Pacemaker activity and ionic currents in mouse atrioventricular node cells

Abstract

It is well established that pacemaker activity of the sino-atrial node (SAN) initiates the heartbeat. However, the atrioventricular node (AVN) can generate viable pacemaker activity in case of SAN failure, but we have limited knowledge of the ionic bases of AVN automaticity. We characterized pacemaker activity and ionic currents in automatic myocytes of the mouse AVN. Pacemaking of AVN cells (AVNCs) was lower than that of SAN pacemaker cells (SANCs), both in control conditions and upon perfusion of isoproterenol (ISO). Block of I_Na by tetrodotoxin (TTX) or of I_Ca,L by isradipine abolished AVNCs pacemaker activity. TTX-resistant (I_Nar) and TTX-sensitive (I_Nas) Na⁺ currents were recorded in mouse AVNCs, as well as T-(I_Ca,T) and L-type (I_Ca,L) Ca²⁺ currents. I_Ca,L density was lower than in SANCs (51%). The density of the hyperpolarization-activated current, (I_f) and that of the fast component of the delayed rectifier current (I_Kr) were, respectively, lower (52%) and higher (53%) in AVNCs than in SANCs. Pharmacological inhibition of I_f by 3 µM ZD-7228 reduced pacemaker activity by 16%, suggesting a relevant role for I_f in AVNCs automaticity. Some AVNCs expressed also moderate densities of the transient outward K⁺ current (I_to). In contrast, no detectable slow component of the delayed rectifier current (I_Ks) could be recorded in AVNCs. The lower densities of I_f and I_Ca,L, as well as higher expression of I_Kr in AVNCs than in SANCs may contribute to the intrinsically slower AVNCs pacemaking than that of SANCs.

Introduction

The sino-atrial node (SAN) controls the heart rhythm and rate under physiological conditions.¹ The AVN is also endowed with automaticity. It is generally accepted that the AVN drives the heartbeat as a “secondary” supraventricular pacemaker in case of SAN failure.^2,3 Due to its pivotal role in cardiac conduction and as a secondary pacemaker, the structure and the function of the AVN have been extensively studied (recently reviewed in ref. 3). However, little is known about mechanisms underlying AVN automaticity. Some ion channels known to be involved in SAN pacemaker activity have been found in the AVN,^4,5 but their importance in the generation of automaticity in AVNC has not been investigated in depth at the cellular level. Recently, the phenotypes of genetically-modified mouse strains in which genes coding for ion channels involved in cardiac automaticity have been described (reviewed in ref. 6). Genetically-modified mice constitute a promising approach to understand the bases of AVN automaticity, but the electrophysiological properties of spontaneously active mouse AVN cells (AVNCs) have not been investigated directly and we have limited knowledge on the functional role of ionic currents in mouse AVNCs. The aim of the present study was to isolate spontaneously active mouse AVNCs to characterize the properties of AVNCs pacemaker activity, as well as to record the major ionic currents potentially involved in AVNCs automaticity.

Mouse AVNCs displayed an intrinsically slower pacemaker activity than that of SANCs. In comparison to adult rabbit⁷ and mouse⁸ SANCs, automaticity of mouse AVNCs depended on I_Na. A higher density than in SANCs of the fast component of the delayed rectifier current (I_Kr) also characterized mouse AVNCs. Densities of the L-type Ca²⁺ currents (I_Ca,L) were lower in AVNCs than in SANCs. Mouse AVNCs expressed lower densities of the hyperpolarization-activated current (I_f) than SANCs. I_f showed also a more negative half-activation voltage than in SANCs. However, I_f inhibition eventually decreased the slope of the diastolic depolarization in AVNCs, indicating that f-channels contribute to automaticity of these cells. The lower expression of key inward ionic currents in AVN (namely I_f and I_Ca,L), together with an increased expression of I_Kr may explain, at least in part, the slower intrinsic pacing rate of AVNCs. We expect that our work will provide a useful methodological, electrophysiological and functional framework for improving the study of the mechanisms underlying automaticity of AVNCs in genetically engineered mice.

Results

Functional signature of mouse AVNCs.

We first attempted to develop a standard experimental approach to isolate the mouse AVN by using several anatomical landmarks (for details see Material and Methods). Beside these landmarks, we wished to verify independently that our AVN preparations did contain cells with a potentially high degree of automaticity, by using expression of hcn4 as a marker.⁹ We thus stained whole SAN-AVN tissue preparations with an anti-rat HCN4 antibody (Fig. 1A and B). HCN4 positive (presumably pacemaker) AVNCs were present in our AVNC preparations (Fig. 1A and B). In particular, anti-HCN4 immunoreactivity was identified in the SAN as stretches of HCN4-positive cells extending into the AVN region below the inferior vena cava (Fig. 1A). This is consistent with what was reported by Dobrzynski et al.¹⁰ who delineated expression of HCN4 protein in rabbit AVN. Beside the posterior nodal extension of the conduction system identified by Dobrzynski et al.¹⁰ our preparations included also the enclosed node, and a AVN sub-region defined by Boyett et al.¹¹ that was composed by “loosely-packed” atrial cells. Our preparations did not include cells of the mouse His bundle, because the caudal cutting edge of our tissue samples was located on the tricuspid valve. We also checked in pilot experiments using Cx40^EGFP/+ mice¹² that the His bundle was left intact after cutting out the AVN region (see Fig. 2B in the paper by Miquerol et al.¹²). We have thus used all these landmarks to perform our dissections so to include the full mouse AVN region.

Figure 1.

(A) Whole SAN-AVN preparation stained with an anti-rat HCN4 antibody effectively showed the organization of HCN4 positive SANCs and AVNCs. HCN4 positive cells are visible as stretches of cell groups along the crista terminalis and then extending in the posterior part of the lumen of the inferior vena cava. The coronary sinus was cut open for clarity. Distribution of HCN4 immunoreactivity corresponds to the extension of the HCN4-positive conduction system recently described by Dobrzynski et al. in the rabbit AVN.¹⁰ (B) HCN4-positive spindle AVNCs can be seen in a close-up view of the AVN region shown in (A). Abbreviations: right atrium (RA), left atrium (LA), aorta (Ao), crista terminalis (CT); superior vena cava (SVC); inferior vena cava (IVC); coronary sinus (CS); Fossa ovalis (FO); atrioventricular node (AVN); sino-atrial node (SAN). (C) Morphology of myocytes isolated from the mouse AVN. Rod-shaped cells (middle part) and atrial cells (right part) displayed no pacemaker activity. Spindle shaped cells (left part) were beating and were named AVNCs in this study. Note the morphological similarity between the morphology of isolated AVNCs and HCN4 positive cells in (B).

Figure 2.

Automaticity of WT AVNCs is slower than that of SANCs. (A) Examples of spontaneous action potentials in ANVCs and SANCs; consecutive action potentials are superimposed for clarity. The dotted line sets the zero-voltage level. (B) Histograms of pacing rate of WT mouse SANCs (filled boxes) and AVNCs (empty boxes). Perfusion of a maximal dose of ISO (0.1 µM) increased pacemaking of SANCs (C) and AVNCs (D), but AVNCs pacemaking was still slower that of SANCs under the same condition.

Our next objective was to identify the functional signature of “pacemaker” AVNCs in mice. Three cellular morphologies could be identified after enzymatic digestion of the AVN region: a spindle, rod and atrial cell morphology (Fig. 1C). In a typical AVN cell preparation, 49% of the total cell population was comprised of rod shaped cells. These cells have a length of approximately 150 µM and were quiescent. Atrial cells that were similar to myocytes isolated independently from the right atrium constituted 41% of the cellular preparation. Spindle shaped cells accounted for 10% of the total cell number. Spindle cells measured about 50 µM and displayed automaticity.

Spindle- and rod-shaped cells have been reported in the rabbit and guinea-pig AVN by Munk et al. and by Hancox and Levi,⁴ respectively. In our mouse AVN preparations, rod-shaped and atrial-like cells did not appear to be spontaneously active. However, we cannot exclude the possibility that rod-shaped cells can become automatic under particular physiological conditions or in situ in the intact AVN.

Only spontaneously active spindle-shaped cells were used for this study (Fig. 1C). The mean input capacitance of n = 147 spindle shaped cells was 20 ± 1 pA/pF, a value similar to that reported for primary SANCs (20 pF).¹³ Spindleshaped cells from AVN of the mouse (this study), the rabbit⁵ and guinea pig¹⁴ have comparable cellular input capacitance (22 pF in mouse, 29 pF in rabbit and 25 pF in guinea pig). The presence of rod-shaped and atrial-like cells in preparations is consistent with the cellular heterogeneity described in the rabbit AVN.

Pacemaking of mouse AVNCs is slower that that of SANCs and depends on both I_Na and I_Ca,L.

Comparison between pacemaker activity of isolated AVNCs and SANCs showed that the beating rate of mouse AVNCs was significantly lower than that of SANCs (Table 1). The maximum diastolic potential and the action potential threshold were significantly more negative in AVNCs than in SANCs (Table 1) indicating that the diastolic depolarization occurs at a more negative voltage range in AVNCs than in SANCs. To test if ISO application could compensate for the slower basal rate of AVNCs, we applied ISO at 0.1 µM, a concentration that largely saturated the positive chronotropic response to this agonist of mouse SANCs pacemaking (data not shown). The cellular firing rate measured after application of ISO was still higher in SANCs (C) than in AVNCs (D). In ISO, the beating rate of SANCs was 348 ± 7 bpm, while that of AVNCs measured in the same conditions was 222 ± 24 bpm (p < 0.05).

Table 1.

Pacemaker activity and action potential parameters of mouse AVNCs and SANCs

	SANCs	n	AVNCs	n	p
Rate (bpm)	260 ± 21	8	173 ± 27	8	0.0248
MDP (mV)	−53 ± 1	8	−57 ± 1	8	0.0201
Eth (mV)	−35 ±1	8	−41 ± 2	8	0.0119
SDD (mV/ms)	0.09 ± 0.1	8	0.05 ± 0.01	8	0.0322
APA (mV)	88 ± 7	8	91 ±7	8	0.8051
dV/dt (mV/ms)	14 ± 5	8	13 ± 3	8	0.8988
APD (ms)	134 ± 12	8	152 ± 18	8	0.4004

AVNCs displayed slower beating rate (bpm), a more negative maximum diastolic potential (MDP) and action potential threshold (Eth). The slope of the diastolic depolarization (SDD), was lower in AVNCs. No significant differences were observed in the action potential amplitude (APA), upstroke velocity (dV/dt) and action potential duration (APD).

Because genetic mutations affecting the cardiac I_Na channel isoform Na_v1.5 and the L-type Ca_v1.3 channel isoform cause AV conduction dysfunction,^15,16 we investigated the impact of inhibition of I_Na and I_Ca,L on AVNCs pacemaking. Application of 20-µM tetrodotoxin (TTX) blocked action potential discharge (Fig. 3A). The membrane potential of AVNCs exposed to 20 µM TTX was stable at −59 ± 2 mV (n = 8). Inhibition of I_Ca,L by 0.3 µM of the L-type channel blocker isradipine stopped pacemaker activity of AVNCs and the cell membrane potential depolarized to −35 ± 3 mV (n = 6). Only low amplitude oscillations of the membrane potential could be observed in isradipine treated AVNCs (Fig. 3B). These results indicated that pacemaking of mouse AVNCs required both I_Na and I_Ca,L for action potential discharge.

Figure 3.

Automaticity of WT AVNCs is dependent on both I_Na and I_Ca,L. (A) Representative recordings of consecutive action potentials of WT mouse AVNCs in control conditions and after inhibition of I_Na by 20 µM TTX (B).

Voltage-dependent Na⁺ currents in mouse AVNCs.

Both TTXsensitive (I_Na,s) and TTX-resistant (I_Na,r) Na⁺ currents have been recorded in mouse SANCs.^8,17 To test functional expression of I_Na,s we applied 0.1 µM TTX to inhibit this current.⁸ At this concentration, total peak I_Na was reduced by about 20% (Fig. 4). The net peak density of I_Na,s blocked at 0.1 µM TTX at a test potential of −25 mV was (56 ± 12 pA/pF, n = 8), while that of residual I_Na was 287 ± 57 pA/pF, n = 8. Residual I_Na after application of 0.1 µM TTX was not completely blocked at 20 µM TTX (Fig. 3B), demonstrating that this current is attributable to I_Na,r. I_Na,r and I_Na,s displayed distinct V_1/2. Activation of I_Na,r was 11 mV negative to that of I_Na,s. This is consistent with a previous study in mouse SANCs.⁷ The V_1/2 of I_Na,s was significantly more positive than that of I_Na,r (−32 ± 1 mV for I_Na,s and −43 ± 3 mV for I_Na,r, p < 0.05, Fig. 4C).

Figure 4.

TTX-sensitive (I_Na,s) I_Na, in mouse AVNCs. Examples of the effect of 0.1 (A, filled circles) and 20 µM TTX (B, empty circles) on I_Na traces. (C) Examples of the I–V curve of total I_Na (filled circles), I_Na, in presence of 0.1 TTX (hollow circles), I_Na,s (filled squares) in AVNCs. I_Na,s is calculated by subtracting traces recorded in 0.1 µM TTX to total I_Na, traces.

I_Ca,L and I_Ca,T in mouse AVNCs.

Beside I_f, another typical hallmark of pacemaker SANCs is co-expression of “low-voltage activated” T- and L-type Ca²⁺ channel isoforms.^18,19 We thus measured I_Ca,T and I_Ca,L in AVNCs (Fig. 5). The peak current density of I_Ca,T measured at the test potential of −30 mV was 7.9 ± 0.7 pA/pF, a value which is significantly higher that that of the peak I_Ca,L (1.7 ± 0.1 pA/pF, n = 4, V_t −5 mV). The V_1/2 of I_Ca,T was −45 ± 1 mV, while that of I_Ca,L was −22 ± 2 mV. Current densities did not significantly differ between WT C57B/6J mice (this study) and that recorded in AVNCs of 129Sv mice.²⁰

Figure 5.

Recordings of I_Ca,T and I_Ca,L in isolated AVNCs of WT mice. (A) Sample ICa (I_Ca,T + I_Ca,L) traces obtained using a holding potential (Vh) of −90 mV and −55 mV (B) at different test potentials (V_t) as indicated by arrows. The voltage protocol is shown at the top of (A). (C) I–V curves of I_Ca,L and I_Ca,T obtained on WT AVNCs. I_Ca,T + I_Ca,L was recorded using a HP at −90 mV (filled symbols) and I_Ca,L from a HP of −55 mV (open symbols). We calculated the I–V curve of I_Ca,T by subtracting records at HP of −55 mV from that at HP of −90.

Properties of I_f in mouse AVNCs.

I_f was present in all spontaneously-beating AVNCs recorded. The density of I_f was significantly lower in AVNCs than that recorded in SANCs (Fig. 6A). No significant I_f density was recorded in myocytes from the right atrium (Fig. 6A). In comparison, density of the Ba²⁺ sensitive I_K1 component did not significantly vary between AVNCs and SANCs (Fig. 6B). Beside specific differences in current densities, I_f in AVNCs was recorded at more negative potentials than in SANCs (Fig. 6C). Proper measurement of the I_f half-activation voltage (V_1/2),²¹ showed that AVNCs I_f activated about 10 mV more negatively than SANCs I_f (−111 ± 5 n = 5 and −101 ± 1 mV,¹³ respectively, n = 7 p < 0.05, Fig. 7A). The slope factor (v) was not significantly affected (13 ± 1 n = 5 in AVNCs and 12 ± 1, n = 6 in SANCs, p > 0.05). We then tested if, in spite of its more negative activation curve, I_f contributed to pacemaking of AVNCs. We thus used 3 µM of the I_f inhibitor ZD-7228,²² on pacing AVNCs. Under voltage-clamp conditions, a 32% reduction of I_f was observed (Fig. 7B). At this concentration ZD-7228 slowed AVNC pacemaker activity by about 16% (Fig. 7C). We did not test higher concentrations of ZD-7228 because of possible interference with I_Kr inhibition.

Figure 6.

I_f (A) and I_K1 (B) in SANCs, AVNCs and right atrial cells (RACs). Currents were evoked by applying hyperpolarizing voltage steps from a HP of −30 mV, as indicated in the voltage protocol inset. (C) Isochronal I–V curves for I_f in AVNCs (open circles), SANCs (filled circles) and RACs (filled boxes). The peak I_f density measured at −130 mV was lower in AVNCs than in SANCs (16.8 ± 3 n = 5 and 42 ± 4 pA/pF, n = 6 p < 0.01). I_K1 was similarly expressed in AVNCs and SANCs.

Figure 7.

Properties of I_f in mouse AVNCs. (A) I_f activation curve in mouse AVNCs, measured according to DiFrancesco and Mangoni.²¹ Red circles indicate curve fit simulation. (B) 3 µM of ZD-7288 inhibited I_f by 34.5 ± 7.3% (at −115 mV, n = 5). I_f was activated by hyperpolarizing steps as in Figure 6. (C) ZD-7288 reduced AVNCs beating rate by 16 ± 2% (from 286 ± 22 to 242 ± 23 bpm, n = 5, p < 0.05). ZD-7288 reduced the slope of the diastolic depolarization from 2.56 ± 0.5 to 2.17 ± 0.5, (p < 0.05). The action potential duration was not significantly affected by ZD-7288 (from 301 ± 81 to 336 ± 77 ms at 3 µM ZD-7288, p > 0.05).

I_Kr, I_to and I_Ks in AVNCs.

In mouse SANCs, I_Kr plays a fundamental role in controlling the action potential duration and is present during the diastolic depolarization as an outward current.²³ I_Kr was found in all AVNCs tested (n = 20). The peak I_Kr density was higher in AVNCs than that reported in a previous study on mouse SANCs by Clark et al. (3.4 ± 0.4 in AVNCs and 1.5 ± 0.1 pA/pF in SANCs at a test potential of +10 mV) (Fig. 8). The expression of I_to was investigated in AVNCs. In n = 5 AVNCs, a net 4-AP-sensitive I_to current could be recorded (Fig. 8C) (2.1 ± 0.3 pA/pF at a test potential of +10 mV). These cells showed also a relatively large E-4031 and 4-AP insensitive outward K⁺ current component. This latter current is similar to the Ca²⁺ activated K⁺ current mediated by SK2 channels previously described in mouse AVNCs.²⁴ In other cells (n = 9), 4-AP had only a limited effect on outward currents and no statistically significant 4-AP sensitive current could be recorded (not shown). Finally, we studied if mouse AVNCs expressed also the slow component of the delayed rectifier K⁺ current (I_Ks). We used the I_Ks blocker 293B to detect a 293B-sensitive current component. In 11 cells tested, 293B was without significant effect on deactivating tail currents at −45 mV (not shown). This observation indicates that I_Ks is poorly expressed or even absent in automatic mouse AVNCs investigated in this study.

Figure 8.

I_Kr, and I_to in AVNCs. (A) I_Kr in AVNCs. Sample traces in control conditions (top part) and in presence of E-4031 (middle part). I_Kr is evaluated by subtracting traces in E-4031 from control ones (bottom part). The (*) symbol indicates the voltage protocol for records in (A). (B) Averaged IV-curves of I_Kr (open circles) in AVNCs, compared with total current (filled boxes) and current after E-4031 (open boxes). The E-4031-sensitive current started to activate between −40 and −30 mV and peaked at 0 mV. (C and D) Sample traces recorded in control conditions (C, top part) and in presence of 4-AP (C, middle part). I_to is evaluated by subtracting traces in 4-AP from control ones (C, bottom part). (**) symbol indicates the voltage protocol for records in (C). (D) Averaged IV-curve of total current (filled boxes) and current after 4-AP (open boxes). No significant difference was observed between currents in these conditions.

Discussion

The AVN structure and function have been extensively studied (reviewed in ref. 3) however, little is known about the mechanisms underlying AVN automaticity. In this study we present, for the first time, an overview of the electrophysiological properties of spontaneously active mouse AVNCs. We expect that our work will improve the accessibility of this preparation for future studies on AVN pacemaker activity and conduction using genetically engineered mice. Furthermore, our results provide a functional signature of automatic mouse AVNCs and can be useful for studies aiming to characterize AVNCs precursors from the developing mouse heart⁹ or from differentiating stem cells.

Pacemaking of mouse AVNCs.

The AVN is a highly heterogeneous structure containing different cell types that probably serve both SAN impulse conduction and secondary pacemaker activity. Since little is known about the specific cellular mechanisms underlying AVNCs automaticity, we focused on AVNCs that displayed pacemaker activity once isolated from the intact AVN. In addition to widely accepted anatomical landmarks established in rabbit hearts (see the Materials and Methods section), we used also anti-HCN4 staining of intact AVN to identify regions of the AV junction containing automatic cells. This approach maximizes the probability of including all automatic AVNCs in our preparations, even if the fine structure of mouse AVN has not been fully explored. Even if we did not perform a pilot structure-to-function study of mouse AVNCs, we found consistency between the anatomical pattern of anti-HCN4 immunoreactivity in our preparations and previous studies on rabbit conduction system.²⁵ Dobrzynski et al.¹⁰ have reported that the site of origin of automaticity in the rabbit AVN is predominantly located in the posterior nodal extension, even if automaticity can also originate in the enclosed node region. In our preparations, we included the mouse posterior nodal extension, which is HCN4 positive and is thus homologous to the rabbit AVN rhythmogenic centre (Fig. 1A). Also, the gross morphology of HCN4 positive cells in situ is similar to the spindle-like shaped AVNCs used for recordings (Fig. 1B and C). In conclusion, we are confident that spontaneously beating AVNCs described in this study come from rhythmogenic centers of AVN and underlie its automaticity in vivo.

The intrinsic pacemaker activity of mouse AVNCs was significantly slower than that of SANCs (Fig. 2). This difference was maintained after perfusion with ISO. This observation indicates that the slower rate of pacing of mouse AVNCs cannot be explained with a lower basal cAMP concentration, but is due to intrinsic properties of the AVNCs pacemaker mechanism.

I_Na in mouse AVNCs.

Automaticity of mouse AVNCs was highly sensitive to I_Na block by TTX. I_Na thus appears as the predominant mechanisms underlying the action potential upstroke phase. In mouse SANCs high doses of TTX do not stop pacemaker activity, but slow the action potential upstroke velocity as well as the diastolic depolarization phase.⁸ This suggests that the functional role of I_Na in mouse AVNCs can be quantitatively different from that in SANCs. Similarly to SANCs, AVNCs expressed both I_Na,r and I_Na,s. Activation of I_Na,r was 11 mV negative to that of I_Na,s. This is consistent with a previous study in mouse SANCs.⁸ Similarly to SANCs, preliminary data suggest that I_Na,s contributes to the diastolic depolarization of AVNCs, because block of TTX-sensitive channels by 0.1 µM TTX affected the slope of this phase (data not shown). At present, it is difficult to compare the densities of I_Na,r and I_Na,s between AVNCs and SANCs, because in the previous study by Lei et al.⁸ Na⁺ currents were recorded at room temperature and under reduced extracellular Na⁺ concentrations (30 mM). However, the high density of I_Na,r that we recorded in AVNCs and the prominent role of I_Na, in AVNCs action potential discharge are consistent with the severe dysfunction in AVN conduction observed in haplo-insufficient Na_v1.5 mice.¹⁵ Expression of I_Na in AVNCs is also consistent with the fast impulse conduction in mouse AVNCs.

I_Ca,L and I_Ca,T in mouse AVNCs.

Automaticity of AVNCs also depended from I_Ca,L (Fig. 3B). Perfusion of isradipine blocked action potential discharge and depolarized the membrane potential. AVNCs thus responded to I_Ca,L block in a similar way as central SAN tissue strips from the rabbit heart.²⁶ We propose that depolarization of membrane voltage upon I_Ca,L block can be explained, at least in part, by the loss of coupling between L-type channels and SK2 K⁺ channels. Such a functional coupling has been reported in mouse AVNCs,²⁴ as well as in atrial myocytes.²⁷ The density of I_Ca,L recorded in AVNCs was significantly lower than that of SANCs (Fig. 4). However, in spite of the strong expression of I_Na, I_Ca,L is functionally important in AVNCs, because isradipine completely stopped action potential discharge. The absolute density of I_Ca,T in mouse AVNCs was significantly higher than that of I_Ca,L (Fig. 4). This contrasts to what previously observed in rabbit SANCs, where I_Ca,L density is higher than that of I_Ca,T.^18,28 High expression of I_Ca,T in AVNCs, can be an adaptative mechanism to sustain both high basal pacemaking and fast AV conduction in mouse AVN. Use of genetically-modified mice will be useful to understand the functional role of I_Ca,L and I_Ca,T in AVNCs automaticity.

I_f in mouse AVNCs.

In AVNCs, the density of I_f was significantly lower than that recorded in SANCs (Fig. 5). Activation of I_f was shifted negatively in AVNCs by about 10 mV compared to I_f in SANCs. In contrast, I_f activation kinetics did not differ from those recorded in mouse SANCs.^13,29 In a previous study, we showed that the relative abundance of HCN channel isoforms does not significantly differ between the SAN and AVN.³⁰ It is thus possible that subtler differences exist in the regulation of HCN channels in SANCs and AVNCs. In spite of the weaker density and the more negative activation, I_f still contributes to automaticity of AVNCs, because ZD-7228 reduced the slope of the diastolic depolarization. The precise contribution of I_f is difficult to evaluate pharmacologically, because higher doses of ZD-7228 would also affect I_Kr. The use of genetically modified mice is thus important to evaluate the importance of I_f in mouse AVNCs.

In mouse SANCs, I_Kr plays a fundamental role in controlling the AP duration and is present during the diastolic depolarization as an outward current that influences the time course of this phase.²³ We found I_Kr in all AVNCs tested, an observation that suggests that this current can be important also for pacemaking of AVNCs. The predominance of I_Kr upon I_Ks in mouse AVNCs may be a mechanism of adaptation for regulating action potential duration at high heart rates typical of this species. Some AVNCs also expressed moderate quantities of 4-AP sensitive I_to current. The exact location of I_to-expressing cells in the mouse AVN is not known. However, this current can be important in regulating the action potential duration in AVNCs involved in impulse conduction. This hypothesis would be consistent with a previous study showing that mice lacking both the fast and the slow component of I_to display second degree AV blocks.³¹

Materials and Methods

The investigation conforms to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996) and European directives (86/609/CEE).

Morphological identification and definition of mouse AVN.

The mouse AVN was identified as the region between the inferior vena cava (IVC), the coronary sinus (CS) and the tricuspid valve² (Sup. Fig. 1A and B). The anatomical location of tissue samples used for cell isolation is from the mouse atrioventricular junction corresponding to the Koch's triangle as defined in other mammals^2,3,11 (Sup. Fig. 1A and B). This region corresponded to the “AVN region” identified by Rentschler et al. at the inferior part of the interatrial septum and above the mitral annulus. Indeed, in the Rentschler et al. study, hearts from transgenic mice expressing an Engrailed-LacZ reporter construct delineated the mouse cardiac conduction system from the SAN to the distal Purkinje fibers. Finally, we considered as the AVN, the entire region that has been shown to express a pattern of mRNA coding for ion channels similar to that of SAN³⁰ and containing HCN4-positive cells (see Results section). Indeed, Marionneau et al. have reported that ion channel mRNA expression pattern in the mouse AVN region is more similar to that of SAN, a finding that is consistent with the functional expression of ionic currents identified in the present study (see above).

Staining of AVN tissue.

Isolated mouse heart node tissue was placed immediately into 4% paraformaldehyde for 20 minutes at room temperature (RT), followed by soaking in 1x PBS for several hours. The tissue was suspended in 250 ml of 2% Bovine Serum Albumin (Sigma), 500 mg/ml digitonin (Sigma) containing rat HCN4 monoclonal antibody (SHG 1E5, Santa Cruz, 1:100) and sodium azide (0.002%, to prevent bacterial growth) for 66 hrs at RT with very gentle shaking in a LAB-TEK II Chamber Slide. The tissue was rinsed 5 × 500 ml (1x PBS) over a 5-hour period at RT. The procedure was repeated again but this time with donkey anti-rat Alexa 488 secondary (Molecular Probes, 1:500). The tissue was placed in prolong gold (Invitrogen) and a glass coverslip gently placed over the sample. Images were taken with a Leica confocal macroscope (Leica TCS LSI) at the Montpellier RIO Imaging facility of the Arnaud de Villeneuve Campus (Montpellier, France). The final images represent a snapshot of a 3D reconstructed tiled image using IMARIS software package (PC, version 7).

Mouse AVNCs were isolated using the same protocol as used for patch clamping. AVNCs were placed into chambers and allowed to attach to Cell-Tak coated wells (3.5 µg/cm²) for 1 hr (LAB-TEK II Chamber Slide, NUNC). Cells were then fixed with 4% paraformaldehyde for 20 minutes at room temperature (RT), followed by soaking in 1x PBS for several hours. 250 µl of 2% Bovine Serum Albumin (Sigma) containing rat HCN4 monoclonal antibody (SHG 1E5, Santa Cruz, 1:100), anti mouse HA monoclonal (clone HA7, 1:250) were added and allowed to incubate overnight at 4° in a humid chamber. The tissue was rinsed for 4 × 15 minutes (1x PBS). Goat anti-rat FITC secondary (Molecular Probes, 1:500), donkey anti mouse Alexa-647 (Molecular Probes, 1:500), DAPI was incubated for 90 minutes at RT followed by 4 × 15 minutes (1x PBS). Prolong gold (Invitrogen) was overlaid and a glass coverslip gently placed over the sample. Images were taken with a Leica confocal microscope (Leica SPE) at the Montpellier RIO Imaging facility of the Arnaud de Villeneuve Campus (Montpellier, France). Image analysis was carried out on Metamorph Image Analysis software.

The specificity of Rat HCN4 antibody was confirmed using a mouse monoclonal HCN4 antibody (Neuromab), which had the identical overlapping staining pattern. Furthermore, SANCs expressing HCN4-HA tagged subunits showed overlapping staining using rat HCN4/mouse HA antibodies or mouse HCN4/rat HA antibodies (data not shown).

Isolation of mouse SANCs and AVNCs.

Beating hearts were removed under general anesthesia, consisting of 0.01 mg/g of Xylazine (Rompun 2%, Bayer AG, Leverkusen Germany), 0.1 mg/g of Ketamine (Imalgène, Merial, Bourgelat France) and 0.2 mg/g of Na-Pentobarbital (CEVA, France). The SAN and the AVN were exposed by using the landmarks shown in Supplemental Figure S1A and B and then excised in pre-warmed (35°C) Tyrode solution containing (mM/L): NaCl, 140; KCl, 5.4; CaCl₂, 1.8; MgCl₂, 1; Hepes-NaOH, 5; and D-glucose, 5.5; (adjusted to pH = 7.4 with NaOH). SAN tissue strips were then transferred into a “low-Ca²⁺-low-Mg²⁺” solution containing (in mM/L): NaCl, 140; KCl, 5.4; MgCl₂, 0.5; CaCl₂, 0.2; KH₂PO₄, 1.2; taurine, 50; D-glucose, 5.5; bovine serum albumin (BSA), 1 mg/ml; Hepes-NaOH, 5; (adjusted to pH = 6.9 with NaOH). SAN and AVN tissues were digested by adding collagenase type II or IV (229 U/ml, Worthington Biochemical Corporation, Lakewood, NJ USA), elastase (1.9 U/ml, Worthington Biochemical Corporation, Lakewood, NJ, USA), protease (0.9 U/ml, Sigma, St. Quentin Fallavier, France), BSA 1 mg/ml and 200 µM CaCl₂. Digestion was carried out for a variable time of 20–25 minutes at 35°C. Tissue strips were then washed and transferred into a modified “Kraftbrühe” (KB) medium³³ containing (in mM/L): L-glutamic acid, 70; KCl, 20; KOH, 80; (±)D-b-OH-butyric acid, 10; KH₂PO₄, 10; taurine, 10; BSA, 1 mg/ml; and Hepes-KOH, 10; (adjusted to pH = 7.4 with KOH). Single SAN and AVN cells were isolated by manual agitation in KB solution at 35°C for 2–5 minutes.

Cellular automaticity was recovered by re-adapting the cells to a physiological extracellular Ca²⁺ concentration by addition of a solution containing (in mM/L): NaCl, 10, CaCl₂, 1.8 and normal Tyrode solution containing BSA (1 mg/ml). The final storage solution contained (mM/L): NaCl, 100; KCl, 35; CaCl₂, 1.3; MgCl₂, 0.7; L-glutamic acid, 14; (±)D-b-OH-butyric acid, 2; KH₂PO₄; 2; taurine, 2; BSA 1 mg/ml; (pH = 7.4) and gentamycin (50 µg/ml). All chemicals were from Sigma (St. Quentin, Fallavier), except for the (±)D-β-OH-butyric acid that was from Fluka Chemika (Buchs, CH), TTX and ZD 7288 that were from Tocris Bioscience (Ellisville, USA), 293B was a generous gift by Dr. Lang and E-4031 that was a generous gift by Dr. Flavien Charpentier (L' Institut du Thorax, Nantes).

Electrophysiological recordings.

For electrophysiological recordings, aliquots of the cell suspension were harvested in custom made recording chambers (working volume 500 µL) allowing unidirectional flow of solutions and mounted on the stage of an inverted microscope (Olympus, X71), and continuously perfused with normal Tyrode solution. The recording temperature was set to 36°C. The whole-cell variation of the patch-clamp technique³⁴ was used to record cellular ionic currents, by employing an Axopatch 200A (Axon Instruments Inc., Foster USA) patchclamp amplifier, connected to the ground by an agar bridge filled with 150 mM KCl. Cellular automaticity was recorded by the perforated patch technique with escin (50 µM). Recording electrodes were fabricated from borosilicate glass, by employing a Flaming-Brown microelectrode puller (Sutter, Novato, CA USA). For recording cell automaticity, as well as I_f and K⁺ currents, we used an intracellular solution containing (mM/L): K⁺-aspartate, 130; NaCl, 10; ATP-Na⁺ salt, 2; creatine phosphate, 6.6; GTP-Mg²⁺, 0.1; CaCl₂, 0.04 (pCa = 7); Hepes-KOH, 10; (adjusted to pH = 7.2 with KOH). Electrodes had a resistance of about 5 MΩ. Seal resistances were in the range of 2–5 GΩ.

For recording of calcium currents (I_Ca,T, I_Ca,L), we replaced K-Aspartate and KCl in the intracellular solution, with an equal amount of CsCl. The extracellular solution contained (in mM/L): tetraetylammonium-chloride (TEA-Cl), 130; CaCl₂, 2; MgCl₂, 1; 4-amino-pyridine, 10; Hepes, 25; (adjusted to pH = 7.4 with TEAOH). I_Kr was recorded in Tyrode, after addition of 10 µM TTX and 0.3 µM isradipine. I_f was routinely recorded Tyrode solution containing 5 mM BaCl₂ to block I_K1.³⁵ We performed data acquisition by using the pClamp software (ver. 9, Axon Instruments Inc.).

Data analysis.

The AP parameters were calculated as reported previously in reference ^{8, 36} and ³⁷. The following action potential (AP) parameters were calculated: the AP duration (APD), the slopes of the diastolic depolarisation (SDD), the maximum diastolic potential (MDP), the AP threshold (E_th), the maximum upstroke velocity (the peak of the first derivative of the AP waveform dv/dt), and the AP amplitude (APA). The cell membrane capacitance has been monitored by applying brief (10 ms) voltage steps of ±10 mV amplitude from a holding potential (HP) of −35 mV.

The I_f activation curve was measured according to a protocol described by DiFrancesco and Mangoni²¹ and fitted according to the equation:

• P(V) = 1/1 + exp[(V − V_1/2/v]

where P(V) is the degree of activation of macroscopic I_f and is equivalent to the voltage dependency of the probability of opening of single f-channels,²¹ V_1/2 is the voltage for current half activation, and v is the slope factor. Cav-mediated Ca²⁺ currents and I_Na were analyzed as previously reported in references 20 and 37.

We performed analysis by employing the Origin Lab software (ver. 7.5, Microcal Inc., Northampton, MA USA). Results are presented as means ± the standard error of the mean (SEM, number of cells). For calculating the level of significance, Student's t-tests, the one- or two-way ANOVA tests followed by Tukey's post-hoc tests and non-parametric Kruskal-Wallis tests were employed. When testing statistical differences results were considered significant with p < 0.05. In all figures *p < 0.05, **p < 0.01 and ***p < 0.001, respectively.

Abbreviations

AVN

atrioventricular node

SAN

sino-atrial node

TTX

tetrodotoxin

I_Na

voltage-dependent Na⁺ current

I_Ca,L

L-type Ca²⁺ current

I_Ca,T

T-type Ca²⁺ current

I_f

hyperpolarization-activated f-current

I_Kr

delayed rectifier K⁺ current

I_to

transient outward K⁺ current

Financial Support

This work has been supported by CavNet a Research Training Network (RTN) funded through the European Union Research Programme (6FP) MRTN-CT-2006-035367, the INSERM National Program for Cardiovascular Diseases (PNRC), the Fondation de France (Cardiovasc 2008002730), the Agence Nationale pour la Recherche (ANR-06-PHISIO-004-01), the NIH R01HL087120-A2 grant, the Deutsche Forschungsgemeinschaft (DFG, IS63/1-1/2) and the Austrian Science Fund (P-20670). Pietro Mesirca and Angelo Torrente are Ca_vNet fellows.