Introduction

The catecholamine noradrenaline (NA) is released by sympathetic (adrenergic) post-ganglionic nerves terminating on cardiac myocytes. It is substantially involved in regulating cardiac excitation–contraction and the fight-or-flight response, and sometimes in the generation of cardiac arrhythmias including the most common, atrial fibrillation (AF) [39, 46]. The generation of AF by NA probably involves its marked effect to increase atrial L-type Ca2+ current, ICaL, as shown in human [4, 6] and rat [3] atrial myocytes, in turn contributing to triggered activity from atrial spontaneous depolarisations or afterdepolarisations [9, 19, 44, 46]. This ICaL increase results, in large part, from activation of cell surface beta-adrenoceptors (β-AR), supported by numerous studies showing marked effects of the synthetic broad action (β1-, β2- and β3-AR) agonist isoprenaline (ISO) on human atrial ICaL (e.g. [4, 21, 31]). Furthermore, ISO infusion in patients produced AF [29]. β-AR antagonists are used in the pharmacological treatment of patients with AF, primarily for controlling the associated rapid ventricular rates (rate control), but they may also be effective in suppressing AF (rhythm control) when adrenergic tone is elevated, e.g. β1-AR sub-type antagonists in patients with postoperative AF (bisoprolol, metoprolol) [8] or with heart failure (metoprolol) [37]. However, NA activates α- as well as β-ARs, and each main AR sub-type has been identified in human atrial myocardium [46]. Moreover, since the mixed α1-, β1-, β2-AR antagonist, carvedilol, was more effective at preventing postoperative AF than the β1-antagonists metoprolol or atenolol [15, 23], this suggests the possibility of identifying specific mixed AR sub-type antagonism profiles for optimising rhythm control drug efficacy during adrenergic AF.

To do so, however, requires an improved understanding of the contributions of activation of the individual AR sub-types to the effect of NA on human atrial ICaL. So far, this has been addressed using AR sub-type selective agonists, with β2-agonism (salbutamol) increasing human atrial ICaL [50], β3-agonism (BRL37344) having no effect [5] and selective β1-agonism not yet studied for human atrial ICaL. Reports of α-AR agonism in human atrium are so far restricted to contraction, e.g. positive inotropic effect of the α1-agonist phenylephrine [14], although ICaL has been studied in other species, with a marked increase in the current by phenylephrine in cat atrial myocytes [41], and no effect of phenylephrine or methoxamine in rabbit or rat atrial myocytes [12, 18]. Mixed effects of α-AR agonists have also been reported for ventricular ICaL [36]. It is important, however, to investigate the AR sub-type contribution to the ICaL response when using the naturally occurring catecholamine, NA, because this will stimulate all the ARs, simultaneously as would occur in vivo if desired, with consequent physiologically relevant relative activation levels amongst the different AR sub-types, as well as physiologically relevant interactions amongst their associated signalling pathways.

However, there are no reports, to our knowledge, of studies investigating the independent contributions to the ICaL response of the different AR sub-types in this way, i.e. using NA in the presence of AR sub-type selective antagonists, in human atrial myocytes. Potential species-differences in ICaL responses to NA and AR antagonists should also be considered, in order that data from animal species used in models of AF from adrenergic stimulation and/or altered pathology can be adequately compared with those from human. Rabbits have been studied previously to investigate atrial cellular electrophysiological mechanisms of AF promotion by β-AR stimulation with ISO [20], but ICaL responses to NA with AR antagonists have yet to be studied in this species.

The aim, therefore, is to investigate effects, on NA-stimulated ICaL, of various broad-action and sub-type-specific α- and β-AR antagonists, alone or in combination, in human and rabbit atrial myocytes.

Methods

Patients and animals

Right atrial tissues were obtained from 15 adult patients who were undergoing cardiac surgery, predominantly for coronary artery bypass grafting. All patients were in sinus rhythm on the day of surgery, and none had a history of AF. See Table 1 for patients’ clinical characteristics and drug treatments. Rabbits (n = 15; strain: New Zealand White; supplier: Envigo UK; sex: male; age (mean ± SE [range]): 20 ± 1 [15,16,17,18,19,20,21,22,23,24,25,26,27,28,29] weeks; weight: 3.2 ± 0.1 [2.9–3.7] kg; feeding: ad libitum) were humanely killed by intravenous injection of anaesthetic (100 mg/kg Na+-pentobarbital, via the left marginal ear vein) and removal of the heart, which was retrogradely perfused via the aorta before isolating cardiomyocytes.

Table 1 Patients’ clinical characteristics
Full size table

Atrial cardiomyocytes and voltage-clamp technique

Human and rabbit atrial cardiomyocytes were isolated by enzymatic dissociation (Collagenase Type 1, Lorne Laboratories, Lower Earley, UK) and mechanical disaggregation [22, 47] and stored (≤ 9 h, ~ 20 °C) in cardioplaegic solution (mM): KOH (70), KCl (40), L-glutamic acid (50), taurine (20), KH2PO4 (20), MgCl2 (3), glucose (10), HEPES (10), EGTA (0.5), pH 7.2. The whole-cell-patch voltage-clamp technique was used to record membrane current, in ruptured-patch mode, with an AxoClamp 2B amplifier (Axon Instruments) and WinWCP software (J Dempster). Cardiomyocytes were superfused at 35–37 °C with a physiological salt solution containing (mM) NaCl (140), KCl (4), CaCl2 (1.8), MgCl2 (1), glucose (11) and HEPES (10); pH 7.4. Microelectrodes (1.5–3.0 MΩ resistance) contained (mM) K-aspartate (130), KCl (15), NaCl (10), MgCl2 (1), HEPES (10) and EGTA (0.1); pH 7.25. The resulting liquid–liquid junction potential (+ 9 mV; bath relative to pipette) was compensated for a priori [26]. The low [EGTA]i allows physiological oscillations in cytosolic [Ca2+] during ICaL recordings [20]. ICaL was stimulated with 300 ms duration voltage steps to 0 mV from a holding potential (HP) of − 50 mV (to avoid Na+ current), delivered at 0.33 Hz. Signals were low-pass filtered at 10 kHz. 4-aminopyridine (4-AP; 5 mM) and niflumic acid (0.1 mM) were added to the superfusion solution to suppress contaminating K+ currents (mainly ITO and IKur) and ICl(Ca), respectively.

Drugs and reagents

Noradrenaline (Merck Life Science, Glasgow, UK) was used at 0.01–10 µM [4]; propranolol (Merck) at 0.2 µM [13]; phentolamine (Abcam, Cambridge, UK) at 1 µM [17]; CGP20712A (Merck) at 0.3 µM [5]; ICI118551 (Tocris Bioscience, Bristol, UK) at 0.1 µM [1]; prazosin (Merck) at 0.5 µM [2] and yohimbine (Merck) at 10 µM [16]. Propranolol was racemic, which may affect additionally INa and IK; although this was largely mitigated by the HP and [4-AP] used. The AR sub-type antagonists were chosen for their high selectivity (e.g. CGP: ~ 500-fold selectivity for β1 > β2; ICI > 500-fold β2 > β1 [1]), to effectively dissect individual sub-type responses rather than to mimic clinically used drugs. Indeed, metoprolol, atenolol and bisoprolol may have rather poor selectivity for β1 > β2 [1]. All reagents for storage-, pipette- and superfusion-solutions were supplied by Merck, except for niflumic acid (Tocris).

Data and statistical analysis

Data are expressed as mean ± SEM. Comparisons amongst three or more groups were made using (for matched, parametric data) repeated measures one-way ANOVA or (for un-matched, parametric data) ordinary one-way ANOVA, each followed by Tukey’s multiple comparisons test or (for matched, non-parametric) Friedman test and Uncorrected Dunn’s test. Comparisons between two groups of un-matched data were made using either an un-paired t-test (for parametric data) or Mann–Whitney (non-parametric). P < 0.05 was regarded as statistically significant. All statistical and curve fitting analyses were done using Graphpad Prism 7.00.

Results

Noradrenaline increases human atrial L-type Ca2+ current in a concentration-dependent manner

In human atrial myocytes, NA produced a marked, concentration-dependent increase in ICaL, shown by the original current traces and concentration–response curve in Fig. 1A and B, respectively. At maximally effective concentration, NA increased ICaL ~ 2.5-fold (Fig. 1B). A near-maximally effective, but not saturating, NA concentration (EC75) was chosen for use in all subsequent AR antagonist experiments, calculated from Fig. 1B, i.e. 310 nM.

Fig. 1
figure 1

Noradrenaline (NA) increases human atrial L-type Ca2+ current (ICaL) in a concentration-dependent manner. A Superimposed original representative peak ICaL traces, recorded from a single human atrial myocyte by stepping voltage to 0 mV (from HP − 50 mV) following acute superfusion of NA at concentrations shown. B NA concentration-ICaL density relationship. Values are means ± SE; n = 8–31cells, 4–8 patients. Curve-fit: sigmoidal; variable slope, 4 parameters, no constraints, accounting for n and scatter amongst replicates. Emax and Emin: ICaL at maximally and minimally effective [NA]; EC50 and EC75: 50% and 75% maximally-effective [NA], respectively 

Full size image

Effects of broad action β- and α-adrenoceptor antagonists on NA-stimulated ICaL in human and rabbit atrial myocytes

Broad-action β-, and α-, AR antagonism of ICaL-responses to this NA concentration was studied using propranolol, and phentolamine, applied in a step-wise cumulative fashion, in atrial cells from patients, and also from rabbits for direct comparison (Fig. 2). Following a control period to allow for stabilisation of the normal rate of ICaL run-down (time-dependent decrease following cell rupture), NA superfusion caused a rapid and substantial increase in peak ICaL in all cells studied, with the response stabilising within ~ 1–1.5 min (e.g. Figure 2A and C). In two representative human atrial cells (Fig. 2Ai and ii), propranolol, still in the presence of NA, caused a rapid and substantial decrease in ICaL to below the NA-stimulated response, and subsequently applied phentolamine caused a rapid, and relatively smaller, decrease in ICaL to below the NA + propranolol response. In one of these cells (Fig. 2Aii), propranolol and phentolamine were simultaneously washed off; the resulting increase in ICaL (itself reversible) shows that the NA-stimulatory effect had been preserved throughout the preceding superfusion of the antagonists. In each of nine human atrial cells studied in this way, propranolol decreased, then phentolamine further decreased, the NA-stimulated ICaL. The mean data (Fig. 2B) show that both propranolol and phentolamine significantly decreased ICaL, and that the degree of reduction from phentolamine was significantly smaller than that from propranolol. In rabbit atrial cells, NA also produced a rapid and significant increase in ICaL, and propranolol then caused a rapid, substantial and significant decrease in NA-stimulated ICaL, in each of 7 cells studied (Fig. 2C and D). However, by contrast with the human atrial cells, phentolamine (following propranolol) produced a mixed response, either decreasing (e.g. Figure 2Ci) or increasing (e.g. Figure 2Cii) ICaL, in both cases reversible upon phentolamine-washout. The spread of these phentolamine responses can be seen in Fig. 2D: with a decrease in 4/7 cells (by 21, 29, 44 and 48%; reversible in 3/4), and an increase in 3/7 cells (by 16, 159 and 331%; reversible in each). Moreover, there was no significant effect of phentolamine on average, i.e. in contrast to its significant antagonistic effect in human atrial cells under the same conditions.

Fig. 2
figure 2

Effects of broad action β- and α-adrenoceptor (AR) antagonists on NA-stimulated ICaL in human and rabbit atrial myocytes. A Typical time course of changes in peak ICaL recorded from two human atrial myocytes (i and ii), before (control: “Con”) and during step-wise cumulative addition of noradrenaline (310 nM; EC75: “NA”), propranolol (0.2 µM; β1 + β2-AR antagonist: “Pro”) and phentolamine (1 µM; α1 + α2-AR antagonist: “Phe”). Vertical dashed lines: start time of NA or antagonist addition to (or washout from) perfusion bath. (i and ii) NA-stimulated ICaL is decreased by Pro, then further by Phe; ii shows partial recovery of NA effect after drug washout. B Corresponding average (mean ± SE; with individual points shown) magnitude of responses to NA and AR antagonists as used in A. * = P < 0.05 (ANOVA); n = 9 cells, from 3 patients. C Corresponding rabbit atrial ICaL time courses, in two myocytes (i and ii), with Phe having opposite effects (both reversible) between them. D Average effects of NA and AR antagonists as used in C, in 7 cells, from 4 rabbits. NS = not significant

Full size image

The bi-exponential time course of ICaL inactivation was also examined. NA (310 nM) had no significant effect on the fast (τ1) or slow (τ2) time constants in either species: in human, control τ1 and τ2 were 9.77 ± 1.26 and 112.1 ± 23.2 ms, respectively, vs 7.30 ± 0.64 and 236.9 ± 66.4 ms with NA (P = 0.087 and 0.126, respectively; n = 9 cells); in rabbit: control τ1 and τ2 were 12.58 ± 4.40 and 91.9 ± 20.1 ms, vs 13.54 ± 2.96 and 88.7 ± 12.4 ms with NA (P = 0.781 and 0.797; n = 7 cells).

Comparison of independent anti-adrenergic effects of propranolol and phentolamine

In rabbit atrial cells, effects of broad-action α- and β-antagonism were also studied independently of one other, by using phentolamine in the absence of propranolol (for α-antagonism without concurrent β-antagonism) and, in a different group of cells, vice versa. Propranolol alone again caused a consistent, marked and significant decrease in NA-stimulated ICaL (Fig. 3Ai and Bi). However, phentolamine alone (Fig. 3Aii and Bii), by contrast with phentolamine in the continued presence of propranolol (Fig. 2C and D), also caused a consistent (i.e. in each of 9 cells studied), marked and significant decrease in NA-stimulated ICaL. Furthermore, the degree of the inhibitory effect of phentolamine was not significantly different from that of propranolol.

Fig. 3
figure 3

Comparison of independent anti-adrenergic effects of propranolol and phentolamine. A Representative time courses of ICaL change, in two rabbit atrial myocytes (i and ii), after adding 310 nM NA then either (i) 0.2 µM Pro or (ii) 1 µM Phe. B Corresponding mean ± SE (with individual points shown) responses in (i) Pro group (n = 6 cells, 3 rabbits) and (ii) Phe group (9 cells, 2 rabbits). * = P < 0.05: ANOVA within Pro and Phe groups; un-paired t-test between them (NS = not significant)

Full size image

Comparison of β-AR sub-type contributions to ICaL-stimulation by NA, between human and rabbit atrial myocytes

Having established a substantial β-AR contribution to the stimulatory effect of NA on atrial ICaL, the relative contributions to this of the main β-AR subtypes (β1 and β2) were investigated using CGP20712A (CGP) and ICI118551 (ICI), respectively, again applied in a step-wise cumulative fashion and compared between species. In each of 6 human atrial cells studied (e.g. Figure 4Ai and ii), CGP caused a rapid and marked decrease in NA-stimulated ICaL, with a significant average inhibitory effect (Fig. 4B) similar to that from propranolol (β1 + β2-antagonist) earlier (Fig. 2B). In rabbit atrial cells, CGP had similar effects, both in terms of the comparison with human (Fig. 4C and D vs A and B) and with propranolol (Figs. 4C  and D vs 2 C and D). However, the effects of ICI on NA-stimulated ICaL differed substantially, both when compared with CGP, and between species. Thus, amongst 5 human atrial cells studied with ICI, there was a mixed response: a reversible (upon ICI-washout) increase in 3 cells (e.g. Figure 4Ai), by 12, 35 and 37% (Fig. 4B), and a marked and reversible decrease in 2 cells (e.g. Figure 4Aii), by 72 and 78% (Fig. 4B). There was no significant effect of ICI on average, contrasting with the consistent and significant inhibitory effect of CGP in the same cells (Fig. 4B). In the rabbit atrial cells, by contrast with the human atrial cells under the same conditions, ICI consistently and reversibly (in each of 5 cells studied) increased ICaL (e.g. Figure 4Ci and ii), an effect which was significant on average (Fig. 4D). The degree of ICaL increase by ICI was not significantly different (P = 0.391) to the degree of ICaL decrease by CGP in these cells.

Fig. 4
figure 4

Comparison of β-AR sub-type contributions to ICaL-stimulation by NA, between human and rabbit atrial myocytes. A ICaL time course changes in two human atrial cells (i and ii) in response to a β1-antagonist (CGP20712A at 0.3 µM: “CGP”), then a β2-antagonist (ICI118551 at 0.1 μM: “ICI”), both in the presence of 310 nM NA. (i and ii) NA-stimulated ICaL is decreased by CGP, then either reversibly increased (i) or decreased (ii) by ICI. B Mean effects of interventions in A. n = 5–6 cells, 2 patients; * = P < 0.05, NS = not significant (ANOVA). C Corresponding ICaL time course changes in two representative rabbit atrial cells: in i and ii, CGP again decreased ICaL, but ICI consistently increased it, confirmed by D mean data (showing individual points; n = 5 cells, 4 rabbits)

Full size image

α-AR sub-type contributions to NA-stimulation of ICaL in human and rabbit atrial myocytes

The relative contributions of the main α-AR subtypes (α1 and α2), to the broad α-AR contribution to the stimulatory effect of NA on ICaL, were investigated using prazosin and yohimbine, respectively, again compared between the two species. In each of 6 human atrial cells studied (e.g. Figure 5Ai and ii), prazosin decreased NA-stimulated ICaL, and with a significant effect on average (Fig. 5B). By contrast, yohimbine (still in the presence of NA + prazosin) produced a mixed ICaL response: a moderate decrease in 4 cells (e.g. Figure 5Ai), by 19, 36, 43 and 49% (Fig. 5B); a marked increase in one cell (Fig. 5Aii), by 78%, and no effect in the other cell. There was no significant effect of yohimbine on average, contrasting with the consistent and significant inhibitory effect of prazosin in the same cells (Fig. 5B). The degree of reduction in NA-stimulated ICaL by prazosin in these human atrial cells was significantly smaller (P = 0.002) than that observed with CGP earlier (compare Fig. 5B with Fig. 4B). In rabbit atrial cells, similar to human, prazosin consistently (in each of 7 cells studied) decreased NA-stimulated ICaL (e.g. Figure 5Ci and ii), also significant on average (Fig. 5D). The degree of the ICaL-decrease by prazosin was not significantly different (P = 0.073) from that by CGP earlier (compare Fig. 5D with Fig. 4D). Yohimbine, by contrast with prazosin (and also similarly to the finding in human), produced a mixed ICaL response: a decrease in 6 of these 7 cells (e.g. Figure 5Ci), an increase in the other (Fig. 5Cii) and no significant effect on average (Fig. 5D).

Fig. 5
figure 5

α-AR sub-type contributions to NA-stimulation of ICaL in human and rabbit atrial myocytes. A Typical ICaL changes in two human atrial cells (i and ii) in response to an α1-antagonist (prazosin, 0.5 µM: “Pra”), then an α2-antagonist (yohimbine, 10 µM: “Yoh”), both with NA at 310 nM. B Mean responses to interventions in A. n = 6 cells, 3 patients; * = P < 0.05, NS = not significant (ANOVA). C Corresponding ICaL changes in two rabbit atrial cells (i and ii). D Mean responses (n = 7 cells, 3 rabbits) to same interventions as in C.

Full size image

Discussion

Investigation of independent AR sub-type contributions to NA’s effect on human atrial ICaL first required establishing the NA-ICaL concentration–response relationship, to select a suitable NA concentration for testing with the AR sub-type selective antagonists. We found NA to have a marked, concentration-dependent stimulatory effect on ICaL, with an EC50 of 156 nM, comparable with that in another human atrial study (200 nM) [4], although a markedly higher value has also been reported [6]. Whilst NA circulates in the sub- to low-nanomolar range in humans [35], it is expected to be substantially more concentrated at the adrenergic nerve endings and in cardiac tissues [51]. We selected our EC75 for use in all subsequent experiments (in human and rabbit for their direct comparison) because whilst near maximally effective, this would not saturate the stimulatory response, therefore permitting the antagonists to readily exert their effects. Whilst NA consistently increased ICaL, its subsequent “rundown” (line graphs, Figs. 2, 3, 4 and 5), an accepted limitation of the ruptured-patch technique (due to “a decrease in channel activity with time during recording in dialyzed cells” [43]), required the antagonist responses to be normalised with respect to the previous intervention (bar graphs, Figs. 2, 3, 4 and 5) to compensate for this rundown and thus adequately assess average antagonist effects. Broad action β-AR antagonism (with propranolol) revealed a substantial and consistent contribution to NA’s stimulatory effect on human atrial ICaL from either β1- or β2-ARs or both (since β3-ARs are not expected to be involved in this response [5, 24]). This is congruent with numerous studies in which the broad action AR agonist ISO substantially increased human atrial ICaL [4, 21, 31], although no previous atrial ICaL study could be found in which propranolol was applied following either ISO or NA. In the continued presence of NA plus propranolol, i.e. with the β1- and β2-ARs still antagonised and the α-ARs thus adrenergically activated and solely (independently) amenable to antagonism, broad action α-AR antagonism with phentolamine revealed a substantial and consistent contribution to NA’s stimulatory effect on human atrial ICaL from either α1- or α2-ARs or both. Furthermore, we found that the α-AR contribution to the stimulatory effect of NA on ICaL was significantly smaller (at 37%) than that of the β-AR contribution (at 60%), in human atrial cells. Use of the same protocol in the rabbit atrial cells, i.e. stepwise cumulative addition of NA, propranolol and phentolamine, revealed important species similarities, but also a curious difference regarding the contribution of α-ARs. Thus, whilst propranolol consistently, markedly and significantly antagonised NA’s stimulatory effect on rabbit as well as human atrial ICaL, in rabbit, by contrast with human, phentolamine had a mixed response following propranolol, producing increases in ICaL in some cells, as well as the decreases as seen in human. These ICaL increases by phentolamine were clear, marked and reversible and occurred in approximately half of the rabbit atrial cells studied. By contrast, no ICaL increase was produced by phentolamine in any of the nine human atrial cells studied in this way. Since only the α-ARs were noradrenergically activated at this point in these experiments (β-AR activation prevented by propranolol in both species), such ICaL increases by the α-AR antagonist indicate an inhibitory contribution of independent α-AR activation to the effect of NA on ICaL in those rabbit atrial cells, i.e. attenuating, but not overcoming, the overall effect of NA to increase ICaL. The reason for this mixed effect of phentolamine in the rabbit atrial cells is unknown, but the resulting net (average) absence of effect, as presumably would occur in the syncytium (multicellular), suggests a potentially important species difference that whilst noradrenergic activation of human atrial ICaL involves a significant contribution from α-ARs, this may not be the case in rabbit, at least when the α-ARs are activated independently of the β-ARs. To assess the α-AR contribution to NA’s effect on rabbit atrial ICaL, this time in the presence of simultaneously activated β-ARs, phentolamine was applied in the absence of propranolol and, in a different group of cells, propranolol in the absence of phentolamine for comparison. In this case, we found either α- or β-AR antagonism to consistently (in every cell), markedly and significantly decrease (and by a similar degree between α- and β-) NA-stimulated ICaL, suggesting that the attenuating influence of independent α-AR activation on the stimulatory influence of NA on ICaL as seen above is prevented when α- and β- ARs are simultaneously activated. This finding likely relates to the highly complex interactions which can occur between α- and β-ARs and their signalling pathways [48]. It also highlights another complex, potentially limiting, yet intriguing, aspect of this type of study, the relevance of the order of application of AR-antagonist(s) following NA.

Having established a substantial broad β-AR contribution to NA’s stimulatory effect on atrial ICaL in both species, we then dissected the β1- versus β2-AR involvement, using CGP and ICI, respectively, and showed β1-AR activation to mediate a consistent, substantial and significant contribution to noradrenergic activation of human and rabbit atrial ICaL. The similarity in the magnitude of effect of CGP with that of propranolol, in both species, indicated the prominence of the β1-AR involvement. By contrast, we found β2-AR activation, amongst human atrial cells, to have a mixed, and on average negligible, involvement in the overall β-adrenergic activation of ICaL. This mixed response could relate to stimulatory and inhibitory responses known to result from β2-activation, via Gs and Gi signalling pathways, respectively [45]. In the only similar human atrial ICaL study, in which a synthetic agonist rather than NA was used to activate β2-ARs [50], salbutamol increased the current, which would suggest a stimulatory contribution of β2-activation to its adrenergic activation under their conditions. We found an important species difference regarding β2, since in each of the rabbit atrial cells, independent β2-AR antagonism with ICI (since β1-AR activation prevented by CGP in both species) produced a consistent, reversible, substantial and on average significant increase in ICaL. This indicated a significant inhibitory contribution of β2-AR activation to the effect of NA on rabbit (but not human) ICaL, attenuating the overall effect of NA to increase ICaL, presumably relating to a relatively enhanced Gi signalling response to β2-AR activation [45]. Consistent with this, in rat atrial tissues, β2-antagonism (butoxamine) potentiated the effect of ISO to produce spontaneous contractions [2]. Furthermore, and also in line with the present data, recent studies comparing effects of β1- and β2-AR agonism on rat ventricular ICaL, intracellular Ca2+-cycling and action potentials found that initial β2-AR stimulation suppressed most of the well-characterised changes of cardiac excitation–contraction coupling commonly seen when adding a β1-AR agonist [27, 49].

Dissection of the respective α1- versus α2-AR involvement in NA’s effect on human atrial ICaL (with prazosin and yohimbine) revealed α1-AR activation to mediate a consistent, substantial and significant contribution to noradrenergic activation of the current, but an overall negligible contribution from α2-AR activation. The stimulatory contribution from this α1-AR activation was, nevertheless, significantly smaller (at 37%) than that observed from the β1-AR activation (at 71%). Although no studies of effects of synthetic α-AR agonists on human atrial ICaL could be found, the α1-AR agonist phenylephrine had positive inotropic effects on human atrial muscle strips [14]. These could potentially be explained, at least in part, by the presently observed stimulatory contribution of α1-AR activation on ICaL. However, it should be noted that such inotropic effects could also be due, at least in part, to inhibition of repolarising K+ current, as shown with phenylephrine for human atrial IK1, ITO and IKur [33], or to increased IP3-dependent sarcoplasmic reticular Ca2+ release [41]. No human atrial ICaL studies using prazosin or yohimbine could be found, although there are reports of attenuation by prazosin of NA-induced positive inotropy [32], again congruent with the observed effects of prazosin on NA-stimulated ICaL. In the rabbit atrial cells, we also found a consistent, substantial and significant stimulatory contribution of α1-AR activation to the NA-stimulation of ICaL and a negligible contribution from α2-AR activation. Previous atrial ICaL studies, using synthetic α1-agonists rather than NA, showed either no effect (in rabbit [12] and rat [18]), or a stimulatory effect, in cat [41]. In mice, NA-induced AF was inhibited by prior injection of the α1-antagonist prazosin [34]. Both NA and α1-agonism inhibit rabbit atria ITO [12], carried prominently by Kv1.4 [42]. We blocked ITO using 4-AP, to avoid contaminating ICaL recordings. However, in vivo, ITO decrease from α1-stimulation could exert an action potential prolonging influence additional to that from the present ICaL increase, and other effects of α-stimulation, including pre-synaptic, should also be considered.

Taking our results together, we find that stimulation of atrial ICaL by NA is mediated, based on responses to AR sub-type-antagonists (applied in a set order: sub-type1, followed by sub-type2), mainly by activating β1- and α1-ARs, in both human and rabbit. Whilst α2-AR involvement was negligible in both species and β2-AR involvement negligible in human, in rabbit, β2-activation can attenuate the stimulatory effect of NA on ICaL. Finally, in human (but not rabbit), the contribution of β1-activation to the ICaL stimulatory response to NA was larger than that of α1-activation. An overview of these AR sub-type contributions, with a qualitative estimation of their relative weights, and differences between human and rabbit, is given in Table 2.

Table 2. Relative contributions of AR sub-types to NA-stimulated atrial ICaL
Full size table

These findings have relevance to the electrophysiological mechanisms and potential inhibition of NA-induced AF. Delayed afterdepolarisations (DADs) were produced by catecholamines in dog atria [19], identified as such by their rate-dependent increase in amplitude and decrease in coupling interval [19, 44]. Furthermore, afterdepolarisations of various types were produced or facilitated by ISO in human atrial tissues or cells [28, 31, 40]. DADs are caused by increased inward Na+/Ca2+ exchange current (INa/Ca) associated with increased intracellular Ca2+ loading and Ca2+ waves [10], and it may be argued that NA-induced increase in ICaL could contribute to such Ca2+ loading and thus facilitate DADs. In support, in human atrial myocytes, β-AR stimulation (ISO) increased intracellular Ca2+ spark frequency [25], systolic intracellular [Ca2+] and Ca2+ transient amplitude [6, 38], and Ca2+ waves occurred when intracellular [Ca2+] was elevated by increasing extracellular [Ca2+] [25]. In dog atrial cells, ISO also increased the number of pacing-induced spontaneous Ca2+ transients [7]. Furthermore, NA, which dose-dependently increased the duration of pacing-induced AF in mice [34], also increased intracellular Ca2+ leak and spontaneous sarcoplasmic reticular Ca2+ release in the isolated atrial myocytes in the same study. Perhaps such mechanisms also contribute to an observed concentration-dependent increase in arrhythmic contractions by NA in human [6] and rat [2] atrial tissues. The low [Ca2+]i-buffering used here should allow assessment of NA effects on the atrial ICaL bi-exponential inactivation time course including any influence of Ca2+-induced inactivation of ICaL. We found that NA (310 nM) had no significant effect on either τ1 or τ2 in human or rabbit. No previous studies of NA on atrial ICaL inactivation τs could be found, although ISO was tested in human atrial cells [30]. Despite relatively high [Ca2+]i-buffering (10 mM [EGTA]i) and low temperature (22 °C), τ1 and τ2 were comparable with the present study and, also in agreement, ISO (1 μM) had no significant effect on either [30].

The present data suggest that potential therapeutic targeting of AR sub-types as a means of inhibiting NA-evoked atrial arrhythmias should be most effective with β1-AR antagonism, and potentially more effective with concurrent α1-AR antagonism. This would be consistent both with the clinical use of β1-AR antagonists for preventing postoperative AF [8], and the observation that carvedilol (α1-, β1-, β2-AR-antagonist) was more effective at preventing this arrhythmia than β1-AR antagonists [15, 23], although extra-AR actions of carvedilol [11] might also contribute. However, since α1-AR activation might exert various cardioprotective effects, α1-AR antagonism should nevertheless be considered with caution [52]. Furthermore, potentially therapeutic targeting of selected AR sub-types must be considered in the context of highly complex, dynamic and pathology-dependent interactions between each of the various AR sub-types and their associated signalling pathways [48].