Ca2+-Currents in Human Induced Pluripotent Stem Cell-Derived Cardiomyocytes Effects of Two Different Culture Conditions

Human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CM) provide a unique opportunity to study human heart physiology and pharmacology and repair injured hearts. The suitability of hiPSC-CM critically depends on how closely they share physiological properties of human adult cardiomyocytes (CM). Here we investigated whether a 3D engineered heart tissue (EHT) culture format favors maturation and addressed the L-type Ca2+-current (ICa,L) as a readout. The results were compared with hiPSC-CM cultured in conventional monolayer (ML) and to our previous data from human adult atrial and ventricular CM obtained when identical patch-clamp protocols were used. HiPSC-CM were two- to three-fold smaller than adult CM, independently of culture format [capacitance ML 45 ± 1 pF (n = 289), EHT 45 ± 1 pF (n = 460), atrial CM 87 ± 3 pF (n = 196), ventricular CM 126 ± 8 pF (n = 50)]. Only 88% of ML cells showed ICa, but all EHT. Basal ICa density was 10 ± 1 pA/pF (n = 207) for ML and 12 ± 1 pA/pF (n = 361) for EHT and was larger than in adult CM [7 ± 1 pA/pF (p < 0.05, n = 196) for atrial CM and 6 ± 1 pA/pF (p < 0.05, n = 47) for ventricular CM]. However, ML and EHT showed robust T-type Ca2+-currents (ICa,T). While (−)-Bay K 8644, that activates ICa,L directly, increased ICa,Lto the same extent in ML and EHT, β1- and β2-adrenoceptor effects were marginal in ML, but of same size as (−)-Bay K 8644 in EHT. The opposite was true for serotonin receptors. Sensitivity to β1 and β2-adrenoceptor stimulation was the same in EHT as in adult CM (−logEC50: 5.9 and 6.1 for norepinephrine (NE) and epinephrine (Epi), respectively), but very low concentrations of Rp-8-Br-cAMPS were sufficient to suppress effects (−logEC50: 5.3 and 5.3 respectively for NE and Epi). Taken together, hiPSC-CM express ICa,L at the same density as human adult CM, but, in contrast, possess robust ICa,T. Increased effects of catecholamines in EHT suggest more efficient maturation.


INTRODUCTION
The L-type Ca 2+ -current (I Ca,L ) is central for cardiac electrophysiology. It contributes to the shape of the cardiac action potential and its regulation plays an important role in cardiac excitability and contractility (Tsien, 1983). L-type Ca 2+ -currents are activated upon depolarization while their activity can be increased by catecholamines (Hofmann et al., 2014). Since effects of norepinephrine on action potential precede the effects on tension it is assumed that stimulation of I Ca,L is related to inotropy (Reuter, 1974). Therefore, I Ca,L is expected to be crucial for adapting heart function to actual needs. T-type Ca 2+ -currents are typically found in pacemaking heart cells, but are absent from the working myocardium of many adult mammalians including man (Beuckelmann et al., 1991). HiPSC-CM provide a unique opportunity to study human heart electrophysiology in vitro and are believed to offer a model for pharmacological drug testing as well as disease modeling (Dick et al., 2010;Hoekstra et al., 2012;Navarrete et al., 2013). Yet, hiPSC-CM display an immature cardiac phenotype, and current efforts are directed toward means to unfold the full potential of these cells by increasing their maturity (Yang et al., 2014). One such strategy could be culture in engineered heart tissue (EHT) under conditions in which hiPSC-CM form a 3-dimensional network and perform auxotonic contractile work against elastic silicone posts (Schaaf et al., 2011). Here we directly compared the biophysics and regulation of Ca 2+ -currents in hiPSC-CM cultured either in standard monolayer format (ML) or as EHT and compared the data to our previous data on human adult CM obtained under identical patch-clamp protocols.

Human Adult Atrial and Ventricular CM
Adult myocardial tissue was obtained with informed consent from patients undergoing cardiac surgery at the Department of Heart Surgery, Dresden University of Technology. These studies were approved by the Medical Faculty Ethics Committee of Dresden University of Technology (document EK790799). Experiments were performed at the Department of Experimental Pharmacology and Toxicology, Medical Faculty, Dresden University of Technology between 2008 and 2011. Atrial and ventricular CM were isolated and prepared as previously described (Dobrev et al., 2000). Data about atrial CM displayed in Figures 1, 5D (experiments with NE and Epi) were obtained from raw data, used for publication recently (Christ et al., 2014). Data presented in Figures 2, 4, 5D (only experiments with Bay K) are from new experiments done in atrial cells from 6 patients. Patients were in stable sinus rhythm and 61.7 ± 2.7 years old. The majority of patients were treated with acetylsalicylic acid, ACE-inhibitors and β-blockers (for details see Table S1).

Western Blot
For protein extraction 28-day-old EHT were frozen in liquid nitrogen and stored at −80 • C. Each EHT was subjected to lysis with 70 µl 1x M-PERTM Mammalian Protein Extraction Reagent (Thermo Scientific) including protease and phosphatase inhibitor (Roche). The tissues were homogenized and supplemented with 1x Laemmli buffer prior to heating (95 • C, 5 min). SDS polyacrylamide gel (8%) was loaded with 3 µl non-failing human heart tissue lysate or 10-20 µl of EHT lysate. After electrophoresis proteins were blotted onto nitrocellulose membrane using the wet blot technique. Membranes were cut, washed with TBS-Tween 0.1%, blocked in 10% in low fat milk powder solution (1-2 h) and immediately transferred for incubation with the primary polyclonal rabbit antibodies against cardiac troponin T (1:1000; Abcam, 45932) and T-type calcium channel (1:100; alomone labs, SCC-021). After washing (TBS-Tween 0.1%, 3 × 10 min) membranes were transferred to anti-rabbit IgG peroxidase-conjugated secondary antibody (1:5000, Sigma, A0545; 1 h, room temperature in 1% low fat milk powder solution). Membranes were washed with TBS-Tween (0.1%, 3 × 15 min and 0.05% 2 × 15 min) and Pierce R ECL  Table S2). (B) Ca 2+ -current amplitudes (measured at +10 mV) vs. cell capacitance, dotted lines indicate linear regression fit (compare Table S3). Cells without ICa were not plotted. Same n-numbers as in 1C. (C) Frequency distribution of Ca 2+ -currents in human CM. Ca 2+ -currents are expressed as current densities, measured 3.5 min after membrane rupture (test-pulse potential +10 mV, for detailed analysis compare Table S4). (D) Mean values for absolute peak Ca 2+ -currents vs. cell capacitance. Please note that error bars may be smaller than symbols. Same n-numbers as in 1C. N/N indicates number of cells vs. number of isolation for ML and EHT and number of cells/number patients for atrial and ventricular CM.
Western Blotting Substrate (Thermo Scientific) used for band visualization.
Whole-Cell Recording of I Ca I Ca was measured at 37 • C using the whole-cell configuration of the patch-clamp technique (Axopatch 200B, Axon Instruments, Foster City, CA, USA), ISO 2 software was used for data acquisition and analysis (MFK, Niedernhausen, Germany). Heatpolished pipettes were pulled from borosilicate filamented glass (Hilgenberg, Malsfeld, Germany). Tip resistances were 2.5-5 M , seal resistances were 3-6 G . Cell capacitance (C m ) was calculated from steady-state current during depolarizing ramp pulses (1V/1s) from −40 to −35 mV. Ca 2+ -currents were elicited by applying test-pulses from −80 to + 10 mV (0.5 Hz). Extracellular Ca 2+ -concentration was 2 mM. The cells were investigated in a small perfusion chamber placed on the stage of an inverse microscope. Drugs were applied with a system for rapid solution changes (Cell Micro Controls, Virginia Beach, VA, USA; ALA Scientific Instruments, Long Island, NY, USA; Christ et al., 2004a). In order to avoid contaminating currents, K + -currents were blocked by replacing K + with Cs + and tetraethylammonium-chloride in the bath solution. The experiments were performed with the following Na + -free bath solution (in mM): Tetraethylammonium-chloride 120, CsCl 10, HEPES 10, CaCl 2 2, MgCl 2 1 and glucose 20 (pH 7.4, adjusted with CsOH). The pipette solution (pH 7.2, adjusted with CsOH) included (in mM): Cesium methanesulfonate 90, CsCl 20, HEPES 10, Mg-ATP 4, Tris-GTP 0.4, EGTA 10, and CaCl 2 3 (Christ et al., 2001). For some experiments (Figures 4A,B) we used the technique of perforated-patch. Amphotericin B was added to the pipette solution in a concentration of 1 µM. Current amplitude was determined as the difference between peak inward current and current at the end of the depolarizing step. Steadystate inactivation curves for I Ca were obtained by plotting the normalized current amplitude at the test potential as a function of the conditioning potential (V m ). A Boltzmann function was fitted to the normalized values: I/Imax = 1/(1 + exp((V m − V 0.5inact )/k inact )), where V 0.5inact and k inact are the voltage of half-maximum inactivation and the slope factor, respectively. Activation curves were calculated from current-voltage relations . Numbers next to the arrows indicate respective test pulse potential, for clarity some steps were omitted). (F) τ-values for the fast and slow inactivation phase and respective current amplitudes obtained by curve fitting at different test-pulse potentials A (G) and B (H). Please note that in human atrial CM no Ca 2+ -currents could be recorded at Vm < −10 mV. For details see Table S5. N/N indicates number of cells vs. number of isolation for ML and EHT and number of cells/number patients for atrial CM.
Frontiers in Pharmacology | www.frontiersin.org (IV-curves) using the equation G = I/(V m − E rev ), where G and I are peak Ca 2+ -conductance and current at the test potential V m , respectively. The apparent reversal potential E rev was obtained by linear regression of four data points close to E rev . The relation between normalized peak conductance G/G max and membrane potential V m could be described by a bi-exponential Boltzmann equation: is the voltage at half-maximum activation and k act is the slope factor.

Statistics
Results are presented as mean ± SEM. Curve fitting was performed by using the GraphPad Prism Software Version 5.02 (GraphPad Software, San Diego, CA, USA). Statistical differences were evaluated by using the Student's t-test (paired or unpaired) or repeated measures ANOVA, where appropriate. A value of p < 0.05 was considered to be statistically significant. Analyses of frequency distribution (Figures 1A,C) were performed using R (ver. 3.1.1) (R Core Team, 2013). Please note that the statistical term "Kernel density estimation" is used in Figure 1. In the first place, a "kernel" is defined as a probability density function which must possess particular properties. These are that it must be even, non-negative, real-valued and its definite integral over its support set must equal to 1. So "kernel density estimation" is a method which is non-parametric and enables to estimate the probability density function of a random variable (http://scikit-learn.org/ stable/modules/density.html).

Cell Size and Current Density
HiPSC-CM were consistently reported to be smaller than adult CM. We could confirm this finding under our cell culture conditions. When we plotted individual cell capacitance as an electrophysiological correlate of cell size in hiPSC-CM and human adult CM, we found substantial overlap between hiPSC-CM and adult CM, but mean values in hiPSC-CM were twoto three-fold smaller than human atrial and ventricular CM, respectively (Figure 1). Cells from EHT were not larger than from ML. While all cells from EHT and also every adult CM we measured possessed robust Ca 2+ -currents (at +10 mV), a remarkable number of cells in ML (35 out of 289) did not show I Ca . In adult CM I Ca depends on cell capacitance. Therefore, Ca 2+ -currents were normalized to cell capacitance and expressed as "current density" in order to minimize cell-to-cell variability of I Ca . Mean Ca 2+ -current density amounted to 9.9, 12.2, 7.1, and 5.7 pA/pF in ML, EHT, atrial and ventricular CM, respectively ( Figure 1C). Despite the fact that human adult CM were clearly larger than hiPSC-CM, both human adult and hiPSC-CM showed a positive correlation between cell size and I Ca . However, the regression curves for I Ca vs. cell capacitance were two times steeper in hiPSC-CM than in adult CM ( Figure 1B). Surprisingly R 2 values were unexpectedly low in all 4 populations investigated, questioning normalization of currents to cell size. In order to minimize Ca 2+ -current data scattering because of different cell size within the groups and to facilitate comparison to literature data, we will present current data as current density. As hiPSC-CM are expected to undergo maturation during culture, we have tested whether different populations of Ca 2+ -currents may exist and analyzed frequency distribution of current density in hiPSC-CM vs. adult CM. However, as in human adult CM, I Ca revealed a monophasic distribution in hiPSC-CM, indicating a homogenous population. Next we measured voltage-dependent activation of I Ca in hiPSC-CM and adult CM. In human adult CM, Ca 2+ -currents activated at test pulse potentials above −30 mV. In contrast, hiPSC-CM showed robust Ca 2+ -currents at much more negative test pulse potentials (starting already at −50 mV) giving a typical "shoulder" (Nemtsas et al., 2010). To get exact values for voltagedependency we constructed activation curves and fitted a biexponential Boltzmann function to the data points. About 11-13% of total Ca 2+ -conductivity consisted of T-type Ca 2+ -current (I Ca,T ) with activation voltage not different between ML and EHT. Steady state inactivation curves in hiPSC-CM also differed from adult CM (Mewes and Ravens, 1994;Christ et al., 2004b). Low-voltage activated I Ca inactivated at more negative potentials. The data are compatible with the expression of functional Ttype Ca 2+ -channels in hiPSC-CM. Accordingly, western blots with an antibody against Ca V 3.1 showed single bands at ∼270 kDa in both hiPSC-CM and a non-failing human heart sample, providing evidence for the presence of T-type Ca 2+ channel α-subunits in hiPSC-CM ( Figure 2D).

Time-Dependent Inactivation of I Ca (Figures 2E-H)
Total Ca 2+ -influx critically depends not only on voltagedependency of activation and resulting peak current amplitude but also on time-dependent inactivation at different membrane potentials. Therefore, we fitted Ca 2+ -current decay. As shown previously in human adult CM (Christ et al., 2004b), Ca 2+current inactivation could be fitted at positive potentials with two time constants: τ fast of ∼4 ms and a τ slow of ∼50 ms. Like in human adult CM, the quickly inactivating component of I Ca was larger than the slowly inactivating component. In summary we found the same biophysical properties of L-type Ca 2+ -currents as in adult CM.

Pharmacological Block of I Ca
Nifedipine Does not Completely Block I Ca in hiPSC-CM (Figure 3) To further evaluate the hypothesis that I Ca in hiPSC-CM contains I Ca,T we applied pharmacological blockers. We used nifedipine to block L-type and mibefradil to block T-type Ca 2+ -currents. Since the selectivity of mibefradil is moderate compared to the selectivity of nifedipine, cells were exposed first to nifedipine and mibefradil was added on top (Figures 3E,F). Exposure to high concentrations of nifedipine (10 µM) did not block Ca 2+ -currents completely (Figure 3). Scrutinizing IV-curves in the presence of nifedipine revealed that a large amount of the remaining current measured at +10 mV should mainly result from T-type Ca 2+ -currents. In line with this assumption, mibefradil blocked the nifedipine-insensitive current at +10 mV. The data suggest that a substantial amount of Ca 2+currents at +10 mV relates to I Ca,T , both in ML and EHT. This interpretation is complicated by the so-called slip-mode  conduction in which, in a Na + -free environment, the Na +channel can also conduct Ca 2+ ions, thereby mimicking I Ca,T (Mitra and Morad, 1986;Heubach et al., 2000). Since mibefradil not only blocks I Ca,T but also slip-mode conduction (Heubach et al., 2000) we employed tetrodotoxin (TTX; 30 µM). However, the Ca 2+ -currents at low test pulse potentials were insensitive to TTX, confirming the assumption of I Ca,T . Furthermore, I Ca,L was also insensitive to TTX ( Figure S1).

Run-Down and Activation of I Ca,L by (−)-Bay K 8644
Ca 2+ -currents in adult CM typically decrease over time, often with an initial rapid phase and stabilization over time even when the technique of perforated-patch was used. In hiPSC-CM we could not reach stable current densities in the time frame of 5 min. Since hiPSC-CM were found extraordinary fragile compared to human adult cardiomyocytes, we looked for a compromise. Therefore, we decided to measure basal current characteristics already 3.5 min after membrane rupture. In order to evaluate whether L-type Ca 2+ -currents in hiPSC-CM share the typical response to a Ca 2+ -channel opener we employed the dihydropyridine derivative (−)-Bay K 8644. The (−)-enantiomer was used as the (+)-enantiomer blocks I Ca,L and may thereby impair the maximal drug response (Ravens and Schoepper, 1990;Ji et al., 2014). High concentrations of (−)-Bay K 8644 (10 µM) increased Ca 2+ -currents in ML, EHT and adult atrial CM by 57.0, 29.9, and 91.1%, respectively ( Figure 4D). The effect size did not differ between ML and EHT when expressed as delta values (increase by 4.8 and 4.7 pA/pF, respectively).

Activation of I Ca,L by Catecholamines
Both β 1 -and β 2 -AR Stimulation Increase Currents in hiPSC-CM In the human heart I Ca,L is under regulation by β 1 -and β 2 -AR. To activate β 1 -AR we used norepinephrine (NE) in the presence of the selective β 2 -AR antagonist ICI118,551 (50 nM) and to activate β 2 -AR epinephrine (Epi) in the presence of the β 1 -AR antagonist CGP 20712A (300 nM; Kaumann et al., 1989). To assess maximum effects we used first very high concentrations of NE and Epi (100 µM each) and compared effect size to direct activation by (−)-Bay K 8644 (See Run-down and activation of I Ca,L by (−)-Bay K 8644). Activation of β 1 -and β 2 -AR increased I Ca,L in both ML and in EHT. The onset of the effect was as fast as in adult CM which is within ∼20 s (Christ et al., 2006). The effect of β 1 AR equaled that of β 2 AR stimulation in ML, EHT, and adult atrial CM. However, while β 1 -and β 2 -AR effects matched that of (−)-Bay K 8644 in EHT and atrial CM (Figures 5C,D), they were much lower in ML ( Figure 5B). Of note, even in EHT the absolute increases induced by catecholamines or (−)-Bay K 8644 were smaller than in adult CM. One explanation might be that L-type Ca 2+ -currents in hiPSC-CM are already high at baseline and cannot be increased any further. However, when plotted against each other, larger basal current densities were associated with larger current increases ( Figure S2), arguing against this hypothesis. Steeper linear regression curves in adult CM than in both hiPSC-CM groups also indicate that the smaller I Ca responses in hiPSC-CM are not related to basal properties.

Catecholamine Sensitivity
Sensitivity in hiPSC-CM is not Different From Adult CM ( Figure 6) Increases in I Ca by high concentrations of catecholamines were smaller in EHT than in adult CM. Either maximum effects could be diminished or sensitivity is so low that even 100 µM did not induce maximum effects. Therefore, we have measured concentration-response curves for increase of I Ca,L by β 1 -and β 2 -AR stimulation in EHT. Due to small effect size we refrained from doing so in ML. To avoid complications because of run-down and desensitization of receptors catecholamines were applied in a non-cumulative manner. The threshold concentration for both NE-and Epi-induced increases in I Ca was 100 nM and calculated −log EC 50 values amounted to 5.9 and 6.1, respectively (Figures 6C,D). These values do not differ from those determined in human adult ventricular CM and adult atrial CM (Christ et al., 2014).
CAMP-Dependency of I Ca,L -Increase due to β 1 -and β 2 -AR Stimulation Patch-clamp experiments in isolated cells give the unique opportunity to intracellularly apply inhibitors of signal transduction. In order to estimate cAMP levels relevant for I Ca activation we employed Rp-8-Br-cAMPS, a compound competitively inhibiting the effects of native cAMP on PKA (Figure 7). We measured effects of different concentrations of Rp-8-Br-cAMPS on basal currents and currents activated by maximum concentrations of catecholamines (100 µM). Since Ca 2+ -current increases by catecholamines in ML were small and hardly detectable in many experiments, we restricted the analysis to EHT. The presence of Rp-8-Br-cAMPS in the patch pipette did not decrease basal currents (data not shown). Both β 1 -and β 2 -AR mediated increases in EHT could be suppressed concentration-dependently with calculated −logEC 50 values of   Figures 7C,D). Compared to previous data from human adult atrial CM (Christ et al., 2014), −logEC 50 values were 20 times lower in EHT.

5-HT Increases I Ca in hiPSC-CM Less Than in Atrial CM (Figure 8)
Serotonin (5-HT) exerts positive inotropic effects in ventricular preparations from newborn, but not adult pigs (Jahnel et al., 1992;Schoemaker et al., 1992). In contrast, 5-HT-inotropy persists in atrial preparations from pigs (and even humans; Christ et al., 2014). The lack of positive inotropy in human ventricle could be related to the inability of 5-HT to increase I Ca , albeit evidence is based on very preliminary data (Jahnel et al., 1992). We therefore evaluated the effect of 5-HT (100 µM) on Ca 2+currents in hiPSC-CM, indicative of an immature and/or atriallike phenotype (Figure 8). First we confirmed the failure of 5-HT to raise I Ca in human ventricular CM in a larger number of cells. In atrial CM 5-HT-evoked increases were similar compared to direct Ca 2+ -channel activation by (−)-Bay K 8644. HiPSC-CM showed increases in I Ca (Figure 8C), but at much smaller size than in atrial CM. While increases in ML cells accounted to 60% of (−)-Bay K 8644-effects, they amounted to only 20% in EHT, indicating more advanced maturation or ventricular differentiation.

DISCUSSION
Here we evaluated whether Ca 2+ -channels in hiPSC-CM share typical properties and regulatory mechanisms of Ca 2+ -channels in human adult CM and whether advanced culture conditions in EHT favor maturation. We found many similarities but also distinct differences between hiPSC-CM and adult human CM: (I) Basal I Ca,L density was not smaller in hiPSC-CM than in adult CM. (II) HiPSC-CM, other than adult human CM, express Ttype Ca 2+ -currents and the necessary pore forming α-subunit ( Figure 2D). (III) Voltage-dependency of activation, steady state inactivation and inactivation kinetics of I Ca,L in hiPSC-CM were not different from those found in adult CM (IV) I Ca,L in hiPSC-CM was increased upon β 1 -and β 2 -AR stimulation with the same sensitivity as in adult human CM, but maximum effect size was smaller. (V) Catecholamine-induced increases of I Ca,L in hiPSC-CM were PKA-dependent, but the amount of cAMP related to Ca 2+ -increase was less.

Cell Size
Smaller than normal mean cell size is considered as one of the hallmarks of hiPSC-CM (Yang et al., 2014). Mean cell size in hiPSC-CM as determined by cell capacitance in a large number of cells was approximately two-fold and three-fold smaller than atrial and ventricular CM, respectively (45.2 vs. 86.7 and 125.7 pF). While the data principally confirm previous conclusions (Yang et al., 2014), absolute values were almost three-fold higher in our hiPSC-CM (45.2 vs. 17.5 pF). The reasons are not clear, but may be related to different differentiation and culture protocols. Of note, cell capacitance/size showed substantial overlap between hiPSC-CM and adult atrial and even ventricular CM, indicating that at least some hiPSC-CM reach an adult-like size. Identical values in ML and EHT imply that the EHT-format did not have favorable effects on this parameter of CM maturity. Small size represents technical limitations in patch-clamp as discussed by Wilson et al. (2011). However, analyzing cell size and Ca 2+currents revealed a rather monotonic frequency distribution and Ca 2+ -currents showed linear dependency on cell size also in the low cell size range. Therefore, we believe that the relatively small cell size of hiPSC-CM is not a major obstacle for the measurement of large membrane currents such as Ca 2+ -currents.

Ca 2+ -Current Density in hiPSC-CM
Few studies measured Ca 2+ -currents in hiPSC-CM and reported current densities of 3.3-17.1 pA/pF (Ma et al., 2011;Yazawa et al., 2011). Keeping in mind that results are hardly comparable due to different methodology, we compared our measurements from hiPSC-CM to previous data on human adult atrial as well as ventricular CM under identical experimental conditions. I Ca,L was undetectable in 10% of cells isolated from conventional ML, indicating either a very low level of cardiac differentiation or a by-chance picking of a non-cardiac cell, present in our differentiation protocol at ∼10-15%. In contrast, all cells from EHT showed robust Ca 2+ -currents. Absolute Ca 2+ -current amplitudes in hiPSC-CM did not differ significantly from adult atrial and ventricular CM, but, at smaller mean cell size, mean current density tended to be larger in hiPSC-CM (9.9-12.2 vs. 5.7-7.1 pA/pF), giving a potentially wrong impression of basal "overactivity" of Ca 2+ -currents in hiPSC-CM.

Biophysical and Pharmacological Properties of I Ca
Low-Voltage Activated I Ca : I Ca,T or I Ca,TTX ?
The contribution of Ca 2+ -currents to depolarization and to transsarcolemmal Ca 2+ -influx depends on the voltagedependency of every individual Ca 2+ -channel type expressed and their relative amplitude. Our data in hiPSC-CM showed Ca 2+ -currents activating at low and high voltage, suggesting contribution of T-type to total Ca 2+ -influx over the physiological voltage range. Several observations support this hypothesis. Besides the typical voltage for half-maximum activation between 32 and 37 mV (Hansen et al., 2004), western blots from hiPSC-CM showed a robust signal for Ca v 3.1, the ion channel subunit carrying I Ca,T (Hansen et al., 2004). The highly specific Ltype blocker nifedipine reduced total I Ca only by ∼80% and the remainder was sensitive to the non-selective I Ca blocker mibefradil. However, measuring Ca 2+ -currents at low voltages can raise complications since Na + -channels in the absence of Na + can conduct Ca 2+ to some extent (slip-mode conduction) and thereby give a wrong impression of I Ca,T (Mitra and Morad, 1986). Even worse, slip-mode conduction is at least in part sensitive to mibefradil (Heubach et al., 2000). In order to discriminate whether currents we measured at low voltages may represent an artifact or not, we employed tetrodotoxin (TTX; Lemaire et al., 1995). Thirty microliter of TTX is expected to block slip-mode conduction completely, but does not affect I Ca,T (Mitra and Morad, 1986;Heubach et al., 2000). In our experiments Ca 2+ -currents at low voltage were completely insensitive to TTX, providing further evidence for the existence of functional T-type Ca 2+ -channels in ML and EHT ( Figure   S1). Unexpectedly, in contrast to many classic pharmacological studies, I Ca,L was reported to be sensitive to high concentrations of TTX (>30 µM). However, results were obtained in canine cardiomyocytes only (Hegyi et al., 2012). There are no data on human cardiomyocytes.

I Ca,T in hiPSC-CM
The finding of T-type Ca 2+ -currents in hiPSC-CM is important. Former studies characterizing Ca 2+ -currents in hiPSC-CM did either not support the existence of T-type Ca 2+ -currents (Ma et al., 2011) or did not address this question (Yazawa et al., 2011). T-type currents were found in ventricular CM from fishes (Maylie and Morad, 1995;Nemtsas et al., 2010), dogs, guinea pig and sinoatrial node cells from rabbit (Mitra and Morad, 1986;Hagiwara et al., 1988;Hirano et al., 1989). In human adult atrial and ventricular myocardium, T-type Ca 2+ -currents were consistently absent (Beuckelmann et al., 1991;Ouadid et al., 1991;Li and Nattel, 1997;Bosch et al., 1999). Data about T-type Ca 2+channel in development of human heart are understandably rare. Qu and Boutjdir (2001) found a decline of T-type Ca 2+ -channel mRNA expression by RT-PCR in fetal ventricular tissue over time of development. DeHaan (1989, 1990) found large T-type Ca 2+ -channel amplitudes in chicken embryos, but no change over time. T-type Ca 2+ -currents in rat atrial CM dropped only slightly during postnatal development (Xu and Best, 1992). Therefore, it remains unclear whether I Ca,T is an indicator of cardiac myocyte immaturity (Ono and Iijima, 2010). While the role of I Ca,T in pacemaking is established (Marger et al., 2011;Mesirca et al., 2014), its relevance in the working myocardium is less clear. Effects of β-AR stimulation seem to be speciesdependent, with stimulation in dog and guinea pig ventricular CM (Mitra and Morad, 1986;Tseng and Boyden, 1991), but no effects in shark (Maylie and Morad, 1995). In our cells the Ttype Ca 2+ -current was insensitive to β-AR stimulation ( Figure  S3). Taken together, the co-existence of T-type and L-type Ca 2+channels in hiPSC-CM suggests a permanent inward current at low potentials. Such a current very likely contributes to the (abnormal) pacemaking in these cells. Further studies have to clarify the long-term functional relevance of I Ca,T in hiPSC-CM.

L-Type Ca 2+ -Currents in hiPSC-CM Show Classic Pharmacological Properties: (−)-Bay K 8644
In cellular electrophysiology individual membrane currents were frequently identified by selective blockers. I Ca,L cannot only be blocked but also be activated, for example by (−)-Bay K 8644 (Schramm et al., 1983). Because the compound circumvents activation via intracellular signaling cascades it is often used to estimate maximum activity of L-type Ca 2+ -channels (Ouadid et al., 1991). Accordingly, Ji et al. (2014) employed the agent recently in commercially available hiPSC-CM (Cor.4U, Axiogenesis, iCell, Cellular Dynamics International) and could detect increases in currents only when Ba 2+ was used as charge carrier (eliminating Ca 2+ -dependent inactivation of Ca 2+ -channels) and cells were hold at low potentials to increase the affinity of (−)-Bay K 8644. In contrast, (−)-Bay K 8644 robustly increased I Ca,L both in ML and EHT in our hands. The discrepancy could be due to differences in the experimental protocol. For example, application of (−)-Bay K 8644 early after patch rupture may interfere with the initial fast run-down phase. More likely, however, the discrepant results with (−)-Bay K 8644 indicate a different biology of Ca 2+ -channels in cells from Cor.4U and iCell compared to our hiPSC-CM. Since we found (−)-Bay K 8644 effect sizes to be independent from culture condition (EHT vs. ML), differences in the differentiation protocol [e.g., growth factor-based (Burridge et al., 2012)] vs. small molecule-based (Burridge et al., 2014) may be more likely underlying. Head-to-head investigations are needed to clarify this issue.

Catecholamine Responses are Smaller in ML than in EHT
In human (in contrast to rat and mouse) adult myocardium activation of β 1 -and β 2 -AR increases I Ca,L to the same extent (Christ et al., 2014). Absolute current increase depends on temperature (Christ et al., 2011). We could confirm both findings for hiPSC-CM ( Figure S4). Culture conditions had a main impact on catecholamine responses. While catecholamine effects equaled that of (−)-Bay K 8644 in EHT, effects in ML were clearly smaller. Both hiPSC-CM responses were smaller than in adult atrial or ventricular CM. It should be noted, that effect size is smaller in ventricles compared to atria. The differences are not due to higher baseline current density, because current density was positively associated with the β-AR response in all preparations. The sensitivity for activation of β 1 -AR by NE and β 2 -AR by Epi was identical to adult ventricular CM (this study) and adult atrial CM (Christ et al., 2014). Smaller maximum effect size but preserved sensitivity could indicate proper AR function but reduced ability of adenylate cyclase to generate cAMP. In order to estimate the amount of cAMP activating I Ca,L , we measured maximum effects in the presence of different concentrations of Rp-8-Br-cAMPS, which inhibits binding of cAMP to PKA competitively (Van Haastert et al., 1984). As shown before for human adult CM (Christ et al., 2014), basal current activity was independent from PKA activity, but catecholamine effects were concentrationdependently suppressed. The observation that 20 times less Rp-8-Br-cAMPS was sufficient to inhibit catecholamine-induced effects in EHT than in adult atrial CM (Christ et al., 2014), indicates that the small β-AR-effect on I Ca,L in EHT could be related to lower β-AR-induced cAMP-generation by an immature β-AR/Gs-protein/adenylyl cyclase signaling complex.

5-HT Increases I Ca,L in hiPSC-CM:
Indicator for Immaturity or Just Atrial-Like Phenotype?
Expression of 5-HT-receptor transcripts (RT-PCR) is higher during fetal development, and increased expression of 5-HTreceptors in adult myocardium under pathological conditions is interpreted as fetal (Brattelid et al., 2012). Data for I Ca,Lreponses to 5-HT from fetal heart cells are lacking. In human adult heart 5-HT-responses are restricted to atrium (Kaumann et al., 1990;Ouadid et al., 1992). 5-HT effects on I Ca,L in hiPSC-CM indicate an atrial and/or immature phenotype. Smaller 5-HT-effects together with larger catecholamine effects indicate that EHT favor functional maturation compared to standard ML.

CONCLUSIONS
In hiPSC-CM from both ML and EHT we found I Ca,L not smaller than in human adult CM. Basal current densities as well as current increases to (−)-Bay K 8644 did not differ between ML and EHT. However, only hiPSC-CM from EHT showed robust catecholamine responses, suggesting maturation of the β-AR/Gs-protein/adenylyl cyclase signaling complex.

LIMITATIONS
We used ventricular CM obtained from patients with end-stage heart-failure only. I Ca,L responses are reported to be reduced in heart failure by Chen et al. (2002), whereas Mewes et al. (Mewes and Ravens, 1994) demonstrated unchanged responses. In both studies basal currents were not affected (Mewes and Ravens, 1994;Chen, 2002). HiPSC-CM lose their rod-shaped appearance during digestion very quickly. Freshly isolated hiPSC-CM do not adhere to the bottom of a recording chamber even if coated. We are sorry to report that we have failed to handle freshly isolated hiPSC-CM in electrophysiological experiments. Re-culturing hiPSC-CM from EHT in monolayer format could reverse some effects of EHT culture.

AUTHOR CONTRIBUTIONS
AU, AHo, IM, and KB performed research. AU, AHo, AHa, TE, and TC planned experiments. AU, IM, KB, and SJ analyzed results. AU, TE, and TC wrote the manuscript. All authors approved the final version of the manuscript.