Conserved Role of the Large Conductance Calcium-Activated Potassium Channel, K_Ca1.1, in Sinus Node Function and Arrhythmia Risk

Supplemental Digital Content is available in the text.

Abstract

Background:

KCNMA1 encodes the α-subunit of the large-conductance Ca²⁺-activated K⁺ channel, K_Ca1.1, and lies within a linkage interval for atrial fibrillation (AF). Insights into the cardiac functions of K_Ca1.1 are limited, and KCNMA1 has not been investigated as an AF candidate gene.

Methods:

The KCNMA1 gene was sequenced in 118 patients with familial AF. The role of K_Ca1.1 in normal cardiac structure and function was evaluated in humans, mice, zebrafish, and fly. A novel KCNMA1 variant was functionally characterized.

Results:

A complex KCNMA1 variant was identified in 1 kindred with AF. To evaluate potential disease mechanisms, we first evaluated the distribution of K_Ca1.1 in normal hearts using immunostaining and immunogold electron microscopy. K_Ca1.1 was seen throughout the atria and ventricles in humans and mice, with strong expression in the sinus node. In an ex vivo murine sinoatrial node preparation, addition of the K_Ca1.1 antagonist, paxilline, blunted the increase in beating rate induced by adrenergic receptor stimulation. Knockdown of the K_Ca1.1 ortholog, kcnma1b, in zebrafish embryos resulted in sinus bradycardia with dilatation and reduced contraction of the atrium and ventricle. Genetic inactivation of the Drosophila K_Ca1.1 ortholog, slo, systemically or in adult stages, also slowed the heartbeat and produced fibrillatory cardiac contractions. Electrophysiological characterization of slo-deficient flies revealed bursts of action potentials, reflecting increased events of fibrillatory arrhythmias. Flies with cardiac-specific overexpression of the human KCNMA1 mutant also showed increased heart period and bursts of action potentials, similar to the K_Ca1.1 loss-of-function models.

Conclusions:

Our data point to a highly conserved role of K_Ca1.1 in sinus node function in humans, mice, zebrafish, and fly and suggest that K_Ca1.1 loss of function may predispose to AF.

Atrial fibrillation (AF) is the most common cardiac arrhythmia and a major cause of morbidity and mortality. Genetic factors were first implicated in AF pathogenesis when a landmark study identified a novel locus on chromosome 10q22-q24 in 3 kindreds.¹ Two additional cardiac disorders (dilated cardiomyopathy and hypoplastic left heart syndrome) and 3 neurological disorders were subsequently mapped to the same chromosomal region (Figure I in the Data Supplement).^2–6 KCNMA1, a gene that encodes the α-subunit of the large-conductance calcium (Ca²⁺)-activated potassium (K⁺) channel, K_Ca1.1, lies within all 6 linkage intervals. KCNMA1 variants have been identified in patients with neurological defects⁴ but early studies failed to detect KCNMA1 transcripts in the heart,^7,8 and it has not been considered as a candidate gene for the mapped cardiac disorders.

K_Ca1.1 channels are comprised of 4 pore-forming α-subunits that each contain a short extracellular N-terminus, 7 transmembrane domains including pore-gate and voltage-sensing domains, and a long intracellular C-terminal tail with high affinity binding sites for Ca²⁺ and multiple ligands (Figure 1). K_Ca1.1 activation results in K⁺ efflux and membrane hyperpolarization that leads to a reduction of L-type Ca²⁺ channel activity and Ca²⁺ influx, protecting cells from Ca²⁺ overload. K_Ca1.1 channels are sensitive to Ca²⁺ and voltage and can be modulated by interactions with β and γ subunits, alternative splicing of KCNMA1 transcripts, post-translational modification, and exogenous factors.^9,10 They contribute to Ca²⁺ homeostasis in multiple cell types and have been implicated in diverse processes, including neurotransmission, hearing, circadian rhythms, bladder control, vascular smooth muscle tone, obesity, and cancer.^9,10

Figure 1.

KCNMA1 sequence variant identified in a family with tachy-brady syndrome. A, KCNMA1 sequence traces showing variation identified in the family FF proband, FF-II-3. B, This sequence change (AGCA deletion, G insertion) results in loss of 2 serines at amino acid positions 11 and 12, with the addition of a glycine. C, Family pedigree, with phenotypes denoted as: tachy-brady syndrome (solid symbols), suspected atrial fibrillation (AF; hashed symbol), no AF (open symbols); FF-II-2 had a single episode of AF associated with myopericarditis (gray symbol). The proband is indicated by an arrow. The presence (+) or absence (−) of the p.S11_S12delinsG variant are indicated. D, Schematic showing K_Ca1.1 (α-subunit, blue) with its pore and 2 RCK (regulator of potassium conductance) domains and a regulatory β-subunit (pink). The location of the p.S11_S12delinsG variant in the short extracellular amino terminus of the α-subunit is indicated (red cross).

The role of K_Ca1.1 in the heart is incompletely understood. Cardiac K_Ca1.1 channels have mainly been studied with respect to mitochondrial properties and protection against ischemia-reperfusion injury.^11–13 K_Ca1.1 expression has been seen in cardiac fibroblasts, telocytes, coronary artery smooth muscle and intracardiac neurons, suggesting pleiotropic functional effects.^14–17 In recent years, there has also been accumulating evidence that K_Ca1.1 channels are involved in heart rate regulation.^13,18–20 Given established links between sinus node function and atrial arrhythmogenesis,^21,22 these findings refocus interest on KCNMA1 as a potential candidate gene for AF. Here, we report genetic screening of KCNMA1 in a cohort of patients with familial AF. We evaluated K_Ca1.1 distribution and function in the hearts of humans, mice, zebrafish, and fly, and a novel KCNMA1 mutation was functionally characterized.

Methods

An expanded Methods section is provided in the Data Supplement. Study data and materials are available from the corresponding authors upon reasonable request. All participants provided informed written consent, and protocols were approved by the Human Research Ethics Committees of St Vincent Hospital, Prince Charles Hospital, and the University of Manchester. Protocols for animal studies were approved by Animal Ethics Committees of St Vincent Hospital/Garvan Institute and the University of Sydney.

Results

Identification of KCNMA1 Sequence Variants

The KCNMA1 gene was evaluated in 118 probands with familial AF. Six coding sequence variants were identified, including 4 synonymous variants and a 3-nucleotide insertion that did not alter the reading frame (Table VI in the Data Supplement). A complex variation comprised of a 4-nucleotide deletion and a 1-nucleotide insertion was identified in the proband of family FF, II-3 (Figure 1). This sequence change, p.S11_S12delinsG, results in loss of an MspA1I enzyme restriction site and was independently confirmed in the proband and evaluated in family members using both Sanger sequencing and restriction enzyme digestion (Figure VI in the Data Supplement). It was present in all affected family members as well as in 3 asymptomatic children and was absent in older unaffected family members and in the gnomAD population database (v3; accessed October 2020). The proband had previously undergone genetic screening of 14 AF-associated K⁺ channel genes^23,24 as well as the GJA1, GJA5, GJA7, HCN1, HCN2, HCN4, LMNA, NPPA, and SCN5A genes, with no other pathogenic variants found.

FF-II-3 was diagnosed with paroxysmal lone AF at 34 years of age. He subsequently developed features of tachy-brady syndrome, with sinus bradycardia, frequent sinus pauses (up to 3 seconds) and permanent AF (Table VII in the Data Supplement). His affected brother, FF-II-1, had sinus node dysfunction, that required pacemaker insertion and permanent AF. Their father had evidence of sinus bradycardia on a 24-hour Holter monitor. He had a history of palpitations but had no documented AF. To investigate how this KCNMA1 variant might predispose to sinus node dysfunction and AF, we next undertook a series of experiments to evaluate K_Ca1.1 channels in normal cardiac structure and function before evaluating variant effects.

Cardiac Expression of K_Ca1.1 in Humans and Mice

The subcellular localization of K_Ca1.1 in cardiac tissues was evaluated using immunostaining. In human right atrial tissue and isolated atrial cardiomyocytes from adult wild-type mice, K_Ca1.1 showed a weak striation pattern with relatively stronger expression in the intercalated discs and sarcolemma (Figure 2A). K_Ca1.1 staining overlapped with RyR2 (ryanodine receptor) and connexin-43 (Figure 2A). Similar patterns were observed in the human right ventricle (Figure II in the Data Supplement). In humans and in mice, there was robust expression of K_Ca1.1 in the sinoatrial node where its expression overlapped considerably with RyR2 and the L-type calcium channel, Ca_v1.3 (Figure 2A through 2C). Further evaluation of K_Ca1.1 in human right atrial tissue was undertaken using immunogold electron microscopy (Figure 3). Gold-labeled K_Ca1.1 epitopes were abundant in the cardiomyocyte T-tubules and sarcoplasmic reticulum and were also present in the nucleus, sarcolemma, intercalated discs, mitochondria, coronary vascular endothelium, and fibroblasts. Negative controls for anti-K_Ca1.1 APC-021 antibody are shown in Figure IV in the Data Supplement.

Figure 2.

K_Ca1.1 in the atrium. A, Immunostaining of human right atrial tissue sections shows colocalization of K_Ca1.1 with the RyR2 (ryanodine receptor) and Cx43 (connexin-43), with strong K_Ca1.1 expression in the sinus node (right); scale bar=5 μm. B, Whole mount immunostaining of the murine sinoatrial node (SAN) complex. K_Ca1.1 is expressed throughout the SAN and paranodal connective tissue (CT) and co-localizes with the L-type calcium channel, Ca_v1.3; scale bar=100 μm. C, Isolated murine atrial cardiomyocytes (left column) and pacemaker cells (center and right columns) show K_Ca1.1 colocalization with RyR2 and Ca_v1.3; scale bar=10 μm. Antibody optimization and a negative control for immunostaining are shown in Figures II and III in the Data Supplement.

Figure 3.

Immunogold electron microscopy of K_Ca1.1 in human atrial tissue. A, In atrial cardiomyocytes, colloidal gold-labeled K_Ca1.1 (black dots) was present in the nucleus (nu), T-tubules (T), and sarcoplasmic reticulum (red arrows). Double labeling of K_Ca1.1 (15 nm particles, large black dots) and L-type calcium channel, Ca_v1.2 or ryanodine receptor-2 (10 nm particles, small block dots), shows K_Ca1.1 in (B) the sarcolemma (SL) and T-tubules (T), and (C) sarcoplasmic reticulum. K_Ca1.1 was also present in (D) intercalated discs, (E) mitochondria, (F) coronary vascular endothelium, and (G) fibroblasts. Negative controls for immunolabeling are shown in Figure IV in the Data Supplement. Scale bars=(A, D, and E) 0.2 μm; (B and C) 0.1 μm; (F) 0.5 μm; (G) 1 μm.

The presence of KCNMA1 transcript in human right atrial tissue was confirmed by quantitative polymerase chain reaction with relatively higher levels of expression in the sinus node than in paranodal regions and atrial myocardium (Figure VIII in the Data Supplement). Total KCNMA1 transcript levels were significantly higher in atrial tissue from patients with a history with AF when compared with those without AF (P=0.023; Figure 4A). These differences appeared to be driven by the subset of older patients (≥70 years) who had atrial dilatation and chronic AF (Figure 4B and 4C; Figure IX in the Data Supplement). We then evaluated expression of a number of KCNMA1 isoforms (Figure 4C). Levels of the mitochondrial-specific isoform, DEC, were significantly higher in older patients with AF when compared with age-matched patients without AF (P=0.019) and younger patients with AF (P=0.014). Levels of the stretch-activated isoform, STREX, were also significantly higher in older patients with (P=0.028) or without AF (P=0.018; Figure 4C). On Western blotting, 2 K_Ca1.1 bands were detected that were ≈100 kD and 50 kD (Figure 4D), and K_Ca1.1 protein levels were significantly increased in patients with AF (P<0.0001; Figure 4E).

Figure 4.

Quantitative assessment of atrial K_Ca1.1 expression. A, Quantitative polymerase chain reaction evaluation of KCNMA1 mRNA in right atrial appendage tissue from patients in sinus rhythm with normal atrial size (red, n=11) and patients with atrial dilatation and documented atrial fibrillation (AF; blue, n=8). B, Dot blot representation of total KCNMA1 mRNA levels in individual patients (from A) showing effects of AF status and age. C, Relative mRNA levels of total KCNMA1, constitutive transcript (insertless) and the 4 isoforms, DEC, STREX, SV27, and del23, in patients with and without a history of AF, aged <70 y or ≥70 y at the time of study; data are expressed as mean mRNA levels (of multiple experiments) normalized to expression of housekeeping genes (hprt, b-actin) and an internal control. One-way ANOVA with the Tukey method correction for multiple comparisons identified significant differences between groups (P<0.0001). In addition, the unpaired Student t test was performed to evaluate the effects of age and AF status on expression levels of specific isoforms; P values for statistically significant differences are shown. D, Western blot showing total K_Ca1.1 protein in patients with and without AF (n=3 in each group); β-tubulin was used as a loading control. Negative controls for Western blotting are shown in Figure V in the Data Supplement. E, Mean data for levels of total K_Ca1.1 protein in patients with and without AF (n=3 each group, 2 replicates/sample). AU indicates arbitrary units.

Effects of Adrenergic Receptor Stimulation and Paxilline on Sinoatrial Node Function

To investigate the role of K_Ca1.1 in sinoatrial node function, we made use of an ex vivo preparation of the murine sinoatrial node (Figure 5A). Addition of the K_Ca1.1 antagonist, paxilline 10 µmol/L, alone did not significantly reduce the beating rate (control: 359±33; paxilline: 321±18, P=0.36; n=8). When the sinoatrial node was subjected to an adrenergic stimulus, isoproterenol 500 nmol/L, the Ca²⁺ transient frequency increased from 340±23 to 520±19 bpm (P=0.0003; Figure 5B). This adrenergic-induced rate increase was blunted in the presence of paxilline 10 µmol/L, with an average of 440±18 bpm (P=0.0.03; n=6; Figure 5B through 5F), suggesting K_Ca1.1-dependent effects.

Figure 5.

Effect of K_Ca1.1 blockade on ex vivo murine sinoatrial node (SAN) function. A, Images of Cal520-loaded whole SAN preparation showing the right atrium (RA), crista terminalis (CT), inferior vena cava (IVC), and superior vena cava (SVC). Intensity of the intracellular Ca²⁺ signal is shown at baseline (between beats [top], zoom [right]) and at the peak of a Ca²⁺ transient (during beat [lower], zoom [right]). B, Raw traces of Ca²⁺ transients from the Ca²⁺ signal in arbitrary fluorescent units (AU) for control, isoproterenol 500 nmol/L, and isoproterenol 500 nmol/L plus paxilline 10 µmol/L. C, Timeline showing the experimental protocol with incubation times. A 5 s imaging measurement was taken at each of 7 time points (T1–T7). D, Effects of isoproterenol 500 nmol/L (Iso) and paxilline 10 µmol/L (Pax) shown as change in beating rate Δbpm compared with control recordings (n=6; data shown as mean±SEM). E, Difference in absolute beating rate for SAN treated with isoproterenol 500 nmol/L alone or isoproterenol 500 nmol/L plus paxilline 10 µmol/L (n=6). F, Difference in Δ bpm between SAN treated with isoproterenol 500 nmol/L alone and isoproterenol 500 nmol/L plus paxilline 10 µmol/L (n=6).

K_Ca1.1 Knockdown Alters Cardiac Function in Zebrafish

The zebrafish orthologous genes, kcnma1a and kcnma1b, share ≈90% sequence identity and have ≈85% homology to human KCNMA1.²⁵ Both kcnma1 genes were expressed in zebrafish whole embryos and embryonic heart at 3 dpf (Figure 6A). Morpholinos were designed to target the translation start sites of the kcnma1a and kcnma1b genes with effective protein depletion achieved for kcnma1b (Figure 6B; Figure X in the Data Supplement). Overall size and morphology of the kcnma1b morphant was grossly normal. There were significant differences in atrial phenotypic features in Kcnma1b morpholino-injected embryos when compared with control-injected embryos with lower mean heart rate (Kcnma1b, 123±4 bpm versus controls, 140±2 bpm, P=0.0004), larger atrial diameter (Kcnma1b, 108±5 μm versus controls, 79±3 μm, P<0.0001), and reduced atrial fractional area change (Kcnma1b, 27±2% versus controls, 47±2%, P<0.0001; Figure 6C and 6D). Ventricular end-diastolic diameter (P<0.0001) and fractional shortening (P<0.0001) were also significantly lower in kcnma1b morpholino-injected embryos than in controls (Figure XI in the Data Supplement). These observations indicate that kcnma1b depletion affects heart rate, as well as atrial and ventricular chamber formation and function in zebrafish.

Figure 6.

Characterization of zebrafish kcnma1. A, Expression of kcnma1 genes in whole embryo and isolated heart of 3 dpf wild-type zebrafish, detected by RT quantitative polymerase chain reaction; β-actin was used as a control for cDNA quality. B, Western blot showing kcnma1 protein levels in control morpholino (MO)-injected and kcnma1b morpholino-injected embryos; samples were run in duplicate, β-tubulin was used as a loading control. Western blot was performed with anti-K_Ca1.1 antibody specific for the kcnma1b isoform. C, Brightfield images of hearts of 3 dpf zebrafish embryos injected with either control or kcnma1b morpholino showing relative sizes of the atria (yellow dashed line) and ventricle (green dashed line). D, Mean data for heart rate, maximal atrial diameter, and atrial fractional area change (FAC) in control morpholino-injected and kcnma1b morpholino-injected embryos at 3 dpf (n=17–20 each group).

K_Ca1.1 Is a Determinant of Cardiac Function in Drosophila

The Drosophila slowpoke (slo) gene has ≈60% sequence identity to the human KCNMA1 gene and, along with homologs of KCNA (Kv1.1, the shaker related voltage-dependent K+ channel), KCND (Kv4.3, the voltage-dependent K+ channel, sub-family D), KCNQ (Kv7.1, the slow delayed rectifying K+ channel), hERG (Kv11.1, the rapid delayed rectifying K+ channel), KCNJ (Kir, the inwardly rectifying K+ channel subfamily J), and K_ATP channels, is present in the fly heart.^26–28 We confirmed the cardiac expression of slo using nanofluidic quantitative polymerase chain reaction (Figure XII in the Data Supplement). To determine the effects of K_Ca1.1 deficiency in the Drosophila model, we reduced systemic or cardiac slo function and evaluated cardiac parameters using semi-automated optical heartbeat analysis^27,29 in young (1 week), middle-aged (3 week), and old (5 week) mutants and age-matched controls.

We first generated flies with 2 loss-of-function slo alleles: one of these, slo⁴ is a null allele produced by a gamma ray-induced chromosomal inversion,³⁰ and the other, Df(3R) BSC 397, contains a transposon-mediated deletion. These trans-heterozygote flies (slo⁴/Df(3R) BSC 397) showed increased heart period (P<0.0001) and a higher arrhythmia index (defined as irregularly irregular heart period; P=0.016)²⁷ at 3 weeks of age when compared with control flies (Figure 7A and 7B; Figure XII in the Data Supplement). To show that slo deficiency is the cause of these heart phenotypes, we crossed a wild-type genomic copy of slo into the slo4/Df(3R) BSC 397 combination. In these flies, the slow heartbeat and arrhythmic phenotype were rescued (P<0.0001 and P=0.002, respectively; Figure 7A and 7B). Interestingly, the presence of 1 or 2 extra copies of genomic slo in the wild-type fly background also markedly increased the heart period (P<0.0001; Figure XIIIA through XIIID in the Data Supplement). Taken together, these results show that heart rate regulation in flies is sensitive to K_Ca1.1 dose, with both deficiency and excess resulting in bradycardia.

Figure 7.

Cardiac function in Drosophila slo mutants at 3 wk of age. A, Heart period and (B) arrhythmia index are increased in flies with cardiac-specific RNAi knockdown of slo (Heart Gal4>Slo RNAi) and a trans-heterozygous configuration of 2 mutant slo alleles (slo⁴/Df(3R) BSC 397) compared with controls. An extra copy of the wild-type slo genomic locus (PBac slo/+; slo⁴/Df(3R) BSC 397) was able to rescue the cardiac phenotype of the trans-heterozygous mutant. C and D, Cardiac-specific slo knockdown in adult flies resulted in a dramatic increase in arrhythmia index compared with controls. E and F, Cardiac expression of the human p.S11_S12delinsG KCNMA1, but not the wild-type form, resulted in a significantly increased heart period (E) with a trend towards a higher arrhythmia index (F) in a sensitized background. For all genotypes, n=20–50 each group. Statistical comparisons between groups were made using 1-way ANOVA and Tukey test; P values for statistically significant differences are shown.

To determine whether slo acted directly in the heart, we crossed the cardiac Gal4 driver, Hand4.2-Gal4, and UAS-slo-RNAi lines to generate flies with heart-specific knockdown of slo (Heart Gal4>slo RNAi). When compared with controls, the Heart Gal4>slo RNAi flies, like systemic slo-deficient flies, had significant prolongation of systolic intervals (P=0.007) that have been shown to correspond 1:1 to the duration of the underlying action potentials in this system.²⁸ In addition, there was an increase in the diastolic intervals between contractions, resulting in an overall increase in the heart period (slower heart rate; P=0.003) and increased arrhythmias (P=0.047) at 3 weeks of age (Figure 7A and 7B; Figure XIIA through XIID in the Data Supplement). Similar changes in heart period were seen at 5 weeks (P<0.0001). For arrhythmia index, the control values increased with age as expected,²⁷ and the differences between groups did not achieve statistical significance (P=0.058; Figure XIIA through XIID in the Data Supplement).

The Heart Gal4>slo RNAi flies have slo knockdown in the heart throughout development and a significant proportion died during aging (P<0.0001; Figure XIIE in the Data Supplement). To determine if slo knockdown during development or at adult stages was important, we used a conditional heart driver, HandGS-Gal4 to restrict knockdown to adult stages.³¹ Adult-only slo knockdown in the heart also caused a slower heartbeat (P=0.03) and substantial arrhythmias (P<0.0001; Figure 7C and 7D; Figure XIV in the Data Supplement). These data suggest that adult-only slo knockdown is sufficient to cause cardiac abnormalities.

To exclude the possibility that neural slo function might be responsible for this cardiac phenotype, we crossed a pan-neural Gal4 driver line (elav-Gal4) with the UAS-slo-RNAi line. We found no significant differences in either heart period (P=0.20) or arrhythmia index (P=0.07) between mutant (Neural Gal4>slo RNAi) and control flies (Figure XIIIE through XIIIF in the Data Supplement), suggesting a cardiac autonomous effect of slo deficiency.

Electrophysiological evaluation of fly hearts was performed at 3 weeks of age. Intracellular tracings in wild-type (w1118) and in Heart Gal4/RNAi control hearts typically showed a resting membrane potential of −40 to −60 mV, with single action potential peaks of 60 mV amplitude (Figure 8; Table VIII in the Data Supplement). There were no changes in resting membrane potential, maximum action potential amplitude, or time to maximum action potential amplitude in the slo mutants when compared with w¹¹¹⁸ flies. However, both Heart Gal4>slo RNAi and slo⁴/Df(3R) BSC397 flies showed multiple peaks and a longer duration of depolarization in each action potential (P<0.0001 for both groups versus w¹¹¹⁸) suggestive of diminished repolarization reserve and increased propensity to early after depolarizations. In particular, flies with conditional adult-only slo knockdown showed frequent bursts of action potentials (Figure 8E), consistent with the increased systolic intervals (P<0001) and high incidence of fibrillatory contractions evident in M-mode tracings (Figure XIII in the Data Supplement).

Figure 8.

Electrophysiological evaluation of Drosophila slo mutants. Ten second representative traces of the electrical activity within control lines: (A, D, F, and G), slo-deficient lines: (B, C, E) and human transgene expression in slo background (H and I); slo knockdown lines showed increased peaks per event and higher event duration compared with controls (see Table VIII in the Data Supplement); y axis: voltage (mV), x axis: time (ms). Cardiac-specific human wild-type KCNMA1 expression showed a prolonged contraction phenotype, but this was less pronounced than seen with the human KCNMA1 mutant in a null background.

Functional Characterization of Human KCNMA1 Mutation

To study the p.S11_S12delinsG mutation identified in family FF, wild-type and mutant human KCNMA1 constructs were generated and injected into w¹¹¹⁸ flies as UAS transgenes, then crossed into the slo⁴ mutant background. This strategy created a sensitized background to evaluate the human KCNMA1 constructs in the absence of the endogenous slo gene. At 3 weeks of age, there were no differences in heart period (P=0.48), arrhythmia index (P=0.78), or action potential duration (P=0.76) between flies with cardiac-specific overexpression of wild-type human KCNMA1 (Heart Gal4>KCNMA1^wt) compared with slo mutants alone (Figures 7E, 7F, and 8H). In contrast, flies overexpressing the human KCNMA1 mutant (Heart Gal4>KCNMA1^mut) showed a significantly longer heart period (P=0.027) compared with flies overexpressing the human KCNMA1 wildtype in the mutant background (Heart Gal4>KCNMA1^wt). The Heart Gal4>KCNMA1^mut also showed relatively longer action potentials (P=0.03 versus w¹¹¹⁸; P=0.079 versus slo⁴/Df(3R) BSC 397) and a greater number of peaks per burst when compared with slo control groups (Figures 7E, 7F, and 8I). Collectively, these results indicate that wild-type human KCNMA1 is unable to rescue flies with slo deficiency, but mutant human KCNMA1 is functionally deleterious and exacerbates the effects of slo deficiency.

Discussion

Here we show that K_Ca1.1 is highly expressed in the sinus node in humans as well as in mice, and that K_Ca1.1 inactivation affects heart rate in mice, embryonic zebrafish, and adult flies. A KCNMA1 mutation with a loss-of-function effect was found in a family with tachy-brady syndrome and AF. Collectively, our data provide new perspectives on the role of K_Ca1.1 in sinus node pathophysiology and atrial arrhythmia susceptibility.

The sinus node is a crescent-shaped structure, comprised of specialized pacemaker cells interspersed with supportive connective tissue and neural cells, that is located at the junction of the superior vena cava and the right atrium.^21,22,32 Its major role is the cyclical generation of action potentials that initiate electrical impulse generation and excitation-contraction coupling in the heart. This pacemaker activity is dependent on the coordinated activity of multiple cardiac ion currents.³³ Although a number of potassium channel currents have been implicated, the contribution of K_Ca1.1 channels is incompletely understood.

Early work in Drosophila larvae suggested that K_Ca1.1 was involved in heart rate regulation,³⁴ but subsequent studies yielded inconclusive data. Initial observations in kcnma1 knockout (Kcnma^−/−) mice showed that resting heart rates were similar to those of age-matched wild-type mice.¹⁸ Pharmacological inhibition of K_Ca1.1 had no effects in the Kcnma^−/− animals but did reduce heart rates in wild-type mice and isolated rat hearts.¹⁸ Similarly, injection of paxilline elicited dose-dependent reductions of heart rate in wild-type rats.²⁰ Subsequent analysis of isolated sinus node cells from Kcnma^−/− mice showed that baseline firing rates were reduced when compared with wild-type cells, and it was proposed that the absence of bradycardia in intact animals was due to compensatory activation of the sympathetic nervous system.¹⁹ In contrast, mice with cardiomyocyte-specific K_Ca1.1 deletion did show modest reductions in heart rate as well as reductions of ventricular ejection fraction and mean arterial pressure.¹³ Our murine data extend these findings and suggest that K_Ca1.1 is required for normal chronotropic responses to adrenergic stimulation. Our data also implicate K_Ca1.1 as a determinant of heart rate in zebrafish and fly.

How do these findings fit into current concepts of sinus node pacemaker automaticity? In the Ca²⁺ clock model, impulse generation is dependent on cyclical RyR2-mediated Ca²⁺ release from sarcoplasmic reticulum stores coupled to Ca²⁺ inflow though voltage-gated Ca²⁺channels.³² The latter include L-type Ca_v1.3 channels that are activated at negative voltages at the beginning of diastolic depolarization. The localization of K_Ca1.1 in the vicinity of RyR2 and Ca_v1.3 in pacemaker cells suggests that activation of K_Ca1.1 might be a key component of this cycle, linking Ca²⁺ release, membrane hyperpolarization and Ca_v1.3 activation. Membrane hyperpolarization is also required for activation of the HCN4 (hyperpolarization-activated cyclic nucleotide-gated 4) channels that are involved in regulation of sinus node action potentials at rest and following autonomic stimulation.³⁵ K_Ca1.1 deficiency could result in bradycardia by (1) prolonged diastolic depolarization due to limitation of intracellular Ca²⁺ repletion or (2) reduced HCN4 activation. It is notable that bradycardia is seen in humans and animal models that have excessive RyR2 activity, reduced Ca_v1.3 activity, or HCN4 mutation.^35–39 Previous studies have shown that K_Ca1.1 is highly expressed in fibroblasts, which are particularly abundant in the sinus node.⁴⁰ Interactions between fibroblasts and myocytes are an important component of mechano-electric feedback in the atrium and can influence the spontaneous depolarization rates of pacemaker cells. Hence, K_Ca1.1 deficiency could also impact on fibroblast-myocyte coupling with effects on basal heart rate and on chronotropic responses to atrial stretch.

Sinus node dysfunction (also known as sick sinus syndrome) is a common condition that has an estimated prevalence of 1 in 1000 person-years in adults aged >45 years.²² By 65 years of age, 1 in every 600 people will be affected.²² This disorder can be manifest in several ways, including sinus bradycardia, sinus pauses, sinus exit block, sinus arrest, or impaired heart rate responses to physiological stressors.²² It is frequently associated with intraatrial conduction delay and atrial tachyarrhythmias, particularly AF. The combination of bradycardia and AF (tachy-brady syndrome) may be due to ectopic impulse generation from other sites in the atrium, or reflect a diffuse process that extends beyond the sinus node to involve atrial myocardium.^21,22,41

Our data are in keeping with the latter scenario, as we found that K_Ca1.1 has an extensive subcellular distribution in human and murine atrial cardiomyocytes, fibroblasts, and coronary vessels. K_Ca1.1 knockdown resulted in cardiac chamber dilatation and impaired contractile function in zebrafish embryos, with prolonged action potential duration and increased action potential spikes seen in flies. We previously reported that K_Ca1.1 is a downstream effector of interactions between the Rho-GTPase Cdc42 and the cardiac transcription factor Nkx2.5. Compound heterozygous Cdc42-Nkx2.5 mutant flies and mice showed reduced K_Ca1.1 expression and changes in myofibrillar architecture, cardiac output, and conduction.⁴² Taken together with work by others, these findings point to numerous ways in which K_Ca1.1 deficiency could contribute to an atrial substrate for arrhythmogenesis, including atrial electrical and structural remodeling, myocardial ischemia, oxidative stress, and altered autonomic tone.

For the human KCNMA1 mutation identified in Family FF, the segregation analysis was consistent with disease association, with the lack of overt phenotypic features in young variant carriers likely due to age-related penetrance. The p.S11_S12delinsG variation is located in the extracellular N-terminus of K_Ca1.1 that is required for interactions with the channel’s β-subunits,⁴³ and the S12 residue has been demonstrated to be an in vivo K_Ca1.1 phosphorylation site.⁴⁴ Hence, mutation in this region could impair K_Ca1.1 channel activation due to changes in β-subunit modulation or post-translational modification. We posit that K_Ca1.1 loss of function is an important risk factor for AF in family FF. Although no variants in known AF disease genes were identified, it remains possible that other deleterious genetic variants may be contributing to disease in this family. No spontaneous episodes of AF were observed in our zebrafish model, but this may be due to the young age at which the fish were studied, small heart size, baseline fast heart rate, or lack of an appropriate trigger. Young wildtype and even KCNQ mutant flies also do not exhibit significant arrhythmia. Overt arrhythmia develops with age in flies, is correlated with age-related declines in expression of multiple potassium channel genes, and can be rescued by KCNQ overexpression.^27,45 K_Ca1.1-deficient flies did not show an increased incidence of arrhythmia until 3 weeks of age (middle age for flies), which is similar to findings in Drosophila long QT models and suggestive of an age-dependent deficit in repolarization capacity.

Increased levels of K_Ca1.1 transcript in older subjects with chronic AF, when compared with young subjects with AF and age-matched individuals without AF, was an unexpected finding and most likely represents a compensatory response to mitigate against intracellular Ca²⁺ overload. This may be a contributing factor to the substantial reductions in atrial I_CaL that are typically seen with normal ageing and in patients with persistent AF.^22,46 While initially adaptive, increased expression of human K_Ca1.1 has been shown to shorten the action potential duration in HL-1 atrial cardiomyocytes,⁴⁷ and this could increase the propensity for initiation and maintenance of AF via re-entry mechanisms. The phosphodiesterase III inhibitor, milrinone, is a heart failure therapy that has recently been shown to cause K_Ca1.1-dependent dilatation of the pulmonary veins.⁴⁸ AF is a recognized complication of milrinone therapy⁴⁹ and may be attributable, at least in part, to K_Ca1.1 activation. In flies, we found that extra copies of the slo locus had similar effects to K_Ca1.1 knockdown with increased heart period and arrhythmia index. Taken together, these findings indicate that a tightly regulated range of K_Ca1.1 levels is required to maintain normal cardiac function and rhythmicity (ie, a Goldilocks effect). This has direct implications for administration of K_Ca1.1 activating drugs, which have been proposed as potential therapies for long QT syndrome and ischemia-reperfusion injury.^47,50 Pharmacological manipulation of K_Ca1.1 activity warrants investigation as a new treatment option for cardiac arrhythmias and myopathies, but as for many other cardiac ion channels, there may be a narrow beneficial therapeutic range.

Taken together, our data show that K_Ca1.1 channels have a highly conserved role in sinus node function and arrhythmia risk. Our findings provide support for KCNMA1 as a candidate disease gene for AF and other human cardiac disorders that map to chromosome 10q22-q24. Further studies to explore the potentially diverse roles of K_Ca1.1 in cardiac function are likely to be fruitful.

Acknowledgments

We thank Jamie Vandenberg for helpful discussions, and the Victor Chang Cardiac Research Institute Innovation Centre, funded by the NSW Government.

Sources of Funding

The authors were supported by the National Health and Medical Research Council of Australia (1019693, 1025008, 1074386, 459419, and 573732), National Heart Foundation of Australia (PB06S2916), Estate of the Late RT Hall, St Vincent Clinic Foundation, NSW Health Early Career Fellowship, Prince Charles Hospital Foundation (EN2018-01), British Heart Foundation, Fondation Leducq (TNE FANTASY 19CV03), Hunter Medical Research Institute, American Heart Association (14GRNT20490239), National Institutes of Health (HL054732, HL132241, and HL098053).

Disclosures

None.

Supplemental Materials

Online Methods

Online Tables I–VIII

Online Figures I–XIV

References ^51–56