The specialized network of the conduction tissues mediates coordinated propagation of electrical activity through the adult vertebrate heart. The heartbeat originates in the pacemaker that is situated in the sinus venosus, or the sinoatrial node in the case of higher vertebrates. Following activation of the atria, the activation wave is slowed down in the atrioventricular canal or node, then spreads rapidly into the left and right ventricles via the His-Purkinje system (HPS). As a consequence of this arrangement, activation of the working muscle of the ventricle, and its ensuing contraction, proceed both initially from the ventricular apex in higher (Chuck et al., 1997; Nygren et al., 2000; Rentschler et al., 2001) as well as lower (Arbel et al., 1977; Dillon and Morad, 1981; Sedmera et al., 2003) vertebrates. This apex-to-base sequence of excitation and contraction is thought to increase ventricular pumping efficiency and is used as a marker of HPS function.
The development of mature HPS function during embryogenesis follows significant phases of cardiac morphogenesis (Gourdie et al., 2003). Initially, cardiac impulse propagates in a slow and apparently isotropic fashion from the sinus venosus, at the most caudal portion of the tubular heart, toward the cranially located primitive outflow tract (Kamino, 1991). While the apparent speed of impulse propagation gradually increases (based on total activation time inferred from isochronal epicardial maps), the sequence of ventricular activation as the heart loops follows the flow of blood (Chuck et al., 1997; Rentschler et al., 2002). Eventually, the immature base-to-apex sequence of ventricular activation undergoes an apparent reversal, altering to a mature apex-to-base pattern (Rentschler et al., 2001; Reckova et al., 2003).
The chick embryo is a popular model system for studying morphology and physiology of the developing heart (Hu and Clark, 1989; Sedmera et al., 1997). The electrophysiological properties of chick embryonic heart have been studied using microelectrodes (Arguello et al., 1986; de Jong et al., 1992; Chuck et al., 1997) as well as optical mapping with voltage-sensitive dyes (Kamino, 1991; Reckova et al., 2003; Hall et al., 2004). The anatomy of avian conduction system is well characterized (Davies, 1929; Vassall-Adams, 1982; Cheng et al., 1999). One interesting but seldom studied feature is the anterior septal branch, which came sharply to focus (together with the rest of ventricular conduction system) in our birthdating studies (Thompson et al., 2000, 2003). It is not, however, unique to avian hearts, since “dead-end tracts” were described in the human heart (Kurosawa and Becker, 1985) in similar location, and an analogous fascicle was demonstrated by HNK-1 antibody staining in the embryonic rat and confirmed by our optical mapping studies (data not shown). Furthermore, an anterior conduction pathway was demonstrated electrophysiologically in some univentricular human hearts along the anterior free wall of the ventricle (Krongrad and Malm, 1979). Recently, potential functionality of this embryonic conduction pathway was described in cultured mouse embryos after neuregulin treatment (Rentschler et al., 2002). This evidence led us to detailed examination of activation sequence of the anterior surface of the developing chick ventricle using optical mapping, and its correlation with conduction system markers on histological sections of relevant stages. We found that between Hamburger-Hamilton stages 16 and 25, the anterior ventricular surface is activated in a nonuniform pattern consistent with functionality of anterior conduction shortcut, independent of the posteriorly forming His bundle. These results were corroborated by birthdating and immunohistochemical studies, detailing the natural history and eventual disappearance of this temporary conduction pathway. Attention to this conduction system branch could advance our understanding of conduction system patterning and help explain some arrhythmias associated with its incomplete regression.
MATERIALS AND METHODS
This study conforms to the principles of the Declaration of Helsinki and the Guiding Principles for Research Involving Animals and Human Beings of the American Physiological Society.
Fertilized White Leghorn chicken eggs (ISE, Newberry, SC) were incubated blunt end up in a 38°C forced-draft incubator HH (Hamburger-Hamilton) stages (Hamburger and Hamilton, 1951) 12–29 (incubation day 2–6). These stages were selected based on our previous study (Reckova et al., 2003) to cover the period prior to HPS system maturity. Hearts (at least six per stage) were isolated under a dissecting microscope and stained with voltage-sensitive dye di-4-ANEPPS (0.002% solution in tyrodes-hepes buffer, pH 7.4) for 3–5 min at room temperature (20°C). Some hearts were dissected in frontal plane to obtain endocardial views of the interventricular septum and trabeculae and to verify independence of anterior and posterior conduction pathways. Anterior and posterior views were obtained from each heart analyzed. The hearts were placed in a custom-made silicone rubber-lined copper dish with oxygenated yyrodes-hepes solution on a temperature-controlled stage (Biostage 600, 35°C) of an upright epifluorescence microscope (Leica DML-FS) fitted with a 12 bit intensified high-speed digital CCD camera (80 × 80 pixels, Perkin Elmer/Olympus). To reduce undesirable motion artifacts, we used Cytochalasin D (Biermann et al., 1998) at 40 μmol/l during the recordings. Because of possible influence of this substance on calcium channels (Lader et al., 1999), alternative recordings were made with mechanical inhibition of heart movements (Raddatz, 1997). No significant differences in ventricular activation patterns were observed between these two approaches. Data were acquired at 500–1,000 frames/sec. For technical reasons, simultaneous acquisition of an electrocardiogram was not possible. Digital data were processed for construction of spatiotemporal activation maps and movies as described previously (Reckova et al., 2003; Sedmera et al., 2003).
The embryonic hearts were examined histologically using several techniques. Scanning electron microscopy (Sedmera et al., 1997) and immunohistochemistry (Sedmera et al., 2002) were used to obtain a general idea of the trabecular architecture. Thick acrylamide sections (Germroth et al., 1995) were used to follow trabecular structures of interest in three dimensions and to measure myofiber orientation (Thompson et al., 2003). Combination of double-labeling technique with [3H]-thymidine and 5-bromodeoxyuridine in both saturation and label dilution modes was used to outline areas of cell quiescence that correspond to nascent specialized conduction tissues (Reckova et al., 2003; Thompson et al., 2003). With the label dilution technique, times of differentiation of the cells forming the conduction fascicles is inferred from their retention of radiolabel administered at a certain time point, which makes them stand out against their neighbors, which took the label as well, but diluted it by a series of cellular division while in the proliferative and thus less differentiated stage (Cheng et al., 1999).
Between stages 12 and 16, the activation of the looping tubular heart followed direction of the blood flow in essentially isotropic fashion as previously described at stages prior to stage 11 (Kamino et al., 1981; Kamino, 1991). We observed slowing of action potential propagation in the regions of the differentiating atrioventricular canal and outflow tract, which became apparent from stage 15, in agreement with data from other studies (Arguello et al., 1986; de Jong et al., 1992).
The transitions in ventricular activation that mark the emergence of His-Purkinje system functionality, as well as dependence of maturation of this pathway on hemodynamic loading, were recently described using the optical mapping approach by our laboratory (Reckova et al., 2003; Hall et al., 2004) and essentially confirmed earlier data of other investigators obtained using electrode mapping (Chuck et al., 1997). Here we focus on the ventricular activation patterns between stages 16 and 27, well before the emergence of mature His-Purkinje function.
At stage 16, five of six examined hearts presented with activation sequence following the axis of the blood flow with apparent acceleration along the outer curvature, resulting in isochrones being perpendicular to the axis of the cardiac loop. In the remaining heart, activation proceeded in a narrow band down the center of the ventricle, then turned up toward the atrioventricular junction and the outflow tract. In most hearts, no ventricular trabeculae were present; in 20% of cases, a few ridges were apparent in the future ventricular apex (Sedmera et al., 1997). The outlier heart presented incipient ventricular trabeculae, but no clearly defined band corresponding to observed preferential conduction pathway was found histologically.
At stage 17/18, trabeculation in the ventricular apex was distinct, and a prominent ridge in the place of future interventricular septum identified previously by others (de la Cruz et al., 1997) was clearly visible, extending almost all the way to the atrioventricular canal anteriorly (Fig. 1). This correlated with a central preferential activation pathway on the anterior ventricular surface in 50% of examined hearts (Fig. 2). Interestingly, in a subset of hearts with an apparent anterior conduction pathway, a similar posterior pathway (as shown in Fig. 3 at a later stage) was present, resulting in a pattern slightly different from that described previously by us (Reckova et al., 2003).
The anterior conduction pathway continued to be present and was detected in all nine stage 21 hearts examined. As at previous stages, 50% of these hearts presented a similar posterior pathway (Fig. 3). Independence of these pathways was demonstrated by frontal dissection and imaging of each half separately. Similar to our previous observations (Thompson et al., 2003), the activation patterns were identical to those observed in the intact heart. Imaging of the cut endocardial surface confirmed that the activation was carried by the trabecular bands, in agreement with our previously reported data (de Jong et al., 1992; Reckova et al., 2003). A prominent anterior trabecular band was easily identified both histologically and by label dilution (Thompson et al., 2003)). This tight trabecular strand stained positively with the polysialic acid-neural cell adhesion molecule (PSA-NCAM) antibody, as did the rest of trabecular network. However, no similarly organized posterior counterpart was detected at this or other stages examined.
At stages 24/25, an anterior conduction pathway was present in 9 of 12 examined hearts (Fig. 2). Its posterior counterpart, however, was present in only three hearts. Histologically, trabecular bands forming the muscular interventricular septum stretched from the anterior atrioventricular junction all the way to the apex (Figs. 1 and 4). The anterior and posterior activation patterns were independent, as we demonstrated previously by independent imaging of anterior and posterior halves of frontally dissected hearts (Thompson et al., 2003). Similar to stage 21, mapping of endocardial surfaces clearly showed propagation of the action potential from the atrioventricular junction through the ventricular trabecular network (Fig. 3).
At stage 27, the ventricular activation patterns became simpler, and the activation proceeded on both the anterior and posterior surface from the atrioventricular junction toward the apex, then looped upward toward the outflow tract (Fig. 2). Neither anterior nor posterior preferential conduction pathway was detected (Fig. 5). The trabeculae were well developed and filled most of the ventricular cavity, while the compact myocardium remained thin (Sedmera et al., 2000). Immunohistochemical analysis with PSA-NCAM antibody showed stronger staining of a subset of trabeculae destined to become left and right bundle branches, in agreement with data previously reported by others (Chuck and Watanabe, 1997).
This same base-to-apex pattern was present at stage 29 (Fig. 2) and corresponded with our previous observations (Reckova et al., 2003). In a subset of hearts analyzed at that stage (2 of 21), we detected a second apical center of activation corresponding to a right bundle branch breakthrough site that becomes a single source of epicardial activation at later stages. This site corresponded to insertion of developing moderator band (Fig. 6), which carries the right bundle branch. Morphologically, the interventricular septum became more compact in appearance, and future His-Purkinje system components became clearly identifiable both by more intense PSA-NCAM immunostaining by others (Chuck and Watanabe, 1997) and by lack of proliferation by our laboratories (Chuck and Watanabe, 1997; Cheng et al., 1999; Reckova et al., 2003).
Despite differences in size, the morphological plan of the avian and mammalian heart is quite similar (Wessels and Sedmera, 2003), and the mature ventricular activation patterns also show remarkable homology (Durrer et al., 1961; Witkowski et al., 1997; Rentschler et al., 2001; Reckova et al., 2003). During development, these hearts go through a period of rapid increase in myocardial mass prior to establishment of coronary circulation, which necessitates the formation of a sponge-like trabecular meshwork to prevent myocardial hypoxia (Blausen et al., 1990; Sedmera et al., 2000). During that period, these trabeculae are the most differentiated part of the ventricular myocardium and are part of the primordial conduction system network (de Jong et al., 1992).
Numerous factors contribute to differentiation of trabecular myocytes toward the conduction phenotype. The first factor is simply their strand-like geometry, which favors electrical impulse propagation along the long axis (Rohr et al., 1997). Another influence is their mechanical conditioning stemming from the fact that inner layers of the tubular heart are subjected to higher strain and work harder (Thompson et al., 2003). Their close proximity to the endocardium exposes them to signaling molecules released by endothelial cells in response to stress, such as endothelin-1, which has been implicated in conduction system induction and patterning (Bond et al., 2003; Hall et al., 2004). Regardless of the stimulus, trabeculae in the chick heart predate mature His-Purkinje function by several days. It was suggested previously by de Jong et al. (1992) that trabeculae conduct faster and form the primitive conduction system of the embryonic heart. In the heart close to completion of septation (stage 31), they found that the endocardial surface (trabeculae) is activated earlier (with the first upstroke of the fractionated activation pattern seen on the epicardial surface) than the outer subepicardial myocardium. We have observed similar inside-out activation pattern at this stage using optical mapping of a partially microdissected heart (Reckova et al., 2003). Our present study shows directly that the trabeculae are indeed the primitive specialized ventricular conduction system even at earlier stages by endocardial mapping and by demonstration of an early and temporary anterior conduction pathway that functions prior to His-Purkinje system.
As the developing heart transforms from a tubular to looped morphology, the pattern and speed of ventricular activation also changes. The pattern of universally slow propagation along the primitive tubular heart develops two areas of fast conduction: atria and ventricles. While studying the atrial activation patterns is challenging in itself due to its thin wall and complex 3D morphology (data not shown), the ventricle is more amenable to study, including experimental dissection. It is worth noting that although the early chicken hearts are more physically and electrophysiologically delicate than later stages (past stage 17), they are still far more robust than comparable stages in mouse or rat (data not shown). In the ventricular loop, activation still follows the direction of the blood flow through the tubular heart. Coincident with the emergence of ventricular trabeculae and primordium of the interventricular septum (de la Cruz et al., 1997; Sedmera et al., 1997), preferential activation pathways that correspond with those myocardial bands are established anteriorly and posteriorly. These temporary conduction pathways were not discussed in our previous study (Reckova et al., 2003), which was focused on the effects of hemodynamic loading on transition to mature apex-to-base activation that occurs at later stages. As development proceeds, these pathways are masked by the appearance of more trabeculae, and finally superseded by early His-Purkinje system, which manifests by the emergence of apical breakthrough sites corresponding to right and left bundle branches (Reckova et al., 2003). It is worth noting that the appearance of apical epicardial breakthrough was recently reported by our colleagues (Chuck et al., 2004) 1 day earlier (stage 27) than in our experience; the most likely explanation is the difference in temperature setting (35°C in our conditions, to prevent hypoxia-related arrthyhmias at later stages, vs. more physiological 37°C by them). This breakthrough site corresponds to termination of forming right bundle branch identified by the same investigator at stage 27 by PSA-NCAM (Chuck and Watanabe, 1997).
The anterior activation pathway is not unique to the chick heart. We have been able to document similar patterns in embryonic day 11.5 rats (data not shown), and it was induced by neuregulin treatment in early embryonic mouse heart (Rentschler et al., 2002); however, no morphological substrate for this pathway was described by these authors. In the settings of two competing pathways, it is interesting to note that our attempts to image both surfaces simultaneously using angled mirrors or prisms showed that the anterior pathway at stages 17/18 is sometimes (but not always) activated prior to the posterior one (data not shown); however, the functional consequences of a few milliseconds of difference for ventricular contraction are not clear.
The temporary nature of the anterior conduction pathway was confirmed by its birthdating studies (Thompson et al., 2003). Although this is the earliest component of the ventricular conduction system to differentiate (the oldest cells making their last round of DNA synthesis prior to stage 10), follow-up studies showed that this pathway normally disappears during later fetal development or shortly after hatching. Its function becomes undetectable even earlier, probably because of concurrent differentiation of posteriorly located His bundle and its branches. We have seen disruption of this band in the settings of experimental ventricular septal defects (data not shown). Persistence of similar anterior conduction fascicles was observed in some malformed human hearts (Krongrad and Malm, 1979; Kurosawa and Becker, 1985), raising the possibility that it has, as a tight band of nondividing muscle, an important role in ventricular septation and proper alignment of the great arteries and ventricles. Future experiments will be directed at determining whether retention of this pathway in chick results in cardiac dysmorphogenesis and abnormal cardiac activation.
Our previous work has shown that timing of developmental transitions in ventricular activation sequence can be influenced by external factors changing hemodynamics (Reckova et al., 2003; Hall et al., 2004). In the settings of increased pressure load, the maturation of His-Purkinje function was accelerated, correlating with increased levels of conduction system markers connexin40 and PSA-NCAM. This effect is probably mediated by endothelin signaling, as suggested by accompanying upregulation of endothelin-converting enzyme. There is a period of sensitivity and time frame required for its effects, since no influence of conotruncal banding on early ventricular conduction (stages 24/25) was observed (data not shown). Similarly, maturation of bundle branches was delayed by decreased workload in experimental left heart hypoplasia as well as by inhibition of stretch-sensitive cation channels by gadolinium. In both cases, changes in activation sequence correlated with reduced levels of connexin40 and endothelin-converting enzyme (limited to the left ventricle in the case of left heart hypoplasia).
In conclusion, changes in ventricular activation sequence are as complex as its morphology. The persistence of primitive conduction pathways appears to correlate with heart dysmorphogenesis, warranting future investigation. The chick embryo is a suitable model for experimental testing of signaling cascades that direct induction and patterning of the cardiac conduction system.
Supported by National Institutes of Health grants HL50582 (to R.P.T.), HL67150 (to T.M.), HD39946 (to R.G.G.), RR16434 (to D.S.), and March of Dimes 5-FY02-269 Basil O'Connor Award (to D.S.).