How does β-adrenergic stimulation increase the heart rate? The role of intracellular Ca²⁺ release in amphibian pacemaker cells

Abstract

1. The mechanism by which sympathetic transmitters increase the firing rate of pacemaker cells was explored in isolated cells from the sinus venosus of the cane toad Bufo marinus. Intracellular calcium concentration ([Ca²⁺]_i) was measured with indo-1 and membrane potential and currents were recorded with the nystatin perforated-patch technique.

2. Adrenaline or isoprenaline (2 μM) increased the transient rise in [Ca²⁺]_i and increased the firing rate; these effects were blocked by propranolol (2 μM).

3. To determine whether the changes in [Ca²⁺]_i might influence the firing rate we studied agents which affect either the loading or the release of Ca²⁺ from the sarcoplasmic reticulum (SR). Rapid application of caffeine (10 mM) to spontaneously firing cells caused a large Ca²⁺ release from the SR and the cells were then quiescent for 24 s. In the presence of β-adrenergic stimulation the caffeine-induced [Ca²⁺]_i was 14 % larger but the period of quiescence after application was reduced to 12 s.

4. Ryanodine, at either low (1 μM) or high (> 10 μM) concentration, stopped firing. However, when the SR store content of Ca²⁺ was tested with caffeine, at low ryanodine concentration the SR Ca²⁺ store was empty whereas at the high concentration the SR store was still loaded with Ca²⁺. β-Adrenergic stimulation was not able to restore firing at the low concentration of ryanodine but did restore firing at the high ryanodine concentration.

5. An SR Ca²⁺ pump blocker, 2,5-di(tert-butyl)-1,4-hydroquinone (TBQ) which depletes the SR store of Ca²⁺, also rapidly and reversibly stopped spontaneous firing.

6. The relation between the amplitude of the [Ca²⁺]_i transient and firing rate established in the presence of ryanodine was similar when firing was restored by β-stimulation.

7. In both spontaneously firing and voltage-clamped cells, depleting the SR store with either ryanodine or TBQ suggested that about half of the Ca²⁺ which contributes to the calcium transient is released from the SR.

8. These results show that the amplitude of the [Ca²⁺]_i transient is an important factor in the firing rate of toad pacemaker cells and consequently agents which modify SR Ca²⁺ release influence firing rate. The effects of β-stimulation on firing rate seem to be largely mediated by changes in amplitude of the [Ca²⁺]_i transient.

It is generally thought that the increase in heart rate after β-adrenergic stimulation is caused by modulation of ionic channels located in the surface membrane (for review see DiFrancesco, 1993; Irisawa et al. 1993). The channels most likely to be involved are those which contribute to the pacemaker activity and which are affected by β-adrenergic stimulation. These include the L-type calcium current I_Ca(L) (Noma et al. 1980), the delayed rectifier current I_K (Brown et al. 1979) and the hyperpolarization-activated cation channel I_f (DiFrancesco, 1981). In particular DiFrancesco (1995) has argued that I_f is a major pacemaker current in mammalian pacemaker cells and that the increase in magnitude of I_f caused by β-adrenergic stimulation (Yatani et al. 1990) underlies the acceleration of firing rate caused by β-adrenergic stimulation.

β-Adrenergic stimulation also increases the amplitude of the systolic rise in intracellular Ca²⁺ concentration (the [Ca²⁺]_i transient) in atrial cells (Allen & Blinks, 1978), ventricular cells (Hussain & Orchard, 1997) and atrioventricular pacemaker cells (Hancox et al. 1994). This raises the possibility that the increase in [Ca²⁺]_i might contribute to the increase in firing rate. A number of studies have shown that interventions which alter the [Ca²⁺]_i in pacemaker cells can cause changes in the firing rate (Rubenstein & Lipsius, 1989; Rigg & Terrar, 1996; Ju & Allen, 1998; Li et al. 1998). For example, Ju & Allen (1998) found that ryanodine, an agent that interferes with sarcoplasmic reticulum (SR) Ca²⁺ release, gradually reduced the amplitude of the [Ca²⁺]_i transient while slowing, and finally stopping, the spontaneous firing of action potentials. It is known that pacemaker cells contain a Na⁺-Ca²⁺ exchanger (Zhou & Lipsius, 1993; Ju & Allen, 1998) and that extrusion of Ca²⁺ from the cell will generate an inward Na⁺-Ca²⁺ exchange current I_NaCa (Kimura et al. 1986; Callewaert et al. 1989). Thus it is proposed that [Ca²⁺]_i modulates I_NaCa which acts as a pacemaker current (Zhou & Lipsius, 1993; Li et al. 1998; Ju & Allen, 1998).

In the present study we have attempted to determine whether some part of the effect of β-adrenergic stimulation on the firing rate of pacemaker cells arises through the changes in [Ca²⁺]_i. We measured [Ca²⁺]_i and membrane potential in single pacemaker cells isolated from the sinus venosus of cane toads (Bufo marinus) and used agents which modify the Ca²⁺ loading of the SR. It is important to note that this preparation appears to have no I_f (Ju et al. 1995), thus the ‘increased I_f mechanism’ (DiFrancesco, 1995) is unlikely to be important in this tissue. Our main findings are that in the toad pacemaker cells the firing rate was dependent on the amplitude of the [Ca²⁺]_i transient and that consequently agents which modify SR Ca²⁺ release affect the firing rate. β-Adrenergic stimulation increased the amplitude of the [Ca²⁺]_i transient and much of the increase in firing rate caused by β-stimulation seems to occur through this mechanism.

METHODS

Preparation of pacemaker cells

Single pacemaker cells were enzymatically isolated from the sinus venosus of toads (Bufo marinus) as previously described (Ju et al. 1995). These experiments were approved by the Animal Ethical Committee of the University of Sydney. Briefly, the toad was anaesthetized by immersion in 0.5 % tricaine methanesulphonate in water and killed by decapitation followed by pithing. The heart was then rapidly removed and the sinus venosus region was dissected free. In order to isolate pacemaker cells, the sinus venosus was incubated at 28°C in a low Ca²⁺ solution first with collagenase and then with elastase. Mild trituration produced many long, spindle-shaped cells some of which survived addition of Ca²⁺ to the perfusate. We studied only those cells that beat spontaneously and displayed a smooth surface membrane, regarding these as pacemaker cells. We have found no evidence for different classes of pacemaker cells such as is described in mammalian sino-atrial node. The cells were routinely superfused with the following (standard) solution (mM); NaCl, 110; KCl, 2.5; MgSO₄, 0.5; CaCl₂, 2; Na-N-2-hydroxyethylpiperazine-N’-2-ethanesulphonic acid (Hepes), 10; glucose, 10; pH 7.3; equilibrated with air. Drugs were applied from a fine tube positioned within 200 μm of the cell to ensure a rapid onset of action. All experiments were performed at room temperature (22°C).

Fluorescence measurements

After isolation, cells were loaded with 5 μM indo-1 AM for 10-15 min as previously described in detail (Ju & Allen, 1998). Loaded cells were illuminated at 360 ± 5 nm with a UV light source whose intensity was reduced 30-fold with a neutral density filter. The emitted fluorescence was guided to two photomultiplier tubes with either a 400 ± 5 nm or a 510 ± 5 nm interference filter in front of their photocathodes. The light signal at each wavelength was filtered at 10 Hz and the background subtracted. The analog signals were digitized and the ratio of fluorescence signals at 400 nm/510 nm (R) was calculated and converted to [Ca²⁺]_i using the following equation (Grynkiewicz et al. 1985):

An in vivo calibration method was used and gave the following values: R_min, 0.11; R_max, 1.94; β, 2.84; K_d, 606 nM (Ju & Allen, 1998).

Patch-clamp procedure

The nystatin perforated-patch technique was used to record spontaneous action potentials. The pipette solution contained (mM): KCl, 100; KH₂PO₄, 10; N-tris(hydroxymethyl)methyl-2-aminoethanesulphonic acid (Tes), 10; MgSO₄, 2; pH 7.3. Nystatin (240 μM) was added to the pipette solutions. An Axopatch 200A (Axon Instruments, Foster City, CA, USA) was used in current-clamp mode to record membrane potentials. The series resistance of perforated patches was < 20 MΩ; cell capacitances was 30-50 pF. Membrane currents and fluorescence signals were sampled at 2 kHz; membrane potentials were sampled at 0.5 kHz.

Statistics

All statistical data are presented as means ±s.e.m. with the number of cells studied as n. Statistical tests were either Student's paired or unpaired t tests and P < 0.05 was taken as the level of significance.

RESULTS

β-Adrenergic stimulation increases peak [Ca²⁺]_i and the firing rate

The amplitude of the [Ca²⁺]_i transient and the firing rate were recorded from single pacemaker cells loaded with indo-1. To explore some of the effects of sympathetic activation we used either adrenaline (2 μM), one of the sympathetic transmitters in the frog, or isoprenaline (2 μM), a predominantly β-agonist. Both drugs had similar effects in all situations we tested and we will report their results together. β-Adrenergic stimulation increased both the amplitude of [Ca²⁺]_i transients and the firing rate and a representative example of the effects of adrenaline is shown in the first two panels of Fig. 1A. In 23 cells systolic [Ca²⁺]_i was increased from 470 ± 35 to 864 ± 102 nM (P < 0.001) while the firing rate (action potentials min⁻¹) was increased by 41 ± 3 % (from 22 ± 1 to 31 ± 2 min⁻¹; P < 0.01). Changes in diastolic [Ca²⁺]_i were very small and not apparent in all experiments; in 12 experiments the mean of control and recovery diastolic [Ca²⁺]_i was 128 ± 11 nM while diastolic [Ca²⁺]_i in the presence of β-stimulation was 144 ± 15 nM (P < 0.05). β-Stimulation also increased the rate of decline of [Ca²⁺]_i from its peak. In control conditions the decay of [Ca²⁺]_i was roughly exponential with a time constant of 470 ± 30 ms; after β-stimulation this time constant decreased to 200 ± 33 ms.

Figure 1. Bath-applied adrenaline increases the firing rate and the amplitude of [Ca²⁺]_i transients in isolated toad sinus venosus cells.

A, spontaneous [Ca²⁺]_i transients recorded under control conditions (left panel), after application of 2 μM adrenaline (ADR) (middle panel) and after the addition of 2 μM propranolol (PROP) (right panel). Note the changes of amplitude and firing rate of the [Ca²⁺]_i transients in the presence of adrenaline and that both effects are largely reversed by propranolol. B, mean results from 5 cells showing an increase in firing rate caused by adrenaline and reversal by propranolol. * Significantly different from control with P < 0.05. C, mean results from 5 cells showing increase in the peak of the [Ca²⁺]_i transient caused by adrenaline and reversal by propranolol. * Significantly different from control with P < 0.05.

In order to confirm that adrenaline and isoprenaline were acting via β-adrenergic receptors we used the β-antagonist propranolol (2 μM). The third panel of Fig. 1A shows that application of propranolol to a cell already stimulated by adrenaline largely reverses the increase in [Ca²⁺]_i transients and firing rate. Addition of 2 μM propranolol to the control solution (not shown) had no effect on the firing rate and systolic [Ca²⁺]_i. Figure 1B and 1C shows mean data from five cells on firing rate and amplitude of the [Ca²⁺]_i transient and confirm that the effects of adrenaline are reversed by propranolol. Propranolol also reversed the effects of isoprenaline. Thus bath-applied adrenaline and isoprenaline are acting via β-receptors in contrast to the effects of sympathetic nerve stimulation in the intact sinus venosus of the toad where the receptors are not of the α- or β-type (Bramich et al. 1993).

The possible role of I_f in β-adrenergic stimulation of firing rate is unclear in amphibian pacemaker cells (Irisawa et al. 1993). The presence of I_f seems to be variable among different amphibian species; it has been reported as absent in Rana catesbeiana (Giles & Shibata, 1985) and Bufo marinus (Ju et al. 1995) while present in Rana esculenta (Bois & Lenfant, 1991). To further exclude the presence of I_f in the present preparation we examined the effect of Cs⁺ (2 mM) which is an established blocker of I_f (Irisawa et al. 1993). In four cells the mean firing rate was 23.2 ± 0.8 min⁻¹ and Cs⁺ had no significant effect on the rate (-0.7 ± 0.6 min⁻¹). In the presence of adrenaline the rate increased significantly to 33 ± 3 min⁻¹ and Cs⁺ again produced no significant change in the rate (0 ± 1.2 min⁻¹). These data provide further evidence that the present preparation had no functional I_f.

Firing of pacemaker cells requires Ca²⁺ in the sarcoplasmic reticulum

We show above that β-stimulation increases the [Ca²⁺]_i transients and this raises the possibility that the increased [Ca²⁺]_i transients have some role in the increased firing rate. To explore this possibility we first used caffeine which increases the opening probability of SR Ca²⁺ release channels (Rousseau & Meissner, 1989) and depletes the SR of Ca²⁺ (Callewaert et al. 1989).

Figure 2 shows the [Ca²⁺]_i response to rapid application of caffeine (10 mM) in a spontaneously firing pacemaker cell. In Fig. 2A the first two peaks are spontaneous [Ca²⁺]_i transients under control conditions. Caffeine (10 mM) caused a large and rapid rise in [Ca²⁺]_i which then fell spontaneously in the continuing presence of caffeine. This result is similar to the effect of caffeine on cells voltage clamped at -60 mV (Ju & Allen, 1998). However in this non-voltage-clamped cell, spontaneous [Ca²⁺]_i transients ceased for about 20 s before spontaneous firing recovered, initially at a slower rate. In nine cells, the peak of the caffeine-induced [Ca²⁺]_i signal was 1410 ± 202 nM so that the amplitude (peak - diastolic) was 1250 ± 210 nM. The amplitude of the caffeine-induced signal was 5.3 ± 0.8 times the amplitude of the [Ca²⁺]_i transient induced by the action potential. The time for the recovery of firing after caffeine application was 23.9 ± 4.5 s. This time may reflect the duration of SR refilling with Ca²⁺ (Hussain & Orchard, 1997) and suggests that spontaneous firing may be partly dependent on SR Ca²⁺ content.

Figure 2. Effect of rapid application of caffeine on [Ca²⁺]_i and firing rate in the presence and absence of β-adrenergic stimulation.

A, continuous record of [Ca²⁺]_i in a spontaneously firing cell. Caffeine caused a large increase in [Ca²⁺]_i which spontaneously declined in the continuing presence of caffeine. After caffeine was washed off the cell did not fire spontaneously for about 20 s. B, the same cell in the presence of adrenaline (2 μM). Note that this had increased the amplitude and firing rate of the [Ca²⁺]_i transients. Caffeine produced a bigger [Ca²⁺]_i increase but a shorter period of quiescence afterwards.

Figure 2B shows that adrenaline (2 μM) increased the spontaneous firing rate and the amplitude of [Ca²⁺]_i transients in the same cell as shown in Fig. 2A. Rapid application of caffeine caused a [Ca²⁺]_i signal which was usually somewhat bigger than in control. In nine cells the peak caffeine-induced [Ca²⁺]_i was 1605 ± 187 nM (amplitude = 1450 ± 200 nM) which is significantly larger than control (P < 0.05). However, the ratio of the peak amplitude of the caffeine-induced [Ca²⁺]_i to the spontaneous [Ca²⁺]_i transient was significantly smaller than control at 2.8 ± 0.3 (P < 0.01). In the presence of adrenaline the time required for recovery of spontaneous firing after caffeine application was 12.5 ± 2.6 s, which was faster than control (P < 0.01). The mechanism for faster recovery of spontaneous firing during β-stimulation is not clear; it may be related to the modulation of pacemaker currents (DiFrancesco, 1995), the rate of SR refilling with Ca²⁺ (Hussain & Orchard, 1997), or to an increased sensitivity of the Ca²⁺ release channels to the triggering Ca²⁺(Patel et al. 1995).

The effects of β-adrenergic stimulation on [Ca²⁺]_i in ryanodine-treated cells

The above experiment shows that depleting the SR Ca²⁺ store by caffeine application causes a temporary halt to spontaneous firing in the absence or presence of β-stimulation. However, caffeine has a variety of pharmacological effects, so we investigated ryanodine as a second agent which can affect SR Ca²⁺ content. The effect of ryanodine on SR Ca²⁺ release channels is dependent on its concentration (for review see Fleischer & Inui, 1989). High concentrations of ryanodine (> 10 μM) directly block SR Ca²⁺ release channels whereas low concentrations of ryanodine (< 1 μM) open channels in a subconductance state (Rousseau et al. 1987).

In the first series of experiments a low concentration of ryanodine (1 μM) was applied to spontaneously firing cells. A representative experiment is shown in Fig. 3 in which panel A shows [Ca²⁺]_i transients recorded before the application of ryanodine. As previously described (Ju & Allen, 1998), there was often a small initial increase in [Ca²⁺]_i and firing rate but after 30 min of exposure to ryanodine [Ca²⁺]_i transients were abolished (Fig. 3B). Rapid application of caffeine (Fig. 3C) showed a small increase in noise that was presumably due to the quenching of the fluorescence signal by caffeine (O'Neill et al. 1990), but there was no substantial increase in [Ca²⁺]_i. This result indicated that the SR Ca²⁺ store was empty as expected. When isoprenaline (2 μM) was added to the perfusate of the ryanodine-treated cells, there was no change in the resting [Ca²⁺]_i, the cells remained quiescent and Ca²⁺ transients were not detectable (Fig. 3D). Similar results were observed in five cells.

Figure 3. Low concentration of ryanodine abolishes spontaneous firing and empties the SR of Ca²⁺.

A, spontaneous [Ca²⁺]_i transients under control conditions. B, 1 μM ryanodine abolishes [Ca²⁺]_i transients after 30 min exposure. C, rapid exposure to caffeine (10 mM) produces no substantial increase in [Ca²⁺]_i. The increase in noise with the same time course as the caffeine exposure is probably caused by the quenching of indo-1 by caffeine. D, exposure to isoprenaline (2 μM) produced no obvious change in the [Ca²⁺]_i.

In the second series of experiments, in which the cells were treated with a high concentration of ryanodine (10-20 μM), the results were quite different. Although spontaneous [Ca²⁺]_i transients were also abolished after ryanodine treatment (10 μM) (Fig. 4A and B), rapid application of caffeine was still able to trigger a large Ca²⁺ release as shown in Fig. 4C. The amplitude of the caffeine-induced [Ca²⁺]_i was 1010 ± 170 nM (n = 6) which compares with no detectable rise in the cells treated with the low concentration of ryanodine (Fig. 3C). This result indicates that the Ca²⁺ locked in the SR Ca²⁺ store by the high concentration of ryanodine is able to be released by caffeine, confirming previous reports (Kondo, 1988). Surprisingly, we found that when these quiescent cells were superfused with 2 μM isoprenaline, there was first a small increase in resting [Ca²⁺]_i from 94 ± 6 to 104 ± 7 nM (P < 0.05). Thereafter [Ca²⁺]_i transients became apparent and gradually increased in amplitude despite the presence of ryanodine as shown in Fig. 4D. In six cells, the peak of [Ca²⁺]_i transient restored by β-stimulation in the presence of ryanodine was 300 ± 26 nM.

Figure 4. High concentrations of ryanodine abolish spontaneous firing but the SR remains loaded with Ca²⁺ and isoprenaline can cause spontaneous firing to recover.

A, spontaneous [Ca²⁺]_i transients under control conditions. B, 10 μM ryanodine abolishes [Ca²⁺]_i transients after 10 min. C, rapid exposure to caffeine causes a large increase in [Ca²⁺]_i. D, exposure to 2 μM isoprenaline causes a resumption of spontaneous [Ca²⁺]_i transients.

The effects of ryanodine and β-adrenergic stimulation on the action potential

In our previous study (Ju & Allen, 1998) we showed that action potentials and [Ca²⁺]_i transients always accompanied one another. However when the [Ca²⁺]_i transients become very small, as occurs in ryanodine, it becomes harder to reliably detect [Ca²⁺]_i transients so it was important to establish that similar results were obtained in experiments in which the membrane potential was recorded.

Spontaneous action potentials were measured in the eight cells which were not indo-1 loaded and had a maximum diastolic potential of -49 ± 2 mV, a peak potential of 27 ± 4 mV and the firing rate was 40 ± 4 min⁻¹. Note that the lower firing rate in the indo-1-loaded cells (22 ± 1 min⁻¹) compared with the unloaded cells probably reflects the buffering effect of indo-1 (Ju & Allen, 1998). In the presence of adrenaline or isoprenaline the maximum diastolic potential was -55 ± 4 mV, and the peak of the action potential was 44 ± 5 mV, which are similar to previously reported results on the effects of bath-applied adrenaline on the action potential in the toad sinus venosus (Bramich et al. 1993). Ryanodine, at concentrations between 1 and 20 μM, arrested action potentials in all cells within 7-30 min with the higher concentrations taking the shorter times. The resting membrane potentials in arrested cells averaged -46 ± 5 mV. Application of isoprenaline (2 μM) restored action potentials in all cells (5 out of 5) treated with high concentrations of ryanodine (10-20 μM) but in none of the cells (0 out of 3) treated with a low concentration of ryanodine (1 μM). Figure 5A shows the membrane potential recorded in a cell under control conditions. The action potentials were arrested after 8 min ryanodine (20 μM) application (Fig. 5B). After application of isoprenaline (2 μM) the first noticeable effect was an increase in membrane noise after about 2 min (Fig. 5C) which was followed within 20-30 s by small action potentials which rapidly increased in amplitude (Fig. 5D). The spontaneous action potentials rescued by isoprenaline in the presence of ryanodine had a maximum diastolic potential of -51 ± 3 mV and peak potentials of 29 ± 5 mV, values which are not significantly different to control measurements.

Figure 5. Recording of action potentials only in the presence of drugs which modify SR Ca²⁺ release.

A, action potential firing under control conditions. B, 8 min after application of 20 μM ryanodine. C, 2 min after application of isoprenaline (2 μM) the membrane potential starts to show an increase in noise. D, follows C with a 5 s gap and shows the first clear action potentials appearing after the addition of isoprenaline. E, after the action potentials had fully recovered, application of TBQ, an SR Ca²⁺ pump inhibitor, caused action potentials to decline. F, after 1 min exposure to TBQ firing has ceased. G, 1 min after washing off TBQ membrane potential starts to show increased noise and firing restarts. H, stable spontaneous action potentials again in the continuing presence of ryanodine and isoprenaline.

To further establish the role of the SR Ca²⁺ store in pacemaker activity we used 2,5-di(tert-butyl)-1,4-hydroquinone (TBQ), which reversibly blocks the SR Ca²⁺ pump in muscle (Westerblad & Allen, 1994). Since the SR Ca²⁺ leak persists after the inhibition of the SR Ca²⁺ pump this agent leads to loss of SR Ca²⁺. Figure 5E shows that 1 min after application of TBQ (10 μM) the action potentials started to decline in amplitude and shortly after the cell stopped firing (Fig. 5F). When TBQ was washed off, again there was some evidence of an increase in membrane potential noise before the cell started to fire action potentials once more. Similar results were observed in two other cells and we also found that TBQ could reversibly prevent firing in control cells (not shown) confirming earlier observations in guinea-pig sino-atrial tissue (Rigg & Terrar, 1996).

In three experiments we showed that the β-adrenergic antagonist propranolol prevented the resumption of firing by isoprenaline in the presence of ryanodine (20 μM). In each of these experiments isoprenaline restored firing after the propranolol was washed off.

We also wished to confirm our previous observation (Ju & Allen, 1998) that action potentials and [Ca²⁺]_i transients always occur together even in the presence of ryanodine and isoprenaline. In a single experiment we simultaneously recorded spontaneous action potentials and [Ca²⁺]_i (Fig. 6). Figure 6A shows control spontaneous firing. After 6 min of application of 20 μM ryanodine, both the amplitude and the frequency of the [Ca²⁺]_i transients were substantially reduced. Simultaneously recorded action potentials also showed a reduced amplitude but each [Ca²⁺]_i transient was still associated with an action potential (Fig. 6B). After 13 min of ryanodine application, neither [Ca²⁺]_i transients nor action potentials were detectable (Fig. 6C). However, following a 2 min application of isoprenaline in the continuing presence of ryanodine, both [Ca²⁺]_i transients and action potentials reappeared (Fig. 6D). After a further 4 min, the firing rate and the amplitude of the [Ca²⁺]_i transients had both partially recovered towards control values (Fig. 6E). This experiment confirms that action potentials and [Ca²⁺]_i transients are equivalent markers of pacemaker firing.

Figure 6. Simultaneous recording of action potentials and [Ca²⁺]_i during exposure to ryanodine and then isoprenaline.

A, spontaneous action potentials and [Ca²⁺]_i transients under control conditions. B, after 6 min exposure to ryanodine (20 μM). C, 13 min exposure to ryanodine. D, 2 min of exposure to isoprenaline (2 μM) in the continuing presence of ryanodine. This trace shows the first reappearance of action potentials. E, 6 min exposure to isoprenaline. [Ca²⁺]_i transients and firing rate are both partially recovered.

Contribution of increased [Ca²⁺]_i to the increase in firing rate after β-adrenergic stimulation

The results so far show that depleting or preventing the release of Ca²⁺ from the SR store is associated with cessation of the firing of pacemaker cells and that only in the presence of a high concentration of ryanodine can β-stimulation reverse this effect. However, the results do not directly determine whether the effects of β-stimulation on firing rate arise through the changing properties of pacemaker currents or through changes in the amplitude of the [Ca²⁺]_i transient. To address this question we compared the slowing of firing rate caused by ryanodine, which we assume to operate solely by changes in Ca²⁺ release, with the changes in firing rate caused by β-stimulation. If both interventions change firing rate by means of the amplitude of the [Ca²⁺]_i transient then the relation between firing rate and the amplitude of the [Ca²⁺]_i transient should be similar for both. Conversely if β-stimulation works partly or largely through effects on pacemaker channels then the relation between firing rate and [Ca²⁺]_i should be different since firing rate could change in part because of Ca²⁺-independent mechanisms.

Figure 7 shows the sample records (A) and the resulting plot of firing rate vs. amplitude of the [Ca²⁺]_i transient (B) for one such experiment. Initially the [Ca²⁺]_i transient amplitude and firing rate were measured under control conditions (○). Adrenaline (2 μM) was applied first to establish that the Ca²⁺ transients and firing rate increased (□). Then adrenaline was washed off and ryanodine (10 μM) was added, which caused a slow reduction in the amplitude of the [Ca²⁺]_i transient and the firing rate until the preparation became quiescent (•). Finally adrenaline was added in the continuing presence of ryanodine (▪), which increased the firing rate as previously shown in Figs 4, 5 and 6. Linear regression lines have been plotted through the control and ryanodine points (continuous line) and through the adrenaline and adrenaline + ryanodine points (dashed line). It is clear from the plot in Fig. 7B that the slope of the β-stimulated data is very similar to that in the absence of β-stimulation. In five such experiments there was no significant difference between the slopes of the regression lines.

Figure 7. Relation between firing rate and the amplitude of [Ca²⁺]_i transients.

A, sample records of [Ca²⁺]_i under control conditions, 2 μM adrenaline (ADR), 20 μM ryanodine (RYA) before firing ceased and, after firing ceased, adrenaline caused firing to restart in the presence of ryanodine. B, plot of firing rate vs. amplitude of [Ca²⁺]_i transient from the same cell as A. ○, control data; □, adrenaline alone; •, presence of ryanodine; ▪, adrenaline and ryanodine present. Continuous line is the regression fit to control and ryanodine alone data; dashed line is the regression fit to the adrenaline and adrenaline + ryanodine data.

For further investigation of whether β-stimulation affected the relationship between firing rate and the amplitude of the [Ca²⁺]_i transient, we compared the firing rate under control conditions (○) with the firing rate in the presence of β-stimulation but at the same [Ca²⁺]_i transient amplitude. For this we measured the vertical difference between the regression lines at the [Ca²⁺]_i transient amplitude of the control. There was no significant difference in these firing rates. These results show that the relationship between calcium transient amplitude and firing rate is unaffected by the presence or absence of β-stimulation.

Contribution of SR Ca²⁺ release to the [Ca²⁺]_i transient

It is clear from Fig. 1 that β-stimulation increases the amplitude of the [Ca²⁺]_i transient but the source of the additional Ca²⁺ is not apparent from such experiments. The experiments showing that agents which affect SR Ca²⁺ storage (caffeine, ryanodine, TBQ) suggest that SR Ca²⁺ is important to firing rate but conflict with the long-standing view that the major source of Ca²⁺ for contraction is from I_Ca in amphibian hearts (Chapman, 1979). We have explored this issue in two ways. First we measured the amplitude of the [Ca²⁺]_i transient in control spontaneously firing cells (315 ± 46 nM, n = 9) and compared that to the amplitude of the last [Ca²⁺]_i transient before firing ceased in the presence of either ryanodine or TBQ (131 ± 22 nM). The difference (184 nM) is an estimate of the contribution of the SR to the [Ca²⁺]_i transient and appears to provide 59 ± 2 % of the Ca²⁺. In the presence of β-stimulation, the [Ca²⁺]_i amplitude was 857 ± 140 (n = 4). Application of 10 μM TBQ caused firing to slow (one cell) or stop (3 cells) and the smallest [Ca²⁺]_i amplitude was 356 ± 90 nM. The difference (501 nM) suggests that the SR provides 58 ± 5 % of the Ca²⁺.

The above experimental approach is not completely satisfactory because the action potential may change in the presence of drugs altering Ca²⁺ release and the SR may still be partially loaded when the cell stops firing. These problems can be overcome by voltage clamping the cells and measuring the [Ca²⁺]_i transient before and after application of agents which deplete the SR. Figure 8 shows a representative example. In these experiments we used the perforated-patch technique and the pipette contained a K⁺-based solution to avoid the additional effects of Cs⁺ on SR Ca²⁺ release (Hussain & Orchard, 1997). The voltage protocol was: hold at -60 mV, depolarize to +10 mV for 300 ms, then return to the holding potential. In some experiments a series of conditioning depolarizations was given but this did not affect the size of the resulting [Ca²⁺]_i transient. When 2 μM isoprenaline was applied a very large increase in inward current occurred and the [Ca²⁺]_i transient was larger and shorter in duration. Then TBQ (10 μM) was applied and after about 5 min exposure the [Ca²⁺]_i transient was reduced but the inward current was little affected. In ten experiments the inward current increased by a factor of 3.7 ± 1.2 in the presence of isoprenaline and the amplitude of the [Ca²⁺]_i transient increased from 464 ± 75 to 643 ± 102 nM. Application of TBQ had no significant effect on the peak inward current but reduced the [Ca²⁺]_i amplitude to 373 ± 48 nM; thus the component of [Ca²⁺]_i provided by the SR appears to be (100 ×[643 - 373]/643) = 42 ± 5 %.

Figure 8. Membrane currents and [Ca²⁺]_i in a voltage-clamped pacemaker cell showing effects of isoprenaline and TBQ.

A, control voltage-clamp conditions; perforated patch, cell loaded with indo-1 AM in the normal way. B, 2 μM isoprenaline present (ISO); note very large increase in inward current in this cell and moderate potentiation of the [Ca²⁺]_i transient. The slow decline of inward current observed in this cell was not seen in other cells. C, after 5 min application of 10 μM TBQ in the continuing presence of isoprenaline.

In identical control experiments, TBQ had no significant effect on the peak inward current but the amplitude of the [Ca²⁺]_i transient was 464 ± 75 nM (control) and 226 ± 31 nM (TBQ). Thus the proportion of Ca²⁺ released from the SR was 53 ± 8 % under these conditions. Three similar experiments in control conditions were performed using ryanodine (20 μM) instead of TBQ with similar results.

DISCUSSION

Ca²⁺ release from the SR plays a role in cardiac pacemaking

The spontaneous electrical activity of pacemaker cells is generally attributed to a combination of currents active during the pacemaker period whose net result is a small inward current. While it is accepted that [Ca²⁺]_i modulates some of the potential pacemaker currents (Irisawa et al. 1993), there is no consensus as to whether [Ca²⁺]_i has an identifiable role in pacemaker activity.

In the present study we show that several different interventions which reduce Ca²⁺ release from the SR led to the abolition of spontaneous pacemaker activity. These interventions were (1) the rapid application of caffeine, (2) the application of ryanodine at a low or at high concentration, and (3) the application of TBQ. Caffeine increases the frequency and duration of SR Ca²⁺ channel opening (Rousseau & Meissner, 1989) and therefore rapidly depletes the SR of Ca²⁺ (Callewaert et al. 1989). Ryanodine has different effects at low and high concentrations; at low concentrations it opens the release channel in a sub-conductance state whereas at high concentrations it closes the release channels (Rousseau et al. 1987). It follows that the effect on the SR Ca²⁺ store is different at the low and high concentrations and we confirmed this with the caffeine exposures. TBQ inhibits the SR Ca²⁺ pump (Westerblad & Allen, 1994) which will allow the SR Ca²⁺ leak to deplete the SR of Ca²⁺. Although these compounds are structurally and functionally very different and they are not generally thought to have direct effects on the pacemaker channels, their common effect would be to prevent Ca²⁺ release from the SR either because it was empty (caffeine, low concentration of ryanodine, TBQ) or because the release channel is closed (high concentration of ryanodine). Thus we believe that the present study strongly supports earlier claims (Rubenstein & Lipsius, 1989; Zhou & Lipsius, 1993; Rigg & Terrar, 1996; Ju & Allen, 1998) that the release of Ca²⁺ from the SR store plays an important role in pacemaker activity.

At intermediate concentrations both caffeine and ryanodine have been shown to cause a transient acceleration of firing before depressing the firing rate (Satoh, 1993; Ju & Allen, 1998). These results can be explained if the initial opening of SR Ca²⁺ channels leads to a rise in [Ca²⁺]_i transients but later the reduced Ca²⁺ content of the SR leads to a reduced [Ca²⁺]_i transient.

What is the mechanism by which SR release of Ca²⁺ influences firing rate? The most obvious possibility is that the amplitude of the [Ca²⁺]_i transient influences the magnitude of one or more of the pacemaker currents. Note that the experiments with ryanodine show that it is not the state of the SR Ca²⁺ store which is important but whether or not release occurs. However, it is not yet possible to unequivocally identify the Ca²⁺-sensitive current(s). A likely contender is I_NaCa which is present in pacemaker cells and whose calculated amplitude would make it an important contributor to the net pacemaker current (Zhou & Lipsius, 1993; Ju & Allen, 1998; Cousins & Bramich, 1998). However, contributions from other Ca²⁺-sensitive currents, such as the calcium-sensitive chloride current (Zygmunt, 1994), cannot be excluded.

β-Adrenergic agents modulate the [Ca²⁺]_i transient in pacemaker cells

We show that β-adrenergic stimulation increases the amplitude of the [Ca²⁺]_i transient confirming earlier studies on atrioventricular pacemaker cells (Hancox et al. 1994). In addition the time course of decline is accelerated presumably because of stimulation of the SR Ca²⁺ pump (Tada et al. 1974). Part of the increase in the [Ca²⁺]_i transient will be caused by β-adrenergic stimulation of I_Ca(L) (Noma et al. 1980) and our own data suggest that the component of Ca²⁺ released from the SR also increases (see later section).

Given that we have shown that the firing rate is dependent on the amplitude of the [Ca²⁺]_i transient and that β-stimulation increases the [Ca²⁺]_i transient, it is appropriate to ask whether these changes in Ca²⁺ handling contribute to the changes of firing rate caused by β-stimulation. One line of evidence in support of this hypothesis is that interventions which reduce the Ca²⁺ content of the SR still abolish or slow the rate of firing in the presence of β-stimulation. Thus β-stimulation of pacemaker channels is not sufficient to cause spontaneous firing in the absence of Ca²⁺ release from the stores. More importantly, in Fig. 7 we demonstrate that the relationship between firing rate and amplitude of the [Ca²⁺]_i transient is indistinguishable in the presence and absence of β-stimulation. This result strongly suggests that the effect of β-adrenergic stimulation on the [Ca²⁺]_i transient appears to explain much of its effect on rate. This probably arises because one of the key effects of β-stimulation is the increase in I_Ca(L) which contributes to the [Ca²⁺]_i transient both directly and by increasing SR Ca²⁺ release. In this interpretation β-stimulation enhances the established ‘pacemaker currents’ but, in addition, it has an important effect on both Ca²⁺ influx and SR Ca²⁺ release which stimulate an additional Ca²⁺-sensitive pacemaker current.

However, the most convincing evidence that the major effects of β-stimulation are through [Ca²⁺]_i rather than effects on surface membrane channels is the ability of β-stimulation to induce pacemaker activity in preparations made quiescent by high concentrations of ryanodine. Why does β-stimulation cause resumption of firing in high rather than low concentrations of ryanodine? If the β-induced change of pacemaker channel current such as I_Ca(L) or I_f was the most important action involved, then one would expect firing to be induced at either ryanodine concentration. If the reason that ryanodine slows firing is that it directly inhibits one of the pacemaker currents (Li et al. 1998) and β-stimulation reverses this effect, then one would expect β-stimulation to be more effective at low ryanodine concentrations than at high.

Instead, the fact that SR stores are still loaded with Ca²⁺ in the presence of high concentrations of ryanodine and the fact that the initiation of firing coincides with the detection of small [Ca²⁺]_i transients suggests that β-stimulation is able to overcome the ryanodine-induced inhibition of Ca²⁺ release. It seems that only when SR Ca²⁺ release can occur does spontaneous pacemaker firing restart. These findings strongly support our general conclusion that Ca²⁺ release from the SR induces an inward pacemaker current which is necessary for firing in toad pacemaker cells. One surprising aspect of these results is that the normally irreversible block of SR Ca²⁺ release caused by a high concentration of ryanodine can apparently be reversed so quickly by β-stimulation. We are not aware of any other reports of this phenomenon which suggest that β-stimulation-induced phosphorylation of the SR Ca²⁺ release channel (Yoshida et al. 1992) can overcome the channel block caused by ryanodine.

Another possible interpretation of these results is that β-stimulation leads to release of Ca²⁺ through another class of SR channels which are not blocked by high concentrations of ryanodine. These release channels might be inositol trisphosphate-sensitive, as recently proposed by Cousins & Bramich (1998). To explain why β-stimulation has no effect at low concentrations of ryanodine, the second class of release channels would need to be associated with the SR so that when the SR was depleted stimulation of this channel was without effect.

These experiments with adrenaline in the presence of high concentrations of ryanodine give some insight into how firing first starts in quiescent preparations. It has frequently been observed that firing in quiescent preparations is preceded by oscillatory potential changes (for review see Irisawa, 1978). Our observation that [Ca²⁺]_i can rise before the membrane oscillations suggests that activation of Ca²⁺ release from the SR leading to a Ca²⁺-dependent pacemaker current is a possible triggering mechanism.

Contribution of calcium current and SR Ca²⁺ release to the [Ca²⁺]_i transient

The [Ca²⁺]_i transient of pacemaker cells is often considered to be dominated by Ca²⁺ entry across the surface membrane based on the narrow diameter of the cells, absence of T-tubules and a paucity of SR (Masson-Pévet et al. 1978). The present experiments suggest that the SR contributes about 40-60 % of the Ca²⁺ required for activation. The estimates from spontaneously firing cells were higher (59 % control; 60 % during β-stimulation) but are open to the criticism that the action potential may change during application of agents which affect SR Ca²⁺ release and the possibility that some SR Ca²⁺ may still be present when the cells stop firing. The estimates from the voltage-clamped cells (53 % control; 44 % during β-stimulation) are somewhat lower than those obtained from spontaneously firing cells. However, we noted, as previously described by Hussain & Orchard (1997), that the potentiation produced by β-stimulation under voltage-clamp conditions was considerably smaller than in spontaneous cells so that the lower value of SR contribution to the [Ca²⁺]_i transient may represent some unphysiological aspect of voltage clamp. On the basis of these data we conclude that SR release provided about half of the Ca²⁺ required for activation and that this fraction is not greatly changed by β-stimulation. Thus the SR clearly makes an important contribution to the [Ca²⁺]_i transient but the contribution from this source is clearly smaller than in mammalian ventricular myocytes where the fraction of Ca²⁺ provided by the SR is often much higher (94 %: Varro et al. 1993).

Comparison with stimulation of the sympathetic nerve to the toad sinus venosus

Our results show that bath application of β-adrenergic agents to isolated pacemaker cells causes increases in firing rate by means of β-receptors. In contrast, earlier studies on the intact sinus venosus have demonstrated that stimulation of the sympathetic nerve produced an increased firing rate and that this effect was not blocked by α- or β-blockers but was reduced by ergotamine (Bramich et al. 1993). These authors proposed that bath-applied adrenaline activated extrajunctional receptors which were conventional β-receptors acting via adenyl cyclase and cAMP. In contrast the nerve stimulation acted via junctional receptors which are not α- or β-adrenoceptors and are not coupled to adenyl cyclase (Bramich et al. 1993). Since all our results are inhibited by β-blockers the implication is that we are studying the extrajunctional receptors. Recently Cousins & Bramich (1998) showed that sympathetic nerve stimulation could trigger oscillatory Ca²⁺ release from the sinus venosus rendered quiescent with Ca²⁺ channel blockers. These effects were eliminated or reduced by caffeine, ryanodine or thapsigargin (another SR Ca²⁺ pump blocker) suggesting that the store of calcium is the same as that from which action potential-triggered Ca²⁺ release occurs. They postulate that the junctional receptors activated by sympathetic nerve stimulation are linked to phospholipase C and inositol trisphosphate production which triggers the release of Ca²⁺ from the stores. Despite these differences in the receptors involved there are interesting similarities in our results. In both their study and ours the increase in firing rate in the toad sinus venosus appears to be triggered by increased [Ca²⁺]_i, although the receptors and the intracellular pathways involved are different.

Conclusion

We show that the firing rate of toad pacemaker cells is dependent on Ca²⁺ release from stores and that when Ca²⁺ release is prevented spontaneous firing slows or ceases. Bath application of the sympathetic transmitter adrenaline acts through β-receptors and causes increased [Ca²⁺]_i transients, which appear to be the major cause of the increase in firing rate. The probable mechanism by which the increased [Ca²⁺]_i transients affect the firing rate is as follows. I_NaCa is increased by elevated [Ca²⁺]_i causing an increased inward current during diastole. This mechanism operates in a tissue which lacks I_f and it remains to be seen how important this type of mechanism is in mammalian pacemaker cells which contain I_f.