Inhibition of cardiac potassium currents by oxidation-activated protein kinase A contributes to early afterdepolarizations in the heart

Abstract

Reactive oxygen species (ROS) have been shown to prolong cardiac action potential duration resulting in afterdepolarizations, the cellular basis of triggered arrhythmias. As previously shown, protein kinase A type I (PKA I) is readily activated by oxidation of its regulatory subunits. However, the relevance of this mechanism of activation for cardiac pathophysiology is still elusive. In this study, we investigated the effects of oxidation-activated PKA I on cardiac electrophysiology. Ventricular cardiomyocytes were isolated from redox-dead PKA-RI Cys17Ser knock-in (KI) and wild-type (WT) mice and exposed to H₂O₂ (200 µmol/L) or vehicle (Veh) solution. In WT myocytes, exposure to H₂O₂ significantly increased oxidation of the regulatory subunit I (RI) and thus its dimerization (threefold increase in PKA RI dimer). Whole cell current clamp and voltage clamp were used to measure cardiac action potentials (APs), transient outward potassium current (I_to) and inward rectifying potassium current (I_K1), respectively. In WT myocytes, H₂O₂ exposure significantly prolonged AP duration due to significantly decreased I_to and I_K1 resulting in frequent early afterdepolarizations (EADs). Preincubation with the PKA-specific inhibitor Rp-8-Br-cAMPS (10 µmol/L) completely abolished the H₂O₂-dependent decrease in I_to and I_K1 in WT myocytes. Intriguingly, H₂O₂ exposure did not prolong AP duration, nor did it decrease I_to, and only slightly enhanced EAD frequency in KI myocytes. Treatment of WT and KI cardiomyocytes with the late I_Na inhibitor TTX (1 µmol/L) completely abolished EAD formation. Our results suggest that redox-activated PKA may be important for H₂O₂-dependent arrhythmias and could be important for the development of specific antiarrhythmic drugs.

NEW & NOTEWORTHY Oxidation-activated PKA type I inhibits transient outward potassium current (I_to) and inward rectifying potassium current (I_K1) and contributes to ROS-induced APD prolongation as well as generation of early afterdepolarizations in murine ventricular cardiomyocytes.

INTRODUCTION

Cardiac arrhythmias are a major contributor to cardiovascular morbidity and mortality. At the cellular level, triggered activity, which is a main driver of ventricular tachycardia (VT) (46, 47), results from early and delayed afterdepolarizations (EADs and DADs) (64). In heart failure (HF), it was shown that triggered activity may be a consequence of reduced repolarization reserve with significant prolongation of action potential duration partly due to reduced transient outward potassium current (I_to) and inward rectifying potassium current (I_K1) (34). The mechanisms are incompletely understood but may involve reactive oxygen species (ROS)-dependent signaling (58, 59). Increased levels of ROS are frequently found in animal models (6, 10, 24) and human HF (4, 28, 36, 37) and have been shown to result in inhibition of I_K1 (9, 32, 33, 56, 57) and I_to (11, 38, 45). ROS-dependent signaling may not only involve direct posttranslational modification (i.e., oxidation) of ion channels and transporters (12, 15) but may also be mediated by oxidative activation of second messengers (14, 59, 60).

Interestingly, recent evidence delineated a novel mechanism of cAMP-dependent protein kinase (PKA) activation by oxidation of cysteine 17 and 38 at the regulatory RIα subunit. This results in intermolecular disulfide bond formation (between cysteine 17 and 38) of two adjacent RI subunits consistent with dimerization (5), which disinhibits the catalytic subunits leading to increased PKA-dependent phosphorylation. This oxidative PKA activation has been shown to be important for angiogenesis (5) and platelet-derived growth factor (PDGF)-dependent signaling (13). However, its relevance for cardiac (patho-)physiology has not been characterized. Since PKA has been shown to regulate I_to and I_K1, we tested the hypothesis that redox-dependent activation of PKA I may be important for arrhythmogenesis by inhibition of repolarizing potassium currents under conditions of increased oxidative stress.

MATERIALS AND METHODS

Study approval.

Experiments with murine cardiomyocytes conformed to Directive 2010/63/EU of the European Parliament, to the Guide for the Care and Use of Laboratory Animals (8^th ed., 2011), and to local institutional guidelines.

Generation of Cys17Ser PKA RI mice.

Mice that constitutively express PKA RI Cys17Ser (knock-in or KI) were generated as previously described (5). The point mutation Cys17Ser, which renders PKA type I redox-dead, was introduced into exon 1 of the Prkar1a gene by site-directed mutagenesis. These mice behave normally under physiological conditions.

Isolation of murine cardiomyocytes.

Homozygous PKA RI Cys17Ser (KI) and wild-type (WT) littermates (male, aged 10–18 wk, BL/6J background, housed according to German animal laws, kindly provided by Prof. Philip Eaton) were anesthetized with isoflurane. After death by cervical dislocation, hearts were quickly excised, mounted on a Langendorff perfusion apparatus, and retrogradely perfused with nominally Ca-free solution containing (in mmol/L) 113 NaCl, 4.7 KCl, 0.6 KH₂PO₄, 0.6 Na₂HPO₄, 1.2 MgSO₄, 12 NaHCO₃, 10 KHCO₃, 10 HEPES, 30 taurine, 10 2,3-butanedione monoxime (BDM), 5.5 glucose, and 0.032 phenol red for 4 min at 37°C (pH 7.4). Then, 7.5 mg/mL liberase (Roche Diagnostics, Mannheim, Germany), trypsin 0.6%, and 0.125 mmol/L CaCl₂ were added to the perfusion solution. Perfusion was continued for 3–4 min until the heart became flaccid. Ventricular tissue was collected in perfusion buffer supplemented with 5% bovine calf serum, cut into small pieces, and dispersed by repeatedly pipetting until no solid cardiac tissue was left. Ca reintroduction was performed by stepwise increasing [Ca] from 0.1 to 0.8 mmol/L (43).

Patch-clamp experiments.

A ruptured-patch whole cell voltage clamp was used to measure I_to and I_K1. For these experiments, microelectrodes (2–3 MΩ) were filled with a solution containing (in mmol/L) 40 K-aspartate, 90 KCl, 1 MgCl₂, 1.8 CaCl₂, 5 EGTA, 5 Mg-ATP, 0.3 Li-GTP, 0.1 niflumic acid, and 10 HEPES. The pH was adjusted to 7.2 using KOH. Bath solution contained (in mmol/L) 135 tetramethylammonium chloride as the substitute for Na, 4 KCl, 1 MgCl₂, 10 glucose, 10 HEPES, and 0.3 CdCl₂ to block calcium channels (pH 7.4 with KOH). After patch rupture, the access resistance was typically <7 MΩ. Currents were elicited using a voltage step protocol (−130 mV to +90 mV, steps of 10 mV, holding potential of −90 mV, and duration of 300 ms for each step). Following each recording, cells were superfused with bath solution containing BaCl₂ (200 µmol/L) to block I_K1. The resulting trace was then digitally subtracted from the original trace to determine the Ba-sensitive current, which is hence referred to as I_K1. I_to was defined as the difference between the peak outward current and the steady-state current. Time constants of I_to inactivation (τ_slow and τ_fast) were calculated from a triple exponential fit

$$f\left(t\right)={\displaystyle \sum }_{i=1}^n{A}_i{e}^{-t/{\tau }_i}+C,$$

where f(t) is the current at time t, n = 3, A is the respective current amplitude, τ is the time constant of inactivation, and C is the steady-state current.

For action potential measurements, the pipette solution contained (in mmol/L) 122 K-aspartic acid, 10 NaCl, 8 KCl, 1 MgCl₂, 5 Mg-ATP, 0.3 Li-GTP, and 10 HEPES (pH 7.2 with KOH). The bath solution contained (in mmol/L) 140 NaCl, 4 KCl, 1 MgCl₂, 10 glucose, 1 CaCl₂, and 5 HEPES (pH 7.4 with NaOH). APs were elicited using square current pulses of 1 nA and a duration of 1–4 ms at a frequency of 1 Hz. Cells with access resistance >15 MΩ, spontaneous activity under basal conditions, and need for injection of negative currents with amplitudes more negative than −40 pA were excluded from analysis to prevent factors like bad cell access, inclusion of damaged cells, and cells with increased leak currents to bias our results. Only cells with excellent membrane potential stability and adequate overshoot upon each electrical stimulus were included in the analysis.

To inhibit PKA and Ca/calmodulin-dependent protein kinase II (CaMKII), cells were preincubated in bath solutions containing Rp-8-Br-cAMPS (10 µM; Biolog, Bremen, Germany) and myristoylated autocamtide-2-related inhibitory peptide (AIP, 1 µM; AnaSpec, Fremont, California) for 10 min, respectively. For inhibition of late I_Na, tetrodotoxin (TTX, 1 µM, Bristol, UK) was added to the superfusate. Oxidative stress was induced by preincubation with H₂O₂-containing bath solution (200 µM) for at least 4 min for current measurements or by continuous superfusion for AP recordings in a light-protected environment. The time points of patch-clamp registrations were selected based on our previous experience. We have previously shown that Ca spark frequency or late Na current was increased as early as 4 min after onset of H₂O₂ exposure in a CaMKII-dependent fashion (60). Moreover, EADs and DADs also occurred as early as 4–8 min after onset of H₂O₂ exposure (60). All experiments were conducted at room temperature (22°C–24°C).

Western blot analysis.

Murine cardiomyocytes were isolated as described earlier (Isolation of murine cardiomyocytes). Cells were then harvested and lysed in Tris buffer containing (in mmol/L) 20 Tris·HCl, 200 NaCl, 20 NaF, 1 Na₃VO₄, 1% Triton X-100, 1 DTT (pH 7.4) and a complete protease inhibitor cocktail and a PhosSTOP phosphatase inhibitor cocktail (both Roche Diagnostics) by trituration. Protein concentration was determined by bicinchoninic acid assay (Pierce Biotechnology). Denatured cell lysates (30 min, 37°C in 2% β-mercaptoethanol) were subjected to Western blot analysis (8% SDS polyacrylamide gels) using primary antibodies against Kir2.1 (polyclonal rabbit, 1:1,000, Alomone Laboratories) and GAPDH (monoclonal mouse, 1:20,000, BIOTREND). Primary antibodies were incubated at 4°C overnight. Secondary antibodies were horseradish peroxidase (HRP)-conjugated sheep anti-mouse IgG (1:10,000, Amersham Biosciences) for GAPDH and HRP-conjugated donkey anti-rabbit IgG (1:10,000, Amersham Biosciences) for Kir2.1 (incubation for 1 h at room temperature). Incubation with Immobilon Western Chemiluminescent HRP Substrate (Millipore) for 5 min at room temperature enabled detection of the protein bands, which were developed onto Super XR-N X-ray films (Fujifilm) and scanned using ChemiDoc MP Imaging System (Bio-Rad). ImageJ was used to analyze mean densitometric values.

PKA RI dimerization assay.

PKA RI dimerization was assessed from isolated cardiomyocytes treated with H₂O₂ (100 µmol/L) for 10 min by Western blot analysis following protein separation by SDS-PAGE, with addition of maleimide (100 mmol/L) to the lysis buffer to prevent disulfide exchange by thiol alkylation. Immunoblots were probed with primary antibodies to PKARI (monoclonal mouse, 1:500, BD Transduction Laboratories). Secondary antibody was horseradish peroxidase (HRP)-conjugated sheep anti-mouse IgG (1:10,000, Amersham Biosciences). PKA oxidation status was determined by the ratio of its dimeric (oxidized) (120 kD) RI subunit to GAPDH.

Data analysis and statistics.

All data are presented as means ± SE. Data were analyzed using one-way or two-way ANOVA with post hoc test (Holm–Sidak) or Student’s t test for unpaired data and mixed-effects analysis with post hoc test (Holm–Sidak or Tukey) for paired data as appropriate. P values <0.05 were considered statistically significant.

RESULTS

Dimerization of PKA regulatory subunit I by oxidative stress.

To test if exposure of isolated murine ventricular cardiomyocytes to H₂O₂ (100 µmol/L) results in oxidation of PKA RI subunits, we applied a PKA RI dimerization assay using Western blot analysis under nonreducing conditions. In WT, the PKA RI dimer and monomer were almost equally abundant in vehicle-treated cells. However, H₂O₂ exposure resulted in a significant increase in PKA RI dimer at the expense of RI monomer consistent with RI oxidation and disulfide bridge formation. The dimer/GAPDH ratio increased by approximately threefold from 0.7611 ± 0.1012 to 2.228 ± 0.1464 (N = 7 vs. 7 mice, P < 0.0001, Fig. 1, A and B). Importantly, in KI cardiomyocytes, formation of relevant amounts of RI dimers could not be observed under basal conditions or after addition of H₂O₂ (0.0017 ± 0.0012 vs. 0.0047 ± 0.0019, N = 2 vs. 2 mice, P = 0.99) (Fig. 1, A and B).

Fig. 1.

PKA RI dimerization assay in murine hearts. A: original Western blot for the assessment of dimerization of the PKA regulatory subunit RI upon exposure to H₂O₂ (100 µmol/L). B: analysis of mean densitometric values reveals a threefold increase in PKA RI dimer/GAPDH ratio in WT myocytes exposed to H₂O₂ (n= 5 vs. 5; *P = 0.0003 vs. control, unpaired two-tailed t test; values are means ± SE). In contrast, KI myocytes were devoid of any H₂O₂-dependent RI dimerization (n = 2 vs. 2). PKA, protein kinase; RI, regulatory subunit; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; WT, wild-type; n.s., not significant (P > 0.05).

H₂O₂ prolongs action potential duration by oxidative PKA activation.

Action potentials (APs) were measured by whole cell voltage clamp. H₂O₂ exposure resulted in a marked increase in AP duration in WT cardiomyocytes. Compared with the baseline, 2 min after onset of H₂O₂ exposure, the AP duration at 90% repolarization (APD₉₀) increased from 53.42 ± 4.44 ms to 107.6 ± 19.59 ms [N (n) = 9 (16) vs. 8 (11); P = 0.0006, Fig. 2, A and C]. In contrast, H₂O₂ did not increase APD₉₀ in myocytes lacking oxidative PKA activation. In KI myocytes, APD₉₀ was 46.81 ± 4.71 versus 58.13 ± 9.07 ms [N (n) = 6 (17) vs. 5 (12); P = 0.6686; Fig. 2, A and C], suggesting that redox-activated PKA may be important for regulation of the repolarization reserve. Importantly, in WT myocytes, APD₂₀ significantly increased from 6.79 ± 0.33 ms to 8.12 ± 0.74 ms [WT Veh vs. WT H₂O₂, N (n) = 9 (16) vs. 8 (9), Holm–Sidak’s post hoc test for WT Veh vs. WT H₂O₂ P = 0.0407] and APD₅₀ significantly increased from 10.45 ± 0.63 ms to 12.96 ± 1.04 ms [WT Veh vs. WT H₂O₂, N (n) = 9 (16) vs. 8 (9), Holm–Sidak’s post hoc test for WT Veh vs. WT H₂O₂ P = 0.0380] at 8 min of H₂O₂ superfusion, where significant differences in EAD frequency are observed (Fig. 3, A and B). Contrary to this, H₂O₂ had no effect on these parameters in KI myocytes [APD₂₀: 6.36 ± 0.30 vs. 6.46 ± 0.30, KI Veh vs. KI H₂O₂, N (n) = 6 (16) vs. 5 (8), Holm–Sidak’s post hoc test for KI Veh vs. KI H₂O₂ P = 0.7764; APD₅₀: 8.76 ± 0.65 vs. 8.70 ± 0.59, N (n) = 6 (16) vs. 5 (8), Holm–Sidak’s post hoc test for KI Veh vs. KI H₂O₂ P = 0.8586] (Fig. 3, A and B).

Fig. 2.

Effects of H₂O₂ on cardiac action potentials in WT and KI cardiomyocytes. A and B: superimposed original traces of APs in WT (A) and KI (B) cardiomyocytes under control conditions (Veh) and upon exposure to H₂O₂ (200 µmol/L). C: action potential duration at 90% of repolarization (APD₉₀) is significantly prolonged upon exposure to H₂O₂ for 2 min in WT cardiomyocytes compared with control conditions and to H₂O₂-treated KI cardiomyocytes (ANOVA P for factor H₂O₂ = 0.0005, ANOVA P for factor genotype = 0.0131, Tukey’s post hoc test for WT Veh vs. WT H₂O₂ P = 0.0006 (*), WT H₂O₂ vs. KI H₂O₂ P = 0.0108 (#), and for KI Veh vs. KI H₂O₂ P = 0.6686). D: resting membrane potential remains unaltered upon H₂O₂ treatment (ANOVA P for factor H₂O₂ = 0.3637, ANOVA P for factor genotype = 0.8599, Tukey’s post hoc test for WT Veh vs. WT H₂O₂ P = 0.8586 and for KI Veh vs. KI H₂O₂ P = 0.9575). E: action potential amplitude (APA) is increased in WT and KI cardiomyocytes upon oxidative stress (ANOVA P for factor H₂O₂ = 0.0018, ANOVA P for factor genotype = 0.4612, Tukey’s post hoc test for WT Veh vs. WT H₂O₂ P = 0.0124 and for KI Veh vs. KI H₂O₂ P = 0.3639). F: H₂O₂ treatment results in an increase in AP upstroke velocity in both WT and KI cardiomyocytes (ANOVA P for factor H₂O₂ = 0.0042, ANOVA P for factor genotype = 0.9645, Tukey’s post hoc test for WT Veh vs. WT H₂O₂ P = 0.0316 and for KI Veh vs. KI H₂O₂ P = 0.4063). All values are means ± SE. The numbers of mice/cells were: WT (Veh), 9/16; WT (H₂O₂), 8/11; KI (Veh), 6/17; and KI (H₂O₂), 5/12. Mixed-effects analysis with Tukey’s post hoc correction. WT, wild-type; KI, knock-in; Veh, vehicle; AP, action potential.

Fig. 3.

Effects of H₂O₂ on APD₂₀ and APD₅₀ in WT and KI cardiomyocytes. In contrast to KI cardiomyocytes, APD₂₀ (A) as well as APD₅₀ (B) are significantly increased in WT cardiomyocytes upon exposure to H₂O₂ (200 µmol/L) for 8 min (APD₂₀: ANOVA P for factor H₂O₂ = 0.0644, ANOVA P for factor genotype = 0.0377, Holm–Sidak’s post hoc test for WT Veh vs. WT H₂O₂ P = 0.0407 (*) and for KI Veh vs. KI H₂O₂ P = 0.7764; APD₅₀: ANOVA P for factor H₂O₂ = 0.0711, ANOVA P for factor genotype = 0.0024, Holm–Sidak’s post hoc test for WT Veh vs. WT H₂O₂ P = 0.0380 (*) and for KI Veh vs. KI H₂O₂ P = 0.8586. Values are means ± SE. The numbers of mice/cells were: WT (Veh), 9/16; WT (H₂O₂), 8/9; KI (Veh), 6/16; and KI (H₂O₂), 5/8. APD, action potential duration; WT, wild-type; KI, knock-in; Veh, vehicle.

Interestingly, H₂O₂ treatment significantly increased the action potential amplitude (APA) and also the upstroke velocity (V_max) in both WT and KI myocytes (ANOVA P for factor H₂O₂ 0.0018 and 0.0042, respectively), with no difference between the genotypes (ANOVA P for factor genotype 0.4612 and 0.9645, respectively). This suggests that redox activation of PKA may not be required for H₂O₂-dependent stimulation of depolarizing ion currents like voltage-gated Na current. In WT in the presence of H₂O₂, APA increased from 107.6 ± 2.52 mV to 118.6 ± 2.03 mV [WT Veh vs. WT H₂O₂, N (n) = 9 (16) vs. 8 (11); Tukey’s post hoc test P = 0.0124, Fig. 2E] and V_max increased from 85.469 ± 4.503 V/s to 102.564 ± 3.073 V/s [for WT Veh vs. WT H₂O₂, N (n) = 9 (16) vs. 8 (11); Tukey’s post hoc test P = 0.0316, Fig. 2F]. For KI Veh versus KI H₂O₂, APA was 108.1 ± 2.65 mV versus 114.2 ± 3.28 mV [N (n) = 6 (17) vs. 5 (12); Tukey’s post hoc test P = 0.364 and P = 0.561 for KI H₂O₂ vs. WT Veh] and V_max was 88.947 ± 4.568 V/s versus 98.983 ± 5.340 V/s [N (n) = 6 (17) vs. 5 (12); Tukey’s post hoc test P = 0.406 and P = 0.323 for KI H₂O₂ vs. WT Veh, Fig. 2, E and F].

In contrast to AP duration, there were no differences in the resting membrane potential after exposure to H₂O₂ both in WT and KI myocytes [ANOVA P for factor H₂O₂ = 0.3637; ANOVA P for factor genotype = 0.8599, WT Veh vs. WT H₂O₂: −69.84 ± 1.71 mV vs. −71.93 ± 1.92 mV, N (n) = 9 (17) vs. 8 (12), Tukey’s post hoc test P = 0.8586; KI Veh vs. KI H₂O₂: −70.54 ± 1.84 mV vs. −71.88 ± 1.87 mV, N (n) = 6 (17) vs. 5 (12), Tukey’s post hoc test P = 0.9575] (Fig. 2, A and D).

H₂O₂ enhanced early afterdepolarizations by oxidative activation of PKA.

Early afterdepolarizations (EADs) were measured using a provocation protocol that was initiated at different time points after the start of H₂O₂ superfusion. In detail, the protocol consisted of three trains of 30 consecutive stimuli at a frequency of 1 Hz, with a 10-s rest interval in between trains 1 and 2, and a 30-s rest interval in between trains 2 and 3. EADs were defined as a positive voltage deflection of ≥5 mV following the initial repolarization (phase 1 of the action potential) of a stimulated AP (Fig. 4,A and B).

Fig. 4.

Early afterdepolarizations in WT and KI cardiomyocytes upon exposure to H₂O₂. A: original traces of action potentials elicited at a frequency of 1 Hz before and 8 min after onset of H₂O₂ (200 µmol/L) superfusion. In contrast to KI cardiomyocytes, H₂O₂ exposure induces frequent early afterdepolarizations (indicated by arrows) in WT myocytes. B: superimposed original traces of action potentials with and without early afterdepolarizations (EAD). C: summary data for EAD frequency. The time points refer to the time at the end of the respective acquisition train interval (90 s). APs were analyzed under basal conditions (0 min = from −130s to 0 s), 4 min after start of H₂O₂ superfusion (= from 120 s to 250 s) and again 8 min after start of H₂O₂ superfusion (= 360 s to 490 s). For calculation of EAD frequency, for each interval, the number of EADs was normalized to the interval duration. At 8 min after onset of H₂O₂ exposure, WT cardiomyocytes exhibited a significantly increased EAD frequency compared with redox-dead KI cardiomyocytes (*P = 0.0035, mixed-effects analysis with Tukey’s post hoc correction). Interestingly, treatment of WT as well as KI myocytes with TTX completely abolished EAD formation even after 8 min of H₂O₂ treatment (WT + H₂O₂ vs. WT + H₂O₂ + TTX, #P = 0.0126; KI + H₂O₂ vs. KI + H₂O₂ + TTX, P = 0.7045; mixed-effects analysis with Tukey’s post hoc correction). At the time points 0, 4, and 8 min, respectively, the number of mice/cells were: 8/16, 8/11, and 8/9, respectively, for WT; 2/2, 2/2, and 2/2, respectively, for WT + TTX; 5/16, 5/11, and 5/8, respectively, for KI; and 2/4, 2/4, and 2/3, respectively, for KI + TTX. The reduced number of cells at later time points is a function of H₂O₂-dependent cell death. Values are means ± SE. WT, wild-type; KI, knock-in; AP, action potential; TTX, tetrodotoxin.

Consistent with increased AP duration, H₂O₂ exposure resulted in significantly more EADs in WT compared with in KI myocytes. At 8 min after onset of H₂O₂ exposure, EAD frequencies were 0.258 ± 0.082/s in WT and 0.079 ± 0.048/s in KI myocytes [N (n) = 8 (9) vs. 5 (8); P = 0.0035] (Fig. 4,A and C). There were no significant differences between EAD take-off potentials or EAD delay between WT and KI cardiomyocytes, neither at 4 min nor at 8 min of treatment with H₂O₂ (Supplemental Fig. S2, A and B; all Supplemental material is available at https://doi.org/10.5283/epub.41754). Importantly, when WT or KI myocytes were additionally treated with TTX to inhibit late I_Na, no EADs could be observed [WT: 0.258 ± 0.082/s vs. 0.0 ± 0.0/s; H₂O₂ vs. H₂O₂ + TTX N (n) = 8 (9) vs. 2 (2), P = 0.013; KI: 0.078 ± 0.048/s vs. 0.0 ± 0.0/s; N (n) = 5 (8) vs. 2 (3), H₂O₂ vs. H₂O₂ + TTX, P = 0.705] (Fig. 4C).

H₂O₂-dependent inhibition of I_to is mediated by oxidative activation of PKA.

Transient outward current (I_to), a major contributor to early repolarization of the action potential (phase 1), was measured by application of a voltage step protocol from −50 mV to +90 mV (10-mV steps) in ruptured whole cell voltage-clamp configuration. In WT cardiomyocytes, H₂O₂ treatment resulted in a significant decrease in I_to density from 10.01 ± 0.82 A/F to 8.19 ± 1.01 A/F [N (n) = 8 (16) vs. 8 (11); P = 0.02, Fig. 5, A and B]. Importantly, this decrease could be completely abolished by preincubation with the PKA-specific inhibitor Rp-8-Br-cAMPS (10 µM), suggesting that PKA-dependent regulation of I_to may be involved. In the presence of H₂O₂ and Rp-8-Br-cAMPS, I_to density was 11.42 ± 1.26 A/F [N (n) = 9 (12) vs. 8 (11); P < 0.0001 vs. WT + H₂O₂, Fig. 5, A and B].

Fig. 5.

Effect of H₂O₂ on transient outward potassium current (I_to) in WT and KI cardiomyocytes. A: original traces of outward currents in ventricular cardiomyocytes isolated from WT and redox-dead KI mice under basal conditions and after exposure to H₂O₂ (6–12 min) with or without pretreatment with the PKA-selective inhibitor Rp-8-Br-cAMPS. The inset shows the applied voltage protocol. B and C: I–V curves of I_to in WT (B) and KI (C) cardiomyocytes. B: in WT cardiomyocytes, H₂O₂ causes a marked decrease in I_to density, which could be completely abolished by Rp-8-Br-cAMPS (*P < 0.05 vs. Veh, #P < 0.05 vs. H₂O₂ + Rp-8-Br-cAMPS). Numbers of mice/cells were: 8/16 (Veh), 8/11 (H₂O₂), and 9/12 (H₂O₂ + Rp-8-Br-cAMPS). C: in contrast, H₂O₂ did not affect I_to in KI cardiomyocytes. Numbers of mice/cells were: 7/12 (Veh), 5/10 (H₂O₂), and 6/13 (H₂O₂ + Rp-8-Br-cAMPS). Values are means ± SE. Two-way ANOVA with Holm–Sidak’s post hoc correction. WT, wild-type; KI, knock-in; PKA, protein kinase A.

In contrast to WT, induction of ROS generation by H₂O₂ did not affect I_to density in myocytes lacking redox-dependent activation of PKA. In KI myocytes, I_to density was 10.50 ± 1.37 A/F (vehicle) versus 10.68 ± 1.58 A/F [H₂O₂; N (n) = 7 (12) vs. 5 (10); P = 0.91, Fig. 5, A and C].

In accordance, preincubation of the KI myocytes with Rp-8-Br-cAMPS did not alter I_to current density under oxidative stress [KI H₂O₂ + Rp-8-Br-cAMPS vs. KI H₂O₂: 11.02 ± 1.31 A/F vs. 10.68 ± 1.58 A/F, N (n) = 5 (10) vs. 5 (10), P = 0.78; KI vehicle vs. KI H₂O₂ + Rp-8-Br-cAMPS 10.50 ± 1.37 vs. 11.02 ± 1.31 A/F, N (n) = 7 (12) vs. 5 (10); P = 0.91] (Fig. 5, A and C).

It is worth noting that time constants of I_to inactivation (τ_slow and τ_fast) were not influenced by H₂O₂, neither in wild-type nor in knock-in myocytes (data not shown).

H₂O₂ causes inhibition of I_K1 by oxidative activation of PKA.

Similar to I_to, inward rectifier potassium current (I_K1) is a major contributor to cardiac repolarization. We measured I_K1 by application of a voltage step protocol in ruptured whole cell patch clamp (from −130 mV to −30 mV, 10-mV steps). I_K1 was blocked by exposure to BaCl₂, and the Ba-sensitive current was used as a measure of I_K1. Similar to inhibition of I_to, exposure of isolated WT cardiomyocytes to H₂O₂ resulted in a significant reduction of I_K1 density. At −130 mV, compared with vehicle, H₂O₂ reduced I_K1 from −11.67 ± 1.198 A/F to −8.47 ± 0.848 A/F [N (n) = 8 (16) vs. 8 (14); P = 0.007, Fig. 6, A and C]. This H₂O₂-dependent inhibition of I_K1 could be prevented by preincubation with Rp-8-Br-cAMPS, pointing toward a PKA-dependent I_K1 regulation. In the presence of H₂O₂ and Rp-8-Br-cAMPS, I_K1 density was 11.74 ± 1.049 A/F [N (n) = 8 (14) vs. 9 (15); P = 0.006 vs. WT + H₂O₂, Fig. 6, A and C].

Fig. 6.

Effect of H₂O₂ on inward rectifying potassium current (I_K1) in WT and KI cardiomyocytes. A and B: original traces of inward currents before (top) and after (middle) BaCl₂ superfusion. Digital subtraction results in the Ba²⁺-sensitive current I_K1 (bottom). C: I–V curves of I_K1 in WT cardiomyocytes reveal a significant inhibition of I_K1 upon exposure to H₂O₂ (6–12 min) (*P < 0.05), which could be completely abolished by preincubation with the PKA-selective inhibitor Rp-8-Br-cAMPS (#P < 0.05). Numbers of mice/cells were: 8/16 (Veh), 8/14 (H₂O₂), and 9/15 (H₂O₂ + Rp-8-Br-cAMPS). The inset shows the applied voltage protocol. D: in contrast, I_K1 increased in KI cardiomyocytes in response to H₂O₂, which could not be prevented by Rp-8-Br-cAMPS (*P < 0.05 vs. Veh). Numbers of mice/cells were: 7/11 (Veh), 5/10 (H₂O₂), and 6/13 (H₂O₂ + Rp-8-Br-cAMPS). Values are means ± SE. Two-way ANOVA with Holm–Sidak’s post hoc correction. WT, wild-type; KI, knock-in; PKA, protein kinase A.

To further investigate the role of redox activation of PKA I in this context, we repeated the experiments in redox-dead PKA RI Cys17Ser knock-in mice. Interestingly, compared with WT, basal I_K1 density was reduced in KI myocytes to −8.046 ± 0.915 A/F [at −130 mV N (n) = 8 (16); P < 0.05 vs. WT, Fig. 6, B and D] despite no difference in resting membrane potential or AP duration.

In accordance with reduced I_K1 under basal conditions, Western blotting revealed that the expression of the main pore-forming subunit Kir2.1 was significantly diminished in KI versus WT myocytes (0.198 ± 0.014 vs. 0.323 ± 0.024 arbitrary units, N = 3 vs. 4, P = 0.0094, Supplemental Fig. S1).

Surprisingly, in contrast to WT cardiomyocytes, exposure to H₂O₂ caused a significant increase in I_K1 to −11.000 ± 0.919 A/F in KI cardiomyocytes [at −130 mV, N (n) = 7 (11) vs. 5 (10); P = 0.01, Fig. 6, B and D]. This increase in I_K1 density, however, was unaffected by preincubation with Rp-8-Br-cAMPS, suggesting that PKA is not involved. In the presence of H₂O₂ and Rp-8-Br-cAMPS, I_K1 density was −10.866 ± 1.302 A/F [at −130 mV, N (n) = 7 (11) vs. 6 (13); P = 0.999 vs. KI + H₂O₂, Fig. 6, B and D].

In contrast, the H₂O₂-dependent increase of I_K1 in KI myocytes was abolished by preincubation with the selective inhibitor of Ca/calmodulin-dependent protein kinase II (AIP). In the presence of H₂O₂ and AIP, I_K1 density was −8.175 ± 0.711 A/F [N (n) = 7 (11) vs. 6 (10); P < 0.05 vs. KI + H₂O₂, data not shown]. Together with the lack of H₂O₂-dependent I_to inhibition, the H₂O₂-dependent increase of I_K1 in KI myocytes may partly explain the lack of AP prolongation after H₂O₂ exposure.

DISCUSSION

In this study, we investigated the role of oxidative activation of PKA for action potentials and early afterdepolarizations in murine ventricular cardiomyocytes. We demonstrate that redox-activated PKA is involved in the ROS-induced inhibition of transient outward potassium current (I_to) and inward rectifying potassium current (I_K1), which are crucial for membrane repolarization and stabilization of the resting membrane potential (41, 67, 68). Moreover, we showed that lack of oxidative PKA activation protected myocytes from H₂O₂-dependent APD prolongation and induction of EADs, further underscoring the pathophysiological relevance of this novel mechanism of PKA activation.

ROS signaling in the heart involves oxidative activation of PKA I.

ROS are centrally involved in cellular signaling processes mediating adaptive responses in the healthy heart (48, 50). However, under pathological conditions, for example, heart failure, ROS production exceeds the antioxidative capacity of cardiomyocytes resulting in oxidative stress, which is implicated in the development of contractile dysfunction and arrhythmias (25, 28, 60).

Interestingly, a novel cAMP-independent mechanism of PKA I activation via oxidation of its regulatory subunits (RIα) has been delineated (5). The relevance of this pathway has been confirmed in ischemia-dependent angiogenesis (5) and in PDGF-dependent signaling (13). Here, we show that ROS, which are generated by external application of H₂O₂, are capable of oxidizing PKA RIα subunits with a consequent increase in RI dimerization (Fig. 1) in isolated murine ventricular cardiomyocytes, suggesting a potential relevance of this pathway also in cardiac pathophysiology. It is worth noting that in contrast to KI cardiomyocytes, there was a detectable level of RI dimerization also in untreated WT cells (vehicle), probably owing to an intrinsic steady-state ROS generation in these healthy ventricular myocytes. This suggests that there may be a potential role of oxidation-activated PKA also for physiological ROS signaling.

Redox-activated PKA regulates cardiac action potential and contributes to arrhythmogenesis.

Since PKA is involved in the regulation of several ion channels, we next sought to analyze the impact of oxidation-activated PKA on the cardiac action potential. Therefore, we induced oxidative stress in WT ventricular cardiomyocytes, which, in contrast to the redox-dead KI myocytes, resulted in a marked increase in action potential duration at 90% of repolarization. Furthermore, also APD₂₀ and APD₅₀ were significantly increased in WT cardiomyocytes upon H₂O₂ treatment, pointing to alterations in ionic currents, which shape the early phase of cardiac action potential (e.g., I_to). In contrast, this H₂O₂-dependent effect was not observed in KI myocytes.

Importantly, H₂O₂ treatment significantly increased the action potential amplitude and also the upstroke velocity in both WT and KI myocytes (ANOVA P for factor H₂O₂ 0.0018 and 0.0042, respectively), with no difference between the genotypes (P for factor genotype 0.4612 and 0.9645, respectively). Consequently, although regulation of the phase I action potential depolarization could affect AP duration, it is unlikely to explain the lack of APD prolongation in KI myocytes and may thus be a consequence of lacking oxPKA-dependent regulation of K currents.

Intriguingly, APD prolongation in WT cardiomyocytes was associated with an increased occurrence of early afterdepolarizations (EADs), as compared with in KI cardiomyocytes. EADs are defined as positive voltage deflections during repolarization of the cardiac action potential (phases 2 and 3) (62) and constitute substantial proarrhythmic events at the cellular level (64). To date, afterdepolarizations induced by oxidative stress were mainly explained by CaMKII-mediated augmentation of the late Na current (late I_Na) with consequent dysregulation of Na⁺ and Ca²⁺ homeostasis via Na-Ca exchanger (NCX) activity (53, 61, 69). In the present article, we tested the contribution of ROS-dependent enhancement of late I_Na to the generation of EADs. Interestingly, both WT and KI myocytes exposed to the selective late I_Na inhibitor TTX did not show any H₂O₂-dependent EAD formation, suggesting that late I_Na is critical for ROS-induced generation of EADs. There are only few and discrepant data regarding PKA-dependent regulation of late I_Na, suggesting that PKA is either not involved in the regulation of this current (1, 16, 27, 52) or that it may directly stimulate late I_Na early during the AP plateau at most positive membrane potentials (23). However, our EADs occur at much more negative membrane potentials, where late I_Na has been shown to be regulated by CaMKII (23). Furthermore, whether these data from rabbit myocytes can be transferred to rodent models that lack AP plateau remains elusive. Thus, oxPKA probably facilitates EAD formation by reducing the repolarization reserve, which would favor late I_Na-dependent EAD formation. These data are in accordance with previous publications, suggesting that redox-enhanced late I_Na (by oxidation of CaMKIIδ) is a major proarrhythmogenic mechanism in situations of increased ROS formation (60). Consequently, other mechanisms are likely to be involved in the PKA-dependent induction of early afterdepolarizations.

Of note, β-adrenergic signaling has been shown to stimulate EADs mainly by activation of L-type Ca²⁺ current (LTCC) in the setting of concomitant slow activation of the delayed rectifier potassium current I_Ks (35, 66). Importantly, H₂O₂ treatment of cardiomyocytes has been shown to rather decrease LTCC (18, 19), possibly by direct thiol-oxidation of its α_1C subunit (7, 15). In parallel, the β-adrenergic responsiveness of these channels is mitigated upon oxidative stress downstream of cAMP (26), making a specific contribution of PKA-dependent dysregulation of LTCC to EAD formation in our model unlikely. Furthermore, although data on LTCC in heart failure are discrepant, most studies suggest LTCC current density to be unaltered in failing cardiomyocytes (22, 34, 40), but failing cardiomyocytes show a huge increase in the propensity for EAD formation. This suggests that a direct LTCC dysregulation by PKA is unlikely to be the main contributor to EAD formation in our model. A lack of LTCC dysregulation, however, does not prevent the LTCC-dependent generation of EADs. In contrast, since EAD take-off potential and the time of reactivation from the initial stimulus (EAD delay) are in the range of the LTCC window current (Supplemental Fig. S2, A and B), it is very likely that LTCC is involved in the EAD formation in our model.

Besides increased depolarizing currents (e.g., late I_Na), a reduced repolarization reserve has been found in HF and conditions of increased ROS generation (3, 34, 42). It is widely accepted that the dysregulated transient outward current I_to is critically involved in EAD formation (8, 20, 62, 65).

In accordance, our data demonstrate that PKA mediates a ROS-dependent impairment of repolarization reserve by inhibition of I_to and I_K1. In consequence, depolarizing currents are more likely to induce early afterdepolarizations, especially since AP duration was also prolonged. A PKA-dependent inhibition of I_to has been described in several studies (17, 44, 49). Intriguingly, several consensus PKA phosphorylation sites were identified within the Kv4.2 and Kv1.4 α subunits, constituting I_to,fast and I_to,slow channels, respectively, with PKA-dependent phosphorylation verified in heterologous expression systems and neurons (2, 21, 44, 49, 55).

However, an A-kinase-anchoring protein targeting PKA type I to the channel subunits has not been detected to date. Of note, the fact that PKA-dependent regulation of I_to was abolished in our KI myocytes argues for a cAMP-independent mechanism involving oxidative activation of PKA.

Noteworthy, it is still heavily debated whether a decrease in I_to resulting in APD prolongation with reactivation of L-type Ca current (LTCC) and late I_Na or an increase in I_to, which lowers AP plateau voltage into LTCC window current range (71), is more arrhythmogenic. Importantly, species differences must be taken into account. In murine ventricular cardiomyocytes, where the contribution of I_to to membrane repolarization is much more pronounced than in rabbit or human ventricular myocytes, resulting in a more negative plateau voltage, inhibition of I_to rather than its stimulation might set the AP plateau voltage into LTCC window current range and thus facilitate EAD formation, as seen in our model.

Besides inhibition of I_to, we show here that oxidative stress also inhibits I_K1 in a PKA-dependent manner, which is in accordance with several prepublished studies (9, 32, 33, 56, 57). Interestingly, Kir2.x subunits exhibit only one PKA consensus site (Ser425 and Ser430 for Kir2.1 and Kir2.2, respectively). Although PKA-dependent phosphorylation of Ser425 of Kir2.1 subunits with consequent marked inhibition of I_K1 was demonstrated in a heterologous expression system (COS-7 cells) (63), there is evidence that the respective consensus site of Kir2.2 is not relevant for its PKA-dependent regulation (72), suggesting an indirect effect, for example, by phosphorylation of associated molecules within a macromolecular complex to which PKA could be targeted via A-kinase-anchoring proteins (AKAPs) (51). Importantly, the ROS-induced inhibition of I_K1 was abolished in our KI myocytes lacking oxidation-sensitive RI subunits, again pointing to a direct oxidative activation of PKA instead of a cAMP-dependent pathway. Interestingly, reduced I_K1 is also encountered in heart failure (3, 30, 34). Although I_K1 is critical for stabilization of the resting membrane potential, it also substantially contributes to action potential repolarization (39, 68). Consequently, in a Kir2.1 knockout mouse model, action potential duration was markedly increased with an enhanced propensity for EADs (39). Noteworthy, although I_K1 density was reduced by almost 95% in these ventricular myocytes, resting membrane potential was unaltered compared with wild-type (39), which is in parallel to our action potential data.

Limitations.

Although our data show a significant reduction of I_K1 and I_to as well as a concomitant increase in EAD frequency, all of which are mediated by oxidation-activated PKA, we cannot preclude a possible contribution of other pathways to increased formation of EADs upon oxidative stress. EAD generation first requires a substrate, that is, APD prolongation, which renders the AP plateau phase vulnerable to the “second hit,” which is the trigger that actually induces the EAD. A simple APD prolongation alone, due to decreased potassium currents, is not sufficient to trigger EADs (31). On the other hand, any APD shortening will suppress EAD formation. We show here that the AP duration can be prolonged by oxPKA due to inhibition of repolarizing potassium currents such as I_to and I_K1. If, however, oxPKA also influences the second hit—that is, the trigger—which manifests as enhanced activity of the depolarizing currents late I_Na, LTCC, I_NCX, or a combination is unclear and merits further investigation. Importantly, H₂O₂ exposure does not result in a selective oxidation of PKA. There are many other targets for ROS-dependent oxidation including CaMKII, which is proarrhythmic via stimulation of late I_Na and dysregulation of cellular Ca handling (54, 59, 60), as well as ryanodine receptors (70) with consequent increase in sarcoplasmic reticulum (SR) Ca leak, which all can contribute to EAD formation. In accordance, our data underscore the importance of late I_Na in the generation of EADs, as they are completely abolished in both WT and KI myocytes upon treatment with TTX. Furthermore, LTCC, which also has been shown to be critically involved in EAD formation, might also play a role in our experimental setup, since all the observed EADs take place in the LTCC window current range.

Another limitation of our study is the fact that there is apparently no significant difference in I_K1 current density in the voltage range where all of the observed EADs occur. Although there is a clear trend to a reduction of I_K1 in WT cardiomyocytes treated with H₂O₂, which cannot be detected in KI cardiomyocytes, these results do not reach statistical significance. This could in part be a result of the artificial protocol using BaCl₂ to inhibit I_K1. It has been shown that Ba²⁺-induced block of I_K1 is voltage dependent, resulting in a smaller degree of block at more positive membrane potentials (29). In combination with the relatively small Ba²⁺ concentration used to avoid off-target effects (in other studies, concentrations of up to 1 mM are used), this could lead to an underestimation of I_K1 especially at more positive membrane potentials, possibly concealing the clear effect shown at more negative potentials.

SUMMARY

Our data indicate that oxidation-activated PKA type I is critically involved in the ROS-dependent regulation of I_K1 and I_to and thus contributes to APD prolongation and potentially facilitates early afterdepolarizations under conditions of oxidative stress, making it a potential target for antiarrhythmic therapy in patients with heart failure.

GRANTS

The German Cardiac Society (DGK) funded a research grant to M. Trum. S. Wagner is funded by DFG Grants WA 2539/4-1, 5-1, 7-1, and 8-1. L. S. Maier is funded by DFG grants MA 1982/5-1 and 7-1. L. S. Maier. and S. Wagner are also funded by the DFG SFB 1350 grant (Project number 387509280, TPA6). L. S. Maier and S. Wagner are supported by the ReForM C program of the faculty of medicine of the University of Regensburg.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

M.T. and S.W. conceived and designed research; M.T., M.M.T.I., S.L., M.B., and P.H. performed experiments; M.T. and S.W. analyzed data; M.T. and S.W. interpreted results of experiments; M.T., and S.W. prepared figures; M.T. and S.W. drafted manuscript; M.T., M.M.T.I., P.E., L.S.M., and S.W. edited and revised manuscript; all authors approved final version of manuscript.