Catecholamines and development of cardiac pacemaking: An intrinsically intimate relationship

Abstract

A generation ago, a melding of imagination and experimental evidence led to the hypothesis that catecholamines were essential in establishing basal cardiac pacemaking rhythm. Subsequent discoveries of depolarizing “pacemaker” currents and viable adult catecholamine-deficient animals raised serious doubts about the necessity of catecholamines in pacemaking. However, the findings that catecholamines are produced in pacemaking regions prior to innervation, and that they are required for embryonic survival during a defined “critical period” of embryonic development have revitalized the original hypothesis. Recent results have further suggested that intrinsic cardiac adrenergic cells can differentiate into pacemaking myocytes, and that protein kinase A, a prominent downstream mediator of β-adrenergic signaling, is required for pacemaking activity. Here, we discuss how catecholamines and the intrinsic cardiac adrenergic cells that produce them may influence ontological development of cardiac pacemaking.

Epinephrine, Norepinephrine, Cardiac development, Pacemaker

Nearly 30 years ago, Gerald Pollack put forward a detailed mechanistic hypothesis explaining how catecholamines may be required for cardiac pacemaking activity [1]. This hypothesis proposed that catecholamines are made and secreted from pacemaking cells whereupon they act in autocrine and/or paracrine fashion to stimulate membrane receptors, bringing about transient elevations in intracellular cyclic adenosine monophosphate (cAMP) levels, which leads to triggered intracellular Ca²⁺ release and membrane depolarization, which, in turn, activates voltage-gated sarcolemmal Ca²⁺ channels. The cycle is completed when all of the vesicular catecholamine “fuel” is depleted, and the vesicles are then reloaded following re-uptake of catecholamines during repolarization of the cell.

One of the hallmark characteristics of pacemaker myocytes is slow diastolic depolarization. The catecholamine hypothesis did not address this key issue directly, and for this and other reasons, [2] was soon overshadowed by an electrophysiological model of pacemaking activity [3,4]. Although there is still considerable debate about whether membrane currents (I_f) or intracellular Ca²⁺ oscillations are the major driving force responsible for generating pacemaking activity, [5–8] it is now generally accepted that cardiac pacemaking activity is primarily driven by electrophysiological mechanisms involving coordinated activities between multiple membrane ion channel and intracellular Ca²⁺-handling proteins [9,10]. In all scenarios, catecholamines are thought to play a modulatory, but not integral, role in cardiac pacemaking. Consequently, the catecholamine-dependent hypothesis of cardiac pacemaking has been generally dismissed.

Accumulating new evidence now provides compelling reasons to revisit key aspects of Pollack's hypothesis, which had two main parts: (i) “…pacemaker cells store and secrete catecholamines”, and (ii) “they [pacemaker cells] are unable to pace when deprived of…catecholamines” [1]. This review will re-examine both parts of this hypothesis in light of evidence published since its inception, including (i) genetic studies in mice that revealed that catecholamines are in fact essential during early embryonic development, [11,12] (ii) identification of intrinsic cardiac adrenergic (ICA) cells in embryonic rodent and human hearts at early developmental stages when the heart first begins to beat, [13,14] (iii) transient and progressive localization of ICA cells in cardiac pacemaking and conduction system regions, [15] (iv) new data suggesting that ICA cells within the sinoatrial node (SAN) differentiate into pacemaking myocytes, [16] and (v) the recent demonstration that adult rabbit SAN myocytes appear to require protein kinase A (PKA), an important intracellular mediator of β-adrenergic receptor signaling effects, for pacemaking activity [17].

1. Are catecholamines required for cardiac pacemaking activity?

The bulk of the evidence cited by Pollack [1] to back an “obligatory” role for catecholamines in cardiac pacemaking stemmed from studies done with isolated SAN tissue and cultured cardiomyocytes in which beating rates were examined before and after treatment with reserpine, a drug that blocks storage and re-uptake of catecholamines. The studies typically showed that reserpine causes a dose-dependent decrease in beating rate, often to the point of cessation that can then be “recovered” by the addition of catecholamines and/or drugs that mimic their actions. These data suggested that the endogenous supplies of cardiac catecholamines were necessary for cardiac pacemaking activity.

It has been known since at least the late 1960s that the heart is able to take-up and store catecholamines, [18] but few studies specifically evaluated pacemaking cells in this regard. One that did was a study in dogs which showed that despite both chemical (6-hydroxydopamine) and surgical denervation of the heart, sinoatrial epinephrine (EPI) levels were not depleted by more than ∼50% [19]. In addition, catecholamine histofluorescence was observed in a “mosaic-like array” within the sinus node, [20] and electron micrographic analysis of nodal myocytes suggested that “numerous” synaptic-like vesicles are present in pacemaker cells [21]. The catecholaminergic cells in the heart have been characterized as small, roundish, granule-containing cells found in isolated groups or clusters near ganglia and often in close proximity to blood vessels [22,23]. These findings suggested that there are local, non-neural supplies of catecholamines in the heart, consistent with early studies in chicks, which showed that the embryonic heart was capable of producing catecholamines long before innervation occurred [24]. Based on these observations, Pollack argued that pacemaker cells likely store and secrete catecholamines.

Despite Pollack's arguments, the catecholamine hypothesis of cardiac pacemaking was soon displaced in favor of electrophysiological models that developed shortly after the discovery and characterization of the cardiac “pacemaker” current, I_f (sometimes referred to as “I_h”) [3,4,7]. I_f is activated during hyperpolarization of the cell, thereby producing an inward current that helps to trigger a slow depolarization. Recent evidence has also revealed potentially depolarizing intracellular Ca²⁺ oscillations from the sarcoplasmic reticulum (SR) Ca²⁺ stores [6,17]. In adult pacemaking myocytes, this diastolic depolarization activates low-voltage-activated channels such as Ca_V3.1 T-type and Ca_V1.3 L-type Ca²⁺ channels, [25–28] which in turn depolarize the cytoplasm sufficiently to open the primary, rapid Ca_V1.2 L-type Ca²⁺ channels and induce SR Ca²⁺ release, leading to full depolarization [29]. An inherent feature of this model is that I_f itself is sensitive to stimulation by catecholamines, [3] as are the L-type Ca²⁺ current, I_Ca,L, the type II ryanodine receptor, RyR2, and the SR re-uptake Ca²⁺-ATPase (SERCA) via its regulator, phospholamban (PLB), which leads to faster repolarization and greater Ca²⁺ cycling [17]. A mathematical model describing how catecholamines and acetylcholine neuroeffectors influence cardiac pacemaking activity has previously been described [30].

Subsequent work has shown that a family of subunit proteins known as hyperpolarization-activated nucleotide-gated (HCN) channels is responsible for mediating I_f. HCN4, one of the major isotypes expressed in cardiac pacemaking cells, appears to be critical for the development of pacemaking activity in mice since mice with HCN4 deficiency fail to develop ‘mature’ pacemaking action potentials, have significantly slower heart rates than their wild-type siblings, and die between embryonic day 9.5 (E9.5) and E11.5 [31]. This phenotype was apparent following both global and cardiomyocyte-directed deletion of the HCN4 gene, thereby demonstrating that cardiomyocyte-specific expression of HCN4 is essential for survival during embryogenesis. Interestingly, HCN4 is activated by both hyperpolarization and direct binding of cAMP to its carboxy terminus, [32,33] so it was perhaps not too surprising that HCN4-deficient embryonic hearts failed to demonstrate increased cardiac contractions in response to cAMP [34]. In contrast, RyR2^−/− embryos, while similarly bradycardic like HCN4^−/− embryos, still responded strongly to β-adrenergic stimulation [35]. Thus, HCN4 and hence, I_f, appear to be key mediators of β-adrenergic-induced increases in pacing frequency.

The possibility that catecholamines were required to help drive basal pacemaker activity was laid to rest in 1995, however, by gene knockout experiments that targeted tyrosine hydroxylase (Th) and dopamine β-hydroxylase (Dbh), resulting in catecholamine-deficient mice (Fig. 1) [11,12]. Most of the resulting homozygous mutant embryos died from apparent cardiac failure during prenatal development beginning at E10.5, with a majority of mutants dead within the ensuing next few days [11,12]. It may be important to note that the cause of death, the embryonic period of lethality, and the minimal morphological defects associated with catecholamine deficiency in these mice were remarkably similar to the phenotype observed for the HCN4 knockout mice [11,12,34]. The few catecholamine-deficient pups that survived to term usually perished within the first few days after birth [36]. Importantly, the catecholamine deficiency in these mice could be overcome by administration of synthetic precursors and/or adrenergic agonists to the maternal drinking water during the latter half of gestation [11]. This “rescue” strategy was effective such that Dbh^−/− pups were born, weaned, and matured into viable adult mice without the further need for drug treatment. Although there are significant differences between catecholamine-deficient and wild-type mice, [37] catecholamine-deficient hearts were clearly able to beat autonomously. These data are consistent with a recent report showing that rhythmic internal Ca²⁺ store oscillations and spontaneous beating of cardiac pacemaking cells continue unabated in isolated rabbit SAN myocytes when challenged with β-adrenergic receptor blockers [17]. Thus, catecholamines are not required for the generation of pacemaker activity per se. However, they are necessary for cardiovascular function and survival during prenatal development.

Fig. 1

Catecholamine biosynthetic pathway. The enzymes that catalyze each reaction are shown in italics above the arrows between precursors and products. TH, tyrosine hydroxylase; L-AAAD, l-aromatic amino acid decarboxylase; DBH, dopamine β-hydroxylase; PNMT, phenylethanolamine N-methyltransferase.

2. Which catecholamines are required for cardiac development?

The genetic knockout studies in mice have been informative in terms of identifying which of the endogenous catecholamines are important and when they are needed during cardiac development. To date, targeted disruption of the Th,[12],Dbh, [11] and Phenylethanolamine n-methyltransferase (Pnmt)[16] genes have been described. The developmental phenotypes of Th^−/− and Dbh^−/− mice are similar in that both result in embryonic/fetal cardiovascular failure. Thus, dopamine, which was elevated in Dbh^−/− embryos, [11] was unable to compensate for the loss of norepinephrine (NE) and EPI, thereby indicating that NE and/or EPI are the critical catecholamines necessary for cardiovascular development.

More recently, it has been shown that selective loss of EPI via targeted knockout of Pnmt[16] had no apparent effect on cardiac development or survivability, presumably because endogenous NE is still present in these animals. Thus, NE appears to be the key catecholamine molecule required for cardiac development. It is likely that there is some redundancy of function imparted by EPI, which can activate the same receptors as NE. ICA cells appear to produce both NE and EPI, and it is assumed that these two catecholamines can substitute for each other in many cases. There are some differences in binding affinity for certain receptors, especially β₂-adrenergic receptors where EPI is thought to be the primary physiological ligand. Differential activation of G_i and G_s pathways by β₁- vs. β₂-adrenergic receptors likely accounts, in part, for some of the differential actions of EPI and NE in the cardiovascular system [38].

There are multiple subtypes of adrenergic receptors, and many of them, including β₁, β₂, α_1b, α_1c, and α_1d have been detected in the heart [39]. The spatial and temporal distribution of these receptor subtypes in the developing heart is not known, though most of them are expressed at early embryonic stages of development [39]. Importantly, prenatal lethality associated with catecholamine deficiency can be rescued by administering the β-adrenergic receptor agonist, isoproterenol, via the maternal drinking water [36]. The α-adrenergic agonist, l-phenylephrine, had no apparent effect by itself, but did marginally enhance the ability of isoproterenol to rescue the Dbh^−/− embryos [36]. Thus, it appears that the embryonic lethality is primarily associated with signaling through β-adrenergic receptors, with α-adrenergic receptors playing a supportive role. It is somewhat perplexing that targeted disruption of mouse β-adrenergic genes either alone or in combination produced little or no lethality [40–42] whereas deletion of the β-adrenergic receptor kinase (βARK) resulted in cardiovascular failure and embryonic lethality with a phenotype remarkably similar to that described for the Dbh^−/− embryos [43]. It is possible that activation of α₁-adrenergic receptors may also compensate for the loss of β-adrenergic receptors. Taken together, these results suggest that catecholamines critically influence cardiac development primarily via β-adrenergic receptor signaling pathways.

3. A “critical period” of heart development for catecholamine sensitivity

The idea that endogenous catecholamines are important regulators of heart rate even at early, pre-innervation stages of development has been in existence at least since the early 1930's when the actions of adrenaline and acetylcholine were first observed to influence chick embryo hearts within the first day or so after contractions ‘spontaneously’ initiated [44]. Soon after making such an observation, for example, Hsu stated that, “Should it be possible, however, to detect the presence of acetylcholine and adrenaline in the embryo of the age we have worked with, our result would indicate a possible humoral control of heart rate before the inception of nervous control” [45].

In 1968, Ignarro and Shideman [24] showed that the major catecholamines, dopamine (DA), NE, and EPI, as well as the enzymes responsible for their production (Fig. 1) were present in the developing chick heart beginning on the 3rd day of incubation, shortly after the heart first starts to beat. Of particular note, chick embryos exposed to single doses of either EPI or isoproterenol at an early stage of development (day 4, Hamburger–Hamilton [HH] Stage 24 or day 5, HH Stage 26) were found to have a 20- to 25-fold higher incidence of serious cardiac anomalies than control embryos when assessed 9–10 days later (HH Stage 41) [46]. The types of defects induced by such exposure to β-adrenergic agonists included “aortic arch defects, ventricular septal defects, double outlet right ventricle, aortic hypoplasia, and (persistent) truncus arteriosus” [46]. These types of cardiac malformations are reflective of those typically found in humans born with congenital heart defects [47]. Administration of adrenergic drugs a few days later or earlier in development had little, if any, influence on cardiac development. Thus, it appears that there is a critical window of development in the chick where the heart is susceptible to catecholamine-induced cardiac malformations.

Intriguingly, this window of catecholamine hypersensitivity in the chick roughly corresponds to a period in mice when catecholamines are essential for development. For example, when catecholamine-deficient (Dbh^−/−) dams (obtained via maternal drug rescue experiments as discussed earlier) [11] were crossed with heterozygous (Dbh^+/−) males, none of the Dbh^−/− embryos survived past embryonic day 13.5 (E13.5) [11]. There was no apparent lethality of Dbh^−/− embryos at E9.5 and only slight (10%) lethality associated with Dbh^−/− embryos 1 day later (E10.5). However, there was a marked drop in survivability between E10.5 and E11.5, with more than 60% of Dbh^−/− embryos dying between E10.5 and E12.5, with the remainder dying between E12.5 and E13.5. These data demonstrate that the minimal “critical period” for catecholamine action during mouse embryonic development occurs between E9.5 and E13.5. It may extend further into later stages of fetal development, but additional experiments are required to ascertain this. Extrapolating to human development, this critical period would occur at approximately 4–6 weeks after fertilization (Fig. 2) [48].

Fig. 2

Critical period of catecholamine sensitivity in chick, mouse, and (hypothetically) human embryonic development. The beginning of the “Critical Period” occurs shortly after the appearance of the first myocardial contractions and the initiation of blood flow through the embryonic heart. It extends through the entire classical embryonic period of development and ends in the middle of the major phase of organogenesis (metamorphosis). Notably, the end of this critical period approximately corresponds with the completion of the major phases of cardiac conduction system development and the onset of functional innervation of the heart. Comparative staging was extrapolated from Sissman [48] and Kaufman [76].

The pharmacological studies in chick embryos suggest that excessive exposure to catecholamines during this critical period leads to congenital heart malformations, while the genetic knockout studies in mice suggest that catecholamine insufficiency results in heart failure and perinatal lethality. Thus, an imbalance that results in too much or too little catecholamine exposure during this critical period of embryonic development can have serious and sometimes grave consequences.

4. A role for endogenous cardiac catecholamines in pacemaker cell development?

In 2001, adrenergic cells were identified in the SAN region of the developing embryonic rat heart, where they were found to make a transient appearance early in development (Fig. 3) [15]. The timing of this localization, which occurs as early as embryonic day 8.5 (E8.5) in mice, [16] demonstrates that these adrenergic cells are likely of cardiac, rather than neural, origin. Indeed, the patterns for both the timing and distribution of adrenergic cells in the early embryonic heart are inconsistent with a neural crest origin, reflecting instead an intrinsic cardiac origin for these cells [14,16]. These intrinsic cardiac adrenergic (ICA) cells were first seen clustered in the SAN and atrioventricular canal regions of the developing rat heart around E11.5. They largely disappear from these regions within 1–2 days, and re-appear several days later (E15.5/E16.5) clustered in a knot-like formation along the crest of the ventricular septum where the bundle of His is thought to form [15]. At these later stages of pre- and early postnatal development, ICA cells are also found sporadically throughout both ventricles, particularly along the sub-endocardial surfaces and within the septum. These regions generally comprise the ventricular conduction system. Thus, ICA cells were found to be transiently and progressively associated with development of the major components of the cardiac pacemaking and conduction systems including the SAN, atrioventricular node, bundle of His, and ventricular conduction system.

Fig. 3

Sagittal section of an embryonic rat heart (E11.5) showing the presence of catecholamine-producing cells in the presumptive SA node region prior to innervation. The section was co-stained for the epinephrine biosynthetic marker, Pnmt (green), and a muscle-specific marker, α-actinin (red). Staining was performed using indirect immunofluorescence histochemistry, and the captured images were combined and artificially colorized using Adobe Photoshop software. Reprinted from the cover of Circulation Research[15].

Using a genetic fate-mapping strategy in mice, it was subsequently shown that ICA cells eventually become myocytes found throughout the heart, notably including all of the major pacemaking and conduction system areas [16]. Within the SAN, for example, approximately half of the pacemaker myocytes were derived from the adrenergic lineage as could be seen in cells co-stained for β-galactosidase, marking adrenergic lineage cells due to its activation by expression of Cre-recombinase that had been “knocked-in” to the gene locus encoding the adrenergic biosynthetic enzyme, Pnmt, and a pacemaker cell marker, the I_f-gating HCN4 channel (Fig. 4) [33]. HCN4 is highly concentrated in the SAN where the highest densities of I_f are also found [49]. Although HCN4 expression is not restricted to SAN myocytes within the heart, these pacemaking myocytes nevertheless represent the location where cardiac HCN4 expression is clearly most abundant (Fig. 4A) [16,49]. The fact that up to half of the HCN4⁺ myocytes in the SAN (determined by HCN4 immunofluorescent staining and anatomical location) were also positive for β-galactosidase in these mice (Fig. 4F and H), indicates that a substantial number of pacemaking myocytes in the SAN were derived from cells with an adrenergic lineage, presumably represented by the ICA cells that had populated the SAN region during earlier stages of cardiac development. These data provide the strongest evidence to date in support of adrenergic cells within pacemaking centers of the heart.

Fig. 4

Many HCN4⁺ pacemaker cells (green fluorescence) in the E15.5 mouse SAN region express β-galactosidase (blue XGAL staining) in Pnmt-Cre/R26R reporter mice. Co-immunofluorescent histochemical staining of wild-type (Pnmt^+/+) E15.5 mouse heart sections for (A) HCN4 (arrow depicts SAN myocytes, and arrowhead indicates presumptive AVN myocytes), and (B) sarcomeric α-actinin. RA, right atrium; RV, right ventricle. Scale bar, 1.0 mm. (C–H) Frontal series of E15.5 Pnmt–Cre/R26R heart sections stained for expression of LacZ (C–E) and the pacemaker channel protein, HCN4 (panels G and H). An expanded view of the boxed region (SAN) in panel E is shown in panel F. Upon switching to dark-field fluorescence microscopy, HCN4 green fluorescent staining can be seen in this section (panel G). To evaluate co-staining for XGAL and HCN4 in this section, the images in panels F and G were combined to produce the Overlay image depicted in panel H. Please note that the green fluorescence surrounding the vessel shown in panels G and H appears to be due to autofluorescence. Abbreviations: Ao, aorta; AoV, aortic valve; MiV, mitral valve; SAN, sinoatrial node (boxed region of panels E and F). Scale bar (for panels F–H), 0.1 mm. Reprinted from Developmental Dynamics[16].

It is important to note that ICA cells appear to be only transiently associated with pacemaking and conduction tissue in the developing heart [13,15]. Thus, although it is difficult to determine active from past expression status of the endogenous catecholamine biosynthesis genes from the β-galactosidase fate-mapping data shown here (Fig. 4), [16] previously published data suggest that there is relatively little active Pnmt gene expression in the SAN during late stages (e.g., E15.5 to birth) of prenatal development [13,15,16]. It follows, then, that the β-galactosidase expression found in the SAN at E15.5 and later stages of development is most likely the result of past Pnmt–Cre recombinase expression in those cells.

These observations are not restricted to pacemaking and conduction system regions of the heart, as ICA cells appear to give rise to myocytes found throughout the heart. By E15.5, for example, most of the ventricular septum, right ventricle, and to lesser extent, the left ventricular free wall and both atrial chamber walls contained myocytes derived from an adrenergic cellular phenotype (Fig. 4C and D) [16]. Although some of these cells retain an adrenergic phenotype as development proceeds into the postnatal period, the vast majority appear to have differentiated into cardiomyocytes in all parts of the heart.

Based on this new information, we propose that ICA cells differentiate into pacemaking and other types of cardiomyocytes. During the process of differentiation, catecholamine biosynthetic machinery would be down-regulated as expression of myocardial-specific genes would be up-regulated. In support of this hypothesis, a subset of ICA cells in the developing SAN of the rat heart was observed to also express α-actinin, a structural component of the sarcomere, [15] perhaps reflecting cells that were “transitioning” from an adrenergic to a myocardial phenotype. Similarly, a recent study showed that ICA cells “generate spontaneous [Ca²⁺]_i transient activity”, [50] a characteristic of automaticity shared with primitive embryonic and SAN pacemaking myocytes. Taken together, these results suggest that ICA cells can develop an intermediate phenotype that has characteristics of both adrenergic and myocardial cells, with the vast majority of these becoming specialized myocardial cells that serve to generate, conduct, and respond to electrical signals. Thus, ICA cells may serve as novel cardiomyocyte progenitor (CMP) cells. This concept is diagrammatically displayed in Fig. 5.

Fig. 5

Hypothetical model for how ICA cells may contribute to cardiomyocyte development. This picture represents a simplified view of how cardiomyocytes may differentiate from embryonic stem (ES) cells. The question marks in the block arrows indicate uncertainty about the number of steps required in between the specified cell types shown. Abbreviations: CS, cardiac stem; CMP, cardiomyocyte progenitor; SAN, sinoatrial node; AVN, atrioventricular node; His, His bundle; Pur, Purkinje fiber.

5. What are the essential physiological functions of catecholamines during development?

The short answer to this question is that we do not yet know exactly what catecholamines do during the normal processes of development to ensure survivability. Perhaps the simplest hypothesis for cardiac catecholamines is that they may primarily serve as stress hormones, analogous to their role in adults. A prominent in utero stressor is hypoxia, which appears to occur sporadically throughout gestation [51]. Recent evidence showed that catecholamine-deficient (Th^−/−) fetal mouse heart rates fail to compensate during an applied hypoxic stress in ex vivo cultures. Application of the β-adrenergic agonist, isoproterenol, led to full recovery of heart rate in the Th^−/− hearts [51]. Interestingly, ICA cells themselves appear to be sensitive to hypoxia because it was recently shown that intracellular Ca²⁺ transient activity was dampened in ICA cells during hypoxia and re-appeared during re-oxygenation of the cells [50]. Thus, ICA cells can ‘sense’ hypoxia and may respond by secreting catecholamines.

It is important to note, however, that Th^−/− hearts were unable to maintain stable heart rates over time compared to heterozygous (Th^+/−) and wild-type (Th^+/+) hearts even under normoxic conditions, [51] and a slight, but significant bradycardia was observed under normoxic conditions in a similar mouse embryonic model in another study [12]. Hearts that are unable to maintain physiological beating activity during prenatal development would be destined to fail.

Clues to other roles for catecholamines may be gained by considering the timing of their developmental importance. For example, the appearance of catecholamines in the developing heart roughly coincides with the onset of beating activity while the heart is still essentially a tubular structure. Initially, cardiac contractions are weak, sporadic, and often unsynchronized. They are also unresponsive to the actions of catecholamines [52]. Over the next few days, the contractions become more regular, forceful, and sensitive to the chronotropic effects of catecholamines [52,53]. At the same time, the heart undergoes tremendous morphological development as it folds, septates, and grows during this most dynamic phase of cardiovascular development. Studies with catecholamine-deficient mice indicated that there were no gross morphological cadiac abnormalities associated with the deficiency, though some thinning of the atrial chamber walls and disorganization of ventricular myofibrils was noted [11]. In addition to these relatively subtle morphological abnormalities, substantial blood congestion was found in the liver and major vessels. Taken together with the aforementioned bradycardia, the cause of death in catecholamine-deficient embryos was deemed to be cardiovascular failure [11,12].

Thus, one explanation for the relevance of catecholamines during early stages of cardiac development is that they may be needed to stimulate and synchronize cardiac contractions shortly after the heart starts to beat. Catecholamines are ideal candidates for such a role since they are secreted and presumably act on nearby myocytes in a paracrine and autocrine manner. Adrenergic receptors would be simultaneously stimulated on multiple cells, leading to coordinated intracellular rises in cAMP and cytoplasmic Ca²⁺. In pacemaker cells, I_f would be stimulated by direct binding of cAMP to HCN4 channel subunits [33]. Cyclic AMP elevations lead to activation of PKA, which in turn, mediates phosphorylation of key proteins involved in intracellular Ca²⁺-handling, including the major membrane Ca²⁺current, I_Ca,L, [27,28,54,55] RyR2, [6] and phospholamban, PLB, [17] resulting in enhanced membrane I_Ca,L, greater Ca²⁺ release into the cytoplasm from the SR, and more efficient reloading of the SR stores, respectively. These actions produce an overall increase in the rate of cardiac contractions and may facilitate synchronization [56].

A highly simplified schematic view of this idea is shown in Fig. 6. Although not shown in this diagram (Fig. 6), one could imagine that multiple nearby cells would be ‘simultaneously’ stimulated by local release of catecholamines from ICA cells. Thus, catecholamines may directly contribute to intercellular as well as intracellular synchronization of pacemaking activity.

Fig. 6

Revised model for the role of catecholamines in cardiac pacemaking. Simplified model depicting how catecholamines (NE and/or EPI) may influence the development and maintenance of pacemaking activity in SAN myocytes. The “Inactive” cell shown on the left has all the necessary cellular ‘machinery’ for pacemaking function, but shows only weak, asynchronous, and sporadic pacemaking activity due to lack of external stimulation by catecholamines (in the extreme case, these cells may not beat at all). Developmental cues (presently unknown), stochastic release, and/or uterine stressors such as hypoxia induce ICA cells to secrete NE and EPI which act in paracrine fashion on nearby primitive pacemaking cells to activate them, leading to transient increases in cAMP and PKA that, in turn, lead to stimulation of HCN membrane channels by direct cAMP binding, and phosphorylation of key Ca²⁺ regulatory molecules including the L-type calcium channel, RyR2, and PLB, all of which have the net effect of enhancing the intracellular Ca²⁺ oscillations that are thought to be crucial for pacemaking activity. This “Active” pacemaker cell is induced by NE and/or EPI, and can revert to an “Inactive” state upon removal of the catecholamines. The rightmost diagram also depicts an “Active” cell, but it no longer requires the continuous presence of catecholamines as intracellular concentrations of cAMP and PKA are already constitutively activated [17]. These constitutively activated pacemaking myocytes can still be further induced by NE and EPI which leads to increased pacing rates, but they are no longer required for basal pacing activity. The signal(s) that induce constitutive activation of pacemaking cells and how these signals are transduced intracellularly are undefined (represented by question marks). Characteristics of the beating activity associated with each of the proposed pacemaker states are indicated below each cell diagram (e.g., ‘asynchronous’ vs. ‘synchronous’). For the sake of simplicity, α-adrenergic receptor pathways and other potentially important signal transduction molecules were omitted in these diagrams, though we recognize that they may also play significant roles in this process. Abbreviations: AC, adenylate cyclase; β-AR, beta-adrenergic receptor; EPI, epinephrine; HCN, hyperpolarization-cyclic nucleotide gated channel; ICA, intrinsic cardiac adrenergic cell; L-type Ca²⁺, L-type calcium channel; NE, norepinephrine; PLB, phospholamban; RyR2, Type 2 ryanodine receptor; SERCA, sarcoplasmic reticulum Ca²⁺ ATPase; SR, sarcoplasmic reticulum.

One intriguing related hypothesis that has recently been proposed suggests that adult pacemaker cells have constitutively elevated levels of cAMP and PKA, [17] thereby ‘bypassing’ or ‘short-circuiting’ the need for catecholamines (see Fig. 6, right panel). Please note that even in this “constitutive” state of activation, these cells are still able to respond to exogenous catecholamines by increasing their firing rate [17]. It may be interesting to see if catecholamines are involved in the developmental induction (i.e., differentiation) of pacemaker cells during the critical period of embryonic development for cardiac catecholamine action. This testable hypothesis would predict that transient exposure to catecholamines and/or other intercellular signaling molecules during the critical period of embryonic sensitivity would induce a subset of cardiomyocyte progenitor cells within the SAN region to constitutively activate cAMP and PKA. Such actions are in line with the idea that “prenervous” (i.e., before development of a nervous system) neurotransmitters may serve as important intercellular signaling molecules during early stages of development [57].

One key component missing from the diagram shown in Fig. 6 is the sodium–calcium exchanger, NCX1, which is expressed early in cardiac development [58] where it is thought to play a critical role extruding Ca²⁺ and providing an inward current in primitive pacemaking cells [59]. Indeed, genetically-altered mice with NCX1 deficiency die early in development (∼E9) with no detectable heart beats [60–62]. Although catecholamines do not appear to influence mammalian NCX1 activity directly, [63] they may nevertheless affect NCX1 by stimulating expression of the NCX1 gene through α₁-adrenergic receptors [64,65].

As indicated above, NCX1 appears to be one of the critical components of early pacemaking activity in the developing embryo. As also indicated earlier, this primitive pacemaking activity can occur in the absence of HCN4, and is not sensitive to cAMP [31]. These results demonstrate that the onset of cardiac pacemaking activity not only does not require catecholamines, but appears to be insensitive to their actions. The mechanisms underlying the early catecholamine-insensitive embryonic pacemaking activity are unclear at present, but may involve the actions of inositol 1,4,5-triphosphate receptors, which are expressed at the earliest developmental time periods and are capable of modulating Ca²⁺ oscillations from intracellular stores [66–68]. Within 1 or 2 days after the initiation of the heartbeat, however, expression of the key components of adult-like pacemaker cells, including I_Ca,L, HCN4, RyR2, SERCA, and PLB become functional and display steady increases in expression through the remainder of prenatal and early postnatal development [35,69]. I_Ca,L together with the transient outward K⁺ current, I_to, are some of the first membrane currents to be expressed in embryonic cardiomyocytes [70]. Although I_Ca,L may be initially unresponsive to catecholamine stimulation, [44,52] it can be enhanced by β-adrenergic stimulation as early as E9.5 in mice, [71] which appears to mark the beginning of catecholamine influence in the developing heart and is coincident with the onset of the “Critical Period” for catecholamines in early development (see Fig. 2).

Of course, catecholamines have widespread actions, and it is certainly possible that other mechanisms may also be important. Surprisingly little is known, for example, about how ICA cells may contribute to innervation of the heart. They could also act indirectly by influencing metabolism, [72] apoptosis, [73] proliferation, [74] conduction, [75] and many other physiological processes. Clearly, much work still needs to be done before we fully understand how catecholamines specifically influence cardiovascular development.

6. Conclusions

Based on the data reviewed here, we propose a modified version of Pollack's hypothesis: Pacemaker cells transiently produce, store, and secrete catecholamines which act via autocrine and paracrine mechanisms to synchronize pacemaking activity and facilitate transition from the weak, sporadic, and arrhythmic contractions that appear when the heart first begins to beat during embryonic development to the more forceful, regular, and rhythmic contractions that are characteristic of adults. A corollary to this hypothesis would be that once transient exposure to catecholamines has helped facilitate the maturation of pacemaker cells, they are no longer needed to maintain pacemaking activity.

Although catecholamines may not play strictly “obligatory” roles with respect to pacemaking as Pollack originally described, there is little doubt that they have an intimate relationship with cardiac pacemaking activity that begins shortly after the heart first begins to beat and continues until the end of life in most vertebrate animals. Given the prevalence of cardiovascular disease and the need for new and more effective therapies, including the current interest in the therapeutic potential of embryonic and adult stem cells, uncovering the physiological nature of the relationship between catecholamines and the development of cardiac function is more imperative now than ever.

Acknowledgements

This work was supported by grants (S.N.E.) from the National Institutes of Health (HL 78716), The Stem Cell Research Foundation (a program of the American Cell Therapy Research Foundation), and a postdoctoral fellowship (D.G.T.) from the American Heart Association.