Regionally diverse mitochondrial calcium signaling regulates spontaneous pacing in developing cardiomyocytes

Abstract

The quintessential property of developing cardiomyocytes is their ability to beat spontaneously. The mechanisms underlying spontaneous beating in developing cardiomyocytes are thought to resemble those of adult heart, but have not been directly tested. Contributions of sarcoplasmic and mitochondrial Ca²⁺-signaling vs. I_f-channel in initiating spontaneous beating were tested in human induced Pluripotent Stem cell-derived cardiomyocytes (hiPS-CM) and rat Neonatal cardiomyocytes (rN-CM). Whole-cell and perforated-patch voltage-clamping and 2-D confocal imaging showed: 1) Both cell types beat spontaneously (60-140/min, at 24°C); 2) Holding potentials between −70 to 0mV had no significant effects on spontaneous pacing, but suppressed action potential formation; 3) Spontaneous pacing at −50mV activated cytosolic Ca²⁺-transients, accompanied by in-phase inward current oscillations that were suppressed by Na⁺-Ca²⁺-exchanger (NCX)- and ryanodine receptor (RyR2)-blockers, but not by Ca²⁺- and I_f-channels blockers; 4) Spreading fluorescence images of cytosolic Ca²⁺-transients emanated repeatedly from preferred central cellular locations during spontaneous beating; 5) Mitochondrial un-coupler, FCCP at non-depolarizing concentrations (~50nM), reversibly suppressed spontaneous pacing; 6) Genetically encoded mitochondrial Ca²⁺-biosensor (mitycam-E31Q) detected regionally diverse, and FCCP-sensitive mitochondrial Ca²⁺-uptake and release signals activating during I_NCX oscillations; 7) I_f -channel was absent in rN-CM, but activated only negative to −80mV in hiPS-CM; nevertheless blockers of I_f-channel failed to alter spontaneous pacing.

Graphical Abstract

1. Introduction

In studies exploring the mechanisms of cardiac pacemaking much attention has been focused on processes that underlie the slow diastolic depolarization, that in SA-nodal cells start at −60 mV and terminates with upstroke of the action potential mediated by Ca²⁺ current. The primary candidates for slow diastolic depolarization have been the activation of: 1) inward I_f current (HCN2 and HCN4 genes [1-5]); 2) inward Na⁺-Ca²⁺ exchanger current (I_NCX) activated by release of Ca²⁺ from ryanodine receptor (RyR2) [6-8]; 3) turnoff of K⁺ current in Purkinje fibers [9-11] and activation of Na⁺-K⁺ ATPase or leak-type inward current [12, 13].

In considering the mechanisms of pacing, it is possible that one process may dominate in sino-atrial pacing while other mechanisms become critical in Purkinje fibers, [9-11] or in developing myocytes and pathological conditions. But, it is also plausible that the robust, persistent heartbeat requires interactions of number of entrained Ca²⁺- and membrane potential-dependent oscillators [14], which may be modulated by mechanical stimuli [15, 16], mitochondrial Ca²⁺ fluxes, [17, 18], and oxidative stress [19].

We chose to examine spontaneous pacing in developing myocytes by quantifying their pacing frequency under control conditions and when action potential generation were prevented by voltage-clamping the cells at fixed holding potential of −50mV, allowing pacing to be monitored through cell shortening, I_NCX oscillations, and cytosolic and mitochondrial Ca²⁺-transients. After finding that generation of action potentials had little effect on pacing frequency, we proceeded to probe the role in Ca²⁺ signaling at −50mV to determine: 1) How robust were the Ca²⁺ oscillations in rN-CM and hiPS-CM; 2) How did they compare with SA-nodal Ca²⁺ oscillations, reported to rise at random sub sarcolemmal locations and be suppressed by clamping the membrane potential [14, 20], and 3) What were the contributions of mitochondria, where Ca²⁺ fluxes throughmitochondrial Ca²⁺-uniporter and the Na⁺-Ca²⁺-exchanger are suggested to buffer cytosolic Ca²⁺, thereby modulating pacing frequency [17, 18]. Others have suggested, however, that mitochondrial Ca²⁺ fluxes are too small to have significant effects on cellular Ca²⁺ signaling physiologically [21].

To address these issues we used two models of developing myocytes, rN-CM and hiPS-CM. Compared to adult rodent cardiomyocytes, neonatal myocytes are reported to have altered functional protein expression (increased I_NCX and T-type Ca²⁺ current and decreased I_K1, I_Ks, and I_Kr etc. [22], which might predispose them to spontaneous pacing. In general, hiPS-CMs have an immature structure (rounded or polygonal shape, few myofibrils and z-lines) and electrophysiologically express lower levels of I_K1 and I_to. Most cell-line colonies appear to contain myocytes with ventricular, atrial and nodal action potential phenotypes. In addition, variability between different hiPS-CM cell-lines is even larger driving the ongoing effort to guide cellular differentiation [23]. We have previously reported on the hiPS-CMs used in the present study and found their Ca²⁺-signaling properties to be stable and dominated by I_Ca-triggered release of Ca²⁺ from the SR [24, 25, 26]. While both cell types continued to beat spontaneously under a broad range of experimental conditions, the critical role of subcellular Ca²⁺ signals was seen most clearly when they were dialyzed with Fluo-4, or infected with genetic probes targeted to RyR2 (GCamP6-FKBP [27]) or cytochrome-C (mitycam-E31Q [28]), and were voltage-clamped at −50mV to prevent the activation of any voltage-gated ion channels, but allowed activation of currents (I_NCX) generated by release of Ca²⁺.

To suppress motion-induced imaging artifacts or possible mechanical feedback mechanisms, we used firmly attached cultured myocytes and, in addition, incubated the cells with blebbistatin to desensitize the myofilaments to Ca²⁺ [29]. The voltage-clamped myocytes were also dialyzed with 5mM ATP to prevent exhausting them metabolically or when the mitochondrial un-coupler FCCP was used at concentrations (~50nM) that only stimulates mitochondrial oxidative metabolism, without depolarizing the mitochondria [30].

This is the first study to compare the details of Ca²⁺ signaling in two different spontaneously beating developing cell-types. The data suggests that spontaneous beating in hiPS-CM or rN-CM is triggered by rhythmic Ca²⁺ oscillations that activate in-phase NCX-currents, are independent of I_f expression, brief (~1min) suppression of I_Ca, or changes (~20s) in holding potential, and point to cross-talk between SR and mitochondria as a critical pacing mechanism. Imaging showed no indication that cellular Ca²⁺ transients start as random mutually reinforcing local sub-sarcolemmal Ca²⁺ releases. Rather, they rose from preferred central location and spread throughout the cell in a repeatable pattern. Mitycam-E31Q signals confirmed the presence of substantial mitochondrial Ca²⁺ transients, but also showed unexpected regional diversity between peripheral mitochondria that sequestered Ca²⁺ (Cf. [18, 30]), and peri-nuclear mitochondria that released Ca²⁺ accompanying the cytosolic Ca²⁺ transients.

2. Methods

2.1 General experimental approach

Experiments with spontaneously beating hiPS-CM [25, 26] and rN-CM [26, 31] cultures were carried out in accordance with national and institutional guidelines. The beating was examined at 24 and 35°C in intact cells and in single cells that were voltage- or current-clamped in configurations where the membrane under the patch pipette was either subjected to amphotericin B perforation or ruptured to allow cell dialysis. Holding potentials of −50 or −60mV were used to measure spontaneous oscillations in membrane current I_NCX, without activating other voltage-dependent channels, I_f and I_ca. Ca²⁺ oscillations were recorded fluorometrically using dialyzing solutions with 0.1mM Fluo-4 or transient expression of either FKBP-linked GCamP6 [27] or a novel mitochondrially-targeted probe (mitycam-E31Q [28]).

2.2. Neonatal cardiomyocyte (rN-CM) isolation

Rat neonatal CMs (rN-CM) were isolated using an isolation kit from Worthington Biochemical Corporation (Lakewood, NJ 08701). One to six day-old neonatal rats were decapitated and the beating hearts were surgically removed and placed in chilled Hank's Balanced Salt Solution (HBSS). The main vessels and atria were removed and the ventricles were minced with a razor blade to pieces <1mm³ that were incubated in HBSS with trypsin (50μg/ml) for 14-16h at 4°C. The digestion was then arrested by exposure to trypsin inhibitor (200μg/ml) for 20min in 37°C. Thereafter collagenase (100U/ml) was used for 30min to isolate single rN-CM, which were filtered through a cell strainer and centrifuged at 1000rpm for 3min. Cells were re-suspended in Dulbecco's Modified Eagle's Medium (DMEM) containing 10% fetal bovine serum (FBS) with 1% penicillin/streptomycin and 1% non-essential amino acids, plated on 100mm dishes and placed in the incubator for 1-1.5h to eliminate fibroblasts. rN-CM overall viability was ~80%. Isolated single rN-CM were plated onto non-treated glass cover slips and used for electrophysiological experiments.

2.3. Cultivation of hiPS cells and preparation of hiPS-CM

Human iPS-CMs were produced by transfecting somatic cells from a healthy control individual with a set of pluripotency genes (OCT4, SOX2, cMYC, and KLF4) [32]. The generation, cardiac differentiation and characterization of these cells have been reported [24-26].

2.3.1. Culture of undifferentiated human induced pluripotent stem (hiPS) cells

The hiPS cells were maintained on mitomycin C-treated murine embryonic fibroblasts (MEF) prepared in our laboratory in DMEM/F12 medium supplemented with Glutamax, 20% knockout serum replacer, 1% nonessential amino acids (NAA), 0.1mM β-mercaptoethanol (βME, Invitrogen, Darmstadt, Germany), 50ng/ml FGF-2 (PeproTech, Hamburg, Germany). Cells were passaged by manual dissection of hiPS cell colonies every 5-6 days.

2.2.2. Cardiac differentiation

Cardiac differentiation of hiPS cells was carried out on the murine visceral endoderm-like cell line END2, which was provided by C. Mummery (Leiden University Medical Center, The Netherlands) [33]. END2 cells were mitotically inactivated for 3 hours with 10μg/ml mitomycin C (Sigma-Aldrich Chemie GmbH, Munich, Germany) and 1.2×10⁶ cells were plated on 6cm dishes coated with 0.1% gelatin one day before initiation of hiPS cell differentiation. To initiate co-cultures, hiPS cell colonies were dissociated into clumps by using collagenase IV (Invitrogen, 1 mg/ml in DMEM/F-12 at 37°C for 5-10 minutes). The differentiation was carried out in knockout-DMEM containing 1mM L-glutamine, 1% NAA, 0.1mM βME and 1% penicillin/streptomycin (100U/ml and 100μg/ml, respectively). The co-culture was left undisturbed at 37°C/5% CO₂ for 5 days. Medium change was performed first on day 5 and later on days 9, 12 and 15 of differentiation.

2.3.3. Preparation of hiPS-CM for patch-clamp experiments

Beating areas were micro-dissected mechanically at day 40-45 of differentiation and dissociated with collagenase B. Single cardiomyocytes derived from hiPS cells (hiPS-CM) were then plated on fibronectin (2.5μg/ml) -coated 12 or 25mm glass cover slips and incubated for 36-72h before their use in electrophysiological experiments.

2.3.4. Cardiomyocytes derived from mouse iPS cells

(miPSC-CM) were prepared by methods described previously [34, 35] and were used in experiments probing the pacemaker current, I_f (Fig. 7). This iPS cell line (cloneAT25) was genetically modified to express EGFP and puromycin resistance genes under the control of α-myosin heavy chain (α-MHC) promoter to enable visualization and generation of pure CM. For the purpose of this study, murine iPS-CMs were not subjected to antibiotic selection and were analyzed between days 35-40 of differentiation.

Figure 7.

I_f and its lack of effect on pacing at fixed membrane potential. A: Voltage-dependence of I_f in mouse (miPS-CM) and human (hiPS-CM) stem cell-derived cardiomyocytes. B: Average I_f at −110mV is larger in mouse than in hiPS-CM, and is virtually absent in rN-CM. C: Sample traces of membrane current in the three cell types during 2.5s hyperpolarizations from −60mV to potentials from −70 to −130mV. D: Sample recordings with the same protocol in hPS-CM showing block of I_f by 2mM Cs⁺. E & F: Addition of 10 μM ivabradine to suppress I_f had no effect on oscillations of I_NCX in hiPS-CM (E) or rN-CM (F). The cells were voltage-clamped at 24°C and dialyzed with a solution containing 100μM EGTA.

2.4. Voltage-clamp procedures

Cells were voltage-clamped in the whole-cell configuration at −50 or −60mV tomeasure spontaneous oscillations in membrane current using a Dagan amplifier and pClamp software (Clampex 10.2). Action potentials were recorded in the current-clamp mode. Borosilicate patch pipettes were prepared using a horizontal pipette puller (Model P-87, Sutter Instruments, CA). The pipettes had a resistance of 3–5MΩ. In most experiments the membrane under the patch pipette was ruptured and the cells dialyzed with Ca²⁺-buffered pipette solutions containing (in mM): 110 Cs-aspartate, 15 NaCl, 20 TeaCl, 5 Mg-ATP, 0.1 EGTA with or without 0.05 or 0.1 Fluo-4, 0.1 CaCl₂ ([Ca²⁺]_i ≅ 100nM), 10 glucose and 10 HEPES (titrated to pH 7.2 with CsOH). In some experiments the cells were examined with the perforated patch technique using pipette solutions containing (in mM): 145 Cs-Glutamate, 9 NaCl, 1 MgCl₂, 10 HEPES (titrated to pH 7.2 with CsOH), and amphotericin B (Fisher Scientific) 0.69mg/ml. The extracellular solution used during experiments contained (in mM): 137 NaCl, 5.4 KCl, 2 CaCl₂, 1 MgCl₂, 10 glucose and 10 HEPES (titrated to pH 7.4 with NaOH). Rapid applications of 3mM caffeine and other compounds were used in the external K⁺-free solutions to probe the cellular Ca²⁺ stores or spontaneous Ca²⁺ releases from SR. While most experiments were carried out at room temperature, 24°C, some were performed at 35°C as indicated in the legends.

2.5. Fluorometric Ca²⁺ measurements in voltage-clamped cells

Cytosolic Ca²⁺, [Ca²⁺]_i, was measured routinely by including 0.1mM Fluo-4 pentapotassium salt in the dialyzing pipette solution. Whole-cell fluorescence was measured using excitation with 460nm light from an LED-based illuminator (Prismatix, Modiin Ilite, Israel) and detection of emitted light (>500nm) with a photomultiplier tube that was placed behind a moveable, adjustable diaphragm, which served to limit the area of detection to the voltage-clamped cell. Fluo-4 loaded, voltage- or current-clamped cells were also imaged using rapid 2-D confocal microscopy (Odyssey XL, Noran Instruments, Middleton, WI) to chart the spread of spontaneous cytosolic Ca²⁺ transients (Figs. 4 & 5; [36]). The stacks of fluorescence images were collected at 30 or 120Hz, using a 63x objective (Zeiss Plan-achromat, NA 1.4, oil) providing sampling on a 0.13μm grid.

Figure 4.

The beating rates of a rN-CM are similar whether it generates action potentials under current-clamp control (A1, A2, A3) or Na⁺-Ca²⁺ exchanger current is recorded at −50mV holding potential (B1, B2, B3). A1, B1, A2, B2: Time course of color-coded regional Ca²⁺ signals (ΔF/F₀), membrane current (I_m), and membrane potential (V_m). A3 & B3: Two sequences of 10 ratiometric fluorescence images measured at the times indicated in panels A2 and B2. F₀: Baseline fluorescence. ROI: Color-coded regions of interest (Red: approximate origin of spreading Ca²⁺ signals; purple: nucleus). Cells were voltage-clamped at 24°C, and dialyzed with a solution containing 50μM Fluo-4 to permit 2-D confocal imaging of cytosolic Ca²⁺ at 30 Hz.

Figure 5.

Confocal images of spreading cytosolic Ca²⁺ oscillations in hiPS-CM (Movie_S2) and rN-CM (Movie_S3). The larger images in each panel show the baseline fluorescence (F₀) measured with dialyzed Fluo-4 and color-coded regions of interests (ROI) that were chosen to correspond to the shifting origins of Ca²⁺ waves and the larger expanses between them. The traces show the time course of the normalized fluorescence changes (ΔF/F₀) in the ROIs and the simultaneously measured I_NCX. For each cell, the onset of each of 5 Ca²⁺ transients (1-5) is illustrated by 4 ratiometric images (+0, +1, +2, and +5) that were displayed at 30Hz. The last (+5) was delayed 100ms from the previous (+2) to show to what extend the Ca²⁺-waves spread throughout the cell. As shown on the color scale, elevation of [Ca²⁺]_i is indicated by successive transitions from a greenish hue to yellow and red colors. Notice that some beats do not develop fully (rN-CM: 2, 3 and 4; hiPS-CM: 2 and 5) and that this is associated with reduced I_NCX and shifting sites of origination of the Ca²⁺ waves. Cells were voltage-clamped at 24°C at a holding potential of −50mV, dialyzed with a solution containing 0.1mM Fluo-4, and imaged confocally at 120Hz.

In other experiments we used adenoviral expression of genetically engineered Ca²⁺ probes. Cytosolic Ca²⁺ was measured in intact cells with a probe, GCamP6-FKBP that was targeted to ryanodine receptors with the FKBP-12.6 binding domain [27]. Mitochondrial Ca²⁺ signals were measured confocally with a novel mitycam probe, mitycam-E31Q [28] that is targeted to the mitochondrial matrix space by subunit VIII of human cytochrome c oxidase, contains mutant calmodulin with a relatively high K_d (~1.6μM) for binding of Ca²⁺, and uses yellow fluorescent protein as the reporting fluorophore. Unlike Fluo-4 and GCamP6-FKBP it responds to increasing Ca²⁺ concentrations with a decrease in fluorescence.

2.6. Image processing

The collected stacks of fluorescence images showing spontaneous cytosolic and mitochondrial Ca²⁺ oscillations were low-pass filtered to display significant temporal and spatial variations. Individual frames were binned and filtered (2×2 or 3×3 pixel).

Cytosolic Ca²⁺ distributions measured with Fluo-4 were displayed ratiometrically (Figs. 4 & 5) as F/F₀ meaning F(x,y,t)/<F₀(x,y)> where F(x,y,t) is the low-pass filtered x-y image of a single frame measured at time t, and divisor is the baseline fluorescence image, F₀ = <F₀(x,y)> which was calculated from several frames collected between beats. This point-by-point division was carried out only out to the edge of the cell where the fluorescence intensity fell below 10-15 % of maximum. The ratiometric images were displayed at 30Hz on a rainbow scale where the baseline (F/F₀ ≅ 1) was shown in blue and elevated [Ca^2}+]_i as warmer green, yellow, and red hues.

The mitochondrial Ca²⁺ distributions measured with mitycam-E31Q (Figs. 13 & 14) were displayed differentially as ΔF/F₀ meaning (F(x,y,t)-<F₀(x,y)>)/<F₀> where <F₀> is a single number calculated as the average of <F₀(x,y)> over the entire cell. Generally short sequences of differential images (4-8) were pooled to clearly show local variation in mitochondrial Ca²⁺ signaling.

Figure 13.

Suppression of mitochondrial Ca²⁺ oscillations by FCCP in myticam-E31Q infected hiPS-CM (A1, A2) and rN-CM (B1, B2). The tracings show changes in regional mitochondrial Ca²⁺ signals (top) and I_NCX (bottom) before and 30 s after exposure to 50nM FCCP (A1, B1). For each cell type the images (A2, B2) show baseline fluorescence (F₀) and color-coded regions of interest (ROI, with “n” indicating nuclei) and 4 differential fluorescence images (measured at the times indicated along the traces) showing repeatable mitochondial Ca²⁺ signals before FCCP application (1, 2, 3) and smaller, more localized responses after (4). Increases in mitycam-E31Q fluorescence (reddish hues, arrows) correspond to reduced mitochondrial Ca²⁺, i.e. mitochondrial Ca²⁺ release. The cells were voltage-clamped at 24°C and dialyzed with a solution containing 100μM EGTA, while 2-D confocal imaging at 30 Hz was used simultaneously to measure mitochondrial Ca²⁺ reported by myticam-E31Q.

Figure 14.

Suppression of regional caffeine-induced regional mitochondrial Ca²⁺ signals by FCCP in a hiPS-CM (A1-3) and rN-CM (B1-3). A1 and B1: Time course of changes of fluorescence intensities in color-coded ROIs and I_NCX before (Control) and after addition of 50nM FCCP. A2 and B2: Baseline fluorescence (F₀), ROIs (ROI), and changes in mitochondrial Ca²⁺ the times (a-d) indicated along the I_NCX trace. The cell was incubated in 10μM blebbistatin to suppress cell shortening. A3 & B3: Average values of I_NCX and of mitochondrial Ca²⁺ uptake and release measured in the most active regions of rN-CM and hiPS-CM. The cells were voltage-clamped at 24°C and dialyzed with a solution containing 100μM EGTA, while 2-D confocal imaging at 30 Hz was used simultaneously to measure mitochondrial Ca²⁺ reported by myticam-E31Q with a decrease in fluorescence.

The ratiometric images of cytosolic Ca² distributions and the differential images of mitochondrial Ca²⁺ signals formed the foundations for the identification of color-coded regions of interest (ROI) with distinct kinetics. The image analysis was carried out using a custom designed program, Con2i [36].

2.6. Statistical analysis

Average values are presented in histograms and in the text as the mean ± the standard error of the mean for “n” cells. T-test was used to determine statistical significance. Significant findings are labeled with one (p< 0.05, *) or two stars (p < 0.01, **).

3. Results

3.1. Spontaneous beating: Membrane current and Ca²⁺ transients

Figure 1 illustrates simultaneous measurements at room temperature, 24°C, of membrane current (I_ncx) and whole-cell cytosolic Ca²⁺ (Fluo-4) in hiPS-CM (A) and rNCM (B) that were voltage-clamped at −50 mV, dialyzed with an internal solution containing 100μM Fluo-4, and incubated with 10μM blebbistatin to suppress contractions [29]. Both cell types produced rhythmic Ca²⁺ oscillations that activated in-phase transient inward currents that were attributed to the Na⁺-Ca²⁺ exchanger since they occurred in the absence of membrane potential excursions and had nearly the same time course as the measured Ca²⁺ transients. Interruption of spontaneous rhythmic activity by rapid application of caffeine also generated transient rises in global Ca²⁺ that were accompanied by large transient inward I_NCX and transiently blocked pacing suggesting involvement of SR Ca²⁺ stores. Quantification of the beating frequency and the amplitude of caffeine-induced I_NCX in rN-CM and hiPS-CM, suggested somewhat higher beating rates in rN-CM (Fig. 1C), but larger density of I_NCX in hiPS-CM (Fig. 1D).

Figure 1.

Simultaneous measurements of I_NCX current and Ca²⁺-signals in hiPS-CM (A) and rN-CM (B) voltage-clamped at −50 mV. A & B: Signals during spontaneous beating (left) and exposure to 3mM caffeine (middle) followed by microscopic bright field images of the voltage-clamped cell with labels indicating membrane capacitance (right). The bar graphs show differences in beating rate (C) and caffeine-induced I_NCX (D). Cells were voltage-clamped at 24°C, and dialyzed with a solution containing 100μM Fluo-4 to obtain whole-cell measurements of cytosolic Ca²⁺.

The spontaneous beating continued uninterrupted under broad array of experimental conditions. The visible contractions that guided the selections of cells had the same frequency as the Ca²⁺- and I_NCX -signals that generally continued for minutes after establishing the whole-cell voltage-clamp configuration. Recordings from rN-CM in the perforated-patch configuration and without Ca²⁺-dyes showed faster pacing when the temperature was raised to 35°C, but the frequency was the same whether the cells generated I_NCX transients at a holding potential of −50mV or generated action potentials in the current-clamp mode (Fig. 2 A & B). This was also the case when genetically engineered Ca²⁺ probes (Fig. 2 C) were used as indicators of beating, and it was observed under current-clamp conditions that the Ca²⁺-transients started to develop prior to the upstroke of the action potential (Fig. 2D &E). Similar action potentials were recorded in hiPS-CM in the current-clamp mode.

Figure 2.

Spontaneous beating in rN-CM measured at 35°C in the perforated patch configuration. A & C: Tracings of membrane current (I_m) and membrane potential (V_m) measured while rapidly switching back and forth between current- and voltage-clamp modes in cells without (A) or with Ca²⁺ measurements (C-E, ΔF/F_o) using the biologically engineered Ca²⁺ probe GCamP6-FKBP. B: Comparison of average beating frequency of action potentials (AP) and I_NCX transients in 7 cells without expression of GCamP6-FKBP. Panel E shows a single beat from panel D on an expanded time scale illustrating that the onset of the Ca²⁺ signal coincides with the diastolic depolarization that precedes the upstroke of the action potential.

To examine the effects of membrane potential more fully we performed voltage-clamp experiments, with membrane rupture and cell dialysis, on hiPS-CM and rN-CM where the beating was assessed by measuring oscillations in the exchanger current, I_NCX, during clamp pulses lasting 15s. In both cell types the beating rate was insensitive to potentials ranging from −60 to 0mV, but often slowed or stopped at more positive voltages (Fig. 3). Spontaneous membrane current oscillations also continued to occur at hyperpolarizing potentials between −60 to −90mV, but the frequency of the oscillation decreased somewhat as their magnitudes became larger and longer (data not shown).

Figure 3.

Spontaneous beating of hiPS-CM (A) and rN-CM (B) are insensitive to changes in holding potential. Oscillations in I_NCX were insensitive to 15s voltage-clamp depolarizations from −60 to −30 and 0mV. Depolarization to +30mV stopped oscillations of I_NCX in rN-CMs, but not in hiPS-CM. Capacitance: rN-CM 35pF, hiPS-CM 36pF. Cells were voltage-clamped at 35°C and dialyzed with a solution containing 0.1mM EGTA.

To analyze these robust and persistent I_NCX and Ca²⁺ oscillations we performed 2-D confocal fluorescence microscopy at 30 or 120 Hz of voltage-clamped cell that were dialyzed with 50μM Fluo-4. Again we found that spontaneous Ca²⁺ transients (ΔF/F₀) occurred with similar frequency whether the cell generated action potentials (V_m) with maximal diastolic potentials of ~-65mV (Fig. 4 A1), or generated inward I_NCX current at −50mV (Fig. 3 B1). Detailed examination, on an expanded time scale, revealed that the diastolic depolarization period coincided with local increases in Ca²⁺, which became endemic once the action potential was triggered (Fig. 4 A2, ↑).

The spreading Ca²⁺ signals could be followed in sequential ratiometric confocal images (Fig. 4 A3 & B3) and the time-course of the normalized fluorescence intensities (corresponding to selected color-coded regions of interest, ROI; Fig. 4 A2 & B2). At 30Hz imaging speeds, both action potential- and voltage-clamp-activated image sequences show emergence of a local Ca²⁺ signal near the center of the cell (Frame 0, red ROI) and similar spreading of Ca²⁺ waves in the adjoining orange-colored ROI in the following 4 frames (33, 67, 100 and 133ms). Perhaps more significantly, under current-clamp conditions, large rises in cytosolic Ca²⁺ appear to take place (red ROI) during the slow diastolic depolarization period, prior to triggering of action potential and further rapid rise of Ca²⁺ that spreads to the ends of the cell (Fig. 3 A3, frame 167ms). Under voltage-clamp condition, however, the Ca²⁺ wave, in the same cell, continues to spread slowly and with decreasing amplitudes (Fig. 4 B2, 4 o'clock arrow) and appears not to invade the ends of the cell by 300ms (Fig. 4, B3, frame 300ms, blue ROI). Consistent with this pattern, most voltage-clamped, and dialyzed hiPS-CM and rN-CM generated Ca²⁺ waves that typically originated from a relatively fixed central, peri-nuclear location, spread to the entire cell and were accompanied by significant activation of I_NCX (10-30pA, but occasionally generated smaller Ca²⁺ signals that remained localized and generated little or no detectable I_NCX (Fig. 5). Using less invasive procedures, we used GCamP6-FKBP to measure Ca²⁺ signals in spontaneously beating, intact, non-dialyzed, non-patched cells and found persistent Ca²⁺ oscillations where synchronous rises in Ca²⁺ most likely reflect generation of action potentials, while the localized or slowly spreading Ca²⁺ signals resulted from failure to reach the thresholds for electrical excitation (Fig. 6).

Figure 6.

Spontaneous Ca²⁺ signals at 35°C in an intact, non-voltage clamped rN-CM expressing GCamP6-FKBP probe. A: Spreading cytosolic Ca²⁺ signals recorded confocally at 120Hz. The curves show the time course of the fluorescence changes in selected ROIs. The same records are illustrated in Movie_S1. B: Color-coded regions of interest (left) and second beat on an expanded time scale (right). The diagram at the bottom shows the design of the expressed Ca²⁺-sensing probe GCamP6-FKBP.

Since pacing at distinct frequencies was robustly present in both rN-CM and hiPS-CM irrespective of methodology used (perforated-patch vs. cell dialysis vs. intact cells; dim light for whole-cell measurements vs. strong light for confocal imaging, 24 vs. 35°C, voltage- vs. current-clamp and Fluo-4 vs. GCamP6-FKBP as Ca²⁺ probe), in the next set of experiments we focused on Ca²⁺- and voltage-dependence of spontaneous pacing by: a) using holding potential of −50mV to prevent activation of voltage-gated ion channels; b) dialyzing with of 0.1mM Fluo-4 to allow detailed and fast Ca²⁺ measurements and c) experimenting at 24°C (instead of 35°C) to slow pacing frequency in order to more clearly monitor the sequence of events during the diastolic interval. With this general approach we assessed beating by measuring membrane current, I_NCX, and cytosolic Ca²⁺ transients where contractions were suppressed by blebbistatin [29].

3.2. Hyperpolarization-activated current, I_f, and spontaneous beating

Since spontaneous beating activity in the SA-nodal cells has been associated with the hyperpolarization-activated cyclic nucleotide-gated current, I_f, we examined the possible activation of this channel during spontaneous beating activity. We found significant expression of I_f channel in hiPS-CM, but not in spontaneously pacing rN-CM. Figure 7 A & B quantify the current density of I_f in hiPS-CM at different voltages and compares them to those recorded in mouse iPS-CM (miPS-CM) and rN-CM. It is clear that there was little or no significant activation of I_f at voltages positive to −80mV in either miPS-CM and hiPS-CM and none in rN-CM at any hyperpolarizing voltage. Furthermore, the small amounts of I_f activated at −80 or −90mV had activation kinetics that were too slow (seconds) to contribute significantly to rapid spontaneous beating (~2s⁻¹). Figure 7 D, confirms that 2mM Cs⁺, reported to block I_f in SA-nodal cells [37], rapidly and reversibly blocked I_f, in hiPS-CM. Thus neither the voltage-dependence and kinetics, nor the pharmacology of I_f channels is consistent with its contribution to the spontaneous beating activity of rN-CM or mouse iPS-CM. Consistent with this idea, 10μM ivabradine, a potent and specific blocker of I_f in SA-nodal cells [38] had no effect on spontaneous beating frequency or on the current oscillations activated at −50mV in hiPS-CM or rN-CM (Fig. 7E & F).

3.3. Ca²⁺ current and spontaneous beating

Since L-type Ca²⁺ channels are critically important in regulation of pacemaking activity in SA-nodal cells [37, 39], and in the light of recent reports that nifedipine suppresses spontaneous beating in hiPS-CM [40, 41], we measured the effectiveness of nifedipine in suppressing the spontaneous Ca²⁺ transient and membrane current oscillations recorded at −50mV. We found that 5-10μM nifedipine while fully suppressing I_Ca decreased the rate of spontaneous oscillation and their amplitude by only ~20% in hiPS-CM or rN-CM (Fig. 8 A & C) after ~1min. Under these conditions, significant rise in cytosolic Ca²⁺, though with slower kinetics, was activated with depolarizing pulses to 0mV in the I_Ca-suppressed cells (Fig. 8 B & D; upper traces), suggesting significant influx of Ca²⁺ on NCX and/or its release from the SR, independent of I_Ca-triggered release, underlies the spontaneous beating of these developing cardiomyocytes. Longer exposures (> ~5min) of nifedipine were accompanied by irreversible loss of I_Ca and irregular beating suggesting run-down or slow changes in the Ca²⁺ balance of the voltage clamped cells.

Figure 8.

The Ca²⁺ channel blocker nifedipine blocks I_Ca (B, D), but has only a minor effect on spontaneous oscillations of I_NCX and [Ca²⁺]_i (A, C) in hiPS-CM (A, B; 10μM, n=3) or rN-CM (C, D; 5μM, n=1). Panel B and D show that 5 or 10μM nifedipine effectively blocked I_Ca activated by depolarization to 0mV (lower traces) and delayed the development of Ca²⁺-dependent fluorescence (upper traces). Cells were voltage-clamped at 24°C using a holding potential of −50mV and dialyzed with a solution containing 0.1mM Fluo-4, which was excited with low intensity 460nm light for cytosolic whole-cell Ca²⁺ measurements.

3.4. Na⁺-Ca²⁺ exchanger and spontaneous beating

To identify the molecular entity responsible for generation of spontaneous rhythmic oscillations in the membrane current in these minimally Ca²⁺-buffered (0.1mM EGTA or Fluo-4 plus 0.1mM Ca²⁺) cells, we tested the sensitivity of this current to NCX-blockers KB-R7943 (10μM), Ni²⁺ (5mM) and Cd²⁺ (2mM). KB-R7943 had little immediate effect (Fig. 9 A & D), but in 15-30s it decreased by 60-80% (histograms) the magnitude of spontaneously occurring transient inward currents and the currents activated by caffeine-triggered Ca²⁺-release in both hiPS-CM and rN-CM (Fig. 5 A & D). In voltage-clamped cells dialyzed with Fluo-4, we found that the greatly attenuated spontaneous oscillations of I_NCX that still persisted in the presence of KB-R7943 (*s in Fig. 9 A & D) were accompanied by cytosolic Ca²⁺ transients that were often somewhat irregular in frequency, but of near normal amplitude (Fig. 9 B & E). To achieve a faster and more complete and reversible block of I_NCX, we used Ni²⁺ (5mM) or Cd²⁺ (2mM). Under voltage-clamp conditions, Ni²⁺ effectively blocked I_NCX, but left strong spontaneous and caffeine-induced Ca²⁺ signals that, as expected, showed marked suppression of rate of relaxation (Fig. 9 C; Cf. [42]). Similarly, spontaneously developing Ca²⁺ signals continued unabated when rapid application of Cd²⁺ blocked the generation of action potentials in cells that were current-clamped at 35°C in the perforated patch configuration (Fig. 9 E). It is plausible that longer lasting exposures to KB-R7943 may block both sarcolemmal and mitochondrial Na⁺-Ca²⁺ exchangers while divalent cations under current-clamp conditions may block the action potentials by blocking voltage-gated ion channels. Collectively the results illustrated in Fig. 9 suggest that rapid block of I_NCX does not abolish spontaneous Ca²⁺ oscillations in hiPS-CM and rN-CM.

Figure 9.

KB-R7943 (A, B, D, E, 10μM), Ni²⁺ (C, 5mM), and Cd²⁺ (F) suppress I_NCX in hiPS-CM (A-C) and rN-CM (D-F), but do not abolish spontaneous Ca²⁺ oscillation or caffeine-induced Ca²⁺ release. A & D: Exposures to KB-R7943 for 15-30s suppress spontaneously oscillating (upper panels) and caffeine-induced I_NCX (lower panels) only by 60-80% (histograms) thereby leaving indications of continued Ca²⁺ oscillations (*). B & E: Simultaneous measurements with Fluo-4 confirm the continuation of spontaneous cytosolic Ca²⁺ oscillations when I_NCX is suppressed by KB-R7943. C: Rapid exposure of hiPS-CM to 5mM Ni²⁺ blocks I_NCX while accentuating associated spontaneous [Ca²⁺]_i oscillations (top) and prolonging the duration of caffeine-induced [Ca²⁺]_i transients (bottom). F: Rapid exposure of a current-clamped rN-CM reversibly blocked action potentials without affecting cytosolic Ca²⁺ oscillations measured with GCamp6-FKBP (Cf. Fig. 6). The cells were voltage-clamped at 24°C (A-E) using dialyzing solutions containing 100μM EGTA (A, D) or Fluo-4 (B, C, E) or current-clamped in the perforated patch configuration at 35°C (F).

3.5. I_Ca-gated Ca²⁺ release from the SR and spontaneous beating

To probe the role of SR in the generation of spontaneous beating, we used agents that either blocked RyRs or depleted the SR Ca²⁺ content. We found that application of 100-200μM tetracaine, known to suppress SR Ca²⁺ release [43], arrested spontaneous beating rapidly and reversibly in both cell types clamped at −50mV (Fig. 10 A & C). Using longer exposures to tetracaine, spontaneous beating often recovered and was accompanied by sporadic larger releases of Ca²⁺ (Fig. 10 B). The first beat after washout of tetracaine was often accompanied by larger Ca²⁺ release and I_NCX, reflecting increased SR Ca²⁺ content during tetracaine block of RyRs.

Figure 10.

Transient suppression of spontaneous oscillations of Ca²⁺-dependent fluorescence (F/F₀) and I_NCX in voltage-clamped hiPS-CM (A, B) and rN-CM (C) by the ryanodine receptor-blocker tetracaine (100 or 200 μM) during short-term (A, C) and long-term (B) exposure. The first beat after washout of tetracaine was generally larger, indicative of an increased SR Ca²⁺ load during tetracaine exposure. In some cells, during the tetracaine exposure and cessation of spontaneous activity, sporadic but large transient inward currents were activated (B), that were thought to be related to unblocking of tetracaine-block as the Ca²⁺ load of SR increases in the presence of the drug (A). Similar data were also obtained in rN-CM. Cells were voltage-clamped at 24°C at a holding potential of −50mV, and dialyzed with a solution containing 0.1mM Fluo-4, which was excited with low intensity 460nm light for whole-cell measurements of cytosolic Ca²⁺.

Thapsigargin, a widely used inhibitor of Ca²⁺-ATPase in mammalian cells, suppressed the spontaneously activating I_NCX in 3-5min and caffeine triggered I_NCX in 5-7min in rN-CM, when used in a concentration of 5μM (Fig. 11 E-H). The effects developed more slowly with 0.5μM thapsigargin in hiPS-CMs (Fig. 11 A-D). Thus both sets of data support the idea that Ca²⁺ release from the SR plays a critical role in supporting spontaneous beating in the developing cells.

Figure 11.

Effect of thapsigargin, an inhibitor of Ca²⁺-ATPases (SERCA2a) on membrane current, I_NCX, in hiPS-CM (A-D) and rN-CM (E-H). Each panel shows I_NCX during spontaneous oscillations at the top and responses to paired caffeine exposures (3mM) at the bottom. The four panels illustrating each cell type show results under control conditions (A, E) and the gradual effects of exposure of hiPS-CM to 0.5μM thapsigargin for 5min (B), 8min (C) and 10min (D) or rN-CM to 5μM thapsigargin for 3min (F), 5min (G) and 7min (H). Cells were voltage-clamped at 24°C at a holding potential of −50mV, and dialyzed with a solution containing 0.1mM EGTA.

3.6. Mitochondrial un-coupler FCCP arrests spontaneous pacing

Since mitochondria constitute ~30% of the myocyte volume [44] and a number of reports suggest a critical role for mitochondria in cardiac Ca²⁺ signaling [18, 45], we probed the extent to which mitochondria regulate spontaneous beating. Using short exposures to low concentrations of FCCP (50-100nM), we found that this mitochondrial uncoupler abolished I_NCX and Ca²⁺ transients associated with spontaneous pacing (Fig. 12 A1 & B1). The suppressive effect of FCCP was accompanied by significant slowing in the rates of relaxation of the caffeine-induced I_NCX and Ca²⁺ transients (Fig. 12 A2 & B2), but amplitudes remained nearly intact (Fig. 12 A3 & B3), suggesting that SR Ca²⁺ loads and NCX activities were not affected substantially by FCCP in the low concentrations of drug used. These low concentrations of FCCP are reported to enhance oxidative mitochondrial metabolism without depolarizing the mitochondria [30] and were fully reversible after 2-3 minutes of drug washout.

Figure 12.

Suppression of spontaneous beating by the mitochondrial uncoupler FCCP. A1-3 (hiPS-CM, 50nM FCCP) & B1-3 (rN-CM, 200nM FCCP): Matched tracings of Ca²⁺-signals (Fluo-4) and membrane current (I_NCX) showing strong suppression of spontaneous Ca²⁺ and I_NCX oscillations (A1, B1), but prolongation of the caffeine-induced (3mM) signals (A2, B2) with only minor changes in their amplitudes (A3, B3). C: Ramp-clamp data suggesting that 100nM FCCP slightly attenuated I_Ca (in the range −30 to 40mV) without affecting I_Ncx. The histograms show that 100nM FCCP had no significant effects on the current measured at −100 and +80mV. D & E: I_NCX measured in response to Na⁺-withdrawal (137 → 5mM; TEA substitution) before (black) and after (red) treatment with FCCP in hiPS-CM (D) and rN-CM (E). The cells were voltage-clamped at 24°C and dialyzed with a solution containing 100μM Fluo-4 (A, B) or EGTA (C-E).

We also tested directly whether I_NCX might be suppressed by FCCP by using depolarizing/repolarizing ramp pulse protocols, from −50 to +80mV and back to −100mV, to measure both the influx and efflux of Ca²⁺ on NCX. Fig. 12 C shows that 100nM FCCP had no significant effect on I_NCX measured at −100 and +80mV even though it suppressed I_Ca as seen during the ramp-clamps at potentials between −30 and +60mV. This finding was further supported by the results where influx of Ca²⁺ on NCX was measured using the Na⁺ withdrawal procedure. Figure 12 D & E shows that the outward I_NCX activated on rapid withdrawal of Na⁺ (red and black traces) were not significantly affected by exposure to 50nM FCCP in either hiPS-CM or rN-CM (histograms), consistent with Fig. 6 C. The prolongation of relaxation time of caffeine-induced Ca²⁺ transients, and suppression of spontaneous pacing by short exposures of FCCP (Fig. 12 A & B), suggests a direct role for mitochondria in the cycling of cytosolic Ca²⁺ and regulation of pacing.

3.7. Measurements of mitochondrial Ca²⁺ in mitycam-E31Q infected cells

In spontaneously beating hiPS-CM or rN-CM, infected for 2-3 days with a genetically encoded mitochondrial Ca²⁺-sensing probe, mitycam-E31Q [28], we measured oscillations in the global mitycam signals that accompanied membrane current oscillations. Such global signals were, however, small in amplitude, and varied in kinetics as also reported earlier for adult myocytes [46]. To examine whether regional heterogeneity in mitochondrial populations contributes to the variability of global mitycam signals, we used confocal microscopy to identify regions of interest (ROI) with different Ca²⁺-signaling profiles and analyzed the results along the same lines used for imaging of cytosolic Ca²⁺ with Fluo-4. Figure 13 illustrates how this approach was used to analyze the patterns of mitochondrial Ca²⁺ signals in hiPS-CM and rN-CM in control conditions and where spontaneous beating was suppressed by 30s exposures to 50nM FCCP. The images of baseline fluorescence (F₀) showed that the mitycam-E31Q probe, targeted to mitochondrial subunit VIII of human cytochrome C, clustered around the nuclei of the cells (n). Under control conditions, as spontaneous beating activated large I_NCX transients (Figs.1-5, 7-13), the mitochondrial probe signal unexpectedly showed regionally diverse patterns, such that while some mitochondrial populations showed decrease in fluorescence (uptake of Ca²⁺), others displayed an increase in fluorescence corresponding to Ca²⁺ release. The mitochondrial populations showing Ca²⁺ release signals were essentially conserved in hiPS-CM (Fig. 13 A1 & A2, red trace and ROI) and rN-CM (Fig. 13 B1 & B2, red and orange traces and ROIs) from one beat to the next, consistent with findings of dominant pacing sites, (Figs. 4-6). The release sites were generally surrounded by mitochondria that appeared to concurrently take up Ca²⁺ (Fig. 13, green, blue, purple ROI in hiPS-CM; primarily blue ROI in rN-CM). Addition of FCCP suppressed the dominant pacing site and regular spontaneous beating, but left greatly reduced and discordantly activating I_NCX that were associated with more locally confined oscillatory mitochondrial Ca²⁺ signals where Ca²⁺ releases and uptakes were seen to switch back and forth between neighboring ROIs. For instance, as seen in the illustrated hiPS-CM (Fig. 13 A1), the release was seen to switch from the purple ROI to the red and back again. These patterns of mitochondrial Ca²⁺ signaling are characteristic of 4 recordings in rN-CM and 3 recordings in hiPS-CM and are consistent with the Fluo-4 measurements of cytosolic Ca²⁺ (Fig. 12 A & B) suggesting that mitochondria play an active role in Ca²⁺-mediated pacing in these developing cells.

Releases and uptakes of Ca²⁺ by distinct mitochondrial populations were also observed in both hiPS-CM and rN-CM in response to rapid puffs of 3mM caffeine that also activated large I_NCX (Fig. 14). As anticipated from Fig. 12, short (30-60s) exposures to 50nM FCCP significantly slowed the decay of I_NCX, (Fig. 14 A1 & B1, lower traces), but produced only minor reduction in its magnitude (Fig. 14 A3 & B3, right histograms). Simultaneously measured mitochondrial Ca²⁺ signals were similarly prolonged (Fig. 14 A1 & B1, upper traces) while their amplitudes corresponding to both uptake and release were reduced relatively little initially (Fig. 14 A3 & B3, left histograms). Similarly, baseline mitycam fluorescence after addition of FCCP was 80±3% (rN-CM, n=6) or 83±7% (hiPS-CM, n=6) percent of control, indicating that the mitochondrial Ca²⁺ content remained largely intact. The effects of exposure to 50-100nM FCCP for 1-3 min were fully reversible. When used in higher concentrations (1μM), FCCP irreversibly damaged mitochondrial Ca²⁺ handling.

4.1. Discussion

The quintessential property of developing cardiomyocytes is their ability to beat rhythmically in culture, a property also observed in early stages of embryonic development, long before the heart evolves its multi-chambered structure. Here, we have attempted to probe the mechanisms underlying the spontaneous rhythmic beating of cardiomyocytes in two cell models: 1) Neonatal rat cardiomyocytes isolated from 2-3 day old rat hearts, and 2) spontaneously beating human iPS cell-derived cardiomyocytes [25]. Because we have already reported that Ca²⁺-signaling properties of rN-CM and hiPS-CM were similar to those of freshly isolated adult cardiomyocytes, [26, 31], our approach was to image focal changes in Ca²⁺ profiles in spontaneously beating cells either dialyzed with Fluo-4 or infected with genetically encoded Ca²⁺ probes directed at ryanodine receptors (GCamP6-FKBP) or mitochondria (mitycam-E31Q [28]). Fluo-4 confocal imaging showed that while Ca²⁺ signaling started at a preferred central location and spread throughout the cell in a repeatable pattern, the mitycam E31Q mitochondrial probe showed divergent patterns of Ca²⁺ release and uptake during spontaneous pacing. The finding that spontaneous pacing rate in developing myocytes were not significantly affected by triggering of action potential, but was highly sensitive to low concentrations of mitochondrial un-coupler FCCP, suggests that the primary pacing mechanism may be independent of activation of ion channels, but highly dependent on metabolic state of mitochondria.

4.2. Expression of hyperpolarization-activated current, I_f, and spontaneous rhythmic beating

Fig. 7 suggests significant differences in the level of expression of I_f in rN-CM and human or mouse iPS-CM. Interestingly, even though rat neonates do not express I_f (Fig. 7 A-C), the frequency of its spontaneous beating was significantly higher than that of human iPS-CM (Fig. 1). Expression of I_f in iPS-CM was more robust and was present in every cell examined, but I_f activated only at potentials negative to −80 mV in human and −70 mV in mouse iPS-CM, and was rapidly and reversibly blocked by 2mM Cs⁺ (Fig. 7 D) without significantly altering the pacing frequency, as reported also for human and rabbit SA-nodal cells [47, 48]. Thus, if I_f were to contribute to generation of spontaneous activity, it would have to carry significant inward current at voltages positive to −70 mV, clearly not the case here, where I_f is either absent or its kinetics too slow and activation-voltage too negative. Furthermore, the pacing rate was unaltered by blockers of I_f, (Fig. 7 E & F), changes in holding potential (Fig. 3, −60 to 0 mV), and the triggering of action potentials (Figs. 2, 4, 6). Pacing frequency did not vary significantly between intact nondialyzed cells (Fig. 6), and cells voltage-clamped at −50mV, but the frequency accelerated significantly (to ~3/s in rN-CM) when the temperature was increased from 24 to 35°C under perforated patch conditions (Fig. 2). Thus, our findings argue for a mechanism independent of membrane ion channels initiating the spontaneous rhythmic beating.

This raises the question as to the role of I_f current in mouse and human iPS-CM, as it is present in these cells (Fig. 7), and in mammalian SA-node, where it is proposed to initiate diastolic depolarization alone [3] or together with Ca²⁺-activated I_NCX [7]. It is a somewhat of over-looked observation that in SA-nodal cells too, there is a discrepancy between diastolic depolarization potentials (−60 to −40 mV) and activation range of I_f (negative to −70 mV), including human SA-node [48]. A combination of rapid 2-D Ca²⁺ imaging and voltage- and current-clamp studies might clarify the role of I_f in SA-nodal cells by examining whether sub-sarcolemmal Ca²⁺ transients always precede the diastolic depolarization. We suggest that I_f is critical in protecting the small pacemaker cells from the large voltage sink of bigger atrial cells by serving as a “functional insulator”. It is likely that without expression of HCN proteins, activating strongly and rapidly at potentials near the atrial resting potentials (~−90 mV), the SA-nodal cells would be hyperpolarized to more negative resting potentials of atrium, thereby preventing them from generating the diastolic depolarization necessary for pacing the heart. This property of I_f along with significantly lower expression of I_K1 channels especially in SA-nodal/atrial junctional cells may serve as a dynamic “push-me pull-you” current oscillations that might amplify the oscillations set by the Ca²⁺ signaling mechanisms in central SA-nodal cells.

4.3. Rhythmic spontaneous Ca²⁺ release and NCX current oscillations

The central observation of this report is that spontaneous rhythmic oscillations of cytosolic Ca²⁺ and the accompanying I_NCX were consistently recorded in myocytes that were voltage-clamped at constant holding potentials of −50mV, where no ionic channels were allowed to activate (Figs. 1-5, 7-14). The sensitivity of these signals to KB-R7943, Ni²⁺ and Cd²⁺ suggest that the current oscillations (I_NCX) were triggered by cytosolic Ca²⁺ changes activating the NCX current (Fig. 9). Although the rhythmic activity was fairly constant, either at the onset of dialysis and before imaging laser light, they became more irregular following exposure to intense light of confocal imaging. Such irregularities were less apparent when imaging measurements were not attempted. Confocal examination of such rhythmic beating activity in cells dialyzed with Fluo-4 and held at −60 or −50 mV suggests multiple Ca²⁺ release sites. The stronger Ca²⁺ triggering sites appeared to invade the cell rapidly and fully (Figs. 4-6) while weaker sites were not able to invade the cell fully (Figs. 5 & 6). It is not clear as yet why the strong triggering sites give way to the weaker site, but since the frequency of spontaneously occurring rhythmic activity is not significantly altered by nifedipine (Fig. 8) or by holding potentials (−60 to 0 mV; Fig. 3), the data is consistent with activation of different Ca²⁺ pools capable of maintaining the rhythmic activity.

The Ca²⁺-dependent mechanism described here for two types of developing cells provides the intriguing possibility that significant rises of Ca²⁺ may occur in specific cellular micro-domains of the spontaneously pacing cell during diastolic depolarization, prior to activation of upstroke of the action potential (Figs. 2 & 4) that may in turn contribute to slow diastolic depolarization of the membrane. This mechanism might be somewhat different from that described for the adult SA-nodal cells where voltage-gated channels are thought to play a mandatory role [7, 18, 49]. It remains to be explored whether these differences are associated with the developmental stages of the examined cells or experimental approaches.

4.4. Nature of Ca²⁺ pools supporting the spontaneous rhythmic beating

The findings that tetracaine and thapsigargin suppressed the rhythmic beating in both hiPS-CM and rN-CM, (Figs. 10 & 11) suggests Ca²⁺ release from the SR is critical in generating the rhythmic beating. Somewhat less expected were the rapid and reversible suppressive effects of small (50-100nM) and short exposures to mitochondrial un-coupler FCCP on the rhythmic activity (Figs. 12 & 13). Since FCCP also suppressed the rate of relaxation of caffeine-triggered Ca²⁺-transients (Figs. 12 & 14) as it inhibited the mitochondrial Ca²⁺ signals associated with spontaneous pacing (Figs. 12 & 13), the data suggests a critical role for mitochondria in spontaneous pacing. The mitochondrial imaging experiments with mitycam-E31Q probe, suggesting that the peri-nuclear population of mitochondria release Ca²⁺ as the peripheral mitochondria take up Ca²⁺ (Fig. 13, spontaneous beating and Fig. 14, caffeine application), were somewhat surprising, but do provide an intriguing possibility that may couple the mitochondrial ATP production to the rhythmic Ca²⁺ signaling and spontaneous beating. It is not as yet clear whether Ca²⁺ sequestered by mitochondrial is re-released directly into the cytosol or is transferred via specialized SR-mitochondrial junctions into the SR, and thereby modulating spontaneous beating. Our data clearly provide some support for mitochondrial Ca²⁺ release (Figs. 12-14), but the release mechanism remains unclear, though proposed candidates range from mitochondrial permeability transition pore (mPTP), and mNCX, to mRyR1 [50-52].

4.5. Physiological implications

Rhythmic activity in the adult mammalian heart originates from the SA-nodal cells. The morphology of SA-nodal cells differs sharply from their neighboring atrial and ventricular cells. They are smaller, ~20pF vs. 70-250pF for atrial and ventricular cells, with large surface to volume ratios, allowing for released Ca²⁺ to produce larger local Ca²⁺ concentrations, activating thereby larger inward I_NCX. Spontaneously beating rNCM and hiPS-CM are similarly small (~20pF), with large surface/volume ratios, and robust Ca²⁺ releases generating large I_NCX (Figs. 1-5, 7-13). Interestingly, unlike hiPS-CM, rN-CM do not express I_f, yet have higher pacing frequency, and otherwise are indistinguishable from hiPS-CM electrophysiologically, supporting the idea of minor role for I_f in spontaneous pacing of developing cardiomyocytes

Our studies provide insight into the mechanism that sets the “Ca²⁺ oscillator” in developing cardiomyocytes. Clearly depletion of Ca²⁺ content of SR, using caffeine or thapsigargin (Fig. 11), or blocking the RyRs (Fig. 10) stops the rhythmic Ca²⁺ oscillations, implicating the SR Ca²⁺ release to be critical in initiation of rhythmic beating. More intriguing, however, is the finding that short exposures of 50nM FCCP, reported to enhance oxidative metabolism, but not to depolarize the mitochondria or reduce ATP levels [30], consistently blocked the rhythmic activity rapidly and reversibly. Since the drug prolonged the rate of re-uptake of Ca²⁺ of spontaneously activated or caffeine-triggered Ca²⁺ transients (Fig. 12), it is likely that mitochondria have a critical role in rhythmic Ca²⁺-cycling thus modulating pacemaker activity in developing myocytes. We suggest that a cross-talk between SR and mitochondria sets the rhythmic cycle of Ca²⁺ release and re-uptake that results in cyclic activation of NCX to generate the inward current necessary to depolarize the myocyte at voltages beyond the activation voltage of I_f (Fig. 7). This possibility is strongly supported by the perforated patch and current-clamped experiments where the pacing cell generates significant rises of Ca²⁺ in specific cellular micro-domains prior to the upstroke of the action potential at −50mV, see (Fig. 2).

Experiments using mitycam-infected cells show diversity in mitochondrial populations vis-à-vis Ca²⁺ release and uptake. In this respect FCCP was effective in suppressing both the mitochondrial release and uptake signals as it suppressed the spontaneous beating and I_NCX accompanying them (Fig. 13). It is likely that this diversity in Ca²⁺-signaling of different mitochondrial populations serves as an additional oscillatory mechanism contributing to spontaneous pacing. Whether the processes that generate spontaneous pacing are also coupled to the energy producing mechanisms remains to be explored. This possibility may reinforce the notion that Ca²⁺ signaling parameters are relevant to normal and abnormal cardiac impulse formation where mitochondrial dysfunction may play a critical role in pathologies like arrhythmias and dilated cardiomyopathy [53].

4.6. Conclusions

Regionally diverse mitochondrial Ca²⁺-signaling, in addition to RyR2/NCX co-activation, may provide the critical negative feedback mechanism that initiates and regulates spontaneous beating in developing cardiomyocytes, independent of I_f-channels.

Highlights.

• Myocytes derived from newborn rat heart & human induced pluripotent stem cells pace spontaneously

• Ca²⁺-mediated pacing mechanisms are similar in both developing cell types

• Pacing depends on activation of I_NCX & Ca²⁺-transients, but not on I_f and acute I_Ca block & V_m change

• Regionally diverse mitochondrial Ca²⁺ signals may initiate and regulate pacing in both cell types

Bullet.

Ca²⁺ release from SR and mitochondria, not If, controls beating in developing cardiomyocytes.

Abbreviations

HCN channel

hyperpolarization-activated cyclic nucleotide-gated channel

hiPS-CM

cardiomyocytes derived from human induced pluripotent stem cells

I_f

“funny” current through HCN channel

I_NCX

Na⁺-Ca²⁺ exchanger current

miPS-CM

cardiomyocytes derived from mouse induced pluripotent stem cells

NCX

Na⁺-Ca²⁺ exchanger

rN-CM

rat neonatal cardiomyocytes

RyR

ryanodine receptor

SR

sarcoplasmic reticulum