.Omgml1 retinoic acid, for each stage
Fig. 9.

Bar, chart showing the incidence of normal, retarded, partially and fully bifurcated hearts resulting from treatment with l.0mg ml-1 retinoic acid for each developmental stage (data in Table 3). There is a higher incidence of full cardia bifida in those treated at the earlier stages, while more of those treated at later stages display a partial bifurcation.

Fig. 9.

Bar, chart showing the incidence of normal, retarded, partially and fully bifurcated hearts resulting from treatment with l.0mg ml-1 retinoic acid for each developmental stage (data in Table 3). There is a higher incidence of full cardia bifida in those treated at the earlier stages, while more of those treated at later stages display a partial bifurcation.

The effects of retinoic acid applied locally to the precardiac mesoderm, via an anion exchange bead, were very similar to those resulting from the removal of an area of precardiac mesoderm. The consistency with which retinoic acid produced various degrees of cardia bifida suggests that the retinoid achieves by chemical means, what the disruption of the heart-forming region achieves mechanically. When a large piece of precardiac mesoderm is removed, there is a discontinuity in the cardiogenic crescent which cannot heal, so the divided portions form two smaller separate hearts. With the bead implants, retinoic acid appears to inhibit the craniomedial migration of the surrounding precardiac cells, resulting in a functional, rather than a physical discontinuity in the cardiogenic crescent.

There were both differences and similarities between the bifurcations produced by precardiac mesoderm extirpation and the bead implants. In most cases of complete cardia bifida, the retinoic acid produced hearts that were of about the same size, rather than the left one being smaller. This is because none of the precardiac cells are removed and there is approximately the same amount of tissue on both sides (see Fig. 10), which indicates that the retinoic acid, at these concentrations, is not killing precardiac cells. In previous experiments, we have stained retinoic acid-treated embryos with Nile Blue Sulphate, which specifically stains dying cells (Saunders et al. 1962; Mills and Bellairs, 1989) and cell death in the precardiac mesoderm was never seen (unpublished observations).

Fig. 10.

Diagrams illustrating the similarity between removing a piece of the heart-forming area and the proposed action of retinoic acid on the precardiac mesoderm. (A) A rectangle is cut through the left heart-forming area and the mesoderm of this area (a) is removed. The remaining precardiac cells caudal to this explant area are prevented from joining the rest of the cardiogenic crescent and thus form a small heart on the left side. The right heart-forming area forms a larger heart on the right with a possible contribution of precardiac cells from above the explant in the left heart-forming area. (B) A bead (b) soaked in retinoic acid is implanted into the left heart-forming area, affecting the precardiac mesoderm within a certain radius and thus setting up a zone (z) of non-migrating cells. These cells are prevented from contributing to the cardiogenic crescent and, together with the precardiac cells caudal to the zone, form a left heart that is about the same size as that produced by the right heart-forming area. The size of the non-migrating zone depends on the concentration of retinoic acid. A high concentration appears to inhibit the migration of the whole left heart-forming area, while a lower concentration may not affect the more cranial precardiac cells, allowing them to contribute to the rest of the cardiogenic crescent.

Fig. 10.

Diagrams illustrating the similarity between removing a piece of the heart-forming area and the proposed action of retinoic acid on the precardiac mesoderm. (A) A rectangle is cut through the left heart-forming area and the mesoderm of this area (a) is removed. The remaining precardiac cells caudal to this explant area are prevented from joining the rest of the cardiogenic crescent and thus form a small heart on the left side. The right heart-forming area forms a larger heart on the right with a possible contribution of precardiac cells from above the explant in the left heart-forming area. (B) A bead (b) soaked in retinoic acid is implanted into the left heart-forming area, affecting the precardiac mesoderm within a certain radius and thus setting up a zone (z) of non-migrating cells. These cells are prevented from contributing to the cardiogenic crescent and, together with the precardiac cells caudal to the zone, form a left heart that is about the same size as that produced by the right heart-forming area. The size of the non-migrating zone depends on the concentration of retinoic acid. A high concentration appears to inhibit the migration of the whole left heart-forming area, while a lower concentration may not affect the more cranial precardiac cells, allowing them to contribute to the rest of the cardiogenic crescent.

One similarity between the precardiac mesoderm removal and retinoid implant experiments, was the beating pattern. In all cases of full cardia bifida and most cases of partial cardia bifida, the bilateral hearts were beating independently. In some specimens, the left and right hearts beat at about the same rate, but in most, the left heart beat faster. These beating patterns agree with the findings of others (Kamino et al. 1981; Yada et al. 1985; Satin et al. 1988) that there is a ‘left-heart dominance’ with respect to beat rate. The heart beat originates in the sinuatrial region at the caudal end of the heart tube which is derived from the most caudal portion of each heart-forming area, but is stronger in the left (Van Mierop, 1966; Kamino et al. 1981).

Development of the anterior intestinal portal and foregut in retinoid-induced cardia bifida was similar to that in precardiac mesoderm-deprived embryos. In most instances of full bifurcation of the heart, the anterior intestinal portal and foregut was reduced to a small, barely visible pocket of endoderm. The portal and foregut are formed by an infolding of endoderm in the mid-line, just caudal to the head fold. The apex of the cardiogenic crescent normally lies over the anterior intestinal portal and, as the bilateral heart tubes fuse in a craniocaudal direction, the portal is drawn further caudally, thus lengthening the foregut. By preventing the formation of the cardiogenic crescent, retinoic acid causes the hearts to develop on either side of the anterior intestinal portal, so there is no fusion of the bilateral heart tubes in the mid-line and the portal is not drawn caudally. When incomplete bifurcations occur, there is sometimes enough cardiac tissue in the mid-line to allow partial fusion of the heart tubes, thus drawing the anterior intestinal portal caudalward. In cases of complete cardia bifida, one heart appeared to have an associated portal and, unlike those seen in the mesoderm-deprived embryos, this could be on either side, rather than only the right. In the mesoderm-deprived embryos, the portal and perhaps the foregut can be shifted only to the right because a large area of endoderm, from which these structures arise, is removed from the left hand side and the remaining endoderm is associated with the more prominent right hand heart. By contrast, in the retinoid-treated embryos, the endoderm is not disrupted, so it may be possible for the portal to be shifted to either side, in association with that one heart.

The fact that the DMSO-treated controls showed a much lower incidence of these heart malformations indicates that the physical presence of the bead implant itself can account for only a small proportion of the anomalies seen in the embryos treated with retinoic acid. Rather, we propose that retinoic acid produces cardia bifida by inducing a zone of non-migrating cells in the precardiac mesoderm surrounding the implanted bead (see Fig. 10). The bead sets up a radial concentration gradient of retinoic acid in the mesoderm, so that the amount of tissue affected depends on the concentration of the retinoid used. A high concentration will inhibit the migration of the entire right heart-forming area and, providing the two heartforming regions have not already joined in the mid-line, will result in two completely separate hearts. A lower concentration, however, may not affect the precardiac cells at the cranial end of the heart-forming area, allowing them to migrate and join up with the left-hand branch. This would result in a link between the two precardiac regions and, thus, a heart that is only partially bifurcated.

The degree of bifurcation also depends on the stage at which the embryo is treated. In the specimens exposed to retinoic acid at stage 5, there was a higher incidence of full than partial cardia bifida, while in those treated at stage 7, there was a lower incidence of both types, but fewer full than partial bifurcations. There could be two different, but related explanations for this stage effect. Firstly, by stage 7 the precardiac cells of the two heart-forming regions have migrated cranially, forming the cardiogenic crescent. Thus, even if further migration is subsequently inhibited by retinoic acid, there is likely to remain at least a slight connection between the two hearts. At stage 5, however, craniomedial migration of the precardiac mesoderm has not yet commenced, so the retinoic acid is more likely to produce two completely separate hearts. In addition to the differences in the degree of advancement of the precardiac cells between these stages, there may be differences in the strength of the forces that direct this migration and thus in their susceptibility to the effects of retinoic acid.

These same two ideas can be used to explain the results of the whole embryo exposure to retinoic acid. Although two separate hearts were not seen in these specimens, the degrees of cranially deficient hearts indicates a similar concentration and stage-dependent inhibition of precardiac cell migration. In addition to the effects on cell-substratum interactions, cell-cell interactions also appear to be affected. The ‘winged’ somites seen in many retinoid-treated specimens appear to be due to their incomplete compaction and epithelia-lisation and a failure of the lateral plate mesoderm (presumptive pericardium) to separate and migrate away from the somitic mesoderm. Again, the degree of this anomaly may depend on the strength of the forces that influence these cell-cell and cell-substratum associations.

Linask and Lash (1986) have suggested that the major force involved in the directed migration of the precardiac mesoderm is a gradient of fibronectin in the extracellular matrix between the endoderm and precardiac mesoderm. They found that a gradient of both extracellular matrix material and fibronectin appeared in the heart-forming region, which increased cranially and coincided with the migratory behaviour of the precardiac cells. In further experiments, when a piece of precardiac mesoderm was removed, rotated 180° and replaced at stage 5, before this gradient had been set up, the heart developed normally, but when the same procedure was carried out at stage 7, after the gradient was well established, cardia bifida resulted (Linask and Lash, 1988b). Similar effects were seen after exposing embryos of stages 5 to 8 to high concentrations of exogenous fibronectin, or antibodies to fibronectin (Linask and Lash, 1988a). In both cases, it was concluded that migration was prevented by blocking the interaction between the precardiac cells and extracellular fibronectin. Embryos exposed to antibodies to other components of the extracellular matrix, such as laminin and collagen types I and II developed normally, suggesting that it is specifically fibronectin which plays an important role in the directional migration of the precardiac mesoderm.

The similarities between the effects of retinoic acid and disrupting the fibronectin gradient, on heart formation, suggest that retinoic acid may interfere with the interaction between the precardiac cells and fibronectin in the extracellular matrix. Retinoids have been found to disrupt the migratory behaviour of other cell populations. Thorogood et al. (1982) cultured quail cranial neural crest in the presence of retinol and found that outgrowth from the explant, cell morphology, fibronectin distribution and actin microfilament organisation were all severely affected. Smith-Thomas et al. (1987) found that neural crest cells responded to retinoic acid in much the same way.

The mechanism(s) by which retinoids have this effect on cell-matrix interactions is not clear. With the discovery of retinoid binding proteins (Liau et al. 1981 ; Chytil and Ong, 1984) and retinoid nuclear receptors (Petkovich et al. 1987; Giguere et al. 1987), there is evidence that retinoids can effect changes in gene expression directly. This may be the route by which retinoids alter the synthesis of cell-surface glycoproteins in vitro, as had previously been shown for fibronectin (Carlin et al. 1983; Grover and Adamson, 1985; Maden, 1985), as well as laminin (Wang et al. 1985) and collagen (Oikarinen et al. 1985). However, retinoids can also act directly on the cell membrane and surface glycoproteins. Dingle and Lucy (1965) demonstrated that the lipid-soluble retinoids can enter the lipid phase of the membrane, causing changes in its viscosity and permeability, whilst both retinol (De Luca, 1977) and retinoic acid (Bernard et al. 1984) have been shown to influence the glycosylation of membrane glycoproteins and, in particular, fibronectin. Retinoic acid alters fibronectin levels in enucleated cells (Bolmer and Wolf, 1982) and in cells where protein synthesis has been inhibited (Zerlauth and Wolf, 1984), providing further evidence for direct retinoid action on a cell-surface target.

Thorogood et al. (1982) suggested that retinoids affect fibronectin distribution in neural crest cells by inserting into the plasma membrane, causing viscosity and fluidity changes, which lead to a loss of organisation of membrane-bound fibronectin receptors. We have shown that explants of precardiac mesoderm exposed to retinoic acid in vitro show a similar disruption of fibronectin receptors and subsequent inhibition of cell-extracellular matrix interactions and migration (Osmond et al. in preparation).

In view of our present findings, it seems likely that many of the retinoid-induced heart malformations seen in mammalian embryos (Taylor et al. 1981; Davis and Sadler, 1981) and humans (Lammer et al. 1985) may be related to a disruption of precardiac cell-matrix interactions and inhibited cell migration during formation of the cardiogenic crescent.

This work was supported by grants from the British Heart Foundation. We are most grateful to Mrs R. Cleevely for her skilled technical assistance and to Dr C. Tickle for the gift of the AG1-X2 beads.

Bernard
,
B. A.
,
De Luca
,
L. M.
,
Hassell
,
J. R.
,
Yamada
,
K. M.
and
Olden
,
K.
(
1984
).
Retinoic acid alters the proportion of high mannose to complex type oligosaccharides on fibronectin secreted by cultured chondrocytes
.
J. biol. Chem
.
259
,
5310
5315
.
Bolmer
,
S. D.
and
Wolf
,
G.
(
1982
).
Retinoids and phorbol esters alter release of fibronectin from enucleated cells
.
Proc. natn. Acad. Sci. U.S.A
.
79
,
6541
6545
.
Carlin
,
B. E.
,
Durkin
,
M. E.
,
Bender
,
B.
,
Jaffe
,
R.
and
Chung
,
A. E.
(
1983
).
Synthesis of laminin and entactin by F9 cells induced with retinoic and dibutyryl cyclic AMP
.
J. biol. Chem
.
258
,
7729
7737
.
Chytil
,
F.
and
Ong
,
D. E.
(
1984
).
Cellular retinoid-binding proteins
. In
The Retinoids
, vol.
2
(ed.
M. B.
Spom
,
A. B.
Roberts
and
D. S.
Goodman
), pp.
90
123
.
Orlando
:
Academic Press
.
Davis
,
L. A.
and
Sadler
,
T. W.
(
1981
).
Effects of vitamin A on endocardial cushion development in the mouse heart
.
Teratology
24
,
139
148
.
DeHaan
,
R. L.
(
1963a
).
Oriented cell movements in embryogenesis
. In
Biological Organization at the Cellular and Supracellular Level
(ed.
R. J. C.
Harris
), pp.
147
164
.
London
:
Academic Press.
DeHaan
,
R. L.
(
1963b
).
Migration patterns of the precardiac mesoderm in the early chick embryo
.
Expl Cell Res
.
29
,
544
560
.
DeHaan
,
R. L.
(
1963c
).
Organization of the cardiogenic plate in the early chick embryo
.
Acta Embryol. Morph, exp
.
6
,
26
38
.
DeHaan
,
R. L.
(
1964
).
Cell interactions and oriented movements during development
.
J. exp. Zool
.
157
,
127
138
.
De Luca
,
L.
(
1977
).
The direct involvement of vitamin A in glycosyl transfer reactions of mammalian membranes
. In
Vitamins and Hormones
, vol.
35
(ed.
P. L.
Munson
,
E.
Diczfalusy
,
J.
Glover
and
R. E.
Olson
), pp. 1–57.
New York
:
Academic Press
.
Dingle
,
J. T.
and
Lucy
,
J. A.
(
1965
).
Vitamin A, carotenoids and cell function
.
Biol. Rev. Cambridge Philos. Soc
.
40
,
422
461
.
Eichele
,
G.
,
Tickle
,
C.
and
Alberts
,
B. M.
(
1984
).
Microcontrolled release of biologically active compounds in chick embryos: Beads of 200 pm diameter for the local release of retinoids
.
Anal. Biochem
.
142
,
542
555
.
Giguere
,
V.
,
Ong
,
E. S.
,
Segui
,
P.
and
Evans
,
R. M.
(
1987
).
Identification of a receptor for the morphogen retinoic acid
.
Nature
330
,
624
629
.
Grover
,
A.
and
Adamson
,
E. D.
(
1985
).
Roles of extracellular matrix components in differentiating teratocarcinoma cells
.
J. biol. Chem
.
260
,
12252
12258
.
Hamburger
,
V.
and
Hamilton
,
H. L.
(
1951
).
A series of normal stages in the development of the chick embryo
.
J. Morph
.
88
,
49
92
.
Hassell
,
J. R.
,
Greenberg
,
J. H.
and
Johnston
,
M. C.
(
1977
).
Inhibition of cranial neural crest development by vitamin A in the cultured chick embyro
.
J. Embryol. exp. Morph
.
39
,
267
271
.
Kalter
,
H.
and
Warkany
,
J.
(
1961
).
Experimental production of congenital malformations in strains of inbred mice by maternal treatment with hypervitammosis A
.
Am. J. Path
.
38
,
1
21
.
Kamino
,
K.
,
Akihito
,
H.
and
Fum
,
S.
(
1981
).
Localization of pace-making activity in early embryonic heart monitored using voltage sensitive dye
.
Nature
290
,
595
597
.
Lammer
,
E. J.
,
Chen
,
D. T.
,
Hoar
,
R. M.
,
Agnish
,
N. D.
,
Benke
,
P. J.
,
Braun
,
J. T.
,
Curry
,
C. J.
,
Fernhoff
,
P. M.
,
Grdc
,
A. W.
JR
,
Lott
,
I. T.
,
Richard
,
J. M.
and
Sun
,
S. C.
(
1985
).
Retinoic acid embryopathy
.
New Eng. J. Med
.
313
,
837
841
.
Liau
,
G.
,
Ong
,
D. E.
and
Chytil
,
F.
(
1981
).
Interaction of the retinol-binding protein complex with isolated nuclei and nuclear components
.
J. Cell Biol
.
91
,
63
68
.
Linask
,
K. K.
and
Lash
,
J. W.
(
1986
).
Precardiac cell migration: Fibronectin localization at mesoderm-endoderm interface during directional movement
.
Devl Biol
.
114
,
87
101
.
Linask
,
K. K.
and
Lash
,
J. W.
(
1988a
).
A role for fibronectin in the migration of avian precardiac cells: I. Dose-dependent effects of fibronectin antibody
.
Devl Biol
.
129
,
315
324
.
Linask
,
K. K.
and
Lash
,
J. W.
(
1988b
).
A role for fibronectin in the migration of avian precardiac cells: II. Rotation of the heartforming region during different stages and its effects
.
Devl Biol
.
129
,
325
329
.
Maden
,
M.
(
1985
).
Retinoids and the control of pattern in regenerating limbs
.
Ciba Found. Symp
.
113
,
132
155
.
Manasek
,
F. J.
(
1968
).
Embryonic development of the heart. I. A fight and electron microscopic study of myocardial development in the early chick embryo
.
J. Morph
.
125
,
329
366
.
Mills
,
C. L.
and
Bellairs
,
R.
(
1989
).
Mitosis and cell death in the tail of the chick embryo
.
Anat. Embryol
.
180
,
301
308
.
Morriss
,
G. M.
(
1975
).
Abnormal cell migration as a possible factor in the genesis of vitamin A-induced craniofacial anomalies
. In
New Approaches to the Evaluation of Abnormal Embryonic Development
(ed.
D.
Neubert
and H. J. Merker), pp. 678–687. Stuttgart: Thieme
.
New
,
D. A. T.
(
1955
).
A new technique for the cultivation of the chick embryo in vitro
.
J. Embryol. exp. Morph
.
3
,
326
331
.
Oikarinen
,
H.
,
Oikarinen
,
A. I.
,
Tan
,
E. M.
,
Abergel
,
R. P.
,
Meeker
,
C. A.
,
Chu
,
M. L.
,
Prockop
,
D. J.
and
Uitto
,
J.
(
1985
).
Modulation of procollagen gene expression by retinoids. Inhibition of collagen production by retinoic acid accompanied by reduced type I procollagen messenger ribonucleic acid levels in human skin fibroblast cultures
.
J. Clin. Invest
.
75
,
1545
1553
.
Pannett
,
C. A.
and
Compton
,
A.
(
1924
).
The cultivation of tissues in saline embryonic juice
.
Lancet
1
,
381
384
.
Petkovich
,
M.
,
Brand
,
N. J.
,
Krust
,
A.
and
ChambóN
,
P.
(
1987
).
A human retinoic acid receptor which belongs to the family of nuclear receptors
.
Nature
330
,
444
450
.
Poswillo
,
D.
(
1975
).
The pathogenesis of the Treacher Collins syndrome (mandibulofacial dysostosis)
.
Br. J. Oral. Surg
.
13
,
1
26
.
Rawles
,
M. E.
(
1943
).
The heart-forming areas of the early chick blastoderm
.
Physiol. Zool
.
6
,
22
44
.
Rosenquist
,
G. C.
(
1970
).
Location and movements of cardiogenic cells in the chick embryo: The heart-forming portion of the primitive streak
.
Devl Biol
.
22
,
461
475
.
Rosenquist
,
G. C.
and
Dehaan
,
R. L.
(
1966
).
Migration of precardiac cells in the chick embryo: A radiographic study
.
Carnegie Inst. Washington Contrib. Embryol
.
38
,
111
121
.
Satin
,
J.
,
Fum
,
S.
and
Dehaan
,
R. L.
(
1988
).
Development of cardiac beat rate in early chick embyros is regulated by regional cues
.
Devl Biol
.
129
,
103
113
.
Saunders
,
J. W.
,
Gasseling
,
M. T.
and
Saunders
,
L. C.
(
1962
).
Cellular death in morphogenesis of the avian wing
.
Devl Biol
.
5
,
147
178
.
Shenefelt
,
R. E.
(
1972
).
Morphogenesis of malformation in hamsters caused by retinoic acid: Relation to dose and stage at treatment
.
Teratology
5
,
103
118
.
Smith-Thomas
,
L.
,
Lott
,
I.
and
Bronner-Fraser
,
M.
(
1987
).
Effects of isotretinoin on the behaviour of neural crest cells in vitro
.
Devl Biol
.
123
,
276
281
.
Sporn
,
M. B.
,
Roberts
,
A. B.
,
Goodman
,
D. S.
, eds (
1984
).
The Retinoids, vol. 1 and 2
.
Orlando: Academic Press
.
Stalsberg
,
H.
and
Dehaan
,
R. L.
(
1969
).
The precardiac areas and formation of the tubular heart in the chick embryo
.
Devl Biol
.
19
,
128
159
.
Taylor
,
I. M.
,
Wiley
,
M. J.
and
Agur
,
A.
(
1981
).
Retinoic acid-induced heart malformations in the hamster
.
Teratology
21
,
193
197
.
Thorogood
,
P.
,
Smith
,
L.
,
Nicol
,
A.
,
Mcginty
,
R.
and
Garrod
,
D.
(
1982
).
Effects of vitamin A on the behaviour of migratory neural crest cells in vitro
.
J. Cell Sci
.
57
,
331
350
.
Van Mierop
,
L. H. S.
(
1966
).
Location of pacemaker in chick embryo heart at the time of initiation of heartbeat
.
Am. J. Physiol
.
212
,
407
415
.
Wang
,
S. Y.
,
La Rosa
,
G. J.
and
Gudas
,
L. J.
(
1985
).
Molecular cloning of gene sequences transcriptionally regulated by retinoic acid and dibutyryl cyclic AMP in cultured mouse teratocarcinoma cells
.
Devl Biol
.
107
,
75
86
.
Yada
,
T.
,
Sakai
,
T.
,
Komuro
,
H.
,
Hirota
,
A.
and
Kamino
,
K.
(
1985
).
Development of electrical rhythmic activity in early embryonic cultured chick double heart monitored with a voltage sensitive dye
.
Devl Biol
.
110
,
455
466
.
Zerlauth
,
G.
and
Wolf
,
G.
(
1984
).
Kinetics of fibronectin release from fibroblasts in response to 12-0-tetradecanoylphorbol-13-acetate and retinoic acid
.
Carcinogenesis
5
,
863
868
.