Heart and Circulatory Physiology

High purity human-induced pluripotent stem cell-derived cardiomyocytes: electrophysiological properties of action potentials and ionic currents

Junyi Ma, Liang Guo, Steve J. Fiene, Blake D. Anson, James A. Thomson, Timothy J. Kamp, Kyle L. Kolaja, Bradley J. Swanson, Craig T. January


Human-induced pluripotent stem cells (hiPSCs) can differentiate into functional cardiomyocytes; however, the electrophysiological properties of hiPSC-derived cardiomyocytes have yet to be fully characterized. We performed detailed electrophysiological characterization of highly pure hiPSC-derived cardiomyocytes. Action potentials (APs) were recorded from spontaneously beating cardiomyocytes using a perforated patch method and had atrial-, nodal-, and ventricular-like properties. Ventricular-like APs were more common and had maximum diastolic potentials close to those of human cardiac myocytes, AP durations were within the range of the normal human electrocardiographic QT interval, and APs showed expected sensitivity to multiple drugs (tetrodotoxin, nifedipine, and E4031). Early afterdepolarizations (EADs) were induced with E4031 and were bradycardia dependent, and EAD peak voltage varied inversely with the EAD take-off potential. Gating properties of seven ionic currents were studied including sodium (INa), L-type calcium (ICa), hyperpolarization-activated pacemaker (If), transient outward potassium (Ito), inward rectifier potassium (IK1), and the rapidly and slowly activating components of delayed rectifier potassium (IKr and IKs, respectively) current. The high purity and large cell numbers also enabled automated patch-clamp analysis. We conclude that these hiPSC-derived cardiomyocytes have ionic currents and channel gating properties underlying their APs and EADs that are quantitatively similar to those reported for human cardiac myocytes. These hiPSC-derived cardiomyocytes have the added advantage that they can be used in high-throughput assays, and they have the potential to impact multiple areas of cardiovascular research and therapeutic applications.

  • patch clamp
  • ion channels
  • iCell cardiomyocytes

human-induced pluripotent stem cell (hiPSC)-derived cardiomyocytes are providing new approaches to studying the cardiac cellular properties of healthy individuals and those with inheritable diseases (5, 16, 29, 49, 52). hiPSC-derived cardiomyocytes, as well as those derived from human embryonic stem cells (hESCs), also provide in vitro platforms for drug discovery and potentially for regenerative therapies (1, 43, 48).

Differentiation of hiPSCs or hESCs into cardiomyocytes can be achieved using embryoid body and various directed differentiation approaches (13, 23, 48, 52). Stem cell-derived cardiomyocytes express cardiac-specific genes and proteins; display atrial-, nodal-, and ventricular-like cardiac action potentials (APs); generate triggered arrhythmia mechanisms of early and delayed afterdepolarizations (EADs and DADs); contract rhythmically; have intact sarcoplasmic reticulum Ca2+ release and β-adrenergic responsiveness; and possess functional excitation-contraction coupling properties resembling adult cardiomyocytes (4, 11, 13, 34, 39, 52). At the same time, there are limitations to present stem cell derived-cardiomyocytes that include the small numbers of myocytes produced, the presence of contaminating proliferative cell types (e.g., fibroblasts, etc.), and their immature nature. A result is limited information about electrophysiological and biochemical properties of hiPSC-derived cardiomyocytes and the small number of research applications for which these cells have been applied.

In the present study, we characterized cardiomyocytes generated from a hiPSC line engineered to permit selection for high cardiomyocyte purity by expressing blasticidin resistance under the control of the cardiac myosin heavy chain 6 (MYH6) promoter. A directed differentiation approach was used for high volume cardiomyocyte generation. APs, responses to canonical drugs, and multiple ionic currents [sodium (INa), L- and T-type and calcium (ICaL and ICaT), hyperpolarization-activated pacemaker (If), transient outward potassium (Ito), inward rectifier potassium (IK1), and the rapidly and slowly activating components of delayed rectifier potassium (IKr and IKs) currents] were then characterized in single hiPSC-derived cardiomyocytes. The key findings are that these hiPSC-derived cardiomyocytes display cellular electrophysiological properties similar to human cardiomyocytes, that they can be grown at high purity and in large numbers, and that these cells can potentially impact many areas of cardiovascular investigation.


hiPSC-derived cardiomyocyte generation and cell culture.

The hiPSC cell line was created by reprogramming a human fibroblast cell line by retroviral expression of the reprogramming factors sox7, oct4, nanog, and lin28 similar to Yu and colleagues (50) using Moloney murine leukemia virus viral constructs (50). This line was used to generate hiPSC clones that were engineered to exhibit blasticidin resistance in cardiomyocytes by inserting the coding region of the BSD gene encoding blasticidin S deaminase from Aspergillus terreus in-frame downstream of the last exon of the native MYH6 gene coding region through homologous recombination (24, 56). The MYH6 and BSD coding regions were separated by picornavirus 2A translational slip site to prevent MYH6 haploinsufficiency (7–8). A neomycin-resistance gene driven by the constitutive phosphoglycerate kinase promoter was included in the targeting vector downstream of the BSD gene to allow selection of genomic integrants by inclusion of neomycin in the hiPSC culture medium. Neomycin-resistant iPSC colonies were subjected to genomic DNA PCR screening to correctly identify targeted homologous recombinants using primer sets that spanned both arms of the homologous recombination sites. Cardiomyocytes were derived from this engineered hiPSC clonal line as follows. hiPSC aggregates were formed from single cells and cultured in suspension in medium containing zebrafish basic fibroblast growth factor and FBS. Upon observation of beating cardiac aggregates, cultures were subjected to blasticidin selection at 25 ug/ml (24) to enrich the cardiomyocyte population. Cardiomyocyte aggregate cultures were maintained in DMEM containing 10% FBS during cardiomyocyte selection through the duration of the culture before cryopreservation. At 30 to 32 days of culture, the enriched hiPSC-derived cardiomyocytes were subjected to enzymatic dissociation using 0.5% trypsin to obtain single cell suspensions of purified cardiomyocytes, which are >98% cardiac troponin-T (cTNT) positive as described below. These were cryopreserved and stored in liquid nitrogen (iCell Cardiomyocytes; Cellular Dynamics International; Madison, WI). For this study, single vials containing >1.5 × 106 cardiomyocytes were thawed by immersing the frozen cryovial in a 37°C water bath, transferring thawed cardiomyocytes into a 15-ml tube, and diluting them with 10 ml of ice-cold plating medium [iCell Cardiomyocytes Plating Medium (iCPM); Cellular Dynamics International].

For single cell patch-clamp recordings, glass coverslips were coated with 0.1% gelatin (Sigma, St. Louis, MO) and placed into each well of a 12-well plate, and 2 ml of iCPM containing 40,000–60,000 cardiomyocytes were added to each coverslip. Plated cardiomyocytes were at a low density to permit culture as single cells and were stored in an environmentally controlled incubator maintained at 37°C and 7% CO2. After 48 h, iCPM was replaced with a cell culture medium [iCell Cardiomyocytes Maintenance Medium (iCMM); Cellular Dynamics International], which was exchanged every other day with the cardiomyocytes maintained on cover slips for 4 to 21 days before use.

Flow cytometry and immunofluorescence.

To analyze hiPSC-derived cardiomyocyte purity, single cell suspensions of cultured cardiomyocytes, freshly thawed or after 2 wk in culture in iCMM, were first stained for cell viability with Invitrogen live/dead stain (Invitrogen, Carlsbad, CA) for 5 min and then fixed in 5.5% formaldehyde for 15 min. After fixation, cell membranes were permeabilized in PBS (Invitrogen) containing 0.5% saponin and 10% FBS. Permeabilized cardiomyocytes were stained with either a mouse anti-cTNT monoclonal antibody (Ab8295; Abcam, Cambridge, UK) or a matching isotype control (M5284; Sigma). Following primary antibody staining, cardiomyocytes were washed in permeabilization buffer and stained with Alexa Fluor 647 goat anti-mouse IgG1 followed by secondary antibody (A21240; Invitrogen). The stained cardiomyocytes were analyzed using an Accuri C6 flow cytometer, and histogram plots quantifying the percentage of live cardiomyocytes that are cTNT positive or control were generated using FCS Express Professional (DeNovo Software, Los Angeles, CA).

For immunofluorescence, cryopreserved cardiomyocytes were thawed and plated into 12-well plate containing gelatin-coated coverslips at a density of 400,000 cardiomyocytes/well (Fisher Scientific, Fitchburg, WI). After incubating for an additional 11 days, cardiomyocytes form a beating monolayer that was fixed with 4% paraformaldehyde for 15 min, permeabilized with 0.1% Triton X-100, and blocked with 3% milk for 15 min. Cardiomyocytes were then stained with mouse anti-sarcomeric α-actinin primary antibody (AB9465; Abcam; 1:200) overnight at 4°C, washed three times with PBS, and then stained with Alexa Fluor-647 donkey anti-mouse IgG (Invitrogen; 1:500) for 30 min in PBS. After being stained, the cardiomyocytes were washed once with PBS, stained with Hoescht 33342 anti-nuclear antibody (Invitrogen, 1:10,000) for 15 min in PBS, and then washed an additional three times in PBS. Immunofluorescence imaging was performed using a ×20 objective on an Olympus IX81 microscope recorded with a QImaging EXi Blue CCD camera (QImaging, Surrey, BC, Canada) using Surveyor software (Objective Imaging, Cambridge, UK).

Electrophysiological recordings.

Cardiac APs and ionic currents were recorded from single cardiomyocytes with perforated (gramicidin) and ruptured whole cell patch-clamp techniques, respectively. Coverslips containing plated cardiomyocytes were transferred to a RC-25-F recording chamber (Warner Instruments, Hamden, CT) mounted on to the stage of an inverted microscope. Extracellular solution perfusion was continuous using a VC-6 valve controller (Warner Instruments) with solution exchange requiring ∼1 min. Temperature was maintained constant by a TC-344B dual channel heating system (Warner Instruments) at 35–37°C except for Ca2+ current recordings that were performed at room temperature. Pipettes were pulled from thin wall borosilicate glass capillaries with a DMZ universal puller (Dagan, Minneapolis, MN) and had resistances between 1.5 and 3.5 MΩ with access resistances of <5 MΩ for ruptured patch recordings and 10–30 MΩ for gramicidin perforated patch recordings. The gramicidin perforated patch method allows for pipette-intracellular exchange of monovalent cations only to minimize the effects of cell dialysis and improve AP waveform stability over time (26). Series resistance and cell capacitance were compensated to between 50 and 80% in all voltage-clamp recordings. Data were acquired at 10 or 100 kHz and filtered at 1 or 10 kHz with the Axon 700B Multiclamp, Digidata 1322A digitizer hardware and pClamp 10.2 software (Molecular Devices, Sunnyvale, CA) depending on experimental requirements. For ruptured whole cell patch-clamp experiments, cardiomyocyte cell capacitance averaged 88.7 ± 5.0 pF (n = 48; cardiomyocytes used for measurement of INa are not included, see below).

For some experiments, planar automated whole cell patch-clamp recordings were obtained with a PatchXpress 7000A (Molecular Devices) that permits the simultaneous recording of up to 16 channels of data. After thawing, cardiomyocytes were plated for 2 to 14 days into 6-well tissue culture plates at a seeding density of 0.5–1 × 106 per well. From there, ∼25 × 103 cardiomyocytes were added to each SealChip recording well and allowed to form gigaohm seals. Data were acquired at 5 kHz and low-pass filtered at 1 kHz for INa or acquired at 1 kHz and low-pass filtered at 0.2 kHz for ICa or IKr with PatchXpress Commander software (Version; Molecular Devices). Recordings were obtained at ambient temperature.

Solutions and protocols.

The pipette and extracellular solution compositions used in different protocols are listed in Supplemental Table S1 (Supplemental Material for this article is available online at the Am J Physiol Heart Circ Physiol website). For perforated patch recording, pipettes were backfilled with solution containing 50 μg/ml gramicidin (diluted from a 50 mg/ml gramicidin DMSO stock solution). For automated planar patch recordings, seals were first obtained in sealing solution before switching to different extracellular solutions. Stock solutions of E4031 (Alomone Labs, Jerusalem, Israel), 4-aminopyridine (Sigma), ZD7288 (Tocris, Ellisville, MO), and tetrodotoxin (TTX; Ascent Scientific, Princeton, NJ) were prepared in distilled water. Nifedipine (Sigma) and (−)-[3R,4S]-chromanol 293B (Tocris) was dissolved in DMSO. All stock solutions were diluted ≥1:1,000 in extracellular solution before use. All other chemicals were purchased from Sigma.

Spontaneous APs were categorized as atrial-, nodal-, or ventricular-like as described in results. The pharmacological sensitivity of ventricular-like APs was assessed at a constant pacing rate of 1 Hz with APs elicited through 5-ms depolarizing current injections of 150–550 pA. Offline data analysis, curve fitting, and statistical evaluations were performed with Origin (OriginLab, Northampton, MA) and Spike2 (CED, Cambridge, UK) software packages. Data are presented as means ± SE. Statistical significance was determined by Student's t-test or two-way repeated-measures ANOVA with post hoc testing using the Tukey method. P < 0.05 was considered statistically significant.


Flow cytometry and immunofluorescence analyses.

Figure 1A shows representative flow cytometry data to assess cardiomyocyte purity. “Live” (live/dead stain) permeabilized cardiomyocytes were stained for cTNT or matching isotype control. For each analysis, the cardiomyocyte cultures contained cTNT-positive cells with a purity of >98% (filled histogram, n = 3 experiments), and after 2 wk in culture cardiomyocyte purity remained unchanged. Figure 1B shows a cardiomyocyte monolayer immunostained for sarcomeric α-actinin (red) and nuclei (blue). Nuclei are evenly distributed and α-actinin staining is present throughout the field. The higher power image (inset) shows the subcellular sarcomeric organization of contractile myofibrils.

Fig. 1.

Purity and structural organization of cardiomyocytes. A: flow cytometry analysis of cardiomyocytes stained for cardiac troponin-T (cTNT; closed histogram) or isotype control (open histogram). Cell count indicates the number of cells at each fluorescence intensity. Bar represents the troponin-T-positive gate with 99.3% of cells in this example stained for cTNT. B: immunofluorescent analysis of a cardiomyocyte monolayer stained for sarcomeric α-actinin (red) and with anti-nucleic acid stain (blue). Scale bar = 50 μm. Sarcomeric structure can be identified at higher magnification (inset, scale bar = 20 μm).

AP properties.

APs were recorded from spontaneously beating cardiomyocytes 10 to 21 days postthaw. Similar to human cardiac myocytes, hiPSC-derived cardiomyocyte APs have different morphologies that can be categorized as atrial-, nodal-, or ventricular-like. Example APs are shown in Fig. 2 (extracellular K = 5.4 mmol/l, ionic conditions given in Supplemental Table S1). APs from 59 cardiomyocytes that beat spontaneously at a constant rate were analyzed for beat rate interval (Interval), maximum diastolic potential (MDP), peak voltage (Peak), amplitude (Amp), maximal rate of depolarization (dV/dtmax), and AP duration (APD) at different levels of repolarization (APD measured at 10% increments of Amp). These data are summarized in Fig. 2. The following criteria were used to distinguish the AP phenotypes: atrial- and ventricular-like APs displayed more negative MDPs with higher dV/dtmax values and larger amplitudes than nodal-like APs. Ventricular-like APs have a distinct plateau phase (phase 2) after which repolarization accelerates (phase 3). This was measured as the time difference between APD30 to APD40 (APD30–40) and the time difference between APD70 to APD80 (APD70–80). Ventricular-like APs had a ratio (APD30–40/APD70–80) of ∼2.5 compared with the atrial-like ratio of ∼1.1; thus APs with a ratio >1.5 were categorized as ventricular like. Ventricular-like APs had longer duration APDs with slower rates of spontaneous beating (increased Interval), whereas atrial-like APs had a minimal plateau phase (APD30–40/APD70–80 close to unity) with shorter durations and higher spontaneous beating rates. Nodal-like APs had less negative MDPs and smaller amplitudes with lower dV/dtmax values (<10 V/s). With the use of these criteria, 54% of the 59 cardiomyocytes displayed ventricular-like APs, 22% showed nodal-like APs, and 24% showed atrial-like APs.

Fig. 2.

Action potentials (APs) were recorded from human-induced pluripotent stem cells (hiPSC)-derived cardiomyocytes at 35–37°C using the perforated patch-clamp method. See text for details. Interval, beat rate interval; MDP, maximum diastolic potential; Peak, peak voltage; Amp, amplitude; dV/dtmax, maximal rate of depolarization, APD, AP duration at different levels of repolarization (APD measured at 10% increments of Amp).

The effects of selective ion channel block were tested on cardiomyocytes with ventricular-like AP waveforms stimulated at a constant rate of 1 Hz. For control experiments, AP parameters were stable over a 15-min recording period in extracellular solution without (Fig. 3A, left) or containing 0.1% DMSO (Fig. 3B, left). Selective block of INa by TTX (22) had no effect on AP parameters at 1 μmol/l, whereas increasing concentrations (3–30 μmol/l) delayed the upstroke and reduced dV/dtmax (Fig. 3A, middle). Selective block of the IKr by E4031 (37) caused concentration-dependent (3–100 nmol/l) slowing of repolarization to “triangulate” the AP (Fig. 3A, right). Selective block of ICa by nifedipine (12) caused concentration-dependent (3–100 nmol/l) shortening of the AP throughout repolarization with minimal effects on the AP peak voltage and dV/dtmax (Fig. 3B, middle). Suppression of the IKs by 3R4S-chromanol 293B (47) had minimal effects on AP properties up to 10 μmol/l sufficient to cause >50% block of IKs (Fig. 3B, right). Data from the above experiments are summarized in Table 1, which shows measurements of AP Peak, MDP, APD (10, 50, and 90%), and dV/dtmax normalized to the initial control (predrug) value. For measurements of APD, the time of dV/dtmax was used as the start time for the AP. TTX caused slowing of dV/dtmax, E4031 caused lengthening of APD50 and APD90, and nifedipine caused shortening of APD10, APD50, and APD90.

Fig. 3.

Effects of ion channel blocking drugs on AP morphology. Data were recorded at 35–37°C with the perforated patch-clamp method. Cardiomyocytes in A and B were stimulated continuously at 1 Hz. A: effects of control extracellular solution after 3-, 6-, 9-, 12-, and 15-min intervals (left) and in different cardiomyocytes exposed to increasing concentrations of tetrodotoxin (TTX; middle) or E4031 (right). B: effects of control extracellular solution after 3 min followed by extracellular solution containing 0.1% DMSO at 6-, 9-, 12-, and 15-min intervals and in different cardiomyocytes exposed to increasing concentrations of nifedipine (middle) and 3R4S-chromanol 293B (right). C: APs were stimulated at 0.5 or 1 Hz in increasing concentrations of E4031 (10, 30, and 100 nmol/l). Pacing at 0.5 Hz in 100 nmol/l E4031 resulted in an early afterdepolarization (EAD; arrow). D: plot of EAD peak voltage vs EAD take-off potential with linear fit to the data. Inset: 2 representative EADs with different take-off potentials and peak voltages.

View this table:
Table 1.

Summary of drug-induced effects on evoked APs

Single EADs were induced by 30–100 nmol/l E4031 in three of five cardiomyocytes paced at 0.5 Hz. Figure 3C shows that lower concentrations of E4031 prolonged the AP and that 100 nmol/l induced single EADs. Similar to cardiac myocytes, EADs in hiPSC-derived cardiomyocytes were rate dependent and disappeared upon increasing the pacing rate from 0.5 to 1 Hz. Furthermore, there was an inverse relation between the EAD peak voltage and EAD take-off potential, as is shown in Fig. 3D for 79 single EADs from the same cardiomyocyte (inset shows 2 EADs with differing amplitudes and take-off potentials). In the three cardiomyocytes that generated EADs, the average EAD amplitude was 20.6 ± 1.1 mV and the EAD peak voltage to take-off potential relation, fit as a linear process, had a mean slope of −2.28 ± 0.11 mV/mV (peak voltage/take-off potential), and the take-off potential at which EAD amplitude became 0 was −27.0 ± 0.9 mV.

INa and ICa.

Recording INa at 35–37°C is technically challenging. To maintain voltage control during INa experiments, we reduced the amplitude of INa by decreasing the transmembrane Na+ gradient, maintaining a fraction of the channels in an inactivated state by not using very negative holding potentials, and by selecting very small cardiomyocytes for study. Nifedipine (1 μmol/l) was present to block ICa, and the ionic conditions are given in Supplemental Table S1. At a peak INa of −3 nA with 2 MΩ of uncompensated series resistance, the cell membrane potential would deviate from command potential by 6 mV. From a holding potential of −80 mV, 40-ms long depolarizing pulses to between −70 and 60 mV were applied in 10-mV increments. Figure 4A, left, shows representative INa traces. INa was first observed at −60 mV and increased to reach a maximal level at −20 mV. Further depolarization progressively decreased INa and at the most positive voltage steps INa became outward. The averaged peak INa current-voltage (I-V) plot (Fig. 4A, middle) shows a graded voltage response suggesting suitable voltage control. Maximal peak INa at −20 mV normalized to cell capacitance (mean 15.8 ± 1.5 pF) was −216.7 ± 18.7 pA/pF (n = 5). The reversal potential obtained by linear regression of the positive slope of the I-V relation was 42.8 mV, which is close to the Na+ reversal potential of 43.0 mV calculated with the Nernst equation. The activation relation is shown in Fig. 4A, right. When fit as a Boltzmann function, the half-activation voltage (V1/2) and slope factor (k) were −34.1 and 5.9 mV/e-fold change, respectively (n = 5). To measure INa steady-state inactivation, 400-ms long prepulses to between −110 and −20 mV in 10-mV increments followed by 40-ms long pulses of 0 mV were used to elicit INa. The current amplitude was normalized to the peak INa obtained with the −110-mV prepulse and fit as a Boltzmann function. The V1/2 and k values were −72.1 and 5.7 mV/e-fold change, respectively (Fig. 4A, right; n = 5). There was a small overlap of the activation and inactivation relations to give a very small Na+ window current.

Fig. 4.

Characteristics of sodium (INa) and L-type and calcium (ICa) currents. A: representative INa traces were elicited by the protocol shown in the inset (left). Peak INa density current density (I-V) plot is shown at middle. Steady-state inactivation and activation relations for INa are shown at right. B: representative ICa traces were elicited by the protocol shown in inset (left). Peak ICa density I-V plot is shown at middle. Steady-state inactivation and activation plots are shown at right. V1/2, half-activation voltage; k, slope factor.

ICa was recorded at room temperature to minimize current rundown. The extracellular solution contained 5 mmol/l Ca2+ along with TTX and was Na+- and K+-free (Supplemental Table S1). To prevent Ca2+ overload, the cardiomyocytes were first perfused with a Na+-, K+-, and Ca2+-free extracellular solution (54). The cardiomyocytes were held at −80 mV to ensure complete recovery from inactivation of ICa. A 3-s long prepulse was then applied to −50 mV to voltage inactivate Na+ and any T-type Ca2+ channels that might be present. This was followed by a 100-ms long test pulse to between −60 and 60 mV in 10-mV increments to activate ICa. Representative current traces are shown in Fig. 4B, left. The averaged peak I-V relation is plotted in Fig. 4B, middle (n = 5). ICa began to activate positive to −50 mV and the current density reached a peak of −17.1±1.7 pA/pF at 0 mV. Figure 4B, right, shows the normalized peak conductance-voltage (activation) relation derived from the peak I-V plot. Ca2+ conductance began to increase at −40 mV and reached a maximum at 10 mV. When fit with a Boltzmann function, the V1/2 and k values were −14.9 and 6.6 mV/e-fold change (n = 6), respectively. To measure ICa steady-state inactivation, 3-s long prepulses to between −50 and −10 mV in 5-mV increments were followed by 400-ms long pulses to 0 mV to elicit ICa. The current amplitude was normalized to the peak ICa obtained with the −50 mV prepulse and fit as a Boltzmann function. The V1/2 and k values were −29.1 mV and 4.9 mV/e-fold change, respectively (Fig. 4B, right; n = 6). Over the voltage range of −40 to −10 mV, there was overlap of the activation and inactivation relations consistent with a Ca2+ window current. The voltage dependence and properties of the ICa we studied suggest that we measured L-type (low voltage activated) Ca2+ current. In some experiments, we employed protocols at more negative voltages to search for T-type (high voltage activated) Ca2+ current (14) but did not find clear functional evidence to support its expression (data not shown).

IKr and IKs.

IKr was isolated as an E4031-sensitive current (37, 53). The ionic conditions are given in Supplemental Table S1. From a holding potential of −40 mV, test steps were applied for 4 s to between −40 to 15 mV in 5-mV increments (Fig. 5A). The data show families of current traces for control conditions (left), after the addition of E4031 (500 nmol/l; middle), and digitally subtracted traces to give the E4031-sensitive current or IKr (right). The E4031 concentration is sufficient to produce complete block of IKr. Data with E4031 were collected 2 min after drug application to minimize possible effects of IKr rundown. An E4031-sensitive current was present in all cardiomyocytes studied (n = 8). Figure 5C shows the I-V relation for IKr measured at the end of the depolarizing step (left). IKr activates at voltages positive to −40 mV and is maximal at approximately −10 mV, and it declines at more positive voltages consistent with inward rectification. The voltage-dependence of IKr peak tail current (Fig. 5C, middle) shows that it activates positive to −40 mV to reach a maximum by ∼0 to 10 mV. The peak tail current density is 0.95 ± 0.02 pA/pF. When fit as a Boltzmann function, the V1/2 and k values for IKr were −22.7 and 4.9 mV/e-fold change, respectively.

Fig. 5.

Characteristics of rapidly and slowly activating components of delayed rectifier potassium (IKr and IKs) currents at 35–37°C. A: representative current traces elicited by the protocol shown in inset. Left, middle, and right traces: control, E4031 (500 nmol/l), and E4031-subtraction, respectively. B: representative current traces elicited by the protocol shown in the inset. Left, middle, and right traces: control, 3R4S-chromanol 293B (10 μmol/l), and 3R4S-chromanol 293B-subtraction, respectively. C: I-V plots of IKr at the end of the depolarizing step (left), peak tail IKr normalized to maximal current following repolarization to −40 mV (middle), and I-V plot of IKs at the end of depolarizing step (right).

IKs was isolated as an 3R4S-chromanol 293B-sensitive current in the continuous presence of 500 nmol/l E4031 to block IKr. Experimental conditions are given in Supplemental Table S1. From a holding potential of −40 mV, test steps were applied to −20, 0, 20, and 40 mV for 5 s. Figure 5B shows families of control current traces (left), 3R4S-chromanol 293B exposure (10 μmol/l; middle), and the digitally subtracted 3R4S-chromanol 293B-sensitive current (right). A 3R4S-chromanol 293B-sensitive current was detected in 5 of 16 cardiomyocytes studied with this protocol. In these five cardiomyocytes, depolarization resulted in activation of a voltage- and time-dependent outward current and at 40 mV the 3R4S-chromanol 293B-sensitive current at the end of the step was 0.31±0.09 pA/pF. Figure 5C, right, shows the I-V relation for activation of IKs.

If, IK1, and Ito.

If undergoes time-dependent activation at hyperpolarizing voltages. If was elicited from a holding potential of −40 mV by 2-s long hyperpolarizing test pulses to between −50 and −120 mV applied in 10-mV increments. BaCl2 (500 μmol/l) was included in the extracellular solution to inhibit IK1 (27) (Supplemental Table S1). Current traces obtained with this protocol are shown in Fig. 6A (left). Beginning at −60 mV, a slowly activating inward current was observed. The current density at the end of each voltage step was normalized to the maximal current density at −120 mV to generate an activation relation (Fig. 6A, right; n = 17). When fit with a Boltzmann function, the V1/2 and k values for If were −84.6 and 8.8 mV/e-fold change, respectively, and the average If density was −4.1±0.3 pA/pF at −120 mV. The I-V plot in Fig. 6B, left (control), shows average peak If at each voltage and demonstrates its inward rectification properties. Figure 6B also shows the effect of 5 mmol/l CsCl, which blocks If (35) in the same cardiomyocytes. We also studied the drug ZD7288, which causes voltage-dependent block of If (3). Averaged data in a different series of cardiomyocytes are shown in Fig. 6B, right (n = 4). The control data again show inward rectification at negative voltages and that ZD7288 (100 μmol/l) inhibited If. The inhibition of ZD7288 at −120 mV (62.8 ± 14.6%) was significantly smaller than the inhibition at −90 mV (99.6 ± 3.9%; n = 4; P < 0.01) consistent with the voltage-dependent blocking properties of ZD7288.

Fig. 6.

Characteristics of hyperpolarization-activated pacemaker (If), transient outward potassium (Ito), and inward rectifier potassium (IK1) currents at 35–37°C. A: representative recording of If (left) and the activation relation for peak If (If/If max, right) with the protocol shown in inset. If max was the current measured at −120 mV. B: I-V relation of peak If in the absence and presence of 5 mmol/l Cs+ (left) and in the absence and presence of 100 μmol/l ZD7288 (right). C: averaged current traces of IK1 elicited by the voltage-ramp protocol shown in inset recorded before and after 500 μmol/l Ba2+ blockade and the Ba2+-sensitive current normalized to cell capacitance. D: representative Ito traces elicited by protocol shown in inset (left) and the I-V plot for peak Ito normalized to cell capacitance (right).

IK1 contributes to the maintenance of the cardiac resting potential and participates in phase 3 of AP repolarization, and it becomes minimal at positive voltages due to its rectification properties. We measured IK1 in Na+- and Ca2+-free extracellular solution that contained 100 nmol/l E4031 and 10 μmol/l nifedipine to block IKr and ICa (Supplemental Table S1). To record IK1 from a holding potential of −73 mV, whole cell current was elicited by a 2-s long voltage ramp applied from −123 to 17 mV as illustrated in Fig. 6C. IK1 was identified as a Ba2+-sensitive current. The I-V plot shows averaged current traces (n = 6) recorded with the voltage ramp before and after the addition of BaCl2 for 3 min. The dashed line shows the digitally subtracted, Ba2+-sensitive current. The IK1 reversal potential is close to −80 mV, and at voltages positive to −35 mV it has a negative slope conductance consistent with inward rectification. The average current density at −123 mV (peak inward) and −35 mV (peak outward) is −2.3 ± 0.6 and 1.0 ± 0.2 pA/pF, respectively.

Ito regulates phase 1 repolarization in mammalian heart to modulate the early AP plateau (31). We studied Ito in Na+- and Ca2+-free extracellular solution that contained 100 nmol/l E4031, 10 μmol/l nifedipine, and 10 μmol/l TTX to suppress IKr, ICa, and INa (Supplemental Table S1). Figure 6D, left, shows a representative family of current traces elicited from a holding potential of −50 mV with 400-ms long depolarizing test pulses to between −40 and 60 mV applied in 10-mV increments every 4 s. Outward current activates within milliseconds of depolarization and then decays over a few hundred milliseconds consistent with Ito to then reach a small nearly steady component of current. The I-V plot in Fig. 6D, right, shows averaged peak Ito measured as the difference between the peak current and the current at the end of the voltage step (n = 8). Ito begins to activate at voltages positive to −40 mV and shows outward rectification with increasing depolarization. Ito density at 60 mV was 2.5 ± 0.3 pA/pF.

Planar automated patch clamp.

A limitation of existing hESC and hiPSC-derived cardiomyocyte differentiation models is that small numbers of cardiomyocytes are produced and contaminating cell types often are present in large numbers. In the present experiments, we generated large numbers of essentially pure cultured of cardiomyocytes (>1.5 × 106 per cryopreserved vial) that have the experimental advantages of ease of use and their application to higher throughput systems. To assess this, we also studied the hiPSC-derived cardiomyocytes using a planar automated patch-clamp system to analyze INa, ICa, and IKr. Experiments were performed at room temperature. An acceptable whole cell recording was defined as having a membrane resistance (Rm) ≥200 MΩ with an access resistance (Ra) ≤15 MΩ for ≥15 min. Success rates for single SealChips were variable but could reach ∼50% with acceptable recordings obtained in 58 wells. For voltage-clamp experiments, the average Rm was 617 ± 53 MΩ, Ra was 6.7 ± 0.4 MΩ, and cell capacitance was 35.0 ± 2.5 pF.

Example currents and summarized data are shown in Fig. 7. For each current, cardiomyocytes were bathed in specific external ionic solutions (Supplemental Table S1). Figure 7A shows example INa recordings obtained using the voltage protocol shown in the inset. INa could be resolved from the capacitance spike at voltages negative to the reversal potential. The I-V plot shows averaged peak INa with a graded response to voltage. Inset shows the concentration-response relation of peak INa to TTX with an IC50 of 0.64 μmol/l. Figure 7B shows example ICa recordings obtained using the voltage protocol shown in the inset. The bath solution was Na+ free (Supplemental Table S1). The I-V plot shows averaged peak ICa with a graded response to depolarization. Inset shows the concentration-response relation of peak ICa to nifedipine with an IC50 of 0.038 μmol/. Figure 7C, top, shows example IKr recordings and the effect of E4031. IKr amplitude is reduced at room temperature, and the current can be difficult to resolve; therefore, the extracellular K concentration was 140 mM to reverse the normal K+ transmembrane gradient, and IKr was measured as an inward tail current (Supplemental Table S1). From a holding potential of −80 mV, a prepulse was applied to 20 mV for 1 s before the cardiomyocyte was repolarized to −110 mV for 500 ms to generate a large amplitude inward tail current (arrow) that gradually decayed. The pulse sequence was applied every 5 s, and the peak tail current amplitude (measured as the difference between peak tail current and the current at the end of the step) was plotted during the experiment (middle). Application of 1 μmol/l E4031 resulted in its complete block, consistent with IKr. The I-V plot (bottom) shows the effect of different prepulse voltages (−120 to 60 mV in 10-mV increments) on peak tail IKr recorded at −120 mV. Control tail current was recorded followed by application of 1 μmol/l E4031, and the E4031-sensitive current was plotted. IKr activates at voltages positive to −50 mV and reaches a maximal value at 10 mV. When fit with as a Boltzmann function, the V1/2 and k values are −17.5 mV and 9.4 mV/e-fold change, respectively.

Fig. 7.

Planar-automated patch clamp. A: representative traces for INa are shown above the I-V plot for peak INa. Inset: concentration-response relation for block by TTX. B: representative traces for ICa are shown above the I-V plot for peak ICa. Inset: concentration-response relation for block by nifedipine. C: representative traces for IKr (arrow) and the voltage protocol are shown during E4031 wash-in (top). Plot of peak tail IKr (middle) shows the rapid development of block with E4031 (arrow). Activation relation for IKr is shown at bottom.


There are several important advances in these experiments. One is that we generated essentially pure cardiomyocyte cultures by engineering blasticidin-resistance gene expression controlled from the cardiac-specific endogenous MYH6 promoter into the parental hiPSC clone.(24) With differentiation of the hiPSCs, this approach allows for the selection of cardiomyocytes with a high purity of >98% that was maintained for ≥2 wk in culture postthaw. Furthermore, the cardiomyocytes were generated using a differentiation approach that avoids manual dissection of embryoid bodies, and this has the advantage of high volume cardiomyocyte production and they can be cyropreserved to facilitate their ease of use. Many research applications, including biochemical assays, gene expression, tissue and arrhythmia modeling, and regenerative therapies, as well as high capacity applications for drug discovery and safety pharmacology, require high purity and large cardiomyocyte numbers that can now be easily achieved. As proof of principle, we studied hiPSC-derived cardiomyocytes using a planar automated patch-clamp analysis. INa, ICa, and IKr were present as TTX-, nifedipine-, and E4031-sensitive currents, respectively. For INa and ICa, peak current densities were lower than those obtained with single cell patch-clamp experiments, whereas for IKr the peak current density was greater, and cell capacitance was reduced. These differences are likely to reflect changes in the experimental conditions and protocols between automated planar and conventional patch-clamp methods such as differences in ionic conditions, temperature, and cellular characteristics of cardiomyocytes studied in the two voltage clamp systems. The experiments provide direct evidence that the hiPSC-derived cardiomyocytes can be studied using high throughput experimental approaches requiring high purity and large cardiomyocyte numbers.

The cardiomyocytes in these experiments were cultured for up to 21 days postthaw and beat spontaneously as either single cardiomyocytes or synchronized monolayers. Current clamp experiments revealed atrial-, nodal-, and ventricular-like APs, and we developed quantitative criteria to assist in their characterization. The presence of atrial-, nodal- and ventricular-like APs agrees with previous reports from both hiPSC- and hESC-derived cardiomyocytes (13, 19, 28, 30, 34, 52). However, the reported percentage of each AP type is highly dependent on the classification criteria (see for example Ref. 34). Compared with some reports with hESC-derived cardiomyocytes, the hiPSC-derived cardiomyocytes we studied have more negative MDP values for atrial- and ventricular-like APs, and the ventricular-like AP durations are within the normal range of the human electrocardiographic QT interval; thus the cardiomyocytes have some properties of more mature human cardiac myocytes. This is supported by the cell capacitance measurements obtained with conventional ruptured whole cell patch clamp that are larger than previous reports with hESC-derived cardiomyocytes (39, 55). Cell capacitance measurements, while variable among isolated mammalian cardiomyocytes, increase with cell maturity (40). At the same time, these cardiomyocytes beat spontaneously; thus they retain pacemaking properties. Many factors may contribute to the variation and maturity of atrial-, nodal- and ventricular-like APs including cardiomyocyte differentiation protocols, culture conditions of the differentiated cardiomyocytes, and recording procedures. Also, in the present study we used a perforated patch method to record APs that minimizes cell dialysis while other reports used ruptured patch or sharp microelectrode recording techniques.

The ventricular-like APs responded to IKr suppression by prolonging APD and generating EADs. Our findings are similar to previous reports in hiPSC- and hESC-derived cardiomyocytes. (13, 16, 32, 34). IKr block first slowed terminal AP repolarization to cause “triangulation,” and this preceded EADs. The EADs were rate (bradycardia) dependent, and EAD amplitudes depended on the EAD take-off potential.(6) The slope of the EAD peak voltage to take-off potential relation averaged −2.28 ± 0.11 mV/mV (peak voltage/take-off potential) and the take-off potential at which EAD amplitude became 0 was −27.0 ± 0.9 mV. Previously, this relation was reported in adult canine cardiac Purkinje fiber short segments to be −1.87 ± 0.26 mV/mV, and the voltage at which the EAD amplitude became 0 was −25.8 ± 2.3 mV (18). Therefore, another important finding is that the EADs in our hiPSC-derived cardiomyocytes exhibit electrophysiological properties nearly identical with those seen in native canine heart cells. Thus the complex ion channel mechanisms, including the arrhythmogenic L-type Ca2+ window current required to generate EADs, are intact in these hiPSC-derived cardiomyocytes.(10, 17, 42, 51).

We compared our hiPSC-derived cardiomyocytes findings with published data from human cardiac myocytes, as summarized in Supplemental Table S2. This comparison is difficult, as reported ionic current properties for human cardiac myocytes are heavily dependent on individual experimental conditions and protocols, as well as tissue heterogeneity and disease status. An additional limitation is that for our voltage clamp experiments, which were performed with the ruptured patch method, we did not record APs to identify whether cardiomyocytes were atrial, nodal, or ventricular like. The largest current densities in our hiPSC-derived cardiomyocytes were for INa followed by ICa. For INa, the current density of −216.7±18.7 pA/pF underestimates the maximal density, as the recording protocol did not fully activate the channels and we used a reduced Na+ concentration gradient to maintain adequate voltage control. The steady-state inactivation V1/2 value of −72.1 mV we obtained is more positive than some previous reports, which may result from our not using very negative holding potentials to fully activate the channels. INa gating properties are similar to those previously reported for human cardiac myocyte Na+ channels.(9, 36) as well as from hESC-derived cardiomyocytes (39). For ICa, the current density of −17.1±1.7 pA/pF is larger than previous reports for human cardiac myocytes. (33, 44) In the present study, ICa was recorded with 5 mmol/l Ca2+ in the extracellular solution, while ICa densities for human cardiac myocytes were recorded with 2 mmol/l Ca2+. The activation and inactivation gating properties of ICa in hiPSC-derived cardiomyocytes are similar to those obtained from human ventricular myocytes and hESC-derived cardiomyocytes (33, 38).

Delayed rectifier K+ channel currents in human cardiac myocytes are reported to be of small amplitude and in some cases have not been detected (2, 15, 21, 25, 41, 4546), and most quantitative data have been obtained from animal and heterologous expression models. In our hiPSC-derived cardiomyocytes, the current density and activation properties for IKr are close to values reported for human cardiac myocytes (9, 20). For IKs, there is one report of a maximal current density of ∼0.18 pA/pF. In these experiments Viraq et al. (46) were able to identify IKs in only ∼50% of the cardiac myocytes they studied (see also Ref. 21). Recently, Moretti et al. (29) demonstrated that both IKr and IKs were present in hiPSC-derived cardiomyocytes. In the present study, IKr was detected in all cardiomyocytes, whereas, IKs was detected in 5 of 16 cardiomyocytes. IKr peak tail current following the activation voltage step was approximately twofold larger than peak current during the depolarizing step. In cardiomyocytes with IKs, current density was less than that found for IKr, IKs tail current was small compared with current at the end of the depolarizing step, and IKs activated at more positive voltages than IKr (37). Thus the currents have the gating properties expected for IKr and IKs. A potential limitation to these experiments is that chromanol 293B and its enatiomers lose selectivity for IKs over other K+ channels at higher drug concentrations, and for this reason we did not study higher (>10 μmol/l) drug concentrations (47). We also found an Ito-like current in hiPSC-derived cardiomyocytes; however, our studies do not distinguish Ito-fast from Ito,slow (31).

The hiPSC-derived cardiomyocytes we studied exhibited phase 4 depolarization and beat spontaneously. This prompted us to study currents important to maintenance of the resting potential and pacemaking. If undergoes time-dependent activation at voltages negative to −60 mV and its gating properties closely resemble those previously reported in human cardiac myocytes (Supplemental Table S2). A new finding is that IK1 is present, which has not been shown previously in hESC- or hiPSC-derived cardiomyocytes. The presence of IK1 is likely to contribute to the more negative MDP values we found, but the robust If prevents complete repolarization to the resting potential and serves to promote automaticity.

In conclusion, this is the first detailed quantitative analysis of hiPSC-derived cardiomyocyte electrophysiological properties including of seven ionic currents from the same clonal line. We found atrial-, nodal-, and ventricular-like APs and developed quantitative criteria for their classification. hiPSC-derived cardiomyocytes with ventricular-like APs had many properties characteristic of human cardiac myocytes (a more negative MDP, APD measurements similar to the normal human QT interval, EAD properties similar to mature cardiac myocytes) but also retain immature properties (spontaneous beating and robust If). For individual ion channel currents, the voltage dependence of channel gating properties are remarkably similar to those found in human cardiac myocytes. These findings support the conclusion that hiPSC-derived cardiomyocytes faithfully reproduce cardiac electrophysiological properties of human cardiac myocytes. These hiPSC-derived cardiomyocytes have the added advantage that they can be grown in large numbers and high purity, and the hiPSC-derived cardiomyocytes can be cyropreserved to facilitate their ease of use. This work has the potential to impact many areas of cardiovascular investigation and drug safety.


J. A. Thomson, T. J. Kamp, and C. T. January are cofounders of Cellular Dynamics International.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
View Abstract