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

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 I_Na 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 I_Na or acquired at 1 kHz and low-pass filtered at 0.2 kHz for I_Ca or I_Kr with PatchXpress Commander software (Version 1.6.0.36d; Molecular Devices). Recordings were obtained at ambient temperature.

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/dt_max), 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/dt_max 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 (APD_30–40) and the time difference between APD₇₀ to APD₈₀ (APD_70–80). Ventricular-like APs had a ratio (APD_30–40/APD_70–80) 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 (APD_30–40/APD_70–80 close to unity) with shorter durations and higher spontaneous beating rates. Nodal-like APs had less negative MDPs and smaller amplitudes with lower dV/dt_max 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.

Open in a separate window

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/dt_max, 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 I_Na 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/dt_max (Fig. 3A, middle). Selective block of the I_Kr by E4031 (37) caused concentration-dependent (3–100 nmol/l) slowing of repolarization to “triangulate” the AP (Fig. 3A, right). Selective block of I_Ca 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/dt_max (Fig. 3B, middle). Suppression of the I_Ks by 3R4S-chromanol 293B (47) had minimal effects on AP properties up to 10 μmol/l sufficient to cause >50% block of I_Ks (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/dt_max normalized to the initial control (predrug) value. For measurements of APD, the time of dV/dt_max was used as the start time for the AP. TTX caused slowing of dV/dt_max, E4031 caused lengthening of APD₅₀ and APD₉₀, and nifedipine caused shortening of APD₁₀, APD₅₀, and APD₉₀.

Open in a separate window

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.

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.

I_Na and I_Ca.

Recording I_Na at 35–37°C is technically challenging. To maintain voltage control during I_Na experiments, we reduced the amplitude of I_Na 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 I_Ca, and the ionic conditions are given in Supplemental Table S1. At a peak I_Na 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 I_Na traces. I_Na was first observed at −60 mV and increased to reach a maximal level at −20 mV. Further depolarization progressively decreased I_Na and at the most positive voltage steps I_Na became outward. The averaged peak I_Na current-voltage (I-V) plot (Fig. 4A, middle) shows a graded voltage response suggesting suitable voltage control. Maximal peak I_Na 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 (V_1/2) and slope factor (k) were −34.1 and 5.9 mV/e-fold change, respectively (n = 5). To measure I_Na 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 I_Na. The current amplitude was normalized to the peak I_Na obtained with the −110-mV prepulse and fit as a Boltzmann function. The V_1/2 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.

Open in a separate window

Fig. 4.

Characteristics of sodium (I_Na) and L-type and calcium (I_Ca) currents. A: representative I_Na traces were elicited by the protocol shown in the inset (left). Peak I_Na density current density (I-V) plot is shown at middle. Steady-state inactivation and activation relations for I_Na are shown at right. B: representative I_Ca traces were elicited by the protocol shown in inset (left). Peak I_Ca density I-V plot is shown at middle. Steady-state inactivation and activation plots are shown at right. V_1/2, half-activation voltage; k, slope factor.

I_Ca 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 I_Ca. 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 I_Ca. Representative current traces are shown in Fig. 4B, left. The averaged peak I-V relation is plotted in Fig. 4B, middle (n = 5). I_Ca 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 V_1/2 and k values were −14.9 and 6.6 mV/e-fold change (n = 6), respectively. To measure I_Ca 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 I_Ca. The current amplitude was normalized to the peak I_Ca obtained with the −50 mV prepulse and fit as a Boltzmann function. The V_1/2 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 I_Ca 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).

I_Kr and I_Ks.

I_Kr 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 I_Kr (right). The E4031 concentration is sufficient to produce complete block of I_Kr. Data with E4031 were collected 2 min after drug application to minimize possible effects of I_Kr rundown. An E4031-sensitive current was present in all cardiomyocytes studied (n = 8). Figure 5C shows the I-V relation for I_Kr measured at the end of the depolarizing step (left). I_Kr 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 I_Kr 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 V_1/2 and k values for I_Kr were −22.7 and 4.9 mV/e-fold change, respectively.

Open in a separate window

Fig. 5.

Characteristics of rapidly and slowly activating components of delayed rectifier potassium (I_Kr and I_Ks) 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 I_Kr at the end of the depolarizing step (left), peak tail I_Kr normalized to maximal current following repolarization to −40 mV (middle), and I-V plot of I_Ks at the end of depolarizing step (right).

I_Ks was isolated as an 3R4S-chromanol 293B-sensitive current in the continuous presence of 500 nmol/l E4031 to block I_Kr. 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 I_Ks.

I_f, I_K1, and I_to.

I_f undergoes time-dependent activation at hyperpolarizing voltages. I_f 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 I_K1 (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 V_1/2 and k values for I_f were −84.6 and 8.8 mV/e-fold change, respectively, and the average I_f density was −4.1±0.3 pA/pF at −120 mV. The I-V plot in Fig. 6B, left (control), shows average peak I_f at each voltage and demonstrates its inward rectification properties. Figure 6B also shows the effect of 5 mmol/l CsCl, which blocks I_f (35) in the same cardiomyocytes. We also studied the drug ZD7288, which causes voltage-dependent block of I_f (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 I_f. 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.

Open in a separate window

Fig. 6.

Characteristics of hyperpolarization-activated pacemaker (I_f), transient outward potassium (I_to), and inward rectifier potassium (I_K1) currents at 35–37°C. A: representative recording of I_f (left) and the activation relation for peak I_f (I_f/I_{f max,} right) with the protocol shown in inset. I_{f max} was the current measured at −120 mV. B: I-V relation of peak I_f 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 I_K1 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 I_to traces elicited by protocol shown in inset (left) and the I-V plot for peak I_to normalized to cell capacitance (right).

I_K1 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 I_K1 in Na⁺- and Ca²⁺-free extracellular solution that contained 100 nmol/l E4031 and 10 μmol/l nifedipine to block I_Kr and I_Ca (Supplemental Table S1). To record I_K1 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. I_K1 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 I_K1 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.

I_to regulates phase 1 repolarization in mammalian heart to modulate the early AP plateau (31). We studied I_to in Na⁺- and Ca²⁺-free extracellular solution that contained 100 nmol/l E4031, 10 μmol/l nifedipine, and 10 μmol/l TTX to suppress I_Kr, I_Ca, and I_Na (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 I_to to then reach a small nearly steady component of current. The I-V plot in Fig. 6D, right, shows averaged peak I_to measured as the difference between the peak current and the current at the end of the voltage step (n = 8). I_to begins to activate at voltages positive to −40 mV and shows outward rectification with increasing depolarization. I_to density at 60 mV was 2.5 ± 0.3 pA/pF.