β-adrenergic Regulation of Late Na⁺ Current During Cardiac Action Potential Is Mediated by Both PKA and CaMKII

Abstract

Late Na+ current (I_NaL) significantly contributes to shaping cardiac action potentials (AP) and increased I_NaL is associated with cardiac arrhythmias. β-adrenergic receptor (βAR) stimulation and its downstream signaling via protein kinase A (PKA) and Ca²+/calmodulin-dependent protein kinase II (CaMKII) pathways are known to regulate I_NaL. However, it remains unclear how each of these pathways regulates I_NaL during the AP under physiological conditions. Here we performed AP-clamp experiments in rabbit ventricular myocytes to delineate the impact of each signaling pathway on I_NaL at different AP phases to understand the arrhythmogenic potential. During the physiological AP (2 Hz, 37°C) we found that I_NaL had a basal level current independent of PKA, but partially dependent on CaMKII. PAR activation (10 nM isoproterenol, ISO) further enhanced I_NaL via both PKA and CaMKII pathways. However, PKA predominantly increased I_NaL early during the AP plateau, whereas CaMKII mainly increased I_NaL later in the plateau and during rapid repolarization. We also tested the role of key signaling pathways through exchange protein activated by cAMP (Epac), nitric oxide synthase (NOS) and reactive oxygen species (ROS). Direct Epac stimulation enhanced I_NaL similar to the PAR-induced CaMKII effect, while NOS inhibition prevented the βAR-induced CaMKII-dependent I_NaL enhancement. ROS generated by NADPH oxidase 2 (NOX2) also contributed to the ISO-induced I_NaL activation early in the AP. Taken together, our data reveal differential modulations of I_NaL by PKA and CaMKII signaling pathways at different AP phases. This nuanced and comprehensive view on the changes in I_NaL during AP deepens our understanding of the important role of I_NaL in reshaping the cardiac AP and arrhythmogenic potential under elevated sympathetic stimulation, which is relevant for designing therapeutic treatment of arrhythmias under pathological conditions.

Graphical abstract:

1. INTRODUCTION

Activation of Na+ channel upon excitation leads to a fast, transient Na+ current (I_NaT) generating the upstroke of action potential (AP). However, under a sustained depolarization such as the plateau phase of AP in ventricular cardiomyocytes, a tiny fraction of Na+ channels may remain open/reopen generating a non-inactivating or persistent Na+ current referred as late Na+ current (I_NaL) [1]. I_NaL significantly contributes to shaping cardiac AP and pathological augmentation of I_NaL is associated with increased risk for cardiac arrhythmias [2]. Several gating modalities of Na+ channels with different voltage-and time- dependent properties have been identified that contribute to increased I_NaL in pathological states, including early channel bursting [3], late scattered opening [4], window current [5] and non-equilibrium gating [6]. Mutations of Na+ channels linked to long QT syndrome 3 (LQT3) in patients cause increased I_NaL, and may preferentially affect one of these gating modalities resulting in distinct molecular determinants [7]. Beside inherited genetic defects, several signaling pathways modulate I_NaL, and are associated with heart diseases like ischemia, cardiomyopathies and heart failure [2]. One such regulatory mechanism is Ca2+/calmodulin-dependent protein kinase II (CaMKII), which causes complex INa gating changes including elevated I_NaL [8, 9]. CaMKII is upregulated in pathologic states and is associated with enhanced I_NaL contributing to arrhythmias and cardiac dysfunction [10–12]. β-adrenergic receptor (βAR) stimulation has also been shown to regulate Na+ channels through both cAMP-dependent and independent pathways[13]. βAR stimulation increases whole-cell Na+ channel conductance by protein kinase A (PKA) phosphorylation [14], predominantly via enhanced Na+ channel trafficking to the sarcolemma [15]. In addition, a PKA-independent Gs protein mediated effect on Na+ channel gating has also been suggested [13, 16]. However, most previous studies of cardiac Na+ channel regulation focused on I_NaT rather than I_NaL [17–19], and most were conducted under non-physiological conditions. Typically, I_NaL was recorded under a rectangular voltage command to −20/−30 mV at very low pulse frequency and with strong intracellular Ca2+ buffering. So, no direct data is available as to βAR regulation of I_NaL during physiological APs, and how different pathways may contribute to those I_NaL changes. Therefore, the goal of our study was to determine how PAR stimulation modulates I_NaL during the physiological AP with physiological ionic conditions in rabbit ventricular cardiomyocytes.

βAR stimulation induces complex signaling, including crosstalk between PKA and CaMKII pathways in cardiac myocytes [20]. Much prior work on PKA-CaMKII crosstalk during cardiac βAR stimulation targeted Ca2+ handling [21–24]. However, because I_NaL is known to be regulated by both PKA[25]and CaMKII [8, 9], we focused here on dissecting PKA- versus CaMKII-dependent mechanisms of I_NaL modulation during the cardiac AP. Moreover, both PKA and CaMKII can be regulated by posttranslational modifications causing autonomous activity. CaMKIIô autophosphorylation (Thr287), oxidation (Met280 281) and S-nitrosylation (Cys273/290) are well-established modulators of CaMKII activity [26, 27]. The exchange protein directly activated by cAMP (Epac) can also induce CaMKII activation upon PAR stimulation [28, 29]. Similarly, oxidation [30] and S-nitrosylation [31] of PKA have been shown to activate type I PKA independent of cAMP via interprotein disulfide bound formation between the two regulatory subunits. However, it is unknown whether such autonomous PKA activation occurs upon acute βAR stimulation physiologically in adult myocytes. It has also been shown that stimulation of Gq proteins by angiotensin II results in PKA-dependent enhancement of I_NaT, but CaMKII-dependent enhancement of I_NaL [32]. However, it remains unknown whether similar mechanism occur upon βAR stimulation and how different posttranslational modifications might affect I_NaL via PKA and CaMKII autonomous activation. Hence, we aimed to study how PKA and CaMKII mediate I_NaL under physiological conditions and whether Epac, reactive oxygen species (ROS) and nitric oxide signaling influence these PAR-induced I_NaL changes during the cardiac AP. Our results indicate that physiological pacing alone causes CaMKII-dependent augmentation of I_NaL, whereas PAR activation induces additional I_NaL that is mediated by both PKA and CaMKII (early and late in the AP, respectively).

2. METHODS

All animal handling and laboratory procedures were in accordance with the approved protocols of the Institutional Animal Care and Use Committee at University of California, Davis confirming to the Guide for the Care and Use of Laboratory Animals published by the US National Institute of Health (8^th edition, 2011).

2.1. Animal model, cell isolation

New Zealand White rabbits (male, 3–4 months old, 2.5–3 kg) were purchased from Charles River Laboratories (Wilmington, MA, USA). Rabbits were first injected with heparin (1000 U/kg) and then anesthetized with isoflurane inhalation (3–5%). After achieving deep anesthesia, a standard enzymatic technique was used to isolate ventricular myocytes at 37°C as previously described [33, 34]. Briefly, hearts were mounted on a Langendorff system and retrogradely perfused for 5 min with an oxygenated solution containing (in mmol/L): NaCl 138, KCl 5.4, CaCl₂ 0.05, MgCl₂ 1, NaH₂PO₄ 0.33, NaHCO₃ 10, HEPES 10, glucose 6, pyruvic acid 2.5; at pH=7.4. When blood was removed from the coronary circulation, we added 1 mg/mL type II collagenase (305 U/mg; Worthington Biochemical Co., Lakewood, NJ, USA), 0.05 mg/ml protease type XIV (Sigma-Aldrich Co., St. Louis, MO, USA) and 1 mg/ml bovine serum albumin, which was perfused for ~30 min to enzymatically dissociate cells. The left ventricle minced and Ca²+ concentration [Ca²+]_o was gradually restored to 1.2 mmol/L.

2.2. Electrophysiology

Isolated myocytes were placed in a temperature-controlled Plexiglas chamber (Cell Microsystems Inc., Research Triangle Park, NC, USA) and continuously perfused with a bicarbonate-containing Tyrode (BTY) solution with the following composition (in mmol/L): NaCl 124, NaHCO₃ 25, KCl 4, CaCl₂ 1.2, MgCl₂ 1, HEPES 10, and glucose 10; pH=7.4. Electrodes were fabricated from borosilicate glass (World Precision Instruments Inc., Sarasota, FL, USA) having tip resistances of 2–2.5 MQ when filled with internal solution. In experiments aimed to preserve the physiological Ca²+ homeostasis during AP, the internal solution contained (in mmol/L): K-Aspartate 110, KCl 25, NaCl 5, Mg-ATP 3, HEPES 10, cAMP 0.002, phosphocreatine dipotassium salt 10, and EGTA 0.01; pH was set to 7.2 with KOH. In another set of experiments intracellular Ca²+ concentration, [Ca²+]_i was buffered to nominally zero by using an internal solution containing (in mmol/L): K-Aspartate 100, KCl 25, NaCl 5, Mg-ATP 3, HEPES 10, cAMP 0.002, phosphocreatine dipotassium salt 10, and BAPTA 10; pH=7.2.

AP-clamp experiments were conducted as previously described [33, 35, 36]. Briefly, the steps are: (1) Recording the cell’s steady-state AP under I-clamp at given pacing frequency. (2) Applying this AP to the same cell as the V-clamp command pulse at the same pacing frequency. The net current (reference current) during the AP at steady-state should be zero (Fig. 1A). (3) Isolation of the current of interest (compensation current; Fig. 1A, third panel) uses specific blockers to remove only that from the net current. (4) The current of interest is obtained by subtraction, (i.e. reference current - compensation current). In some experiments, a pre-recorded “typical” AP waveform was used as the V-clamp command (canonical AP-clamp) and delivered at 2 Hz steady-state frequency to measure I_NaL using tetrodotoxin (TTX, 10 μM) or the specific I_NaL blocker GS-458967 (GS, 1 μM) [37]. Drug-sensitive current (I_GS or I_TTX) was calculated by subtracting the average compensation current when TTX or GS achieved steady- state inhibition (3 min of perfusion) from the reference current right before TTX or GS application (average of 60 consecutive traces in each case). In all cases I_NaL was taken as the I_GS measured in this way, beginning after steady state had been achieved for pretreatment with various inhibitors (next section). This I_NaL density was calculated after normalizing to cell capacitance, determined in each cell using 10 ms hyperpolarizing pulses from −10 mV to −20 mV. Average cell capacitance was 144.98±0.99 pA/pF in the measured 233 cells from 46 animals.

Figure 1. Action potential-clamp measurement of I_NaL using tetrodotoxin and GS-458967.

Overview of the action potential-clamp technique. First, using an AP as voltage command a pre-drug control or reference current is recorded (above). Next, when a drug is applied, a compensation current is recorded specific to the drug action (middle). The drug-sensitive current is obtained as the difference current (i.e. subtracting the compensation current from the reference current) (below). (B) Effect of TTX (10 μM), then GS-458967 (GS, 1 μM) under self AP-clamp in rabbit ventricular myocytes. The cell’s own steady-state AP was applied as voltage command (upper panel). Application of TTX inhibited both the transient Na+ current (I_NaT) and the late Na+ current (I_NaL). GS in the presence of TTX did not cause any additional current inhibition. Peak I_NaT was out of scale, thus I_NaT amplitude was reported 5 ms after the peak of AP when the Na+ channels are already largely inactivated. (C) Effect of GS, then TTX under self AP-clamp in rabbit ventricular myocytes. GS also partially inhibited I_NaT, but it mostly inhibits I_NaL. TTX in the presence of GS further inhibited I_NaT, but it did not have any additional effect on I_NaL. (D) Canonical AP-clamp (using a prerecorded, typical rabbit AP) to measure I_NaL. To ensure that the recorded currents are not contaminated with L-type Ca²+ current and the Na+/Ca²+ exchanger, 10 μM nifedipine in the extracellular solution and 10 mM BAPTA in the pipette solutions were applied, respectively. Just as in panel B, no additional effect of GS was observed when applied after TTX in a cumulative manner. Columns and bars represent mean±SEM. Asterisks denote significant difference using paired, two-tailed Student’s t test. ***p<0.001.

To avoid contamination with I_NaT, GS-458967-sensitive currents were analyzed starting from 10 ms after the AP peak, except for when we compared the effect of TTX and GS-458967 on early I_Na, where inhibited currents are reported starting 5 ms after AP peak. Note that I_NaT peak density cannot be reliably measured under AP-clamp with physiological conditions despite rigorous series resistance compensation, because the capacitive transient (stimulation spike) still overlaps with the AP upstroke. Axopatch 200B amplifier (Axon Instruments Inc., Union City, CA, USA) was used for AP and I_NaL measurements and the signals were digitized at 50 kHz by a Digidata 1440A A/D converter (Molecular Devices, Sunnyvale, CA, USA) under software control (pClamp 10, Molecular Devices). Signal amplification was set to achieve high resolution in the range of I_NaL magnitude. The series resistance was typically 3–5 MQ and it was compensated by 90% to achieve good voltage control. Experiments were discarded if the series resistance was higher or increased by >10%. Reported AP voltages are already corrected for liquid junction potentials. All experiments were conducted at 36±0.1°C.

2.3. Chemicals and cell treatments

Chemicals and reagents were purchased from Sigma-Aldrich if not specified otherwise. GS-458967 was obtained from Gilead Sciences, Inc. (Foster City, CA, USA). Gp91ds-tat peptide was from AnaSpec (Fremont, CA, USA).

Cell pretreatments with different drugs occurred for ~2 hours prior the seal formation and the given drug was also continuously present in both the perfusing and pipette solutions. To inhibit CaMKII, KN-93 (1 pM) and the more specific autocamtide-2-related inhibitory peptide (AIP, 1 μM, myristoylated) were used and compared with KN-92 (1 μM). To inhibit PKA, H-89 (1 μM) and the more specific protein kinase inhibitor-(14–22)-amide (PKI, 1 pM, myristoylated) were used. cAMP-dependent pathways were also examined using the Rp-isomer of adenosine-3’,5’-cyclic monophosphorothioate (Rp-cAMPS, 100 μM) as a competitive inhibitor of cAMP binding. To investigate the influence of ROS pathway on I_NaL modulation, a reductant and ROS scavenger “cocktail” was applied containing reduced glutathione (GSH, 10 mM) and N-acetyl cysteine (NAC, 10 mM). The involvement of NADPH oxidase 2 (NOX2) pathway was further tested using its specific inhibitor gp91ds-tat (1 μM). As positive control, H₂O₂ (100 μM) was applied. To examine the effect of nitric oxide signalling, non-specific nitric oxide synthase (NOS) inhibitor Nω-nitro-L-arginine methyl ester (L-NAME, 1 mM) pretreatment was used. Epac was directly activated using 8-(4-Chlorophenylthio)-2’-O-methyladenosine 3’,5’-cyclic monophosphate (8- pCPT, 3 μM). To examine the [Ca²+]_i-dependence of pathways mediating βAR response, the pipette solution was supplemented with 10 mM BAPTA (with no added Ca²+), and we waited for 10 min after cell break-in to allow the agent to diffuse sufficiently into the cell, meanwhile this effect was monitored using a voltage step pulse to +5 mV arising at every 1 s to follow the loss of Ca²+-dependent inactivation of L-type Ca²+ current and myocyte contraction. In all other AP-clamp experiments the recordings were started 5 min after membrane rupture.

βAR responses were evoked adding 10 nM isoproterenol (ISO) to the appropriate perfusion solution. ISO response reached a steady-state in 2 min which was maintained usually for ~5 min before some desensitization was observed. GS was applied 2 min after ISO application and the GS-sensitive current traces were analyzed following 3 min of perfusion. If any sign of Ca²⁺ current rundown was observed in periodic tests or either ISO or GS effect did not reach steady-state, those experiments were excluded from the analysis.

2.4. Statistical analysis

Data are expressed as Mean±SEM. The number of cells in each experimental group is reported in the figures, and the cells in each group came from three to eight individual animals. Given the biological variability among cells, each cell was treated as independent in the statistical tests, although multiple cells may come from one animal. Statistical significance of differences was evaluated using paired Student’s t test to compare two groups and one-way or two-way ANOVA to compare multiple groups, followed by a Bonferroni posttest for pairwise comparisons. Differences were deemed significant if p<0.05 and denoted *p<0.05, **p<0.01, and ***p<0.001.

3. RESULTS

3.1. Profile of I_NaL under AP-clamp using TTX and GS-458967

I_NaL was measured as specific blocker-sensitive current under AP-clamp (Fig. 1A) using physiological conditions (internal and external solutions mimicking physiological ionic composition, preserved intracellular Ca²+ homeostasis, 2 Hz steady-state pacing frequency, and at 36°C).

First, we compared the effect of the selective I_NaL inhibitor, GS-458967 (GS, 1 pM) with tetrodotoxin (TTX, 10 μM). Accordingly, TTX-sensitive (I_TTX) and GS-sensitive currents (I_GS) were measured under AP-clamp using the cell’s own steady-state AP. TTX inhibited both the transient and late Na+ current (I_NaT and I_NaL) under AP-clamp (Fig. 1B). Because of overlap with the capacitive transient during AP upstroke I_NaT peak cannot be reliably measured using AP-clamp with physiological solutions at 36°C (despite rigorous series resistance compensation). However, the huge I_NaT (hundreds of A/F) inactivates rapidly, but the TTX-sensitive early decaying peak in Fig. 1A-B may include terminal decay of Inst (i.e. the last 1–2%). On the other hand, the much smaller sustained I_NaL was present throughout the entire AP, and achieved a peak density of −0.50±0.02 A/F during AP phase 3, as driving force (E_Na - E_m) increases (Fig. 1B). Cumulative application of GS in the presence of TTX did not inhibit additional current (Fig. 1B). This indicates that under our conditions, GS does not influence other ionic currents not already blocked by TTX. When the order was reversed (Fig. 1C) GS only slightly inhibited I_NaT (9.0% of that for I_TTX at 5 ms after AP peak), whereas I_NaL (as either I_GS or I_GS+I_TTX) late in the AP was identical. Thus, I_GS provides a useful measure of I_NaL during the cardiac AP under our physiological ionic conditions.

While this self AP-clamp technique is the most physiological technique to determine current profile during the cell’s own AP, each myocyte has a slightly different AP duration (APD) which can also alter I_NaL density [33]. Thus, we used a canonical rabbit AP waveform (prerecorded) in subsequent AP- clamp experiments, to obtain more controlled mechanistic insight into I_NaL modulation (Fig. 1D; in this panel only, Ca²+ current and transients were suppressed by nifedipine and with 10 mM BAPTA in the pipette). Again, 1 μM GS had no further effect after prior TTX application.

3.2. Tetrodotoxin-sensitivity of I_NaL under AP-clamp

The TTX-sensitivity of I_NaL measured under AP-clamp can suggest whether it is primarily due to cardiac Na+ channel isoforms that are relatively TTX-resistant (μM range as for the predominant cardiac isoform Na_V1.5) or TTX-sensitive (nM range as for neuronal Na+ channel isoforms). To prevent complications of TTX partial inhibition of L-type Ca²+ current at high concentrations (IC₅₀=55 μM) [38], these experiments included nifedipine. Fig. 2A shows that TTX dose-dependently inhibited I_NaL. The observed IC₅₀ values were 1.18±0.20 μM (for I_NaL net charge) and 1.08±0.07 μM (for I_NaL peak density) under AP-clamp (Fig. 2B). Importantly, ≤5% of total I_NaL was inhibited by 100 nM TTX suggesting that I_NaL is predominantly mediated by TTX-resistant Na+ channel isoforms in healthy rabbit ventricular myocytes.

Figure 2. Tetrodotoxin-sensitivity ofI_NaL under AP-clamp.

Representative tetrodotoxin (TTX)-sensitive current traces under AP-clamp at 2 Hz steady-state pacing. Increasing TTX concentrations (0.1, 0.3, 1, 3, 10, 30 μM) were applied in a cumulative manner. Pipette solution contained 10 mM BAPTA and extracellular Tyrode solution was supplemented with 10 ^M nifedipine. (B) TTX dose-response effect on I_NaL net charge and I_NaL peak. Inhibition of I_NaL was normalized to that obtained with 30 μM TTX in each cell. IC₅₀ values and Hill coefficients were determined by fitting data to the Hill equation, indicated by solid lines.

3.3. CaMKII, but not PKA regulates basal I_NaL under AP-clamp

Next, we studied the modulation of I_NaL by basal PKA or CaMKII activity under AP-clamp with physiological intracellular Ca²+ transients at 2 Hz and 36°C. Inhibition of CaMKII by extracellular application of KN-93 (1 μM) significantly decreased I_GS magnitude throughout the AP (Fig. 3A), most prominently in the phase 3 at −60 mV (−0.45±0.02 A/F in control vs. −0.24±0.04 A/F in KN-93, Fig. 3F). The same reduction in I_GS was observed when KN-93 was applied intracellularly through the patch pipette. Moreover, the more specific autocamtide-2-related inhibitory peptide, AIP (1 pM) caused similar decrease of I_GS (Fig. 3A,C-F). In contrast, the inactive KN-93analogue, KN-92 (1 μM) had no effect on I_GS (Fig. 3C-F). These data indicate that roughly half of basal I_NaL during the physiological AP is secondary to CaMKII activity.

Figure 3. CaMKII-dependent and PKA-dependent regulation of basal I_NaL under AP-clamp.

(A) GS-458967-sensitive current (I_Gs) traces (mean±SEM) under AP-clamp in control and after CaMKII inhibition using KN-93 (1 pM) or AIP (1 μM). The physiological intracellular Ca²+ cycling was preserved, and 2 Hz steady-state pacing was applied. I-V trajectories of I_GS under AP-clamp are shown in inset. (B ) I_GS traces (mean±SEM) under AP-clamp in control and after PKA inhibition using H-89 (1 μM) or PKI (1 μM). I-V relationships are shown in the inset. (C-E) I_GS density at different voltages during AP repolarization. CaMKII inhibition with KN-93 or AIP significantly decreased I_NaL at all voltages, whereas KN-92 (1 μM) had no effect on I_GS. The decrease in I_NaL was most prominent at −60 mV which may reflect the window component of I_NaL. On the contrary, PKA inhibition using PKI or Rp-cAMPS (100 μM) had no effect on I_GS, whereas H-89 may exhibit some off-target effect at −60 mV. Columns and bars represent mean±SEM. Asterisks denote significant difference using two-way ANOVA with Bonferroni posttest. *p<0.05, **p<0.01, ***p<0.001.

We also tested the effect of basal PKA activity on I_NaL under AP-clamp (Fig. 3B). Inhibiting PKA using H-89 had no effect on I_GS except for a slight inhibition in the late phase 3 of the AP (~15% at - 60 mV) (Fig. 3C-F). However, more specific inhibitors of PKA, Rp-cAMPS (100 μM) and protein kinase inhibitor peptide (PKI, 1 μM) had no effect on I_GS (Fig. 3B-F). Thus, we infer that I_NaL is not modulated by basal PKA activity under baseline conditions.

3.4. βAR stimulation upregulates I_NaL under AP-clamp via both CaMKII and PKA

βAR stimulation with 10 nM ISO led to a substantial increase of I_GS under AP-clamp (Fig. 4A). Surprisingly, this increase was most prominent during the early AP plateau phase at relatively positive V_m, where driving force for Na+ entry is low (Fig. 4A), The I_GS-V_m relationship below the I_GS time course for Control is fairly linear, indicating little change in conductance between −60 and +25 mV. ISO more than doubled I_GS density at +30 mV (from −0.34±0.03 A/F in control to −0.78±0.02 A/F in ISO, Fig. 4E,) and also at 0 and −30 mV (Fig. 4F,G). However, ISO did not change I_GS density at −60 mV (Fig. 4H), such that ISO shifted maximal I_GS from −60 mV in control to ~0 mV after βAR stimulation (Fig. 4A, I_GS-V_m).

Figure 4. βAR stimulation upregulates I_NaL under AP-clamp via both CaMKII and PKA signaling.

(A-D) GS-458967-sensitive current (I_Gs) traces (mean±SEM) under AP-clamp with and without isoproterenol stimulation (ISO, 10 nM) in control (A, cycling Ca²+), in the presence of 10 mM BAPTA in the pipette solution (B, buffered Ca²+), in the presence of CaMKII inhibitor KN-93 (C, 1 μM) and in the presence of the PKA inhibitor H-89 (D, 1 μM). Lower panels show the corresponding I-V trajectories of I_GS under AP-clamp. (E-H) I_GS density at different voltages during AP repolarization. ISO significantly upregulated I_GS throughout the AP, except for −60 mV in control and in H-89. Ca²+-buffering using BAPTA decreased basal I_GS at phase 3 of the AP, but not during the plateau phase. However, ISO increased I_GS in BAPTA under the AP plateau similarly as is control. ISO also increased I_GS in KN-93, but the increase in I_GS during the plateau phase of the AP was significantly reduced compared to control. Interestingly, ISO increased I_GS also at −60 mV both in BAPTA and in KN-93. H-89 pretreatment abolished the ISO-induced increase in I_GS at the early plateau phase, and significantly diminished the effect of ISO on I_GS at phase 3 of the AP. Columns and bars represent mean±SEM. Asterisks denote significant difference using two-way ANOVA with Bonferroni posttest. *p<0.05, **p<0.01, ***p<0.001.

Next, we examined how buffering [Ca²+]_i with 10 mM BAPTA affects I_NaL density and profile under AP-clamp (Fig. 4B). BAPTA did not change I_GS significantly at positive V_m (Red vs. Black in Fig. 4E,F), but reduced I_GS significantly at −30 mV and even more at −60 mV (Fig. 4G,H), similar to the effects of CaMKII inhibition (Fig. 3A vs. BAPTA in Figs. 1D and3B). This is consistent with the baseline CaMKII-dependence of I_NaL being due to regular 2 Hz Ca²+ transients at baseline (Fig 3A; absent with 10 mM BAPTA). ISO still enhanced I_GS with BAPTA, but mainly at +30 and 0 mV and much less at negative V_m (−30 and −60 mV) where CaMKII effects were largest (Black/gray vs Red/pink in Fig. 4E- H).

To dissect the contributions of PKA and CaMKII to the PAR-induced I_NaL enhancement, we pretreated cells with KN-93 (1 μM) to inhibit CaMKII or H-89 (1 μM) to inhibit PKA. In KN-93 pretreated cells, ISO significantly increased I_GS, but the magnitude of ISO effect was significantly reduced compared to control (Fig. 4C). Qualitatively, the ISO-induced I_GS in KN-93 resembled that in BAPTA, with larger increases at positive V_m (vs. near −60 mV; Fig. 4B vs. Fig. 4C). In contrast, the ISO-induced I_GS enhancement in the early plateau (at +30 mV) was completely abolished by PKA inhibition by H-89 pretreatment (Fig. 4D, E). Nonetheless, ISO still increased I_GS later during the AP (more negative Vm), with the largest difference near V_m= −30 mV (consistent with CaMKII predominance after PKA inhibition).

3.5. CaMKII and PKA differentially modulates I_NaL in different AP phases

So far, we infer that ISO increases I_NaL via PKA early in the AP plateau, but via CaMKII later in the AP plateau. However, both H-89 and KN-93 might have off-target effects [35]. So we repeated the above experiments with additional PKA and CaMKII inhibitors that differ molecularly and may be more selective. First, we pretreated myocytes with the selective CaMKII inhibitory peptide, AIP (1 μM) that inhibits CaMKII both upon Ca²⁺/CaM-activation and also when CaMKII is autonomously activated (via autophosphorylation, oxidation or S-nitrosylation). ISO still increased I_GS in AIP-treated cells at the early plateau phase (Fig. 5A, C), but smaller increases were observed at −30 mV and no change at −60 mV (Fig. 5E,F). These results agree well qualitatively with the KN-93 and BAPTA results from Fig. 4, and with CaMKII having most prominent effects on I_NaL late in the AP plateau.

Figure 5. βAR stimulation differently modulates I_NaL under AP-clamp via CaMKII and PKA.

(A-B ) GS-458967-sensitive current (I_Gs) traces (mean±SEM) under AP-clamp with and without isoproterenol stimulation (ISO, 10 nM) in the presence of highly selective peptide inhibitors of CaMKII (A, AIP, 1 μM) and PKA (B, PKI, 1 μM). Corresponding I-V trajectories are shown in the inset. (C-E) Effect of ISO on I_GS density at different voltages during AP repolarization in the presence of AIP, PKI and Rp-cAMPS (100 μM). Similar results were obtained using the selective peptide inhibitors as using KN-93 and H-89 (shown in Fig. 3), confirming the different modulation of I_NaL by CaMKII and PKI upon PAR stimulation. CaMKII inhibitor AIP significantly reduced the amplitude of ISO stimulation on I_GS. Importantly, no I_GS increase at −60 mV was observed following ISO application in AIP pretreated cells in contrast to cells pretreated with KN-93 (compare with Fig. 3H). PKA inhibitor PKI and Rp-cAMPS completely abolished the ISO effect on I_GS at +30 mV, and significantly diminished the I_GS upregulation both at 0 and −30 mV. Results in control cells are shown for comparison. The physiological intracellular Ca²+ cycling was preserved, and 2 Hz steady-state pacing was applied. Columns and bars represent mean±SEM. Asterisks denote significant difference using two-way ANOVA with Bonferroni posttest. **p<0.01, ***p<0.001.

We also blocked PKA using the specific inhibitory peptide PKI (1 μM) and cAMP analog Rp- cAMPS (100 μM, Fig. 5B-F). PKI or Rp-cAMPS completely abolished the ISO-induced enhancement in early I_GS, but did not prevent the enhanced I_GS late in the plateau (Fig. 5B,D,E). These findings agree with the H-89 results shown in Fig. 4D, and with PKA effects being most prominent in the early plateau.

These data suggest distinctive effects of PKA and CaMKII on mediating the ISO-induced enhancement of I_NaL. PKA predominantly mediates the ISO effect on increasing I_NaL during the early plateau phase of the AP, whereas CaMKII contributes in the late plateau and the phase 3 of the AP.

3.6. Effect on Epac activation and ROS signaling on I_NaL under AP-clamp

Physiologically, Epac2 (a parallel cAMP target to PKA) has been shown to mediate βAR-induced activation of CaMKII and arrhythmogenic SR Ca leak [29]. Here we tested whether the Epac-selective agonist 8-pCPT-2’-O-Me-cAMP (8-pCPT, 3 μM, Fig. 6A,B) could mimic PAR effects on I_NaL. Indeed, 8- pCPT increased I_GS during the late plateau (at 0 mV) and phase 3 of AP (at 0 and −30 mV, Fig. 6F,G), but not during the early AP plateau (at +30 mV, Fig. 6E). Importantly, BAPTA completely prevented the 8- pCPT-induced I_GS enhancement (Fig. 6C). The 8-pCPT data implicates Epac as a potential mediator of the ISO-induced increase in I_GS late in the AP, likely via activating CaMKII (rather than PKA).

Figure 6. Effect of Epac activation and ROS on I_NaL under AP-clamp.

(A) GS-458967-sensitive current (I_Gs) traces (mean±SEM) selective Epac activator 8-pCPT-2’-O-Me- cAMP (8-pCPT, 3 μM) treatment under AP-clamp with preserved Ca²+ cycling. I-V relationships of I_GS after Epac activation are shown in inset. (B) I_GS traces (mean±SEM) following pretreatment with H₂O₂ (100 μM) under preserved Ca²+ cycling. I-V relationships of I_GS after H₂O₂ application are shown in inset. (C) I_GS traces (mean±SEM) following 8-pCPT pretreatment measured with 10 mM BAPTA in the pipette solution. I-V trajectories are shown in the inset. (D) I_GS traces (mean±SEM) following H₂O₂ pretreatment measured with buffered [Ca²+]_i. I-V trajectories are shown in the inset. (E-H) I_GS density at different voltages during AP repolarization. H₂O₂, but not Epac, increased I_GS at +30 mV (E). H₂O₂ increased I_GS at 0 mV and −30 mV regardless of [Ca²+]_i. 8-pCPT increased I_GS during phase 3 of AP, but only with preserved Ca²+ cycling. (F-G) H₂O₂ increased I_GS at −60 mV in the presence of 10 mM BAPTA (H). Results in control cells are shown for comparison. Columns and bars represent mean±SEM. Asterisks denote significant difference using two-way ANOVA with Bonferroni posttest. *p<0.05, **p<0.01, ***p<0.001.

Increased production of reactive oxygen species (ROS) may also promote autonomous activation of CaMKII and pKa. We used H₂O₂ (100 μM) to examine the effect of increased ROS on I_NaL (Fig. 6B). H₂O₂ significantly increased I_GS both at early and late AP plateau phases (at +30 mV and 0 mV, Fig. 6E-F); however, only a slight increase in I_GS was observed at −30 mV (Fig. 6G) and none at −60 mV under AP-clamp when measured with preserved Ca²+ cycling (Fig. 6B). H₂O₂ caused similar effect at −30 and 0 mV under AP-clamp with BAPTA in the pipette solution (Fig. 6D,E,F). However, in this case, significant increase in I_GS was observed also at −30 mV and −60 mV (Fig. 6G-H). These results suggest that ROS might contribute to early I_NaL enhancement during the AP, but that it may also contribute to CaMKII-dependent effects, especially late component of I_NaL at −60 mV.

3.7. Effect of ROS and NOS inhibition on I_NaL modulation during PAR stimulation

To test the effect of endogenous physiological ROS production on I_NaL, we pretreated the cells with ROS scavengers (reduced glutathione, GSH, 10 mM; and N-acetyl cysteine, NAC, 10 mM). This treatment did not change basal I_GS under AP-clamp, except for a slight decrease at −60 mV (Fig. 7A,F), consistent with most of the basal CaMKII-dependent I_NaL being independent of ROS. However, ROS scavengers limited the ISO-induced I_GS enhancement. The effect was most prominent at the early plateau phase (at +30 mV, Fig. 7C), but smaller effects were observed at −30 mV (Fig. 7E). We also tested the involvement of NADPH oxidase 2 (NOX2) in mediating the ISO-induced I_NaL enhancement. The NOX2 specific inhibitor gp91ds-tat (1 pM) partially inhibited the early I_GS enhancement at +30 mV (Fig. 7A,C), but no significant limitation was found later in the AP at more negative membrane voltages (Fig. 7D-F). Thus, we conclude that endogenous ROS production, partially via NOX2, is involved in the I_GS enhancement in the early phase of the AP (the range where PKA-dependent effects were largest). However, oxidation of CaMKII may also occur upon PAR stimulation, contributing to the ISO-induced I_NaL.

Figure 7. Effect of ROS and NOS inhibition on I_NaL during PAR stimulation.

GS-458967-sensitive current (I_Gs) traces (mean±SEM) under AP-clamp with and without isoproterenol stimulation (ISO, 10 nM) in the presence of reduced glutathione (GSH, 10 mM) + N-acetyl cysteine (NAC, 10 mM), as well as NOX2 inhibitor peptide gp91ds-tat (1 μM). The corresponding I-V trajectories are shown in inset. (B) I_GS traces (mean±SEM) under AP-clamp in L-NAME (1 mM) pretreated cells with and without ISO stimulation. (C-F) I_GS density at different voltages during AP repolarization. Both NOX2 inhibition and NOS inhibition diminished the effect of ISO on I_GS at +30 mV. GSH+NAC further decreased the effect of ISO on I_GS. Results in control cells are shown for comparison. Columns and bars represent mean±SEM. Asterisks denote significant difference using two-way ANOVA with Bonferroni posttest. *p<0.05, **p<0.01, ***p<0.001.

Recent studies have also implicated myocyte nitric oxide signaling in βAR-induced CaMKII activation in myocytes [27, 39–42], so we tested whether inhibition of nitric oxide synthase (NOS) would alter ISO-induced I_NaL enhancement. Pretreatment of myocytes with non-specific NOS inhibitor L-NAME (1 mM) did not alter basal I_GS (Fig. 7B), suggesting that nitrosylation is not involved in the basal CaMKII-induced I_NaL. However, L-NAME significantly reduced the ISO-induced enhancement of I_GS during the AP plateau (Fig. 7B-D), similar to that seen with KN-93, AIP or BAPTA. These data are consistent with a significant role of NOS in mediating the ISO-induced enhancement of I_NaL, especially the CaMKII-dependent component during the plateau phase of the AP. This may be similar to the Epac2- NOS1-CaMKII5 pathway implicated in βAR-induced increase in RyR2-mediated SR Ca²+ leak [43].

4. DISCUSSION

Our study shows that βAR stimulation increases cardiac I_NaL during the physiological AP via both PKA and CaMKII signaling pathways, preferentially at more positive (early) vs. more negative (late) Vm, respectively. The CaMKII-mediated effect on I_NaL were already partially active during normal APs at 2 Hz, 36°C and normal Ca²+ transients, while PKA inhibition had no effect under these baseline physiological conditions. The CaMKII- and PKA-mediated effects on I_NaL appear to be additive and may synergize. This upregulation of I_NaL upon PAR stimulation may significantly alter the balance between the depolarizing and repolarizing currents during the plateau phase of the AP, where overall conductance is low [44]. Moreover, I_NaL during physiological APs can peak during AP repolarization, where early afterdepolarizations (EADs) arise under pathologic conditions [33, 45]. Therefore, under pathological conditions like heart failure - where both I_NaL, and CaMKII activity are elevated, and repolarization reserve is reduced [46] - PAR stimulation may increase I_NaL and lead to AP prolongation and increased arrhythmia risk.

4.1. I_NaL gating differs from I_NaT and is TTX-sensitive and GS-sensitive

Na+ channels exhibit transient openings that generate I_NaT, but during sustained depolarization can also exhibit additional openings (early bursting mode, late scattered mode) that contribute to I_NaL [1, 4]. During AP repolarization, an apparent window component of I_NaL has also been reported in well- controlled biophysical studies [5, 8, 9, 17], and a similar Na+ current was implicated in neuronal pacemaking [47, 48], despite the tiny overlap of steady state activation and availability V_m-dependence. Additionally, the slowly decreasing V_m during the cardiac AP plateau can enhance I_NaL, via a unique nonequilibrium gating scheme [6]. So, the time- and V_m-dependence of Na+ channel gating is complex, may differ widely between I_NaT and I_NaL, and only a small subset of Na+ channels may exhibit gating modes that mediate I_NaL. Furthermore, I_NaL has mostly been studied in conditions far from the physiological AP (e.g. square pulses), intracellular Ca²+ buffering and low stimulation frequency, with limited I_NaL data during physiological AP [33, 44].

Single-channel I_NaL records exhibit burst and late scattered modes of Na+ channel opening. In square V_m steps Maltsev and Undrovinas [4] reported an early larger burst mode, which declines in 50100 ms, and a smaller late scattered opening mode that declines only slightly during 200 ms. Burst mode open probability declines faster with membrane depolarization, but late scattered openings seemed less voltage-dependent. Slow V_m ramps indicated a noninactivating I_NaL (at 0 mV) which at more negative potentials resembled a window current [49, 50]. To account for the complex I_NaL V_m-dependence, we analyzed I_gs at 4 different AP voltages: (1) Early plateau (+30 mV; ~60 ms after AP peak) which likely includes I_NaL burst mode, (2) Late plateau (0 mV) where a transition to more late scattered openings are expected, (3) Early phase 3 (−30 mV) as repolarization accelerates, and (4) Rapid repolarization (- 60 mV) where driving force is rapidly increasing as V_m-dependent deactivation may be progressing. These latter two phases may reflect the non-inactivating current and window type I_NaL. These helped classify I_NaL V_m ranges that were preferentially influenced by PKA or CaMKII activity.

I_NaL measured under AP-clamp was significant throughout the AP plateau, but increased during late the plateau phase as driving force increases (Fig. 1). Using TTX and GS-458967 in AP-clamp experiments produced identical I_NaL traces, as expected if both block I_NaL (and not other currents). In contrast, fast I_NaT was significantly less affected by GS-458967 vs. TTX, in agreement with higher I_NaL selectivity reported for GS-458967 [37]. Measure of the much larger I_NaT was impractical here during physiological AP-clamp, so we used only GS-458967 to study I_NaL regulation by PKA and CaMKII.

The TTX titrations in Fig. 2 were used to identify the main Na+ channel isoforms that likely mediate I_NaL as measured here. TTX inhibited >95% of I_NaL with an IC₅₀ value of ≈1 μM, suggesting that predominantly TTX-resistant Na+ channel isoforms mediate I_NaL, including the predominant “cardiac” Na_v1.5 (but not excluding Na_v1.8 or Na_v1.9). However, TTX-sensitive Na+ channels (Na_v1.1–1.4, Na_v1.6– 1.7) are less likely to contribute to I_NaL here. This agrees with the reported IC₅₀ values of 1–2 ^M for TTX I_NaL inhibition in rabbit [51], guinea-pig [33, 52] and human [1] ventricular myocytes, although TTX- sensitive Na+ channel isoforms have been reported to contribute to I_NaL in heart failure [53]. Moreover, TTX-resistant Na_v1.8 channels have also been suggested to be expressed and contribute to I_NaL in mouse, rabbit [54] and failing human [55] ventricular myocytes. Further studies would be needed to resolve the exact contribution of different Na+ channel isoforms, splice variants and their regulation by CaMKII and PKA to I_NaL in health and disease.

4.2. Nearly half of the basal physiological I_NaL is CaMKII-dependent

We found that [Ca²+]_i and CaMKII affect the magnitude of I_gs under AP-clamp already in control (Fig. 3). Buffering [Ca²+]_i and CaMKII inhibition (either by KN-93 or AIP) strongly decreased I_gs during AP phase 3, especially between −30 and −60 mV. CaMKII inhibition also decreased I_gs earlier during the AP plateau (Figs. 3–4). Calmodulin (CaM) and CaMKII have complex effects on Na+ channel gating, with CaMKII shifting steady-state inactivation to more negative Vm, slowing inactivation, promoting intermediate inactivation and slowing recovery from inactivation, but not altering maximal conductance or activation V_m-dependence [8, 9, 56–58]. In addition to those loss of function effects CaMKII also increases myocyte I_NaL. Moreover, Ca²+/CaM alone can alter steady-state I_Na inactivation [9, 58, 59], but does not alter I_NaL [9], in agreement with our study. Importantly, we demonstrated that basal CaMKII- activity at 2 Hz pacing under physiological conditions (with endogenous Ca²⁺ levels) nearly doubles I_NaL in rabbit ventricular myocytes vs. what is seen without Ca²⁺ transients or when CaMKII is inhibited (Fig 3). This agrees with recent data in guinea pig [52]. The gradual repolarization during the AP also promotes I_NaL that is attributable to non-equilibrium gating. These factors may account for prior underestimates of physiological I_NaL when studies are done under non-physiological conditions (buffered [Ca²⁺] and square pulses).

AIP and KN-93 exerted the same effect on basal I_GS (Fig. 3), despite having different mechanisms of inhibition (KN-93 competes with CaM binding [60], while AIP mimics autoinhibition of basal and autonomous CaMKII [61]). Inhibition of ROS and NOS did not alter basal I_GS appreciably (Fig. 7) suggesting that oxidation and S-nitrosylation that are known to promote autonomous CaMKII [26, 27] are not required for the basal CaMKII effect on I_NaL. In marked contrast, basal PKA activity did not contribute to I_GS under AP-clamp (Fig. 3) in agreement with previous square pulse studies [9].

4.3. βAR-induced I_NaL is dependent on both PKA and CaMKII signaling

Importantly, βAR stimulation upregulated I_GS under AP-clamp and both PKA and CaMKII are required for the full effect (Figs. 4–5). Notably, PKA and CaMKII affected I_GS predominantly in different phases of the AP. I_NaL enhancement early in the AP plateau (+30 mV) upon βAR stimulation was exclusively dependent upon PKA, suggesting that PKA may particularly enhance the early burst mode I_NaL openings. This is consistent with effects of a PKA-dependent long QT3-associated mutation, D1790G, that promotes early burst opening of the Na+ channel [62]. Conversely, PAR activation had no effect on I_NaL measured during rapid repolarization (−60 mV). While BAPTA and KN-93 uncovered a potential CaMKII-independent effect of ISO (Fig. 4H), this was not seen with more selective CaMKII block via AIP (Fig. 5F).

At intermediate V_m during repolarization (0 and −30 mV) the βAR-induced I_GS was progressively less PKA-dependent and more CaMKII-dependent. For some data this is hard to appreciate because KN-93, AIP and BAPTA all reduce basal I_NaL prior to ISO activation. First, we consider the AIP and PKI data in Fig. 5 and assume the ISO effect with AIP is all due to PKA and that with PKI is all due to CaMKII. The I_GS vs. V_m curves for PKA effect are superimposable from −90 to −45 mV and then split progressively despite a decrease in driving force. This indicates that PKA influences opening preferentially at more positive Vm. Conversely, these I_GS - Vm curves for CaMKII effect (with PKI) diverge already below −60 mV but start converging at 0 mV and are identical at +25–40 mV. This indicates that CaMKII promotes I_NaL most strongly at negative V_m and later in the AP plateau. Using Fig. 5C-F we can also infer that the PAR-induced I_NaL increase is entirely PKA-dependent at +30 mV, and declines to 65% and 57% during the plateau (0 and −30 mV) and becomes entirely CaMKII- dependent between −30 and −60 mV. We speculate that PKA promotes preferentially the burst mode, while CaMKII promotes the late scattered openings seen at the single channel level. Further study will be required to test this speculation and also to identify specific amino acids phosphorylated by PKA and CaMKII in this process (and dozens of candidate sites exist [63–66]). Key candidates on Na_V1.5 could include Ser⁵²⁵ and Ser⁵²⁸ for PKA vs. Ser⁵¹⁶ and Ser⁵⁷¹ for CaMKII [64–66].

The CaMKII-dependent activation of I_NaL with ISO was partially dependent on NO production and was mimicked by direct Epac activation (Figs. 6A and7B). This is reminiscent of the recently elucidated pathway by which βAR activates RyR2 and SR Ca leak, mediated by cAMP-dependent Epac2 activation which causes NOS1-dependent S-nitrosylation/activation of CaMKIIS to phosphorylate RyR2 [24, 27, 29, 39, 43]. So that same pathway may impact Na+ channels as well. The I_NaL enhancement with ISO was also partially dependent on ROS and NOX2, especially in the early AP phase where PKA- dependent effects were strongest (Fig. 7A). This might reflect some ROS-dependent modulation of the PAR-PKA-I_NaL pathway, but our data do not resolve a molecular mechanism for such an effect. Angiotensin II was reported to induce PKA-dependent enhancement of I_NaT via NOX2-mediated ROS production [32]. In that study, the angiotensin-II and ROS induced I_NaL was attributed to CaMKII vs. PKA, but I_NaL was measured only late in square pulses which might have favored detection of CaMKII as the mediator. Since oxidation may lead to autonomous activation of both CaMKII and PKA, further studies are needed to better clarify the upstream signaling whereby ROS leads to I_NaL enhancement during βAR stimulation.

4.4. Physiological magnitude of I_NaL in rabbit ventricular myocytes

The magnitude of I_NaL and the contributions of multiple Na+ channel gating components have become a greater focus in cardiac research, because altered Na+ channel gating has been linked to abnormal AP activities and disturbed cellular Na+ homeostasis [2, 67]. Here we examine for the first time the detailed time course of I_NaL under physiological conditions in response to βAR stimulation. Comparing I_NaL amplitude with other studies is complicated by different conditions, which have often been done at sub- physiological temperature, with square pulses at single V_m (vs. AP-clamp), measured only at end of pulse, and non-physiological intracellular solutions (including Ca²+ buffering).

We measured peak I_NaL at −30 mV under AP-clamp as −0.44±0.02 A/F (~65 pA) with preserved Ca²+ cycling, −0.31±0.01 with CaMKII inhibition and −0.33±0.02 A/F (~50 pA) with 10 mM BAPTA (Fig. 4). These values agree well with prior studies in rabbit ventricular myocytes (between 0.25 and 0.4 A/F at −20/−30 mV with strong Ca²+ buffering [10, 68–70]. Clamping [Ca²+]_i at 600–1,000 nM in rabbit myocytes also increased I_NaL to 0.45 A/F and 0.55 A/F, respectively [70], and similar effects of [Ca²+]_i-dependence of I_NaL were seen in canine cardiomyocytes [58]. Human ventricular myocytes from healthy donors had similar I_NaL magnitude [1, 12], but significantly larger I_NaL was reported in guinea-pig [33, 71] and rat [50, 72] ventricular myocytes. Other studies compared I_NaL magnitude to I_NaT peak density and I_NaL was reported to be 0.15–0.3% of I_NaT [8, 9]. We could not directly measure peak I_NaT in our physiological AP-clamp, but prior estimates in rabbit ventricular myocytes at 36°C [73] gave peak I_NaT in the range of −395 to −438±27 A/F. Thus, our I_NaL peak density is ≈0.13% of peak I_NaT during physiological AP in rabbits.

4.5. Study Limitations

We used pharmacological agents in freshly isolated adult rabbit ventricular myocytes, because we sought to use an animal with AP plateau phase resembling the human cardiac AP. This made it impractical to use genetically modified mice (or rabbits) to knockout CaMKII or PKA, NOS, Epac, NOX2. Even gene silencing in culture for 2–3 days can significantly alter ion channels and signaling pathways [74]. To minimize the inherent limitations of small molecule inhibitors, we used multiple agents with different structure and mechanism of action wherever practical to confirm target effects. For example, we used TTX to verify the utility of using 1 GS-458967 to measure I_NaL, AIP and KN-93 to inhibit CaMKII and H-89, PKI and Rp-cAMPS to inhibit PKA in experiments leading to major conclusions. Another limitation arises from the complexity of PAR signaling and its crosstalk with other signaling pathways. Because of these potential complications, we were cautious not to combine two or more inhibitors to further isolate signaling pathway.

4.6. Conclusions

In summary, our data reveal that basal I_NaL during the physiological AP and Ca²+ transients at 2 Hz is already boosted significantly by the basal level CaMKII activity. PAR stimulation further increases I_NaL and is mediated in concert by both PKA-dependent and CaMKII-dependent pathways. The PKA- dependent increase in I_NaL is predominantly in the early AP plateau (at more positive V_m), whereas CaMKII mainly increases I_NaL during the late AP plateau and rapid repolarization phase. The CaMKII- dependent effects appear to involve Epac and NOS signaling. Both CaMKII- and PKA-dependent effects might also include ROS signaling. Furthermore, there is a synergistic crosstalk between PKA and CaMKII signaling that further promotes ßAR-induced I_NaL. However, the precise identification of ROS and crosstalk require further studies. Taken together, our data reveal differential modulations of I_NaL at different AP phases by PKA and CaMKII pathways following β-adrenergic stimulation. This comprehensive and nuanced view on the fine-tuning of I_NaL during different AP phases deepens our understanding of the role of I_NaL in shaping cardiac AP and arrhythmogenic potentials, which will inform therapeutic development for treating arrhythmias in heart diseases whereby increased sympathetic tone, increased CaMKII-activity and oxidative stress are present under pathological conditions [11, 75, 76].

• Measure the dynamic profile of I_NaL under cardiac AP during β-adrenergic stimulation

• Reveal the differential contributions of PKA and CaMKII pathways in modulating I_NaL

• Pacing-induced CaMKII activation upregulates basal I_NaL under AP-clamp

• Both PKA and CaMKII upregulate I_NaL during β-adrenergic stimulation

• PKA increases I_NaL in phase 2 and CaMKII increases I_NaL in phase 3 of the AP

• Reactive oxygen species and nitric oxide may contribute to PKA and CaMKII effects

Abbreviations:

AIP

Autocamtide-2-related inhibitory peptide

AP

Action potential

βAR

β-adrenergic receptor

CaM

Calmodulin

CaMKII

Ca²⁺/calmodulin-dependent protein kinase II

cAMP

Adenosine-3’,5’-cyclic monophosphate

dV/dt_max

Maximal upstroke velocity of AP

GS

GS-458967, late Na⁺ current inhibitor

GSH

Reduced glutathione

Epac

Exchange protein directly activated by cAMP

I_GS

GS-458967-sensitive current

I_NaL

Late Na⁺ current

I_NaT

Transient Na⁺ current

I_TTX

Tetrodotoxin-sensitive current

ISO

Isoproterenol

L-NAME

Nω-nitro-L-arginine methyl ester

NAC

N-acetyl cysteine

NOS

Nitric oxide synthase

NOX2

NADPH oxidase 2

PKA

Protein kinase A

PKI

Protein kinase inhibitor peptide

ROS

Reactive oxygen species

Rp-cAMPS

Rp-adenosine-3’,5’-cyclic monophosphorothioate

TTX

Tetrodotoxin