Large-conductance Ca2+- and voltage-activated K+ (BK) channels play prominent roles in shaping muscle and neuronal excitability. In the cardiovascular system, BK channels promote vascular relaxation and protect against ischemic injury. Recently, inhibition of BK channels has been shown to lower heart rate in intact rodents and isolated hearts, suggesting a novel role in heart function. However, the underlying mechanism is unclear. In the present study, we recorded ECGs from mice injected with paxilline (PAX), a membrane-permeable BK channel antagonist, and examined changes in cardiac conduction. ECGs revealed a 19 ± 4% PAX-induced reduction in heart rate in wild-type but not BK channel knockout (Kcnma1−/−) mice. The heart rate decrease was associated with slowed cardiac pacing due to elongation of the sinus interval. Action potential firing recorded from isolated sinoatrial node cells (SANCs) was reduced by 55 ± 15% and 28 ± 9% by application of PAX (3 μM) and iberiotoxin (230 nM), respectively. Furthermore, baseline firing rates from Kcnma1−/− SANCs were 33% lower than wild-type SANCs. The slowed firing upon BK current inhibition or genetic deletion was due to lengthening of the diastolic depolarization phase of the SANC action potential. Finally, BK channel immunoreactivity and PAX-sensitive currents were identified in SANCs with HCN4 expression and pacemaker current, respectively, and BK channels cloned from SANCs recapitulated similar activation as the PAX-sensitive current. Together, these data localize BK channels to SANCs and demonstrate that loss of BK current decreases SANC automaticity, consistent with slowed sinus pacing after PAX injection in vivo. Furthermore, these findings suggest BK channels are potential therapeutic targets for disorders of heart rate.
- large-conductance channel
- calcium-activated potassium channel
- heart rate
- potassium channels
- sinoatrial node
large-conductance Ca2+- and voltage-activated K+ (BK) channels (KCa1.1), encoded by the Kcnma1 gene (Slowpoke, Slo1) (10, 21, 55), are characterized by their large single channel conductance, selective pharmacology, and allosteric regulation by Ca2+ and voltage (2, 10, 18, 52). BK currents are found in multiple excitable cell types, including myocytes and neurons, where they generally suppress membrane excitability by opposing Ca2+ influx, promoting action potential (AP) repolarization, and hyperpolarizing the membrane potential (36, 60, 70). Several physiological functions are dependent on BK channels, including blood pressure regulation (8, 60, 64), urinary bladder contraction (55), erectile function (78), circadian rhythmicity (56, 58, 59), and locomotor control (55, 65).
BK channels are highly expressed in vascular smooth muscle, where they regulate myogenic tone (6, 60) and therefore influence blood pressure (8, 64), cerebrovascular circulation (23), and coronary blood flow (4). BK channels provide negative feedback on vascular tone by linking membrane depolarization to local increases in intracellular Ca2+ (60). In vascular smooth muscle cells, Ca2+-mediated activation of BK current hyperpolarizes the membrane, reducing Ca2+ entry through voltage-gated Ca2+ channels and resulting in relaxation (77). Alterations in BK currents and channel expression have also been shown to be associated with pathological dysregulation of vascular function in diabetes (54), heart failure (75), and aging (51).
In contrast to their role in the vasculature, BK channels are not well characterized in the heart. BK channel expression is low in rodent and human hearts compared with other tissues (27, 33, 74). A few studies have suggested that BK currents are present in cardiac sympathetic nerves (32, 46), vagal parasympathetic nerves (80, 82), Purkinje fibers (11), and atrial and ventricular myocytes (73). More recent evidence supports the hypothesis that BK channels are also directly involved in heart function. BK agonists promote cardioprotection against ischemia-reperfusion injury (1, 81), an effect that is precluded in the presence of BK antagonists or in Kcnma1−/− mice that lack functional BK channels (67). The basis for these effects has been suggested to rely on mitochondrial localization (BKMito), whereby activation of BKMito leads to a reduction in ROS production and inhibition of mitochondrial permeability transition pore opening (14, 40). However, these studies did not address whether BK channels regulate cardiac excitability.
Supporting a role for BK channels in cardiac excitability, the BK antagonist paxilline (PAX) has been shown to reduce heart rate in intact mice and isolated rat hearts (31). PAX caused acute bradycardia when injected into wild-type (WT) mice but not Kcnma1−/− mice, demonstrating that the effects of PAX require functional BK channel expression. Additionally, PAX and iberiotoxin (IbTX), a highly selective BK antagonist (12, 25), caused bradycardia when perfused through isolated rat hearts (31), suggesting that cardiac-localized BK channels mediate this effect. Furthermore, because IbTX is not membrane permeable (25), the reduction in heart rate was not expected to occur through BKMito, which would require intracellular penetration of the antagonist to produce the observed effects.
In the present study, we sought to identify the mechanism of BK antagonist-induced bradycardia. We examined the effect of PAX on intact, conscious mice by ECG and found slowed cardiac pacing and conduction. Initiation of the heart beat and regulation of heart rate are determined by the sinoatrial node (SAN), the primary cardiac pacemaker. To establish the presence of the BK antagonist-induced effect in the SAN, electrophysiological and expression experiments were performed on acutely isolated SAN cells (SANCs) from the mouse heart. We showed a reduction in SANC excitability by decreased AP firing after the application of BK antagonists and in Kcnma1−/− SANCs. Expression experiments and voltage-clamp recordings were performed to substantiate the presence of BK channels and currents in SANCs. Finally, BK channels expressed in SANCs were cloned, characterized, and compared with native currents.
MATERIALS AND METHODS
WT and Kcnma1−/− mice were maintained on inbred FVB/NJ or C57BL/6J backgrounds. Adult mice (1–5 mo old) were used for experiments. Telemetry recordings (Fig. 1), experiments testing the effect of PAX on WT SANCs (Fig. 2B), and RT-PCR (Fig. 5A) were performed on mice on a FVB/NJ background. All other experiments were performed on mice on a C57BL/6J background because these cells were easier to prepare. All procedures involving mice were approved by the Animal Care and Use Committee from the University of Maryland and University of Iowa School of Medicine.
Telemetric recordings of mouse ECGs.
ECG telemetry transmitters (ETA-F10, Data Sciences, New Brighton, MN) were surgically implanted either subcutaneously or intraperitoneally into mice (FVB/NJ background) in a lead II configuration. Mice were placed under general anesthesia via isoflurane/O2 inhalation during surgeries. Device bodies were inserted ventrally, either subcutaneously or intraperitoneally. ECG leads were tunneled subcutaneously from the device body and sutured on the upper right chest muscle (negative lead) and the upper left abdominal wall, caudal to the rib cage (positive lead) (69). Mice were allowed to recover from surgeries for at least 1 wk before experiments.
PAX (no. 2006, Tocris, Minneapolis, MN) or DMSO vehicle were administered to mice by intraperitoneal injection at a final concentration of 8 mg/kg (PAX) or an equivalent volume of DMSO. ECG waveforms were continuously collected from 1 h before injection to 5 h postinjection from mice in their home cage environment. Resting baseline heart rates were determined using the average of five 10-s segments based on R-R intervals calculated with Dataquest Art 4.0 software (Data Sciences) during a period of inactivity, as determined from telemetric activity measurements. ECG intervals were calculated from one 10-s segment using mice with clearly discernible waveform components. P-P and P-R intervals were determined using the ECG Analysis Module in the Ponemah 5.0 Software package (Data Sciences).
Mouse SANC isolation.
Mouse SANCs were acutely isolated as previously described (79). After euthanasia, mouse hearts were excised and rinsed in normal Tyrode solution containing (in mM) 140 NaCl, 5 HEPES, 5.5 glucose, 5.4 KCl, 1.8 CaCl2, and 1 MgCl2, with the pH adjusted to 7.4 with NaOH. The SAN region was dissected from the right atria, cut into strips, and rinsed in low-Ca2+ Tyrode solution containing (in mM) 140 NaCl, 5 HEPES, 5.5 glucose, 5.4 KCl, 0.2 CaCl2, 0.5 MgCl2, 1.2 KH2PO4, and 50 taurine with 1 mg/ml BSA, with the pH adjusted to 6.9 with NaOH. Tissue strips were digested at 35°C in low-Ca2+ Tyrode solution containing collagenase type II (LS004174, Worthington, Lakewood, NJ), elastase (LS002292, Worthington), and protease type XIV (P5174, Sigma-Aldrich, St. Louis, MO) for 25–30 min. Tissue was transferred to Kraft-Bruhe (KB) solution containing (in mM) 100 potassium glutamate, 5 HEPES, 20 glucose, 25 KCl, 10 potassium-aspartate, 2 MgSO4, 10 KH2PO4, 20 taurine, 5 creatine, and 0.5 EGTA with 1 mg/ml BSA, with the pH adjusted to 7.2 with KOH. Tissue was triturated in KB solution using a wide-bore glass pipette for 5–10 min to dissociate SANCs. Cells were stored at 4°C in KB solution and used within 8 h of isolation.
RT-PCR and immunocytochemistry.
RNA was isolated from WT or Kcnma1−/− cells pooled from the SAN region using cell lysis buffer (Signosis, Santa Clara, CA). RT-PCR was performed with total RNA using the MyTaq One-Step RT-PCR kit (Bioline, Taunton, MA) and primers (listed below) according to the manufacturer's protocol. Reactions were performed on PTC-200 Thermal Cycler (MJ Research, Alameda, CA) as follows: cDNA was synthesized using gene-specific primers (Kcnma1: 5′-TCTGTAAACCATTTCTTTTCT-3′ and Hcn4: 5′-GGATGGAGTTCTTCTTGCCTAT-3′) at 48°C for 1 h. PCR products were amplified with 40 cycles of 95°C for 10 s, 60°C for 10 s, and 72°C for 45 s followed by a final extension at 72°C for 7 min. The amplification primers used in this reaction were as follows: Kcnma1, forward 5′-CATCATACCGGTGACCATGGA-3′ and reverse 3′-TTCTATTTGTTACCGAGGTC-5′; and Hcn4, forward 5′- TGCTGTGCATTGGGTATGGA-3′ and reverse 3′-GTAGTTGAAATTGACGGCTTT-5′. BK and HCN4 PCR products (206 and 337 bp, respectively) were visualized on an ethidium bromide-stained agarose gel (1.5%) by electrophoresis.
To clone BK channels expressed in the SAN, individual SANCs were visually identified by beating, and 20 cells were pooled for lysis and RNA extraction. BK channel cDNAs covering splice sites 1–4 (66) were amplified by RT-PCR as described above using the following sequences: reverse transcriptase primers, 5′-TCTGTAAACCATTTCTTTTCT-3′ and 5′-GCCGCTCTTCCTGAACGTACTT-3′; and PCR primers (both variants), forward 5′-GAGTACAAGTCTGCCAACAG-3′ and reverse 5′-TCACCAGGGTCCGTATTAGG-3′. Products were gel purified and subcloned, and two full-length products were obtained and sequenced: BKVYR (GenBank Accession No. KF530042) and BKQEERL (GenBank Accession No. KF530043). Sequences were subcloned into pcDNA3.1 for expression in human embryonic kidney (HEK)-293T cells.
For immunocytochemistry, SANCs were allowed to settle on fibronectin-coated glass slides for 1 h followed by fixation for 20 min in 2% paraformaldehyde. Cells were then permeabilized and blocked in PBS containing 2 mg/ml BSA and 0.075% Triton X-100. Cells were incubated overnight at 4°C with 1:100 dilutions of rabbit α-BK (APC-021, Alomone Labs, Jerusalem, Israel) and mouse α-HCN4 (N114/10, Neuromab, Davis, CA). α-BK and α-HCN4 immunoreactivity were visualized with 1:1,000 Alexa Fluor 488 and 568 secondary antibodies (Life Technologies, Grand Island, NY), respectively. After three PBS washes, cells were mounted onto glass slides using Vectashield (Vector Laboratories, Burlingame, CA).
Tyramide signal amplification was performed with a TSA kit (T20934, Life Technologies) according to the manufacturer's recommended protocol using a 1:1,000 dilution of rabbit α-BK (Alamone Labs) and 1:1,000 tyramide-labeled Alexa Fluor 568 antibody. α-HCN4 immunoreactivity was performed as described above except with 1:1,000 mouse Alexa 488 secondary antibody (Life Technologies). After three PBS washes, cells were mounted onto glass slides using Vectashield (Vector Laboratories).
Images were collected at ×40 using a LSM 5 Live confocal microscope (Zeiss, Jena, Germany). SANCs were identified by membrane labeling of HCN4 expression. Fluorescence intensity measurements were performed blind of genotype and analyzed using Zen 2009 software (Zeiss). Cell perimeters were outlined using the DIC image, and the average raw pixel intensity of each fluorophore was calculated within this area. Acquisition and gain settings for BK channel expression detection were identical for all cells.
For AP recordings, cells were continuously perfused with normal Tyrode solution at 0.7–1.25 ml/min at 35 ± 1°C. SANCs were identified by their characteristic morphology, spontaneous contractions, and ability to spontaneously fire APs. Perforated-patch AP recordings were acquired in current-clamp mode using a Multiclamp 700B amplifier (Molecular Devices, Sunnyvale, CA) at 20 kHz. Borosilicate glass pipettes (3–6 MΩ) were filled with (in mM) 130 potassium aspartate, 10 NaCl, 10 HEPES, 0.04 CaCl2, 2 MgATP, 7 phosphocreatine, and 0.1 NaGTP with 240 μg/ml amphotericin B, with the pH adjusted to 7.2 with KOH. Stable perforations [access resistance (Ra) <50 MΩ] were obtained within 5–13 min, and only SANCs with stable baseline AP firing patterns (no pausing) were analyzed. Frequency analysis was performed on 1-min segments of data using threshold-based detection methods, 1 min after stable perforation was established (baseline) or within 15 min after PAX or IbTX (no. 1086, Tocris) application, at the time of the peak effect observed. For experiments conducted in the presence of isoproterenol (Iso; I6504, Sigma-Aldrich; Fig. 3), frequency analysis was performed on 1-min segments of data as described within 5 min of drug application. Diastolic depolarization duration measurements were calculated as the interval between the maximum diastolic potential (MDP) and the threshold potential, which was determined to be 10% of dV/dtmax, as previously described (43). Half-maximal AP duration (APD50) was defined as the time between the threshold potential and 50% of repolarization. AP parameters were determined from the average of 5–7 APs/cell.
Macroscopic current recordings from SANCs were performed in the whole cell configuration at 30 ± 1°C during continuous perfusion of Tyrode solution at 1.25 ml/min. Data were acquired with the Multiclamp 700B in voltage-clamp mode at 20 kHz. Patch pipettes (1–3 MΩ) were filled with intracellular solution containing (in mM) 128 potassium aspartate, 6.6 sodium phosphocreatine, 7 KCl, 1 MgCl2, 0.05 CaCl2, 10 HEPES, 5 EGTA, and 4 Mg-ATP, with the pH adjusted to 7.2 with KOH. PAX was prepared in DMSO. Currents were evoked from a holding potential of −75 mV followed by 50-ms voltage steps from −130 to +60 mV. All potentials were corrected for liquid junction potential (14.9 mV). Currents were averaged from four to five voltage families at baseline and after 3 μM PAX application. Ra was <10 MΩ, and only recordings where Ra changed <15% were used for analysis. Series resistance was compensated at 60%. SANCs were distinguished from atrial cells based on the presence of pakemaker current (If), which was evoked by 500-ms voltage steps from −65 to −145 mV, in the presence of 1 mM Ba2+. For representative If traces, the time-independent leak current was subtracted from current evoked from hyperpolarizing voltage steps. Atrial cells had no If activation. For macroscopic BK current recordings from cloned BK channels, BKVYR and BKQEERL cDNAs (in pcDNA3.1) were transiently transfected into HEK-293T cells and recorded 24–48 h later. For whole cell recordings, the external (bath) and internal (pipette) solutions and voltage protocols were identical to those used for BK current recordings in mouse SANCs. For inside-out patch recordings, the external (pipette) solution consisted of normal Tyrode solution with Ca2+ omitted. The internal (bath) solution was identical to the internal solution for whole cell recordings. Free Ca2+ solutions (1 and 50 μM) were buffered with 5 mM EDTA. Currents were elicited from a holding potential of −75 mV followed by 50-ms voltage steps from −130 to +110 mV.
Data are presented as means ± SE. Significance was determined at P < 0.05 using the indicated test. Data within each condition were tested for normality (Prism 5.0, GraphPad, La Jolla, CA). For normally distributed data, ANOVAs and t-tests were used as indicated in the figures. A nonparametric Mann-Whitney test was used when one condition did not meet the criteria for normal distribution: WT and Kcnma1−/− AP data (Fig. 2D) and tyramide signal amplification immunohistochemistry (Fig. 5D). Statistical tests were performed in Origin (version 8.5, OriginLab, Northampton, MA).
BK channel inhibition decreases heart rate and slows cardiac pacing in vivo.
Previous work has established both the requirement for BK channels and the cardiac-specific locus for PAX-induced bradycardia in the rat and mouse (31). Here, to determine the mechanism by which BK channel inhibition slows heart rate, we recorded ECGs from awake, unrestrained mice via implantable telemetry before and after 8 mg/kg PAX injection (Fig. 1, A and B). The maximal reduction in heart rate in WT mice was 19 ± 4% 2 h after PAX injection (from 611 ± 18 to 485 ± 14 beats/min, P = 0.0001 by Student's t-test). No reduction in heart rate was observed after PAX injections into Kcnma1−/− mice, which lack expression of functional BK channels (Fig. 1A), corroborating previous results (31). Prior measurements from Kcnma1−/− mice under restraint also showed similar baseline heart rates compared with WT mice (31), raising the possibility that restraint-induced stress could have provided compensation for loss of BK function in the genetic deletion model. However, in the present study, baseline heart rates were also similar between unrestrained mice of each genotype (WT mice: 611 ± 18 beats/min, n = 13, and Kcnma1−/− mice; 593 ± 34 beats/min, n = 7, P = 0.60; Fig. 1A). These results demonstrate that acute inhibition of BK channels reduces heart rate under basal conditions, identifying a basic role for BK channels in cardiac function, and suggest that chronic loss of BK current induces compensatory heart rate stabilization in Kcnma1−/− mice.
Further examination of ECG recordings revealed that PAX did not induce any type of arrhythmia, such as sinus pauses, in WT or Kcnma1−/− mice (Fig. 1, B and C). Rather, the reduction in heart rate was a stable slowing of the entire cardiac cycle. To identify how BK channel inhibition altered the cardiac cycle, we analyzed specific ECG intervals from mice before and after PAX injection. The decreased heart rate (R-R interval) after PAX injection in WT mice was associated with significant elongation of the P-P sinus rhythm (Fig. 1B). However, atrioventricular node conduction (P-R) was not affected, suggesting that BK channels selectively affect heart rate by actions at the SAN. In support of this, we found that the elongated P-P interval (ΔP-P after PAX: 31 ± 5 ms) matched the slowed R-R interval (ΔR-R after PAX: 31 ± 6 ms; Fig. 1B), accounting for all of the R-R increase. This observation demonstrates that slowed sinus rhythm primarily accounts for the decreased heart rate after PAX injection. Intervals analyzed from Kcnma1−/− mice were equivalent to WT mice (Fig. 1B), revealing that the normal heart rates in Kcnma1−/− mice did not rely on altered cardiac conduction. Given the slowed sinus interval, we focused on the SAN, because its activity sets the P-P interval. We therefore characterized the effect of BK channel inhibition on SANC excitability.
Loss of BK channel function decreases SANC excitability.
The heart rhythm originates from spontaneous depolarization of SANCs, resulting in AP firing (49). To test whether the PAX-induced decrease in cardiac pacing in vivo was due to a direct reduction in SANC firing rate, we tested the effect of PAX on isolated mouse SANCs via perforated-patch recordings (Fig. 2A). SANCs were identified visually based on their elongated spindle shape and spontaneous beating after bath perfusion of Tyrode solution containing 1.8 mM Ca2+ (48). Baseline firing rates were recorded after stable perforation. The average firing rate from isolated WT SANCs was 286 ± 19 beats/min, an autonomous firing rate in the absence of autonomic control that was consistent with a variety of other studies (15, 47, 79). Perfusion of 3 μM PAX decreased firing in seven of eight cells that responded to drug application (124 ± 46 beats/min; Fig. 2, B and C), with no effect on one cell. These data show acute BK inhibition reduces AP firing in isolated SANCs, consistent with the PAX-induced decrease in heart rate after in vivo injection.
AP waveforms were analyzed to determine the mechanism of the slowed firing rate. After PAX application, the major effect was a significant slowing of diastolic depolarization, a critical interval integrating the multiple ionic conductances initiating the SANC AP (49). Diastolic depolarization duration increased almost twofold (Fig. 2G). In contrast, neither AP repolarization, as measured by ADP50, nor MDP were significantly altered by PAX (Fig. 2, H and I). These results suggest that the major effect of BK channel inhibition is an alteration of the rate of SANC depolarization, delaying AP initiation. Furthermore, to verify that the PAX effect on diastolic depolarization did not stem from off-target inhibition of If, which could also produce prolongation of diastolic depolarization, we applied PAX to SANCs and HEK-293T cells stably expressing HCN4. PAX had no effect on either native If or recombinant HCN4 current (data not shown).
Next, we determined whether chronic ablation of BK channels mimicked the decreased SANC excitability observed with acute inhibition of the current. SANCs were isolated from Kcnma1−/− mice, and firing frequency was compared with the baseline firing rates of WT SANCs. In contrast to heart rates measured in vivo, isolated Kcnma1−/− SANCs had significantly lower firing rates (213 ± 33 beats/min, n = 11) compared with WT cells (320 ± 10 beats/min, n = 12; Fig. 2, D–F). The slowed firing rate in Kcnma1−/− SANCs was associated with slowed diastolic depolarization duration, similar to acute BK channel inhibition with PAX (Fig. 2G). Likewise, APD50 was not significantly affected. Although MDP appeared more depolarized in Kcnma1−/− SANCs, the effect was not significant (Fig. 2I). The lower intrinsic firing rate in Kcnma1−/− SANCs suggests that not only can BK channel inhibition slow firing in WT SANCs but also that BK channel activity may be essential to generate normal SANC automaticity. Furthermore, the maintenance of normal heart rates in Kcnma1−/− mice (Fig. 1A) suggests the factors that compensate for reduced SANC automaticity are extrinsic to the SAN.
To determine if mimicking sympathetic nervous system activation could oppose the PAX-induced firing rate reduction, we applied the β-adrenergic agonist Iso (Fig. 3). First, WT SANCs were paced with 1 nM Iso. PAX was applied, which decreased firing and elongated diastolic depolarization duration (Fig. 3, A–D). Next, Iso was increased to 1 μM in the presence of PAX, resulting in increased firing back to baseline levels. Increasing Iso caused a reduction in diastolic depolarization duration in the presence of PAX (Fig. 3, D–F), consistent with its role in enhancing If and Ca2+ currents during diastolic depolarization (49). These results suggest that sympathetic activation could provide functional compensation for defects in SAN automaticity caused by loss of BK currents, contributing to the observed normal resting heart rates in intact Kcnma1−/− animals.
After establishing that elimination of BK currents through pharmacological and genetic methods resulted in decreased SANC firing, we further addressed the cellular location of the channels that mediate this effect. Since PAX is a membrane-permeable compound (35), its effects on SANC firing could be exerted through inhibition of plasma membrane BK channels or by altering of the function of intracellular BK channels (68). To determine whether BK channels expressed on the plasma membrane influenced SANC function, we tested the effect of IbTX, a highly selective peptide toxin that is membrane impermeable and blocks from the extracellular side of the BK channel (12, 25). We used 230 nM IbTX, which previously produced a 42% decrease in the rate of isolated whole hearts, since lower concentrations did not have a significant effect (31). IbTX caused a 28 ± 9% reduction in the firing rate of WT SANCs (Fig. 4A), paralleling the result seen with PAX on firing rate. IbTX also produced a concomitant elongation of diastolic depolarization duration, with no significant change in APD50 or MDP (Fig. 4, A–D). The ability of IbTX to decrease SANC firing suggests that BK channels localized on the plasma membrane mediate this effect.
BK channel expression and currents in SANCs.
BK channels are expressed throughout the cardiovascular system in smooth muscle (8, 53, 72), but there have been few investigations of BK channel expression in cardiac myocytes (37, 67, 81). Because our data demonstrated a functional effect of inhibiting BK channels in isolated SANCs, we looked for BK transcript and protein expression in the SAN. RT-PCR products corresponding to BK and HCN4, the hyperpolarization and cyclic nucleotide gated channel underlying If, were amplified from the WT SAN (Fig. 5A). No BK products were detectable from the Kcnma1−/− SAN or in the absence of reverse transcriptase (Fig. 5A).
To visualize BK expression, immunocytochemistry was performed with dissociated WT and Kcnma1−/− SANCs labeled with an α-BK primary antibody, which recognizes an epitope on the intracellular COOH terminus, and visualized via confocal fluorescence microscopy. SANCs were positively identified by colabeling cells with HCN4, which distinguishes SANCs from atrial cells in the dissociated cell preparation (9, 24). Low levels of green fluorescence (α-BK) were detected in WT SANCs (Fig. 5, B and C). BK expression was not homogenous; instead, it overlapped spatially with HCN4, suggesting a similar plasma membrane expression domain. To quantify expression levels, the average pixel intensity was calculated for BK and HCN4 fluorescence from all cells. WT SANCs had a significantly higher average BK intensity compared with Kcnma1−/− SANCs (Fig. 5C), suggesting that the fluorescence pattern reflects actual BK channel expression. In contrast, HCN4 fluorescence was not different between WT and Kcnma1−/− SANCs (n = 9 and 7, respectively, P > 0.05). WT and Kcnma1−/− atrial cells, which were negative for HCN4, did not exhibit appreciable BK fluorescence (Fig. 5C).
To better visualize BK expression, immunocytochemistry was also performed with a tyramide signal amplification kit (Fig. 5, D and E). The average pixel intensity for cells incubated with the α-BK antibody was significantly higher than cells incubated with the secondary antibody alone. These results corroborate that BK channel protein is expressed in SANCs, overlapping with HCN4.
Next, to determine whether a BK current was detectable in SANCs, whole cell voltage-clamp recordings were performed. Recordings were made with physiological solutions containing 5 mM EGTA in the pipette, a concentration necessary to maintain recording integrity. Macroscopic currents were evoked with depolarizing voltage steps from a holding potential of −75 mV, which elicited inward currents followed by outward currents (Fig. 6A) (61). After baseline currents had been recorded, 3 μM PAX was applied, and the subtracted current was analyzed (Fig. 6B). SANC identity was verified by If activation at negative membrane potentials (Fig. 6C) (19). Five of six SANCs exhibited a PAX-sensitive current. To eliminate contamination from the inward current component, current-voltage relationships were plotted from the steady-state current at the end of the voltage step and normalized to cell capacitance (40.23 ± 4.6 pF, n = 6, Fig. 6D). The PAX-sensitive current activated at depolarized potentials of >0 mV, with a peak value of 1.4 ± 0.3 pA/pF at +60 mV. The apparent right-shifted voltage dependence and small current magnitude, compared with BK currents isolated from other cells, are likely due to the intracellular Ca2+ buffering conditions (5 mM EGTA) necessary to make the SANC recordings. Unfortunately, we could not increase internal free Ca2+ by either lowering EGTA or raising the pipette free Ca2+ (see discussion), both of which resulted in recording instability and seal loss. To verify that this small PAX-sensitive current was not due to rundown, we isolated the vehicle-sensitive current as a control (Fig. 6E), which significantly differed from the PAX-sensitive current and remained flat with depolarizing voltage (Fig. 6F). The isolation of a voltage-dependent PAX-sensitive current suggests that BK currents are present in SANCs, although under these conditions, the current magnitude may be reduced by the intracellular Ca2+ buffering conditions necessary to make the recordings.
To better characterize the currents produced by BK channels expressed in SANCs, BK transcripts were amplified from single SANCs by RT-PCR. Two cDNA products were obtained, each containing no alternative exons at splice sites 1–3 (66) but differing at the last alternative exon (BKVYR and BKQEERL). The alternative exon combinations represented by BKVYR and BKQEERL are not unique to SANCs and have been amplified from other tissues (Ref. 66 and data not shown).
Expression of BKVYR and BKQEERL in HEK-293T cells produced BK currents that activated with increasing Ca2+ (Fig. 7, A and B). In physiological K+, high Ca2+ was required to activate currents in the voltage range experienced during diastolic depolarization (Fig. 7B). However, since BK currents from native SANCs could only be recorded in buffered Ca2+ conditions, we characterized BKVYR and BKQEERL currents using internal and external solutions identical to those used for SANC recordings. BKVYR and BKQEERL produced current-voltage relationships that overlaid the PAX-sensitive current (Fig. 7C), suggesting that these channel variants underlie the native BK current in SANCs.
Despite the established role for BK channels in the vasculature (6, 8, 60) and recently characterized role in cardioprotection (1, 67, 81), a functional role for BK channels has not been previously described in cardiac excitability. Here, we report that pharmacological inhibition of BK channels in vivo slows cardiac pacing (Fig. 1), abolishment of BK currents through both genetic and pharmacological methods reduces intrinsic SANC firing rate (Figs. 2–4), and BK transcript, protein, and currents are detectable in SANCs (Figs. 5 and 6). BK channels cloned from SANCs and recorded under the same conditions produced current-voltage relationships identical to PAX-sensitive currents isolated from native SANCs (Fig. 7). These data reveal that BK channels are involved in cardiac excitability and suggest that plasma membrane BK channels in SANCs are novel ionic regulators of mouse SAN automaticity. Interestingly, certain mutant alleles of slo, encoding the Drosophila BK channel, exhibit a profound decrease in heart rate and rhythmicity (34), suggesting an evolutionarily conserved role for BK channels in cardiac excitability.
Our data propose that the primary mechanism of PAX-induced bradycardia in vivo relies on BK channel targets within the SAN, supported by ECG interval analysis, a reduction in firing rate in isolated SANCs treated with PAX, and the intrinsically lower firing rate observed in Kcnma1−/− SANCs. However, the role of BK channels in cardiac excitability may not be limited to SANCs. BK subunits have been detected at very low levels in atrial and ventricular tissue (7, 62, 67, 74, 81), although the existence of BK channels on the plasma membrane of ventricular myocytes has been controversial (67, 71). Consistent with ventricular expression, in the present study, ECG interval analysis showed slowed ventricular conduction in addition to slowed sinus pacing (data not shown). However, the question of whether BK channels in ventricular myocytes alter this component of cardiac excitability remains unclear. BK currents have also been detected in Purkinje fibers, which are critical for synchronizing ventricular propagation (11).
Further supporting a basic role for BK channels in SAN excitability, in the present study, we also identified a previously unknown phenotype in Kcnma1−/− tissue, i.e., the lower intrinsic firing rate of Kcnma1−/− SANCs. Interestingly, the essentially normal heart rates in Kcnma1−/− mice, with similar ECG intervals to WT mice, suggest that compensation for the lower firing rate of Kcnma1−/− SANCs may occur through systemic means, such as increased sympathetic tone. Consistent with this idea, increasing β-adrenergic activity with Iso raised SANC firing rates back up to baseline levels after they had been decreased by PAX application (Fig. 3). Similarly, other ion channel knockouts with sinus bradycardia have heart rates similar to WT mice during activity or administration of β-adrenergic receptor agonists, potentially implicating elevated sympathetic tone for the heart rate compensation (29, 50, 63).
The reduction in excitability with BK channel inhibition suggests that BK currents play a distinctive role in regulating SANC membrane properties compared with other tissues. Loss of BK current generally increases membrane excitability in most cell types, particularly in muscle (28, 55, 60). However, several diverse examples exist where block or genetic ablation of BK channels facilitates a reduction in firing (26, 65). Furthermore, activation of BK currents under distinct membrane conditions can produce bidirectional effects on excitability, underscoring the context dependence of the role of BK channels in excitability (58). The SANC context is unique due to complex interactions between voltage-gated channels and intracellular Ca2+ cycling (49, 57). In addition, the SANC AP is significantly longer than central nervous system neurons, where the contribution of BK currents during the AP has been more extensively studied (17, 22). Recently, loss of function in other Ca2+-activated K+ channels was also shown to decrease SANC excitability, initiating discussion on possible mechanisms linking Ca2+-activated K+ currents and SANC automaticity. Apamin, a small-conductance (SK) Ca2+-activated K+ channel antagonist, and TRAM 34, which inhibits intermediate-conductance (IK) Ca2+-activated K+ channels, reduced firing rates in acutely isolated SANCs or spontaneously beating (pacemaker) human cardiomyocyte stem cells, respectively (13, 76). Inhibition of both classes of Ca2+-activated K+ channels were associated with depolarization of MDP (13, 76), and MDP is critical for If activation and diastolic depolarization (20). Loss of SK and IK currents has been proposed to increase Ca2+ influx, causing depolarization of MDP and subsequent Ca2+ channel inactivation, reducing AP firing (76).
In the present study, the decreased SANC firing with inhibition or loss of BK channels was primarily correlated with elongated diastolic depolarization. In contrast, the small differences in repolarization, measured as increased APD50 after BK inhibition, could not account for the entire decrease in firing rate. Additionally, reduced SANC firing was not associated with the significant MDP depolarization that occurs with block of other K+ currents, including SK current, IK current, and rapidly activating delayed recifier K+ current (IKr) (13, 38, 76). Since the effect of BK inhibition occurs during diastolic depolarization, it suggests that BK activation could be primarily tied to Ca2+ influx during diastolic depolarization via T- or L-type Ca2+ channels or by local Ca2+ release events from the sarcoplasmic reticulum (30, 49, 57). However, it is also possible that BK activation may occur during other phases of AP cycling, suggested by the trend toward longer repolarization with BK inhibition and in Kcnma1−/− SANCs as well as the depolarization of MDP in Kcnma1−/− SANCs. The mechanism by which BK inhibition slows diastolic depolarization is not clear, but the consistent result across two inbred mouse strains (FVB/NJ and C57BL/6J), and with both pharmacological and genetic manipulation of BK currents, may underscore the fundamental effect of BK inhibition on diastolic depolarization. Similar to the model proposed for SK activation, one could speculate that BK channel openings during diastolic depolarization contribute to the activation of excitatory diastolic currents by hyperpolarizing the membrane, which would be expected to increase If or facilitate the relief from inactivation of voltage-gated Ca2+ or Na+ channels (39, 45, 47, 50). We did find PAX-sensitive currents in SANCs with If (Fig. 6). In addition, the large inward currents observed in SANCs and the rapid upstroke of the AP suggest that many of the cells recorded in this study have prominent Na+ currents (39, 45). Reducing If or Na+ current via BK inhibition would expected to slow diastolic depolarization and AP firing (20, 39, 45). Alternately, BK channels may feedback on Ca2+ influx, and BK inhibition could increase Ca2+-dependent inactivation, leading to reduced firing (13, 76). However, the mechanism of diastolic depolarization elongation and the factors that promote BK activation during AP cycling, as well as the presence of BK currents in pacemaker SANC subtypes (5, 38, 39, 45), will require further detailed experiments specifically addressing these hypotheses.
In the present study, a more comprehensive characterization of BK current properties in SANCs was limited by the inability to resolve single channels (n = 35 attempts in our laboratory) or record from excised patches to raise the internal Ca2+ (performed in collaboration with C. Proenza, personal communication). Both the lack of an appreciable vehicle-control current (Fig. 6F) and the isolation of an IbTX-sensitive current of similar magnitude (1.3 ± 0.4 pA/pF at +60 mV, n = 3) suggest that the PAX-sensitive current is produced by BK channels. However, the PAX-sensitive currents isolated from SANCs were very small, presumably due to the minimal requirement of 5 mM EGTA to make stable whole cell recordings. Although the intracellular Ca2+ concentration is not known in our recording conditions, with heterologously expressed BKVYR and BKQEERL channels in physiological K+, high levels of intracellular Ca2+ are required for appreciable BK current activation (Fig. 7B). With 50 μM Ca2+, BK currents begin to activate at voltages comparable to MDP compared with upward of +25 mV in low Ca2+. Thus, under native conditions in SANCs, tight coupling to an intracellular Ca2+ source would be expected to shift the voltage dependence of BK activation to potentials significantly negative to those presented here from SANCs (Fig. 6F). Despite the caveats with recording, the observed BK currents in SANCs were of similar magnitude to other Ca2+-activated K+ currents recently described in cardiac SANCs and atrioventricular cells (13, 83) or IKr in SANCs (15, 16).
In a broader context, the mechanism(s) of SAN automaticity has been the subject of vigorous and continued debate (41, 44). The locus of the major driving force for pacemaking and the relative contributions of If and other plasma membrane Ca2+ currents versus rhythmic intracellular Ca2+cycling (“Ca2+ clock”) (3, 24, 42) has been a central focus of this discussion. The identification of BK currents in SANCs, the strong functional impact of BK antagonists and genetic deletion on SAN automaticity, and the potential for BK activation by Ca2+ sources at either the plasma membrane or via local intracellular Ca2+ release open up new possibilities for experimental access to address the general underlying mechanisms of cardiac pacing.
This work was supported by American Heart Association Grant 0930292N (to A. L. Meredith), National Institutes of Health (NIH) Grant R01-HL-102758 (to A. L. Meredith), The American Physiological Society's Ryuji Ueno award, sponsored by the S & R Foundation (to A. L. Meredith), Marsden Fund Royal Society of New Zealand Grant AGR302 (to J. E. Dalziel and A. L. Meredith), and NIH Grants R01-HL-079031 (to M. E. Anderson), R01-HL-096652 (to M. E. Anderson), R01-HL-070250 (to M. E. Anderson), and R01-HL-071140 (to M. E. Anderson). M. H. Lai was supported by NIH Grants T32-AA-R07592 and T32-HL-72751.
No conflicts of interest, financial or otherwise, are declared by the author(s).
Author contributions: M.H.L., J.E.D., and A.L.M. conception and design of research; M.H.L., Y.W., Z.G., and A.L.M. performed experiments; M.H.L., Y.W., Z.G., and A.L.M. analyzed data; M.H.L., Y.W., Z.G., M.E.A., J.E.D., and A.L.M. interpreted results of experiments; M.H.L. and A.L.M. prepared figures; M.H.L. and A.L.M. drafted manuscript; M.H.L. and A.L.M. edited and revised manuscript; M.H.L., Y.W., Z.G., M.E.A., J.E.D., and A.L.M. approved final version of manuscript.
The authors thank Ed Lakatta, Alexey Lyashkov, and Wendy Imlach for assistance with preliminary experiments and Jon Lederer and Long-Sheng Song for initial consultations on the project. The authors thank Ling Chen for the implantation of telemeters for ECG recordings, Josh Whitt for performing RT-PCRs, Chris Shelley for assistance with single channel recordings, and Cathy Proenza and Joshua St. Clair for assistance with excised patch-clamp recordings. The authors thank Matthew Trudeau and Cathy Proenza for comments on the manuscript and helpful discussions.
- Copyright © 2014 the American Physiological Society