Regional Organization of Pre-pacemaker Cells in the Cardiac Primordia of the Early Chick Embryo

There is ample evidence that the cells of the cardiac primordia in the chick embryo have undergone substantial chemodifferentiation long before any histological or morphological sign of the heart appears. Rawles (1943) has localized the position of the potential heart mesoderm in the head-process stage embryo (Stage 5, Hamburger-Hamilton series, 1951) to a pair of lateral heartformingregions. It is at this time that Ebert et al. (1953, 1955) have shown that the same two areas react positively for antigenic combining groups of cardiac actin and myosin. Taking advantage of Spratt’s demonstration (1950) of the sensitivity to fluoride ion of embryonic heart, as compared with other embryonic tissues, Duffey & Ebert (1957) have also shown that the cytotoxic effect of this ion on the stage 5 embryo is limited almost entirely to the cells in the heart-forming regions.

These regions of precardiac mesoderm can be seen on time-lapse motion picture films as dark condensed areas silhouetted through the endoderm (DeHaan, 1961a, 1963a). The pre-heart mesoderm is composed primarily of discrete clusters of cells, which migrate with the folding foregut endoderm, using that layer as a substratum for their own independent movements. The migrations of these clusters have been traced on such films from their initial position in the stage 5 heart-forming region, into the primitive tubular heart at stage 10 (DeHaan, 19636). For the first few hours after these clusters condense out of the background mesenchyme as discrete structures at stage 5, they appear to migrate in a random fashion within the heart-forming regions. Gradually, during stages 6 and 7, the mesoderm becomes arranged into a crescentic pattern. The cells destined to form non-cardiac structures, such as extraembryonic vascular tissue or head mesenchyme, leave the heart-forming regions while each cluster remaining within these regions takes up a position which bears a definite and constant relationship with the part of the heart to which that cluster will contribute. The group of clusters in the anteriormost portion of the lateral heart-forming region, for example, migrates into the rostral portion of the heart rudiments which develop first, and ultimately forms conus and cono-ventricular tissue. Those clusters in the middle of each cardiac primordium form the belly of the ventricle of the tubular heart; while the most posterior clusters in the heart-forming regions enter the heart rudiments last, to form atrial and sino-atrial tissue. Thus, in addition to the presumptive heart cells being differentiated, as such, from other mesoderm at these early stages, the various parts of the heart are also represented as localized portions of the heart-forming regions.

This localization suggests that cells in each of the presumptive regions may show distinct differentiative capacities, since in the later embryonic and adult heart, cells in the conus are quite different, histologically and biochemically, from, say, those which form the posterior wall of the definitive atria. In their physiological properties, too, cells in the various parts of the heart exhibit distinct differences. As revealed with intra-cellular electrodes, the shape and characteristics of the action potentials recorded from cells in ventricular, atrial or sinoatrial tissues are distinctly different, in both adult (Hoffman, 1961) and embryonic heart (Meda & Ferroni, 1959). Moreover, it has long been known that these various regions differ in their intrinsic pulsation rate. Barry (1942) has shown that there is a continuous antero-posterior gradient of inherent rhythmicity in the chick heart, such that any fragment of myocardium beats more slowly than those posterior, and more rapidly than those anterior to it. That these differences in rate are a property of individual cells, has been demonstrated by Cavanaugh (1955), who found that isolated cells of the atria beat faster than those from the ventricle, after disaggregation with trypsin. Since the ‘spontaneous’ beat of heart tissue depends upon stimulation of myocardial cells by the specialized pacemaker cells of the conductive tissue (DeHaan, 1961b), the question arises whether, in the early heart-forming regions, pre-pacemaker cells are already localized with the rest of the pre-heart mesoderm, and in accordance with the rate gradient seen in the beating heart. Specifically, are there pre-pacemaker cells in the anterior portions of the heart-forming regions, destined to contribute to the distal branches of the Purkinje system in the cono-ventricular region, which exhibit a low level of rhythmicity ? In contrast, are there pre-pacemaker cells in the posterior portions of the heart primordia, already localized with the presumptive atrial and sino-atrial mesoderm, which are capable of developing high levels of inherent rhythmicity suitable for the sino-atrial node and atrial conduction tissue ? The experiments reported here were designed to answer this question by separating the heart-forming regions into anterior, middle and posterior portions, and allowing each to develop in isolation. Under these conditions, a clear-cut antero-posterior gradient, or differential localization, of pacemaker activity can be seen as early as stage five.

Chick embryos were explanted at stage 4–5 by the method of New (1955), as described previously (DeHaan, 1959,1963a). With this technique the embryo is cultivated, ventral side up, on its own vitelline membrane. Development progresses normally for as long as –3 days. For microsurgery, embryos were allowed to develop in culture to the desired stage (stage 4 to 9). They were then. chilled to 4°C, a maneuver which stops further development and reduces stickiness and the tendency of the tissues to curl after cutting (suggested to me by Dr Mary E. Rawles). Each embryo was flooded with ice-cold Howard-Ringer solution (Howard, 1953) and cut into fragments, as shown in Text-fig. 1. Fragments 1R and IL were meant to include that material which would form (or at later stages, Text-fig. 1c, had formed) conus and cono-ventricular tissue; 2R and 2L included presumptive ventricle, and possibly some atrio-ventricular tissue; while 3R and 3L contained the posterior clusters, destined to form atrial and sinus tissue. Stages 4, 6 and 8 (not shown in Text-fig. 1) were cut in the same fashion, except that at stage 4 one of the posterior pieces (3R or 3L) was cut so as to include Hensen’s node.

Each fragment was washed in sterile culture medium and transferred with a micropipette to a labeled culture-cup containing 0·2 ml. of the medium. Each cup was then filled with a layer of 0·5 ml. of light mineral oil, to prevent evaporation, and the preparations were incubated at 37°C in an atmosphere containing 40 per cent oxygen, 5 per cent carbon dioxide and 55 per cent nitrogen.

The culture medium was composed as follows : 20 per cent chick embryo extract (from 11 day-old embryos), 30 per cent chicken or horse serum, and 50 per cent Earle’s balanced salt solution. Before use, 50 micrograms per ml. of Streptomycin and 100 units per ml. of penicillin were added. The medium was made up fresh each week and stored, frozen, at − 20°C.

At the end of 24 hours and 48 hours of incubation, each culture was examined under the dissecting microscope in a 37°C warm box, for spontaneous pulsatile activity, and the rate of beating was counted with the aid of a stopwatch. Fragments not showing spontaneous activity were stimulated mechanically or electrically (supra-threshold stimulus from a square-wave generator). Cases where brief stimulation initiated regular rhythmic contraction which continued for a substantial period, were counted with those exhibiting spontaneous activity, and the rate was determined. The remaining fragments, which showed no sign of beating even after electrical or mechanical stimulation, were counted as quiescent.

After examination at 48 hours, cultures were fixed overnight in 10 per cent neutral formalin, and then stored in 70 per cent ethanol. For further histological examination, each culture was stained whole in alcoholic cochineal, dehydrated through alcohols, cleared in pine oil, photographed, and then either mounted as a whole-mount, or embedded in paraffin for sectioning.

In all, 178 embryos (20–30 at each stage) were operated upon, yielding a total of 1068 cultured fragments.

After 8–10 hours of incubation in culture medium, explanted fragments tend to round up into solid, irregularly-shaped masses of tissue. At the end of 24 and 48 hours of incubation substantial increases in size are seen, and each culture takes on an appearance which is characteristic of the original position of the fragment and age of the donor. Plate 1 A–F shows three expiants after 48 hours of culture, as whole-mounts and in cross-section. The anterior fragments (1R and IL) taken from embryos younger than stage 8, usually retain a dense, rounded configuration. These cultures are largely opaque to transmitted light, making it difficult to see details of their internal structure when viewed intact (Plate 1A). They almost never show spontaneous contractile activity, and usually show no histological evidence of heart tissue (Plate 1B).

At the other extreme, fragments taken from the posterior part of the heartforming regions (3R and 3L) from embryos at stage 4-plus or older, usually form large, translucent, fluid-filled vesicles (Plate 1E). One part of the vesicle is usually somewhat denser than the other, and may be bright red with hemoglobin. This half encloses several compact masses of tissue, usually including a spontaneously-beating hollow ball or tube of heart tissue. The other half of these cultures is commonly a clear turgid vesicle, covered with a thin, tightly-stretched, almost transparent sheet of epithelial cells. This vesicle usually has fine strands of mesenchymal cells spanning its lumen, but is otherwise empty (Plate 1F).

Plate 1c and D illustrate a culture which is intermediate in character between the two already described. It is less dense and opaque than a typical anterior fragment, but not as large or translucent as the majority of the 3R–3L cultures. Such fragments frequently form histologically recognizable heart tissue, which may or may not exhibit spontaneous activity. As a general rule, if a mass of cardiac tissue is present in a vesicle, it will be surrounded by a fluid-filled translucent blister. Fragments 2R and 2L taken from embryos at stages 5 and 6, and fragments 1R and IL from later embryos (stages 8 and 9) usually develop as cultures of the intermediate type seen in Plate 1C.

Text-fig. 2 summarizes the heart-forming potencies of the clutured fragments, in terms of the fraction of the cultures which developed spontaneously-beating heart tissue, as a function of the stage of the embryonic donor. As noted above, the anterior fragments (1R, IL) destined to form conus and cono-ventricular tissue contain very few pre-heart cells before stage 8, and even at stages 8 and 9 only about half of the cultures are capable of forming contractile heart tissue. The middle fragments (2R, 2L) also exhibit very little pacemaker activity at early stages. However, by stage 6 more than 80 per cent of the cultures of these fragments form beating hearts, and by stage 7 all of them do. The posterior fragments (3R, 3L), containing presumptive atrial and sino-atrial cells, are the first to gain the capacity to form pacemaker tissue. At the first sign of notochordal cells pushing out in front of Hensen’s node (stage 4-plus), 80 per cent of these posterior fragments can develop beating heart masses; and this fraction very quickly increases to 90–100 per cent.

Results from the right and left fragments at each antero-posterior level were not statistically different, and were pooled to produce the curves in Text-fig. 2; except for the data from fragments 3R and 3L explanted from embryos at stage 8 and beyond. At these late stages fragment 3R decreases markedly in its ability to produce beating heart tissue, as compared with fragment 3L. This point will be discussed later.

Out of 66 fragments cultured from stage 4 embryos, none developed beating heart tissue. The inclusion of Hensen’s node in a fragment made no difference in this respect (see, however, Duffey, 1960).

As already shown in Text-fig. 2, the anterior fragments are incapable of forming beating heart tissue until stages 7–8. Moreover, those vesicles which can beat, as indicated in Text-fig. 3 (curve 1RL), do so at the low rate of 40–50 beats/min. The middle fragments (2R, 2L), from the beginning, contain cells with a higher intrinsic rhythmicity, producing vesicles which by stage 7 beat at a rate of 60–85 beats/min. The posterior fragments, in accordance with their content of presumptive atrial and sino-atrial cells, produce heart vesicles even from very early stages with high rates of spontaneous contraction, which, like those from more anterior fragments, gradually increase in rate with age of the donor, levelling off at stages 7 to 9 at 120–130 beats/min.

These results indicate that as early as stages 5–9, each portion of the embryonic heart-forming region has information coded into it which elicits a level of intrinsic rhythmicity of its pacemaker cells that is appropriate to the particular part of the heart which will develop from that portion.

Regional specificity within organ primordia is not an unusual finding in embryonic systems. The well-known local differences in amphibian medullary plate, leading to the differentiation of forebrain, midbrain, and hindbrain structures (see Holtfreter & Hamburger, 1955), or the presence in the embryonic limb disc and limb bud of areas of cells with specific fates as forearm or digits, radial or ulnar tissues (Zwilling, 1961) or again, the amazing mosaic of morphogenetic fields that makes up an imaginai disc of an insect (see, e.g., Hadorn, Anders & Ursprung, 1959; Ursprung, 1962), are but three of many classic examples of this phenomenon. What is perhaps unusual in the present work is the use of an internal physiological property, the intrinsic rhythmicity manifested in specific pacemaker cells, for the demonstration of such regionalization. For the present study, other criteria such as histological or morphological differences in the various parts of the heart would have been difficult or impossible to use as indicators of previous regionalization, since such differences, even in the intact embryo, do not appear until late in development. Conus cells are not histologically distinguishable from those of the atrium, for example, until 8-10 days of incubation; a period far greater than that during which normal histogenesis occurs in cultured fragments. However, within 24 hours after the primitive streak stage, pacemaker cells from each of the cardiac regions have differentiated sufficiently to begin pulsating at a characteristic rate which discloses their site of origin and prospective fate.

The well-defined antero-posterior rate gradient of cells in the heart-forming regions, demonstrated in the present work, accords well with earlier reports of a similar gradient of inherent rhythmicity in the formed heart after it begins beating (Paff, 1935; Barry, 1942). However, this finding should not be interpreted as indicating, necessarily, a distribution within the heart-forming regions at these early stages of pre-pacemaker cells, already determined to be appropriate for conus or ventricle, atrium or sinus. This is one possibility. However, an equally plausible hypothesis is that the organization of the mesoderm is not intrinsic, but derives from endoderm, or other cells in the environment. The heart-forming mesoderm in early stages, like the primordia of brain and limb cited as examples above, may represent an equipotential system in which all parts are competent to form any part of the adult organ. Localized differences in developmental potential would arise as the result of inductive or other epigenetic influences from the environment. The regional differentiation of a specific group of pre-cardiac mesoderm cells would be a function of the milieu provided by the particular region of endoderm and/or ectoderm with which these cells happened to come in contact. Anterior fragments (IR, IL) would produce slowly-beating vesicles, according to this idea, because the endoderm overlying the anterior heart-forming regions induces pre-cardiac cells in contact with it to become pacemaker tissues with low levels of inherent rhythmicity; while fragments 3R or 3L develop a sinuslike, rapid beat in similar fashion as a result of influences on the mesoderm by the surrounding tissues. The information in this case would be contained, not in the precardiac mesoderm itself, but in a set of reciprocal interactions between mesoderm and endoderm. Such inductive relations between pre-heart mesoderm and both endoderm and ectoderm are well documented for other species (Jacobson, 1961).

A critical test to decide between these two ideas is easily conceived : namely, the recombination of specific areas of endoderm with mesoderm from different portions of the heart-forming regions, or from neutral (i.e., non-precardiac) sites, from embryos during the period from stage 4 to stage 9. Such a test awaits the development of practical techniques for recombining tissues from these early stages.

The localization of a peak of heart-forming potency in the posterior part of the stage 5 pre-cardiac regions, illustrated in Text-fig. 2, conforms well with the classic experiments of Rawles (1936, 1943). Her expiants also demonstrated, more clearly than those in the present study, that fragments of the left cardiac primordium form pulsatile cardiac masses more frequently than those from the right, and show better histological differentiation. This was especially noticeable, in her work, when each primordium was divided into a medial and lateral half; in which case the lateral part of the left region showed much greater heart-forming potency than the symmetrical region on the right. In the present experiments, most of the data from the right and left sides were not significantly different. Lateral asymmetry did appear, however, in the posterior fragments at stages 7–9 (Text-fig. 2, curve 3RL). The capacity of the right fragments to form beating heart tissue decreased significantly, while that of the left side remained at 90–100 per cent.

Lateral asymmetry in intrinsic rhythmicity of the two cardiac primordia was reported in an earlier pubheation (DeHaan, 1959). If the two heart-forming regions were surgically prevented from fusing in the mid-fine, but otherwise left intact, the hearts that formed from each side beat with independent rates. When these lateral hearts first began beating, at the 10–11 somite stage, the left heart consistently had a faster rate than the right. As development progressed, both hearts increased gradually in rate, to a peak at about 18 somites (stage 13⁻). Beyond this age, from nineteen to twenty-four somites, the left heart dropped off dramatically in pulsation rate, while the right one remained constant.

In the present experiments, the isolation of small fragments of the heartforming tissue, and the relatively short-term culture period, apparently prevented such lateral differences in rate from appearing. The pulsation rate of equivalent right and left fragments were not significantly different at any stage, as indicated by the small standard errors shown on the curves in Text-fig. 3, which were made from the pooled right and left data.

The ability to localize pre-pacemaker cells in specific regions of the early embryo, before they become functional, suggests the exciting possibility of studying the electrophysiology of such cells during their differentiation. Recording with intracellular electrodes during the transition of a cell from a state of quiescence to one of spontaneous rhythmic activity, might provide new insights into the mechanism of action of pacemaker tissues in general.

1. Les ébauches cardiaques latérales du jeune embryon de Poulet ont été divisées en fragments antérieur, moyen et postérieur. Chacun d’eux a été incubé en milieu de culture à 37°C pour 48 h. Au milieu et à la fin de la période d’incubation chaque culture a été examinée pour déceler la présence de tissu cardiaque battant de façon spontanée, et le rythme de pulsation intrinsèque a été déterminé.

2. Les fragments postérieurs, lesquels, dans l’embryon forment le tissu de l’oreillette et du sinus, acquièrent le plus tôt (stade 4⁺) la capacité de former des vésicules cardiaques contractiles et montrent au plus haut degré la potentialité cardio-formative. Les fragments moyens ne sont capables qu’au stade 5 de former des coeurs qui battent, et ne le font que dans un pourcentage moins élevé de cas. Normalement, les cellules de ce fragment forment le ventricule. Les fragments antérieurs, qui représentent le futur cône artériel et la base ventriculaire de celuici, ne deviennent aptes à former des coeurs qui battent qu’au stade 7, et ne le font alors qu’avec une fréquence très basse.

3. Les vésicules issues des trois catégories de fragments augmentent leur taux spontané de pulsation avec l’âge du donneur. Les fragments postérieurs produisent du tissu cardiaque qui bat de 90 à 135 fois à la minute, avec une moyenne de 115 battements. Les vésicules provenant des pièces moyennes se situent dans les 40 à 45 pulsations par minute, avec une moyenne de 65. Les fragments antérieurs arrivent à donner 10 à 40 battements par minute, avec une moyenne de 36.

4. On peut conclure de ces résultats que les diverses parties du coeur (sinus, oreillette, ventricule et cône artériel), sont représentées dans le jeune embryon (stades 5 à 9) par des parties spécifiques et localisées des territoires cardio-formateurs et que chaque partie comprend des cellules déjà aptes à régler le rythme (éléments pTQ-pacemaker) lesquelles font apparaître un niveau correspondant de contractilité rythmique intrinsèque. Le gradient antéro-postérieur de capacité contractile que l’on observe dans le coeur devenu fonctionnel est déjà défini dans les territoires cardio-formateurs de l’embryon au stade du prolongement céphalique.

5. Les expériences réalisées ne permettent pas de décider si la régionalisation précoce de l’activité contractile est due à la répartition dans les ébauches cardiaques d’éléments régulateurs du rythme (pre-pacemaker cells) dont le taux futur de pulsation est déjà déterminé ou si elle résulte d’un complexe d’interactions locales émergeant épigénétiquement entre le mésoderme cardiogène et les tissus environnants. Un moyen de résoudre cette question est suggéré et d’autres points de vue relatifs aux résultats obtenus sont examinés.

I thank Miss Linda Fuson for her cheerful technical assistance throughout all phases of the work reported here.