Relative position of the atrioventricular canal determines the electrical activation of developing reptile ventricles

Heart muscle cells become electrically excited by action potentials, coupled to which is the calcium release that induces contraction (Burggren et al., 1998). In hearts of mammals and birds, the His–Purkinje system allows rapid conduction of the action potential that originates in the sinus node through specialized, large heart cells arranged in defined bundles that travel from the ventricular base (cranial) to the ventricular apex (caudal) (Davies and Francis, 1946; Dobrzynski et al., 2013). Accordingly, activation of most of the ventricular mass is from the apex towards the base (Sedmera, 2011). Molecularly, the His–Purkinje system is distinguished by the presence of the low-resistance gap junction protein Cx40, and its transcript Gja5, which allows fast conduction (Park and Fishman, 2017). Crocodylians appear to be the only ectotherms with a His bundle-like structure, probably because they are the only ectotherms with a full ventricular septum like mammals and birds (Jensen et al., 2018). The earliest activated part of the ventricular surface, the breakthrough of the activation, is towards the apex in the region of the ventricular septum. Crocodylians do not have a Purkinje system, as key markers, Gja5 and Nppb, are not confined to a small subset of ventricular muscle but are broadly expressed (Jensen et al., 2018).

Anole lizards have a typical squamate reptile heart (Jensen et al., 2014) and ventricular activation proceeds from the base, where the atria connect to the ventricular mass, to the apex (Jensen et al., 2012). No His–Purkinje system was found (Jensen et al., 2012), although this does not rule out a role for the 5–10 dorsoventrally and caudocranially oriented sheets of myocardium that are a common feature in the center of the ventricle of ectotherms (Sedmera et al., 2003). The absence of a His–Purkinje system begs the question whether, in the lizard ventricle, current may propagate at similar velocity from its point of initiation outwards in any direction. If true, the pattern of activation in lizards may be determined largely by the shape of the ventricle and where the atria connect to this shape. Here, we tested this hypothesis on developing lizards and snakes, as their ventricles develop from early left–right broad to caudocranially elongated in late stages. We found that the first activated region was always in the vicinity of the atrioventricular junction and the last point was on the far right in the early stages and at the apex in the later stages, thus always furthest away from the atrioventricular junction.

The handling of anole lizards (Anolis sagrei Duméril and Bibron 1837) and harvest of tissues complied with national and institutional guidelines and were approved by the Institutional Animal Care and Use Committee of the University of Amsterdam (DAE101617). For in situ hybridization, we used anole lizards of which one was adult and four were isolated from eggs and staged according to Sanger et al. (2008) (stages 7, 9, 17, 19). None of the sections shown have been published before, but other sections of the stage 19 specimen and the adult have been used in previous publications (Jensen et al., 2012, 2017). For optical mapping, we used 17 specimens of the central bearded dragon (Pogona vitticeps Ahl 1926) from embryonated eggs ranging in incubated age from 14 to 62 days post-oviposition and staged according to Sanger et al. (2008), and four specimens of the corn snake [Pantherophis guttatus (Linnaeus, 1766)] from embryonated eggs ranging in incubated age from 9 to 32 days post-oviposition and staged according to Boback et al. (2012). Under Czech Animal Protection law, handling of embryos in eggs is exempt from regulation; furthermore, only samplings followed by in vitro measurements of isolated hearts were performed.

We performed non-radioactive in situ hybridization to detect the gene transcript (Gja5) of the low­-resistance gap junction protein Cx40 as described previously (Jensen et al., 2012, 2017). The probe was based on the sequence of the following coordinates using UCSC Genome Browser on Lizard May 2010 (Broad AnoCar2.0/anoCar2) Assembly; Gja5 (chr3:162,064,640-162,065,729). This gene transcript marks the ventricular conduction system, or His–Purkinje system, in mammals and birds (Christoffels et al., 2010; Miquerol et al., 2011). We used one section per heart, cut in the horizontal plane, which included both the atrioventricular canal and the ventricular apex (stage 7, 9, 17, adult), or in the transverse plane of the ventricular base (stage 19). Images of stained hearts were imported into ImageJ version 1.51a, where we first used Enhance Contrast (saturated pixels: 0.4%) and then converted the images to 8-bit and inverted them. Using the Segmented Line tool, we drew a line with 5–10 points, from the ventricular left, over the apical region, to the ventricular right. The grayscale values under the line were retrieved by Plot Profile and then exported. Values below 50 were considered lumen or unstained tissue and were removed. Pearson's correlation was used to correlate staining intensity with anatomical position (in IBM SPSS Statistics version 24). A P-value below 0.05 was considered statistically significant. The correlations, irrespective of significance, were considered of limited biological interest if the regression coefficient (R²) was below 0.1.

Embryonated eggs of the central bearded dragon and the corn snake were incubated at 28°C and 75% humidity without rotation. At regular intervals, the embryos were carefully removed from the eggs, photographed under a dissecting microscope for staging, and rapidly decapitated. The hearts were then extracted from the embryos with the posterior thoracic wall attached and stained at room temperature for 15 min in di-4-ANEPPS (2.5 mmol l⁻¹, Invitrogen, Carlsbad, CA, USA). After brief rinsing, the samples were pinned at various orientations (anterior, posterior, left and right lateral view) on a silicone-covered bottom of a custom-made dish filled with oxygenated reptilian buffer (see Jensen et al., 2012; pH 7.5) with the addition of 40–60 µmol l⁻¹ cytochalasin D (Sedmera et al., 2003) to reduce motion artefacts. Mapping was performed at 0.25–1 kHz using an ULTIMA L setup (SciMedia, Tokyo, Japan) (Sankova et al., 2010). Heart rate, atrioventricular delay and epicardial activation maps were generated with the help of BV_Ana software bundled with the camera as described previously (Sankova et al., 2010).

In the stage 7 embryo, the ventricle was broader (left–right) than it was long (caudal–cranial) and the distance from the atrioventricular canal to the extreme right was longer than the distance to the apical region. The same morphology is seen in other squamates at similar stages of development (Jensen et al., 2013). In mammals and birds, Gja5, which encodes the low-resistance gap junction protein Cx40, is expressed in the atrioventricular bundle, its bundle branches and the Purkinje network of the specialized conduction system (Dobrzynski et al., 2013). In the stage 7 embryonic anole, Gja5 was expressed in the atria and ventricle, but not in the atrioventricular canal and the myocardial outflow tract (Fig. 1A). The level of expression was negatively related to the left–right axis (P<0.001, N=821 pixels), but the squared regression coefficient was miniscule (R²=0.02), showing that almost none of the variation in expression was explained by the left–right axis (Fig. 1A). The stage 9 embryo was analyzed as in Fig. 1A (P<0.001, N=572 pixels; data not shown), and the squared regression coefficient was also negligible (R²=0.02). By stage 17, close to hatching, the ventricle had elongated and developed a pointed apex (Fig. 1B). From the atrioventricular canal, the distance was now greater to the apex than to the ventricular left and right. There was expression of Gja5 in the atria and ventricle, but not in the atrioventricular canal and the myocardial outflow tract (Fig. 1B). The level of expression had a significant positive relationship to the left–right axis (P<0.001, N=1078 pixels), but almost none of the variation in Gja5 staining intensity was explained by the left–right axis (R²=0.04) (Fig. 1B). The stage 19 specimen was analyzed in the transverse plane (P<0.016, N=683 pixels; data not shown) and the level of Gja5 expression did not change from the ventricular left to the right (R²=0.01). In the adult, the shape of the ventricle and the position of the atrioventricular junction was much like that in stage 17 (Fig. 1C). The level of expression of Gja5 did not change from the ventricular left to the right (P=0.17, N=1769 pixels), but was positively related to the base to apex axis (P<0.001, N=1509 pixels), although the squared regression coefficient was again miniscule (R²=0.02) (Fig. 1C). In summary, we did not detect pathways of preferential conduction equivalent to the His bundle or bundle branches of mammals and birds, and propagation of the action potential may spread at similar velocity throughout the ventricular mass. The position of the atrioventricular canal in relation to the ventricular mass, then, may dictate the pattern of ventricular activation (Fig. 1D–F). Next, we tested this hypothesis with optical mapping.

The P. vitticeps specimens ranged from Sanger stage 6 to 19 (pre-hatching), which corresponded to an increment in crown–rump length from 18.2 to 41.1 mm. Size, developmental stage and gestational duration were not related to heart rate, atrioventricular delay or ventricular activation time (Table 1). This suggests that the mechanistic underpinnings of the measured electrophysiological parameters, like the density of sodium channels and gap junctional proteins, increased in proportion to growth. Similarly, in chickens, heart rate changes very little in the last two-thirds of development and most intervals of the ECG have adult-like values by 5 days of development (Bogue, 1932, 1933).

The hearts of the earliest stages of P. vitticeps and P. guttatus were broader than they were long and resembled the A. sagrei stage 7 embryonic heart. In both P. vitticeps and P. guttatus, ventricular activation on the ventral and dorsal surface initiated on the left at the ventricular base, where the atrioventricular canal connects to the ventricle (Fig. 2; 14 and 16 days post-oviposition, dpo). From there, the action potentials were propagated in a sweep towards the ventricular right, which was the part of the ventricle that was farthest from the atrioventricular canal. Later stages of P. guttatus developed a more elongate ventricle with a pointed caudal apex. The apex was farther from the atrioventricular canal than any other part of the ventricle and the ventral activation pattern was base-to-apex rather than the left-to-right sweep of the younger and broader ventricles (Fig. 2; 23 dpo). In P.vitticeps, a base-to-apex pattern of activation also developed in later stages on the ventral surface (Fig. 2; 35 and 62 dpo). On the dorsal surface, there were one or two initial breakthroughs, rather than a single one like on the ventral surface. These breakthroughs were a bit more caudal than on the ventral surface, but still within the region of the ventricular base and far from the apex. Overall, the dorsal pattern of activation was also base-to-apex and resembled the base-to-apex activation patterns reported for the dorsal surface of the adult A. sagrei ventricle (Jensen et al., 2012). The heart in the most developed stage of P. guttatus was very elongate and had a shoulder on the ventricular left, giving it a second apex on the cranial left. Such a cranial left shoulder was previously shown to develop in the corn snake (Jensen et al., 2013) and is a common feature of the formed hearts of snakes and lizards (Jensen et al., 2014). Activation still originated in the vicinity of the atrioventricular canal and spread simultaneously towards the caudal and cranial apex (Fig. 2; 32 dpo).

In all investigated hearts, the activation initiated in the ventricular base in the vicinity of the atrioventricular canal. The last activated part changed during development from the extreme right initially to the extreme caudal at late stages. This change in activation pattern coincided with a change by growth in shape from initially left–right broad to caudocranial elongate. If the shape of the ventricle largely dictates the pattern of activation, the speed of propagation would be expected to be similar across the trabecular ventricular myocardium. Cx40 allows for fast propagation, and we found that Gja5, which encodes Cx40, was essentially homogeneously expressed in the A. sagrei ventricle. We propose the ventricular activation pattern in lizards and snakes is largely determined by the shape of the ventricular mass and where the atrioventricular canal connects to the ventricular mass. Such a manner of activation was probably the basal condition for vertebrates. In contrast, in mammals and birds, the atrioventricular node connects directly to the insulated His–Purkinje network, which ramifies into the apical region predominantly and this part is therefore activated early (Davies and Francis, 1946; Moorman et al., 1998; Sedmera, 2011).

We would like to thank Martin Kotal for providing fertilized reptilian eggs.