Sex‐specific activation of SK current by isoproterenol facilitates action potential triangulation and arrhythmogenesis in rabbit ventricles

Abstract

Key Points

• It is unknown if a sex difference exists in cardiac apamin‐sensitive small conductance Ca²⁺‐activated K⁺ (SK) current (I _KAS).

• There is no sex difference in I _KAS in the basal condition. However, there is larger I _KAS in female rabbit ventricles than in male during isoproterenol infusion.

• I _KAS activation by isoproterenol leads to action potential triangulation in females, indicating its abundant activation at early phases of repolarization.

• I _KAS activation in females induces negative Ca²⁺–voltage coupling and promotes electromechanically discordant phase 2 repolarization alternans.

• I _KAS is important in the mechanisms of ventricular fibrillation in females during sympathetic stimulation.

Abstract

Sex has a large influence on cardiac electrophysiological properties. Whether sex differences exist in apamin‐sensitive small conductance Ca²⁺‐activated K⁺ (SK) current (I _KAS) remains unknown. We performed optical mapping, transmembrane potential, patch clamp, western blot and immunostaining in 62 normal rabbit ventricles, including 32 females and 30 males. I _KAS blockade by apamin only minimally prolonged action potential (AP) duration (APD) in the basal condition for both sexes, but significantly prolonged APD in the presence of isoproterenol in females. Apamin prolonged APD at the level of 25% repolarization (APD₂₅) more prominently than APD at the level of 80% repolarization (APD₈₀), consequently reversing isoproterenol‐induced AP triangulation in females. In comparison, apamin prolonged APD to a significantly lesser extent in males and failed to restore the AP plateau during isoproterenol infusion. I _KAS in males did not respond to the L‐type calcium current agonist BayK8644, but was amplified by the casein kinase 2 (CK2) inhibitor 4,5,6,7‐tetrabromobenzotriazole. In addition, whole‐cell outward I _KAS densities in ventricular cardiomyocytes were significantly larger in females than in males. SK channel subtype 2 (SK2) protein expression was higher and the CK2/SK2 ratio was lower in females than in males. I _KAS activation in females induced negative intracellular Ca²⁺–voltage coupling, promoted electromechanically discordant phase 2 repolarization alternans and facilitated ventricular fibrillation (VF). Apamin eliminated the negative Ca²⁺–voltage coupling, attenuated alternans and reduced VF inducibility, phase singularities and dominant frequencies in females, but not in males. We conclude that β‐adrenergic stimulation activates ventricular I _KAS in females to a much greater extent than in males. I _KAS activation plays an important role in ventricular arrhythmogenesis in females during sympathetic stimulation.

Key Points

• It is unknown if a sex difference exists in cardiac apamin‐sensitive small conductance Ca²⁺‐activated K⁺ (SK) current (I _KAS).

• There is no sex difference in I _KAS in the basal condition. However, there is larger I _KAS in female rabbit ventricles than in male during isoproterenol infusion.

• I _KAS activation by isoproterenol leads to action potential triangulation in females, indicating its abundant activation at early phases of repolarization.

• I _KAS activation in females induces negative Ca²⁺–voltage coupling and promotes electromechanically discordant phase 2 repolarization alternans.

• I _KAS is important in the mechanisms of ventricular fibrillation in females during sympathetic stimulation.

Introduction

Sex is an important biological variable in cardiac electrophysiology. Females and males exhibit differences in electrocardiographic characteristics, morbidities, clinical manifestations and prognosis to cardiac electrical diseases, and drug sensitivities (Salama & Bett, 2014). For instance, women have longer corrected QT interval (QT_c) and higher incidence of Torsade de Pointes than men, while males account for the vast majority of patients with J wave syndrome (Merri et al. 1989; Antzelevitch et al. 2016). Rabbits have highly matched sex‐linked QT prolongation and arrhythmogenic phenotypes as in humans and are commonly used for studying the sex differences in cardiac ion channels (Salama & Bett, 2014). Previous studies in rabbits reported sex differences in several cardiac ionic currents, including L‐type calcium currents (I _Ca,L) (Sims et al. 2008), inward rectifier potassium currents (I _K1) (Liu et al. 1998) and rapidly (I _Kr) (Liu et al. 1998) and slowly (I _Ks) (Zhu et al. 2013) activating delayed rectifier potassium currents. The apamin‐sensitive small conductance Ca²⁺‐activated K⁺ (SK) current (I _KAS) is abundant in atria (Xu et al. 2003; Lu et al. 2007; Tuteja et al. 2010), sinoatrial node (Torrente et al. 2017), atrioventricular node (Zhang et al. 2008) and Purkinje systems (Reher et al. 2017). Although SK channels are less abundantly expressed in ventricles, previous studies have demonstrated that I _KAS exerted important influences on repolarization and arrhythmogenicity of ventricles under pathological conditions such as heart failure (Chua et al. 2011; Chang et al. 2013b), myocardial infarction (Lee et al. 2013; Zhang et al. 2013), atrioventricular block and hypokalaemia (Chan et al. 2015). Due to the fact that I _KAS is only weakly activated under basal conditions in normal ventricles, whether or not sex differences exist in ventricular I _KAS activation is unknown. Sympathetic activity plays important roles in modulation of the physiological fight‐or‐flight response as well as in cardiac arrhythmogenesis (Chen et al. 2014; Shen & Zipes, 2014). Since β‐adrenergic stimulation could enhance I _Ca,L and increase intracellular Ca²⁺ (Ca_i), we hypothesized that the Ca²⁺‐dependent I _KAS may be activated by isoproterenol, thus allowing us to distinguish the potential sex differences. The purpose of the present study is to test the hypotheses that there are sex differences in I _KAS activation during β‐adrenergic stimulation, and that I _KAS activation plays important roles in ventricular action potential (AP) triangulation and arrhythmogenesis.

Methods

Ethical approval

All experimental procedures were approved by the Institutional Animal Care and Use Committee of Indiana University School of Medicine (IACUC no. 10961), and conformed to the Guide for the Care and Use of Laboratory Animals (National Institutes of Health).

Optical mapping

A total of 50 adult (5–6 months old) New Zealand White rabbits (26 females and 24 males) were used for optical mapping studies. We placed the rabbits in a restrainer and euthanized them by intravenous sodium pentobarbitone overdose (160 mg kg⁻¹, i.v.) before heart removal. Hearts were harvested and Langendorff perfused with Tyrode solution (in mmol  L⁻¹: 128.3 NaCl, 4.7 KCl, 20.2 NaHCO₃, 0.4 NaH₂PO₄, 1.8 CaCl₂,1.2 MgSO₄, 11.1 glucose and bovine serum albumin 40 mg L⁻¹) that was bubbled with 95% O₂–5% CO₂ to maintain a pH of 7.40. All chemicals were purchased from Sigma‐Aldrich (St Louis, MO, USA).

Simultaneous optical mapping of the membrane potential (V _m) and Ca_i was performed using techniques similar to that reported elsewhere (Chang et al. 2013a). The hearts were stained with RH237 (10 μmol L⁻¹, from Thermo Fisher Scientific, Waltham, MA, USA) for V _m mapping and Rhod‐2 AM (1.4 μmol L⁻¹, from Thermo Fisher Scientific) for Ca_i mapping. Blebbistatin (15∼20 μmol L⁻¹, from Tocris Bioscience, Minneapolis, MN, USA) was used to inhibit contraction. The epicardial surfaces of right and left ventricles were excited with a laser (Verdi G5, Coherent Inc., Santa Clara, CA, USA) at a wavelength 532 nm, and the emitted fluorescence was filtered at 715 nm long pass. The fluorescence was recorded with 100 × 100 pixels for a spatial resolution of 0.35 × 0.35 mm² per pixel at 2 ms frame⁻¹ sampling rate. Optical signals were processed both spatially (3 × 3 pixels Gaussian filter) and temporally (3 frames moving average).

Seven different protocols were performed as listed below. Optical mapping data were collected at baseline and after each treatment. Protocol I was performed to test ventricular I _KAS at basal condition. Protocols II–IV aimed to examine if I _KAS is differentially activated by isoproterenol (100 nmol L⁻¹) in females and males and to determine the effects of I _KAS blockade on AP duration (APD) and Ca_i transient duration (Ca_iTD). Apamin (100 nmol L⁻¹) blocked I _KAS conducted by all SK channel isoforms while Lei‐Dab7 (20 nmol L⁻¹), a highly specific SK channel subtype 2 (SK2) blocker, differentiated I _KAS conducted by SK2 from that conducted by other isoforms. Chromanol 293B (I _Ks blocker, 10 μmol L⁻¹) was used to exclude the effects of I _Ks activation during β‐adrenergic stimulation. Protocols V and VI aimed to investigate the mechanisms underlying sex differences of I _KAS. Previous studies found that casein kinase‐2 (CK2) interacted with SK2 channels and diminished I _KAS in neurons (Pachuau et al. 2014). Therefore, we used a CK2 specific inhibitor, 4,5,6,7‐tetrabromobenzotriazole (TBB, 10 μmol L⁻¹), to determine if CK2 inhibition can increase I _KAS in males. BayK8644 (I _Ca,L agonist, 1 μmol L⁻¹) was applied to examine the effects of I _Ca,L on I _KAS activation. Protocol VII was performed to determine the effects of I _KAS activation and blockade on ventricular arrhythmogenicity and ventricular fibrillation (VF) dynamics.

Since ventricular pacing and atrioventricular node ablation could activate I _KAS and induce short‐term cardiac memory in normal rabbit ventricles (Chan et al. 2015), we avoided ventricular pacing in Protocols I–IV and VI for APD measurement. In Protocol I, we optically mapped the ventricles with right atrial (RA) pacing at cycle length (PCL) 200, 250 and 300 ms. Since isoproterenol markedly accelerated the sinus rhythm to about 240–290 beats min⁻¹, in Protocols II–IV, we optically mapped the ventricles at a fixed atrial PCL 200 ms in order to reach 1:1 capture. Optical mapping data in Protocols V and VII were collected under ventricular pacing. However, to minimize the ventricular pacing‐induced I _KAS activation, we left the ventricles unpaced except during induction of cardiac alternans and VF. In all protocols, the agents were sequentially added to the perfusate and recirculated until washout.

Protocol I

Baseline–apamin–washout in three females and three males. After baseline mapping, apamin was added to the perfusate. We then waited 15 min to collect post‐apamin optical mapping data.

Protocol II

Baseline–isoproterenol–apamin–washout in eight females and eight males. After baseline mapping, the heart was perfused with isoproterenol for 10 min before the data collection. Apamin was then added to the perfusate while isoproterenol was kept in recirculation. During the washout, isoproterenol can be completely washed out while apamin cannot be fully washed out.

Protocol III

Baseline–isoproterenol–Lei‐Dab7–apamin–washout in four females. Lei‐Dab was administered while isoproterenol was recirculated in the perfusate. Data collection was performed 10 min after Lei‐Dab administration.

Protocol IV

Baseline–chromanol 293B–isoproterenol–apamin–washout in four females and three males. Chromanol 293B was pretreated for 30 min before isoproterenol administration and kept in the perfusate until washout.

Protocol V

Baseline–TBB–apamin–washout in three males. TBB was perfused for 15 min before data collection and was recirculated until washout.

Protocol VI

Baseline–BayK8644–apamin–washout in three females and three males. BayK8644 was perfused for 15 min before data collection and kept in perfusate until washout.

Protocol VII

Baseline–isoproterenol–apamin–washout with ventricular pacing in four females and four males. In this protocol, RV pacing at PCL 150 ms was performed to induce cardiac alternans. RV burst pacing (PCL 50 ms, duration 10 s) was used to induce VF. VF induction was attempted 10 times at baseline and after each treatment. Any VF occurrence was allowed to continue for 2 min before electrical defibrillation. We optically captured the first 100 ms at the initiation of VF. Mapping of phase singularities (PSs) and dominant frequency was also performed. The PS number was manually counted at each time point separated by 10 frames (20 ms) of data (Hayashi et al. 2007). Pseudo‐electrocardiograms (pECG) were simultaneously monitored throughout the entire experiment with widely spaced electrodes in the tissue bath.

Transmembrane potential recording

To verify the optical mapping findings at the cellular level, we performed transmembrane potential (TMP) recording of cardiomyocytes from the epicardium of LV base using the protocol of baseline–isoproterenol–apamin–washout in four hearts (2 females and 2 males). After immobilization of the Langendorff‐perfused hearts by blebbistatin, standard glass capillary microelectrodes filled with 3 mol L⁻¹ KCl with a tip resistance of ∼20 MΩ at the digitization rate of 10 kHz was used to record TMP. The data were stored with Axoscope 10.2 (Molecular Devices, Sunnyvale, CA, USA). Experiments were performed at 38.3°C.

Ventricular cardiomyocytes isolation

Ventricular cardiomyocytes were isolated from an additional eight rabbits (4 females and 4 males) for use in patch clamp, western blot and immunostaining studies. After intravenous sodium pentobarbitone overdose, hearts were quickly excised by thoracotomy and Langendorff perfused for 5 min with Tyrode solution followed by perfusion with an oxygenated buffer containing (in mmol  L⁻¹): 138 NaCl, 5.4 KCl, 0.3 NaH₂PO₄, 1.2 MgCl₂, 10 HEPES, 10 taurine and 10 glucose (pH 7.4 with NaOH at 37°C). This was followed by a 15 min perfusion with the same buffer containing 1 mg mL⁻¹ collagenase (Type II, 270 U mg⁻¹; Worthington Biochemical Corp., Lakewood, NJ, USA) and 0.1 mg mL⁻¹ protease (Type XIV, ≥3.5 U mg⁻¹; Sigma‐Aldrich). The heart was removed from the Langendorff apparatus. Cardiomyocytes from the base of the left ventricles were dissociated from digested ventricles by gentle mechanical dissociation.

I _KAS densities determined by voltage‐clamp techniques

Voltage‐clamp experiments were conducted at room temperature. An Axopatch 200B amplifier and pCLAMP 10 software (Molecular Devices) were used to generate and record all patch experiments. Pipette electrodes were fabricated using Corning 7056 glass capillaries (Warner Instruments, Hamden, CT, USA).

For whole‐cell I _KAS measurements, the pipette solution contained (in mmol  L⁻¹): 144 potassium gluconate; 1.15 MgCl₂, 5 EGTA, 10 HEPES and CaCl₂ yielding a free (unchelated) [Ca²⁺] of 1 μm (pH 7.25 using KOH). The extracellular solution contained (in mm): 140 N‐methylglucamine, 4 KCl, 1 MgCl₂, 5 glucose and 10 HEPES (pH 7.4 using HCl). The voltage‐clamp protocol consisted of a ramp‐pulse protocol (test pulse: between +40 and −120 mV, holding potential: −70 mV; pulse frequency: every 3 s). Pipette resistance ranged from 1 to 2 MΩ. Series resistance was electronically compensated by 70–80%. Currents were recorded at baseline and after the exposure to 100 nmol L⁻¹ apamin. We also recorded the current in the presence of 100 nmol L⁻¹ isoproterenol and after apamin. To obtain whole‐cell I _KAS, currents recorded in the presence of 100 nmol L⁻¹ apamin were digitally subtracted from those recorded in its absence.

To verify the selectivity of Lei‐Dab7 for I _KAS conducted by SK2 but not SK3, we tested the effects of Lei‐Dab7 on human embryonic kidney (HEK) 293 cells transfected with KCNN2 and KCNN3, respectively. HEK 293 cell were cultured in Iscove's modified Dulbecco's medium (Thermo Fisher Scientific) with 10% fetal bovine serum and 1% penicillin/streptomycin in 5% CO₂ at 37°C; 35 mm dishes of HEK 293 cells were transiently transfected using Effectene Transfection Reagent (Qiagen, Germantown, MD, USA) according to the manufacturer's protocol and were harvested for patch clamp experiment 48∼72 h later. The SK2 and SK3 clones were developed in our laboratory. HEK 293 cells were transfected with 2.0 μg of KCNN2/pIRES‐EGFP2 or 2.0 μg of KCNN3/pIRES‐EGFP2 plasmids, respectively. Single cells were picked and propagated in selection medium containing hygromycin 200 μg mL⁻¹. Currents were recorded at baseline and after exposure to 20 nmol L⁻¹ Lei‐Dab 7, followed by addition of 100 nmol L⁻¹ apamin. The LeiDab7‐sensitive current and apamin‐sensitive current (in the presence of LeiDab7) were calculated, respectively.

Western blot

Western blot was performed in isolated rabbit left ventricular cardiomyocytes. A Lowry protein assay was performed before western blot to guarantee that the protein loadings were identical among lanes. Samples were loaded on SDS‐PAGE and transferred to a nitrocellulose membrane. The blot was probed with the following antibodies: anti‐K_Ca2.2 (SK2) antibody (1:500, APC‐028, Alomone, Jerusalem, Israel), anti‐CK2α antibody (1:500, ab181734, Abcam, Cambridge, UK), anti‐CK2β antibody (1:500, ab201990, Abcam) and anti‐SERCA antibody 2A7‐A1 (1:1000). Antibody‐binding protein bands were quantified with a Bio‐Rad (Hercules, CA, USA) Personal Fx phosphorimager. The control sample for SK2 was the heterologously expressed human isoform of SK2 in cultured HEK 293 cells.

Immunofluorescence confocal microscopy

Both isolated myocytes and ventricular tissues were used for immunofluorescence staining and confocal microscopy. Isolated rabbit ventricular myocytes were fixed with 1% paraformaldehyde for 30 min and seeded on the laminin‐coated glass slides for at least 1 h to allow cell attachment. In addition, ventricular tissue samples were fixed with 4% paraformaldehyde for 45 min, followed by storage in 70% alcohol for at least 48 h. The tissue samples were then processed routinely and embedded in paraffin. Tissue sections (5 μm in thickness) were deparaffinized and hydrated by multiple xylene and ethanol washes.

Slides with either isolated myocytes or ventricular tissue were washed with water and PBS in Tween 20 at least 3 times. Samples were then permeabilized and blocked in PBS with 3% BSA and 0.2% Triton X‐100 for 1 h. After blocking, slides were incubated with anti‐K_Ca2.2 (SK2) antibody (APC‐028, Alomone) 1:200 diluted in PBS (BSA 1%) at 4°C overnight. Subsequently, slides were washed 3 times for 5 min with PBS and incubated with protein A conjugated with fluorescent dyes (Alexa 488, Thermo Fisher Scientific, 1:2000) and 4′,6‐diamidino‐2‐phenylindole for 1 h. After this step, the slides were again washed 3 times for 5 min with PBS followed by mounting with coverslips. The slides were allowed to dry at room temperature and then sealed with nail polish overnight. For comparison, samples from females and males were stained simultaneously.

Confocal images were obtained through a DMI6000 adaptive focus automated inverted microscope, Leica TCS SP8 FSU (argon ion laser) spectral confocal system with HyD supersensitivity detection. For comparison of the fluorescence intensities, slides from females and males were microscopically detected on the same day using identical microscopy settings.

Statistics and data analysis

APD at the level of 25% (APD₂₅) and 80% (APD₈₀) repolarization of the AP was optically measured. We used APD₂₅ to characterize the early phases (phases 1 and 2) of repolarization and APD₈₀ to represent the entire repolarization (phase 1, 2 and 3). Thus, the differences between APD₂₅ and APD₈₀ (APD₈₀ − APD₂₅) were used to characterize the repolarization at phase 3. The APD₂₅/APD₈₀ ratio was used as a quantitative measurement of the AP morphology, with a smaller APD₂₅/APD₈₀ ratio reflecting greater AP triangulation and a larger APD₂₅/APD₈₀ ratio representing AP squaring. Intracellular calcium transient duration at 25% and 80% recovery (Ca_iTD₂₅ and Ca_iTD₈₀) were used as measures of early and total Ca_i duration, respectively. We also calculated the differences between Ca_iTD and APD (Ca_iTD₂₅ − APD₂₅ and Ca_iTD₈₀ − APD₈₀) to characterize the Ca_i–V _m coupling. We averaged the APD in the regions of interest and presented them in the summary data.

Continuous variables were expressed as mean ± SEM. Student's paired t test was used to compare two variables from the same rabbit (such as APD₂₅ before and after apamin). Unpaired t tests were used to compare variables between two groups (such as ΔAPD₂₅ between females and males). Multiple t tests were used to compare I _KAS current densities between sexes at different membrane potentials. One‐way or two‐way ANOVA with the appropriate post hoc test, as indicated in each experiment, was used to compare the means among three or more different variables. A two‐sided P value of <0.05 was considered statistically significant.

Results

Isoproterenol activates I _KAS in female rabbit ventricles

Consistent with previous studies (Chua et al. 2011), I _KAS blockade by apamin only minimally prolonged APD (by <5%) in ventricles from females and males at baseline (Protocol I, Fig. 1). The magnitudes of APD prolongation was not significantly different between males and females. We then tested the effects of I _KAS blockade in the presence of isoproterenol in female rabbit ventricles (Fig. 2 A, Protocol II). The AP exhibited a prominent phase 2 plateau at baseline. Isoproterenol shortened and triangulated the AP. Subsequent apamin administration (with isoproterenol in the recirculation) significantly prolonged APD₂₅ (from 29 ± 2 to 40 ± 2 ms, P < 0.001) and APD₈₀ (from 92 ± 4 to 101 ± 3 ms, P < 0.001, Fig. 2 B and C), indicating I _KAS activation during isoproterenol perfusion. In addition to APD prolongation, apamin also affected the AP morphology, i.e. reversing isoproterenol‐induced AP triangulation and restoring the phase 2 plateau of repolarization, by lengthening APD₂₅ (11 ± 1 ms and 41 ± 6%) more prominently than APD₈₀ (9 ± 1 ms and 10 ± 1%, P = 0.018 for absolute value and P < 0.001 for percentage, Fig. 2 D). As a result, the ratio APD₂₅/APD₈₀ was significantly increased (from 32 ± 2% to 40 ± 2%, P < 0.001, Fig. 2 E). Apamin significantly abbreviated differences between APD₈₀ and APD₂₅ from 63 ± 2 to 60 ± 3 ms (P = 0.043, Fig. 2 F), indicating a slight acceleration of phase 3 repolarization. Furthermore, as shown in Fig. 2 G, the prolongation of the entire repolarization (positive ΔAPD₈₀) was associated with the abbreviation of phase 3 repolarization (negative Δ(APD₈₀ − APD₂₅)) in 7/8 females (circles in quadrant IV), indicating the APD prolongation was completely attributable to the prolongation of phase 2 repolarization. After isoproterenol washout, APD was further prolonged towards the baseline level. As shown in Fig. 2 H, TMP recording in cardiomyocytes of female rabbits also verified the optical mapping findings. These results indicate that I _KAS is activated during β‐adrenergic stimulation in female rabbit ventricles.

Figure 1. Effects of I _KAS blockade on APD at basal condition.

A and B, representative V _m traces, ΔAPD₂₅ and ΔAPD₈₀ maps in female (A) and male (B) rabbit ventricles at RA PCL 200, 250 and 300 ms under Protocol I. The apamin‐induced APD prolongation was less than 5% in the absence of β‐adrenergic stimulation. Representative V _m traces were obtained at the LV base. C, summary data showed that no significant difference existed in ΔAPD₂₅ and ΔAPD₈₀ between females and males at baseline (by two‐way ANOVA with Sidak's post hoc test). [Color figure can be viewed at http://wileyonlinelibrary.com]

Figure 2. Effects of I _KAS blockade on APD during isoproterenol infusion in female rabbit ventricles.

A, representative V _m traces, APD₂₅, APD₈₀ and APD₈₀ − APD₂₅ maps at baseline, during isoproterenol, after apamin and after washout (Protocol II). At baseline, the AP exhibited a prominent phase 2 plateau. Isoproterenol markedly shortened APD and more prominently APD₂₅ than APD₈₀, leading to a short and triangular AP. Representative V _m traces were obtained at the LV base. B, C and D, apamin significantly prolonged both APD₂₅ and APD₈₀, but more prominently APD₂₅ (D), consequently reversing the AP triangulation and restoring the AP plateau. After washout, the APD was further prolonged to a level similar to the baseline. E, apamin significantly increased the ratio of APD₂₅/APD₈₀, resulting in AP squaring and plateau restoration. F, apamin significantly abbreviated APD₈₀ − APD₂₅, indicating an acceleration of phase 3 repolarization. Student's paired t tests were performed in C–F. G, the prolongation of the total repolarization by apamin (represented by ΔAPD₈₀) corresponded to the shortening of phase 3 repolarization (Δ(APD₈₀ − APD₂₅)) in 7 out of 8 females (circles in quadrant IV). Therefore, the apamin predominantly lengthened APD₂₅. H, representative TMP recording in ventricular cardiomyocytes from a female rabbit. ISO, isoproterenol. [Color figure can be viewed at http://wileyonlinelibrary.com]

Ca_i mapping was simultaneously performed in females under Protocol II (Fig. 3). Compared with baseline, isoproterenol markedly shortened Ca_iTD₂₅ and Ca_iTD₈₀ (Fig. 3 A) but increased the peak Ca_i (Fig. 3 B). These findings indicate higher Ca_i at phase 2 repolarization. In addition, isoproterenol markedly increased the differences between Ca_iTD and APD (Ca_iTD − APD) and more drastically for Ca_iTD₂₅ − APD₂₅ than for Ca_iTD₈₀ − APD₈₀ (30 ± 2 vs. 23 ± 3, P = 0.041, Fig. 3 C and F), suggesting an early activated outward current in compensation for the amplified I _Ca,L during phase 2 repolarization. Apamin had minimal effect on Ca_iTD and peak Ca_i (Fig. 3 B and D), but produced significant abbreviation of Ca_iTD₂₅ − APD₂₅ (from 30 ± 2 to 20 ± 2 ms, P < 0.001, Fig. 3 E) and elimination of the isoproterenol‐induced differences Ca_iTD₂₅ − APD₂₅ and Ca_iTD₈₀ − APD₈₀ (P = 0.563, Fig. 3 G). Taken together, isoproterenol increases Ca_i, which in turn activates I _KAS to shorten the APD. Therefore, isoproterenol increases the differences between Ca_iTD₂₅ and APD₂₅ during phase 2 (negative Ca_i–APD coupling).

Figure 3. Effects of I _KAS blockade on Ca_i and Ca_i–V _m coupling during isoproterenol infusion in female rabbit ventricles.

A, representative Ca_i traces, Ca_iTD₂₅ and Ca_iTD₈₀ maps at baseline, during isoproterenol, after apamin and after washout (Protocol II). Compared with baseline, isoproterenol markedly shortened Ca_iTD₂₅ and Ca_iTD₈₀. Apamin only slightly prolonged Ca_iTD. After washout, the Ca_iTD was prolonged towards the baseline level. Representative Ca_i traces were obtained at the LV base. B, Ca_i F/F ₀ showed that isoproterenol markedly increased peak Ca_i F/F ₀ and accelerated Ca_i upstroke and decay. Apamin had minimal effect on peak Ca_i F/F ₀ and Ca_i transient. C, overlapped Ca_i and V _m traces and Ca_iTD − APD maps showed that isoproterenol enlarged the differences between Ca_i and APD compared with baseline, especially Ca_iTD₂₅ − APD₂₅. Apamin markedly attenuated the Ca_iTD₂₅ − APD₂₅ attributable to the remarkably prolonged APD₂₅. Ca_iTD₈₀ − APD₈₀ remained similar before and after apamin. Washout further abbreviated Ca_iTD − APD towards baseline. D, apamin had insignificant effect on Ca_iTD₂₅ and Ca_iTD₈₀. E, apamin significantly decreased Ca_iTD₂₅ − APD₂₅ while having an insignificant effect on Ca_iTD₈₀ − APD₈₀. F, during isoproterenol, Ca_iTD₂₅ − APD₂₅ was significantly larger than Ca_iTD₈₀ − APD₈₀. G, after apamin, the differences between Ca_iTD₂₅ − APD₂₅ and Ca_iTD₈₀ − APD₈₀ were eliminated. Student's paired t test was performed in D–G. [Color figure can be viewed at http://wileyonlinelibrary.com]

SK2, but not SK3 or I _KS, is responsible for isoproterenol‐induced AP triangulation in females

Among subtypes of SK channels which generate I _KAS, SK2 and SK3 are important to ventricular repolarization and are completely blocked by 100 nmol L⁻¹ apamin (Yu et al. 2014). Here, we took advantage of the specific SK2 blocker Lei‐Dab7 (20 nmol L⁻¹) (Shakkottai et al. 2001) to distinguish the contributions of I _KAS conducted by different SK subtypes. Firstly, we verified the selectivity of Lei‐Dab7 on SK2 over SK3 (Fig. 4 A and B). Lei‐Dab7 at 20 nmol L⁻¹ inhibited the majority of whole‐cell currents in KCNN2‐transfected HEK293 cells but had only minimal effect on KCNN3‐transfected cells. These results indicate Lei‐Dab7 at 20 nmol L⁻¹ only blocks SK2 channels.

Figure 4. SK2, but not SK3, is responsible for AP triangulation in females.

A and B, representative whole cell currents sequentially recorded from KCNN2‐ (A) or KCNN3‐ (B) transfected HEK293 cells at baseline and in the presence of Lei‐Dab7 (20 nmol L⁻¹) and subsequently apamin (100 nmol L⁻¹). Lei‐Dab‐sensitive current was calculated as the difference between baseline and Lei‐Dab7. Apamin‐sensitive current in the presence of Lei‐Dab7 was calculated as the difference between Lei‐Dab7 and apamin. C, representative optical V _m traces, APD₂₅ and APD₈₀ maps at baseline, during isoproterenol, after Lei‐Dab7, after apamin and after washout at PCL 200 ms (Protocol III). Representative V _m traces were obtained at the LV base. D, compared with isoproterenol, Lei‐Dab7 prolonged both APD₂₅ and APD₈₀, but more prominently APD₂₅, thus leading to the phase 2 plateau restoration. E, compared with Lei‐Dab7, subsequently administered apamin did not further prolong APD. F, summary data showed that APD₂₅ and APD₈₀ were significantly longer after Lei‐Dab7 and after apamin, both in comparison with those at baseline. APD were similar between Lei‐Dab7 and apamin. *P < 0.05 by one‐way ANOVA with Tukey's post hoc test. G, I _KAS conducted by SK2 (represented by ΔAPD_{LeiDab7−isoproterenol}) was significantly more activated than I _KAS by SK3 (represented by ΔAPD_{Apamin−LeiDab7}) measured at both APD₂₅ and APD₈₀ (by Student's paired t test). ISO, isoproterenol. [Color figure can be viewed at http://wileyonlinelibrary.com]

We further tested the effects of Lei‐Dab7 (20 nmol L⁻¹) in females under Protocol III (Fig. 4). In the presence of isoproterenol, SK2 blockade by Lei‐Dab7 significantly prolonged APD₂₅ and to a smaller extent APD₈₀ (Fig. 4 C and D). Subsequently administered apamin, which only blocked SK3 with Lei‐Dab7 pretreatment, failed to further lengthen APD (Fig. 4 E and F). In other words, SK2 blockade (ΔAPD_{LeiDab7−isoproterenol}) had significantly larger effects than SK3 blockade (ΔAPD_{Apamin−LeiDab7}) on both APD₂₅ (11 ± 1 vs. 1 ± 1 ms, P = 0.007) and APD₈₀ (13 ± 2 vs. 1 ± 1 ms, P = 0.004, Fig. 4 G). These results suggest that SK2, rather than SK3, is the predominant channel isoform underlying I _KAS activation during β‐adrenergic stimulation.

Similar to I _KAS, I _Ks is also calcium sensitive and is enhanced during sympathetic stimulation (Aflaki et al. 2014; Banyasz et al. 2014; Bartos et al. 2017). To exclude I _Ks as a cause of AP triangulation in females, we pretreated ventricles with the I _Ks blocker chromanol 293B (10 μmol L⁻¹) in Protocol IV (Fig. 5). With I _Ks blockade, isoproterenol still shortened and triangulated AP. Subsequent apamin administration significantly prolonged APD₂₅ and to a lesser extent APD₈₀, indicating that I _KAS contributes to APD shortening and AP triangulation independently of I _Ks.

Figure 5. I _Ks is not responsible for AP triangulation in females.

A, representative V _m traces, APD₂₅, APD₈₀ and APD₈₀ − APD₂₅ maps at baseline, after pretreatment of I _Ks blocker chromanol 293B, after isoproterenol, after apamin and after washout (Protocol III). Representative V _m traces were obtained at the LV base. B, D and E, compared with baseline, chromanol 293B prolonged APD₂₅ and APD₈₀ (B), but more prominently APD₈₀ (D), by markedly slowing repolarization at phase 3 (APD₈₀ − APD₂₅; E). With chromanol 293B pretreatment, isoproterenol still shortened and triangulated AP. C, F and G, apamin prolonged APD₂₅ and APD₈₀ (C and F) but more prominently APD₂₅ than APD₈₀ (G). Student's paired t tests were performed in D–G. [Color figure can be viewed at http://wileyonlinelibrary.com]

I _KAS is minimally activated by isoproterenol in male rabbit ventricles

To test sex differences in I _KAS activation, we repeated Protocol II in male rabbit ventricles (Fig. 6 A). In contrast to females, apamin prolonged APD very slightly after isoproterenol infusion, although the increment was statistically significant (for APD₂₅, from 30 ± 2 to 33 ± 2 ms and for APD₈₀, from 93 ± 3 to 98 ± 4 ms, P = 0.001 for both, Fig. 6 B and C). The prolongations of APD₂₅ and APD₈₀ were similar (for absolute value: 3 ± 0 ms vs. 4 ± 0 ms, P = 0.154; for percentage: 11 ± 0% and 6 ± 1%, P = 0.060; Fig. 6 D) and barely altered the APD₂₅/APD₈₀ ratio (32 ± 2 vs. 33 ± 2, P = 0.101, Fig. 6 E), indicating no major AP morphology change by apamin. Apamin had little effect on phase 3 repolarization (represented by APD₈₀ − APD₂₅, 64 ± 3 ms vs. 65 ± 4 ms, P = 0.180, Fig. 6 F). Figure 6 G shows that apamin abbreviated phase 3 repolarization but prolonged the APD₈₀ in 4/8 male rabbit ventricles (positive ΔAPD₈₀ with negative Δ(APD₈₀ − APD₂₅), circles in quadrant IV). These data indicate that the APD prolongation, although only slight, was completely attributable to phase 2 prolongation in these four rabbits. For the other four males, however, the APD₈₀ prolongation relied on the phase 3 prolongation (positive ΔAPD₈₀ with positive Δ(APD₈₀ − APD₂₅), circles in quadrant I), which was a different apamin responses from females. As shown in Fig. 6 H, we further verified the findings of optical mapping at the cellular level by TMP recording in cardiomyocytes from male rabbits. Taken together, in contrast to females, males only exhibit minimal I _KAS activation both at baseline and during isoproterenol infusion. I _KAS activation could shorten either phase 2 or phase 3 repolarization of the AP in males.

Figure 6. Effects of I _KAS blockade on APD during isoproterenol in male rabbit ventricles.

A, representative V _m traces, APD₂₅ and APD₈₀ maps at baseline, during isoproterenol, after apamin and after isoproterenol washout (Protocol II). After isoproterenol, APD was markedly shortened and triangulated. Apamin slightly prolonged APD but did not reverse the AP triangulation. After washout, AP plateau was restored and APD was lengthened towards baseline. Representative V _m traces were obtained at the LV base. B and C, apamin prolonged APD₂₅ and APD₈₀ significantly but only slightly. D, apamin prolonged APD₂₅ and APD₈₀ to a similar extent. E, APD₂₅/APD₈₀ ratios were similar before and after apamin. F, APD₈₀ − APD₂₅ was similar before and after apamin. Student's paired t test was performed in C–F. G, APD₈₀ prolongation by apamin was partly attributed to phase 3 prolongation in 4/8 ventricles (circles in quadrant I). While in the other 4 ventricles, the APD₈₀ prolongation corresponded to slight phase 3 acceleration (circles in quadrant IV). H, representative TMP recording in ventricular cardiomyocytes from a male rabbit. [Color figure can be viewed at http://wileyonlinelibrary.com]

To evaluate the mechanism of isoproterenol‐induced AP shortening and triangulation that is independent of I _KAS activation in males, we pretreated male rabbits with chromanol 293B to inhibit I _Ks, which is activated by isoproterenol and exhibits stronger activation in males than females (Zhu et al. 2013). As shown in Fig. 7, chromanol 293B prolonged both APD₂₅ and APD₈₀ and more prominently APD₈₀. Compared with females, males had significantly larger APD prolongation, suggesting stronger I _Ks activation in males than females. However, isoproterenol still shortened and triangulated APD even with I _Ks and subsequent I _KAS blockade. These results indicate that isoproterenol‐induced APD shortening and triangulation in males cannot be explained only by I _Ks and I _KAS activation.

Figure 7. I _Ks is not responsible for AP triangulation in males.

A, representative V _m traces, APD₂₅, APD₈₀ and APD₈₀ − APD₂₅ maps at baseline, after pretreatment with I _Ks blocker chromanol 293B, after isoproterenol, after apamin and after washout (Protocol III). Representative V _m traces were obtained at the LV base. B, compared with baseline, chromanol 293B markedly prolonged APD₂₅ and APD₈₀ but more prominently APD₈₀. With chromanol 293B pretreatment, isoproterenol still shortened and triangulated AP. C, apamin only minimally prolonged APD₂₅ and APD₈₀. D, compared with females, males had significant larger response in chromanol 293B‐induced APD prolongation (by unpaired t test). [Color figure can be viewed at http://wileyonlinelibrary.com]

The differential I _KAS activation after isoproterenol infusion between sexes is summarized in Fig. 8. Females had significantly greater responses to I _KAS blockade than males in APD₂₅ (absolute value: 11 ± 1 vs. 3 ± 0 ms, P < 0.001; percentage: 39 ± 5% vs. 11 ± 2%, P < 0.001; Fig. 8 A) and in APD₈₀ (absolute value: 9 ± 1 vs. 5 ± 1 ms, P = 0.003; percentage: 10 ± 1% vs. 5 ± 1%, P = 0.006; Fig. 8 B). Apamin slightly accelerated phase 3 repolarization in females but not in males (Δ(APD₈₀ − APD₂₅): −2 ± 1 vs. 0 ± 1 ms, P = 0.035, Fig. 8 C). The APD₂₅/APD₈₀ ratio was significantly larger in females than in males (40 ± 2% vs. 34 ± 2%, P = 0.037, Fig. 8 D), indicating differential AP morphology changes by I _KAS blockade between sexes.

Figure 8. Sex differences of the apamin effects (Protocol II).

A and B, in the presence of isoproterenol, female rabbit ventricles exhibited significantly larger apamin‐induced prolongation than male rabbit ventricles for both (A) APD₂₅ and (B) APD₈₀. C, females and males responded to apamin significantly differently on phase 3 repolarization (Δ(APD₈₀ − APD₂₅)). Female rabbit ventricles exhibited an abbreviation of Δ(APD₈₀ − APD₂₅) while male rabbit ventricles did not. D, APD₂₅/APD₈₀ ratio after apamin was significantly larger in females than in males. Unpaired t tests were performed in comparison between sexes.

I _KAS current densities are higher in females than males

To better understand the mechanisms of the sex differential ventricular I _KAS activation, we examined I _KAS at cellular levels by performing patch clamp studies in isolated ventricular cardiomyocytes. As shown in Fig. 9 A, with free Ca²⁺ concentration of 1 μm, mean I _KAS densities were significantly larger in cells from females than males measured at a potential >0 mV (at +10mV: 0.35 ± 0.03 vs. 0.19 ± 0.03 pA pF⁻¹; at +20 mV: 0.47 ± 0.04 vs. 0.21 ± 0.03 pA pF⁻¹; at +30 mV: 0.75 ± 0.05 vs. 0.28 ± 0.05 pA pF⁻¹; at +40 mV: 0.98 ± 0.09 vs. 0.34 ± 0.06 pA pF⁻¹; P < 0.05). We further tested I _KAS densities with pretreatment of 100 nmol L⁻¹ isoproterenol (Fig. 9 B). In the presence of isoproterenol, I _KAS densities were significantly larger in myocytes from female than male rabbits at a potential >0 mV (at +10 mV: 0.42 ± 0.08 vs. 0.13 ± 0.03 pA pF⁻¹; at +20 mV: 0.70 ± 0.09 vs. 0.24 ± 0.02 pA pF⁻¹; at +30 mV: 0.99 ± 0.10 vs. 0.38 ± 0.03 pA pF⁻¹; at +40 mV: 1.30 ± 0.12 vs. 0.51 ± 0.03 pA pF⁻¹; P < 0.05). These patch clamp results supported our observations in optical mapping.

Figure 9. I _KAS current density is larger in females than males.

A and B, I _KAS densities of isolated ventricular myocytes by patch clamp using ramp‐pulse protocol (test pulse: between +40 and −120 mV; holding potential −70 mV; pulse frequency: every 3 s) in the absence (A) or presence (B) of isoproterenol. Left: representative membrane current traces obtained from female (a) and male (b) rabbit ventricular myocytes. Currents were recorded with an intrapipette free Ca²⁺ of 1 μmol L⁻¹ in the absence (I _K‐baseline or I _{K‐isoproterenol}, blue) and the presence (I _K‐apamin, red) of 100 nmol L⁻¹ apamin. Middle: amplified images. Right: I _KAS was calculated as the difference between I _K‐baseline/I _{K‐isoproterenol} and I _K‐apamin. Ac and Bc, I _KAS density–voltage relationships from female and male rabbit ventricular cardiomyocytes. *P < 0.05 by multiple t tests. [Color figure can be viewed at http://wileyonlinelibrary.com]

SK2 channel protein expression is higher in females than males

Since SK2 carried the enhanced I _KAS by isoproterenol in females, we examined the SK2 channel expression by western blot and confocal immunofluorescence microscopy in both sexes. Western blotting in Fig. 10 A showed that SK2 protein expression was significantly higher in female rabbit cardiomyocytes than in males (SK2 normalized to SERCA: 0.075 ± 0.005 vs. 0.043 ± 0.004, P = 0.003). As further shown in Fig. 10 B and C, confocal immunofluorescence microscopy detected the expression of SK2 along the Z‐line of plasma membrane at higher fluorescence intensities in females than that in males from both isolated cardiomyocytes and ventricular tissue sections. These results indicate that sex differences exist in SK2 protein expression, which contribute to sex differences in I _KAS.

Figure 10. SK2 western blotting and immunostaining of rabbit ventricles.

A, western blot showing that SK2 channel protein expression (normalized to SERCA) was significantly higher in females than males (by unpaired t test). hSK2 is heterologously expressed human isoform of SK2 in HEK 293 cells, which serves as the positive control. B, representative confocal immunofluorescence microscopy of SK2 staining in isolated ventricular myocytes. The fluorescence intensity was higher in female than in male ventricular myocytes. Protein A conjugated with Alexa 488 without pretreatment of anti‐SK2 antibody was used as a negative control. The fluorescence signals were detected under an identical confocal microscopy setting. a.u., arbitrary units. C, representative confocal immunofluorescence microscopy of SK2 in left ventricular tissues. The fluorescence signals were stronger in female rabbit ventricular tissues than in males. The fluorescence signals were detected under an identical confocal microscopy setting. [Color figure can be viewed at http://wileyonlinelibrary.com]

CK2 activity correlates with I _KAS suppression in males

In its participation in a broad range of cellular signalling pathways, CK2 phosphorylates calmodulin, thus lowing the Ca²⁺ affinity of SK2 channels and reducing I _KAS in neurons (Pachuau et al. 2014). To evaluate if CK2 activity contributed to I _KAS suppression in male rabbit ventricles, we tested the effect of a CK2‐specific blocker, TBB, on APD (Protocol V). As shown in Fig. 11 A, TBB drastically abbreviated and triangulated APD. Subsequently, apamin significantly prolonged both APD₂₅ (from 15 ± 2 to 22 ± 1 ms, P = 0.034) and APD₈₀ (from 39 ± 2 to 64 ± 4 ms, P = 0.045, Fig. 11 B and C), indicating that CK2 inhibition in males can increase the I _KAS activation. Although western blotting showed CK2α and CK2β had similar expression levels between sexes (Fig. 11 D), the significantly higher CK2/SK2 ratio in males than in females (CK2α/SK2: 12 ± 2 vs. 21 ± 2, P = 0.012; CK2β/SK2 ratio: 11 ± 1 vs. 18 ± 2, P = 0.031, Fig. 10 E) may correlate with smaller I _KAS in males.

Figure 11. Effects of I _KAS blockade during CK2 inhibition in rabbit ventricles.

A, representative V _m traces, APD₂₅ and APD₈₀ maps at baseline, after CK2 inhibitor TBB and after apamin (Protocol V). Compared with baseline, TBB prominently abbreviated APD₂₅ and APD₈₀, leading to a short and triangular AP. B and C, apamin significantly prolonged APD₂₅ and APD₈₀. Representative V _m traces were obtained at the LV base. Student's paired t test was performed in C. D, western blot showing similar protein expression of CK2α and CK2β between cardiomyocytes from female and male rabbit ventricles. E, males had significantly higher CK2α/SK2 and CK2β/SK2 ratios than females (by unpaired t test). [Color figure can be viewed at http://wileyonlinelibrary.com]

Sex differences in I _KAS are not secondary to sex differences in I _Ca,L

SK2 is molecularly coupled with L‐type calcium channels (LTCCs) (Lu et al. 2007). The magnitude of I _Ca,L density is known to be smaller in males than in females (Vizgirda et al. 2002). It is therefore possible that the weaker I _Ca,L in males partially explains the lower of I _KAS. To test this hypothesis, BayK8644 was used to amplify I _Ca,L (Protocol VI, Fig. 12). In males, compared with baseline, BayK8644 led to markedly enlarged Ca_iTD, drastically increased peak Ca_i and prolonged APD. However, subsequently administered apamin failed to further prolong APD, indicating that BayK8644 failed to activate I _KAS in males. As a comparison, apamin prolonged APD in the presence of BayK8644 in females. These results are inconsistent with the hypothesis that the sex differences in I _KAS are simply secondary to the sex differences in I _Ca,L.

Figure 12. Effects of I _KAS blockade during BayK8644 in male (A–F) and female (G and H) rabbit ventricles.

A, representative Ca_i traces, Ca_iTD₂₅ and Ca_iTD₈₀ maps at baseline, after BayK8644, after apamin and after washout in a male rabbit (Protocol VI). BayK8644 markedly prolonged Ca_iTD₂₅ and Ca_iTD₈₀ while apamin had minimal effect on Ca_iTD_, B, overlapped Ca_i traces and ΔCa_iTD maps showed BayK8644 markedly prolonged Ca_iTD compared with baseline. C, BayK8644 markedly increased the peak Ca_i F/F ₀. D, corresponding V _m traces, APD₂₅ and APD₈₀ maps. Compared with baseline, BayK8644 markedly prolonged APD₂₅ and APD₈₀. E, apamin only slightly prolonged APD₂₅ and APD₈₀. Representative V _m and Ca_i traces were obtained at the LV base. F, with BayK8644 pretreatment, apamin prolonged APD₂₅ and APD₈₀, both insignificantly (by Student's paired t test). G and H, representative V _m traces, APD₂₅ and APD₈₀ maps in a female rabbit (Protocol VI). Compared with baseline, BayK8644 markedly prolonged APD₂₅ and APD₈₀. Subsequent apamin administration further prolonged APD₂₅ and APD₈₀, but more prominently APD₂₅. I, female rabbits had significant larger response in apamin‐induced APD prolongation than male rabbits with BayK8644 pretreatment (by unpaired t test). [Color figure can be viewed at http://wileyonlinelibrary.com]

I _KAS blockade eliminates negative Ca_i–V _m coupling at phase 2 and attenuates APD alternans

Cardiac repolarization alternans often precedes life‐threatening ventricular arrhythmias. In alternans, Ca_i and V _m are coupled either positively (a large Ca_i transient prolongs APD of the same beat) or negatively (a large Ca_i transient shortens APD of the same beat) (Kennedy et al. 2017). To test the effects of I _KAS on alternans, rapid pacing was performed in females (Protocol VII, Fig. 13). At basal condition, Ca_i–V _m coupling was positive such that alternans was electromechanically concordant. During isoproterenol (Fig. 13 A), the Ca_i–V _m coupling was still positive and the alternans remained electromechanically concordant at the level of APD₈₀. However, at early repolarization phases at the APD₂₅ level, Ca_i–V _m coupling became negative such that alternans became electromechanically discordant. In other words, during phase 2, a larger Ca_i transient (beat 1) led to a shorter and more triangular AP, while a smaller Ca_i transient (beat 2) corresponded to a longer and less triangular AP. As shown in Fig. 13 B, I _KAS blockade by apamin eliminated the negative Ca_i–V _m coupling and APD₂₅ alternans (represented by ΔAPD_{25,beat1−beat2}) and also markedly attenuated APD₈₀ alternans (represented by ΔAPD_{80,beat1−beat2}). Figure 13 C shows that apamin prolonged APD₂₅ of beat 1 more prominently than that of beat 2 (13 ± 1 vs. 6 ± 2 ms, P = 0.004) while APD₈₀ prolongations were similar between the two adjacent beats (5 ± 1 vs. 7 ± 1 ms, P = 0.250). Figure 13 D shows that apamin eliminated alternans of APD₂₅ (from −7 ± 0 to 0 ± 1 ms, P = 0.002) and attenuated alternans of APD₈₀ (from 21 ± 1 to 17 ± 1, P = 0.010). These results indicate that β‐adrenergic stimulation elicits more abundant I _KAS activation at phase 2 than at phase 3 repolarization and during beats with a larger Ca_i transient than those with smaller Ca_i transient, leading to negative Ca_i–V _m coupling and electromechanically discordant alternans at early repolarization phases. I _KAS blockade mainly prolonged phase 2 repolarization in the beats with larger Ca_i transients and had less effect on the beats with smaller Ca_i transients, resulting in the elimination of negative Ca_i–V _m coupling and electromechanically discordant alternans. As a comparison (Fig. 13 E and F), in male rabbits, isoproterenol induced Ca_i alternans at PCL 150 ms with less prominent APD alternans. Both Ca_i transient and V _m had little response to apamin. Therefore, the elimination of phase 2 repolarization alternans by I _KAS blockade in female rabbits suggests its potential antiarrhythmic effects during β‐adrenergic stimulation.

Figure 13. I _KAS blockade eliminates negative Ca_i–V _m coupling at phase 2 and attenuates APD alternans in female (A‐‐D) but not male (E and F) rabbit ventricles (Protocol VII).

A exhibited electromechanically discordant phase 2 alternans after isoproterenol. Despite the positive Ca_i–APD₈₀ coupling, Ca_i and APD₂₅ became negatively coupled, i.e. larger Ca_i transient (beat 1) shortened APD₂₅ while the smaller Ca_i transient (beat 2) corresponded to longer APD₂₅. Notice a more triangular AP (smaller APD₂₅/APD₈₀) in beat 1. B, apamin eliminated APD₂₅ alternans and Ca_i–APD₂₅ negative coupling (ΔAPD_{25,beat1−beat2}, from −7 ms to 0 ms). In addition, apamin did not eliminate but markedly attenuated APD₈₀ alternans (ΔAPD_{80,beat1−beat2}, from 21 ms to 16 ms). C, apamin prolonged APD₂₅ of beat 1 more prominently than that of beat 2. The prolongation of APD₈₀ was similar between beat 1 and beat 2. D, apamin eliminated APD₂₅ alternans (ΔAPD_{25,beat1−beat2}) and attenuated APD₈₀ alternans (ΔAPD_{80,beat1−beat2}) both significantly (by Student's paired t test). E and F, in male rabbits, isoproterenol induced Ca_i alternans at PCL 150 ms without prominent APD alternans. Both Ca_i transient and V _m had little response to apamin. L, larger Ca_i transient or longer APD; S, smaller Ca_i or shorter APD. [Color figure can be viewed at http://wileyonlinelibrary.com]

Antiarrhythmic effects of I _KAS blockade during isoproterenol infusion

To further assess the antiarrhythmic effects of I _KAS blockade during sympathetic stimulation, we compared VF vulnerabilities and characteristics between females and males (Protocol VII, Fig. 14). None of these ventricles developed spontaneous VF. Ventricular fast pacing was therefore performed to induce VF. As shown in Fig. 14 A, VF inducibility (the number of induced VF episode) was similar between sexes at baseline, during isoproterenol and after washout. However, VF inducibility after apamin became significantly lower in females than in males. In addition, the first 100 ms at the onset of each episode of VF was optically captured. As shown in the phase map of VF (Fig. 14 B), isoproterenol markedly increased the number of PSs in both groups compared with baseline. Apamin prominently decreased the number of PSs in females, but less prominently in males. As summarized in Fig. 14 C, the numbers of PSs were similar between sexes at baseline, during isoproterenol and after washout. However, apamin elicited significant lower PS numbers in females than in males. The dominant frequencies were also similar between sexes at baseline. Isoproterenol infusion led to significantly higher dominant frequencies in females than in males. The differences were eliminated by apamin (Fig. 14 D). These results indicate that I _KAS activation plays important roles in proarrhythmia and VF dynamics during sympathetic stimulation, especially in females.

Figure 14. Effects of I _KAS blockade on VF inducibility and VF dynamics.

A, the numbers of induced VF episodes were similar between females and males at baseline, during isoproterenol and after washout. After apamin, however, the numbers of induced VF episodes became significantly lower in females than in males. VF inductions were attempted 10 times at baseline and after each treatment. B, representative phase maps of the first 100 ms of VF after the onset at baseline, after isoproterenol, after apamin and after washout in females (upper panel) and males (bottom panel) (Protocol VII). Compared with baseline, isoproterenol markedly increased the numbers of phase singularities (PSs, black arrowhead) in both groups. Apamin prominently decreased PSs in females, but the effect was less prominent in males. After washout, the numbers of PSs in all three groups were similar to those at baseline. C, the numbers of PSs were similar between sexes at baseline, after isoproterenol and after washout. After apamin, however, the numbers of PSs became significantly lower in females than in males. D, VF dominant frequencies were similar between sexes at baseline. After isoproterenol, dominant frequency became significantly higher in females than in males. Apamin eliminated the isoproterenol‐induced differences in dominant frequency between sexes. After washout, the dominant frequencies remained similar in females and males. *P < 0.05 by two‐way ANOVA with Sidak's post hoc test. [Color figure can be viewed at http://wileyonlinelibrary.com]

Discussion

We discovered that β‐adrenergic stimulation activates ventricular I _KAS. Because sympathetic tone varies throughout the day, the sympathetic sensitivities indicate that I _KAS plays a role in modulating ventricular repolarization of normal ventricles. This new finding suggests that I _KAS is physiologically important. In addition, sympathetic stimulation unmasks the different I _KAS activation profiles between females and males. I _KAS activation promotes AP triangulation and ventricular arrhythmias during β‐adrenergic stimulation in females while I _KAS blockade is antiarrhythmic. Because of the importance of both sympathetic tone and sex differences in cardiac arrhythmogenesis, these findings may be clinically important.

I _KAS and sympathetic stimulation

Even in patients without apparent heart diseases, abnormal adrenergic regulation increases the susceptibility to ventricular arrhythmias (Jouven et al. 2005). β‐Adrenergic activation is known to regulate cardiac repolarization directly and indirectly via a number of currents including I _Ca,L (Reuter, 1983), I _KATP (Maruyama et al. 2014), I _Ks (Zhu et al. 2013; Aflaki et al. 2014; Bartos et al. 2017), and possibly I _Kr (Karle et al. 2002; Harmati et al. 2011) and I _K1 (Fauconnier et al. 2005; Banyasz et al. 2014). Although increased I _Ca,L tends to prolong APD, isoproterenol results in net shortening of APD by activating multiple compensatory outward currents at different phases of repolarization (Harmati et al. 2011; Zhu et al. 2013; Aflaki et al. 2014; Banyasz et al. 2014; Maruyama et al. 2014; Bartos et al. 2017). In addition, as observed in this study and by others (Banyasz et al. 2014), isoproterenol also gave rise to prominent AP morphology changes. While guinea pigs basally exhibited I _Kr > I _K1 > I _Ks, isoproterenol shaped the AP morphology by reversing the dominance pattern to I _Ks > I _K1 > I _Kr (Banyasz et al. 2014). However, the isoproterenol‐induced AP triangulation, i.e. preferentially shortening at phase 2 rather than phase 3 repolarization, is unlikely to rely on I _Kr, I _K1 or I _Ks because these voltage‐dependent currents exhibit limited activation at early repolarization phases under adrenergic stimulation. As an example, in the absence of adrenergic stimulation, I _Ks activates slowly during phase 1 and 2 repolarization and reaches its peak at the end of phase 2 (around 0 mV), followed by a rapid decline at phase 3 repolarization. However, in the presence of adrenergic stimulation, the peak of I _Ks is postponed from mid plateau to phase 3 repolarization (Banyasz et al. 2014). This was supported by the result that I _Ks blockade failed to reverse isoproterenol‐induced AP triangulation in this study (Figs 5 and 7).

Although weak under basal conditions, I _KAS in females was markedly amplified during β‐adrenergic stimulation. The I _KAS activation happens robustly at the early repolarization phases when I _Ca,L activity, sarcoplasmic Ca²⁺ release and Ca_i are all high, resulting in AP triangulation due to more prominent shortening of APD₂₅ than APD₈₀. I _KAS blockade preferentially prolonged phase 2 repolarization to restore the AP plateau and, surprisingly, even slightly shortened phase 3 repolarization, resulting in a squared AP morphology. The AP squaring after I _KAS blockade may be attributed to the further I _Ks activation facilitated by the longer or perhaps more positive plateau, which subsequently accelerated phase 3 repolarization. These findings indicate possible synergistic action between I _KAS and I _Ks in AP repolarization during adrenergic stimulation, especially in females. However, due to the results that isoproterenol still shortened and triangulated AP with both I _KAS and I _Ks blockade in male rabbits, other mechanisms might also contribute in the AP shortening and triangulation, such as the activation of sodium–calcium exchanger (I _NCX) and cAMP‐dependent Cl⁻ current (I _CFTR‐Cl) (Perchenet et al. 2000; Zhang et al. 2001; Lin et al. 2006).

Sex differences in SK2 channels

Sex differences are widely present in a variety of cardiac ion currents and exert large influences on cardiac electrophysiological properties and arrhythmogenesis (Liu et al. 1998; Vizgirda et al. 2002; Sims et al. 2008; Zhu et al. 2013). In this study, we found that I _KAS also exhibited sex differences but they were only unmasked during isoproterenol infusion. The activation of I _KAS could be attributed to the increased currents conducted by either SK2 or SK3, or both. Our experiments with the selective SK2 blocker Lei‐Dab7 suggest that SK2, rather than SK3, is the dominant isoform mediating the enhanced I _KAS during sympathetic stimulation. SK2 channel protein had significantly higher expression in females than in males, thus contributing to the majority of the sex differences in I _KAS activation. However, it is known that SK2 and SK3 proteins can form heteromeric channels in atria (Tuteja et al. 2010; Hancock et al. 2015). Therefore, we cannot exclude the possibility that the SK2‐specific blocker also inhibits I _KAS conducted by the heteromerically assembled SK2–SK3 channels in ventricles.

The sex differences in I _KAS were also verified by voltage clamp studies showing that outward I _KAS densities were larger in isolated ventricular myocytes from females than males with 1 μmol L⁻¹ intracellular Ca²⁺ at basal condition or in the presence of isoproterenol. Western blot and immunostaining suggested that SK2 protein expression was higher in females than males. However, similar I _KAS densities between sexes were detected at potentials ≤0 mV. Since SK2 channels are known for their versatile function and plasticity, the significant increase in outward I _KAS occurring only at potentials >0 mV suggests the mechanism is unlikely to simply result from differences in SK2 protein expression. The sex‐specific differences in outward I _KAS could be due to differential sensitivities of SK channels to voltage‐dependent block by internal Mg²⁺ or other ions. Another explanation is due to the fixed intracellular Ca²⁺ in the patch clamp study. In intact cardiomyocytes, Ca²⁺ influx and subsequent release through LTCC and ryanodine receptor (RyR) might increase the local subsarcolemmal Ca²⁺ concentration up to 100 μmol L⁻¹, while intracellular Ca²⁺ concentration might only reach 1 μmol L⁻¹. Since I _KAS is very sensitive to local Ca²⁺ between the SK channel and LTCC/RyR (Zhang et al. 2018), it is possible that I _KAS recorded at a fixed global Ca²⁺ concentration was not high enough to differentially activate I _KAS between sexes. Further studies with dynamical Ca²⁺ fluctuations are needed.

The Ca²⁺ sensitivity of SK2 channels is strongly modulated by CK2 in neurons (Adelman et al. 2012). Heart failure decreases the interaction between CK2 and SK2 and consequently enhances the sensitivity of I _KAS to Ca²⁺ (Yang et al. 2015). Consistent with those studies, our results further suggested the potential participation of CK2 inhibition in cardiac I _KAS activation. While male rabbit ventricles expressed lower SK2 channels than those of females, the equally expressed CK2 between sexes resulted in a higher CK2/SK2 ratio in males. While these results might in part contribute to the findings of the present study, more work will be needed to prove a causal relationship between the CK2/SK2 ratio and the sex differences of I _KAS densities.

Notably, the function of SK channels not only relies on the total protein expression, but also depends critically on the subcellular distribution and the cell‐surface membrane expression. Interacting proteins, such as α‐actinin and filamin A, regulate SK channel anterograde trafficking to the surface membrane and recycling (Lu et al. 2009; Rafizadeh et al. 2014; Zhang et al. 2017). It is possible that differential channel trafficking properties exist between sexes and contribute to different I _KAS responses to sympathetic stimulation.

Antiarrhythmic effects of I _KAS blockade

AP triangulation, with either APD abbreviation or APD prolongation, predicts serious proarrhythmia (Hondeghem et al. 2001). Agents that prolong phase 3 repolarization (triangulation) are proarrhythmic, while agents that prolong phase 2 without slowing phase 3 (squaring) are generally antiarrhythmic (Hondeghem et al. 2001). Thus, it is important to develop repolarization‐delaying agents that lengthen APD in a safe fashion. In females, I _KAS blockade prolongs phase 2 repolarization without prolonging phase 3 repolarization, reversing AP triangulation and leading to AP squaring, thereby predicting a potential antiarrhythmic effect. Therefore, I _KAS could be a promising target for the development of antiarrhythmic agents, especially for females.

Cardiac alternans is a precursor to VF. APD and Ca_i transient alternans are caused by instabilities in both V _m and Ca_i cycling dynamics and their coupling (Weiss et al. 2006). V _m and Ca_i are coupled via calcium‐sensitive currents, such as I _Ca,L (Weiss et al. 2006), I _NCX (Weiss et al. 2006), I _Ks (Kennedy et al. 2017) and I _KAS (Kennedy et al. 2017). In silico models in the absence of I _KAS, Ca_i and V _m exhibit positive coupling and electromechanically concordant alternans (Kennedy et al. 2017). After introducing I _KAS with a relatively low calcium affinity, APD is shortened when the Ca_i transient is large, resulting in negative Ca_i–V _m coupling and electromechanically discordant alternans (Kennedy et al. 2017). In the present study, we have experimentally validated this prediction by demonstrating that I _KAS activation by isoproterenol caused Ca_i–V _m coupling to shift from positive to negative, resulting in electromechanically discordant alternans, especially at phase 2 repolarization. I _KAS blockade attenuated negative Ca_i–V _m coupling and phase 2 repolarization alternans. In addition, I _KAS blockade reduced the VF vulnerability and altered VF dynamics by reducing the number of PSs and dominant frequency, also suggesting its potential antiarrhythmic effects in females.

Study limitations

The optical mapping was performed only on the epicardium. These findings may not be applicable to mid and endocardium due to the transmurally heterogeneous distribution of I _KAS (Yu et al. 2015). Due to the unavailability of normal human cardiac tissue, the sex differences of I _KAS in human ventricles remain unknown. However, the rabbit is considered as a good animal model for studying sex differences in cardiac ion currents relevant to humans (Salama & Bett, 2014). Isoproterenol also induces AP triangulation in male rabbit ventricles, but apamin failed to normalize the AP morphology. The mechanism of AP triangulation in males remains unclear.

Conclusions and implications

A variety of cardiac electrical diseases exhibit sexual dimorphism (Salama & Bett, 2014; Antzelevitch et al. 2016). Understanding the arrhythmogenic mechanisms may therefore be important for generating sex‐specific therapies. Although ventricular I _KAS is relatively small, it still plays important roles in regulating APs under a variety of physiological and pathological conditions. In the present study, we show that adrenergic stimulation causes greater activation of I _KAS in females than males. The sex differences of I _KAS are attributable to the differences in the expression and biophysical properties of SK2 channels and the CK2/SK2 ratios. Activation of I _KAS is more abundant during phase 2 than during phase 3 repolarization, leading to AP triangulation, negative Ca_i–V _m coupling, electromechanically discordant phase 2 alternans and increased VF vulnerability. I _KAS blockade is potentially antiarrhythmic in females. Drugs specifically targeting cardiac I _KAS should be effective and safe. Amiodarone, a commonly used antiarrhythmic drug, is known to suppress I _KAS (Turker et al. 2013). More specific I _KAS blockers have also been tested in animal models (Diness et al. 2017; Ko et al. 2018). I _KAS blockers may prove to be a new class of antiarrhythmic drugs with sex‐specific clinical applications.

Additional information

Competing interests

None

Author contributions

M.C.: design, data acquisition and analysis/interpretation; drafting and revision. D.Y., S.G., D.Z.X. and Z.W.: data acquisition and analysis. Z.C., M.R.L., S.F.L., T.H.E. and J.N.W.: design, interpretation and revision. P.S.C.: conception, design, interpretation, drafting and revision. All authors approved the final version of the manuscript. All authors agree to be accountable for the data integrity and ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.

Funding

This work was supported by NIH R01HL139829, P01HL78931, R56HL71140, R42DA043391 and TR0022080, a Charles Fisch Cardiovascular Research Award endowed by Dr Suzanne B. Knoebel of the Krannert Institute of Cardiology, a Medtronic‐Zipes Endowment and the Indiana University Health–Indiana University School of Medicine Strategic Research Initiative.

Biography

Mu Chen received his MD from Shanghai Jiao Tong University and did resident training at Xinhua Hospital Shanghai. He now works as a Postdoctoral Fellow under the supervision of Dr Peng‐Sheng Chen in Krannert Institute of Cardiology, Indiana University. His studies focus on the cardiac small conductance calcium‐activated potassium channel and its regulation on ventricular repolarization and arrhythmias.