Cytoskeletal disarray increases arrhythmogenic vulnerability during sympathetic stimulation in a model of hypertrophic cardiomyopathy

Familial hypertrophic cardiomyopathy (FHC) patients are advised to avoid strenuous exercise due to increased risk of arrhythmias. Mice expressing the human FHC-causing mutation R403Q in the myosin heavy chain gene (MYH6) recapitulate the human phenotype, including cytoskeletal disarray and increased arrhythmia susceptibility. Following in vivo administration of isoproterenol, mutant mice exhibited tachyarrhythmias, poor recovery and fatigue. Arrhythmias were attenuated with the β-blocker atenolol and protein kinase A inhibitor PKI. Mutant cardiac myocytes had significantly prolonged action potentials and triggered automaticity due to reduced repolarization reserve and connexin 43 expression. Isoproterenol shortened cycle length, and escalated electrical instability. Surprisingly isoproterenol did not increase CaV1.2 current. We found alterations in CaV1.2-β1 adrenergic receptor colocalization assessed using super-resolution nanoscopy, and increased CaV1.2 phosphorylation in mutant hearts. Our results reveal for the first time that altered ion channel expression, co-localization and β-adrenergic receptor signaling associated with myocyte disarray contribute to electrical instability in the R403Q mutant heart.

Arrhythmias and premature sudden death are tragic sequelae in patients with inherited heart disease that can occur with increased sympathetic activity [1][2][3] . Familial hypertrophic cardiomyopathy (FHC) is a primary disorder of the myocardium characterized by cardiac hypertrophy in the absence of other loading conditions. It is an autosomal dominant condition caused by defects in many sarcomere protein encoding genes. The majority of disease-causing variants are located in beta myosin heavy chain (MYH7) and myosin binding protein C (MYBPC3) 4 . It is well recognized that the progression of FHC involves altered energy metabolism, myocyte remodeling, disorganization of cytoskeletal proteins and fibrosis, and results in major adverse cardiac events such as heart failure and sudden cardiac death [5][6][7] . Intense exercise is thought to promote ventricular tachyarrhythmias, therefore FHC patients are advised to avoid intense physical activity and competitive sport 8 . Although fibrosis and hypertrophy are recognized substrates for arrhythmias, alterations in the electrical properties of the cardiac myocyte and its response to adrenergic stimulation can also contribute to the genesis of ventricular arrhythmias and sudden cardiac death 1,9 . In humans, the R403Q variant in MYH7 causes a severe form of FHC characterized by early-onset and progressive myocardial dysfunction with a high incidence of sudden cardiac death 10 . Mice expressing the R403Q mutation in MYH6, encode the predominant myosin isoform in the adult mouse heart that is highly homologous in sequence with MYH7, develop hallmark features of hypertrophic cardiomyopathy from 30 weeks of age 5 . Homozygous mice are viable at birth and look anatomically normal, but die by day 7 with severe dilated cardiomyopathy. Mice heterozygous for the R403Q MYH6 mutation (αMHC 403/+ ) have a normal lifespan, and www.nature.com/scientificreports/ preserved cardiac function. Young heterozygous mice demonstrate myofibril disorientation, myocyte disarray 5 , alterations in L-type calcium channel kinetics and altered mitochondrial metabolic activity 7 that precede the development of myocyte hypertrophy, myocyte injury and fibrosis. Regardless of the presence of hypertrophy, hearts exhibit impaired diastolic function, myocyte cytoskeletal disarray and altered energetics 11 . Similar to FHC patients, αMHC 403/+ hypertrophic mice can experience serious arrhythmias with vigorous exercise 5,12 . The pathophysiology contributing to the development of arrhythmias is unclear. A study using highresolution optical mapping of αMHC 403/+ hypertrophic hearts during ventricular pacing, found no direct correlation between the amount or the pattern of fibrosis and inducibility of arrhythmias 13 . Arrhythmia formation at the cellular level centers on two key concepts: altered calcium homeostasis and reduced repolarization reserve 14,15 . In addition to an increase in myofilament calcium sensitivity 16 , αMHC 403/+ mice demonstrate a significant reduction in sarcoplasmic reticulum calcium content 7 due to decreased expression of calsequestrin, triadin, junctin and ryanodine receptor 2 (RyR2), the proteins forming the cardiac calcium release unit 11 . Interestingly, although no differences in diastolic or systolic calcium concentrations were measured in cardiac myocytes isolated from pre-cardiomyopathic αMHC 403/+ mice, calcium channel blockers such as diltiazem prevented development of the hypertrophy 11,17 . In addition an early remodeling of repolarizing K + currents has been reported prior to the development of hypertrophy in the αMHC 403/+ mouse that contributes to alterations in repolarization 18 . However the effect of sympathetic nervous system stimulation on arrhythmia formation is unclear.
The objective of this study was to investigate the mechanisms for induction of arrhythmias in the αMHC 403/+ murine model of FHC with developed hypertrophy in the absence and presence of β-adrenergic receptor stimulation. We performed a suite of in vivo and in vitro studies and found that contrary to effects observed in wt hearts, action potentials in αMHC 403/+ myocytes were prolonged and β-adrenergic receptor stimulation shortened the action potential while increasing the frequency of delayed afterdepolarizations and ventricular tachyarrhythmias. This was recapitulated in αMHC 403/+ mice following in vivo challenge with isoproterenol. Consistent with the cytoskeletal disarray, Ca V 1.2-β1 adrenergic receptor colocalization was altered assessed by super-resolution nanoscopy. Ca V 1.2 was unresponsive to isoproterenol due to increased phosphorylation in mutant hearts and connexin 43 expression was significantly decreased. We conclude that altered ion channel expression, location and function contribute to altered β-adrenergic receptor signaling and increased automaticity. Our data demonstrate for the first time an association between cytoskeletal disarray and arrhythmia formation in the R403Q mutant heart.

Results
Sympathetic stimulation induces sustained arrhythmias in αMHC 403/+ mice with a hypertrophic phenotype. First we examined the effect of the β-adrenergic receptor agonist isoproterenol (ISO) on arrhythmia inducibility in vivo. ECGs were recorded in 35-45 week old αMHC 403/+ and wt mice before and after intraperitoneal injection of 20 mg/kg ISO. Isoproterenol has a half-life of 2.5 to 5 min 19 . To monitor arrhythmia inducibility for an extended time period, a second dose of 20 mg/kg ISO was applied 10 min after the first injection ( Fig. 1A-D). The dose is well tolerated and does not induce myocardial damage 20 . Lead II ECGs were recorded using subdermal needle electrodes with no programmed electrical stimulation. Prior to administration of ISO, αMHC 403/+ mice exhibited significantly longer QT, QT c intervals and T peak T end duration compared to wt mice (Table 1). In the absence of ISO, single premature ventricular contractions (PVC) were occasionally recorded in αMHC 403/+ mice. None of the mice exhibited atrial arrhythmias at baseline or following administration of isoproterenol.
Following ISO treatment, both wt and αMHC 403/+ mice exhibited a significantly higher heart rate shown as an increase in beats per minute (Table 1). Although QT and QT c intervals, and T peak T end duration were slightly shortened in αMHC 403/+ mice following ISO, all remained significantly longer compared to wt mice, while R and T amplitudes became significantly reduced (Table 1, Fig. 1G,H). Arrhythmic events were calculated as number of irregular beats each second of recording and included single PVC's and ventricular tachycardia (Fig. 1D). One of the eight αMHC 403/+ mice suffered cardiac arrest shortly after the ISO injection and significant arrhythmias were recorded in six of the seven surviving αMHC 403/+ mice ( Table 1). One of the αMHC 403/+ mice showed no spontaneous or inducible ventricular arrhythmias. Our data are consistent with reports of arrhythmias and sudden death in αMHC 403/+ mice following vigorous swimming 5 .
A final ECG was taken two hours following administration of ISO. The relative occurrence of arrhythmic events in αMHC 403/+ mice was tenfold higher 2 h post-ISO than wt mice challenged with ISO with a higher heart rate ( Fig. 1E,F,N vs. M and Table 1). In addition we found that the 6 αMHC 403/+ mice that exhibited prolonged exacerbated arrhythmic events demonstrated poor recovery and lethargy 2 h following ISO injection observed as difficulty mobilizing and decreased activity moving around the cage assessed using BAR (bright, alert, responsive) animal monitoring criteria and scoring (see original Monitoring Sheets in Supplementary Information File). This was in contrast to wt mice that remained active and spontaneously groomed and fed following ISO injection.
Consistent with the development of a hypertrophic phenotype, αMHC 403/+ mice displayed a significant increase in left ventricular posterior wall thickness and significant decrease in left ventricular internal diameter compared to wt mice (Table 2 and Fig. S1A,B). Stroke volume and diastolic parameters were reduced in mutant mice consistent with previous reports 21 . In wt mice, ISO treatment induced a significantly greater decrease in left ventricular internal diameter at end systole (LVIDs) and end systolic volume (ESV) compared with mutant hearts. In αMHC 403/+ mice treated with ISO the change in ejection fraction (14% increase) was less marked than wt (22%, Table 2). These findings confirm that the αMHC 403/+ mouse heart has difficulty complying with the increased contractile demands imposed during sympathetic nervous system stimulation.
To further investigate the effect of ISO, mice were pre-treated for 10 min with atenolol, a selective β1 AR antagonist. Low dose (1 mg/kg, i.p.) atenolol significantly reduced the heart rate in both wt and αMHC 403/+ mice  www.nature.com/scientificreports/ ( Fig. 1I,J,O) but the effect was more pronounced in the mutant mice ( Table 1). As expected, selective β-blocker pretreatment also relaxed the left ventricles ( Table 2, Fig. S1F,H vs. E,G). ISO increased the heart rate in the presence of atenolol in both wt and αMHC 403/+ mice ( Fig. 1K and Table 1), but importantly heart rate did not remain elevated in the mutant mice treated with atenolol after 2 h' recovery ( Table 1, Fig. 1L,O). The reduction in R amplitude and increase in the relative occurrence of arrhythmic events was less pronounced in the presence of the β 1 -AR blocker in αMHC 403/+ mice ( Table 1, Fig. 1O). The recovery from in vivo ISO treatment when atenolol was present was similar to wt mice (Table 1, Figs. 1L,O, S1I,J). αMHC 403/+ mice were bright, alert, mobile and active similar to wt mice when pre treated with atenolol followed by ISO. Our data confirm that in addition to facilitating ventricular filling, cardiac selective β1AR blockers can reduce arrhythmogenic activity in αMHC 403/+ hypertrophic hearts and decrease lethargy post sympathetic nervous system stimulation.
αMHC 403/+ ventricular myocytes exhibit prolonged action potential duration that shortens in the presence of ISO. Next we assessed the AP characteristics of cardiac myocytes isolated from adult hypertrophic αMHC 403/+ mice in the absence and presence of acute exposure to ISO (100 nM). Under control conditions, at 1 Hz the resting membrane potential of αMHC 403/+ myocytes was slightly depolarized and AP duration significantly prolonged (APD90: 165.1 ± 12.7 vs. 47.2 ± 4.1; n = 62 and n = 63, respectively) ( Fig. 2A-C and Table 3). No significant difference in the amplitude of the AP was recorded. Consistent with this the expression level of the cardiac sodium channel protein Na V 1.5 was unchanged in αMHC 403/+ hearts (Fig. S2). wt cardiac myocytes showed no triggered spontaneous automaticity, while 88.2% of αMHC 403/+ cardiac myocytes developed triggered activity, and delayed afterdepolarizations (DADs) at low (1 Hz) stimulation frequency (Table 3). Exposure to ISO caused a prolongation in action potential duration in wt cardiac myocytes (APD50 and APD90, Fig. 2A, Table 3) but shortened αMHC 403/+ myocyte APD90 (Fig. 2B,C and Table 3). ISO also significantly increased the probability of DADs in αMHC 403/+ myocytes ( Fig. 2E,G, Table 3 and Fig. S3).
It is well known that APD depends on heart rate or stimulation frequency 15 . Pacing αMHC 403/+ ventricular myocytes at their baseline heartbeat frequency (9 Hz or 540 beats/min) revealed irregular AP patterns. At the high frequency the cycle lengths of the impulses were shorter than the triggered APD, resulting in ineffective repolarization, early afterdepolarizations (EADs) and consequently depolarized resting membrane potential ( Fig. 2E and inset vs. wt control on Fig. 2D). ISO had no effect on the pacing pattern at low stimulation frequency (Fig. S3B), but high frequency stimulus aggravated the effect of ISO in αMHC 403/+ cardiac myocytes (Fig. 2G). Furthermore ISO induced DADs in the αMHC 403/+ cardiac myocytes, at all stimulation protocols (at 9 Hz in Fig. 2G vs. F and at 3 Hz showed on right panel in Fig. S3A vs. B). A long APD also results in a long refractory period, leading to impaired impulse conduction and reentry in the heart 15 . Overall, these data indicate that ISO increased the excitability and induction of arrhythmias in αMHC 403/+ myocytes.
To confirm that ISO was increasing susceptibility to arrhythmias in αMHC 403/+ ventricular myocytes via protein kinase A, we pre-treated cells for 30 min with a cell-permeable protein kinase A inhibitor PKI (myristoylated PKI 14-22 amide, Tocris). 3 μM PKI attenuated the arrhythmogenic effect of ISO, and DAD frequency in αMHC 403/+ myocytes in the presence of 100 nM ISO (0.216 ± 0.116 + PKI vs. 0.525 ± 0.0192 p < 0.05 Table 3). Isoproterenol, in the presence of the protein kinase A inhibitor only slightly shortened the APD (Table 3). Our results indicate that PKA phosphorylation activated by the β-adrenergic signaling cascade is responsible for increased arrhythmic activity in mutant ventricular myocytes.
Cardiac myocytes isolated from αMHC 403/+ mice exhibit distinct electrophysiological features in the absence and presence of ISO. Arrhythmia formation in mouse cardiac myocytes can occur as a result of alterations in potassium or calcium currents 22 . Significant decreases in I to , I Kslow and I sust components of repolarizing potassium currents have been previously reported 18 in pre-hypertrophic αMHC 403/+ cardiac myocytes. In this study we recorded decreases in I peak, I to , I Kslow and I sust components of the repolarizing potassium currents in left ventricular myocytes isolated from hypertrophic hearts (Fig. 3A vs. B,F vs. E). Corresponding with the electrophysiological changes, significantly decreased expression levels of K V 4.2, K ir 2.1 and K ir 6.2 channel proteins were detected in αMHC 403/+ versus wt hearts. Localization of the ATP-sensitive K + channel K ir 6.2 was also altered in αMHC 403/+ myocytes (Fig. 3G). The expression and localization of other potassium channel proteins: K V 1.4, K V 1.5, K V 2.1, K V 11.1 and K V 7.1, or their auxiliary subunit proteins TASK1, and KCNE1/MinK were not significantly altered in αMHC 403/+ hearts (Fig. S2).
Interestingly αMHC 403/+ cardiac myocytes exhibited a significant increase in sensitivity of I to and I Kslow currents to 100 nM ISO (Fig. 3D,F) versus wt myocytes (Fig. 3C,E). Reduced repolarization reserve and increased sensitivity of I to and I Kslow to ISO may contribute to the APD shortening observed in αMHC 403/+ cardiac myocytes (Fig. 2B,C). Nevertheless this would only explain in part the increased arrhythmogenicity during sympathetic activation.
It is well recognised that Synapse Associated Protein 97 (SAP97) co-localises with K ir 2 and K V channel proteins, anchoring them to the plasma membrane, and aiding correct folding and function 23 . In addition to AKAP proteins, SAP97 participates in β1-adrenergic receptor localization and PKA phosphorylation 24 . We measured a significant decrease in SAP97 protein expression in cardiac tissue and myocytes isolated from αMHC 403/+ mice versus wt mice (Fig. 4). Furthermore confocal imaging revealed differences in SAP97 protein localization with a relatively preserved surface membrane presentation of the protein, and reduced intracellular SAP97 content. Our data (Figs. 3G, 4) are consistent with changes in cell size and myofilament organization that are characteristic when structural changes such as hypertrophy are present. We measured a profound decrease (~ 40%) in connexin 43 protein expression in αMHC 403/+ hearts (Fig. 4). This likely contributes to altered conductivity causing impulse propagation heterogeneity which in combination with decreased repolarization and impaired calcium handling provides a substrate for reentry arrhythmias 15,22 . Altered calcium handling in the αMHC 403/+ model has been reported as a consequence of reduced SR calcium content and calcium accumulation by mutant myofilaments 11,16,21 . However systolic and diastolic calcium concentrations measured in wt and hypertrophic αMHC 403/+ hearts are similar 17 . Consistent with our previous findings 7 , in the absence of ISO, quiescent adult αMHC 403/+ cardiac myocytes demonstrated a small, but significant decrease in the kinetics of calcium current inactivation (Fig. 5B vs. 5A and E) of the L-type calcium channel (I Ca ), with no change in peak amplitude (Fig. 5A,B), current density (Fig. 5D), activation or deactivation assessed as the integral of the current (Fig. S4A-C). As expected cell capacitance was increased in αMHC 403/+ myocytes consistent with a hypertrophic Table 3. Action potential parameters of wt and αMHC 403/+ ventricular myocytes and frequency of delayed afterdepolarizations. Action potential parameters recorded on wt and αMHC 403/+ ventricular myocytes in the presence of 100 nM isoproterenol, in the presence or absence of 3 μM myristoylated protein kinase A inhibitor-(14-22)-amide (PKI), or 10 μM forskolin. APD50, APD90 are action potential durations at 50% or 90% repolarization, respectively; DAD frequency: delayed afterdepolarizations recorded at 1 Hz pacing frequency in current clamp mode. Values are means ± SEM, (n) number of cells. *Data were analyzed by 2 Way ANOVA followed by a Tukey's test for multiple comparisons p < 0.05 versus wt ctrl, $p < 0.05 versus100 nM ISO αMHC 403/+ a p < 0.05 versus wt 100 nM ISO b p < 0.05 versus ctrl αMHC 403/+ . www.nature.com/scientificreports/ phenotype (Fig. 5C). Immunoblot studies indicated that there was no significant difference in Ca V 1.2 protein expression in heart homogenates from αMHC 403/+ versus wt mice (Fig. 5F). However, most surprisingly acute ISO exposure (100 nM) had no effect on current density (Fig. 5D), activation or deactivation parameters of I Ca in αMHC 403/+ (Fig. S4A-C), although it substantially increased I Ca in wt cardiac myocytes (Figs. 5D,E, S4A-C). Under paced conditions (1 Hz), no significant difference in calcium transients was recorded in αMHC 403/+ versus wt myocytes (Fig. 5H-L), in agreement with previous reports 17 .  D). Whole-cell voltage-gated outward K + (K V ) currents were evoked in response to 500 ms depolarizing voltage steps to test potentials between − 60 and + 40 mV, in 10 mV increments, from a holding potential of − 70 mV (− 70, 0 mV and positive test potentials shown). Inward rectifying K + currents (I K1 ) evoked in response to hyperpolarization to − 120 mV. (E-F) Scatter plot with bar graphs show I K density (pA/pF) values for different K V current components as means ± SEM. wt n = 8-15, N = 5 αMHC 403/+ n = 10-16 N = 5*p < 0.05 control versus ISO, † p < 0.05 wt versus αMHC 403/+ (G) Representative immunofluorescence and corresponding Western blot images for K V 4.2, K ir 2.1, K ir 6.2 potassium channels from wt and αMHC 403/+ hearts as indicated. Relative optical density (Rel OD) values were calculated using VDAC (voltage dependent anion channel) as loading control. For further details please see section "Materials and methods". www.nature.com/scientificreports/ But in the presence of 100 nM ISO, the peak amplitude of the calcium transient was increased significantly in wt myocytes only (Fig. 5H, J). We examined the phosphorylation levels of immunoprecipitated Ca V 1.2 protein and found increased basal phosphorylation in αMHC 403/+ hearts. Exposure of the immunoprecipitated Ca V 1.2 to PKA increased the phosphorylation level of the Ca V 1.2 channel in wt but not αMHC 403/+ hearts (Fig. 5G), suggesting that Ca V 1.2 protein was already significantly phosphorylated in αMHC 403/+ hearts under basal conditions. This was consistent with the lack of response of I Ca to ISO in αMHC403 myocytes.

αMHC 403/+ myocytes exhibit altered Ca V 1.2 and β1AR clustering and co-localization
In vitro electrophysiological and ex vivo biochemical studies demonstrated no difference in Ca V 1.2 (Fig. 5F) and β1-adrenergic receptor expression (Fig. S2), or basal calcium current in αMHC 403/+ versus wt hearts under control conditions (Fig. 5A,B,D). Recent studies have demonstrated super-clustering of Ca V 1.2 promoted by β1-adrenergic receptor stimulation in mouse cardiac myocytes 25 . To investigate a potential role for altered channel clustering and Ca V 1.2-β1-AR colocalization, we performed super resolution microscopy experiments. Pre-treatment with ISO (100 nM isoproterenol, 10 min) induced a significant increase in the formation of superclusters of Ca V 1.2 in wt but not in αMHC 403/+ myocytes (Fig. 6)). While β1-AR cluster area was not significantly different in αMHC 403/+ cardiac myocytes compared to wt myocytes under control conditions (Fig. 7A,C,E), ISO treatment significantly altered β1AR cluster area size in αMHC 403/+ , but not in wt myocytes (Fig. 7B,D,E).
Colocalization analysis of αMHC 403/+ and wt cardiac myocytes labelled with Ca V 1.2 and β1AR antibodies revealed a relatively small number of Ca V 1.2 colocalized with β1ARs in wt cardiac myocytes and a significantly higher proportion of the Ca V 1.2 colocalized with β1ARs in αMHC 403/+ myocytes under control conditions ( Fig. 7A-D,F). ISO treatment resulted in more Ca V 1.2 protein colocalizing with β1ARs in wt cardiac myocytes, but under the same conditions, comparatively less Ca V 1.2 were found to be localized with β1ARs in αMHC 403/+ myocytes ( Fig. 7A-D,F). Colocalization of the Ca V 1.2 with the β1AR may explain the higher phosphorylation level of the channel protein in αMHC 403/+ cardiac myocytes (Fig. 5G). There was no change in β1AR co-localizing with Ca V 1.2 ( Fig. 7A-D,G).
Previous studies have demonstrated a role for caveolin-3 in forming macromolecular complexes with β-adrenergic receptors and in β-adrenergic signaling 26,27 . Although αMHC 403/+ myocytes displayed a reduced level of caveolin-3 expression (that was not statistically significant) versus wt myocytes (Fig. 4), no alterations in cluster area or colocalization with Ca V 1.2 protein were observed (Fig. S5).
Our super resolution study indicated that Ca V 1.2 protein and β-adrenergic receptor localization are altered in αMHC 403/+ myocytes. To further confirm the altered Ca V 1.2-β1AR co-localization, we examined the effect of forskolin on I Ca . In contrast to the lack of response to ISO, the addition of 10 µM forskolin significantly increased the amplitude, activation and inactivation of I Ca in αMHC 403/+ myocytes similar to that of wt myocytes (Figs. 5M-P, S4D-F). These data indicate that altered ion channel and β1AR localization are responsible for altered calcium handling in αMHC 403/+ myocytes.

Ca V 1.2 is near maximally phosphorylated in αMHC 403/+ mice.
To further investigate the mechanisms for the poor recovery of αMHC 403/+ mice following ISO treatment, hearts were collected following in vivo isoproterenol challenge, and phosphorylation of Ca V 1.2 assessed. Untreated αMHC 403/+ mouse hearts demonstrated phosphorylation of both Ca V 1.2 and total protein compared to wt, that remained unchanged following ISO treatment (Fig. S6A-C). These data suggest that Ca V 1.2 is phosphorylated under control conditions in αMHC 403/+ mice. Creatine kinase (CK) activity was also significantly increased in hearts from ISO treated αMHC 403/+ mice versus wt mice suggesting ongoing myocardial injury (Fig. S6D).

Discussion
Hypertrophic cardiomyopathy is characterized by disorganization of cytoskeletal proteins and myofibrils, myocyte remodeling, fibrosis and altered energy metabolism 10 . We and others have demonstrated that αMHC 403/+ mice develop clinical features similar to those found in human disease at both cellular and whole heart level. 7,21 .
Alterations in the electrical properties of the cardiac myocyte that occur in response to stressors such as increased adrenergic stimulation can contribute to the genesis of ventricular arrhythmias and lead to sudden cardiac death 9 . The objective of this study was to investigate the mechanisms for increased arrhythmogenic activity resulting from the human FHC disease causing mutation during sympathetic stimulation.
Our data indicate that AP characteristics in αMHC 403/+ myocytes were significantly prolonged and contrary to responses in wt cells, the AP shortened under conditions of increased adrenergic stimulation. At normal murine www.nature.com/scientificreports/ heartbeat-frequency (9 Hz) in the presence of ISO, membrane repolarization was incomplete in αMHC 403/+ myocytes, the resting membrane potential became more depolarized, resulting in more early and delayed afterdepolarizations. In addition the AP refractory period was prolonged, which is a recognized substrate for impaired impulse conduction and reentry in the heart 15 . The most apparent difference between human and murine AP kinetics is in the plateau phase. In mouse heart depolarizing I Ca is less pronounced, while repolarizing I to and I Kur are more prominent, and as a result, murine APs demonstrate a more rapid repolarization. Nevertheless similarities in structure, excitation-contraction coupling, recovery and propagation of excitation can be investigated at the molecular, cellular, tissue, organ, and whole-animal level in the mouse 22,28 . To clarify any differences in APs between mouse and man, we performed additional experiments and assessed the AP characteristics of hiPSC-CMs (see SI methods) from a hypertrophic cardiomyopathy patient carrying R403Q MYH7 mutation (MYH7 403 /+ , see family pedigree, Fig. S7C). Action potential measurements were performed with the kinetic imaging cytometry (KIC) platform. Similar to the mouse mutant myocytes, the AP duration of MYH7 403 /+ was significantly prolonged (Fig. S7B,E,F) versus the isogenic CRISPR corrected control, MYH7 403 /− (Fig. S7A,D,F). Also similar to the αMHC 403/+ myocytes, application of 1 µM ISO significantly shortened MYH7 403 /+ AP duration while ISO slightly prolonged the isogenic CRISPR corrected control AP. Therefore action potential alterations in hiPSC-CMs are similar to αMHC 403/+ myocytes and electrical remodeling at the cardiac myocyte level occurs early in FHC. In support of this, a high incidence of arrhythmias and sudden cardiac death in cardiac troponin T (TnT-I79N) mutant mice have been reported, in the absence of a hypertrophic phenotype. Introducing the TnT-I79N mutation into human induced pluripotent stem cells with CRISPR/Cas9 technique reproduced key features of FHC including myofilament disarray, hypercontractility and diastolic dysfunction, as well as alterations in the ventricular AP 29 . However the proposed mechanisms often differ between specific mutations. In the TNNT2-R92Q mouse model 30 , a decrease in Na V 1.5 and increase in Ca V 1.2 expression and late sodium current (I NaL ) contribute to the phenotype. Isoproterenol further prolongs the AP. In our model, we found no difference in Na V 1.5 or Ca V 1.2 expression and isoproterenol shortened the AP duration.
In addition to a decrease in the expression and function of some potassium channels, we report here an increase in the sensitivity of I to and I Kslow to ISO that appeared to contribute to the APD shortening observed in αMHC 403/+ cardiac myocytes. SAP97 protein expression was decreased in mutant hearts consistent with an alteration in the trafficking and localization of K + channels. SAP97 protein also localized to the β1-adrenergic signaling complex. Importantly we demonstrated a substantial decrease in connexin 43 protein expression that correlates with morphological and histological changes in the hypertrophic myocardium. This can contribute to impaired impulse conduction leading to impulse propagation heterogeneity 15,22 .
Consistent with previous studies, we found no difference in diastolic or systolic intracellular calcium levels in αMHC 403/+ myocytes or in Ca V 1.2 expression compared with wt hearts 7,17 . However, despite the small, but significant difference in the inactivation rate of the channel and calcium transients, we could not increase the Ca v 1.2 current when αMHC 403/+ myocytes were exposed to β1-adrenergic stimulation. Consistent with this we revealed an elevated phosphorylation state of the Ca V 1.2 channel protein extracted from cardiomyopathic αMHC 403/+ mice. We cannot rule out the possibility that altered phosphatase activity contributes to the elevated phosphorylation state of the channel.
It is well documented that β1-adrenergic stimulation enhances the cardiac L-type calcium channel activity. To facilitate co-operative gating the Ca V 1.2 molecules form super-clusters 25 . Assessed with super-resolution microscopy we demonstrate that in ventricular myocytes isolated from hypertrophic αMHC 403/+ hearts Ca V 1.2 proteins form clusters, similar to wt cells, but isoproterenol treatment only increased the size of these clusters significantly in wt cardiac myocytes. The Ca V 1.2 molecules not only clustered differently, but there were more Ca V 1.2 localized together with β1AR's in mutant myocytes. This demonstrates for the first time a direct interaction between Ca V 1.2 and β1ARs, which can explain the higher phosphorylation level of the channel protein in αMHC 403/+ cardiac myocytes. In vitro ISO treatment stimulated colocalization of Ca V 1.2 and β1ARs in wt cardiac myocytes, but in αMHC 403/+ myocytes it decreased the number of the Ca V 1.2 localized with β1ARs.
Here we demonstrate that β-adrenergic stimulation alone is sufficient to increase the probability of arrhythmic activity in αMHC 403/+ mice. Hypertrophic hearts of the αMHC 403/+ mice not only demonstrated ISO-induced arrhythmias, but the mice did not recover well from the ISO challenge resulting in fatigue and tissue damage www.nature.com/scientificreports/ assessed as a significant increase in creatine kinase activity in the αMHC 403/+ hearts but not wt hearts. FHC patients reportedly experience serious cardiovascular events and fatigue following vigorous exercise 12 .
Our results suggest that cytoskeletal disarray contributes to the alterations in ion channel and β1 adrenergic receptor localization and function in the αMHC 403/+ heart. Reduced repolarization reserve and altered conduction velocity are associated with the generation of arrhythmias during β1-adrenergic receptor stimulation in the αMHC 403/+ heart. However, pro-arrhythmic mechanisms may vary depending on the underlying gene mutation reinforcing the need to individualize treatment options with the genetic mutation. We find that inhibition of the β1AR with atenolol relaxed ventricular muscle and improved filling, but also significantly reduced the occurrence of arrhythmic events and allowed the mice to recover fully from the adrenergic challenge. Our data indicate that treatment with selective β1AR blockers may be sufficient to manage arrhythmias in patients carrying an R403Q mutation. Furthermore we highlight the significance of cytoskeletal disarray in altering ion channel location and function and β-adrenergic receptor signaling, leading to electrical instability in the FHC heart.

Materials and methods
Mouse model. Male 35-45 wk old heterozygous αMHC 403/+ mice expressing the human disease-causing mutation R403Q in MYH6 were used. We used male mice because female mice carrying the αMHCR403Q /+ mutation develop hypertrophic cardiomyopathy less consistently than males. Mice were used to establish a colony received as a gift from C and J Seidman (Department of Genetics, Harvard Medical School, MA). Negatively genotyped male age-matched littermates were used as wild type (wt) controls. A total number of 68 wt and 92 αMHC 403/+ mice were used in the study. In the text N indicates number of animals, n indicates number of cells.
All experiments were approved by The Animal Ethics Committee of The University of Western Australia in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes (NHMRC, 8th Edition, 2013; updated 2021) and all methods reported in accordance with the ARRIVE guidelines.
Electrocardiography and echocardiography studies. Mice were anesthetized using methoxyflurane and placed on a warming plate (37 °C). Electrocardiograms (ECGs) were recorded with s.c. bipolar leads (lead II) using a PowerLab data acquisition system with Animal BioAmp for 10 min prior to (control) and following i.p. isoproterenol (ISO, two doses of 20 mg/kg administered 10 min apart) or atenolol (1 mg/kg). Parameters were measured on signal-averaged complexes derived from 10 s of contiguous data. QT interval was corrected for heart rate using the Mitchell method 31 . The relative occurrence of arrhythmic events was measured as number of irregular beats per second of recording (LabChart ADInstruments).
In parallel experiments, echocardiograms were recorded using i13L probe on a Vivid 7 Dimension (GE Healthcare) as previously described 32 .

Electrophysiology
Action potential recordings. Left ventricular cardiac myocytes were isolated as described 32    . Ca 2+ currents were recorded in bath solution (in mM): 140 NaCl, 5.4 CsCl, 1 CaCl 2 , 0.5 MgCl 2 , 5.5 HEPES, 11 glucose, pH 7.4 at 37 °C as previously described 32 . Pipette solution contained (in mM): 115 CsCl, 1 CaCl 2 , 20 TEA-Cl, 10 HEPES, 10 EGTA, 5 MgATP, 0.1 Tris-GTP, 10 phosphocreatine, pH7.05. Ca 2+ currents were monitored by applying a 100 ms test pulse to 10 mV after a 50 ms prepulse to − 30 mV once every 10 s. Kinetics of calcium current inactivation was analysed by fitting current decay after channel activation with a bi-exponential function (yielding tau 1 and tau 2). Current activation and inactivation were also assessed by calculating the integral or "area under a curve" as the area between the graph of y = f(x) and the x-axis.
Both APs and whole cell currents were recorded using an Axopatch 200B voltage-clamp amplifier (Molecular Devices) and an IBM compatible computer with a Digidata 1400 interface and pClamp10 software (Molecular Devices). Cardiac myocytes isolated from the same animal were used for AP and Ca 2+ or K + current measurements. Data analyses were executed with Clampfit10 and GraphPad Prism8, results are reported as mean ± SEM.

Immunprecipitation and in vitro phosphorylation of Ca V 1.2 protein.
Anti-Ca V 1.2 antibody preincubated with Dynabeads Protein G (Thermo Fisher Scientific) used to pull out the Ca V 1.2 protein from tissue homogenates. 2 Unit PKA catalytic subunit (Promega) was used per each μg of immunprecipitated protein to perform in vitro phosphorylation as previously described 33 . All immunoblot experiments were run as triplicate, representative images were shown on figures. For more details and for original blots see SI.
Super-resolution nanoscopy. Cells were imaged on a super-resolution Ground State Depletion (GSD) microscope (Leica Microsystems, Wetzlar, courtesy of Dr. F Santana) in TIRF mode with 150 nm penetration depth as previously described 34 . Fluorescence was detected through a Leica high-power TIRF quad filter cube (QGS HP-T) with emission band-pass filters. The collected frames were reconstructed into super-resolution localization maps using Leica Application Suite (LAS AF) software. Cluster area size was measured from binary masks of the localization maps with a 10 nm pixel size in ImageJ/Fiji as previously described 35 . Statistical analysis. Results are reported as means ± SEM. Statistical analysis was performed using Prism GraphPad software. The Shapiro-Wilk normality test was used to assess whether the data were normally distrib- www.nature.com/scientificreports/ uted. If the data were normally distributed a Brown-Forsyth and Welch ANOVA was used to analyse differences between wt and αMHC 403/+ groups. Where data were not normally distributed a Kruskal-Wallis ANOVA was performed. A Dunn's test was used to correct for multiple comparisons. All chemicals and reagents were purchased from Sigma-Merck unless otherwise specified.