How does the shape of the cardiac action potential control calcium signaling and contraction in the heart?

The function of the heart is to pump blood through the pulmonary and systemic vasculature. To do this, the atria and ventricles must contract in a precise sequence in response to a conducted electrical signal, the action potential (AP).

The cardiac AP originates in the sinoatrial node and is conducted to the atria and ventricles through electrical synapses: the gap junctions. The AP morphology varies with species, heart rate, location within the heart, developmental stage, and in response to neurohormones and drugs. Unlike the brief APs of skeletal muscle and neurons, which typically last ≈3-5 ms, the cardiac action potential is 100’s of milliseconds long and has five distinct phases (Figure 1). One of these components, sometimes considered one of little importance, the “rapid repolarization” period (also called “phase 1”) is the focus of a clever paper presented in the current issue of Journal of Molecular and Cellular Cardiology [1].

Figure 1.

A typical action potential (AP) waveform of a healthy adult human ventricular myocyte (modified from Cooper et al. [1]).

In Fig. 1, we use a color-coded human AP to illustrate its five phases. Phase 0 (red) is the leading edge of the AP and corresponds to the period of rapid depolarization (depolarization rate ≈ 250 V/s) from the diastolic membrane potential of ≈ −90 mV to +50 mV. It is produced primarily by the activation of cardiac Na⁺ channels with a contribution from cardiac Ca²⁺ channels. Phase 1 (blue), also referred as the “notch”, is produced by termination of the depolarizing inward Na⁺ current due to inactivation and the concomitant activation of the transient outward K⁺ (I_to) and Na⁺/Ca²⁺ exchanger (I_NCX) currents. This results in a transient repolarization period (i.e. the notch) as the membrane potential goes from ≈ +50 to +30 mV (or more negative). Phase 2 (green) is the long (hundreds of milliseconds) plateau phase of the action potential during which the membrane potential changes little. This is produced by the balance of small, but non-inactivated Na⁺ and L-type Ca²⁺ current (I_Ca) components and hyperpolarizing K⁺ currents. In phase 3 (orange), reductions of Ca²⁺ and Na⁺ currents and increasing K⁺ current contribute to the repolarization of the myocyte. Towards the end of phase 3 the repolarizing actions of these K+ currents are opposed by an inward I_NCX produced by the Na⁺/Ca²⁺ exchanger operating in its Ca²⁺ extrusion mode. Phase 4 (black) is the diastolic membrane potential of quiescent ventricular myocytes of (≈ −90 mV). The detailed molecular identities of the membrane currents that contribute to each phase depend on the cardiac tissue type and anatomic location of the tissue as well as the specie and developmental age of the animal [2]. Similarly, the exact AP morphology depends on these same variables.

The coupling of cardiac membrane depolarization to Ca²⁺ release from the sarcoplasmic reticulum (SR) in myocytes (i.e. excitation-contraction (EC) coupling) takes place during the initial phases of the action potential when the L-type Ca²⁺ channels open. EC coupling is dependent on the positioning of the L-type Ca²⁺ channels within 15 nm of the junctional sarcoplasmic reticulum (jSR) to form a functional unit called a “couplon” [3]. The Ca²⁺ influx activates the nearby cluster of SR Ca²⁺ release channels in the couplon, the ryanodine receptors (RyRs) to produce Ca²⁺ sparks [4]. Synchronization of the Ca²⁺ sparks during the AP produces the cell-wide [Ca²⁺]_i transient that activates contraction. Voltage and Ca²⁺-dependent inactivation rapidly terminate I_Ca. Because of the exquisite nano-anatomy of the couplon, the opening of single or small cluster of L-type Ca²⁺ channels could activate a Ca²⁺ spark [5-7]. This central component of EC coupling may thus link changes in AP morphology to EC coupling and the [Ca²⁺]_i transient.

Recent studies suggest that modification of the ventricular action potential is a critical early step in the chain of events linked to the development hypertrophy and heart failure [8-12]. Indeed, changes in the AP waveform can be observed as early as 48 hours after myocardial insult, even in the absence of measurable cardiac hypertrophy. These changes include reductions in the rate of depolarization and peak depolarization during phase 0, a decrease in the repolarization during phase 1 (or even complete loss of the notch), and an increase in the duration of the action potential.

Decreased [Ca²⁺]_i transients and contractility are also a hallmark of pathological hypertrophy and heart failure. The hypothesized mechanisms underlying these changes in function can be subdivided in two general categories. The first involves a decrease in the coupling strength between L-type Ca²⁺ channels and ryanodine receptors [13]. The second mechanism involves decreased Ca²⁺ storage in the sarcoplasmic reticulum. The later mechanism is likely due to decreased SERCA function and enhanced Na⁺/Ca²⁺ exchanger expression.

In the current issue of the Journal of Molecular and Cellular Cardiology, an article from Mark Cannell’s laboratory [1] revisited the following important question: what is the relationship between changes in action potential waveform and EC coupling during heart failure? In this study, they focused on the importance of the loss of the notch during phase 1 of the human cardiac AP during heart failure (HF) with respect to the magnitude and the kinetics of local and global [Ca²⁺]_i transients. Although the functional importance of phase 1 of the action potential on EC coupling has been examined before under physiological [14] and pathological [15] conditions, the elegant study by Cooper et al. [1] provides novel mechanistic insights into the mechanisms underlying changes in EC coupling during HF.

By voltage clamping single cardiac myocytes from rat and rabbit using both normal (notched AP) and HF human AP (without notches), the authors show that EC coupling is significantly impaired during a failing AP. I_Ca during a failing AP shows a reduced peak amplitude and slower inactivation and the cell-wide (global) [Ca²⁺]_i transient also shows reduced peak amplitude and slower rise-time. The authors show that the observed change in I_Ca and [Ca²⁺]_i transient amplitude is not caused by action potential prolongation during a failing AP. By varying an artificially simulated notch voltage while fixing the action potential duration, the authors show that both, peak and integrated I_Ca, as well as [Ca²⁺]_i transient amplitude decrease with decreased notch repolarization.

Using confocal microscopy, the authors further show how EC coupling is impaired when a failing AP causes spatial heterogeneity in Ca²⁺ transient and Ca²⁺ “spike” activation [16]. Due to impaired E-C coupling, the SR load is increased, which the authors demonstrate when a Ca²⁺ transient invoked by a normal AP is larger in amplitude when it is immediately preceded by an AP with HF morphology. Increased SR Ca²⁺ can lead to increased risk of arrhythmogenesis [17].

The recent work by Dong et al [18] is at odds with the studies by Cooper et al. [1], Sah et al. [14], and Harris et al. [15]. Dong et al. used the dynamic clamp to quantitatively examine the influence of I_to on [Ca²⁺]_i and contraction of canine left ventricular endocardial and epicardial myocytes. In endocardial myocytes, where the native I_to is small, simulation of a large I_to increased phase 1 repolarization, but significantly decreased [Ca²⁺]_i transient and cell shortening. Accordingly, subtraction of the native I_to via the dynamic clamp enhanced contractility in epicardial cells. On the basis of these data, Dong et al [18] concluded that there is an inverse correlation between I_to levels and myocyte contractility and [Ca²⁺]_i transient amplitude in canine epicardial and endocardial myocytes. Further experiments may be needed to resolve the conflicts.

Is the decrease in I_Ca density associated with the absence of the repolarizing notch seen by Cooper et al. sufficient to explain the significant impairment in EC coupling observed during HF? The authors use a modified Monte Carlo stimulation of L-type Ca²⁺ channel and RyR gating, which takes into account Ca²⁺ dependent inactivation of L-type Ca²⁺ channels, to show that the simulated Ca²⁺ sparks agree well with the experimental data. The peak rate of Ca²⁺ spark production is halved, and the rate of initial spark triggering is slowed, while the appearance of late Ca²⁺ sparks due to stochastic L-type Ca²⁺ channel openings is increased. The authors report that the simulation shows that the change in EC coupling caused by the failing AP, with its absence of the repolarizing notch, can be entirely explained by the change in I_Ca flux.

In combination with the work by Harris et al. [15], the study by Cooper et al. [1] is likely to change all current models of EC coupling since there are so many influences in AP morphology. This is particularly relevant for our understanding of the development of EC coupling dysfunction during the onset of HF. Decreased I_to is a seemingly general response to myocardial insults, which has been detected as early as 48 hours following myocardial infarction [10]. Down-regulation of I_to precedes changes in SERCA, Na⁺/Ca²⁺ exchanger, and RyR proteins. The findings by Cooper et al. [1] would suggest that this, by itself, would directly translate into a change in EC coupling, which would later be exacerbated by functional decoupling of I_Ca and RyRs, down-regulation of SERCA, and/or up-regulation of Na⁺/Ca²⁺ exchanger.

Could up-regulation of I_to restore EC coupling during HF? A recent study by Jin et al. [19] suggest an answer to this important question. They showed that increasing I_to density by expressing the accessory subunit KChIP2 in failing ventricular myocytes attenuated HF and improved EC coupling. Thus, I_to could be a key, bi-directional regulator of EC coupling and a potential target for restoration of cardiac function during HF.

Although Cooper et al. [1]focused on HF, their findings also provide insights into the mechanisms underlying regional variations in EC coupling in the heart. I_to and hence phase 1 of the action potential varies throughout the heart. I_to is larger in epicardial than in endocardial myocytes of the left ventricular free wall. Yet, [Ca²⁺]_i transients are larger in endocardial than in pericardial myocytes. In combination with the findings by Cooper et al. [1], this suggests that while the changes in the magnitude of phase 1 of the action potential could alter EC coupling, they are not sufficient to explain regional variations in [Ca²⁺]_i transients in the heart. Under physiological conditions, regional and developmental variations in action potential waveform and the expression and function of Ca²⁺ handling proteins (e.g. Na⁺/Ca²⁺ exchanger, RyR2) are the likely culprits for heterogeneous EC coupling under physiological conditions [20].

The study by Cooper et al. [1] has broad implications on EC coupling. In ventricular myocytes, the probability of Ca²⁺ spark occurrence (P_Spark) in a couplon during the action potential is proportional to the open probability (P_o) of Cav 1.2 channels and the square of the Ca²⁺ current and local [Ca²⁺]_i [7]. A recent study suggested that P_Spark approaches unity in specific sites within a cell at the relatively positive membrane potentials (≈ +50 mV) attained during the plateau of the ventricular action potential [5]. This is surprising because, at +50 mV, P_o is less than 1 (≈0.7) and the current through a single L-type Ca²⁺ channel is predicted to be very small (i.e. ≈ 10 fA) at physiological levels of extracellular Ca²⁺. However, as suggested by Cooper et al. [1], by depolarizing to +50 mV the myocyte promotes a high P_o for L-type Ca²⁺ channels, which if followed by a pronounced notch to more negative potentials could enhance the driving force for Ca²⁺ influx and hence the unitary current of L-type Ca²⁺ channels. Indeed, according to the GHK equation repolarization from +50 to +30 during phase 1 would triple the Ca²⁺ influx (≈ 3.09-fold), which would increase P_Spark by nine-fold. Although the opening of a single L-type Ca²⁺ channel can activate a couplon at less positive potentials, a larger number of L-type Ca²⁺ channels are likely to be needed at more positive potentials [6-Santana reference][5]. Indeed, experiments and modeling support the larger number of L-type Ca²⁺ channels activated to trigger a couplon at more positive potentials [5, 21][add Ramay & Sobie ?2009]. Phase 1 activation of Ca²⁺ sparks can thus be attributed to increased P_o of the L-type Ca²⁺ channels and also the larger single channel current that would be observed without the notch. Coupled gating of L-type Ca²⁺ channels may also be a mechanisms for the concerted activation of multiple L-type Ca²⁺ channels and thereby contribute to SR Ca²⁺ release and a high P_Spark [22]. A note of caution is in order, however: There are distinct cellular and molecular structural difference in human and rat ventricular myocytes (e.g. T-tubule organization, RyR2 number, relationship between L-type Ca²⁺ channels and RyR2 clusters, and more). To some extent excitation-contraction coupling in rat ventricular myocytes may not faithfully reproduce the local and cellular [Ca²⁺]_i signals of human ventricular myocytes even when the AP is identical. Nevertheless, Cooper et al. [1] have provided important new evidence suggesting that the notch of the ventricular AP could be an important determinant of P_Spark under diverse and varied physiological and pathological conditions.