Pflugers Arch - Eur J Physiol DOI 10.1007/s00424-015-1769-2

ION CHANNELS, RECEPTORS AND TRANSPORTERS

Spontaneous inward currents reflecting oscillatory activation of Na+/Ca2+ exchangers in human embryonic stem cell-derived cardiomyocytes Seong Woo Choi 1 & Hyang-Ae Lee 1,3 & Sung-Hwan Moon 4 & Soon-Jung Park 4 & Hae Jin Kim 1,2 & Ki-Suk Kim 1,3 & Yin Hua Zhang 1,2 & Jae Boum Youm 5 & Sung Joon Kim 1,2

Received: 6 September 2015 / Revised: 30 November 2015 / Accepted: 3 December 2015 # Springer-Verlag Berlin Heidelberg 2015

Abstract Na+/Ca2+ exchanger current (INCX) triggered by spontaneous Ca2+ release from sarcoplasmic reticulum (SR) has been suggested as one of the cardiac pacemaker mechanisms (BCa2+ clock model^). In human embryonic stem cellderived cardiomyocytes (hESC-CMs) showing spontaneous action potentials (APs), we found that substantial population (35 %) showed regular oscillation of inward currents (SICs) in nystatin-perforated voltage clamp between −40 and 40 mV (−80 ± 10.6 pA, at −20 mV). SICs were similarly observed between nodal, atrial, and ventricular hESC-CMs. Oscillations of [Ca2+]i synchronized with SICs were observed under voltage clamp. SICs were eliminated by lowering [Ca2+]e, L-type Ca2+ channel (VOCCL) blocker (nifedipine, 10 µM), ryanodine receptor (RyR) agonist (caffeine, 10 mM), or NCX inhibitor (1 µM SN-6 and 10 µM KB-R7943). Plasma membrane expression of NCX1 was confirmed using immunofluorescence

* Jae Boum Youm [email protected] * Sung Joon Kim [email protected] 1

Department of Physiology, Department of Biomedical Sciences, College of Medicine, Seoul National University, Seoul, Republic of Korea

2

Ischemic/Hypoxic Disease Institute, College of Medicine, Seoul National University, Seoul, Republic of Korea

3

Next-generation Pharmaceutical Research Center, Korea Institute of Toxicology, Daejeon, Republic of Korea

4

Department of Stem Cell Biology, School of Medicine, Konkuk University, Seoul, Republic of Korea

5

Department of Physiology, College of Medicine, Inje University, Busan, Republic of Korea

confocal microcopy. Both caffeine and SN-6 slowed the pacemaker potential but did not abolish the AP generation. The inhibitors of funny current (3 µM ivabradine) or voltagegated K + channel currents (1 µM E4031 and 10 µM chromanol-293B) also did not abolish but slowed the pacemaker potential. In a computational model of cardiac pacemaker by Maltsev and Lakatta (2009), after modifying the spatial distribution of RyR, VOCCL, and NCX by using our multiparameter adjust algorithm, we could successfully reproduce spontaneous SR Ca2+ release and SICs under voltage clamp. It was proposed that, under the membrane depolarization activating VOCCL, oscillatory Ca2+ releases via RyR induce sharp increases in subsarcolemmal [Ca2+]i and inward INCX (SICs). Since the hESC-CMs without SICs still showed spontaneous APs, the putative BCa2+ clock^ would provide a redundant pacemaker or augmenting mechanism in hESC-CMs. Keywords Embryonic stem cell . Heart . Pacemaker current . NCX . Ca2+ release

Introduction Human embryonic stem cells (hESCs) have a potency to differentiate into various types of cells including cardiomyocytes. The hESC-derived cardiomyocytes (hESC-CMs) could provide multiple applications: in vitro model for studying human heart development, platform for the cardiotoxicity test in drug development, applications for regenerative medicine, and model for basic physiological research [20]. A cardinal phenotype of hESC-CM is the spontaneous beatings and repetitive generation of action potentials (APs). The shapes of AP and the levels of diastolic potential (resting membrane potential) in hESC-CMs are heterogeneous; the duration and shapes of

Pflugers Arch - Eur J Physiol

APs reflect the nodal-, atrial-, and ventricular-like properties [8, 18]. Voltage clamp studies have revealed that hESC-CMs have comparable expression of many excitable channels as mature cardiac cells [8, 19, 25, 27]. In addition, the molecules and functional properties of Ca2+ handling mechanisms in hESCCMs are being understood; Ca2+ release channels (ryanodine receptor (RyR), InsP3 receptor (IP3R)) and Ca2+ pump in sarcoplasmic reticulum membrane (SERCA) as well as the Ca2+ removal mechanisms in the plasma membrane such as Na+/ Ca2+ exchanger (NCX) [13, 17, 27]. The spontaneous AP of hESC-CMs has been reported to be more significant in their early phase of culture time (<4 weeks) [7, 10, 25]. Although the electrical automaticity is useful as a phenotype of cardiac differentiation of hESCs, their electrical heterogeneity potentially causes arrhythmia when transplanted into hearts [15, 34], which lead to the trials of improving maturation status to enhance the therapeutic benefits [31]. On the other hand, the robust spontaneous APs of hESC-CM draw the attention of physiologists and pharmacologists, in expectation of providing insight of cardiac pacemaker mechanisms of the human heart and testing the cardiotoxicity of novel drug candidates. Timely interaction between voltage-gated K+ channel (Kv) deactivation, background depolarizing conductance, and hyperpolarization-activated cyclic nucleotidedependent cation channel (HCN) current (funny current, If) comprise the conventional mechanism of cardiac pacemaker, called the Bmembrane clock^ mechanism. In addition, recent studies by Lakatta’s group have provided evidences that spontaneous local Ca2+ release (LCR) from sarcoplasmic reticulum (SR) accompanies the ion channel-dependent pacemaker mechanism. According to the BCa2+ clock^ model, the LCR induces forward-mode activation of NCX that generates an inward current (INCX) for the pacemaker potential. Experimental data supporting the role of Ca2+ clock mechanism have been shown in the sinoatrial node cells (SANCs) from various species including rabbit, mouse, dog, and toad, while there is no data from human SANCs yet [4, 6, 9, 12, 32]. In our previous study, successful differentiation of hESCCM and various types of APs consistent with nodal, atrial, and ventricular ones have been confirmed under zero current clamp conditions [18]. In the subsequent pilot experiment of voltage clamp analysis, we observed oscillation of spontaneous inward currents (SICs) that could not be simply overlooked or regarded as an artifact. Since the phenomenon like SICs might be associated with the pacemaker depolarization, we were tempted to investigate the mechanisms and physiological implications in hESC-CMs. Mathematical modeling for the computational simulation is highly useful to understand the complex nature of cardiac electrophysiology and has provided insights of

pacemaker mechanisms as well. However, even in the recent model by Lakatta’s group that incorporates the Ca2+ clock mechanism [16], no spontaneous oscillation of membrane current was reproduced under the virtual voltage clamp condition. In the present study, through modifications of the parameters for Ca2+ handling mechanisms, we attempted to simulate SICs based on the MaltsevLakatta (M&L) model (Fig. 1).

Materials and methods Preparation of hESC-CMs hESCs H9 (WiCell, Madison, WI) were grown on inactivated MEF in Dulbecco’s modified Eagle’s medium (DMEM) F/12 (50:50 %; Gibco BRL, Gaithersburg, MD) supplemented with 20 % serum replacement (Gibco) and 1 % nonessential amino acids (Gibco), 1 mM L glutamine (Gibco), 100 mM betamercaptoethanol (Gibco), and 4 ng/mL basic fibroblast growth factor (FGF) (R&D System, Inc. Minneapolis, MN). Medium was changed daily, and hESCs were transferred to fresh feeder cells every 6∼7 days with dissecting pipettes. To induce embryoid body formation, hESCs were detached from feeder cells with dispase (Gibco) then transferred to ultra-low attachment plates for suspension in basic FGF-free and BMP4 20 ng/mL in DMEM F/12 with 10 % SR for 2 days. The primary EBs were cultured in DMEM supplemented with 5 % fetal bovine serum (FBS) (Hyclone) for 12–14 days to differentiate into contracting EBs, then plated on 0.1 % gelatin-coated plates DMEM-high glucoses supplemented 2 % FBS for 8 days. Beating clusters were isolated and dissociated to a single-cell suspension by 0.25 % trypsin-EDTA treatment for 5 to 10 min at 37 °C. Subsequently, the cells were replated on a dish that was coated with a mixture of Matrigel and 0.1 % gelatin solution at a ratio of 1:100, respectively, in low glucose DMEM supplemented with 2 % FBS for 10 days. Contracting cardiomyocytes were cultured at 37 °C and 5 % CO2 in a humidified incubator, and media was refreshed every 3 to 4 days. Preparation of human-induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) The hiPSC-CMs used in this study were purchased from Cellular Dynamics International, Inc. (CDI) (iCell Cardiomyocytes, CDI, Madison, WI, USA). These cells were thawed and maintained according to the protocols recommended by the supplier. Briefly, hiPSC-CMs were thawed at 37 °C and resuspended in plating medium (iCell Cardiomyocyte Plating Medium), and then plated onto 0.1 % gelatin (Sigma-Aldrich)-coated glass coverslips in four-well culture plates. After 2 days of culture at 37 °C and 7 % CO2, the plating medium was replaced with

Pflugers Arch - Eur J Physiol

Electrophysiology

Fig. 1 Schematic illustration of cell compartments, ion channels, and transporters in our model of hESC-CMs. a Main architecture of our model is the same as the model of rabbit sinoatrial node cells (SANC) developed by Maltsev and Lakatta (2009) except that our model has additional population of VOCCL and NCX which are localized on the membrane facing bulk cytosol. Ca2+ in the subsarcolemmal space (Casub) oscillates by concerted actions of RyR in SR membrane and Ca2+permeable proteins such as VOCCL, NCX, and background-type Ca2+ channels in the membrane facing subsarcolemmal space. Casub diffuses into the cytosolic space and is pumped back to SR by network SR Ca2+ pump (SERCA) or pumped out by NCX. Finally, Ca2+ in the network SR diffuses into junctional SR to complete Ca2+ cycle. Cai, Ca2+ in bulk cytosol; IKr, rapid component of delayed rectifier current; IKs, slow component of delayed rectifier current; Ito, 4-aminopyridine sensitive transient outward K+ current; Isus, 4-aminopyridine sensitive sustained K+ current; Ist, sustained inward current; INCX, NCX current; INaK, Na+/ K+ pump current; ICaT, T-type Ca2+ current; ICaL, L-type Ca2+ current; If, hyperpolarization-activated current; IbNa, background Na+ current

culture medium (iCell Cardiomyocyte Maintenance Medium). Cell culture maintenance media were changed three times a week for 1 month.

Immunocytochemistry Cardiomyocytes were fixed in 4 % paraformaldehyde for 20 min at 4 °C and permeabilized with 0.1 % Triton X100 in phosphate-buffered saline (PBS; hyclone) for 5 min. After treatment with 5 % normal goat serum for 30 min at RT, the cells were incubated with primary antibodies that recognize lineage-specific markers for 12 h at 4 °C, such as the cardiac lineage markers cardiac-specific troponin T (cTnT, Abcam), sarcomeric myosin heavy chain (sMHC; Millipore, Billerica, Mass., USA), NCX1 (Abcam, ab135735). The cells were washed three times with 0.03 % Triton X-100 in PBS and then incubated with Rhodamine- and FITC-conjugated secondary antibodies (Molecular Probes Inc., Eugene, Oreg., USA) for 2 h, after three washes with PBS and staining with DAPI. The slides were mounted with a glycerol-based mounting solution containing 2.5 % polyvinyl alcohol (Vector). All images were analyzed with an LSM 510 META confocal microscope (Carl Zeiss Inc., Oberkochen, Germany).

Every voltage clamp and current clamp experiments were carried out using nystatin-perforated whole-cell configuration except for analysis of action potential characterization (Figs. 2 and 3). hESC-CMs and iPSC-CMs were perfused at physiological temperature (37 °C) with normal Tyrode’s solution containing 145 mM NaCl, 4.3 mM KCl, 1 mM MgCl2, 1.8 mM CaCl2, 10 mM HEPES, and 5 mM glucose, adjusted to pH 7.4 with NaOH. Intracellular pipette solution contained 120 mM K-aspartate, 20 mM KCl, 10 mM HEPES, 0.1 mM EGTA, 1 mM MgCl2, 3 mM Mg-ATP, adjusted pH 7.25 with KOH. Nystatin was dissolved in DMSO at a high concentration (60∼70 mg/ml). The nystatin stock was sonicated and used only at the day of work. Final concentration of 60∼70 g/ml was applied through pipette solution to perforate cell membrane. Caffeine, ivabradine, isoproterenol, E4031, chromanol-293B, and nifedipine were purchased from Sigma. SN-6 and KB-R7943 (Tocris Bioscience) were applied to inhibit NCX. Micro glass patch pipettes (World Precision Instruments, Sarasota, FL, USA) were pulled by PP-830 puller (Narishige, Tokyo, Japan) with resistance between 2.5∼3.0 M. Axopatch 200B amplifier, Digidata 1440A, and pClamp software 10.1 (Axon Instruments, Foster, CA, USA) were used for recording and analysis. Simultaneous measurement of intracellular Ca2+ with patch clamp hESC-CMs were loaded with 5 µM of fura-2AM and 2 µM of Pluronic F-127 in the 5 % CO2 incubator for 30 min. The fura2 loaded cells were transferred into a microscope bath and perfused with NT. Spontaneously beating hESC-CMs were selected for patch clamp recording and Ca2+ measurement. Fluorescence was monitored with fluorescence spectrophotometer (DeltaRAM V, Photon Technology International, Edison, NJ, USA) at excitation wavelengths of 340/380 nm and emission wavelength of 510 nm. To synchronize the fluorescence Ca2+ recording and patch clamp recording, a TTL output signal was transferred from FeliX32 ™ software/ BryteBox™ interface (PTI) to START input of Digidata 1440A. Mathematical modeling A modified cell model of M&L model [16] was employed to reproduce SICs observed in hESC-CMs. The original model did not show any oscillation of membrane current or [Ca2+]i under virtual voltage clamp conditions (see Results, Fig. 10a). In order to make Ca2+ released from the SR oscillate in voltage clamp mode, following modifications from the original M&L model were applied.

Pflugers Arch - Eur J Physiol Fig. 2 Heterogeneity of AP morphology in spontaneously beating hESC-CMs. a Spontaneously beating hESCCMs dissociated from contracting EB clusters were plated on a dish for the whole-cell patch clamp. b Expressions of cardiomyocyte markers, cTnT and sMHC. c Beating hESC-CMs show heterogeneous AP morphology (nodal, atrial, and ventricular types). d, e Heterogeneous distributions of maximal diastolic potential (MDP) and AP duration (APD50)

The volume of cytoplasm between the pairs of plasma membrane and junctional SR is named as subsarcolemmal space, functionally separated from the bulk cytosolic space of virtual cardiomyocytes (Fig. 1). In the original M&L model, L-type Ca2+ channels (VOCCL) and NCX were assumed to be exclusively located on the plasma membrane facing the junctional SR fraction. Since the hESC-CMs investigated in the present study reflected an early period of cardiac development, here, we assumed that VOCCL and NCX are relatively evenly distributed along the whole membrane considering the previous studies of neonatal (embryonic) cardiomyocytes [2, 13]. Then, we divided population of channels or exchangers into groups A and B; VOCC L or NCX facing the subsarcolemmal space belong to group A, and the others facing the cytosolic space belong to group B. Those two groups contribute to ion flux in their own space. They are also activated or inactivated by Ca2+ in their own space. For example, if we set the fraction of group A in VOCCL to f, the fraction of them in group B is 1 − f. Therefore, the following modifications were applied to the equations in the original M&L model: I CaL ðAÞ ¼ f *g CaL ⋅d L ⋅ f L ⋅ f Ca ⋅ðV −E CaL Þ I CaL ðBÞ ¼ ð1− f Þ*g CaL ⋅d L ⋅ f L ⋅ f Ca ⋅ðV −E CaL Þ

where gCaL is maximum conductance of VOCCL, d L is activation variable, f L is voltage-dependent inactivation variable, f Ca is Ca2+-dependent inactivation variable, V is membrane voltage, and ECaL is apparent reversal potential of VOCCL. As f Ca is dependent on local Ca 2+ concentration in the subsarcolemmal (group A) or bulk cytosolic space (group B), additional modifications were applied as follows: K mfCa  
K mfCa þ Ca2þ A K mfCa  
f Ca;∞ ðBÞ ¼ K mfCa þ Ca2þ B f Ca;∞ ðAÞ ¼

where f Ca∞ is steady-state inactivation variable, and K mfCa is   dissociation constant of Ca2+-dependent inactivation. Ca2þ  2þ  2+ A and Ca B is [Ca ] in the subsarcolemmal space (group 2+ A) and [Ca ] in the bulk cytosolic space (group B), respectively. The same modifications were applied to the model of NCX. Ca2+ uptake rate by SR and Ca2+ release rate by RyR were also modified from the original M&L model. The Ca2+ uptake rate between 0.01 and 0.2 mM ms−1 and Ca2+ release rate constant between 0.2 × 103 and 100 × 103 ms−1 were tested. In addition, the Kci, a parameter for intracellular Ca2+ binding to NCX, was modified to 0.00621 from the original

Pflugers Arch - Eur J Physiol

Fig. 3 Spontaneous inward current (SIC) observation and voltagedependency in conventional and perforated whole-cell voltage clamp. a–c Different oscillatory patterns of SICs (a no oscillation; b small amplitude of current oscillation; c larger and sustained current oscillation) under −20 mV voltage clamp condition. d, e Observation of SICs under conventional whole-cell (CWC) mode patch (17/56) and perforated whole-cell (PWC) mode patch (23/46). Closed squares and

circles indicates APs having SIC activity (closed square: CWC with SIC and closed circle: PWC with SIC). f SICs under various ranges of voltage-clamped condition and its frequency changes (1.24 ± 0.36, 2.37 ± 0.54, 2.65 ± 0.56, 2.76 ± 0.6, and 2.97 ± 0.59 Hz; −40, −20, 0, 20, and 40 mV, respectively). g Normalized frequency of SICs by −20 mV. Frequency of SICs at −40 mV was significantly slower than AP frequency (1.98 ± 0.36 Hz, n = 5).

value of 0.0207 to allow more stable SR Ca2+ oscillation. Finally, since the original M&L model is for the SA nodal cells, we included the voltage-operated Na+ channel current (INa) to be more consistent with the properties of atrial or ventricular cell types of hESC-CMs. The model of INa is from Kurata et al. [11], and its maximum conductance was set to 0.1 nS pF−1.

showed positive signals for cTnT and sMHC, markers of cardiomyocytes (Fig. 2b). Spontaneously contracting and isolated cells are selected for membrane voltage recording under the zero current clamp mode of conventional whole-cell patch clamp experiment. Representative cases of nodal, atrial, and ventricular type APs are displayed (Fig. 2c). However, the distributions of AP duration at the half of the total amplitude (APD50) and maximal diastolic potential (MDP) revealed that the boundaries between the three cell types were not clear (Fig. 2d, e). In general, hESC-CM with more hyperpolarized MDP showed higher amplitude of AP (Fig. 2d). The higher AP amplitude seemed to be also related with longer APD50 (Fig. 2e).

Statistical analysis Data was managed and analyzed using Origin 8.0 software (Microcal, Northampton, MA, USA). Statistical results are expressed as mean ± SEM. Student’s t test and ANOVA was used to test for significance at the level of <0.05.

SICs and their mechanism in hESC-CMs

Results The hESC-CMs used for patch clamping were cultured in dispersed state that enabled reliable voltage clamp without electrical input from adjacent cells (Fig. 2a). Spontaneous contractile beatings were observed under the microscopic view of experimental bath. Immunocytochemical staining

In the conventional whole-cell configuration, converting from zero current clamp to voltage clamp mode (e.g., −20 mV of holding voltage) showed three types of membrane current recordings: (1) sustained holding currents with noisy fluctuations (39/56 of tested cells, Fig. 3a), (2) decaying small oscillations of inward currents (9/56, Fig. 3b), and (3) regular oscillations of inward currents (8/56, Fig. 3c). With nystatin-

Pflugers Arch - Eur J Physiol

perforated whole-cell clamp configuration, the proportion of cells showing the regular SICs was significantly higher (23/ 56) than the conventional whole-cell clamp (17/56, Fig. 3d). The SICs appeared almost immediately on converting to the voltage clamp mode (Fig. 3c), and their amplitudes and frequencies were relatively stable during the recording time up to 20 min. There were no clear correlations between the AP amplitudes, MDP, and APD50 with SIC generation (Fig. 3e). SICs were observed at the clamp voltages ranging from −40 to 40 mV (Fig. 3f). At 40 mV, however, the regular SICs were not sustainable. At −60 mV, SICs disappeared after a few times of oscillation (data not shown). The frequency of SICs generally became higher by depolarization up to 0 mV (n = 7). The frequency of SICs at −40 mV was 1.24 ± 0.36 Hz that was significantly slower than the frequency of APs (1.98 ± 0.36 Hz, n = 5). We tested whether changes in [Ca2+]i is associated with the generation of SICs. In fura-2 loaded hESC-CMs, nystatinperforated patch clamp was applied. Not only the APs but also the SICs were observed with synchronized increases in [Ca2+]i (Fig. 4a). The repetitive generation of SICs and [Ca2+]i transients under voltage clamp suggested an involvement of cyclic Ca2+ release from SR. Consistently, SICs were inhibited by treatment with 10 mM caffeine that sensitize RyR, resulting to depletion of SR Ca2+ store (Fig. 4b). The frequency of SIC was transiently increased on applying caffeine, and eventually disappeared (Fig. 4b, d). The initial increase in the frequency is thought to reflect the sensitization of RyRs by caffeine, a known pharmacological mechanism of caffeine [1]. It was supposed that the amount of SR Ca2+ store would decrease gradually due to the excessive activation of RyRs by 10 mM caffeine. The SICs were also abolished by reducing [Ca2+]e to 0.18 mM (Fig. 4c). The disappearance of SICs took almost a minute after applying caffeine or lowering [Ca2+]e, which was almost completely reversible (Fig. 4d). Pharmacological inhibition of VOCCL (nifedipine, 10 µM) also inhibited SICs (Fig. 4e, n = 3). Beta-adrenergic stimulation is known to enhance the Ca2+ uptake and CICR processes in SR as well as VOCCL in the plasma membrane of cardiac myocytes, resulting to inotropic effects [28]. Interestingly, bath application of isoproterenol could sometimes induce SICs in the hESC-CMs that initially showed no SIC (Fig. 4f, n = 2 of 3 trials). The persistent SICs at positive clamp voltage (e.g., 20 mV) suggested that, rather than the passive inward current through Ca2+-activated Na+ or Cl− channels, an electrogenic transport mechanism coupled with [Ca2+]i transient (e.g., NCX) might be involved. Consistent with this idea, NCX inhibitor (1 M SN-6) effectively abolished SICs (Fig. 5a, n = 4). The fluctuation of small membrane current was still observed after SN-6 application, probably because of insufficient inhibitory concentration (IC50 for forward NCX, 1.9–2.3 µM) [21]. Another NCX inhibitor, 10 µM of KB-R7943, also inhibited SICs

(Fig. 5b, n = 3). Furthermore, substitution of extracellular Na+ with Li+ abolished SICs (Fig. 5c, n = 3). The surface expression of cardiac type of NCX1 was confirmed in hESCCMs using immunofluorescence confocal microscopy (Fig. 5d). The forward mode of NCX currents, i.e., inward INCX, might contribute to AP generation in hESC-CMs. Thus, we tested the effect of NCX inhibitor on diastolic depolarization (DD) and AP of the hESC-CMs under the zero current clamp mode. Application of 1 µM SN-6 slowed DD, decreased APD50 and decreased MDP. Application of 10 mM caffeine also decreased DD and MDP (Fig. 6). It has to be noted that neither SN-6 nor caffeine treatment abolished the spontaneous APs in hESC-CMs. Then, we tested the effects of pharmacological inhibitors for the cardiac ion channels that are known to be responsible for DD; ivabradine, E4031, and chromanol-293B for If, IKr, and IKs, respectively. The DD of hESC-CMs was slowed by 1 µM ivabradine (Fig. 7a, b). The presence of ivabradinesensitive If-like slow inward currents was confirmed by the hyperpolarizing step pulses (Fig. 7c). Pharmacological inhibition of IKr (1 µM E4031) or IKs (10 µM chromanol-293B) increased APD 50 and decreased the frequency of APs (Fig. 7d, n = 3). Then, we tested the effects of combined application of the channel inhibitors. After confirming the effects of ivabradine on DD and AP frequency, E4031 and chromanol-293B were added, which induced substantial depolarization of MDP. Interestingly, even with the three inhibitors, spontaneous fluctuations of membrane potential were persistently observed (Fig. 7e, n = 2). SICs were also observed in the commercially available human-induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs, iCell Cardiomyocytes from Cellular Dynamics International). Among the hiPSC-CMs tested for conventional whole-cell mode voltage clamp study, 9 % of the cells (12/ 135; lot nos. 1293522 and 1033176) showed SICs with large amplitudes (Fig. 8). The SICs of hiPSC-CM were also suppressed by SN-6 and KB-R7943 (Fig. 8b, c). Computational modeling of SICs In the modified M&L model, to find the conditions demonstrating spontaneous SR Ca2+ oscillations (i.e., Ca2+ releases from SR) under voltage clamp, we tested 10,000 set of combinations (100 × 100) for the parameters related to Ca2+ handling and its localization in silico. As described in the BMaterials and methods^ section, we varied either the distribution of VOCCL and NCX in the plasma membrane of M&L model, or RyR Ca2+ release and pumping rates by Ca2+-handling proteins using our own paired parameter optimization (PPO) protocol. In the process of PPO, the SR Ca2+ oscillations at −20 mV of voltage clamp condition were monitored for as long as 30 s in the simulation time scale at each

Pflugers Arch - Eur J Physiol

Fig. 4 Ca2+-dependent regulation of SICs activity. a Simultaneous recordings of electrical changes (membrane potential or current) and [Ca2+]i transient in spontaneously beating hESC-CMs. The initial recordings were under zero-current clamp mode (initial two APs), then converted to the voltage clamp (a vertical artifact of current) at −20 mVof holding potential (HP). b, c Both SR Ca2+ depletion by bath application of 10 mM caffeine and lowering external Ca2+ level (0.18 mM) abolished SIC activity at −20 and 0 mV of HP, respectively. The activity was

gradually recovered by washout. d The amplitude and frequency of SICs during caffeine (n = 4) and low Ca2+ treatment time (n = 3). Disappearance of SIC activity was taken at least >80 s of caffeine and low Ca2+ application. Among reducing of SICs, the frequency of SICs was increased transiently. e The inhibitor of VOCCL, 10 µM nifedipine also abolished SICs at 0 mV HP (n = 3). f Beta-adrenergic agonist, 1 µM of isoproterenol induced SICs that initially showed no SIC (n = 2 of 3 trials)

Pflugers Arch - Eur J Physiol

Fig. 5 Evidence of NCX roles for SIC activity. a SIC activity was gradually decreased via bath application of NCX inhibitor at −20 mV HP (1 µM SN-6, n = 4). b Another inhibitor of NCX, 10 µM KB-R7943 immediately abolished SICs at −40 mV HP (n = 3). However, relatively small current fluctuation was still observed after applying SN-6, probably

because of insufficient inhibitory concentration (IC50, 1.9–2.3 µM). c Replacement of bath Na+ by Li+ abolished SICs (n = 3). d Representative confocal microscopy images shows the expressions of NCX1 (green) and sMHC (red) in single hESC-CM

parameter set, and parametric space with sustained oscillation was drawn as the color map (Fig. 9a, b). As we simultaneously

varied the Ca 2+ pumping rate by SERCA from 0.01 to 0.2 mM ms−1 and Ca2+ release rate constant by RyR from 0

Fig. 6 Changes of pacemaker potentials and AP shapes by SR Ca2+ depletion or by NCX inhibition. a, b Both 10 mM caffeine and 1 µM SN-6 slowed diastolic depolarization (DD) in hESC-CMs (dotted traces). c, d, e Summary of the changes in DD velocity (caffeine; from

59.4 ± 4.4 to 38.6 ± 5.8 mV/s, SN-6; from 42.0 ± 4.6 to 21.2 ± 4.5 mV/s), APD50 (caffeine; from 185.2 ± 15.9 to 188.2 ± 33.5 ms, SN-6; from 226.4 ± 53.1 to 172.6 ± 44.5 ms), and MDP (caffeine; from −68.9 ± 3.9 to −61.2 ± 4.1 mV, SN-6; from −63.7 ± 3.0 to −61.0 ± 3.5 mV)

Pflugers Arch - Eur J Physiol

Fig. 7 Effects of inhibitors of If, IKr, and IKs on pacemaker potentials and AP shapes. a, b Slowed DD by 1 µM of ivabradine, an inhibitor of If (n = 6). c Ivabradine-sensitive If current was recorded by hyperpolarizing step pulse from −40 to −100 mV. Ivabradine (3 µM) significantly reduced the slowly activating inward current by the membrane hyperpolarization. d, e Effects of E4031 (1 µM, IKr inhibitor) and chromanol-293B (10 µM, IKs

inhibitor); APD was increased with decreased frequency of APs. A combined application of If, IKr and IKs inhibitors markedly reduced MDP, and the AP generation became less regular. Note the increased frequency of APs at the depolarized state. Representative traces from three experiments are shown

to 106 ms−1 under the condition where the subsarcolemmal fractions of VOCCL and NCX are both fixed to 0.1, the parametric space with sustained oscillation was successfully obtained, which was in the area where the Ca2+ pumping rate is increased and Ca2+ release rate is reduced when compared with those in the original M&L model (Fig. 9a). Under the condition where the Ca2+ pumping rate is 0.1 mM ms−1 and the Ca2+ release rate constant is 2000 ms−1 (area i of Fig. 9a), we simultaneously varied the subsarcolemmal fractions of VOCCL and NCX from 0 to 1 to find another parametric space with sustained oscillations. Spontaneous SR Ca2+ oscillation was successfully induced in the restricted ranges of parametric space (Fig. 9b, VOCCL 0 to 0.8, NCX 0 to 0.5). In the zero current clamp mode, the original M&L model faithfully reproduces pacemaker potentials of SA nodal type, while the SIC-like current oscillations were not observed in the voltage clamp condition (Fig 10a). For comparison with

the original M&L model, we tested the best set of parameter combinations with sustained oscillation where the subsarcolemmal fractions of VOCCL and NCX were both set to 0.3. The Ca2+ pumping rate and Ca2+ release rate constant were 0.1 mM ms−1 and 1000 ms−1, respectively. When the best set of parameter combinations was applied to the modified M&L model, both spontaneous APs and SICs are consistently observed in the current clamp and voltage clamp modes, respectively (Fig 10b; solid line). The total amplitude, MDP, peak, and APD50 measured from AP of modified M&L model are 98.8, −63.7, 35.1 mV, and 148.3, which lie in the center of distribution of experimentally obtained APs from hESC-CMs shown in Fig. 2. The SICs from the modified M&L model are also similar to experimentally obtained SICs in that they show both slow and fast increase phases. Since the hESC-CMs also represent atrial and ventricular myocytes, we tested whether the addition of INa component still shows stable

Pflugers Arch - Eur J Physiol

Discussion

Fig. 8 Recordings of SICs and their inhibition by NCX inhibitors in hiPSC-CMs. a A representative recording of SICs by converting from zero current clamp (AP recording) to the voltage clamp at −20 mV. b, c. Application of NCX inhibitors (10 µM KB-R7943 and 10 µM SN-6) eliminated SICs. The clamp voltage (HP) was 0 mV

APs and SICs. The inclusion of INa (fraction of current in comparison with the conventional human atrium = 0.154, [24]) did not affect the frequency and amplitudes of APs (Fig. 10b, dashed line). Along with the virtual SIC, associated changes in INCX, subsarcolemmal [Ca 2+] ([Ca 2+] sub), global [Ca 2+ ] i , and [Na+]i could be simulated (Fig. 10c). The amplitude and pattern of INCX was almost the same with SIC. A sharp transient increase in [Ca2+]sub associated with the fast activation of INCX reached close to 25 μM, while the global [Ca2+]c increase was around 1 μM. In contrast to the [Ca2+]i transient, the changes of [Na+]i was much slower than [Ca2+]i changes (Fig. 10c, lowermost panel). Similar with the recordings in hESCCMs, the virtual SICs could also be reproduced from −20 to 40 mV, and the frequency at depolarized voltages became faster. However, different from hESC-CMs, SICs were not reproduced at −40 mV (Fig. 10d).

The present study shows persistent generation of SICs in the significant population of voltage-clamped hESC-CMs. The pharmacological tests indicate that the SICs reflect electrogenic activities of NCX (forward INCX) triggered by cyclic release of SR Ca2+ via RyRs. The NCX inhibition decreased the slope of DD and frequency of spontaneous APs. Not only the hESCCMs prepared by our hands, but also the commercially available hiPSC-CMs showed SICs. The non-decaying SICs with relatively large amplitudes (60–150 pA) are unprecedented findings; previous voltage clamp studies of SANCs and mouse ESC-CMs showed only a few times of decaying fluctuations of membrane currents with small amplitudes [26, 29, 33]. Considering that NCXs are not ion channels but the electrogenic transporters with slower kinetics, the amplitudes of SICs in the present study suggested significant level of NCX expression in hESC-CMs, which was confirmed by confocal microscopy (Fig. 5d). It has been reported that the translational as well as transcriptional expressions of NCX1 in hESCs are increased during cardiac stem cell differentiation [3, 14]. In addition, the NCX1 expression in human ventricle reached its peak in the fetal stage, and then decreased in adult [23]. Unfortunately, we have not performed a quantitative assay of NCX proteins in hESC-CMs yet. The generation of SICs was tightly associated with repetitive Ca2+ oscillation (Fig. 4a). How could hESC-CMs and hiPSC-CMs maintain SICs under the voltage-clamped states? The forward mode of INCX implies removal of cytosolic Ca2+, which would diminish the SR Ca2+ store unless compensated by antiparallel Ca2+ influx mechanisms. In fact, the lowering of [Ca2+]e to 10 % of normal level effectively abolished SICs (Fig. 4c). As for the Ca2+ influx pathway, VOCCL seem to be highly likely one; the SICs were observed from above −40 mVand were abolished by nifedipine (Fig. 4e). However, one cannot exclude other mechanisms such as Ca2+-release activated Ca2+ channels (CRAC) that might be activated due to the repetitive Ca2+ release from SR. Recent study showed that hESC-CMs also express STIM1 and ORAI1 [30]. Physiological implication of SICs The shapes of SIC show at least three phases: (1) slowly incremental inward current (phase I), (2) fast, large inward current (phase II), and (3) fast decay of inward current (phase III). Because SICs would reflect the changes in [Ca2+]sub, phases I, II, and III might be the changes of INCX caused by (1) spillover leak of SR Ca2+, (2) fast regenerative CICR, and (3) fast Ca2+ removal by SERCA and NCX, respectively. According to the Ca2+-clock model [12], part of DD is due to inward INCX triggered by spontaneous local calcium release from SR (LCR). Then, phase I of SIC might be associated with the

Pflugers Arch - Eur J Physiol

Fig. 9 Color map of paired parameters with persistent SR Ca 2+ oscillation in voltage clamp condition. The color map shows the result of 100 × 100 = 10,000 simulations. The length of time of each simulation is 30 s, and the holding potential is −20 mV. The frequency of SR Ca2+ oscillations are presented as colors from blue to red tiles. a A color map of SR Ca2+ oscillation in terms of amplitude with parametric combinations of Ca 2+ pumping rate (P up ) and Ca 2+ release rate constant (k s ). Subsarcolemmal fractions of VOCCL and NCX are both fixed to 0.1. i–

iv Representative traces of SR [Ca2+] corresponding to each point indicated in a. b A color map of SR Ca2+ oscillation in terms of amplitude with parametric combinations of subsarcolemmal fractions of VOCCL and NCX. Ca2+ pumping rate (Pup) and Ca2+ release rate constant (ks) are fixed to 0.1 mM ms−1 and 2000 ms−1, respectively. v– viii Representative traces of SR [Ca2+] corresponding to each point indicated in b

putative LCR in hESC-CMs. In fact, the NCX inhibitor and SR emptying condition (caffeine) slowed the DD of hESCCMs (Fig. 6). The regenerative CICR is critical for the excitationcontraction coupling in cardiac myocytes. The significant amplitudes of phase II initially suggested that AP kinetics such as APD50 might be affected by the NCX. In fact, the APD50 appeared to be decreased by NCX inhibitor, SN-6 (Fig. 6d). However, the slight depolarization of MDP, and slowed DD by SN-6 might have indirectly attenuated the activation kinetics of VOCCL, which would also result to APD50 shortening. The persistent SICs initially drew our attention since they might play a critical role in the pacemaker potential (DD) in hESC-CMs. However, the hESC-CMs with no SICs still showed spontaneous APs. Furthermore, the generation of SICs is slow or inconsistently observed at the negative membrane voltages close to the MDP (e.g., −40 mV). Finally, the spontaneous APs were still observed in the presence of SN-6, an NCX inhibitor. Thus, the SIC observed in the present study is not the pacemaker mechanism per se. Nevertheless, the slowing of DD by SN-6 suggested that the forward INCX could facilitate DD. In the present study, we also tested the effects of

inhibitors specific to If (ivabradine), IKr (E4031), and IKs (chromanol-293B). However, none of them abolished the AP generation in hESC-CMs (Fig. 7). Interestingly, even the triple combined application of the three channel inhibitors did not abolish the APs (Fig. 7e). Under the depolarized state with multiple channel inhibitors, the spontaneous Ca2+ releases and SICs might induce the AP-like fluctuation of the membrane potential. Considering the critical physiological role of cardiac automaticity, multiple and redundant pacemaker mechanisms are not surprising. Interestingly, in the hESC-CM with calm membrane current initially, beta-adrenergic stimulation with isoproterenol could recover the SICs (Fig. 4f), suggesting that the INCX and spontaneous Ca2+ release from SR and INCX might be more important under the sympathetic tone. The betaadrenergic stimulations might have facilitated the Ca2+ influx and SR filling sate. A previous study showed that high basal PKA-dependent protein phosphorylation states are intrinsically significant in SANC because the treatment with PKA inhibitor reduced local Ca2+ release for the Ca2+ clock mechanism [29]. It was also reported that the sympathetic chronotropic effect became weakened in the NCX1 (−/−) mice [5].

Pflugers Arch - Eur J Physiol

Fig. 10 Reproduction of SICs in the modified M&L model. a Simulated membrane voltage (upper panel) and current (lower panel) in the original M&L model. Dynamic changes of membrane conductance along with the AP generation disappeared on converting to the voltage clamp condition. b Simulated membrane voltage and current in our modified model. The parametric combinations for persistent Ca2+ oscillation (see Fig. 9) reproduced SICs (middle and lower panels). Addition of voltage-gated

Na+ channel component (maximum conductance, 0.1 nS pF−1) did not significantly change the generation of SICs (dotted traces). c Simulations of virtual SIC (uppermost), INCX, [Ca2+]sub, [Ca2+]i, and [Na+]i (lower panels). d Simulations of virtual SICs at different clamp voltages ranging from −40 to 40 mV. Amplitudes become smaller while frequency becomes higher as voltage is depolarized

Computational modeling of SICs

cells than in the developed cardiomyocytes [2]. In the modification of L&M model to show SICs, it was also necessary to reduce the subsarcolemmal fraction of NCX less than 0.5 that may be consistent with the absence of t-tubules in hESC-CMs. Unfortunately, we do not have precise microstructural information of the putative colocalization state of NCX and RyR in hESC-CMs yet. Although the in silico generation of SICs is reproducible, the current computational model still requires modifications reflecting functional and cell biological data from further investigation. Our present study has another limitation because the modified M&L model is based on SA nodal cells, whereas hESCCMs also show the phenotypes of atrial and ventricular myocytes. A previous dedicated paper on mathematical modeling of AP of hESC-CMs had used ventricular cell model as a parental model, but the authors had to add If and VOCCT to generate the spontaneous APs in silico and did not separate the APs into nodal-like, atrial-like, and

The frequent observations of SICs under the nystatinperforated patch condition indicate that intact physiological systems and their interactions are critical. Such complex phenomenon could be better understood with simulations using computational modeling. To simulate the SICs, some parameters were modified from the M&L model of cardiac pacemaker cell [16]: (1) even distributions of VOCCL and NCX, (2) augmentation of SR Ca2+ uptake rate, and (3) reduction of SR Ca2+ release rate etc. Are the modifications just arbitrary changes to simulate SICs or physiologically relevant ones? Interestingly, experimental investigations of hESC-CMs and neonatal cardiac cells have reported similar characteristics; T-tubules are not developed whereas the NCX expression in the surface membrane is high [13]. In addition, the colocalization of NCX and RyR at cell surface layers (0 to 4 layers in cross section of a 3D myocyte) is higher in immature

Pflugers Arch - Eur J Physiol

ventricular-like APs [22]. Lacking complete series of electrophysiological studies on individual ion channels in our hESCCMs, we used M&L model as a parent model and modified it to simulate nodal-like APs recorded in our hESC-CMs. Although the amplitude, MDP, and APD50 of model-generated APs closely match those of experimentally obtained APs, it is essential to build three different types of mathematical model of hESC-CMs based on complete electrophysiological studies and test them before generalizing our idea on the mechanism of SICs. In summary, we report that both hESC-CMs and hiPSCCMs frequently show robust and regular SICs under the voltage clamp condition. Oscillatory Ca2+ release from SR and the associated forward INCX would underlie the SICs. Although the prominent SIC per se would not directly trigger AP, the slowed DD by SN-6 or caffeine treatment suggested that INCX might partly contribute to DD in hESC-CMs. In the human cardiomyocytes where guaranteeing the healthy live sample is extremely limited, further investigation of hESC-CMs might provide helpful knowledge of pacemaker mechanisms.

8.

9.

10.

11.

12.

13. 14.

Acknowledgments This work was supported by a National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIP) (Nos. 2007-0056092 and 2011-0017370) and also by the Bio and Medical Technology Development Program of the NRF (MSIP No. 2015M3A9C7030091)

15.

References

17.

Bhat MB, Zhao J, Zang W, Balke CW, Takeshima H, Wier WG, Ma J (1997) Caffeine-induced release of intracellular Ca2+ from Chinese hamster ovary cells expressing skeletal muscle ryanodine receptor. Effects on full-length and carboxyl-terminal portion of Ca2+ release channels. J Gen Physiol 110:749–62 2. Dan P, Lin E, Huang J, Biln P, Tibbits GF (2007) Threedimensional distribution of cardiac Na+-Ca2+ exchanger and ryanodine receptor during development. Biophys J 93:2504–2518 3. Fu JD, Jiang P, Rushing S, Liu J, Chiamvimonvat N, Li RA (2010) Na+/Ca2+ exchanger is a determinant of excitation-contraction coupling in human embryonic stem cell-derived ventricular cardiomyocytes. Stem Cells Dev 19:773–782 4. Gao Z, Chen B, Joiner ML, Wu Y, Guan X, Koval OM, Chaudhary AK, Cunha SR, Mohler PJ, Martins JB, Song LS, Anderson ME (2010) If and SR Ca2+ release both contribute to pacemaker activity in canine sinoatrial node cells. J Mol Cell Cardiol 49:33–40 5. Gao Z, Rasmussen TP, Li Y, Kutschke W, Koval OM, Wu Y, Wu Y, Hall DD, Joiner ML, Wu XQ, Swaminathan PD, Purohit A, Zimmerman K, Weiss RM, Philipson KD, Song LS, Hund TJ, Anderson ME (2013) Genetic inhibition of Na+-Ca2+ exchanger current disables fight or flight sinoatrial node activity without affecting resting heart rate. Circ Res 112:309–317 6. Groenke S, Larson ED, Alber S, Zhang R, Lamp ST, Ren X, Nakano H, Jordan MC, Karagueuzian HS, Roos KP, Nakano A, Proenza C, Philipson KD, Goldhaber JI (2013) Complete atrialspecific knockout of sodium-calcium exchange eliminates sinoatrial node pacemaker activity. PLoS One 8:e81633 7. Hazeltine LB, Simmons CS, Salick MR, Lian X, Badur MG, Han W, Delgado SM, Wakatsuki T, Crone WC, Pruitt BL, Palecek SP

16.

1.

18.

19.

20.

21.

22.

23.

24.

(2012) Effects of substrate mechanics on contractility of cardiomyocytes generated from human pluripotent stem cells. Int J Cell Biol 2012:508294 He JQ, Ma Y, Lee Y, Thomson JA, Kamp TJ (2003) Human embryonic stem cells develop into multiple types of cardiac myocytes: action potential characterization. Circ Res 93:32–39 Ju YK, Allen DG (1998) Intracellular calcium and Na+-Ca2+ exchange current in isolated toad pacemaker cells. J Physiol 508(Pt 1): 153–166 Kim C, Majdi M, Xia P, Wei KA, Talantova M, Spiering S, Nelson B, Mercola M, Chen HS (2010) Non-cardiomyocytes influence the electrophysiological maturation of human embryonic stem cellderived cardiomyocytes during differentiation. Stem Cells Dev 19:783–795 Kurata Y, Hisatome I, Imanishi S, Shibamoto T (2002) Dynamical description of sinoatrial node pacemaking: improved mathematical model for primary pacemaker cell. Am J Physiol Heart Circ Physiol 283:H2074–2101 Lakatta EG, Maltsev VA, Vinogradova TM (2010) A coupled SYSTEM of intracellular Ca2+ clocks and surface membrane voltage clocks controls the timekeeping mechanism of the heart’s pacemaker. Circ Res 106:659–673 Li S, Chen G, Li RA (2013) Calcium signalling of human pluripotent stem cell-derived cardiomyocytes. J Physiol 591:5279–5290 Liu J, Fu JD, Siu CW, Li RA (2007) Functional sarcoplasmic reticulum for calcium handling of human embryonic stem cell-derived cardiomyocytes: insights for driven maturation. Stem Cells 25: 3038–3044 Macia E, Boyden PA (2009) Stem cell therapy is proarrhythmic. Circulation 119:1814–1823 Maltsev VA, Lakatta EG (2009) Synergism of coupled subsarcolemmal Ca2+ clocks and sarcolemmal voltage clocks confers robust and flexible pacemaker function in a novel pacemaker cell model. Am J Physiol Heart Circ Physiol 296:H594–615 Mery A, Aimond F, Menard C, Mikoshiba K, Michalak M, Puceat M (2005) Initiation of embryonic cardiac pacemaker activity by inositol 1,4,5-trisphosphate-dependent calcium signaling. Mol Biol Cell 16:2414–2423 Moon SH, Kang SW, Park SJ, Bae D, Kim SJ, Lee HA, Kim KS, Hong KS, Kim JS, Do JT, Byun KH, Chung HM (2013) The use of aggregates of purified cardiomyocytes derived from human ESCs for functional engraftment after myocardial infarction. Biomaterials 34:4013–4026 Mummery C, Ward-van Oostwaard D, Doevendans P, Spijker R, van den Brink S, Hassink R, van der Heyden M, Opthof T, Pera M, de la Riviere AB, Passier R, Tertoolen L (2003) Differentiation of human embryonic stem cells to cardiomyocytes: role of coculture with visceral endoderm-like cells. Circulation 107:2733–2740 Mummery CL, Zhang J, Ng ES, Elliott DA, Elefanty AG, Kamp TJ (2012) Differentiation of human embryonic stem cells and induced pluripotent stem cells to cardiomyocytes: a methods overview. Circ Res 111:344–358 Niu CF, Watanabe Y, Ono K, Iwamoto T, Yamashita K, Satoh H, Urushida T, Hayashi H, Kimura J (2007) Characterization of SN-6, a novel Na+/Ca2+ exchange inhibitor in guinea pig cardiac ventricular myocytes. Eur J Pharmacol 573:161–169 Paci M, Sartiani L, Del Lungo M, Jaconi M, Mugelli A, Cerbai E, Severi S (2012) Mathematical modelling of the action potential of human embryonic stem cell derived cardiomyocytes. Biomed Eng Online 11:61 Qu Y, Ghatpande A, El-Sherif N, Boutjdir M (2000) Gene expression of Na+/Ca2+ exchanger during development in human heart. Cardiovasc Res 45:866–873 Sakakibara Y, Wasserstrom JA, Furukawa T, Jia H, Arentzen CE, Hartz RS, Singer DH (1992) Characterization of the sodium current in single human atrial myocytes. Circ Res 71:535–546

Pflugers Arch - Eur J Physiol 25.

26.

27.

28.

29.

Sartiani L, Bettiol E, Stillitano F, Mugelli A, Cerbai E, Jaconi ME (2007) Developmental changes in cardiomyocytes differentiated from human embryonic stem cells: a molecular and electrophysiological approach. Stem Cells 25:1136–1144 Sasse P, Zhang J, Cleemann L, Morad M, Hescheler J, Fleischmann BK (2007) Intracellular Ca2+ oscillations, a potential pacemaking mechanism in early embryonic heart cells. J Gen Physiol 130:133–144 Satin J, Itzhaki I, Rapoport S, Schroder EA, Izu L, Arbel G, Beyar R, Balke CW, Schiller J, Gepstein L (2008) Calcium handling in human embryonic stem cell-derived cardiomyocytes. Stem Cells 26:1961–1972 Viatchenko-Karpinski S, Gyorke S (2001) Modulation of the Ca2+induced Ca2+ release cascade by beta-adrenergic stimulation in rat ventricular myocytes. J Physiol 533:837–848 Vinogradova TM, Lyashkov AE, Zhu W, Ruknudin AM, Sirenko S, Yang D, Deo S, Barlow M, Johnson S, Caffrey JL, Zhou YY, Xiao RP, Cheng H, Stern MD, Maltsev VA, Lakatta EG (2006) High basal protein kinase A-dependent phosphorylation drives rhythmic internal Ca2+ store oscillations and spontaneous beating of cardiac pacemaker cells. Circ Res 98:505–514

30.

31.

32.

33.

34.

Wang Y, Li ZC, Zhang P, Poon E, Kong CW, Boheler KR, Huang Y, Li RA, Yao X (2015) Nitric oxide-cGMP-PKG pathway acts on orai1 to inhibit the hypertrophy of human embryonic stem cell-derived cardiomyocytes. Stem Cells 33: 2973–84 Yang X, Pabon L, Murry CE (2014) Engineering adolescence: maturation of human pluripotent stem cell-derived cardiomyocytes. Circ Res 114:511–523 Yaniv Y, Sirenko S, Ziman BD, Spurgeon HA, Maltsev VA, Lakatta EG (2013) New evidence for coupled clock regulation of the normal automaticity of sinoatrial nodal pacemaker cells: bradycardic effects of ivabradine are linked to suppression of intracellular Ca2+ cycling. J Mol Cell Cardiol 62:80–89 Zahanich I, Sirenko SG, Maltseva LA, Tarasova YS, Spurgeon HA, Boheler KR, Stern MD, Lakatta EG, Maltsev VA (2011) Rhythmic beating of stem cell-derived cardiac cells requires dynamic coupling of electrophysiology and Ca cycling. J Mol Cell Cardiol 50:66–76 Zhang YM, Hartzell C, Narlow M, Dudley SC Jr (2002) Stem cellderived cardiomyocytes demonstrate arrhythmic potential. Circulation 106:1294–1299