Perturbed atrial calcium handling in an ovine model of heart failure: potential roles for reductions in the L-type calcium current.

Heart failure (HF) is commonly associated with reduced cardiac output and an increased risk of atrial arrhythmias particularly during β-adrenergic stimulation. The aim of the present study was to determine how HF alters systolic Ca(2+) and the response to β-adrenergic (β-AR) stimulation in atrial myocytes. HF was induced in sheep by ventricular tachypacing and changes in intracellular Ca(2+) concentration studied in single left atrial myocytes under voltage and current clamp conditions. The following were all reduced in HF atrial myocytes; Ca(2+) transient amplitude (by 46% in current clamped and 28% in voltage clamped cells), SR dependent rate of Ca(2+) removal (kSR, by 32%), L-type Ca(2+) current density (by 36%) and action potential duration (APD90 by 22%). However, in HF SR Ca(2+) content was increased (by 19%) when measured under voltage-clamp stimulation. Inhibiting the L-type Ca(2+) current (ICa-L) in control cells reproduced both the decrease in Ca(2+) transient amplitude and increase of SR Ca(2+) content observed in voltage-clamped HF cells. During β-AR stimulation Ca(2+) transient amplitude was the same in control and HF cells. However, ICa-L remained less in HF than control cells whilst SR Ca(2+) content was highest in HF cells during β-AR stimulation. The decrease in ICa-L that occurs in HF atrial myocytes appears to underpin the decreased Ca(2+) transient amplitude and increased SR Ca(2+) content observed in voltage-clamped cells.

Introduction

Heart failure (HF) remains a leading cause of morbidity and mortality [1] as well as a major risk factor for the development of atrial fibrillation (AF) [2]. In the healthy heart atrial contraction increases ventricular filling and thus cardiac output by approximately 20% [3]. In HF however, the atrial contribution to ventricular filling is reduced and thus reduced atrial contractility contributes to the reduction in cardiac output observed in HF [4]. Whilst the cellular mechanisms leading to reduced contractility in the ventricle in HF have been extensively studied e.g. [5-9], the cellular changes that occur in the atria as result of the development of HF, and how these may contribute to atrial dysfunction in HF, are less well-understood.

Atrial remodelling as a result of HF could lead to reduced atrial contractility and an increased propensity for arrhythmias in at least three general ways; i) through changes to cellular Ca2 + homeostasis reducing the amplitude of the systolic Ca2 + transient or via facilitation of delayed after-depolarizations (DADs) [10], ii) via action potential shortening which could arise as a consequence of changes in the systolic Ca2 + transient or ion channel remodelling (reviewed in [11]) and, iii) as a result of extracellular matrix remodelling leading to fibrosis, increased tissue stiffness and conduction block [12]. There are several reports that the atrial L-type Ca2 + current (ICa-L) is reduced in HF [13-16]. Given the importance of ICa-L in triggering Ca2 + release from the sarcoplasmic reticulum (SR) [17,18], the reduction of atrial ICa-L is potentially significant as it will firstly reduce the amplitude of the systolic Ca2 + transient and thus contractility as well as shortening action potential duration (APD) [19] and atrial effective refractory period thus potentially facilitating re-entrant arrhythmia formation [11].

Despite the extensive data showing that ICa-L is reduced in the atria in various disease settings, there remains a paucity of information regarding how cellular Ca2 + handling is quantitatively altered in the atria in HF and the impact that changes in ICa-L have on SR Ca2 + content and the systolic Ca2 + transient. Using a brief period of rapid ventricular pacing in the dog, Yeh et al. [20] demonstrated larger Ca2 + transients that were attributed to increased SR Ca2 + loading (measured qualitatively) and a prolonged action potential duration in ‘HF’ myocytes; however, a role for changes in ICa-L, or other Ca2 + homeostatic mechanisms such as SERCA activity and Ca2 + buffering [5,21] was not examined. Moreover, adrenergic stimulation plays a critical role in regulating cardiac contractility and the initiation of arrhythmias [22-24]. However, apart from limited studies on the effects of β-adrenergic (β-AR) stimulation on atrial ICa-L in cardiac disease states [16,25,26], how β-AR stimulation modulates systolic Ca2 +, SR Ca2 + content and cellular Ca2 + homeostasis in the atria during HF is unknown.

Thus, the objectives of the present study were to assess the impact of HF on atrial myocyte intracellular Ca2 + homeostasis, systolic Ca2 + and SR Ca2 + content and how these are modified by β-AR stimulation. We used an ovine model of dilated cardiomyopathy produced by rapid ventricular pacing and found that HF leads to a shortening of the atrial action potential duration (APD) and reduction in Ca2 + transient amplitude but SR Ca2 + content, determined from the amplitude of the caffeine evoked rise of [Ca2 +]i, was unaltered. Under voltage clamp conditions the Ca2 + transient remained smaller in HF and this occurred due to a reduction in ICa-L rather than a decrease in SR Ca2 + content, which was paradoxically increased. During β-AR stimulation the Ca2 + transient was increased more in HF than control cells. However, whilst SR dependent Ca2 + removal and ICa-L were both augmented by β-AR stimulation in HF cells, they remained reduced compared to control cells. Interestingly, SR Ca2 + content was highest in voltage-clamped HF atrial cells during β-AR stimulation. In action potential clamped myocytes the shortened HF APD does not explain the smaller systolic Ca2 + transient. We therefore conclude that the decreased ICa-L in HF atrial myocytes is the major factor causing the smaller systolic Ca2 + transient.

Methods

All procedures were in accordance with The United Kingdom Animals (Scientific Procedures) Act, 1986 and European Union Directive 2010/63. Institutional approval was received from The University of Manchester Animal Welfare and Ethical Review Board.

A detailed description of the experimental methods used in this study is available in the online supplement.

Experimental heart failure

HF was induced in 25 female Welsh sheep (35.2 ± 1.3 kg, ~ 18 months of age) by transvenous right ventricular tachypacing as described in detail previously [5,27,28]. For surgical procedures anaesthesia was induced and maintained by isoflurane inhalation (1–4% v/v) and perioperative analgesia (meloxicam, 0.5 mg/kg) and antibiosis (enrofloxacin, 2.5 mg/kg) were provided. Animal welfare and the clinical signs indicating the onset of HF were monitored at least once daily. On development of signs of HF (lethargy, dyspnoea and cachexia), ventricular tachypacing was stopped and para-sternal echocardiography was performed on conscious unsedated animals gently restrained in a sitting position. Animals were killed by intravenous administration of pentobarbitone (200 mg/kg) mixed with heparin (10,000 IU) to prevent coagulation in the coronary vasculature. Age and weight-matched animals were used as controls for the study. Animals were randomly assigned to the control or HF arms of the study and once enrolled, no animals were subsequently excluded from analyses.

Isolated myocyte studies

All experiments (except cell volume measurements, Supplemental Fig. S.I) were performed at 37 °C. Changes in intracellular Ca2 + concentration ([Ca2 +]i) were monitored using the acetoxymethyl ester of either Fluo-5F (5 μM; Figs. 2, 3 & 5) or Fura-2 (5 μM; Figs. 1, 4 & 7). Where Fluo-5F was used to measure [Ca2 +]i we were unable to reliably obtain Ca2 + saturated fluorescence values and therefore changes in [Ca2 +]i were calibrated by determining the resting [Ca2 +]i in Fura-2 loaded cells [29] (following in vitro calibration of the experimental system; Molecular Probes, Invitrogen UK), and then calibrated as originally described by Cheng et al. [30] assuming a Kd for Ca2 + of 1035 nmol/l [31]. Single atrial myocytes were current or voltage-clamped using the perforated patch technique with amphotericin-B (240 μg/ml) [7,32,33]. The surface area to volume ratio (Supplementary data Fig. S.I) was calculated in calcein-AM (20 μmol/l) loaded cells as described previously [32].

SR Ca2 + content was calculated by the rapid application of 10 mM caffeine and integration of the resulting Na+–Ca2 + exchange (NCX) current [7,32-34]. The Ca2 + buffering capacity of cells was also obtained from the caffeine evoked Ca2 + transients [21,32,35].

Immunoblotting

Left atrial tissue was snap frozen and stored in liquid N2 until use. Samples were prepared in RIPA buffer and subject to denaturing SDS-PAGE as described previously [5,32,36]. The expression of the so-called ‘house-keeping’ proteins β-actin and GAPDH were altered in this model of HF (Supplementary data Fig. S.II). We therefore used an internal standard on all blots. This internal standard was obtained from a single control animal and used to normalize protein expression on each gel [5,28]. Three replicate gels were obtained and data averaged for each sample. Details of protein loading, antibody concentrations and sources are provided in the online supplementary information.

Statistics

Initial estimates of sample sizes were determined by power analysis using Sigmaplot 11 (Systat Software Inc., USA) assuming α = 0.05 and β = 0.8 and were based on the level of effects and standard deviations observed previously in this model [5,27,28,37]. Normality of data distribution was assessed using the Kolmogorov–Smirnov statistic (IBM SPSS Statistics v20). Data are expressed either as mean ± standard error of the mean (SEM) for normally distributed data or as the median (inter-quartile range) for non-normally distributed data from n cells and N animals (e.g. n/N control; n/N HF). Continuous scale data was log10 transformed [38] and differences between control and HF animals determined using linear mixed modelling to account for instances where multiple cellular observations (n) were obtained from each experimental subject (N) (IBM SPSS Statistics v20). Where treatments/effects were within the same animal, cellular differences were assessed using a paired Students t-test and considered significant when P < 0.05.

Results

Cardiac and cellular remodelling in heart failure

Clinical signs of HF (lethargy and dyspnoea) were present after a median 33.5 (29.8–37) days of tachypacing. Due to the anatomic orientation of the cardiac apex over the sternum in sheep we were unable to obtain trans-thoracic four-chamber views and thus measures of atrial dimensions or E/A ratios. However ventricular dilatation was present as we have found previously [5]. In the animals used in the present study left ventricular internal diastolic dimension increased by 47.2 ± 8.3% and fractional shortening decreased by 57.1 ± 1.9% (both P < 0.001). Atrial cellular hypertrophy was noted in isolated left atrial myocytes (Supplementary Fig. SI.A) with an increase in planar width from 15.0 ± 0.2 to 18.1 ± 0.4 μm and length from 127.5 ± 1.7 to 168.5 ± 3.5 μm (both P < 0.001; 347/23 control; 110/9 HF). Membrane capacitance also increased from 59.6 ± 2.1 to 125.9 ± 9.3 pF (Supplementary Fig. SI.B, P < 0.001). From simultaneously derived confocal 3-dimensional rendered cell volumes and membrane capacitance measurements [32,39], the depth of the cells increased in HF (control, 9.5 ± 0.4 μm; HF 16.5 ± 1.1 μm, P < 0.001. 19/5 control; 15/5 HF). However, the calculated surface area to volume ratio was unaltered (control, 5.06 ± 0.17 pF/pl; HF, 4.89 ± 0.26 pF/pl. P = 0.58). The respective surface area to volume ratio was used to express all integrated cellular Ca2 + fluxes and SR Ca2 + contents to total cellular volume [7,33].

Decreased atrial Ca2 + transient amplitude and action potential duration in heart failure

We first sought to determine if the cellular hypertrophy and chamber dilatation were associated with alterations to the systolic Ca2 + transient. Under current clamp conditions, action potential evoked systolic Ca2 + transients were elicited at 0.5 Hz and representative traces shown in Fig. 1A. Diastolic [Ca2 +]i was reduced in current-clamped HF myocytes control, 60.7 ± 4.5; HF, 37.3 ± 4.8 nmol/l, P = 0.001. However, the amplitude of the action potential evoked rise of systolic Ca2 + was reduced in HF atrial myocytes by 46 ± 17% (Fig. 1B.i, P < 0.05; 18/6 control; 16/5 HF). Conversely, despite the smaller systolic rise of [Ca2 +]i, the amplitude of the caffeine evoked rise of [Ca2 +]i, used as a qualitative measure of SR Ca2 + content, was unaltered following current clamp stimulation (Fig. 1B.ii). The reduction in systolic Ca2 + was however associated with a shortening of the action potential duration in HF atrial myocytes (Fig. 1C, P < 0.001) with the time to 90% repolarization decreasing from 357 ± 26 ms in control to 280 ± 15 ms in HF.

Fig. 1

Action potential duration and Ca2 + transient amplitude are reduced in HF atrial myocytes. A. Representative systolic and caffeine evoked Ca2 + transients from Fura-2 loaded control (left) and HF (right) atrial myocytes following current clamp stimulation at 0.5 Hz, 37 °C. B. Mean data summarizing Ca2 + transient amplitude (i) and, the amplitude of the caffeine evoked rise of [Ca2 +]i (ii). C. Example action potentials from a control and HF cell (i) and, summary data for action potential repolarization times (ii). ⁎P < 0.05; ⁎⁎⁎P < 0.001.

We then sought to determine the potential mechanisms by which the systolic Ca2 + transient amplitude was reduced in HF atrial myocytes. In the first series of experiments the action potential clamp technique was used to apply the averaged control and HF action potential waveforms to myocytes isolated from control atria (Supplementary material, Fig. S.III). Following stimulation with the HF atrial action potential there was no detectable change in the amplitude of the systolic Ca2 + transient or caffeine-evoked rise of [Ca2 +]i or SR Ca2 + content measured from the integral of the NCX current following caffeine application (Supplementary data, Fig. S.III). Thus, changes in action potential duration do not explain the smaller atrial systolic Ca2 + transients observed in this model of HF.

We then performed voltage clamp experiments to facilitate quantitative analysis of ICa-L, cellular Ca2 + fluxes and SR Ca2 + content. Representative Ca2 + transients following steady-state voltage clamp stimulation are illustrated in Fig. 2A. On average the amplitude of the atrial systolic Ca2 + transient was reduced by 28 ± 9% (Fig. 2B.i, P < 0.05; 43/16 control; 21/10 HF). Given the strong dependence of Ca2 + transient amplitude on SR Ca2 + content and the indication, from the slowed decay of [Ca2 +]i in HF (Fig. 2B.ii) that SERCA activity is impaired, we next investigated if the decrease of Ca2 + transient amplitude was due to a fall of SR Ca2 + content. Following steady-state stimulation, caffeine (10 mmol/l) was applied rapidly to the cell and the resulting NCX current integrated. This data is summarized in Fig. 3A. In HF atrial myocytes SR Ca2 + content was increased from 83.9 ± 4.5 to 100.1 ± 6.7 μmol/l (P < 0.05; 24/8 control; 17/10 HF). Therefore, the reduced Ca2 + transient amplitude observed in HF atrial cells cannot be explained by a decrease of SR Ca2 + content.

Fig. 2

Reduced Ca2 + transient amplitude and slowed relaxation in voltage-clamped HF atrial myocytes. A. Representative systolic Ca2 + transients from a control (left) and HF atrial (right) myocyte. Cells were voltage-clamped, stimulated from a holding potential of − 40 mV with a 100 ms, 50 mV step at 0.5 Hz, 37 °C and changes in [Ca2 +]i measured with Fluo-5F. B. (i) Mean data summarizing changes in Ca2 + transient amplitude. (ii) Superimposed Ca2 + transients from representative control and HF atrial cells demonstrating a slowed rate of decay of [Ca2 +]i. ⁎P < 0.05.

Fig. 3

SR Ca2 + content is increased and cellular Ca2 + buffering power reduced in HF atria. A. (i) Quantification of SR Ca2 + content in voltage-clamped control (left) and HF (right) atrial cells. Following steady state-stimulation membrane potential was held at − 40 mV and 10 mmol/l caffeine applied rapidly as indicated to discharge the SR store resulting in an inward NCX current (middle panels) which was integrated (lower panels) to quantify SR Ca2 + content. (ii) Mean data for SR Ca2 + content. B. (i) Representative Ca2 + buffer slopes from the cell types indicated. Data obtained by plotting total Ca2 + ([Ca2 +]T, ordinate) as a function of [Ca2 +]i (abscissa). Total Ca2 + and [Ca2 +]i are both derived from caffeine evoked Ca2 + transient. The broken lines fitted through the original data are best-fit linear regressions. (ii) Summary data showing mean data for cellular Ca2 + buffering power (β). C. (i) Representative data showing the relationship between NCX current (INCX) and [Ca2 +]i obtained during the decay phase of the caffeine-evoked Ca2 + transient. The broken lines through the original data are best-fit linear regressions. (ii). Mean data for the slope of the linear regression applied to the INCX–[Ca2 +]i relationship. ⁎P < 0.05; ⁎⁎⁎P < 0.001.

Another possible explanation for the smaller systolic Ca2 + transients observed in HF atrial cells would be an increase in the Ca2 + buffering power of the cells. Representative Ca2 + buffering curves are illustrated in Fig. 3B.i. The relationship between total and free Ca2 + is shallower in the HF atrial cells (Fig. 3B.ii, control, 565 ± 35; HF, 318 ± 42). On average the Ca2 + buffering power of the cells (ratio of total to free Ca2 +) was reduced by 43.7 ± 8.2% in HF (P < 0.001; 24/9 control; 17/10). Therefore, as with SR Ca2 + content, changes in Ca2 + buffering power do not explain the smaller systolic Ca2 + transient in HF atrial cells.

The dependence of NCX on [Ca2 +]i was also assessed from the caffeine evoked rise of [Ca2 +]i. The representative traces and summary data (Fig. 3C) show that less NCX current is produced for a given change in [Ca2 +]i in HF atrial myocytes indicating a down-regulation of NCX activity in HF. On average the slope of the relationship between INCX and [Ca2 +]i was decreased by 24.9 ± 8.7% (Fig. 3C.ii, P < 0.05; 24/9 control; 15/9 heart failure).

Decreased L-type Ca2 + current and role in altered systolic Ca2 + and SR Ca2 + content in heart failure

Given that the increase of SR Ca2 + content and reduction in Ca2 + buffering power in HF atrial cells are unable to account for the decreased Ca2 + transient amplitude observed in HF atrial myocytes, we next investigated a potential role for changes in the L-type Ca2 + current (ICa-L) as this has both trigger and loading roles. The representative ICa-L records and summary data of Fig. 4A show that HF was associated with a reduction in the peak ICa-L density (by 36.1 ± 7.8%; control, 2.27 ± 0.14 pA/pF; HF, 1.45 ± 0.15 pA/pF. P < 0.005; 47/20 control; 21/10 heart failure). However, the amount of Ca2 + entry via the L-type Ca2 + current calculated by integrating ICa-L was unchanged in HF (control, 0.74 ± 0.06 μmol/l; HF, 0.86 ± 0.09 μmol/l. P = 0.827). The unchanged integrated Ca2 + entry is associated with a slowed time course of inactivation of ICa-L in HF (control, 404 ± 50 s− 1; HF, 171 ± 23 s− 1, P < 0.01).

Fig. 4

Decreased ICa-L in HF reduces systolic Ca2 + and increases SR Ca2 + content. A. (i) Representative ICa-L traces elicited by step depolarizations as indicated above the current record. (ii) Mean data summarizing peak ICa-L in control and HF atrial myocytes. B. (i) Experimental time course: the application of nicardipine (5 μmol/l) decreases the systolic Ca2 + transient. [Ca2 +]i measured using Fura-2. Data obtained on different experimental apparatus and expressed as pseudo-ratio (R/Rrest) relative to the resting ratio of emitted light excited at 340 nm and 380 nm. (ii) Mean data summarizing the effect of nicardipine on peak ICa-L. (iii) Mean data summarizing the effect of nicardipine on Ca2 + transient amplitude. C. (i) The effect of nicardipine on SR Ca2 + content quantified by the rapid application of caffeine (10 mmol/l, solid bars) to a control cell following steady-state stimulation in the absence (left) and presence (right) of nicardipine (5 μmol/l, open bar). (ii) Mean data summarizing the effect of nicardipine on SR Ca2 + content. ⁎P < 0.05; ⁎⁎P < 0.01; ⁎⁎⁎P < 0.001.

The role that the decreased ICa-L has in producing the smaller Ca2 + transient in HF atrial cells was then investigated using the approach illustrated in Fig. 4B. In control cells, nicardipine (5 μmol/l) was applied to reduce ICa-L to a similar extent to that observed in HF (Fig. 4B.ii, by 30 ± 15%). This reduction of ICa-L caused a 27 ± 15% decrease in Ca2 + transient amplitude (Fig. 4B.iii, P < 0.01). However, unlike the maintained integrated Ca2 + entry via ICa-L in HF atrial myocytes, the integrated Ca2 + entry in response to nicardipine was reduced (from 0.9 ± 0.2 to 0.7 ± 0.2 μmol/l; P < 0.01; 8 cells from 4 control hearts). The reduced integrated Ca2 + entry presumably reflects the combined effects of the decreased peak and unaltered rate of inactivation of the current in nicardipine. A potential role for altered L-type Ca2 + channel single channel kinetics in HF, specifically increased open probability and availability [40], would also contribute to the differential effects on ICa-L inactivation in HF and nicardipine treated myocytes. Furthermore, the favoured mode 0 gating of the channel induced by the dihydropyridine antagonists [41] would also lead to the differential effects on ICa-L inactivation between HF and nicardipine treated cells. The effect that the pharmacological reduction of ICa-L had on SR Ca2 + content is shown in Fig. 4C. The amplitude of the caffeine evoked rise of [Ca2 +]i and inward NCX current are increased following nicardipine exposure (Fig. 4C.i) and this resulted in an increase of SR Ca2 + content from 74.1 ± 7.6 μmol/l in control to 89.9 ± 10.5 μmol/l in nicardipine (Fig. 4C.ii, P < 0.01). This data therefore suggests that it is the reduction of ICa-L in HF atrial cells that is responsible for both the decrease of Ca2 + transient amplitude and increase of SR Ca2 + content.

Whilst the decrease in ICa-L therefore explains the smaller Ca2 + transients and increased SR Ca2 + content observed in voltage-clamped cells we also sought to determine if changes in SERCA or NCX activity also occur in HF and thus may contribute to the observed changes in systolic Ca2 +. Firstly we determined if SERCA activity was altered in HF using the approaches outlined in Fig. 5A. Normalized steady-state systolic Ca2 + transients from a control and HF atrial cell are shown in Fig. 5B.i where it is clear that the systolic Ca2 + transient decays more slowly in HF. This was quantified by fitting a single exponential to the decay phase of the Ca2 + transient to obtain the rate constant of decay of systolic Ca2 + (ksys, Fig. 5B.ii). On average, ksys was reduced by 27.1 ± 6.3% in HF atrial cells (Fig. 5B.iii, Control, 8.99 ± 0.36 s− 1; HF, 6.55 ± 0.5 s− 1, P < 0.001; 43/18 control; 21/10 HF). The underlying mechanism for the reduction of ksys in HF atrial cells was then determined by also fitting the decay phase of the caffeine evoked Ca2 + transient with a single exponential (kcaff). The SR dependent rate of Ca2 + removal (kSR) was then derived by subtraction (kSR = ksys − kcaff). In HF, sarcolemmal dependent Ca2 + extrusion by NCX and PMCA (kcaff) was increased by 25.2 ± 10% (control, 0.745 ± 0.024 s− 1; HF, 0.933 ± 0.069 s− 1; P < 0.01). Conversely, the SR dependent rate of Ca2 + removal (kSR) was decreased (Fig. 5B.iv, Control, 8.1 ± 0.4 s− 1; HF, 5.4 ± 0.5 s− 1, P < 0.001). Thus the increase in surface membrane-mediated Ca2 + extrusion (kcaff) is inconsistent with the decrease in the slope of the relationship between INCX and [Ca2 +]i (Fig. 3C). There are at least two possible explanations for this discrepancy; i) an increase in plasmalemmal Ca2 +-ATPase mediated Ca2 + extrusion in HF and, ii) an effect of reduced cellular Ca2 + buffering power (Fig. 3B). The effect of the reduced Ca2 + buffering power is examined in Supplementary Fig. S.IV. We find that during the systolic Ca2 + transient, in representative cells, CaT falls more slowly in the HF cell whereas CaT falls more quickly in the HF cell during the caffeine-evoked Ca2 + transient. In Fig. S.IV.B, the relationship between kcaff and cellular Ca2 + buffering power is examined in control and HF cells. An inverse correlation (Pearsons correlation coefficient, P = 0.01) exists indicating that as Ca2 + buffering power decreases then k increases. We have also determined, by blocking NCX with 10 mmol/l Ni2 + and measuring the rate of decay of the caffeine evoked rise of [Ca2 +]i, that the contribution of PMCA to Ca2 + extrusion is no different between control and HF atrial myocytes (kPMCA: control, 1.17 ± 0.04 s− 1; HF, 1.09 ± 0.01 s− 1). The unaltered kPMCA, despite reduced Ca2 + buffering, is further evidence that the decrease in Ca2 + buffering power in HF accounts for the observed increase in kcaff.

Fig. 5

Slowed rates of decay of systolic Ca2 +, reduced SR mediated Ca2 + uptake and enhanced sarcolemmal Ca2 + extrusion in HF atrial myocytes. A. Experimental time course illustrating how SR and non-SR dependent rates of Ca2 + removal are calculated in voltage-clamped atrial myocytes. Following steady-state stimulation caffeine is applied to discharge the SR store. B. (i) Normalized systolic Ca2 + transients from a representative control and HF atrial cell as indicated. (ii) Method for determining the rate of decay of the systolic Ca2 + transient (ksys) by fitting (broken line) a single exponential function to the data. (iii) Summary data for the calculated rates of decay of the systolic Ca2 + transient. (iv) Summary data for rate constants of decay of [Ca2 +]i due to sarcolemmal extrusion (kcaff) and SR mediated uptake (ksys). ⁎⁎P < 0.01; ⁎⁎⁎P < 0.001.

Molecular correlates of decreased SERCA function and reduced Ca2 + transient amplitude

The next experiments were designed to determine if changes in the expression and/or phosphorylation status of various proteins involved in controlling SERCA activity explain the reduced SERCA function noted in Fig. 5. In each case we used either 6 or 7 hearts from control and HF animals for biochemical analyes. There was no change in SERCA protein expression, total PLN expression, the SERCA to PLN expression ratio (Fig. 6A) or CSQ expression (Supplementary data, Fig. S.V). Using phospho-specific antibodies we also examined the phosphorylation status of PLN (Fig. 6B) and observed that whilst the phosphorylation status of the PKA-dependent residue (Ser-16) was unaltered in HF, there was increased phosphorylation at the adjacent Ca2 +-calmodulin kinase II- (CAMKII) dependent Thr-17 residue (to 221 ± 73% of control, P < 0.05); an effect most likely due to the 320 ± 40% increase (relative to control) in CAMKII expression in HF (Fig. 6C, P < 0.001). The unaltered PKA-dependent phosphorylation of PLN was associated with an increase in protein phosphatase 1 (PP1; to 159 ± 31%; P < 0.05; Fig. 6D.i), protein phosphatase 2a (PP2a, to 198 ± 38%; P < 0.01; Fig. 6D.ii) and G-protein receptor kinase 2 (GRK-2; to 161 ± 32%; P < 0.05; Fig. 6E) expression HF.

Fig. 6

Molecular alterations impacting on atrial cellular Ca2 + handling in HF. A. Representative Western blots (upper panel) and summary data (lower panel) for (i) SERCA2a, (ii) PLN and, (iii) summary of the SERCA2a:PLN. B. Representative Western blots and summary data for (i) Ser16 and (ii) Thr17 phosphorylated PLN. C. Determination of CAMKIIδ expression in heart failure atria showing Western blot (upper panel) and summary data (lower panel). D. Representative Western blots (upper panel) and summary data (lower panel) for (i) PP1 and, (ii) PP2a expression in HF. E. Increased GRK-2 expression in HF showing representative Western blot (upper panel) and summary data (lower panel). ⁎P < 0.05; ⁎⁎P < 0.01; ⁎⁎⁎P < 0.001.

We then examined additional molecular correlates of the reduced Ca2 + transient amplitude observed in HF atria and investigated RyR expression and phosphorylation at PKA- and CAMKII-dependent sites (Supplementary data; Fig. S.V). No differences in RyR expression, PKA-dependent Ser-2808 RyR phosphorylation or CAMKII-dependent Thr-2814 RyR phosphorylation were observed between control and HF atrial tissues.

Restoration of the systolic Ca2 + transient in HF atrial cells by β-adrenergic stimulation; an effect via the L-type Ca2 + current and SR Ca2 + content

To determine if the reduced Ca2 + transient observed in HF atrial cells could be ‘rescued’, we examined the effects of the non-specific β-AR agonist isoprenaline (100 nmol/l) on Ca2 + transient amplitude (Fig. 7A.i & ii). Isoprenaline increased Ca2 + transient amplitude in both control and HF atrial cells although to a greater extent in HF cells (Fig. 7B.i; control, by 66 ± 8%; HF, by 137 ± 15%; P < 0.001; 24/11 control; 23/5 HF) such that during β-AR stimulation, Ca2 + transient amplitude in HF was indistinguishable from that in control (not shown). Isoprenaline also accelerated the rate of decay of the systolic Ca2 + transient in both control and HF atrial cells (ksys, Fig 7B.ii); an effect attributable to an increase in SERCA activity with kSR increasing from 1.80 ± 0.36 s− 1 to 5.19 ± 0.68 s− 1 in control cells (P < 0.001) and, 0.93 ± 0.17 s− 1 to 3.47 ± 0.50 s− 1 in HF atrial cells (Fig. 7B.iii, P < 0.001). Again, during β-AR stimulation HF and control cells were statistically indistinguishable (ksys, P = 0.145; kSR, P = 0.187). The increase in SERCA activity during β-AR stimulation also increased SR Ca2 + content in both control (97.0 ± 4.7 to 113.8 ± 7.6 μmol/l, P < 0.001) and HF (145.2 ± 6.4 to 171.1 ± 15.1 μmol/l, P < 0.01) atrial cells (Fig. 7C). However, during β-AR stimulation, SR Ca2 + content was greatest in HF atrial cells (P < 0.001). The isoprenaline-mediated augmentation of Ca2 + transient amplitude was also associated with an increase of L-type Ca2 + current (Fig 7D.i) in both cell types (Fig. 7D.ii, control, 4.26 ± 0.35 pA/pF; HF, 2.81 ± 0.36 pA/pF; P < 0.001) although peak ICa-L was smaller in HF cells than in control cells (P < 0.05). In ISO the integrated Ca2 + entry on ICa-L was indistinguishable between control and HF mycoytes. Finally, by Western blotting we were unable to detect differences in the expression of the G-proteins Gαs and Gαi(1/2/3) (Supplementary data; Fig S.VI).

Fig. 7

Normalization of the systolic Ca2 + transient and impact of β-AR stimulation on cellular Ca2 + homeostasis in HF atrial myocytes. A. Representative experimental time course from representative control (i) and HF (ii) cells. Cells were voltage-clamped and stimulated in the presence of 100 nmol/l isoprenaline as indicated above the records. Panels show steady-state effects. [Ca2 +]i measured with Fura-2. Data obtained on different experimental apparatus and expressed as a pseudo-ratio (R/Rrest) relative to the resting ratio of emitted light excited at 340 nm and 380 nm. B. (i) Isoprenaline mediated increase in Ca2 + transient amplitude. (ii) Summary data for effect of β-AR stimulation on the rate of decay of the systolic Ca2 + transient (ksys). (iii) Summary data for effect of β-AR stimulation on the SR dependent rate of Ca2 + removal (kSR). C. Summary data for increase in SR Ca2 + content during β-AR stimulation. D. (i) Representative ICa-L records from control and HF atrial myocytes under basal stimulation and following β-AR stimulation as indicated. (ii) Mean data summarizing the effect of β-AR stimulation on peak ICa-L density. ⁎⁎P < 0.01; ⁎⁎⁎P < 0.001.

The effects of HF-induced changes to intracellular Ca2 + homeostasis on cellular Ca2 + economy in the atria

The previous sections highlight multiple changes in cellular Ca2 + homeostasis in atrial cells occurring in HF. How these impact on the total Ca2 + economy of the cell are summarized in Table 1. The systolic Ca2 + transient decreases by 28 ± 9% in HF. However, as a consequence of the decreased Ca2 + buffering power (β) of the cell in HF the total Ca2 + requirement to generate the systolic Ca2 + transient is decreased by 59 ± 7% (P < 0.001). This total Ca2 + requirement is met from two sources; i) Ca2 + entry via ICa-L and, ii) the release of Ca2 + from the SR. Given that the integrated Ca2 + entry via ICa-L is unaltered and SR Ca2 + content is increased in HF, it is not surprising therefore that the fractional release of Ca2 + from the SR is reduced in HF by 64 ± 7% (P < 0.001). Finally, the gain of excitation contraction coupling can be calculated from the integrated Ca2 + entry and release of Ca2 + from the SR during the systolic Ca2 + transient is reduced by 67 ± 7% (P < 0.001) in HF.

Table 1

Modelling the effects of Ca2 + handling alterations in heart failure on systolic Ca2 +. Summary of how HF affects the total Ca2 + economy of the cell. , not significant; *, P < 0.05; #, P < 0.01 and §, P < 0.001. The following equations were used to calculate the indicated parameters: Ca2 + buffering power (β) = Δ Ca2 + total/Δ [Ca2 +]i; Total Ca2 + transient = β ∗ Δ [Ca2 +]i; systolic Δ SR Ca2 + content = SR Ca2 + content − (Total Ca2 + − ʃICa-L); SR fractional release = Systolic Δ SR Ca2 +/SR Ca2 + content; EC coupling gain = Systolic Δ SR Ca2 +/ʃICa-L.

	ControlAtrial cell (n = 43)	Heart failureAtrial cell (n = 21)
Δ systolic [Ca2 +]i (nmol/l)	83 ± 6	60 ± 6#
Ca2 + buffering power (β)	565 ± 41	318 ± 29§
Total Ca2 + transient (μmol/l)	46.9 ± 4.8	19.1 ± 2.6§
Integrated ICa-L (μmol/l)	0.74 ± 0.06	0.86 ± 0.09ns
SR Ca2 + content (μmol/l)	83.9 ± 4.5	100.1 ± 6.7⁎
Δ SR Ca2 + systole (μmol/l)	46.2 ± 4.8	18.2 ± 2.6§
SR fractional release	0.55 ± 0.06	0.18 ± 0.03§
EC coupling gain	62.4 ± 8.2	21.2 ± 3.8§

Discussion

The main findings from the present work are that in failing hearts atrial myocytes, i) are hypertrophied, ii) have smaller systolic Ca2 + transients that decay more slowly, iii) have a reduction in ICa,L, iv) have a reduced Ca2 + buffering power and, iv) have a greater increase in Ca2 + transient amplitude in response to β-AR stimulation. The reduction of ICa-L is sufficient to explain the smaller systolic Ca2 + transient and the accompanying increased SR Ca2 + content observed in voltage clamped cells. The reduced Ca2 + buffering capacity of HF atrial myocytes is insufficient to maintain systolic Ca2 +, but it does facilitate the enhanced sarcolemmal-mediated rate of Ca2 + extrusion observed in HF atrial myocytes. The restoration of systolic Ca2 + during β-AR stimulation likely occurs as a result of increases in ICa-L and SR Ca2 + content. The above changes in cellular Ca2 + homeostasis provide a mechanism for the reduction in Ca2 + transient amplitude observed in atrial myocytes isolated from failing hearts.

Cellular remodelling in heart failure

Two major considerations in the present manuscript are how HF induced by right ventricular tachypacing mediates changes in cellular geometry and atrial cellular Ca2 + homeostasis. As such it is important to note (Supplementary material, Fig S.VII) that the high ventricular pacing rate used to initiate HF does not increase atrial rate. The changes in Ca2 + homeostasis and cellular remodelling are therefore not due to atrial tachypacing in contrast to those observed in various models of atrial fibrillation e.g. [42,43]. In order to quantify the HF-induced changes in cellular Ca2 + cycling, a measure of the surface area to volume ratio is required to relate sarcolemmal Ca2 + fluxes and SR Ca2 + content to the volume of the cell. In this model of HF, atrial myocytes are markedly hypertrophied showing a ~ 20% increase in cell diameter, ~ 30% increase in cell length and ~ 70% increase in cell depth. Both the increase of cell dimensions and loss of t-tubules that occurs in atrial cells in this HF model [27] would be expected to decrease the surface area to volume ratio. However, such a decrease is not observed in experiments where cell volume is derived from confocal z-stacks of calcein loaded cells patch-clamped to obtain a simultaneous measure of cell volume and membrane capacitance (surface area). Indeed, the HF and control cell data fit on precisely the same surface area (capacitance) to volume relationship even though their surface area and volume values overlap minimally. There are at least two potential factors that could explain the apparent constancy of the surface area to volume ratio in HF despite the cellular hypertrophy and loss of t-tubules [27]; i) the depth of the cells increased fractionally more than cell width consequently changing the expected shift in surface area to volume ratio (which is ‘based’ on the assumption that cells are ‘cylindrical’) and, ii) atrial dilatation may have stretched the cells and incorporated sub-membrane caveolae into the surface membrane and thus increased the apparent surface area to volume ratio. Sub-membrane caveolar incorporation has been suggested previously in acutely stretched myocytes [44]. Such effects would result in an increase in the observed surface area to volume ratio over that predicted from the changes in cellular planar dimensions.

Decreased ICa-L can explain the smaller systolic Ca2 + transient and paradoxical increase of SR Ca2 + content in voltage-clamped HF atrial cells

We find that a decrease of ICa-L accompanied the observed decrease of the systolic Ca2 + transient. Such a decrease of atrial ICa-L has been observed previously in various models of cardiac disease e.g. [42,43,45,46]. That the decrease in ICa-L is causal of the decreased Ca2 + transient is demonstrated by the observation that pharmacologically inhibiting ICa-L in control cells with nicardipine reproduces the effect of HF on ICa-L, Ca2 + transient amplitude and SR Ca2 + content. Whilst we do not have data for changes in dihydropyridine binding or L-type Ca2 + channel α-subunit expression in the atria in HF, the loss of atrial t-tubules that occurs in this model of HF [27] is likely to contribute to the fall in ICa-L [47].

The decrease of ICa-L may also explain the observed increase of SR Ca2 + content in voltage-clamped myocytes. The L-type Ca2 + current has two effects on SR Ca2 +; i) it triggers release and therefore an increase of ICa-L would be expected to decrease SR Ca2 + and, ii) on the other hand it loads the cell and therefore the SR with Ca2 +. It is therefore difficult to predict the net effect of change in ICa-L on SR Ca2 + content. We find that in the atria, as in the ventricle [5,18,48], reducing ICa-L leads to an increase of SR Ca2 + content. However, the increase in SR Ca2 + content is insufficient to restore the systolic Ca2 + transient which remains smaller in HF atrial myocytes despite the increase in SR Ca2 + content.

Altered cellular Ca2 + buffering and t-tubule loss in heart failure

Whilst the smaller systolic Ca2 + transient observed in HF atrial cells can be mimicked in control cells by inhibition of ICa-L with nicardipine, additional factors regulating the systolic Ca2 + transient also change in HF. These include, i) reductions in the Ca2 + buffering capacity and, ii) a decrease in the gain of excitation contraction coupling. The net effect of these changes however remains a smaller rise of Ca2 + in HF atrial cells. A key factor likely contributing to the reduction in excitation contraction coupling gain is the reduction in peak (trigger) ICa-L in HF. Again, the loss of t-tubules in HF atrial myoyctes [27,42,49] and hence reduction in ICa-L [47,50] is therefore likely an important factor in the decrease in Ca2 + transient amplitude we observe in HF atrial cells in the sheep.

The finding in the present study that, despite the reported loss of t-tubules in HF atrial myocytes [27], the relationship between NCX current and [Ca2 +]i is reduced in HF is consistent with either a reduction of NCX expression at the cell surface in HF or, if this is unaltered, an increased density of NCX on the t-tubules which have subsequently been lost in HF. Given the importance of NCX in arrhythmogenesis via Ca2 + dependent mechanisms we would suggest that, on balance, the decrease in NCX current we see in this model of end-stage heart failure might mean that other mechanisms e.g. as a result of action potential shortening or K+ current remodelling, may be more important in increasing the susceptibility of the atria to arrhythmias in HF. However, it is important to note that the data presented here is from a model of end-stage heart failure where contractile dysfunction rather than arrhythmias may be more pathologically relevant.

β-Adrenergic stimulation and restoration of the systolic Ca2 + transient in HF atrial myocytes: potential importance in arrhythmia initiation

In the present study we find that, during β-AR stimulation, the amplitude of the systolic Ca2 + transient in HF atrial cells is indistinguishable from that of control cells. However, neither the rate of decay of the systolic Ca2 + transient (ksys) nor SR dependent rate of Ca2 + removal (kSR) is fully restored by β-AR stimulation. Nevertheless, SR Ca2 + content is increased by β-AR stimulation in both control and HF cells and importantly, in HF cells, SR Ca2 + content remains greater than in control cells. It is therefore likely that the combined effect of increasing SR Ca2 + content and ICa-L during β-AR stimulation is responsible for normalizing systolic Ca2 + in HF atrial myocytes. Thus, during β-AR stimulation the increase in excitation contraction coupling gain and fractional release of Ca2 + from the SR [51] facilitate the propagation of the peripheral Ca2 + transient to the cell centre and restore systolic Ca2 + despite the loss of atrial t-tubules in HF [27,52].

Although not addressed specifically in the present study, due to the difficulty in quantitatively measuring changes of SR Ca2 + content from the amplitude of the caffeine evoked Ca2 + transient, should SR Ca2 + content preferentially increase during β-AR stimulation in HF atrial cells following current clamp stimulation (the in vivo situation) this would also provide a conducive pathway to DAD formation and act as an important precursor for new onset AF [10]. However, despite the potential effects of changes in SR Ca2 + content following β-AR stimulation during current clamp stimulation, the shorter action potential duration noted in HF atrial cells, via reductions in effective refractory period would also serve to promote arrhythmias although via re-entrant mechanisms.

Molecular correlates of altered cellular Ca2 + homeostasis and β-AR responsiveness in heart failure atrial myocytes

A major observation in the present study is that SERCA activity is reduced in HF atrial cells resulting in reductions to the rate of decay of the systolic Ca2 + transient and SR dependent Ca2 + uptake. Whilst changes in SERCA expression or activity have only modest effects on SR Ca2 + content [53,54] (due to the shallow dependence of SR Ca2 + content on SERCA activity [55]), we noted that SR Ca2 + content was increased in voltage-clamped HF atrial cells. Importantly, the changes in SERCA activity were noted despite the decrease in Ca2 + buffering power and therefore, as presented, likely represent an underestimate of the actual decrease in SERCA function. Whilst we were unable to detect changes in SERCA, total PLN, Ser-16 phosphorylated PLN or the ratio of PLN to SERCA, we did observe an increase in Thr-17 phosphorylated PLN, CAMKII, PP1, PP2a and GRK-2 expression. Thus the reported changes in protein expression in HF atria do not readily account for the observed reduction in SERCA activity. In the atria SERCA is also regulated by an additional protein, sarcolipin [56]. However, due to a lack of suitable antibodies we have been unable to investigate whether or not sarcolipin expression alters in HF atrial myocytes and can explain the decrease in SERCA activity we observe.

The failure of β-AR stimulation to completely restore SERCA function and ICa-L in HF atrial cells is consistent with the increased expression of the G-protein receptor kinase GRK-2 [57]. Similar findings have been reported previously in the ventricle in various models of cardiac disease e.g. [5,58]. However, given the effect that reducing ICa-L has on SR Ca2 + content, the increases in GRK-2, PP1 and PP2a may be important in maintaining the increased SR Ca2 + content in HF atrial cells during β-AR stimulation because they attenuate any β-AR mediated increase in ICa,L.

Study limitations

We have examined how intracellular Ca2 + homeostasis and the response to catecholamine stimulation are altered in the atria in HF. Whilst the changes in cellular Ca2 + homeostasis and, in particular SR Ca2 + content, may provide a substrate for the generation of Ca2 + dependent arrhythmias we have not established if the threshold SR Ca2 + content is altered in HF, if HF atrial myocytes are more excitable due to concomitant alterations in K+ channel expression or function e.g. [15] and whether ultimately AF is more readily induced and maintained in this model of HF. However, in a comparable model of tachypacing induced HF in the dog, an increased susceptibility to, and stability of, AF has been reported [59] and, at the single cell level, spontaneous SR Ca2 + release and DADs were more prevalent in HF atrial myocytes [20]. For the measurements of ICa-L a holding potential of − 40 mV was used and we did not correct for any liquid junction potential artefacts (~ 8 mV).

For the in vivo assessments of cardiac function by ECG and echocardiography animals were gently restrained. As such we cannot exclude the possibility that adrenergic outflow may be influencing our measures of cardiac function in both control and HF animals; and potentially differentially. However, the use of sedation or performing such measurements in anaesthetized animals would also be expected to influence these measurements and again may do so differentially.

In the present study we have used Western blotting to measure the expression of protein phosphatases and kinases and have not investigated their activity directly. However, the expression of protein phosphatases and G-protein receptor kinases has previously been demonstrated to correlate directly with their activity [60,61]. Thus, our observation of increased PP1, PP2a and GRK-2 expression is consistent with the reduced SR Ca2 + uptake rate and impaired ICa-L that are present under basal and β-AR stimulation conditions in HF atrial myocytes. Moreover, we did not determine if changes in PMCA activity occur in this model of HF and therefore whether the correction factor non-NCX mediated Ca2 + extrusion differs between control and HF myocytes. However, even if PMCA activity were to change by 50% this would result in an under or overestimation of SR Ca2 + content by less than 10%.

Conclusions

In summary we describe a number of changes to atrial Ca2 + homeostasis that occur as a consequence of HF induced by right ventricular tachypacing. Specifically, the decrease in ICa-L and previously reported loss of atrial t-tubules in this model of HF serve to decrease the systolic Ca2 + transient amplitude and increase SR Ca2 + content in voltage-clamped cells. However, the changes in SR Ca2 + content observed in voltage-clamped cells, are insufficient to prevent the smaller rise of systolic Ca2 +. Similarly, the observed decrease in cellular Ca2 + buffering capacity is also insufficient to maintain systolic Ca2 + although it does contribute to an increased rate of Ca2 + removal from the cell via the sarcolemmal pathways despite the reduction in NCX current in the atria in heart failure.

Funding

This work was supported by grants from the British Heart Foundation (FS/12/57, PG/11/16, PG10/89, PG/09/062, FS/09/036, PG/08/078, PG/07/099, the Michael Frazer PhD studentship FS/07/003), European 6th Framework Award (Normacor), University of Manchester Biomedical Research Centre Award (George Lancashire Award) and Wellcome Trust Institutional Strategic Support Fund (ISSF) award (097820).

Disclosures

None.