Early Mesodermal Cues Assign Avian Cardiac Pacemaker Fate Potential in a Tertiary Heart Field

See allHide authors and affiliations

Science  10 May 2013:
Vol. 340, Issue 6133, pp. 744-748
DOI: 10.1126/science.1232877

Setting the Pace

The heart beats rhythmically throughout life. Highly specialized cardiac pacemaker cells control the timing of this beating. Bressan et al. (p. 744, published online 21 March) identified the embryonic location of the pacemaker precursors in early avian development and traced the cells throughout their incorporation into the heart. The events that establish the pacemaker lineage occur prior to the initiation of heart formation, and are governed, at least in part, by a class of Wnt signaling molecules.


Cardiac pacemaker cells autonomously generate electrical impulses that initiate and maintain the rhythmic contraction of the heart. Although the majority of heart cells are thought to originate from the primary and secondary heart fields, we found that chick pacemaker cells arise from a discrete region of mesoderm outside of these fields. Shortly after gastrulation, canonical Wnts promote the recruitment of mesodermal cells within this region into the pacemaker lineage. These findings suggest that cardiac pacemaker cells are physically segregated and molecularly programmed in a tertiary heart field prior to the onset of cardiac morphogenesis.

The rhythm of the heart is maintained by a specialized subclass of myocytes known as cardiac pacemaker cells (PCs). These cells generate action potentials (APs) in a cyclic manner to stimulate cardiac contractions. The anatomic position of mature PCs, the sinoatrial node (SAN), was described more than 100 years ago (1), however, little is known regarding the ontogeny or molecular mechanisms that specify PCs during development. This study was designed to address the timing, location, and mechanisms of PC cell fate acquisition.

Electrophysiological studies (24) have mapped cells that initiate cardiac APs to the inflow region at heart tube, looping, and septation stages. However, recent evidence indicates that as the heart matures, it continually expands, with cells being added to both the inflow and outflow segments [reviewed in (5)]. To determine which, if any, of the previously identified developmental pacing centers give rise to the mature SAN, we used optical mapping to image the AP initiation site and propagation pattern in embryonic chick hearts (4). Coincident with the heart’s first contractions at stage 10 [St10; Hamburger and Hamilton staging (6)], the AP initiation site was preferentially associated with the left posterior inflow segment of the heart [Fig. 1, A (red region) and F]. Left-sided pacing remained dominant through the process of dextral looping (Fig. 1, B and F). By late heart looping (St18), the AP initiation site shifted to the ventral surface of the right inflow, juxtaposed to and outside of the forming atria (Fig. 1C). From this stage on, all hearts displayed a right-sided AP initiation site (Fig. 1, D to F; see also movies S1 to S5).

Fig. 1 PCs begin pacing heart rhythm during late looping stages.

(A to E) Isochronal maps denoting AP initiation site (red) and propagation pattern from St10 to St35. (F) Pacemaker side spanning five progressive embryonic stages (white, left-sided pacemaker; black, right-sided pacemaker). (G and H) Labeling of St10 AP initiation site (arrow) developed for 36 hours to St18. (I and J) Labeling of St18 AP initiation site (arrow) developed 144 hours to St35. Insets show low-magnification images of labeled embryo (a, anterior; p, posterior; r, right; l, left; ht, heart tube; v, ventricle; at, atria; avj, atrioventricular junction; r-at, right atria; l-v, left ventricle; r-v, right ventricle; oft, outflow tract; svc, superior vena cava; ivc, inferior vena cava).

As the AP initiation site shifted from left to right, AP morphology changed substantially, displaying pronounced slow diastolic depolarization and shorter AP duration than earlier pacing centers (fig. S1). Right-sided pacemakers additionally exhibited a unique expression profile, becoming enriched for genes associated with mature PC AP generation, including Hcn4, Serca2, and Ryr2 (710), and coexpressing the atrial and ventricular muscle markers Amhc1 and Vmhc1 (fig. S2). To determine whether these differences were due to the maturation of migrating earlier pacing cells or were caused by the differentiation of a new cell population, we used vital lipophilic fluorescent dyes to trace the fates of early pacing cells. In no case did dye-labeled cells from left-sided AP initiation sites contribute to older pacing centers (Fig. 1, G and H, and fig. S3). In contrast, cells from the postlooping right inflow remained associated with the heart’s pacing region at all subsequent stages examined, eventually integrating into the SAN region at the back of the right atria (Fig. 1, I and J, and fig. S4). These data demonstrate that PC precursors emerge from a population of cells that are electrically inactive during the initial stages of heart development and begin pacing the heart within about an 8-hour developmental window that coincides with late dextral looping.

To identify when the assignment of PC fate occurs, we determined the location of PC precursors at stages prior to their electrical activation. Because none of the above markers displayed similar enrichment at earlier time points (fig. S2), we used nonmarker, direct cell labeling to create a series of geometric fate maps. Fate maps were scored according to labeled cell incorporation into the late looping stage PC region; scores were verified by whole-mount in situ hybridization for Hcn4 (figs. S5 and S6). At each of the stages examined, including gastrulation (St5), neurula (St8), and heart tube (St10) stages, PC precursors were mapped to a discrete region of the right lateral plate mesoderm (Fig. 2A). This region was posterior to the classical primary and secondary heart fields, as previously defined by cell tracing experiments and Nkx2.5 and Isl1 expression (Fig. 2B) (1117). Because of the limitations of vital dye labeling, we cannot rule out that at each of these stages other regions contribute to the PCs. However, labeled cells occupied the majority of the available area (fig. S5) when present at the ventral right inflow, so any outside contribution would have to be minor.

Fig. 2 PCs originate from mesoderm posterior to the heart fields.

(A) Fate maps depicting the location of PC progenitors at St5, St8, and St10 (St5 and St8, dorsal view; St10, ventral view). Red data points denote sites with >10% PCs; blue, all other tagged sites. Scale bar, 250 μm per division. Lower panel: Quantification of labels in the boxed regions (see fig. S5). (B) In situ hybridization (dorsal view) and schematic (ventral view) for heart field markers Nkx2.5 and Isl1 at St8. Asterisks denote PC region. (C) Best-fit contour plot of data from St8 fate mapping. (D) Overlays of surface plots indicating location of atrial, atrioventricular junction, proepicardial, and pacemaker precursors. A, anterior; P, posterior; R, right; L, left; PE, proepicardium.

PC precursors mapped to a region of the embryo that had not previously been identified as cardiogenic. Given the relatively large distance between PC precursors and the Nkx2.5 and Isl1 expression domains currently associated with cardiac mesoderm, we wanted to determine the distribution of cell fates within this region where Nkx2.5 and Isl1 were undetectable. A contour plot of the St8 labeling data revealed that the region with the highest probability of PC fate was 100 μm in diameter and centered 300 μm lateral to somite 3 (Fig. 2C). The surrounding Nkx2.5- and Isl1-negative mesoderm generated atria, atrioventricular junction, and the proepicardium (Fig. 2D and fig. S5); this finding indicates that large portions of the cardiogenic mesoderm do not express detectable levels of Nkx2.5 or Isl1 at St8. PC precursors did not substantially overlap with adjacent cardiac cell types, which suggests that heart precursors spatially segregate very early during lateral plate mesoderm formation.

Using the above fate mapping information, we sought to determine when PC fate became specified. PC precursors and primary and secondary heart field cells (Nkx2.5- and Isl1-positive domain) were isolated at the embryonic stages outlined above and allowed to differentiate ex vivo. The physiological and molecular identity of explants was then monitored. St18 explants spontaneously initiated APs with a periodicity of 256 ± 29 ms and displayed phase 4 (diastolic) depolarization (Fig. 3, A and B, fig. S7, and movie S7) distinct from that of stage-matched working myocardial explants. PC AP characteristics were not present in St5 explants (Fig. 3, A and B, fig. S7, and movie S7). Explants from St5 had large variations in interbeat intervals and slow beating rates, lacked phase 4 depolarization, and displayed an expression profile inconsistent with St18 PCs (fig. S8). PC precursors from both St8 and St10, however, did differentiate into PC-like cells after 72 hours of culture. Both stages displayed clear phase 4 depolarization, which was absent in age-matched primary and secondary heart field–derived cells (Fig. 3, A and B, fig. S7, and movie S7), and spontaneous depolarization was observed at intervals of 303 ± 53 ms and 294 ± 45 ms, respectively. Expression profiles of explants from St8 were consistent with in vivo PC as described above, showing enrichment for Hcn4, Serca2, and Ryr2, as well as coexpressing Amhc1 and Vmhc1 (fig. S8).

Fig. 3 PC fate is specified by St8.

(A) Membrane depolarization recorded from heart field. (B) PC precursors isolated from indicated stages. (C) Schematic indicating sites of mesodermal isolation from St8 embryos relative to Nkx2.5 and Isl1 expression. HF, heart field; A, atria; AVJ, atrioventricular junction. (D) Representative optical tracings of membrane potential from regions indicated in (C) after 72 hours of culture.

These data suggest that by St8, PC fate is already established in the Nkx2.5- and Isl1-negative lateral plate mesoderm. To determine the spatial restrictions of PC specification, we isolated mesoderm directly adjacent to the PC precursors from the presumptive atria, atrioventricular junction, and proepicardium (see Fig. 3C). Only the PC region displayed elevated phase 4 depolarization and a high rate of AP production (Fig. 3D and fig. S7). These findings suggest that PC fate is specified in a highly restricted subdomain of the right lateral plate mesoderm by St8, and that the initial events dictating the functional divergence of PC fate from the adjacent working myocardium must occur before this stage.

Additionally, our data indicate that a large region of mesoderm outside of the primary and secondary heart fields is already specified into working myocardial and PC fates by St8. To distinguish this mesodermal subdomain from the more classically defined heart fields, we refer to it as a tertiary heart field in chick. Although the precise boundaries of the primary and secondary heart fields remain controversial (18), our high-resolution fate mapping reveals the distribution and boundaries of several subtypes of cardiac precursors within the tertiary heart field. Conservation of this field in other model systems will require further validation.

Many studies have identified factors necessary for inducing myocyte specification in the primary and secondary heart fields. To determine factors that may play a role in inducing PC fate within the tertiary heart field, we examined the expression of several factors thought to positively or negatively influence myocyte specification during the developmental window outlined above. Expression of a canonical Wnt, Wnt8c, was detected in the region of prespecified PCs but not in the more anterior heart fields (fig. S9). Additionally, PC precursors, but not heart field cells, displayed nuclear accumulation of β-catenin suggestive of active Wnt signaling (fig. S9, I to L). We found this surprising because Wnts have previously been identified as inhibitory for heart field specification (19, 20).

To determine whether Wnt signaling is required to promote PC fate, we microinjected cells expressing the soluble Wnt antagonist Crescent (19, 20) adjacent to PC precursors before their specification. After 8 hours, PC precursors were explanted and allowed to differentiate ex vivo. Exposure to Crescent decreased the slope of PC phase 4 depolarization by 65% relative to control injections (Fig. 4, B and D). These experiments could not rule out the possibility that Crescent is interacting with factors not associated with canonical Wnt signaling. Therefore, to further demonstrate that Wnt signaling was capable of inducing PC fate, we injected Wnt-expressing cells into the presumptive heart fields. This resulted in a 69% increase in phase 4 slope (Fig. 4, C and D). We then used Bio, a pharmacological inhibitor of glycogen synthase kinase 3 (GSK3) that has been shown to stabilize β-catenin (21, 22) to activate Wnt signaling in the heart field. Consistent with the findings above, 10 μM Bio increased diastolic slope in heart field explants relative to control cells (fig. S10).

Fig. 4 Canonical Wnt signaling promotes PC fate.

(A) Schematic of stage 5 embryo indicating Crescent and Wnt8c expression domains. (B and C) After injection of Crescent- or Wnt-expressing cells, embryos were developed to St8, when PC or heart field precursors were isolated and placed in culture. Shown are optical membrane potential recordings from PC or heart field cultures after culture. (D) Quantification of diastolic slope from PCs (solid diamonds) or heart field cells (open circles). (E to H) Nkx2.5 expression is expanded into the PC region after injection of Crescent-expressing cells [dashed area of high-magnification insets from (E) and (F)] and is lost in the heart after Wnt cell injection [arrow in (H)] relative to injection of control cells. (I and J) Location of Wnt8C expression relative to ectopic PC-like AP generation.

When we allowed injected embryos to develop to late looping stages, aberrant Wnt signaling led to severe morphological defects, consistent with previous reports (Fig. 4, F and H) (23). Crescent injection adjacent to PC precursors led to the ectopic expression of Nkx2.5 in PC at St18, which is in agreement with a conversion of PC into a more working myocardial fate (24) (Fig. 4, E and F). About 35% of Wnt-injected embryos survived to heart looping stages. Wnt introduction into the primary and secondary heart field mesoderm resulted in irregularly contracting hearts, with decreased Nkx2.5 expression on the injected side of the embryo (Fig. 4, G and H). To confirm that these Nkx2.5-negative regions were still electrically active, we performed optical mapping. Consistent with a Wnt-based conversion of working myocardium into PC-like cells, we detected retrograde propagation (outflow toward inflow) as well as ectopic pacemaker sites (movie S8). These ectopic sites were restricted to the Wnt-injected side of the embryo and displayed AP shapes similar to those of control PCs (Fig. 4, I and J, and movie S8).

These findings suggest that early mesodermal Wnt-mediated cues are sufficient to induce pacemaker-like fates that do not manifest until late looping stages. However, Wnts are broadly and bilaterally expressed in the posterior mesoderm, so it is likely that additional cues are required to restrict PC fate, including laterality genes (25, 26). The early diversification of PC fate from the working myocardium suggests that fate specification is assigned directly in the lateral plate mesoderm, and is not the result of the specialization of an already functional embryonic myocyte. These data establish a framework through which PC development should be viewed, thereby providing a foundation for tissue engineering and stem cell–based approaches for PC generation.

Supplementary Materials

Materials and Methods

Figs. S1 to S10

Tables S1 and S2

Movies S1 to S8

References (27, 28)

References and Notes

  1. Acknowledgments: We thank T. Kornberg, D. Stainier, S. Coughlin, and R. Shaw for their comments, and Mikawa lab members for their suggestions. All data reported in this paper can be found in the main text or supplementary materials. Supported by NIH grants R01HL093566 and R01HL112268 (T.M.) and T32HL007544 (M.B.).
View Abstract

Navigate This Article