Inositol-1,4,5-trisphosphate-mediated spontaneous activity in mouse embryonic stem cell-derived cardiomyocytes

Abstract

Embryonic stem cell-derived cardiomyocytes (ESdCs) have been proposed as a source for cardiac cell-replacement therapy. The aim of this study was to determine the Ca²⁺-handling mechanisms that determine the frequency and duration of spontaneous Ca²⁺ transients in single ESdCs. With laser scanning confocal microscopy using the Ca²⁺-sensitive dye Fluo-4/AM, we determined that spontaneous Ca²⁺ transients in ESdCs at the onset of beating (day 9) depend on Ca²⁺ entry across the plasma membrane (50%) whereas Ca²⁺-induced Ca²⁺ release is the major contributor to Ca²⁺ transients in ESdCs after 16 days (72%). Likewise, Ca²⁺ extrusion in 9-day-old ESdCs depends on Na⁺–Ca²⁺ exchange (50.0 ± 8%) whereas Ca²⁺ reuptake by the sarco(endo)plasmic Ca²⁺ ATPase (72 ± 5%) dominates in further differentiated cells. Spontaneous Ca²⁺ transients were suppressed by the inositol-1,4,5-trisphosphate (IP₃) receptor (IP₃R) blocker 2-aminoethoxydiphenyl borate (2-APB) and the phospholipase C blocker U73122 but continued in the presence of caffeine. Stimulation of IP₃ production by phenylephrine or endothelin-1 had a positive chronotropic effect that could be reversed by U73122 and 2-APB. The presence of Ca²⁺-free solution and block of L-type Ca²⁺ channels by nifedipine also resulted in a cessation of spontaneous activity. Overall, IP₃R-mediated Ca²⁺ release in ESdCs is translated into a depolarization of the plasma membrane and a whole-cell Ca²⁺ transient is subsequently induced by voltage-dependent Ca²⁺ influx. Although ryanodine receptor-mediated Ca²⁺ release amplifies the IP₃R-induced trigger for the Ca²⁺ transients and modulates its frequencies, it is not a prerequisite for spontaneous activity. The results of this study offer important insight into the role of IP₃R-mediated Ca²⁺ release for pacemaker activity in differentiating cardiomyocytes.

In early embryonic heart as well as in the adult sinoatrial (SA) node, spontaneous pacemaker activity strongly depends on the presence of Ca²⁺ signalling mechanisms (Sakai et al. 1983; Komuro et al. 1985; Bogdanov et al. 2001; Lipsius et al. 2001). In the early embryonic heart, spontaneous activity was shown to be initiated by Ca²⁺ influx, whereas in the rabbit or cat SA node, the rhythmic increase in spontaneous ryanodine receptor (RyR)-mediated Ca²⁺ release events and subsequent activation of the Na⁺/Ca²⁺ exchanger (NCX) seem to be the driving force for the diastolic depolarization (Huser et al. 2000; Bogdanov et al. 2001). One reason for the differences in the pacemaker mechanism could be the developmental changes in the Ca²⁺ handling mechanisms. While voltage-dependent L-type Ca²⁺ channels are expressed from day 9.5 days post coitum (dpc) (Takemura et al. 2005), T-type channels have been observed around embryonic day 14 (Cribbs et al. 2001). Influx through L-type channels is thought to be the major source of Ca²⁺ for excitation–contraction coupling (ECC) in early embryonic cardiomyocytes, whereas the main source of Ca²⁺ in adult cells during ECC is RyR-mediated Ca²⁺-induced Ca²⁺ release (CICR) (Escobar et al. 2004; Perez et al. 2005). The cause for this seems to be the rudimentary development of the sarcoplasmic reticulum (SR) and the low expression of RyRs (Gorza et al. 1997; Rosemblit et al. 1999; Escobar et al. 2004). The role of inositol-1,4,5-trisphosphate (IP₃) receptor (IP₃R)-mediated Ca²⁺ release in cardiac muscle is still debated. Recent studies have suggested a role in the generation of arrhythmic activity in atrial and ventricular myocytes or in the defined activation of cardiac transcription factors (Zima & Blatter, 2004; Proven et al. 2006; Wu et al. 2006). However, it seems that IP₃-mediated Ca²⁺ release through IP₃R channels may play a more prominent role in embryonic cardiomyocytes. IP₃ production has been linked to a specific perinuclear Ca²⁺ release and shown to regulate mitochondrial Ca²⁺ uptake in neonatal rat cardiomyocytes (Jaconi et al. 2000).

Embryonic stem cells differentiate in vitro into spontaneously active multicellular cardiomyocyte preparations while recapitulating developmental stages of the early embryonic cardiomyogenesis (Boheler et al. 2002; Hescheler et al. 2002; Banach et al. 2003). They express cardiac-specific marker proteins (e.g. αMHC and MLC-2v) and transcription factors (e.g. Nkx-2.5, MEF2C and GATA-4) (Narita et al. 1997; Boheler et al. 2002; Wobus & Boheler, 2005). Shortly after the onset of spontaneous synchronized beating, the proliferation of the embryonic stem cell-derived cardiomyocytes (ESdCs) ceases and the multicellular aggregates expand by cell growth (Meyer et al. 2000; Banach et al. 2003; He et al. 2003; Puceat et al. 2003). Intracellular recordings from individual ESdCs dissociated from aggregates of beating cells show that different cellular phenotypes such as sinus node-like, atrial-like and ventricular-like cells develop (Maltsev et al. 1999; Abi-Gerges et al. 2000; Fijnvandraat et al. 2003; He et al. 2003). With the in vitro differentiation of ESdCs, expression of the Ca²⁺-handling proteins such as RyR (cardiac type-2 isoform), NCX1, sarcoplasmic reticulum Ca²⁺ ATPase (SERCA)2 and phospholamban has been described (Boheler et al. 2002; Fijnvandraat et al. 2003). Furthermore, the following voltage-dependent currents have been shown in ESdCs: I_f (Gryshchenko et al. 1999; Abi-Gerges et al. 2000), I_to and I_sus (Gryshchenko et al. 2000), and I_Na, I_Ca,T and I_L,Ca (Abi-Gerges et al. 2000; Zhang et al. 2003). Pacemaker activity in ESdC aggregates, connexin expression and conduction velocity also follow closely the development of the embryonic heart (Banach et al. 2003) and pacemaker frequency is responsive to the neurohumoral stimuli isoprenaline and acetylcholine (Ji et al. 1999; Abi-Gerges et al. 2000; Banach et al. 2003).

The spontaneous activity in early embryonic cardiomyocytes is thought to be significantly regulated by I_f (Satoh & Sperelakis, 1993; Robinson et al. 1997); however, it also depends critically on the presence of Ca²⁺ in the extracellular solution (Fujii et al. 1981a,Fujii 1981b). It has been proposed that intracellular Ca²⁺ oscillations in ESdCs are independent of voltage-activated Ca²⁺ influx (Viatchenko-Karpinski et al. 1999; Mery et al. 2005). RyR-mediated Ca²⁺ release has been demonstrated by the presence of Ca²⁺ sparks and caffeine transients in ESdCs (Sauer et al. 2001) and RyR-mediated Ca²⁺ release was shown to contribute to their pacemaker activity (Fu et al. 2006). However, spontaneous whole-cell Ca²⁺ transients continue in the presence of caffeine and in ESdCs deficient for Ry R² expression (Sauer et al. 2001; Yang et al. 2002). By contrast, specific blocking or down-regulation of IP₃R type 1 suppresses Ca²⁺ signalling and automaticity of ESdCs and embryonic heart tubes (Mery et al. 2005). Until now, neither the basis for IP₃R activation nor the mechanism by which IP₃R-mediated Ca²⁺ release results in the spontaneous whole-cell Ca²⁺ transients of the ESdCs has been determined.

In the present study, we tested the hypothesis that the developmental changes in the Ca²⁺ handling mechanisms of ESdCs modulate the frequency of their spontaneous activity, and that IP₃-mediated Ca²⁺ release needs to be translated into a depolarization of the membrane potential (V_m) and subsequent activation of voltage-dependent Ca²⁺ channels to generate a whole-cell Ca²⁺ transient.

Methods

Culture and differentiation of mouse embryonic stem cells

For the experiments we used Mouse embryonic stem cells (mES) of the cell line CMV (Specialty Media; Phillipsburg, NJ, USA) transfected with a neomycin resistance gene expressed under the promoter of αMHC (Klug et al. 1996; Strom et al. 2002). The cells were propagated in culture in the presence of leukaemia inhibiting factor (ESGRO; 1000 U ml⁻¹; Chemicon Int., Temecula, CA, USA) and differentiated as embryoid bodies (EBs) (Nagy et al. 1993; Hescheler et al. 2002; Banach et al. 2003). EBs were plated on day 7 and spontaneously beating aggregates of cardiomyocytes were observed around day 8 (Banach et al. 2003). Neomycin (300 μg ml⁻¹) was added to the differentiating cells at day 9 of culture to select for cardiomyocytes. In early preparations (day 8–10) in which non-cardiomyocytes were not yet entirely removed by neomycin treatment, beating aggregates were microdissected under optical control. For measurements of [Ca²⁺]_i, cardiomyocyte aggregates were collected and dissociated as previously described (Maltsev et al. 1993). Briefly, the aggregates were washed in a Ca²⁺- and Mg²⁺-free phosphate-buffered saline solution (Sigma-Aldrich, St Louis, MO, USA). After 60 min incubation in collagenase (1 mg ml⁻¹ at 37°C; Sigma-Aldrich), the cells were incubated in Kraftbrühe (KB) solution [Klockner et al. 1999; 60 min, room temperature (20°C)]. Dissociated cells were plated onto gelatine-coated coverslips in Iscove medium supplemented with 20% fetal bovine serum (Invitrogen, Frederick, MD, USA). Experiments were carried out on the first day after dissociation. Older ESdC aggregates contain cells with different cardiac phenotypes; however, it was demonstrated that the promoter αMHC is mainly activated in sinus nodal-like and atrial-like ESdCs (Kolossov et al. 2005). By using αMHC–neomycin selected cells we can assume that in our cultures sinus nodal-like cells and atrial-like cells dominate and that the experiments were performed on spontaneously active sinus nodal-like ESdCs.

Intracellular Ca²⁺ measurements

To visualize changes in the intracellular Ca²⁺ concentration, ESdCs were incubated (15 min) at room temperature with fluo-4 acetoxymethyl ester (5 μm; fluo-4/AM; Invitrogen, Eugene, OR, USA). After washout, 15 min were allowed for de-esterification of the dye. Ca²⁺ measurements were performed as previously described (Sheehan & Blatter, 2003). Whole-cell Ca²⁺ transients were obtained from confocal linescan images through single ESdCs by averaging the signal of an individual cell. Ca²⁺ transients are presented as background-subtracted normalized fluorescence (F/F₀).

Solutions

During the experiments cells were superfused with a Tyrode solution containing (mm): NaCl 140, KCl 5.4, CaCl₂ 1.5, MgCl₂ 1.5, glucose 10 and Hepes 5; pH adjusted to 7.38 with NaOH. To induce a depolarization of the cell Membrane potential (V_m) we used a ‘high-K⁺’ solution containing (mm): KCl 140, NaCl 5.4, MgCl₂ 1.5, glucose 10 and Hepes 5; pH adjusted to 7.38 with KOH. To determine the role of the NCX for spontaneous whole-cell Ca²⁺ transients, we superfused the ESdCs with a ‘low-Na⁺’ solution containing (mm): LiCl 128, NaCl 12, KCl 5.4, MgCl₂ 1.5, glucose 10 and Hepes 5; pH adjusted to 7.38 with LiOH. In both cases Ca²⁺ was omitted from the solutions.

Endothelin-1, phenylephrine, tetracaine, nifedipine, NiCl, BAPTA-AM, U73122 and caffeine were purchased from Sigma. Fluo-4/AM and thapsigargin were purchased from Invitrogen. Data sets were statistically evaluated using an unpaired t test. All averaged data are presented as means ± s.e.m.

Results

ESdCs exhibit spontaneous whole-cell Ca²⁺ transients

Embryonic stem cells develop into multicellular beating aggregates of cardiomyocytes. After dissociation, single cells maintain their spontaneous activity and rhythmic changes of [Ca²⁺]_i can be recorded (Viatchenko-Karpinski et al. 1999). To determine the characteristics of this spontaneous activity, we monitored Ca²⁺ transients in ESdCs on concomitant days of differentiation. Changes in [Ca²⁺]_i were recorded in Fluo-4/AM-loaded ESdCs with the linescan mode of the confocal microscope (4.37 ms line⁻¹). The scan line was positioned through the centre of the cell and the fluorescence was integrated over the entire width of the cell for quantification of the transients. Rhythmic spontaneous increases in [Ca²⁺]_i were detected in all ESdCs examined and representative examples of linescans recorded from an ESdCs at 11 days and a 17 days, respectively, are shown in Fig. 1A and B. To distinguish between AP-induced Ca²⁺ transients and spontaneous Ca²⁺ waves we considered a change in Ca²⁺ to be a whole-cell Ca²⁺ transient when Ca²⁺ increased simultaneously throughout the cell and its rise time did not exceed 67 ms. The frequency of spontaneous whole-cell Ca²⁺ transients increased with the differentiation of the ESdCs (day 8, 0.25 ± 0.02 Hz, n = 5; day 17, 1.05 ± 0.09 Hz; n = 19; Fig. 1C) whereas the transient duration at 50% amplitude (TD₅₀) decreased (day 8, 600 ± 77 ms, n = 5; day 17, 212 ± 29 ms, n = 19; Fig. 1D). No significant change in transient amplitude (ΔF/F₀: day 8, 2.39 ± 0.37, n = 7; day 17, 2.67 ± 0.3, n = 19; Fig. 1E) was observed. Whole-cell Ca²⁺ transients exhibited a slow diastolic rise of Ca²⁺ (day 9, 1.2 ± 0.17 d(ΔF/F₀) s⁻¹, n = 11; day 17, 5.8 ± 0.66 d(ΔF/F₀) s⁻¹, n = 11; Fig. 2A and B). The diastolic rise was followed by a rapid upstroke of [Ca²⁺]_i (day 9, 26.1 ± 5 d(ΔF/F₀) s⁻¹, n = 11; day 17, 52.3 ± 4.6 d(ΔF/F₀)*s⁻¹, n = 11; Fig. 2A and B). The velocity of diastolic rise and upstroke significantly increased with differentiation of the ESdCs. The data demonstrate a change of the spontaneous Ca²⁺ transient in differentiating ESdCs.

Figure 1.

Developmental changes in spontaneous whole-cell Ca²⁺ transients in ESdCs Confocal linescan images and spatially averaged Ca²⁺ transients recorded from ESdCs isolated on day 11 (A) or day 17 (B) of in vitro differentiation. A slow diastolic increase of [Ca²⁺]_i preceded the rapid upstroke of the Ca²⁺ transient. While the frequency of the spontaneous Ca²⁺ transients increased with differentiation (C), the transient duration (TD) at 50% amplitude decreased (D). However, no significant change in transient amplitude (TA) could be detected (E).

Figure 2.

The velocity of diastolic increase and upstroke of whole-cell Ca²⁺ transients in control conditions and in the presence of caffeine A, representative Ca²⁺ transients recorded from ESdCs at day 9 (black trace) and day 17 (grey trace) of differentiation. The phases of the diastolic rise and the upstroke are marked underneath the transients. B, comparison of the diastolic rise and upstroke of Ca²⁺ in ESdCs dissociated on day 9 (n = 6) and day 17 (n = 4) of differentiation. Both parameters exhibit a significant developmental increase (**P < 0.001). C, caffeine (10 mm) slows the diastolic rise of [Ca²⁺]_i in ESdCs on day 9 (n = 5) and day 17 (n = 3). In both cases the decrease is significant (*P < 0.05; **P < 0.001). However, no significant difference in the diastolic rise can be determined in ESdCs between day 9 and day 17 in the presence of caffeine.

Early ESdCs already have intracellular Ca²⁺ stores

One significant change in the Ca²⁺ signalling mechanisms is the development of RyR-operated intracellular Ca²⁺ stores and their contribution to CICR (Gorza et al. 1997; Rosemblit et al. 1999; Escobar et al. 2004). We determined the presence of RyR-operated intracellular stores by superfusing spontaneously active ESdCs with the RyR agonist caffeine (10 mm). Already at the onset of beating (day 8), caffeine induced a rapid rise in [Ca²⁺]_i underlining the presence of RyR-operated intracellular stores. Figure 3A and B shows representative examples of caffeine-induced Ca²⁺ transients in ESdCs at day 10 and day 17. The amplitude of the caffeine-induced transients (TA_caff) increased with differentiation of the ESdCs (Fig. 3C). The spontaneous activity of ESdCs continued in caffeine in 53 out of 59 experiments (ESdCs day 8–17). However, the amplitude of spontaneous whole-cell Ca²⁺ transients in caffeine (TA_T,caff) was decreased in comparison to control conditions (TA_T,ctrl). In ESdCs at day 9, the transients continued at 50% of their control amplitude, whereas the amplitude of twitch transients was reduced by 70% at day 17 (Fig. 3D). The data suggest that although RyR-operated stores are present from the onset of spontaneous activity, their contribution to the whole-cell Ca²⁺ transient and ECC increases during differentiation.

Figure 3.

RyR-operated Ca²⁺ stores gain relevance during developmental differentiation of ESdCs Confocal linescan images and spatially averaged Ca²⁺ transients recorded from ESdCs isolated on day 10 (A) and day 17 (B) of in vitro differentiation. Caffeine (10 mm)-induced Ca²⁺ transients document the presence of RyR-operated Ca²⁺ stores. The presence of caffeine does not suppress the occurrence of spontaneous whole-cell Ca²⁺ transients. C, the amplitude of caffeine-induced transients (TA_caff) increases during in vitro differentiation. D, the ratio of transient amplitude in caffeine (TA_T,caff)/transient amplitude control (TA_T,ctrl) decreases. The data support an increasing contribution of RyR-mediated Ca²⁺ release to whole-cell Ca²⁺ transients during differentiation. E, Ca²⁺ transient frequency in older ESdCs is significantly decreased in the presence of caffeine.

Ca²⁺ transients in the presence of caffeine not only exhibited a decrease in amplitude, but a decrease in the slope of the diastolic rise (Fig. 2C). A 38% and 83% reduction in the rise time was determined in ESdCs at day 9 (n = 5) and day 17 (n = 3), respectively, during superfusion with caffeine. This is consistent with a decrease in the transient frequency in the presence of caffeine (Fig. 3E). The data support the idea that although RyR-mediated Ca²⁺ release is not an absolute prerequisite for the maintenance of spontaneous activity, it increasingly contributes to the diastolic rise of the Ca²⁺ transient and thereby modulates the frequency.

Ca²⁺ extrusion mechanisms in ESdCs

The minor contribution of the RyR-mediated Ca²⁺ release to ECC in neonatal cardiomyocytes and ESdCs also suggests a minor role of the intracellular Ca²⁺ stores for the termination of the Ca²⁺ transient (Escobar et al. 2004). To determine the role of SERCA-mediated Ca²⁺ reuptake into the SR and of Ca²⁺ extrusion via NCX in the termination of the Ca²⁺ transient, we analysed the time constants of decay for twitch Ca²⁺ transients in control conditions (τ_T,_ctrl), for twitch transients in the presence of caffeine (10 mm) (τ_T,_caff) (Fig. 4A and B), and for the caffeine-induced Ca²⁺ transient (τ_caff). τ_T,ctrl significantly decreased during differentiation from day 9 (0.47 ± 0.06 s) to day 17 (0.23 ± 0.03 s; P < 0.05) but no significant developmental change could be detected in τ_T,caff (day 9, 0.86 ± 0.07 s; day 17, 0.95 ± 0.14 s, P > 0.5; Fig. 4B). We quantified the contribution of SERCA and NCX to the removal of Ca²⁺ by using the approach described by Fowler et al. (2005) where the contribution of SERCA is described as (k_twitch−k_caffeine)/k_twitch and the contribution of NCX as (k_caffeine−k_caffeine+Ni)/k_twitch (where k is rate constant; 1/τ). Ca²⁺ transients induced in the presence of nickel ([Ni⁺]_o = 10 mm) had a prolonged time constant of decay (τ_caff+Ni; Fig. 4C) due to inhibition of NCX. The analysis reveals that in 9 day old ESdCs, 46.5 ± 5% of Ca²⁺ extrusion is due to Ca²⁺ reuptake into caffeine-sensitive intracellular stores whereas 50.0 ± 8% is extruded via NCX (n = 6). This relation shifted in ESdCs at day 17, where 72 ± 5% of Ca²⁺ is taken up into intracellular stores and only 17 ± 2% is extruded by NCX (n = 9; Fig. 4D). The data support the notion that during developmental differentiation of ESdCs, intracellular Ca²⁺ stores gain importance for Ca²⁺ release during ECC and Ca²⁺ reuptake during the termination of the Ca²⁺ transient.

Figure 4.

Changing contribution of Ca²⁺ extrusion and Ca²⁺ reuptake to the Ca²⁺ transient decay A, whole-cell Ca²⁺ transients recorded under control conditions and in the presence of caffeine were normalized and superimposed to visualize the change in the time course of Ca²⁺ removal. B, the time constants of decay exhibited a significant developmental decrease (*P < 0.05). Superfusion with caffeine in both cases significantly increased the time-dependent decay of Ca²⁺ but abolished differences in time constant (τ)_T,caff between day 9 and day 17. C, confocal linescan image and spatially averaged [Ca²⁺]_i. The time constant of decay of the caffeine transient is further prolonged in the presence of nickel (10 mm). D, estimated contribution of sarco(endo)plasmic Ca²⁺ ATPase and Na⁺–Ca²⁺ exchanger to the removal of Ca²⁺ after a twitch transient in ESdCs at day 9 (n = 6) and day 17 (n = 9). The increased contribution from SERCA and decreased contribution from NCX was significant from day 9 to day 17 (**P < 0.005; *P < 0.05).

Spontaneous activity in early ESdCs depends on IP₃R-mediated Ca²⁺ release

Experiments in early embryonic heart showed that spontaneous activity depends on the presence of [Ca²⁺]_o. However, in previous experiments on early ESdCs the relevance of IP₃R-mediated Ca²⁺ release, and the independence of spontaneous Ca²⁺ transients from voltage-activated Ca²⁺ influx was proposed (Viatchenko-Karpinski et al. 1999; Mery et al. 2005). We aimed to determine the mechanism by which whole-cell Ca²⁺ transients are generated in ESdCs. Previously, a diastolic rise in intracellular Ca²⁺ has been postulated to promote spontaneous activity in atrial pacemaker cells (Huser et al. 2000; Bogdanov et al. 2001). To determine whether there is a correlation between the frequency of spontaneous Ca²⁺ transients and the diastolic rise of Ca²⁺, we buffered Ca²⁺ with the membrane-permeable Ca²⁺ chelator BAPTA/AM (1 μm; Fig. 5A) After 5 min of superfusion, basal [Ca²⁺]_i was decreased by 18.4 ± 1.6% together with a decrease in the frequency of spontaneous Ca²⁺ transients by 29 ± 5% (Fig. 5B; n = 2) indicating that [Ca²⁺]_i can modulate the frequency of the transients.

Figure 5.

Spontaneous activity in ESdCs is maintained by Ca²⁺ release from intracellular stores A, spatially averaged changes of [Ca²⁺]_i in an ESdC at 11 days during superfusion with the membrane-permeable Ca²⁺ chelator BAPTA/AM (1 μm). B, superfusion with BAPTA/AM (5 min) resulted in a significant decrease in the frequency of spontaneous whole-cell Ca²⁺ transients (n = 2). C, spatially averaged changes of [Ca²⁺]_i in ESdCs with the SERCA inhibitor thapsigargin (1 μm). In a total of three experiments, thapsigargin significantly decreased the decay of the spontaneous Ca²⁺ transients (at 5 min) and resulted ultimately (8 min) in a block of spontaneous whole-cell transients. Cessation of Ca²⁺ transients coincided with a depletion of the intracellular stores indicated by the lack of caffeine-induced Ca²⁺ transients.

Changes in the [Ca²⁺]_i can be due to Ca²⁺ influx through voltage-dependent Ca²⁺ channels and/or Ca²⁺ release from intracellular Ca²⁺ stores. To determine whether the rise in [Ca²⁺]_i originates from intracellular Ca²⁺ stores, we superfused spontaneously active cells with the SERCA inhibitor thapsigargin (1 μm) (Thastrup et al. 1990). Figure 5C shows a representative experiment. During the first period of superfusion (∼5 min), spontaneous activity continues in the presence of thapsigargin although a prolonged decay time of the Ca²⁺ transient could be observed. During prolonged application, Ca²⁺ transient frequency decreased and the transients ultimately ceased. At this point, the average increase in basal [Ca²⁺]_i was 5% (n = 3). After 8 min superfusion with thapsigargin and cessation of spontaneous activity, application of caffeine did not result in a Ca²⁺ transient. The experiments show that spontaneous Ca²⁺ transient frequency is modulated by the [Ca²⁺]_i and that the Ca²⁺ release from intracellular stores plays a role in the initiation of the spontaneous Ca²⁺ transients.

As spontaneous activity continues in the presence of caffeine we determined the relevance of IP₃R-mediated Ca²⁺ release by superfusion of spontaneously active ESdCs with the IP₃R blocker 2-aminoethoxydiphenyl borate (2-APB; 2 μm) (Wilcox et al. 1998; Bootman et al. 2002). In 26 out of 26 experiments the application of 2-APB resulted in a suppression (Fig. 6A) of the spontaneous whole-cell Ca²⁺ transients. Application of caffeine in the presence of 2-APB still resulted in a Ca²⁺ transient with an amplitude not significantly different from control conditions (Fig. 10A). The data suggest that the cessation of spontaneous activity was not due to a depletion of intracellular Ca²⁺ stores, or block of RyR-mediated Ca²⁺ release. To test whether 2-APB blocks voltage-gated Ca²⁺ entry (Bootman et al. 2002), we evaluated the effect of the 2-APB concentration used on field stimulation-induced Ca²⁺ transients in rat ventricular myocytes. Amplitude and duration of Ca²⁺ transients remained unchanged, supporting the hypothesis that block of spontaneous activity in ESdCs did not result from block of Ca²⁺ influx through voltage-dependent Ca²⁺ channels (data not shown).

Figure 6.

Whole-cell Ca²⁺ transients require IP₃R-mediated Ca²⁺ release Confocal linescan images and spatially averaged [Ca²⁺]_i in ESdCs superfused with the IP₃R blocker 2-APB (2 μm) (A) or the PLC inhibitor U73122 (1 μm) (B). Suppression of IP₃-mediated signalling in both cases resulted in a suppression of spontaneous activity which in the case of 2-APB was not due to a depletion of Ca²⁺ from intracellular stores, as indicated by the caffeine-induced Ca²⁺ transient

Figure 10.

Role of Ca²⁺ influx on the loading of Ca²⁺ stores and ESdCs spark and beating frequency A, caffeine transient amplitude (TA_caff) is not changed in the presence of 2-APB (2 μm) and nifedipine (10 μm) (ESdCs day 10). B, T-type Ca²⁺ channel blocker nickel (30 μm) only caused a depression of Ca²⁺ transient frequency (n = 4). C, Ca²⁺ spark frequency, as an indicator of the trigger of spontaneous whole-cell Ca²⁺ transients, remains unaffected in Ca²⁺-free solution (n = 2; P > 0.2) or in the presence of nifedipine (n = 2; P > 0.7).

Ca²⁺ transient frequency in ESdCs is modulated by PLC-mediated IP₃ production

Receptor-stimulated increase in IP₃ production depends on the activation of phospholipase C (PLC) where the hydrolysis of phosphatidylinositol (4,5) bisphospate results in the production of the second messengers IP₃ and diacylglycerol. Superfusion of ESdCs with the PLC inhibitor U73122 (1 μm) (Bootman et al. 2002) reversibly blocked spontaneous whole-cell Ca²⁺ transients (n = 6; Fig. 6B). By contrast, receptor-mediated stimulation of IP₃ production by the α-adrenergic receptor agonist phenylephrine (PE; 10 μm, n = 3) or endothelin-1 (ET-1; 100 nm, n = 4) had a positive chronotropic effect on spontaneous Ca²⁺ transients (Fig. 7A and B). PE and ET-1 both increase PLC activity via a G-protein (G_q)-dependent mechanism. In both cases the positive chronotropic effect was reversed in the presence of U73122 (1 μm; Fig. 7A and B). Block of the ET-1-induced positive chronotropic response by 2-APB (n = 3; Fig. 7C) underlines the relevance of IP₃R-mediated Ca²⁺ release in the translation of this signalling pathway. To determine whether ET-1 triggers IP₃-mediated Ca²⁺ release, we analysed its effect on basal [Ca²⁺]_i when spontaneous activity was suppressed by the application of tetracaine (1 mm) in Ca²⁺-free solution; at this concentration tetracaine not only blocks RyR-mediated Ca²⁺ release but also exhibits a non-specific effect on Ca²⁺ influx through voltage-dependent Ca²⁺ channels (Carmeliet et al. 1986). As shown in Fig. 7D, ET-1 increases basal Ca²⁺ in ESdCs under these conditions by 15.8 ± 0.03% (n = 4; Fig. 7E). The data support the notion that the positive chronotropic effect of ET-1 is mediated by IP₃R-mediated Ca²⁺ release.

Figure 7.

Stimulation of IP₃ production has a positive chronotropic effect on spontaneous Ca²⁺ transients in ESdCs A, phenylephrine (10 μm)-induced increase in Ca²⁺ transient frequency (n = 5) is reversed upon simultaneous superfusion with U73122 (1 μm; n = 2). B and C, ET-1 (100 nm) has a significant positive chronotropic effect (n = 4) on spontaneous Ca²⁺ transients that is reversed by superfusion with U73122 (B) (n = 2) or 2-APB (C) (n = 3). D, spatially averaged change of [Ca²⁺]_i induced by the application of ET-1 while spontaneous Ca²⁺ transients are suppressed by superfusion of the cells with tetracaine (1 mm) in Ca²⁺-free solution. E, ET-1 raised basal [Ca²⁺] by 15.8 ± 0.03% (n = 4) in comparison to the control in the presence of tetracaine plus Ca²⁺-free solution.

RyR channels amplify the IP₃R-mediated Ca²⁺ release

In ESdCs of all developmental stages, we detected spontaneous spatially restricted Ca²⁺ release events (Fig. 8A; and see Fig. 1A and B). The events had an amplitude (ΔF/F₀) of 2.1 ± 0.2, a width of 2.3 ± 0.18 μm and a duration of 31.4 ± 3.5 ms at half-maximal amplitude (Fig. 8G). The data are in good agreement with the parameters recorded previously for RyR-mediated Ca²⁺ sparks in ESdCs (Sauer et al. 2001). To test the hypothesis that IP₃R-mediated Ca²⁺ release is amplified by CICR from RyR channels, we determined whether the spark frequency is influenced by PLC activity or IP₃R-mediated Ca²⁺ release. Under control conditions, superfusion of spontaneously active ESdCs with the PLC inhibitor U73122 resulted in a significant decrease of the spark frequency (Fig. 8B and H). By contrast, the spark frequency increased upon stimulation with ET-1 (Fig. 8C and H). Again, this effect can be reversed by 2-APB (Fig. 8D and H). The experimental results support the hypothesis that IP₃R-mediated Ca²⁺ release can trigger RyR-mediated Ca²⁺ sparks by CICR. However, sparks did not seem to be a prerequisite for spontaneous Ca²⁺ transients. We also analysed Ca²⁺ release events that were recorded in the presence of caffeine (10 mm) when presumably RyR-operated stores are depleted. Figure 8F shows such a release event. The release events in the presence of caffeine had an amplitude (ΔF/F₀) of 1.15 ± 0.2, a half-maximal width of 2.03 ± 0.22 μm and a duration of 39.27 ± 2 ms. The release events detected under these conditions exhibited a significantly decreased amplitude (Fig. 8G).

Figure 8.

RyR-mediated Ca²⁺ sparks are promoted by IP₃-dependent Ca²⁺ release A, confocal linescan image and (bottom) local subcellular changes in [Ca²]_i (F/F₀ averaged over the area indicated by the arrow on the left of the linescan). B, the frequency of sparks was suppressed by application of the PLC inhibitor U73122 (1 μm; n = 2) and C, increased when IP₃ production was stimulated by ET-1 (100 nm; n = 2). D, the effect of ET-1 was reversible upon application of the IP₃R blocker 2-APB (n = 2). The data are summarized in H. E and F, confocal linescan images depicting a spatially restricted Ca²⁺ release event in the presence of caffeine (10 mm). G, release events in the presence of caffeine exhibit a significant decrease in the amplitude (*P < 0.002). In comparison to sparks they tend to have a similar width (FWHM) and an increased duration (FDHM); however, this difference did not reach significance in our analysis.

Translation of IP₃R-mediated Ca²⁺ release into whole-cell Ca²⁺ transients

Several studies have addressed the role of Ca²⁺ release from intracellular stores in the generation of spontaneous activity (Huser et al. 2000; Bogdanov et al. 2001; Lipsius et al. 2001). In these experimental models, it is proposed that localized Ca²⁺ release from intracellular stores can contribute in part to the diastolic depolarization due to activation of NCX, a further depolarization of the membrane potential and subsequent activation of voltage-dependent Ca²⁺ channels. We evaluated the role of Ca²⁺ influx by superfusing spontaneously active ESdCs with nominally Ca²⁺-free solution. As demonstrated in Fig. 9A, spontaneous Ca²⁺ transients in ESdCs at day 11 ceased immediately after exposure of the cells to Ca²⁺-free solution (n = 3). The lack of spontaneous Ca²⁺ transients was not due to depletion of the intracellular stores as indicated by the caffeine-induced Ca²⁺ transient after cessation of spontaneous activity. In addition, spontaneous transients resumed immediately with smaller amplitude.

Figure 9.

Whole-cell Ca²⁺ transients in ESdCs depend on Ca²⁺ influx through L-type Ca²⁺ channels A, confocal linescan images and spatially averaged changes in [Ca²⁺]_i of ESdCs during superfusion with Ca²⁺-free solution or in the presence of L-type Ca²⁺ channel blocker nifedipine (B and C). Superfusion with Ca²⁺-free solution (n = 4) or nifedipine (10 μm; n = 14) resulted in a block of spontaneous whole-cell Ca²⁺ transients. The spontaneous activity did not cease in Ca²⁺-free solution because of the depletion of the intracellular stores as indicated by the caffeine-induced Ca²⁺ transient.

To determine the mechanism of Ca²⁺ influx we superfused spontaneously active cells with nifedipine (10 μm; n = 14) or nickel (30 μm; n = 4), which are blockers of the voltage-dependent L-type and T-type Ca²⁺ channels, respectively. Figure 9B and C shows representative examples of spontaneously active ESdCs at day 9 and 14. Application of nifedipine resulted in a block of spontaneous activity. Also in these cases, RyR-regulated intracellular stores were not depleted (Fig. 10A). In contrast to the block of I_Ca,L, nickel had only a small attenuating effect on the Ca²⁺ transient frequency (Fig. 10B). The experimental results support the hypothesis that depolarization-dependent Ca²⁺ influx through L-type Ca²⁺ channels is not essential for maintaining spontaneous whole-cell Ca²⁺ transients in ESdCs.

IP₃R-mediated Ca²⁺ release events or puffs are smaller in amplitude than Ca²⁺ sparks and have a prolonged rise time and duration (Bootman et al. 1997; Zima & Blatter, 2004). Our experimental results suggest that localized RyR-mediated Ca²⁺ release events or sparks are promoted by the stimulation of IP₃R-mediated Ca²⁺ release. Therefore we used the more frequent RyR-mediated sparks as a parameter to determine whether IP₃R-mediated Ca²⁺ release persists even in the absence of whole-cell Ca²⁺ transients. For this purpose we analysed the spark frequency during superfusion of ESdCs with Ca²⁺-free solution (n = 2) or nifedipine (10 μm; n = 2). As shown in Fig. 10C neither in Ca-free conditions nor in the presence of nifedipine, a significant decrease in spark frequency in comparison to control conditions could be detected, although whole-cell Ca²⁺ transients ceased. The data support the hypothesis that IP₃R-mediated Ca²⁺ release and therefore the trigger for the spontaneous activity of the ESdCs persists even in the absence of Ca²⁺ transients.

Because our experimental results indicate that Ca²⁺ influx is required for whole-cell Ca²⁺ transients, we tested the hypothesis that, in contrast to previous reports (Viatchenko-Karpinski et al. 1999), membrane depolarization is required for spontaneous activity in ESdCs. For this purpose we depolarized ESdCs with a high-K⁺ ([K⁺]_o, 140 mm) solution. Superfusion with high- K⁺ solution resulted in an initial Ca²⁺ transient after which oscillatory changes of [Ca²⁺]_i were maintained over time (Fig. 11A). In comparison to the whole-cell Ca²⁺ transients recorded in the cell prior to the change of the extracellular solution, these oscillations had increased amplitudes, a reduced frequency, and a prolonged upstroke velocity (Fig. 11C). The increased rise time of 129.6 ± 19 ms (n = 2) in comparison to 41.2 ± 2.4 ms (n = 10) in transients in control solutions resulted from the non-instantaneous increase of [Ca²⁺]_i throughout the cell reflecting a propagating Ca²⁺ wave rather than a Ca²⁺ transient (Schlotthauer & Bers, 2000). The data indicate that Ca²⁺ signals recorded in ESdCs under depolarizing conditions are distinctly different from spontaneous whole-cell Ca²⁺ transients which we could not identify in high-K⁺ solution.

Figure 11.

Spontaneous whole-cell Ca²⁺ transients are replaced by Ca²⁺ waves when changes of the plasma membrane potential are suppressed A, superfusion of a spontaneously beating ESdC with depolarizing high-K⁺ solution resulted in the suppression of whole-cell Ca²⁺ transients but lead to the appearance of Ca²⁺ waves. B, confocal linescan image of an electrically coupled cell pair and the F/F₀ plot averaged over the entire width of the lower ESdC. Superfusion of the cell with low-Na⁺ solution suppressed whole-cell Ca²⁺ transients (*) that were synchronized between the two adjacent cells. Spontaneous release events transitioned into Ca²⁺ waves (•) that were confined to the boundaries of the individual cell. C, Ca²⁺ waves in the presence of LiCl-containing (n = 5) and KCl-containing solution (n = 1) could be distinguished from whole-cell Ca²⁺ transients under control conditions (n = 3) because of their significantly reduced upstroke velocity (P < 0.05).

To determine whether NCX is the mechanism by which intracellular Ca²⁺ release is translated into a depolarization of the membrane potential, we superfused the cells with a low-Na⁺ solution. Superfusion with low-Na⁺ resulted in the instantaneous arrest of spontaneous Ca²⁺ transients; however, Ca²⁺ release events were observed that transitioned into intracellular Ca²⁺ waves (n = 5). The experiment in Fig. 11B was obtained from a cell pair and exhibits the difference between the two Ca²⁺ signals. Whole-cell Ca²⁺ transients (asterisk in Fig. 11B) were synchronized between the two cells with a propagation delay of less than 4.4 ms. This indicates that Ca²⁺ transient synchronization is supported by the intercellular propagation of an action potential. In contrast, the localized release events or Ca²⁺ waves (circle in Fig. 11B;) that appeared in low-Na⁺ solution ([Na⁺]_o, 12 mm), were restricted to the perimeter of an individual cell underlining the fact that intracellular Ca²⁺ release was not translated into a membrane depolarization and subsequent intercellular action potential propagation. Intracellular Ca²⁺ waves in low-Na⁺ solution were not blocked by 2-APB (data not shown) indicating that they do not depend on the release of Ca²⁺ through IP₃Rs. Our experimental results indicate that spontaneous intercellularly propagating whole-cell Ca²⁺ transients depend on a depolarization of V_m potentially brought about by NCX, and subsequent voltage-dependent Ca²⁺ influx. In the absence of IP₃R-initiated depolarization, RyR-mediated Ca²⁺ release results in Ca²⁺ waves with properties distinct from the pacemaker mechanism.

Discussion

In this study we present experimental evidence that the developmental changes of spontaneous whole-cell Ca²⁺ transients in ESdCs are explained by an increased contribution of RyR-mediated Ca²⁺ release and a shift of the Ca²⁺ removal mechanisms from Ca²⁺ extrusion by NCX to Ca²⁺ reuptake into intracellular stores by SERCA. We further present evidence that spontaneous whole-cell Ca²⁺ transients in ESdCs are initiated by IP₃-mediated Ca²⁺ release which is translated into whole-cell Ca²⁺ transients by a NCX-mediated depolarization and subsequent induction of voltage-dependent Ca²⁺ influx through L-type channels.

In adult cardiac myocytes, the whole-cell Ca²⁺ transient is initiated by Ca²⁺ influx through voltage-dependent L-type Ca²⁺ channels; however, the major increase of [Ca²⁺]_i depends on the subsequent CICR from RyR-operated intracellular Ca²⁺ stores (Fabiato, 1985a,Fabiato, 1985b). The contribution of Ca²⁺ release from the SR for twitch transients constitutes 90% in adult mouse ventricular myocytes (Li et al. 1998), whereas in whole heart experiments with neonatal rat myocytes a contribution as low as 5% has been described (Escobar et al. 2004). In ESdCs at day 17 the twitch transient amplitude decreased by 70% in the presence of caffeine which correlates well with the observation that 72% of the Ca²⁺ removal after a twitch is contributed by SERCA. The data represent a significant increase in the relevance of RyRs and SERCA during ESdC differentiation. One reason for the low contribution of RyRs to the Ca²⁺ transient could be the lack of functional coupling between voltage-dependent Ca²⁺ influx and the RyR channel release site (Franzini-Armstrong et al. 2005). In ESdCs, early RyR channel expression and its developmental up-regulation has been described (Yang et al. 2002). These data support our experimental results indicating an age-dependent increase of the caffeine transient amplitude. As RyR-mediated Ca²⁺ release also contributes to the diastolic increase in [Ca²⁺]_i, its up-regulation could also explain the developmental increases in the frequency of the spontaneous transients.

Although IP₃Rs are present in adult ventricular muscle (Bare et al. 2005), their Ca²⁺ release seems to contribute only to a small degree to the excitation-induced Ca²⁺ transient (Proven et al. 2006). Our experimental results support the hypothesis that spontaneous whole-cell Ca²⁺ transients in ESdCs depend on IP₃ production and IP₃R-mediated Ca²⁺ release (see Figs 6A and B, and 8). This is in agreement with previous reports that in ESdCs spontaneous whole-cell Ca²⁺ transients are inhibited by the IP₃R blocker xestospongin and by IP₃R down-regulation via IP₃R type1 antisense cDNA (Mery et al. 2005). Here we show that promoting (ET-1 and PE) or suppressing (U73122) IP₃ production in ESdCs changes the chronotropic response of the spontaneous whole-cell Ca²⁺ transients. The reported interdependence of IP₃ production and RyR-mediated Ca²⁺ sparks was previously described in adult atrial myocytes of cat and mice where ET-1 increased the spark frequency by IP₃R-mediated Ca²⁺ release (Zima & Blatter, 2004; Li et al. 2005). Spontaneous spatially restricted Ca²⁺ release events from IP₃R channels known as blips or puffs differ from Ca²⁺ sparks by their amplitude and kinetics (Bootman et al. 1997; Zima & Blatter, 2004). The events are difficult to distinguish when they appear simultaneously. Although we can detect the ET-1-mediated increase in basal [Ca²⁺]_i, the appearance of individual release events during suppression of Ca²⁺ sparks by caffeine were rare and had a lower amplitude and longer duration than Ca²⁺ sparks (see Fig. 8G).

Ca²⁺ release from intracellular stores has been postulated previously as a mechanism contributing to spontaneous activity (Viatchenko-Karpinski et al. 1999; Huser et al. 2000; Bogdanov et al. 2001). In contrast to ESdCs (see Figs 2 and 3) (Yang et al. 2002), in these cases RyR channels were described as the source for the increase in diastolic calcium. Continuing spontaneous activity in the presence of ryanodine was recorded in sinus nodal cells (Boyett et al. 2003); however, voltage-activated ion channels are proposed as the driving force for the diastolic depolarization. In ESdCs the presence of voltage-dependent ion channels, such as I_f, I_Ca,L and I_Ca,T, has been described but neither block of I_f nor block of I_Ca,T in early ESdCs suppressed their spontaneous activity (Zhang et al. 2003; Mery et al. 2005). Our results show a cessation of spontaneous whole-cell Ca²⁺ transients upon block of I_Ca,L at all developmental stages (Fig. 9B and C). This result is contrary to prior studies where spontaneous Ca²⁺ transients in early ESdCs were described to be independent of voltage-dependent Ca²⁺ influx. We propose that the spontaneous whole-cell Ca²⁺ transients under depolarizing conditions represent Ca²⁺ waves (see Fig. 11) rather than whole-cell Ca²⁺ transients, and are based on CICR that originates as a result of Ca²⁺ overload of the intracellular stores rather than IP₃R-mediated Ca²⁺ release that promotes spontaneous activity.

The mechanisms that promote the oscillatory changes in spontaneous Ca²⁺ release are still under debate. In latent pacemaker cells the depolarization-induced Ca²⁺ influx through T-type Ca²⁺ channels was proposed as a potential mechanism (Huser et al. 2000). In rabbit SA nodal cells, however, the rhythmic increase in RyR release events seems to be promoted by protein kinase A-modulated filling of the intracellular stores (Bogdanov et al. 2001; Vinogradova et al. 2005; Vinogradova et al. 2006). In comparison to our study and to reports from embryonic cardiomyocytes (Koushik et al. 2001), SR-dependent diastolic increase of [Ca²⁺]_i promotes the depolarization of the plasma membrane by NCX activation (Fig. 9 and 11). The subsequent activation of I_Ca,L then leads to the rapid upstroke of the Ca²⁺ transients. In addition, recent studies on guinea-pig cardiomyocytes have suggested a contribution of NCX to the pacemaker activity as a result of a Na⁺ leak current (Sanders et al. 2006).

It remains to be determined where the IP₃R-mediated Ca²⁺ release occurs to promote sufficient depolarization of V_m to trigger the voltage-dependent Ca²⁺ influx. Release events appear to be sparse and no distinct spatial location could yet be determined. All three IP₃R subtypes (1–3) are expressed in undifferentiated embryonic stem cells (Kapur et al. 2006) and embryonic cardiomyocytes (Rosemblit et al. 1999; Jaconi et al. 2000; Puceat & Jaconi, 2005). Immunostaining revealed that IP₃R-1 is most prominent within the nuclear envelope (Jaconi et al. 2000; Mery et al. 2005). However, one could speculate that a subsarcolemmal location of the Ca²⁺ release site would be more efficient to promote plasma membrane depolarization. Our experiments demonstrate that spontaneous activity can continue even after depletion of RyR-sensitive stores by caffeine (Fig. 3A and B). Although the fact that spark activity strongly depends on IP₃ production indicates a close proximity of the two release channels, the continuation of spontaneous activity in the presence of caffeine would indicate a functional separation of IP₃R- and RyR-controlled stores. In adult ventricular myocytes, the interconnectivity of the store system is supported by experiments showing refilling of stores through intra-SR diffusion (Wu & Bers, 2006). Future experiments will need to address the spatial distribution and the functional relation of IP₃R- and RyR-controlled Ca²⁺ stores.

Conclusion

Our experimental results identify a developmental change in the functional contribution of RyR-mediated Ca²⁺ release to spontaneous whole-cell Ca²⁺ transients in ESdCs and provide evidence that the pacemaker mechanism in early ESdCs depends on IP₃R-mediated Ca²⁺ release that is translated into a depolarization of V_m and subsequent voltage-dependent Ca²⁺ influx by I_Ca,L. The Ca²⁺ handling mechanisms in ESdCs therefore represent a model of how IP₃R-mediated Ca²⁺ release can modulate spontaneous activity in cardiomyocytes or promote the appearance of arrhythmic activity.