Effect of carvedilol on atrial excitation-contraction coupling, Ca²⁺ release, and arrhythmogenicity

Abstract

Carvedilol is an FDA-approved β-blocker commonly used for treatment of high blood pressure, congestive heart failure, and cardiac tachyarrhythmias, including atrial fibrillation. We investigated at the cellular level the mechanisms through which carvedilol interferes with sarcoplasmic reticulum (SR) Ca²⁺ release during excitation-contraction coupling (ECC) in single rabbit atrial myocytes. Carvedilol caused concentration-dependent (1–10 µM) failure of SR Ca²⁺ release. Failure of ECC and Ca²⁺ release was the result of dose-dependent inhibition of voltage-gated Na⁺ (I_Na) and L-type Ca²⁺ (I_Ca) currents that are responsible for the rapid depolarization phase of the cardiac action potential (AP) and the initiation of Ca²⁺-induced Ca²⁺ release from the SR, respectively. Carvedilol (1 µM) led to AP duration shortening, AP failures, and peak I_Na inhibition by ~80%, whereas I_Ca was not markedly affected. Carvedilol (10 µM) blocked I_Na almost completely and reduced I_Ca by ~40%. No effect on Ca²⁺-transient amplitude, I_Ca, and I_Na was observed in control experiments with the β-blocker metoprolol, suggesting that the carvedilol effect on ECC is unlikely the result of its β-blocking property. The effects of carvedilol (1 µM) on subcellular SR Ca²⁺ release was spatially inhomogeneous, where a selective inhibition of peripheral subsarcolemmal Ca²⁺ release from the junctional SR accounted for the cell-averaged reduction in Ca²⁺-transient amplitude. Furthermore, carvedilol significantly reduced the probability of spontaneous arrhythmogenic Ca²⁺ waves without changes of SR Ca²⁺ load. The data suggest a profound antiarrhythmic action of carvedilol in atrial myocytes resulting from an inhibitory effect on the SR Ca²⁺ release channel.

NEW & NOTEWORTHY Here we show that the clinically widely used β-blocker carvedilol has profound effects on Ca²⁺ signaling and ion currents, but also antiarrhythmic effects in adult atrial myocytes. Carvedilol inhibits sodium and calcium currents and leads to failure of ECC but also prevents spontaneous Ca²⁺ release from cellular sarcoplasmic reticulum (SR) Ca²⁺ stores in form of arrhythmogenic Ca²⁺ waves. The antiarrhythmic effect occurs by carvedilol acting directly on the SR ryanodine receptor Ca²⁺ release channel.

INTRODUCTION

The cellular mechanism of excitation-contraction coupling (ECC) reveals marked differences between atrial and ventricular cells (6, 7). Atrial myocytes, in contrast to ventricular myocytes, lack or only have a rudimentary and irregular transverse tubule (TT) system. The absence of TTs causes characteristic spatial and temporal inhomogeneities during action potential (AP)-induced Ca²⁺ release (6–9, 36). Ca²⁺ release initiates in the periphery with the activation of L-type Ca²⁺ channels (LTCCs), where Ca²⁺ influx triggers Ca²⁺-induced Ca²⁺ release (CICR) through ryanodine receptor (RyR) Ca²⁺ release channels from junctional sarcoplasmic reticulum (j-SR). Subsequently, CICR from central nonjunction SR (nj-SR) Ca²⁺ release sites propagates actively in centripetal direction, ultimately leading to global, whole cell elevation of intracellular Ca²⁺ concentration ([Ca²⁺]_i) and atrial contraction (18, 34).

The physiological mechanism of Ca²⁺ release during ECC in atrial cells is reminiscent of the mechanism that sustains pathological arrhythmogenic Ca²⁺ waves. The similarity lies in the fact that both rely on propagating CICR. However, in contrast to the physiological mechanism of atrial ECC, pathological Ca²⁺ waves typically occur spontaneously (i.e., in the absence of an initial electrical trigger in form of an AP and I_Ca), require an increased SR Ca²⁺ load, and RyR and Ca²⁺ release activation by luminal Ca²⁺. Ca²⁺ waves have pathological ramifications and have been linked to cardiac arrhythmias, including atrial fibrillation (10, 13, 20, 24, 25, 35).

β-Adrenergic receptor antagonists (β-blockers), including carvedilol, are clinically used for heart failure treatment to reduce arrhythmia risk, including atrial fibrillation (16), and to prevent sudden cardiac death (15, 40). In vitro, carvedilol has also been described to significantly reduce spontaneous Ca²⁺ waves, presumably by modulating the RyR (43, 45) through reducing mean open time and curtailing Ca²⁺ flux through the channel. Ca²⁺ waves and the physiological mechanism of atrial ECC have in common that both mechanisms involve propagating CICR. Since carvedilol has profound effects on spontaneous Ca²⁺ waves and affects RyR gating, we set out to characterize the effects of carvedilol on atrial ECC and Ca²⁺ release. We tested the effect of the drug on the key elements of ECC, from activation of voltage-gated Na⁺ (I_Na) and Ca²⁺ (I_Ca) currents, AP morphology to activation of RyR-mediated CICR. With the combination of voltage- and current-clamp recordings and spatially resolved Ca²⁺ imaging in atrial myocytes, we demonstrate that carvedilol has inhibitory effects on I_Na and I_Ca, shortens or abolishes the AP, and disrupts physiological Ca²⁺ release in atrial myocytes but protects from spontaneous arrhythmogenic Ca²⁺ waves.

METHODS

Myocyte isolation.

Atrial myocytes were isolated from male New Zealand White rabbits (2.5 kg; 41 rabbits; Charles River, Wilmington, MA). Rabbits were anesthetized with an intravenous injection of pentobarbital sodium (Euthasol; 50 mg/kg) together with heparin (1,000 IU/kg). Hearts were excised and mounted on a Langendorff apparatus and retrogradely perfused with nominally Ca²⁺-free Tyrode (see composition below) solution containing heparin (1,000 IU/kg) for 5 min, followed by perfusion with minimal essential medium Eagle (MEM) solution containing 20 μM Ca²⁺ and 22.5 μg/ml Liberase TH (Roche Applied Science, Indianapolis, IN) for ∼20 min at 37°C. The left atrium was removed and digested for an additional 5 min in the enzyme solution at 37°C. Digested tissue was then minced, filtered, and washed in a MEM solution containing 50 μM Ca²⁺ and 10 mg/ml bovine serum albumin. Isolated cells were kept in MEM solution with 50 μM Ca²⁺ at room temperature (22–24°C) until experimentation. All aspects of animal husbandry, animal handling, anesthesia, surgery, and euthanasia were fully approved by the Institutional Animal Care and Use Committee (IACUC) of Rush University Chicago and comply with United States and United Kingdom regulations on animal experimentation.

Patch-clamp experiments.

Electrophysiological signals were recorded from single atrial myocytes in the whole cell ruptured patch-clamp configuration using an Axopatch 200B patch‐clamp amplifier, the Axon Digidata 1550B interface and pCLAMP 11 software (Molecular Devices, Sunnyvale, CA). Current and voltage recordings were low‐pass filtered at 5 kHz and digitized at 10 kHz. Patch pipettes (1.5–3 MΩ filled with internal solution) were pulled from borosilicate glass capillaries (AM-Systems, Squim, WA) with a horizontal puller (model P‐97; Sutter Instruments, Novato, CA). For Na⁺ (I_Na) and Ca²⁺ (I_Ca) current measurements, myocytes were voltage clamped and held at a membrane potential (V_m) of −80 mV. In experiments testing drug effects on membrane currents, a two-step voltage-clamp protocol was applied to measure I_Na and I_Ca during the same experiment. I_Na was elicited with a 50-ms voltage step from −80 to −50 mV, immediately followed by a voltage step (350 ms) to 0 mV to elicit I_Ca (the voltage-clamp protocol is shown in Fig. 3A, inset). Upon completion of each step command, the command potential was returned to −80 mV for 1.4 s. During extended I_Na and I_Ca recordings, a substantial current rundown was observed, especially of I_Na. To correct for such rundown, the same protocol was used to record I_Na and I_Ca in the absence of the drug. I_Na and I_Ca were normalized for I_max, and the averaged normalized current data from the carvedilol experiments were divided by the averaged normalized data recorded in the absence of carvedilol. To establish the current-voltage (I–V) relationship of I_Ca, the holding potential (V_H) was −80 mV and cells were depolarized by a prepulse to −50 mV (50 ms) immediately followed by depolarization steps of 350-ms duration in increments of 10 mV to a maximum depolarization of +60 mV. For the measurement of the I–V relationship for I_Na, single myocytes were held at −80 mV, and 30-ms voltage steps were applied from −80 to +30 mV in 5-mV increments. For current-clamp measurements, the whole cell fast current-clamp mode of the Axopatch 200B was used, and APs were evoked by 5-ms stimulation pulses of a magnitude ∼1.5 times higher than AP activation threshold. V_m measurements were corrected for a junction potential error of −10 mV.

Fig. 3.

Carvedilol effect on voltage-gated Na⁺ current (I_Na). A and B: effect of 1 µM (A) and 10 µM (B) on I_Na: representative I_Na traces (a) and time course (b) of carvedilol effect on average peak I_Na corrected and normalized for rundown. A,b, insert: voltage-clamp protocol. Myocytes were held at membrane potential = −80 mV. A 2-step voltage-clamp protocol was applied to measure I_Na and L-type Ca²⁺ current (I_Ca) (Fig. 4) during the same experiment. I_Na was elicited with a 50-ms voltage step from −80 to −50 mV, immediately followed by a voltage step (350 ms) to 0 mV to elicit I_Ca. Upon completion of each step command, the command potential was returned to −80 mV for 1.4 s. A,c and B,c: summary data for uncorrected currents in control (Ctrl) and after 1- and 3-min carvedilol exposure. Means ± SE and individual cell data are shown (n = 8). *P < 0.02; **P < 0.003 (ANOVA). NS, not significant.

For I_Na and I_Ca measurements, the internal solution consisted of (in mM) 130 Cs-glutamate, 20 CsCl, 0.33 MgCl₂, 4 MgATP, 5 EGTA, and 10 HEPES with pH adjusted to 7.2 with CsOH. External Tyrode solution was composed of (in mM) 135 NaCl, 4 CsCl, 2 CaCl₂, 1 MgCl₂, 10 HEPES, and 10 d-glucose (pH 7.4 with NaOH). For I–V curve of I_Na, the internal solution consisted of (in mM) 5 NaCl, 20 CsCl, 5 EGTA, 4 MgATP, and 10 HEPES (pH 7.2 with CsOH), and the low-sodium external solution consisted of (in mM) 20 NaCl, 120 CsCl, 1 MgCl₂, 10 HEPES, and 10 d-glucose (pH 7.4 with CsOH). Current-clamp and [Ca²⁺]_i experiments were conducted using external Tyrode solution containing (in mM) 135 NaCl, 4 KCl, 2 CaCl₂, 1 MgCl₂, 10 d-glucose, and 10 HEPES (pH 7.4 with NaOH). For AP measurements, the internal solution contained (in mM) 130 K-glutamate, 5 NaCl, 15 KCl, 0.33 MgCl₂, 4 MgATP, and 10 HEPES (pH 7.2). Experiments were conducted at room temperature (22–24°C).

[Ca²⁺]_i measurements.

To monitor cytosolic Ca²⁺ transients and Ca²⁺ waves, atrial myocytes were loaded with the membrane-permeable Ca²⁺ indicator Fluo-4 AM (10 μM) and 0.2% of pluronic acid (Invitrogen) for 20 min and then washed twice for 10 min with Tyrode solution. Confocal microscopy (Bio-Rad Radiance 2000 and Nikon A1R) were used to image [Ca²⁺]_i. Fluo-4 was excited at 488 nm and emission collected at wavelengths > 515 nm. Fluo-4 [Ca²⁺]_i signals are presented as F/F₀ ratios, where F₀ represents diastolic [Ca²⁺]_i during electrical pacing. Ca²⁺-transient (CaT) amplitudes are quantified as ∆F/F₀, where ∆F/F₀ = (F − F₀)/F₀. [Ca²⁺]_i measurements were conducted in the line scan mode (512 lines/s) using a ×60 water-immersion (Bio-Rad Radiance 2000) and a ×60 oil-immersion (Nikon A1R) objective lens. The scan line was placed along the transversal axis of the cell avoiding the nucleus with a pixel size of 0.08–0.15 µm. Cells were superfused with external Tyrode solution. CaTs were elicited by electrical field stimulation of intact atrial myocytes using a pair of platinum electrodes (voltage set at ∼50% above the threshold for contraction, 3-ms duration). Ca²⁺ waves were recorded during a 20-s period of rest after electrical stimulation at 1 Hz. To increase the probability of spontaneous Ca²⁺ waves in these experiments, extracellular Ca²⁺ concentration ([Ca²⁺]_o) was increased to 7 mM. Experiments were conducted at room temperature (22–24°C).

Chemicals.

Chemicals were from Sigma‐Aldrich (St. Louis, MO), unless otherwise stated. Fluo‐4 AM was obtained from Thermo Fisher Scientific (Waltham MA). Carvedilol and metoprolol were dissolved in DMSO at a concentration of 40 mM. The final DMSO concentration in extracellular solutions was 0.025%.

Data analysis and presentation.

Results are presented as individual observations or as means ± SE, and n represents the number of individual cells. Statistical significance was evaluated using unpaired and paired Student's t-test for single comparisons, and two-way ANOVA, followed by Tukey's post hoc test for multiple comparisons. Statistical analysis was conducted with SigmaPlot (v. 11; Systat Software). Differences were considered significant at P < 0.05. NS indicates statitically not significant.

RESULTS

Carvedilol effect on atrial ECC and SR Ca²⁺ release.

We tested the effect of carvedilol on AP-induced Ca²⁺ release in isolated rabbit atrial myocytes. Cytosolic CaTs were elicited by electrical field stimulation (1 Hz). We found that carvedilol had profound, dose-dependent effects on electrically evoked CaTs. As shown in Fig. 1A, carvedilol diminished the CaT amplitude in a dose-dependent manner and led to disruption of ECC that became more severe with increasing carvedilol concentration. ECC disruption was defined here as the failure of electrical stimulation to elicit CaTs. Figure 1B summarizes the degree of ECC disruption. A single failure of Ca²⁺ release during a test period of 10 s counted as a disruption of ECC. At the lowest concentration tested (0.5 µM), carvedilol did not disrupt ECC (n = 0/10). At higher concentrations, failure of electrical stimulation to elicit CaTs was observed in 14% (1 µM; n = 2/14), 61% (3 µM; n = 13/21), and 85% (10 µM; n = 17/20) of cells, respectively. Figure 1C summarizes the carvedilol concentration-dependent decrease in whole cell CaT amplitude (∆F/F₀). CaT amplitudes were determined by averaging the amplitudes of 10 consecutive CaTs (including disruptions). In control, mean CaT amplitude was ∆F/F₀ = 3.32 ± 0.27, compared with 2.88 ± 0.38 (carvedilol 0.5 μM; n = 10, P = 0.14), 1.79 ± 0.51 (1 μM, n = 14, P < 0.005), 0.63 ± 0.31 (3 μM, n = 21, P < 0.005), and 0.36 ± 0.21 (10 μM, n = 20, P < 0.005). Since carvedilol is a clinically used β-blocker, we tested whether the carvedilol effects on CaTs and ECC resulted from its antiadrenergic action. In contrast to carvedilol, however, metoprolol (FDA-approved β-blocker) did not significantly affect CaT amplitude (3.24 ± 0.52; n = 7, P = 0.21; Fig. 1C), arguing against an antiadrenergic effect.

Fig. 1.

Excitation-contraction coupling (ECC) failure by carvedilol. Carvedilol inhibits action potential-induced Ca²⁺ release during ECC. ECC disruption was defined as the inability of electrical filed stimulation (1 Hz) to elicit Ca²⁺ transient (CaTs) in isolated rabbit atrial myocytes. A: representative whole cell intracellular Ca²⁺ concentration ([Ca²⁺]_i) profiles and corresponding line scan images in control (n = 65) and during exposure to 0.5 µM (n = 10 cells), 1 µM (n = 14), 3 µM (n = 21), and 10 µM (n = 20) carvedilol and the negative control with metoprolol (Met; 10 µM, n = 7). Black circles indicate time of electrical stimulus. B: fraction of cells (%) showing ECC disruption. A single failure of Ca²⁺ release during the test period of 10 s counted as a disruption of ECC. C: average CaT amplitudes from control and carvedilol-treated atrial myocytes. The [Ca²⁺]_i responses to 10 consecutive electrical stimulations were averaged, including failures. The Fluo-4 Ca²⁺ signal was averaged over the entire width of the cell. For presentation purposes, control (Ctrl) CaT amplitudes were pooled; however, for statistical analysis (t-test for paired measurements), each carvedilol concentration and the metoprolol group had its own control, since experiments were conducted as a paired measurement with recording control CaTs followed by drug application to the same cell. Means ± SE and individual cell data (white circles) are shown. *P < 0.05. NS, not significant.

In the following we systematically tested the key steps of ECC that might be critically affected by carvedilol, from AP generation, membrane depolarization-induced activation of voltage-gated Na⁺ and Ca²⁺ currents, to release of Ca²⁺ from the SR via the RyR Ca²⁺ release channel.

Carvedilol effects on AP.

First, we tested the effect of carvedilol (1 μM) on AP morphology using the whole cell current-clamp technique. APs were elicited by applying 5-ms stimulation pulses of a magnitude ∼1.5 times higher than AP activation threshold at 1 Hz. Representative traces of APs recorded before and during application of carvedilol are shown in Fig. 2A. APs failed after carvedilol treatment in 67% of cells (n = 4/6; Fig. 2B). AP failures occurred at an average rate of 0.43 ± 0.15 Hz during a 10-s test period, whereas in control no AP failures were recorded (Fig. 2A). Resting membrane potential (Ctrl, −73.7 ± 1.7; carvedilol, −76.4 ± 3.2 mV, n = 6, P = 0.11) was similar in control and after drug exposure. Typical APs in control (black), during carvedilol treatment (gray), as well as an AP failure (dashed), are shown in Fig. 2C. After reaching steady state, AP duration (APD) was determined at 30, 50, and 90% repolarization (APD₃₀, APD₅₀, and APD₉₀, respectively) using the average of 10 consecutive APs (no failures included). Carvedilol treatment resulted in APD shortening [APD₃₀ decreased by 21 ± 19% (P = 0.19, n = 6); APD₅₀ by 46 ± 18% (P = 0.04), and APD₉₀ by 24 ± 6% (P = 0.04)]. Summary of the AP shortening is shown in Fig. 2D.

Fig. 2.

Action potential (AP) failure by carvedilol. A: representative AP traces in control (top) and after carvedilol (1 µM) exposure (bottom). APs were recorded in current-clamp mode and evoked by 5-ms stimulation pulses of a magnitude ∼1.5 times higher than AP activation threshold. B: summary of AP failure in percentage of cells (left) and AP failure frequency (right) after carvedilol treatment. C: representative traces of AP waveform in control (black), after carvedilol exposure (gray), and AP failure (dashed). D: average AP duration (APD) at 30, 50, and 90% repolarization (APD₃₀, APD₅₀, and APD₉₀, respectively; n = 6) in control (white bars) and in presence of carvedilol (gray bars). AP failures are not included in APD calculations. *P < 0.05 (paired t-test). V_m, membrane potential; NS, not significant.

Carvedilol effects on I_Na and I_Ca.

To investigate the cause of ECC and AP failures and AP shortening, we investigated the effect of carvedilol on I_Na and I_Ca using the whole cell voltage-clamp technique. In a first set of experiments, I_Na and I_Ca were recorded during the same voltage protocol. I_Na and I_Ca were elicited by applying a two-step depolarization protocol (Fig. 3A, inset). From a holding potential of −80 mV, cells were depolarized to −50 mV for 50 ms to elicit I_Na, immediately followed by further depolarization to 0 mV for 350 ms to elicit I_Ca. The advantage of this protocol is that both currents are recorded from the same cell at essentially the same time; however, at a test potential of −50 mV, only ~60% of peak I_Na is activated (see below Fig. 5). The effect of carvedilol on I_Na and I_Ca was tested for two drug concentrations (1 and 10 µM).

Fig. 5.

Current density-voltage (I–V) curves of voltage-gated Na⁺ current (I_Na; A) and L-type Ca²⁺ current (I_Ca; B). A,a and B,a: I–V relationships in control (black) and during carvedilol treatment (gray). Means ± SE. A,b and B,b: family of current traces at all test voltages in control (Ctrl; black) and during carvedilol treatment (gray) recorded from same cell. The following voltage-clamp protocols were used. I_Ca, holding potential (V_H) = −80 mV; depolarization to −50 mV, followed by 350-ms depolarization steps in increments of 10 mV to a maximum depolarization of +60 mV. I_Na, V_H = −80 mV; 30-ms voltage steps −80 to +30 mV in 5-mV increments. I_Na, n = 8 (carvedilol, 1 µM), *P < 0.03; I_Ca, n = 5 (carvedilol, 10 µM), *P < 0.04 (ANOVA).

Figure 3 shows the effect of carvedilol on I_Na. Representative I_Na traces showing the effect of 1 µM (Fig. 3A,a) and 10 µM (Fig. 3B,a) carvedilol are presented for control (a) and 1-min (b) and 3-min (c) carvedilol treatment. Carvedilol had a profound inhibitory effect on I_Na. Figure 3, A,b and B,b, shows the time course of the averaged, run-down corrected I_Na. During these extended I_Na and I_Ca recordings, we observed a substantial current rundown, especially of I_Na. To correct for such rundown, the same protocol was used to record I_Na and I_Ca in the absence of the drug. I_Na and I_Ca were normalized for I_max, and the averaged normalized current data from the carvedilol experiments were divided by the averaged normalized data recorded in the absence of carvedilol. Figure 3, A,c and B,c, shows average I_Na amplitudes in control and at 1- and 3-min exposure to carvedilol before run-down correction. With 1 µM carvedilol, I_Na density in control was 85.5 ± 25.8 compared with 70.3 ± 15.9 (1-min carvedilol) and 17.0 ± 6.3 pA/pF (3 min). At 10 µM, carvedilol mean control I_Na density was 75.8 ± 21.3 and decreased to 16.3 ± 6.4 after 1 min and 1.2 ± 0.6 pA/pF after 3 min of carvedilol exposure. While both carvedilol concentrations eventually inhibited I_Na almost completely, the time course of inhibition was accelerated in 10 µM carvedilol (t_1/2 of maximal inhibition, 1 µM, 163 s; 10 µM, 77 s).

Figure 4 presents the effect of 1 and 10 µM carvedilol on I_Ca. In contrast to I_Na, the inhibitory effect of carvedilol on I_Ca was much less pronounced. At 1 µM, the average run-down corrected current was only minimally affected (Fig. 4A,b), whereas at 10 µM, an ~21% inhibition of I_Ca was observed after 3 min of drug exposure (Fig. 4B,b). With 1 µM carvedilol, I_Ca density before run-down correction (Fig. 4A,c) was 5.5 ± 1.1 pA/pF in control, compared with 5.4 ± 0.7 (1 min) and 5.2 ± 0.7 (3 min) pA/pF in the presence of the drug. At 10 µM (Fig. 4B,c), mean control I_Ca density was 5.1 ± 0.7 pA/pF and decreased to 3.7 ± 0.4 pA/pF after 1 min and to 3.2 ± 0.4 pA/pF after 3 min of carvedilol treatment (n = 8, P < 0.05).

Fig. 4.

Carvedilol effect on L-type Ca²⁺ current (I_Ca). A and B: effect of 1 µM (A) and 10 µM (B) on I_Ca. For voltage-clamp protocol, see Fig. 3. Representative I_Ca traces (a) and time course (b) of carvedilol effect on average peak I_Ca corrected and normalized for rundown. A,c and B,c: summary data for uncorrected currents in control and after 1- and 3-min carvedilol exposure. Means ± SE and individual cell data are shown (n = 8). *P < 0.02 (ANOVA). NS, not significant.

In a second set of experiments, we evaluated the I–V relationship for I_Na and I_Ca (Fig. 5). For I_Na, depolarization steps (30 ms) were applied from a holding potential of −80 mV in 5-mV increments to a maximum depolarization of +30 mV in the presence and absence of carvedilol (1 µM). An inhibition of I_Na was observed across the entire voltage range tested. Peak I_Na was measured at −40 mV where carvedilol decreased the current from −24.7 ± 4.9 in control to −6.7 ± 2.1 pA/pF after drug exposure (72% inhibition; Fig. 5A,a). At −50 mV, I_Na was −13.7 ± 2.7 pA/pF in control (56% of peak I_Na) and decreased by 70% to −4.2 ± 1.9 pA/pF in carvedilol. Representative traces of the I_Na family in control (black) and after carvedilol treatment (gray) are depicted in Fig. 5A,b. In the voltage range from −50 to +10 mV, I_Na inhibition was statistically significant (P < 0.05, n = 8).

The I–V relationship of I_Ca (Fig. 5B) was evaluated by applying a 50-ms prepulse from a holding potential of −80 to −50 mV to inactivate I_Na, immediately followed by depolarization steps of 350-ms duration in increments of 10 mV to a maximum depolarization of +60 mV. The I–V curves of I_Ca measured at peak inward current are shown in Fig. 5B,a. Carvedilol reduced the I_Ca amplitude without affecting the voltage dependence of I_Ca. On average, carvedilol (10 µM) decreased maximum I_Ca measured at 0 mV by 42 ± 6%. I_Ca was significantly reduced at 0, +10, and +20 mV (n = 5, P < 0.04). Figure 5B,b shows representative traces of the I_Ca family over the voltage range tested in control (black) and during carvedilol treatment (gray).

For both currents, the washout of the drug was slow, incomplete, and variable from cell to cell (data not shown). Therefore, reversibility of current inhibition by carvedilol was not further investigated.

Metoprolol effect on I_Na and I_Ca.

Again, we tested whether the observed carvedilol effect on ion currents were due to its β-blocking action. We applied the same two-step protocol, but using the β-blocker metoprolol (Fig. 6) instead of carvedilol. Figure 6A shows the effect on I_Na. Figure 6A,a shows exemplary I_Na traces before and after 1- and 3-min metoprolol (10 µM) exposure. The time course of run-down corrected average peak I_Na (Fig. 6A,b) revealed that metoprolol had no effect on I_Na. Figure 6A,c shows uncorrected average I_Na data. In this set of experiments, I_Na in control was 46.8 ± 12.0 and decreased to 29.7 ± 6.7 pA/pF after 1 min of metoprolol exposure and further to 24.8 ± 5.5 pA/pF after 3 min.

Fig. 6.

Metoprolol effect on voltage-gated Na⁺ current (I_Na; A) and L-type Ca²⁺ current (I_Ca; B). A,a and B,a: representative I_Na and I_Ca traces. For voltage-clamp protocol, see Fig. 3. A,b and B,b: time course of metoprolol (Met) effect on average run-down corrected currents. A,c and B,c: summary data for uncorrected currents in control (Ctrl) and after 1- and 3-min metoprolol exposure. Means ± SE and individual cell data are shown (n = 8). *P < 0.05 (ANOVA). NS, not significant.

Figure 6B shows metoprolol effects on I_Ca. Representative I_Ca traces of the metoprolol effect are shown in Fig. 6B,a. Like I_Na, I_Ca was completely unaffected by metoprolol when data were corrected for rundown (Fig. 6B,b). Uncorrected average data showed that I_Ca in control was 5.3 ± 0.5 and 4.7 ± 0.4 (1 min) and 4.7 ± 0.3 (3 min) pA/pF in the presence of metoprolol (Fig. 6B,c). From these experiments we concluded that the observed effects of carvedilol, in line with metoprolol effects on CaTs (Fig. 1), were not mediated by the β-blocking potency of this compound.

Effect of carvedilol on RyR function and SR Ca²⁺ release.

In the next set of experiments, we turned our attention to Ca²⁺ release from the SR. Due to the lack or only rudimentary developed TT system in atrial cells, Ca²⁺ release during physiological ECC is spatially and temporally inhomogeneous. Triggered by the AP, I_Ca induces CICR from subsarcolemmal (SS) SR Ca²⁺ release sites (j-SR) from where the elevated subsarcolemmal Ca²⁺ initiates a propagating CICR wave from nj-SR toward the center (CT) of the cell. This spatial inhomogeneity caused by the propagating Ca²⁺ release is visualized in transverse line scan images and has previously referred to as a U-shaped CaT (18) (Fig. 7A). Previous studies have demonstrated that carvedilol suppressed spontaneous Ca²⁺ waves (43, 45). Generally speaking, Ca²⁺ waves propagate by an analogous mechanism to the one that is in operation during physiological atrial ECC. Therefore, we explored the effect of carvedilol (1 µM) on the spatiotemporal attributes of AP-induced Ca²⁺ release in isolated atrial cells. Figure 7A shows SS and CT [Ca²⁺]_i profiles (F/F₀ spatially averaged over 2 µm) in the presence and absence of carvedilol. Under control conditions (Fig. 7B), SS CaT amplitude was 67% larger than CT CaTs (P = 0.04). Carvedilol (1 µM) reduced the SS CaT amplitude to the level of the CT CaT (∆F/F₀, Ctrl 3.5 ± 0.6 vs. treated 2.3 ± 0.4; n = 14; P = 0.03) but did not significantly affect CT Ca²⁺ release (∆F/F₀, Ctrl 2.1 ± 0.8 vs. treated 1.7 ± 0.5; n = 14; P = 0.15). Thus, even though 1 µM carvedilol inhibited SS Ca²⁺ release from the j-SR, resulting in a decreased amplitude of the averaged whole cell CaT (Fig. 1C), the local SS elevation of [Ca²⁺]_i was sufficient to initiate robust centripetal propagation of CICR that remained largely unaffected by the drug, although there was a tendency toward at reduced CT CaT amplitude. Furthermore, carvedilol did not affect the delay between SS and CT Ca²⁺ release (Δt between SS half-maximum and CT half-maximum CaT amplitude, control 23.7 ± 3.1 ms; carvedilol 20.9 ± 3.1 ms; Fig. 7C), indicating that propagation velocity of activation was not affected by carvedilol (n = 14; P = 0.12).

Fig. 7.

Carvedilol effect on subcellular intracellular Ca²⁺ concentration ([Ca²⁺]_i) during excitation-contraction coupling. Carvedilol (1 µM) reduces subsarcolemmal (SS) junctional sarcoplasmic reticulum (SR), but not central (CT) nonjunctional-SR Ca²⁺ transient amplitude. A: line scan images and corresponding [Ca²⁺]_i profiles from peripheral (SS, black trace) and central (CT, gray trace) Ca²⁺ release sites in control (a) and after carvedilol treatment (b). Subcellular [Ca²⁺]_i profiles were spatially averaged over 2 µm. B: average SS and CT Ca²⁺-transient amplitudes from control and carvedilol-treated atrial myocytes. *P < 0.04 (ANOVA). C: average delay between peripheral and central Ca²⁺ release in control (Ctrl) and carvedilol (1 µM)-treated atrial myocytes (paired t-test). Means ± SE and individual cell data are shown (n = 14). NS, not significant.

Since carvedilol is known to reduce the mean open time of the RyR Ca²⁺ release channel (43, 45), we explored the effect of the drug on spontaneous Ca²⁺ release events in the form of spontaneous CaTs and Ca²⁺ waves. To increase the propensity of such events, atrial myocytes were electrically paced in elevated extracellular Ca²⁺ (7 mM). Cells were stimulated at 1 Hz for 30 s, pacing was then stopped, and the occurrence of spontaneous activity was recorded during a 20-s interval of rest, using line-scanning confocal imaging. Representative confocal images in control and after carvedilol treatment together with their respective whole cell [Ca²⁺]_i profiles are depicted in Fig. 8A. Carvedilol (1 µM) significantly decreased the frequency of spontaneous Ca²⁺ waves from 0.19 ± 0.05 in control to 0.05 ± 0.01 Hz in the presence of carvedilol (n = 10; P = 0.002; Fig. 8B). However, Ca²⁺ wave latency (time between the last electrically evoked CaT and the first Ca²⁺ wave; Fig. 8C) was not statistically different (Ctrl, 5.5 ± 2.0 s; carvedilol, 4.8 ± 1.6 s; n = 10; P = 0.8). There was a tendency toward slowing of Ca²⁺ wave propagation velocity in the presence of carvedilol (Fig. 8D); however, the difference was not statistically significant (Ctrl, 82.4 ± 11.5 µm/s; carvedilol 69.6 ± 5.3 µm/s; n = 10; P = 0.26).

Fig. 8.

Carvedilol (1 µM) suppresses spontaneous Ca²⁺ waves. Ca²⁺ wave frequency was quantified during a period of rest (20 s) after pacing (1 Hz) in elevated extracellular Ca²⁺ concentration ([Ca²⁺]_o; 7 mM). A; whole cell [Ca²⁺]_i profiles and confocal line scan images in control (top) and carvedilol-treated (bottom) atrial myocytes (n = 10). Black circles indicate stimulus during electrical pacing. B: average carvedilol effect on Ca²⁺ wave frequency. *P < 0.004 (paired t-test). C: average wave latency [Δt between last electrically elicited Ca²⁺-transient (CaT) and first Ca²⁺ wave] in control (Ctrl) and carvedilol-treated atrial myocytes (unpaired t-test). D: average Ca²⁺ wave velocity in control and carvedilol-treated atrial myocytes (unpaired t-test). E: carvedilol effect on sarcoplasmic reticulum (SR) Ca²⁺ load. SR load was quantified as the amplitude of a CaT elicited by rapid superfusion with caffeine (10 mM) in control (top) and in the presence of carvedilol (1 µM; bottom). Whole cell intracellular Ca²⁺ concentration ([Ca²⁺]_i) profiles and corresponding line scan images are shown. F: summary of CaT amplitudes in response to 10 mM caffeine (n = 11; paired t-test). NS, not significant.

Furthermore, we tested for the possibility that the decreased Ca²⁺ wave frequency resulted from a reduced SR Ca²⁺ load. For this purpose, atrial cells were rapidly exposed to caffeine (10 mM; Fig. 8E), and the amplitude of the caffeine-induced CaT served as a measure of SR Ca²⁺ load (Fig. 8F). Carvedilol did not significantly affect releasable SR Ca²⁺ content (∆F/F₀, Ctrl 4.1 ± 0.6; caffeine 3.8 ± 0.7; n = 11; P = 0.15). The significant reduction of the propensity of spontaneous Ca²⁺ waves by carvedilol in the absence of a difference in SR Ca²⁺ load suggests a direct action of carvedilol on the RyR Ca²⁺ release channel and demonstrates an antiarrhythmic effect of carvedilol in atrial myocytes.

DISCUSSION

In this study we investigated the effect of the clinically widely used β-blocker carvedilol on SR Ca²⁺ release and ECC in rabbit atrial myocytes. The main findings are as follows: 1) carvedilol causes failure of SR Ca²⁺ release and ECC in a dose-dependent manner; 2) the primary cause of ECC failure by carvedilol is a dose-dependent inhibition of voltage-gated I_Na and I_Ca; 3) at low carvedilol concentration (1 µM), the decrease of the whole cell CaT amplitude by over 55% results from a specific inhibition of subsarcolemmal CICR from j-SR presumably due to I_Na and I_Ca inhibition and AP shortening, while propagating CICR from nj-SR remains largely unaffected; and 4) the frequency of spontaneous arrhythmogenic Ca²⁺ waves is significantly reduced despite identical SR Ca²⁺ load, suggesting a direct effect of carvedilol on RyR function.

Carvedilol is a nonspecific β-blocker clinically used for the treatment of cardiovascular diseases, including ventricular tachyarrhythmia and atrial fibrillation (16, 19, 26, 32), has vasodilatory effects, and has proven to reduce hospitalization and mortality of patients with congestive heart failure (29). Most clinical benefits stem from the β-blocker drug action (15); however, carvedilol is not entirely specific to β-adrenoceptors. It also has α-adrenoreceptor blocking effects (31), can act as an antioxidant (14, 42), and has antiproliferative properties in vascular smooth muscle (28).

At the cellular level we observed a dose-dependent failure of AP-induced SR Ca²⁺ release in atrial myocytes (Fig. 1), reminiscent of a previously reported reversible loss of responsiveness to field stimulation in ventricular myocytes (40). We demonstrate that carvedilol-induced ECC failure is the result of inhibition of I_Na and I_Ca (Figs. 3, 4 and 5), resulting in AP failure (Fig. 2, A and B) and shortening (Fig. 2C). An inhibitory effect of carvedilol on several membrane currents in cardiac cells (ventricle and pacemaker, but not in the atrial myocardium) has been described before. Inhibition of voltage-gated I_Ca (11, 12, 27, 40, 41), I_Na (1, 2), K⁺ currents (I_K) (11, 22), pacemaker current I_f (41), and NCX1 current (I_NCX1) (37) have been reported. For example, carvedilol was found to inhibit LTCCs in vascular smooth muscle (27) and ventricular cells (11) by 64 and 43%, respectively, compared with our observation that 10 µM carvedilol inhibited I_Ca by 21%. Carvedilol (1 μM) disrupted ECC in 14% of cells (Fig. 1B) and reduced the whole cell CaT amplitude by 55% (Fig. 1C) but apparently failed to significantly inhibit I_Ca (Fig. 5F) in voltage-clamp experiments when the cell membrane was depolarized with rectangular depolarization pulses. However, AP recordings revealed that at 1 µM carvedilol, APD was already shortened in the voltage range where I_Ca is maximally activated. In other words, under physiological conditions when ion channel activation is dictated by the AP, the data are consistent with an I_Ca inhibition already at 1 µM. In other studies, 1 µM carvedilol blocked I_Ca in ventricular myocytes by ∼10%, but by ~50% with 10 µM carvedilol (40), whereas in pacemaker cells (sinoatrial and atrioventricular node cells), the degree of inhibition of I_Ca was higher (45–50%) and the spontaneous firing frequency was reduced (41). Carvedilol (10 μM) had no effect on I_Ca voltage dependence in atrial myocytes (Fig. 5), similar to the results obtained in vascular smooth cells (27), but different from pacemaker cells, where carvedilol shifted the voltage dependence by 10 mV toward more positive values (41). A dose-dependent I_Na inhibition by carvedilol has been reported using a heterologous expression system (HEK-293 cells) for Na_V1.5 and Na_V1.6 channels (1, 2) and in cultured bovine adrenal cells (21). It has been reported that half maximal inhibitory concentration for I_Na and I_Ca inhibition by carvedilol is rather similar [3.59 and 3.83 µM, respectively (1, 11, 27)]. In our hands, however, inhibition of I_Na for any given concentration tested was much more pronounced than the effect on I_Ca. Carvedilol plasma levels for therapeutic doses of the drug have been reported to be in the range of 0.1 to 0.6 µM (23, 38, 41), somewhat lower than the concentration (1 µM) where we started to see an inhibitory action on ECC. Specifically, at 0.5 µM we did not see an acute effect after 3 min of drug exposure; however, we found ECC disruption after 15 min of treatment (data not shown). The reason for this discrepancy between in vitro and in vivo data is unclear. The profound current inhibition in vitro, especially of I_Na at the 1 µM concentration, might put doubts on the clinical usefulness of this drug and appears at odds with the fact that it is widely used clinically and is considered rather safe. Carvedilol is highly lipophilic, and the drug concentration at the target might differ from plasma levels. Furthermore, the I_Na inhibitory action could be advantageous for the use of carvedilol as an antiarrhythmic drug since Na⁺ channel blockers are clinically used for this purpose.

Since carvedilol has well-documented diverse effects, we tested whether the observed inhibition of I_Na and I_Ca was due to the β-blocking effect of the drug. For that purpose, we tested the specific β-blocker metoprolol. In contrast to carvedilol, metoprolol (10 μM) had no effect on CaT amplitude (Fig. 1) and on I_Na and I_Ca (Fig. 6). This result implies that carvedilol inhibition of I_Na and I_Ca is not due to the β-blocker action, consistent with previous studies on I_Na (3, 5, 39) and I_Ca where significantly higher (30 µM) drug concentrations were required to elicit an inhibitory effect (3, 33). These data suggest that carvedilol acts directly on the ion channels involved in ECC, either the relevant voltage-gated surface membrane channels or the SR Ca²⁺ release channel (RyR).

Previous studies have tested the effect of β-blockers on RyR function. Zhou et al. (45) tested over a dozen β-blockers for effects on RyR-mediated Ca²⁺ release. With the exception of carvedilol, none of the blockers tested had any effect on spontaneous Ca²⁺ release events (Ca²⁺ waves). Carvedilol, however, was capable of significantly suppressing spontaneous SR Ca²⁺ release. The suppression of Ca²⁺ release could be pinpointed to a significant reduction of the mean open time of the RyR Ca²⁺ release channel. The reduction in mean open time as the key mechanism underlying the action of carvedilol was further confirmed by the observation that the non-β-blocking R-enantiomer of carvedilol (43) and the compound VK-II-86 (45), which also reduces RyR mean open time, both suppressed spontaneous Ca²⁺ waves. Presumably through its ability to curtail RyR activity, carvedilol can facilitate Ca²⁺ alternans in hearts harboring a suppression-of-function RyR mutation (44). Furthermore, when compared with other β-blockers (e.g., metoprolol), carvedilol had a higher antiarrhythmic effect due to the combination of β-blocking and RyR-modulating actions (45). In addition, the I_Na inhibition observed here adds another facet of antiarrhythmic action of carvedilol since Na⁺ channel blockers are therapeutically used for treatment of supraventricular arrhythmias, including atrial fibrillation.

A key finding in our study was the observation that in adult atrial myocytes, carvedilol also suppressed spontaneous Ca²⁺ waves (Fig. 8) by decreasing their frequency by ~75%. SR Ca²⁺ overload (17) accompanies certain pathological conditions and ultimately generates Ca²⁺ waves (30). Such spontaneous Ca²⁺ waves can become arrhythmogenic because they induce a depolarizing membrane current (via Na⁺/Ca²⁺ exchange) that can result in arrhythmia and contractile dysfunction (4, 10). As shown here, carvedilol has an antiarrhythmic effect in atrial cells by significantly reducing the probability of Ca²⁺ waves. Remarkably, the reduction in Ca²⁺ wave frequency could not be explained by a decreased SR Ca²⁺ load, since SR Ca²⁺ content was the same under both conditions (Fig. 8F), suggesting again that the suppression of spontaneous SR Ca²⁺ release is a consequence of a direct action of carvedilol on the RyR channel.

As detailed in introduction, the mechanism of atrial Ca²⁺ release and ECC differs significantly from ventricular cells due to the absence or paucity of the TT system. Once activated by an AP, CICR propagates through the nj-SR network independent of I_Ca (18, 34), only sustained by the Ca²⁺ gradients established by intracellular SR Ca²⁺ release. Thus, the activation of the majority of Ca²⁺ release sites in atrial cells reside in the membrane of the nj-SR and are activated in a propagating Ca²⁺ wave fashion. We observed that carvedilol reduced the whole cell CaT amplitude in a dose-dependent manner (Fig. 1). Confocal Ca²⁺ imaging studies revealed (Fig. 7) that carvedilol primarily reduced the peripheral CaT, i.e., release from j-SR. This is consistent with the notion that the trigger for CICR from j-SR release site is Ca²⁺ influx through LTCCs, which is reduced by carvedilol. We propose, however, that the elevation of [Ca²⁺]_i in the subsarcolemmal space in the presence of 1 µM carvedilol is still high enough to induce CICR from a nearby nj-SR release site [note that the diffusional distance between j-SR and the most peripheral array of nj-SR release site is only 1 to 2 µm (6)] from where CICR continues to propagate in centripetal direction. Overall, we found a slight reduction in central CaT amplitude (Fig. 7B), but no impediment of CICR propagation velocity (Fig. 7C). In an earlier study we showed that in atrial myocytes, I_Ca and j-SR CaT amplitudes have the well-known, bell-shaped membrane potential dependence (34). In contrast, the relationship between membrane potential and CT nj-SR CaT amplitude is relatively flat, i.e., the atrial ECC machinery is highly reliable in that there is a substantial degree of safety inherent to the system which allows a wide range of variability of peripheral j-SR Ca²⁺ release without jeopardizing nj-SR Ca²⁺ release and CICR propagation. The observation that Ca²⁺ wave propagation velocity is not affected by carvedilol once waves have been triggered is consistent with this hypothesis. The inhibitory carvedilol effect on the RyR, however, renders luminal RyR activation by luminal Ca²⁺ less effective and explains the reduction of spontaneous Ca²⁺ wave frequency.

In summary, as shown here for the first time in atrial cells, carvedilol inhibited I_Na and I_Ca in a concentration-dependent manner, resulting in CaTs of reduced amplitude and ultimately failure of APs and SR Ca²⁺ release during ECC. The effect of carvedilol on electrically evoked Ca²⁺ signals was subcellularly heterogenous where the drug preferentially affected peripheral Ca²⁺ realese from the j-SR, whereas propagating SR Ca²⁺ release from nj-SR, once initiated, was not further affected. Moreover, carvedilol had a prominent inhibitory effect on arrhythmogenic Ca²⁺ waves, an effect that was not associated with differences in SR Ca²⁺ load and is consistent with the known effect of reducing the mean open time of the RyR. These cellular findings are in line with the clinical finding of carvedilol's efficiency of suppressing cardiac arrhythmias, including atrial fibrillation.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants HL-057832, HL-132871, and HL-134781.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

E.M.-H. and L.A.B. conceived and designed research; E.M.-H. performed experiments; E.M.-H. and L.A.B. analyzed data; E.M.-H. and L.A.B. interpreted results of experiments; E.M.-H. and L.A.B. prepared figures; E.M.-H. and L.A.B. drafted manuscript; E.M.-H. and L.A.B. edited and revised manuscript; E.M.-H. and L.A.B. approved final version of manuscript.