Introduction

Catecholaminergic polymorphic ventricular tachycardia (CPVT) is an inherited arrhythmic syndrome characterized by episodic syncope and/or sudden cardiac arrest, typically triggered by adrenergic stimulation as occurs during strenuous exercise or emotional stress [1, 34, 46]. CPVT has been associated with mutations in various Ca2+ handling proteins, which lead to abnormalities in Ca2+ homeostasis; most notably, these mutations are found in the cardiac ryanodine receptor-Ca2+ release channel (RyR2) [25, 35] and the sarcoplasmic reticulum (SR) Ca2+-binding protein calsequestrin 2 (CASQ2) [13, 19]. Ca2+-dependent calmodulin missense mutations also occur in a small number of cases [18]. Some RyR2 mutations also predispose to atrial arrhythmias [4, 34, 39]. For example, the RyR2-P2328S mutation is associated with high incidences of both CPVT and atrial tachycardia (AT) [25, 37, 40]. This RyR2 variant has been associated with a normal luminal SR Ca2+ release sensitivity but an increased sensitivity to cytosolic Ca2+ [31], giving rise to lower cytosolic Ca2+ thresholds leading to Ca2+ release. If reached during increased heart rates, these could be sufficient to elicit a ‘leak’ of SR Ca2+ consequently triggering arrhythmia.

The atrial and ventricular phenotypes are replicated by the RyR2-P2328S (RyR2 S/S) murine model which shows potential arrhythmic triggers in the form of delayed after-depolarizations [15, 22]. However, there remains a requirement for an electrophysiological tissue substrate in order to perpetuate and sustain arrhythmia, which has previously been typified by a reduced conduction velocity (θ) in systems showing reduced Nav1.5 expression [33], connexin 40 and/or 43 (atria) or connexin 43 (ventricle) expression [16, 23], or structural abnormalities, including fibrosis [44]. Interestingly, several reports have indicated the potential for Nav1.5 expression [6, 11, 42] and function [2, 41] to be modulated, both directly and indirectly, by alterations in cytosolic Ca2+. Rat cardiomyocytes showed reductions in Na+ channel activity following imposed increases of intracellular [Ca2+]. Additionally, the Ca2+ channel blocker verapamil and the Ca2+ ionophore calcimycin increased and decreased Nav1.5 mRNA and Nav1.5 protein expression respectively [11, 42]. In agreement with these findings, elevations and reductions of cytosolic [Ca2+], by chronic treatment with high extracellular [Ca2+] and [K+] or BAPTA-AM, decreased and increased Na+ current densities, respectively [6]. More recently, Casini et al. [5] demonstrated that acute increases in pipette [Ca2+] were capable of reducing both Na+ current density and (dV/dt)max.

Biochemical evidence accounting for the potential mechanisms of functional modulation of Nav1.5 by cytosolic [Ca2+] identifies both direct and indirect Ca2+ binding sites on Nav1.5. Direct Ca2+ binding to Nav1.5 is mediated at an EF hand motif resident at the carboxy-terminal region of Nav1.5 [47]. This binding results in a depolarizing shift of the voltage dependence of Na+ channel inactivation with a potential increase in Na+ channel activity [47]. Indirect mechanisms of Ca2+ binding have been attributed to both the presence of an additional binding site, the ‘IQ’ domain, within the C-terminal region of Nav1.5 for Ca2+/Calmodulin (Ca2+/CaM) and multiple phosphorylatable sites (including serines 516 and 571 and threonine 594) within the IDI-II linker region of Nav1.5 targeted by Ca2+/CaM Kinase II (CaMKII). These two mechanisms occur only subsequent to Ca2+ binding to the EF hand motifs of Ca2+/Calmodulin (Ca2+/CaM) or Ca2+/CaM Kinase II (CaMKII) and thus constitutes an indirect interaction of Ca2+ with Nav1.5. Ca2+/CaM binding at the IQ domain and CaMKII-dependent phosphorylation shifts Na+ current availability to a more depolarized membrane potential [2] and enhances slow inactivation of the Na+ current [41].

Recent reports have indeed implicated reduced Nav1.5 expression and Na+ channel function in the increased arrhythmogenicity in RyR2 S/S atria [21, 22, 36]. They also demonstrated a reduced conduction velocity [22], resulting from a reduced Na+ current attributable either to a reduced Nav1.5 expression or the direct inhibitory effect on Na+ channel function of altered Ca2+ homeostasis outlined previously. Slowed conduction resulting from reduced Nav1.5 expression would potentially produce arrhythmogenic substrate, which would compound the arrhythmic effect of Ca2+-mediated triggered activity in the RyR2 S/S [21, 22, 36].

The present study investigates for possible roles of Cx43 expression as well as Nav1.5 expression and function in RyR2 S/S ventricular, as opposed to atrial, arrhythmogenicity. First, the arrhythmogenic properties of the RyR2 S/S ventricle compared to the WT was confirmed in accordance with earlier reports [15] and correlated with measurements of action potential duration (APD), conduction velocity (θ) and their alternans, as well as ventricular effective refractory period (VERP). The stimulation protocols either interposed extrasystolic, S2, stimuli at progressively decremented S1S2 intervals within 8 Hz S1 pulse trains or applied steady stimulus frequencies at progressively decremented basic cycle lengths (BCLs). Second, to assess the underlying mechanism for the slowed conduction and arrhythmic phenotype, we assessed the ventricular expression of Cx43 and Nav1.5, the latter assessed in both the whole ventricle and the membrane fraction compared between WT and RyR2 S/S hearts. Third, the corresponding functional evaluation of Nav1.5 was investigated through peak I Na current recordings of WT and RyR2 S/S ventricular tissue. These comparisons successfully correlated Nav1.5 expression and function, particularly within the functional Nav1.5 containing membrane fraction, with the incidence of ventricular arrhythmia, and the resulting conduction changes in RyR2 S/S ventricles.

Materials and methods

Experimental animals

Homozygous RyR2 S/S and WT mice (aged 4 to 6 months) with an inbred 129/Sv genetic background (supplied initially by Harlan, UK) were generated as described previously [15]. Mice were kept in plastic cages at room temperature in an animal facility, given free access to sterile rodent chow and water and subjected to 12 h light/dark cycles. Mice were killed by cervical dislocation for experimentation. All procedures conformed to the UK Animals (Scientific Procedures) Act 1986, approved by a university ethics review board. Hearts were rapidly excised and submerged in ice-cold Krebs-Henseleit (KH) buffer solution (containing, in mM, NaCl 119, NaHCO3 25, KCl 4, KH2PO4 1.2, MgCl2 1, CaCl2 1.8, glucose 10 and Na-pyruvate 2, pH 7.4, 95 % O2/ 5 % CO2; British Oxygen Company, Manchester, UK) for whole heart electrophysiological studies and Western blot analyses. All chemicals were purchased from Sigma-Aldrich (Poole, Dorset, UK), unless otherwise stated. Six WT and seven RyR2 S/S mice were used in whole heart electrophysiological investigations. Four WT and four RyR2S/S hearts were used for Western blot studies of Cx43 expression. Seven WT and six RyR2 S/S mice were used for Western blot studies of Nav1.5 expression in the whole tissue and membrane fraction. Four WT and three RyR2 S/S mice were used to give n = 6 and 12 patches, respectively, for loose patch-clamp studies of Na+ currents. Male and female mice were used in statistically equal numbers in each group.

Electrophysiological studies in whole heart

Excised hearts were cannulated and retrograde perfused on a Langendorff system as previously described [28, 29, 38, 48]. Prior to electrophysiological testing, hearts were perfused with KH solution for at least 5 min to achieve a steady state. Monophasic action potentials (MAPs) were recorded by microMAP electrodes (HugoSachs, Harvard Apparatus, UK) placed upon the epicardial surface. Recordings were amplified (Neurolog NL100 preamplifier; NL104 amplifier, Digitimer, Welwyn Garden City, UK), band-pass filtered (NL125/126 filter; 0.5 Hz to 1.0 kHz) and sampled at 5 kHz (micro1401 interface) for display using Spike2 software (Cambridge Electronic Design, Cambridge, UK).

Hearts were paced at twice their excitation threshold voltage using a bipolar platinum-coated stimulating electrode placed on the ventricular septum connected to a DS2A-Mk.II stimulator (Digitimer). After pacing at 8 Hz for at least 5 min to attain and confirm a steady state, two types of pacing protocols were applied. A S1S2 protocol first stimulated hearts at frequencies of 10 Hz for 20 s; this was followed by cycles of drive trains of eight S1 beats delivered at 8 Hz followed by an S2 extra-stimulus, at S1-S2 coupling interval successively reduced by 1 ms with each subsequent cycle until either 2:1 block or sustained arrhythmia occurred. A dynamic pacing protocol [24, 28, 29] first assessed action potential (AP) properties at a BCLs of 134 ms duration for 100 stimulations. The BCL was then decremented by 5 ms, and the pacing sequence repeated until the hearts showed either entry into 2:1 block or sustained arrhythmia. Both stimulation protocols yielded incidences of arrhythmia. The S1S2 protocol additionally provided VERPs. The dynamic pacing protocol yielded APDs, and an indication of conduction velocity θ’ (1/latency, which was measured as the time from the stimuli to the peak of the MAP) as a function of BCL, measured as the recovery time from AP peak to 90 % full repolarization, APD90 (Fig. 1). The corresponding diastolic intervals (DIs) were calculated from the BCL and APD90 values using the relationship:

Fig. 1
figure 1

Two typical successive monophasic action potential (MAP) recordings at the LV epicardium of a WT heart obtained during dynamic pacing to highlight the derivation of the various parameters used for analysis; BCL, latency, APD90 and DI. BCL was the time interval between the adjacent stimuli, thus the pacing rate. Latency was measured as the time elapsed from the stimuli to the peak of MAP. APD90 is the time course over which 90 % repolarization of the MAPs obtained. DI was measured as BCL-APD90, thus comprising the final 10 % of MAP repolarization and up to the start of the next stimuli

Full size image
$$ \mathrm{D}\mathrm{I}=\mathrm{B}\mathrm{C}\mathrm{L}-{\mathrm{APD}}_{90} $$

Protein extraction and Western blot analysis

The method of protein extraction was optimized for the protein of interest. For connexin proteins, which are primarily found in the surface membrane in hexameric clusters, a well-established high content sodium dodecyl sulphate (SDS) buffer [7, 12, 43] was used in order to fully solubilize the membrane and maximize release of the connexin proteins from the gap junction channels in the plasma membrane. For Nav1.5 channels, we chose to use a milder buffer, followed by a centrifugation step and solubilization in order to separate out the membrane and non-membrane fractions based on the different distribution and abundance of Nav1.5 channels and their contribution to conduction.

Ventricular tissues of the excised hearts were snap-frozen and crushed into powder by a clamp pre-cooled with liquid N2. The powdered tissue was homogenized in either SB20 (20 % SDS, 2 mM EDTA, 150 mM Tris) [7, 12, 43] and diluted to an appropriate gel loading concentration in sample loading buffer (90 % SB20, 5 % 2-mercaptoethanol, 5 % w/v bromophenol blue) for connexin 43 analysis, or the re-suspension buffer (containing, in mM, Tris–HCl 50, NaCl 10, Sucrose 320, EDTA 5, EGTA 2.5 and Protease inhibitors (1 tablet/20 ml; Roche, West Sussex, UK), pH 7.4) and lysis buffer (containing, in mM, Tris–HCl 20, EDTA 2, NaCl 137 and Triton X-100 1 %, glycerol 10 %, pH 7.4) and then centrifuged for 15 min at 13,000g and 4 °C for Nav1.5 analysis. The supernatant was divided into two parts. One was stored at −80 °C as the whole tissue fraction and the rest was centrifuged at 100,000g at 4 °C for an hour to extract the membrane proteins. The pellet was re-suspended in radioimmunoprecipitation assay (RIPA) lysis buffer and then vortexed and placed on ice for 30 min to harvest the membrane protein.

For Western blot analysis, the protein extracts from whole tissue and the membrane fraction were separated on premade 4–12 % Bis-Tris Gels (Invitrogen, Paisley, UK) and then transferred onto PVDF membranes (Immobilon-P, Millipore, Hertfordshire, UK). Blots were blocked in 5 % skimmed milk in TBST (Tris-buffered saline; Invitrogen, Paisley, UK and Tween 20) and then probed with either anti-Nav1.5 (rabbit anti-mouse IgG, 1:1000, Alomone, Jeruselam, Israel) or anti-Cx43 (rabbit anti-mouse, 1:10,000, C6219 Sigma-Aldrich) and anti-GAPDH (Abcam, Cambridge, UK) or anti-α-tubulin (Cell Signaling Technology, Danvers, MA, USA) antibodies overnight at 4 °C. Horseradish peroxidase (HRP)-conjugated secondary antibody (Abcam, 1:10000 to 1:50000) were detected using an enhanced chemiluminescent system (GE Healthcare, Little Chalfont, Bucks, UK). Specific protein bands were quantified using Image J (National Institutes of Health, Wash., USA).

Loose patch clamp recordings of I Na in ventricular tissue

Loose patch clamp experiments previously described for atrial tissue [21, 36] were performed in a right ventricular tissue preparation. This technique allows for measurement of Na+ currents in whole, perfused ventricular tissue preparations where intracellular Ca2+ homeostasis is not disrupted by cell isolation. Furthermore, the tissue preparation allows a more reliable comparison with APD90 and CV measurements of this study. Activation properties were assessed in order to determine peak I Na. The activation protocol utilized a series of 75 ms duration depolarizing pulses, incremented by 10 mv steps ranging from 20 to 120 mV excursions applied 5 ms after the beginning of the sampling period using a P/4 pulse protocol [3].

Statistical analysis

Statistical analysis for differences between experimental groups was performed using Graphpad Prism software (La Jolla, CA 92037 USA), applying unpaired Student’s t tests. A value of P < 0.05 was considered statistically significant. Data are presented as means ± SEM.

Results

Comparison of ventricular arrhythmogenicity in S1S2 protocols and dynamic protocols

We initially confirmed the arrhythmogenic phenotype of the RyR2 S/S murine heart as previously reported [15]. An S1S2 stimulation protocol was used to determine the incidence, frequency and duration of ventricular arrhythmia in isolated Langendorff-perfused hearts.

The occurrence of ventricular tachycardia (VT) was defined as an occurrence of two or more sequential spontaneous APs, as in previous work [36] in the course of programmed electrical stimulation. Figure 2 illustrates representative left ventricular epicardial traces from WT (A, B) and RyR2 S/S (C-F) hearts, displaying regular activity (A), ectopic (B) and VT (C-E) and ventricular fibrillation (VF) (F) episodes during the S1S2 protocol. Six WT and seven RyR2 S/S hearts were subject to the S1S2 protocol described in methods. None of the WT hearts showed VT, although one heart showed three separate ectopic events (at S1S2 intervals of 33, 31 and 30 ms). In contrast, the RyR S/S hearts showed 30 episodes of arrhythmia, taking the form of either polymorphic or monomorphic VT in three of the hearts. Of these hearts, the first showed 7 episodes of VT, lasting approximately 2.8 s, with an additional VT episode that degenerated into VF lasting approximately 22.5 s. The second heart showed one episode of VT lasting approximately 11.2 s. The third heart showed 21 episodes of VT accounting for a total time of approximately 14 s.

Fig. 2
figure 2

MAP traces obtained from the epicardium of the left ventricle in WT and RyR2 S/S hearts during S1S2 pacing highlighting typical traces of either normal activity or arrhythmogenesis. All WT hearts entered the refractory period without displaying any episodes of arrhythmia, as defined by two or more non-stimulated APs (a); however, one heart displayed the occurrence of a singular ectopic (one non-stimulated AP) (b). Multiple arrhythmic events were observed in RyR2 S/S hearts including short non-sustained ventricular tachycardia (NSVT) (c), polymorphic tachycardia following a previously imposed S2 extrastimulus (d), monomorphic ventricular tachycardia (VT) (e) and episodes of VT which deteriorated into ventricular fibrillation (VF) (F). The small black circles indicate the timing of stimuli

Full size image

Ventricular effective refractory periods (VERPs) were defined as the time period during which the myocardium is incapable of re-excitation in response to the twice-threshold stimulus employed in the S1S2 protocols [10, 32]. It was thus the S1S2 interval at which loss of S2 capture first occurred in an absence of VT. WT and RyR2 S/S hearts typically became refractory at similar S1S2 pacing intervals (VERP: WT: 38.2 ± 1.55 ms (n = 6); RyR2 S/S: 37.5 ± 5.04 ms (n = 5); P = 0.9057). VERP could not be ascertained from all the mice studied due to arrhythmogenesis: sustained arrhythmias occurred in two of the seven RyR2 S/S hearts during the S1S2 protocol.

The differing arrhythmic properties of WT and RyR2 S/S were further confirmed in the dynamic pacing protocol, which subjected hearts to systematically decreasing BCLs. Two of the six WT hearts showed VT at BCLs of 39 and 44 ms. However, these correspond to BCLs which are substantially lower, thus a much higher heart rate than those experienced during normal physiological maximal exercise [8]. RyR2 S/S hearts not only demonstrated higher incidences of VT and VF but they occurred also at higher BCLs than WT (54, 64, 54 and 74 ms in four RyR2 S/S hearts respectively; these necessitated termination of the protocol), suggesting a reduced capacity to tolerate increased heart rates as may be observed during emotional or physical stress such as exercise [8].

Action potential properties and conduction at varying pacing rates

AP propagation and recovery properties at different BCLs were then investigated using the dynamic pacing protocol. Figure 3 illustrates typical APs thus obtained from the LV epicardium of WT (left column) and RyR2 S/S hearts (right column). In both cases, AP amplitude decreased with increasing pacing rate in every heart, independent of genotype (Fig. 3a–e). At lower pacing rates, RyR2 S/S hearts showed a higher incidence of alternans (Fig. 3a, right column) compared with WT hearts (Fig. 3c, left column). Around half of both the WT and the RyR2 S/S hearts had shown either a loss of capture or arrhythmogenesis when the BCL reached 54 ms (Fig. 3e).

Fig. 3
figure 3

Typical MAP recordings obtained from the left ventricular epicardium of WT and RyR2 S/S during dynamic pacing. Traces from WT (left) and RyR2 S/S at progressively decreasing BCLs: 124 (a), 99 (b), 84 (c), 74 (d) and 54 ms (e). If a heart entered 2:1 block, the protocol was terminated (E). Traces are displayed along a common horizontal timescale. The vertical scale was normalized to a standard AP deflection at a BCL of 134 ms. Small black circles above each trace indicate the timing of stimuli

Full size image

Figure 4 plots averaged (mean ± SEM) APD90 (A, C) and θ’ (=1/latency) (B, D) values in WT (filled symbols) and RyR2 S/S (open symbols) hearts against BCL (A, B) and DI (C, D). At BCLs, where significant differences between readings at the two genotypes were obtained, this is indicated (*P < 0.05; **P < 0.01). Both WT and RyR2 S/S showed similar (P > 0.05) values of APD90 at each BCL and DI. These both declined with decreasing BCL and DI. RyR2 S/S and WT hearts additionally showed 2:1 block at similar values of BCL (WT 56.5 ± 5.95 ms, n = 4; RyR2 S/S: 54 ± 2.5 ms, n = 3; P = 0.751). As with VERP for the S1S2 protocol, 2:1 block was not obtained from all the mice studied; this was due to arrhythmogenesis warranting termination of the dynamic pacing protocol in four of the seven RyR2 S/S hearts and two of six WT hearts. In contrast to similar APD90, at equivalent BCLs, RyR2 S/S hearts showed consistently lower θ ’ than WT hearts at equivalent BCLs. Indeed, the highest mean θ’ showed by the RyR2 S/S (0.043 ± 0.003 m s−1), which was observed at the highest BCL, was similar to the lowest θ’ (0.042 ± 0.006 m s−1) shown by the WT, which was observed at the shortest BCLs. These findings were corroborated when the APD90 and θ’ values were plotted against their preceding DIs reflecting recovery times from the preceding APs (C, D). The present findings demonstrate normal AP repolarization characteristics, but compromised AP conduction in the RyR2 S/S arrhythmic phenotype, which could arise from abnormalities in gap junction channels and/or Nav1.5.

Fig. 4
figure 4

Plots of APD90 and θ’ at different BCLs and DIs in WT and RyR2 S/S hearts. a, b Mean (± SEM) values for APD90 and θ’ respectively at different BCLs (134, 129, 124, 119, 114, 109, 104, 99, 94, 89, 84, 79, 74 and 69 ms) for WT (n = 6, filled symbols) and RyR2 S/S (n = 7, open symbols) hearts. c, d Mean (± SEM) values for APD90 and θ’, respectively, at different DIs for WT (n = 6, filled symbols) and RyR2 S/S (n = 7, open symbols) hearts. APD90 is virtually superimposable at all BCLs between WT and RyR2 S/S hearts and thus shows no statistically significant variation. However, θ’ is consistently lower in RyR2 S/S hearts as compared to WT hearts with significant differences between the genotypes denoted by asterisks; *P < 0.05 and **P < 0.01

Full size image

Alternans of electrophysiological parameters, reflecting temporal variability, often presages arrhythmic activity. Figure 5 assesses the average (mean ± SEM) degree of alternans in AP amplitude (A, D) [30], APD90 (B, E) and θ’ (C, F) at different BCLs (A-C) and DIs (D-F) in WT (filled symbols) and RyR2 S/S (open symbols) hearts. Alternans reflects system instability through the mean difference between alternating high and low values of a given parameter normalized to the mean value of the parameter. Both the RyR2 S/S and the WT demonstrated similarly increasing AP amplitude instabilities with either decreasing BCL or decreasing DI. RyR2 S/S and WT showed similar APD90 and θ’ instabilities which similarly varied with decreasing BCL or DI. θ’ instabilities were relatively small in contrast to the large changes in their mean values described.

Fig. 5
figure 5

Plots of alternans at different BCLs and DIs. The mean (± SEM) alternans characteristics of AP amplitude (a, d), APD90 (b, e) and θ’ (c, f) for WT (filled symbols) and RyR2 S/S (open symbols) hearts have been plotted as percentage variation between each beat as a function of BCL (a–c) and DI (d, e) AP magnitude displays an increasing degree of alternans with decreasing BCL and DI; however, this does not vary between genotypes. Similarly, a small degree of alternans is observed in the APD90 and less so the θ’ with decreasing BCL and DI, but these do not vary significantly between the WT and RyR2 S/S hearts

Full size image

These findings implicate abnormal conduction as opposed to abnormal repolarization in the RyR2 S/S ventricular arrhythmic phenotype. The underlying mechanism/cause of this abnormal conduction is thus investigated in the next sections.

Cx43 expression is comparable between the ventricles of WT and RyR2 S/S murine hearts

Abnormal cardiac conduction can arise from one of three factors: abnormal connexin expression/gap junction formation, impaired Na channel function and/or structural abnormalities such as with fibrosis or hypertrophy. Due to the structural similarity of WT and RyR2 S/S hearts [49], we pursued the remaining two factors.

We first assessed the expression of the ventricular gap junction protein, Cx43. Western blots of whole tissue ventricular lysates from WT and RyR2 S/S hearts demonstrate that the overall expression of Cx43, normalized to the housekeeping gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH), was not significantly different between WT and RyR2 S/S ventricles (0.59 ± 0.07; n = 4 and 0.79 ± 0.1; n = 4, respectively; P > 0.05, Fig. 6). This suggests that a loss of Cx43 expression is not a contributory factor to the slowed ventricular conduction and increased arrhythmogenesis observed in the ventricles of RyR2 S/S hearts, in parallel to findings in the atria of the same model [21].

Fig. 6
figure 6

Total Cx43 expression in WT and RyR2 S/S ventricles. a Representative blots of Cx43 and GAPDH expression in WT and RyR2 S/S ventricular tissue and b the mean (± SEM) Cx43 expression normalized to GAPDH expression. Cx43 expression was similar between WT and RyR2 S/S (0.59 ± 0.07 and 0.79 ± 0.1, respectively; P = 0.167, n = 4) ventricles

Full size image

Decreased Nav1.5 expression in the ventricles of RyR2 S/S murine hearts

Western blots of WT and RyR2 S/S ventricular homogenates (Fig. 7) illustrate a decreased Nav1.5 expression in the RyR2 S/S relative to the WT in both the whole tissue fraction (Fig. 7a; left panel) (1.17 ± 0.2; n = 6, and 1.69 ± 0.15; n = 7, respectively; P < 0.05) and within the membrane fraction (Fig. 7a; right panel) (2.06 ± 0.33; n = 4, and 0.91 ± 0.13; n = 4, respectively; P < 0.01). Thus, Nav1.5 expression in RyR2 S/S ventricles was approximately 69 % of that seen in the WT whole tissue fraction and down to only 44 % of WT in the membrane fraction (Fig. 7b). This significant reduction of Nav1.5 expression in the ventricular membrane where the function of Nav1.5 channels is crucial to cardiac excitability, and conduction would be expected to lead to a significant reduction in I Na in the RyR2 S/S heart compared to the WT.

Fig. 7
figure 7

Western blots of Nav1.5 expression in whole tissue and membrane fraction samples from WT and RyR2 S/S ventricles. Ventricular Nav1.5 expression was decreased in RyR2 S/S compared to WT, both in the whole tissue (1.17 ± 0.20; n = 6, vs 1.69 ± 0.15 n = 7, respectively, P = 0.048) and in the membrane fraction (0.91 ± 0.13; n = 4, vs 2.06 ± 0.33; n = 4, respectively, P = 0.006). This suggested a greater proportional reduction in membrane relative to total Nav1.5 expression in RyR2 S/S. Symbols denote significant differences between genotypes *P < 0.05, **P < 0.01

Full size image

Decreased I Na in the ventricles of RyR2S/S murine hearts

To assess whether the reduced expression of Nav1.5 in RyR2 S/S ventricles correlated with a functional alteration of Nav1.5, we measured I Na in both WT and RyR2 S/S ventricles using the loose patch clamp technique. Figure 8a illustrates representative currents elicited by WT and RyR2 S/S ventricles following a series (20–120 mV voltage excursions) of depolarizing test pulses. The peak current elicited at each voltage excursion and the overall peak current for both WT and RyR2 S/S ventricles are shown in Fig. 8b, c, respectively. Currents recorded from the WT ventricle were significantly larger than those recorded in the RyR2 S/S ventricle at depolarizing pulses of 60 mV or greater (P < 0.01). The overall peak current in the RyR2 S/S was −14.45 nA ± 0.88 nA while in the WT it was −21.3 nA ± 1.87 nA (P < 0.01); this equates to a 32 % reduction in peak I Na in the RyR2 S/S.

Fig. 8
figure 8

Loose patch clamp recordings of I Na activation in WT and RyR2 S/S ventricles. a Representative currents in response to depolarizing steps increased from 20 to 120 mV in voltage-clamped WT and RyR2 S/S ventricular tissue. b Peak inward current (mean ± SEM) elicited at each voltage step for WT (n = 6) and RyR2 S/S (n = 12) ventricles. c The maximum current recorded during each complete voltage step protocol (mean ± SEM) was larger in the WT than the RyR2 S/S ventricles, P < 0.0047. The asterisks denote significant differences between genotypes of P < 0.01

Full size image

Discussion

The present experiments demonstrate that reduced Nav1.5 expression and Na current is associated with the reduced conduction velocity and consequent arrhythmic substrate and ventricular arrhythmogenesis in homozygotic murine RyR2-P2328S (RyR2 S/S) hearts. The quantitative changes were compatible with earlier reports of linear relationships predicted between the conduction velocity and the peak I Na of the AP, but a nonlinear (logarithmic) relationship between peak I Na and maximum Na+ permeability [20]. Thus, increased arrhythmogenicity was associated with a reduced conduction velocity of ∼22 % during steady 8 Hz pacing and in the region of a ∼33 % reduction during dynamic pacing, which would correspond to comparable reductions in AP wavelength given an absence of significant changes in repolarization characteristics (VERP and APD90), and determinants of passive conduction reflected in Cx43 expression. These in turn accompanied reductions in membrane Nav1.5 expression of ∼56 % and peak I Na of ∼32 %.

The murine RyR2 S/S heart has proven a useful experimental model for CPVT in reproducing a particular clinically observed human CPVT genotype [25, 40]. RyR2 S/S ventricular myocytes show features of altered Ca2+ homeostasis [15] thought to result from an increased RyR2-mediated Ca2+ leak reflecting an increased sensitivity of Ca2+ release to cytosolic though not to SR levels [Ca2+] [31]. The consequent increase in cytosolic [Ca2+] would result in increased sodium-calcium exchanger (NCX) activity whose electrogenic actions would result in triggering events including delayed after-depolarizations leading to ectopic APs that could potentially initiate ventricular arrhythmia. However, these initial studies did not explore for the presence or otherwise for arrhythmic substrate that could sustain the resulting arrhythmia.

Genetic modifications in RyR2 are also associated with AF phenotypes [17, 31, 37]. This has also been modeled by the RyR2 S/S system which demonstrates abnormal atrial Ca2+ homeostasis, delayed triggering events and atrial arrhythmia [22, 48]. However, they also demonstrated reductions in conduction velocity that could provide an arrhythmic substrate [22]. This was attributed to a reduced Na+ current which could be either attributed to a reduced Nav1.5 expression or a direct inhibitory effect on Na+ channel function of altered Ca2+ homeostasis [21]. This could arise from either increased leak of SR Ca2+ or the consequently elevated diastolic Ca2+. Indeed, recent evidence suggests that altered Ca2+ homeostasis can acutely affect cardiac excitability due to both direct [47] and indirect actions on the Na+ channel [2, 5, 41]. CaMKII has been shown to directly interact with Nav1.5, shifting Na+ current availability to a more depolarized membrane potential, thus enhancing the accumulation of Na+ channels into an intermediate inactivated state [2]. Increases in CaMKII activity additionally is known to phosphorylate RyR2 which itself increases SR Ca2+ leak [45]. Intracellular Ca2+ concentration can also acutely modulate Na+ current density in ventricular myocytes [5]. Atrial conduction slowing has also been observed in further models of RyR2-mediated Ca2+ leak including a CSQ2 mutant [14, 26].

These findings suggest that altered Ca2+ homeostasis following the chronic atrial alterations in SR Ca2+ release in the RyR2 S/S system could compromise Nav1.5 expression or function as a result of the elevated diastolic Ca2+. The present study now demonstrates that RyR2 S/S ventricles similarly displayed a reduced Nav1.5 expression and consequently reduced peak I Na, that could explain similar reductions in their conduction velocities [49]. It further extends these findings in localizing this altered expression to the membrane, as well as the whole tissue, fraction (Fig. 7), leading to a reduced maximum rate of AP depolarization, which would be expected to reduce AP conduction velocity, thus creating an arrhythmic substrate. These findings accompanied a greater arrhythmogenicity of RyR2 S/S murine ventricles, which showed arrhythmic events on extrasystolic (S2) stimulation unlike WT and more frequent arrhythmias that occurred at higher BCLs during dynamic stimulation. These findings took place despite indistinguishable AP recovery characteristics in WT and RyR2 S/S ventricles, as reflected in VERP and APD90 readings, thereby excluding re-entrant mechanisms involving recovery phases of the AP [27]. In contrast, RyR2 S/S showed reduced indices of conduction velocity, θ’ through all BCLs examined compared to WT, despite indistinguishable AP amplitude, APD90 and θ’ alternans and their variation with BCL or DI, particularly at low BCLs.

Our findings therefore suggest that the arrhythmic substrate results from reduced expression of Nav1.5 in the membrane, where a reduced I Na leads to slowed AP conduction velocity, in the ventricles of RyR2 S/S mice. This would be consistent with a situation in which abnormalities in cytosolic Ca2+ exert both short- and long-term effects. In the short term, ectopic activity can follow transient elevations in cytosolic [Ca2+]. In the long term, chronic elevations in cytosolic [Ca2+] can result in a downregulation of either Nav1.5 expression or activity, thereby reducing action potential conduction and resulting in arrhythmic substrate. In such a situation, short-term triggering events could potentially form a means for initiating electrical events then perpetuated by the pre-existing arrhythmic substrate. These findings may have broader implications for the mode of therapeutic intervention in a variety of Ca2+ dependent, and potentially some Nav1.5 dependent, arrhythmia.