Na⁺ currents are required for efficient excitation–contraction coupling in rabbit ventricular myocytes: a possible contribution of neuronal Na⁺ channels

Abstract

Ca²⁺ transients were activated in rabbit ventricular cells by a sequence of action potential shaped voltage clamps. After activating a series of control transients, Na⁺ currents (I_Na) were inactivated with a ramp from −80 to −40 mV (1.5 s) prior to the action potential clamp. The transients were detected with the calcium indicator Fluo-4 and an epifluorescence system. With zero Na⁺ in the pipette I_Na inactivation produced a decline in the SR Ca²⁺ release flux (measured as the maximum rate of rise of the transient) of 27 ± 4% (n = 9, P < 0.001) and a peak amplitude reduction of 10 ± 3% (n = 9, P < 0.05). With 5 mm Na⁺ in the pipette the reduction in release flux was greater (34 ± 4%, n = 4, P < 0.05). The ramp effectively inactivates I_Na without changing I_Ca, and there was no significant change in the transmembrane Ca²⁺ flux after the inactivation of I_Na. We next evoked action potentials under current clamp. TTX at 100 nm, which selectively blocks neuronal isoforms of Na⁺ channels, produced a decline in SR Ca²⁺ release flux of 35 ± 3% (n = 6, P < 0.001) and transient amplitude of 12 ± 2% (n = 6, P < 0.05). This effect was similar to the effect of I_Na inactivation on release flux. We conclude that a TTX-sensitive I_Na is essential for efficient triggering of SR Ca²⁺ release. We propose that neuronal Na⁺ channels residing within couplons activate sufficient reverse Na⁺–Ca²⁺ exchanger (NCX) to prime the junctional cleft with Ca²⁺. The results can be explained if non-linearities in excitation–contraction coupling mechanisms modify the coupling fidelity of I_Ca, which is known to be low at positive potentials.

Introduction

The microscopic theory of excitation–contraction coupling (ECC) in cardiomyocytes is organized around the concept of the couplon (Stern et al. 1997; Franzini-Armstrong et al. 1999), which is viewed as the functional unit that produces a Ca²⁺ spark. Within a single couplon, sarcolemmal L-type Ca²⁺ channels (LCCs) admit Ca²⁺ into a restricted junctional region (the diadic cleft), leading to a significant rise in Ca²⁺ concentration (Peskoff & Langer, 1998). This Ca²⁺ gates a cluster of ryanodine receptors (RyRs) on the apposing membrane of the junctional sarcoplasmic reticulum (SR), allowing Ca²⁺ release from the SR to generate a single Ca²⁺ spark. The spatial separation between couplons is sufficient to permit their local control (Stern, 1992), which explains the graded response of Ca²⁺ transients under voltage clamp by their voltage dependent stochastic recruitment. This embodies the classical ideas of Ca²⁺-induced Ca²⁺ release proposed by Fabiato (1983), as well as the original evidence by London & Krueger (1986) that Ca²⁺ currents (I_Ca) are responsible for triggering Ca²⁺ release from the terminal cisternae of the SR. Although it is widely held that I_Ca is the principal trigger for SR Ca²⁺ release in cardiac cells, there are studies suggesting that the Na⁺–Ca²⁺ exchanger (NCX) may also be involved in this process (Haworth et al. 1991; Nuss & Houser, 1992; Kohomoto et al. 1994; Levi et al. 1994b; Vites & Wasserstrom, 1996a,b; Wasserstrom & Vites, 1996; Litwin et al. 1998).

Recently Inoue & Bridge (2003) established in patch clamped rabbit ventricular myocytes that when action potentials (APs) are used to repetitively stimulate a cell, in those locations where sparks are present, they occur with a probability approaching 100%. This, despite (1) the inherent low open probability (P_o) of LCCs, which is less than spark probability (P_s), and (2) their low coupling fidelity (P_Cpl) (Zhou et al. 1999; Polakova et al. 2008; Sobie & Ramay, 2009). Preliminary data show that when the magnitude of the trigger is increased spikes are not induced at new locations (Tsujii et al. 2005). This suggests that, at least in rabbit, APs activate all couplons. However SR Ca²⁺ release is graded with voltage clamp pulses from −40 to +50 mV by voltage-dependent local recruitment of Ca²⁺ sparks which are under local control (Lopez-Lopez et al. 1995). This indicates that unlike recruitment of couplons by an AP, recruitment by a voltage clamp is not 100% at many potentials including those corresponding to the peak of an AP. How can these apparently contradictory findings be resolved? The obvious difference between voltage clamp pulses from −40 mV and APs is that in the latter there is a Na⁺ current (I_Na). That the extent of triggering seems to differ in the presence and absence of I_Na requires explanation. Leblanc & Hume (1990) suggested that reversal of NCX by I_Na could trigger SR Ca²⁺ release directly. However, it is also plausible that in rabbit the main effect of NCX is to raise P_Cpl of the I_Ca by priming the junctional cleft with Ca²⁺. This would not require a large quantity of Ca²⁺ but could, in principle, produce a disproportionately large effect on triggering (see Discussion).

In view of these rather paradoxical results we used AP clamps both with and without a preceding ramp to inactivate I_Na to investigate two issues. First, the possibility that an I_Na makes a significant contribution to triggering SR Ca²⁺ release in rabbit ventricular cells by activating reverse NCX. Second, that this significant contribution originates from a modest reverse NCX. Finally, we investigated the hypothesis that a neuronal Na⁺ channel is responsible for mediating at least part of the foregoing effects in rabbit.

Methods

Isolation of rabbit ventricular myocytes

We isolated ventricular myocytes from adult New Zealand White rabbits (2.5 kg). The animals were anaesthetized with an intravenous administration of sodium pentobarbital (50 mg ml⁻¹). The heart was quickly excised and the aorta cannulated. The cells were obtained by retrograde perfusion of the heart with Collagenase P (Roche) and protease XIV (Sigma) as previously described (Litwin et al. 1998). Following isolation, we stored the cells for up to 8 h in a 1 mm Ca²⁺-Hepes-buffered saline solution. All procedures were in accordance with the NIH Guide for the Care and Use of Laboratory Animals and animal protection guidelines of the University of Utah.

Electrophysiological recordings

We placed isolated myocytes in a recording chamber coated with mouse laminin (BD Biosciences, Franklin Lakes, NJ, USA) mounted on the stage of an inverted microscope (Nikon Diaphot, Nikon Instruments). The bath was continuously superfused with a modified Tyrode solution containing (in mm): 4.4 KCl, 138 NaCl, 1 MgCl₂, 2 CaCl₂, 11 dextrose, 24 Hepes, 0.5 probenecid (pH 7.4 adjusted with NaOH). Individual cells were voltage clamped (Axopatch 200; Molecular Devices, Sunnyvale, CA, USA) with a pipette of 1–2 MΩ resistance when filled with a solution containing (in mm): 128 KCl, 4.5 MgCl₂, 5.5 dextrose, 10 Hepes, 0.02 EGTA, 5 K₂ATP (pH 7.1 adjusted with KOH) (in some experiments 5 mm Na⁺ was added to this solution). An AP waveform template was first recorded from a cell under the conditions of interest and the stimulus artifact was removed by subtraction (Arreola et al. 1991; Bouchard et al. 1995; Grantham & Cannell, 1996). Five AP clamps were applied from a holding potential of −80 mV at a frequency of 0.2 Hz, followed by an AP waveform clamp with an initial voltage ramp from −80 to −40 mV (1.5 s) to inactivate the I_Na (Fig. 1A).

Figure 1. Effect of the I_Na inactivation on the Ca²⁺ transient.

A, graphic representation of the voltage protocol used to record the associated transients. B, summary of the effect of the ramp on the amplitude and maximum rate of rise of the transient with two different dialysing [Na⁺] (0 and 5 mm). The relative values of the amplitude of the transients were: 90 ± 3% (0 mm Na⁺, n = 9) and 84 ± 5% (5 mm Na⁺, n = 4), and the values of the maximum rate of rise were 73 ± 4% (0 mm Na⁺, n = 9) and 67 ± 4% (5 mm Na⁺, n = 4). *P < 0.05, **P < 0.001, ***P < 0.07. C, representative Ca²⁺ transient obtained with 0 mm Na⁺ in the pipette when applying an AP waveform with and without a ramp. D, a 27% difference in a slope is easier to appreciate in the derivative of the signal. Thus, we show the derivative of the first 120 ms of the transients presented in (C).

We measured I_Ca using a pipette solution containing (in mm): 128 CsCl, 4.5 MgCl₂, 5.5 dextrose, 10 Hepes, 5 BAPTA, 5 MgATP, 10 TEA-Cl (pH 7.2 adjusted with CsOH). After breaking into the cell, the external solution was replaced with a solution containing (in mm): 2 CaCl₂, 4.4 CsCl, 128 TEA-Cl, 1 MgCl₂, 11 dextrose, 24 Hepes, 0.5 probenecid (pH 7.4 adjusted with TEA-OH). I_Ca was measured with square clamp pulses generated by pCLAMP software or with AP waveform clamps, preceded by a ramp when necessary. I_Ca was calculated as the difference between the peak inward current and the value before the change in voltage. Background current was corrected using a solution containing 300 μm Cd²⁺ and 10 μm of nifedipine.

I_Na was measured using a pipette solution with the following composition (in mm): 120 CsCl, 4.5 MgCl₂, 5.5 dextrose, 10 Hepes, 0.02 EGTA, 5 MgATP, 10 TEA-Cl (pH 7.2 adjusted with CsOH). After breaking into the cell, the external solution was replaced with a solution containing (in mm): 4.4 CsCl, 128 NaCl, 1 MgCl₂, 11 dextrose, 24 Hepes, 0.5 probenecid (pH 7.4 adjusted with NaOH).

To measure the loss of voltage control during the AP clamp with a physiological [Na⁺]₀, we recorded the membrane potential in the cell using discontinuous single-electrode voltage clamp (dSEVC, Axoclamp-2B, Molecular Devices). The external solution was the modified Tyrode solution, and the pipette was filled with the 0 mm Na⁺ pipette solution.

In the experiments with 100 nm TTX, APs were evoked at a frequency of 0.2 Hz with current steps delivered by an Axoclamp-2B operating in bridge mode. We recorded steady state APs before and after the application of 100 nm TTX, which blocks TTX-sensitive Na⁺ channels.

All experiments were performed at room temperature (22–24°C).

[Ca²⁺]_i measurements

To record intracellular free Ca²⁺, we loaded the cells with 12.5 μm of fluo-4 AM (Molecular Probes/Invitrogen, Carlsbad, CA, USA) for 15 min at room temperature. Fluorescence transients were obtained in voltage clamped cells using a custom-designed epifluorescence system. We reported self-ratioed transients obtained after background subtraction. We calculated the maximum rate of rise of the transient, which provides a measurement of the release flux.

Expression of results and statistical analysis

All data are presented as means ±s.e.m. When comparing two conditions, we used Student's two-tailed unpaired t test. A P < 0.05 was considered to be significant.

Results

Inactivation of Na⁺ channels reduces the rate of rise and amplitude of the Ca²⁺ transient

To investigate the contribution of I_Na to the Ca²⁺ transient we compared transients evoked by an AP waveform before and after inactivating I_Na with a ramp. In each cell, we measured the average peak and maximum rate of rise of five transients and then compared these with the Ca²⁺ transient that was produced after inactivating I_Na with the ramp. We first established whether I_Na alone contributed to triggering Ca²⁺ release. To accomplish this we clamped a cell with zero Na⁺ in the pipette. Under these conditions inactivation of I_Na reduced the maximum rate of rise of the Ca²⁺ transient by 27 ± 4% (n = 9, P < 0.001) (Fig. 1C and D). Even with this large reduction of SR Ca²⁺ release flux, the peak amplitude of the transient was only diminished by 10 ± 3% (n = 9, P < 0.05). With 5 mm Na⁺ in the pipette, the reduction of the maximum rate of rise of Ca²⁺ released upon inactivation of I_Na was greater than the rise with no Na⁺ in the pipette (33 ± 4%, n = 4, P < 0.05) (Fig. 1B). There was little difference between the reduction in transient amplitude at the two [Na⁺]_i after I_Na was inactivated (Fig. 1B). The slow phase of relaxation of the transients was not altered and the effect of the ramp on the AP waveform clamp was completely reversible in the majority of the cells. The inactivation of I_Na was sufficiently rapid that these results cannot be explained by changes in SR Ca²⁺ content. It appears therefore that I_Na is critically important for producing optimum SR Ca²⁺ release flux.

Inactivation of I_Na has a small effect on reducing transmembrane Ca²⁺ flux

We used the same voltage clamp protocols to investigate the effect of inactivation of I_Na on transmembrane Ca²⁺ flux in the absence of SR function. For these experiments, we depleted SR Ca²⁺ by incubating the cells in 1 μm of thapsigargin and 1 μm ryanodine for 30 min. The cells exhibited a slower maximum rate of rise and a prolonged time to peak of the Ca²⁺ transient, as well as a prolonged time constants of Ca²⁺ decline when compared with control solution (0.009 ± 0.005 vs. 0.009 ± 0.001, 380 ± 2 vs. 264 ± 9 ms, and 486 ± 19 vs. 1252 ± 55 ms respectively). These findings are in agreement with others (Lewartowski et al. 1994). We assumed that, at least in part, this transmembrane flux can trigger SR Ca²⁺ release. We found that in the absence of internal Na⁺, regardless of whether or not I_Na is inactivated, the transients are virtually superimposable (Fig. 2A). Moreover, the maximum rate of rise of the Ca²⁺ transient is almost identical in both cases (Fig. 2B). With I_Na inactivated the maximum rate of rise was 97 ± 2% (n = 3, P > 0.05) of the control value. We infer that any reverse NCX flux activated by I_Na is extremely small when compared with the movement of Ca²⁺ across the membrane through LCCs. This suggests that a very small NCX flux in conjunction with the Ca²⁺ flux through LCCs can significantly affect triggering. This latter effect is consistent with the findings of others that isolated NCX is a small and rather ineffective trigger (Sipido et al. 1997) unless [Na⁺]_i is extremely high (Kohomoto et al. 1994). When we repeated this experiment with a more physiological [Na⁺] in the pipette (Levi et al. 1994a) (5 mm) we observed only modest differences in transmembrane flux with or without inactivated I_Na (Fig. 2C). In the absence of I_Na the maximum rate of rise of the transient and its amplitude were only slightly smaller than the control (94 ± 2%, n = 4, P < 0.1 and 93 ± 3%, n = 4, P > 0.05 respectively). Thus, an extremely small difference in transmembrane Ca²⁺ flux (cause by I_Na inactivation) must explain a significantly larger difference in SR Ca²⁺ release flux. We believe that this is a consequence of the structure of couplons, a topic we will fully discuss later.

Figure 2. Effect of the inactivation of I_Na on transmembrane Ca²⁺ flux.

A and B, when we used 1 μm thapsigargin and 1 μm ryanodine to disable SR Ca²⁺ content, the Ca²⁺ transient (A) and its maximum rate of rise (B) (a measurement of transmembrane Ca²⁺ flux) obtained in both conditions (with and without the ramp) were virtually identical with 0 mm Na⁺ in the pipette. C, with 5 mm Na⁺ in the pipette, the maximum rate of rise of the transients was slightly smaller than in control condition (94 ± 2%, n = 4, P < 0.1) (only the derivative of the Ca²⁺ transient is shown). Note that in both dialysing [Na⁺] the peak of the derivative of F/F₀ is smaller and occurred at later times when compared to a transient with the SR intact (Fig. 1D).

The ramp that precedes the AP waveform clamp effectively inactivates I_Na

To ensure that I_Na was inactivated after the ramp, we recorded the inward current generated with an AP waveform clamp. To prevent Ca²⁺ and K⁺ currents from interfering with our measurements, we removed them from the solutions. It is well known that during an AP waveform clamp I_Na cannot be controlled; therefore we cannot reliably measure its kinetics. However, we can establish whether or not I_Na is inactivated. Our results indicated that this ramp preceding the AP clamp does completely inactivate I_Na (Fig. 3A). It is possible that an undetectable I_Na was still present; however, if this was the case it is unlikely to contribute to the trigger. Since it is important that during the AP waveform clamp there is no significant loss of voltage control, we used dSEVC to demonstrate the extent to which we maintained voltage control while applying an AP waveform. To illustrate how severe loss of control can be when attempting to clamp I_Na, we depolarized the cell with a step pulse from a holding potential of −80 mV to −40 mV (the voltage in which I_Na reaches its peak value). With this step pulse we initially lost all voltage control, as expected (Fig. 3B). When we applied the AP waveform, the loss of control was dramatically reduced, and was confined to the first few milliseconds of the AP (Fig. 3C). The peak of the AP was hardly affected. Moreover, the form of the APs was similar with and without ramps, and therefore with and without I_Na (Fig. 3D). This suggests that there is a relatively small loss of voltage control during the peak of the AP waveform clamp. This is presumably because the peak of the AP corresponds closely to the reverse potential of I_Na. Although we cannot measure the kinetics of I_Na, it seems unlikely that during an AP clamp the amplitude of I_Na is greatly altered.

Figure 3. Effect of the ramp on I_Na and voltage control during a voltage clamp with normal [Na⁺]_o.

A, graphic representation of the voltage protocol used (upper traces) and representative I_Na traces (lower traces) evoked by an AP waveform clamp with (left) and without (right) the ramp preceding it. The ramp inactivates I_Na. This result was obtained in 8 cells. B, record of the membrane voltage while imposing a step pulse from a holding potential of −80 mV to −40 mV. As expected, initially there was a complete loss of voltage control. C, record of the membrane voltage while imposing an AP waveform clamp. The loss of control was reduced and was confined to the first few milliseconds of the AP. D, representative traces of the APs recorded while imposing AP waveform clamps with and without a ramp. The forms of the APs were similar in both conditions.

The ramp that precedes the AP waveform clamp does not affect I_Ca

The ramp that we used to inactivate I_Na could also influence I_Ca, which might, in turn, change the characteristics of the Ca²⁺ transient. For example, if I_Ca is present between −80 and −40 mV and is partially inactivated during the ramp, this could reduce the release flux by reducing Ca²⁺ triggering. We therefore investigated I_Ca with voltage clamp steps in 10 mV increments from −80 to +70 mV. For these experiments we replace K⁺ with Cs⁺, and Na⁺ with TEA-Cl in the bath and internal solution. We could not detect activation of I_Ca between −80 and −40 mV (n = 7 cells) (Fig. 4A). These data are in agreement with other studies that at least partially cover this range of voltage (Osaka & Joyner, 1991; Yuan et al. 1996). To demonstrate that I_Ca is not affected by the ramp, we also measured I_Ca during a series of 10 mV steps from a holding potential of −80 mV to +70 mV, with and without ramps imposed prior to each step. To avoid I_Ca rundown we applied the protocol displayed in Fig. 4B, top panel. After calculating the average I_Ca under the two conditions, we found no difference in the I_Ca density–voltage relationship with and without the ramp (n = 8 cells) (Fig. 4B, lower panel). Finally, we measured I_Ca during an AP waveform clamp with and without a ramp. There was no significant difference in the peak, time to peak or inactivation of I_Ca under the two conditions (n = 9 cells) (Fig. 4C). Note that the I_Ca we obtained has the shape that is usually associated with an AP waveform clamp (Arreola et al. 1991; Grantham & Cannell, 1996). Therefore it seems that there is a genuine effect of I_Na inactivation on the release flux that is not related to a modification of I_Ca.

Figure 4. Effect of the ramp on I_Ca.

A, representative I_Ca (middle panel) evoked by the applied protocol (upper panel). In the lower panel the average current density–voltage relationship between −80 and +70 mV is shown. There was no detectable I_Ca activation over the range of voltage covered by the ramp (−80 to −40 mV). Symbols represent means ±s.e.m. for 6 cells. B, representative I_Ca (middle panel) evoked by voltage steps in each of the sections (a–d) of the applied protocol (upper panel). In the lower panels are the representative current density–voltage relationships with ramps (empty squares) and without ramps (filled circles). We found no difference in the I_Ca density–voltage relationships under the two conditions. C, I_Ca recordings during an AP waveform clamp with ramp (grey) and without ramp (black). There is no difference in the traces obtained under both conditions.

The effect of 100 nm TTX on SR Ca²⁺ release is similar to the effect of inactivating I_Na

We tested the effect of 100 nm TTX on the Ca²⁺ transient. It is known that neuronal Na⁺ channels (e.g. Na_v1.1, Na_v1.2, Na_v1.3 and Na_v1.6) are inhibited by nanomolar concentrations of TTX (Goldin, 2001; Catterall et al. 2005). Indeed, 100 nm TTX provides a way of defining the existence of these channels, which have been identified in heart in different species (Rogart et al. 1989; Haufe et al. 2005). The heart isoform Na_v1.5 is, however, not affected by this concentration of TTX (Cribbs et al. 1990; Goldin, 2001; Catterall et al. 2005). We evoked APs with current pulses and recorded these and the transients that they produced in both control and 100 nm TTX. Initially, the APs in both conditions were identical. As with the ramp, neuronal I_Na inactivation by 100 nm TTX produced a similar though somewhat larger decline in both the maximum rate of rise and the amplitude of the Ca²⁺ transient (reduction by 35 ± 3%, n = 6, P < 0.001 and by 12 ± 2%, n = 6, P < 0.05 respectively) (Fig. 5A and B). This suggests that neuronal Na⁺ channels influence SR Ca²⁺ release. After continued APs in the presence of 100 nm TTX, in the three cells in which the transient did not deteriorate, there was a further reduction in peak amplitude and in release flux, which were subsequently restored to some extent (Fig. 5C). We will discuss this later. However, it appears consistent with the homeostatic mechanisms that control SR Ca²⁺ content and release. These mechanisms have been extensively investigated and described elsewhere (Diaz et al. 2005).

Figure 5. Effect of 100 nm TTX on the Ca²⁺ transient.

A, representative trace of Ca²⁺ transients during evoked APs in control Tyrode solution and in presence of 100 nm TTX. With 0 mm Na⁺ in the pipette, the effect of 100 nm TTX is comparable to the effect of inactivation of I_Na. B, derivative of the first 120 ms of the transient presented in A. C, a sequence of Ca²⁺ transients. In 100 nm TTX, there is an initial decline and a subsequently partial restoration of transient amplitude.

Discussion

Controversy surrounds the question of whether or not reverse Na⁺–Ca²⁺ exchange activated by I_Na makes a contribution to the trigger mechanism (Leblanc & Hume, 1990; Sipido et al. 1995). Nevertheless several reports have supported this hypothesis. The first clear demonstration of NCX involvement in triggering by Nuss & Houser (1992) indicated that when intracellular Na⁺ was high in voltage clamped isolated ventricular cells the magnitude of the Ca²⁺ transient was higher at positive potentials than one could attribute to triggering by I_Ca alone. These authors therefore suggested that reverse mode NCX could contribute to triggering SR Ca²⁺ release. Other investigators showed that when voltage clamped ventricular myocytes were dialysed with 20 mm Na⁺, in the presence of a Ca²⁺ channel blocker transients or shortening were far larger than one could expect from the residual I_Ca, again suggesting the involvement of NCX in triggering (Kohomoto et al. 1994). Additional studies obtained by a number of authors further suggested that NCX could be involved in triggering SR Ca²⁺ release (Haworth et al. 1991; Nuss & Houser, 1992; Kohomoto et al. 1994; Levi et al. 1994b; Vites & Wasserstrom, 1996a,b; Wasserstrom & Vites, 1996; Litwin et al. 1998; Sobie et al. 2008). Under voltage clamp exchange current is quite small even at positive potentials but makes what appears to be a significant contribution to triggering. To explain this, Litwin et al. (1998) proposed that NCX acted synergistically with the I_Ca. Recently Sobie et al. (2008) proposed an explanation for this synergism by suggesting that I_Ca could activate the regulatory site on the exchanger, which would increase its activity and therefore explain its significant contribution to triggering. While this indicates that reverse NCX can function as a trigger, it provides little information on whether or not during an action potential with normal cellular Na⁺, triggering by reverse NCX is significant. This issue was further clouded by the finding of Sipido et al. (1997) that in the absence of I_Na the exchanger produced little triggering or did so very slowly. This led to the view that NCX was unimportant in triggering (Bers et al. 1990; Lopez-Lopez et al. 1995; Sipido et al. 1997). These authors did not take into consideration the possibility that the I_Ca and NCX could sum their effects in a non-linear fashion. Weber et al. (2003) have suggested that exchange cannot function as a trigger because entry of I_Ca into a restricted domain would prevent the exchanger from reversing. However, we must consider the possibility that during the 4 ms or so required for triggering, early exchange activity could prime a restricted domain with Ca²⁺ prior to activation of I_Ca (Cannell & Soeller, 1997). Weber et al. considered the effect of I_Na on the exchanger 5 ms after its activation. It is likely that this is far too late to draw any useful conclusions.

There are far fewer studies designed to test the involvement of I_Na in ECC. Leblanc & Hume (1990) produced the first evidence (although the result has been controversial) that an I_Na could produce reverse NCX that would contribute significantly to triggering. Additional voltage clamp experiments by Lipp & Niggli (1994) supported this result, but experiments by Sipido et al. (1995) dampened enthusiasm for this hypothesis. More recently, other investigators have obtained evidence that a TTX-sensitive neuronal I_Na could be involved in the triggering process (Maier et al. 2002; Lines et al. 2006). This too has proved controversial. For example, Brette & Orchard (2006) found that while in rat ventricular cells contained TTX-sensitive Na⁺ channels, blockade of these did not affect SR Ca²⁺ release. Thus it seems clear that the mechanism that triggers SR Ca²⁺ release in ventricular cells is incompletely understood.

In view of these rather contradictory results we undertook a study to determine the contribution of I_Na to triggering SR Ca²⁺ release in rabbit ventricular cells. We adopted an approach that has not been attempted before. We measured SR Ca²⁺ released during AP shaped voltage clamps with and without voltage-dependent inactivation of I_Na. Inactivation of I_Na in such a way that SR content is unaffected produced a significant decline in SR Ca²⁺ release flux (Fig. 1). In addition, we found that a similar decline in release flux was produced by low doses of TTX (Fig. 5A and B), thus implicating the involvement of a neuronal (TTX-sensitive) Na⁺ channel in ECC. Some of these results are supported by a study in mice (Larbig et al. 2010) indicating that inactivation of I_Na, while significantly reducing SR Ca²⁺ release flux in wild-type mice, has no effect on release flux in cardiac specific NCX knock-out mice.

We employed AP shaped voltage clamps to investigate the effect of inactivating I_Na on SR Ca²⁺ release because these are more likely to represent ECC under ‘physiological’ conditions. We inactivated I_Na with a ramp because the significant loss of voltage control during a square pulse (Fig. 3B) can activate I_Ca (data not shown) and confound the interpretation of our results. Control is also lost on the upstroke of the AP, and for this reason even when other currents are removed we cannot reliably measure the time course of I_Na because the measurement would be distorted. However we can decide whether or not I_Na is present or absent after inactivation (Fig. 3A), which is what is required if we are to draw conclusions about the contribution of the I_Na to ECC. We should be clear that during our AP clamps the amplitude of I_Na is unlikely to be significantly distorted. It is our ability to measure its kinetics that is compromised. There appears to be no difference in the I_Ca density–voltage relationship obtained with square voltage clamp pulses from −80 or −40 mV to +70 mV (Fig. 4B), provided the voltage pulses from −40 mV were preceded by a ramp. We were also able to demonstrate that no I_Ca is generated between −80 and −40 mV (Fig. 4A) and that the ramp does not alter the I_Ca when measured with an AP clamp (Fig. 4C).

It is clear that in the absence of I_Na the SR Ca²⁺ release flux is significantly reduced. This result was obtained with no Na⁺ in the pipette. Since we assumed that intracellular Na⁺ is very low, the result indicates that it is the I_Na alone that can produce a sufficient Na⁺ accumulation in the vicinity of the NCX to promote its reversal and contribute to triggering release flux. As one may expect, with 5 mm Na⁺ in the pipette the loss of release flux is increased. It is perhaps surprising in view of the large alterations in release flux that the peak of the Ca²⁺ transient is only 10% smaller than when the I_Na is inactivated. These experiments were conducted at room temperature so the APs were prolonged. Presumably when the AP is prolonged, the transient is also prolonged until NCX contributes to its reduction. It is likely that at physiological temperatures the shorter AP will terminate the release flux earlier. In the mouse where APs repolarize early, the peak of the Ca²⁺ transient is much smaller after inactivation of I_Na (Larbig et al. 2010). Although these results were produced by a different approach to that used by Leblanc & Hume (1990) they still are consistent with their results. Moreover, the result is qualitatively similar to that obtained by Larbig et al. (2010) who showed that inactivation of I_Na in wild-type mice also led to a significant reduction in release flux.

By eliminating the contribution of the SR to Ca²⁺ release we could measure the transmembrane Ca²⁺ flux as the maximum rate of rise of the Ca²⁺ transient. This is clearly displayed in plots of the derivative of the time course of the Ca²⁺ transient (Fig. 2B and C). There is no observable difference in this flux in the presence or absence of I_Na. We infer from this that the NCX contributes very little during an AP to the total transmembrane Ca²⁺ flux, which we assume is largely mediated by Ca²⁺ flux through LCCs. We cannot conclude that this is also true locally in couplons because we do not know the ratio of NCX to LCCs in these structures. However, without evidence to the contrary, we assume that the extent of NCX mediated transmembrane Ca²⁺ flux is smaller than that mediated by LCCs in couplons. Since the junction is extremely small this does not preclude the possibility that a very small reverse NCX will produce a significant increase in junctional Ca²⁺. Despite this, triggering by NCX in isolation is insufficient for significant triggering (Sipido et al. 1997).

It therefore appears that we confront a paradox. On the one hand NCX stimulated by I_Na is likely to be very small, but on the other hand, its effect on triggering is large. This must stem from non-linear processes that exist in the ECC mechanism. We will attempt to provide a resolution to this paradox. However before discussing a mechanism that can explain this non-linearity we must first consider some properties of the coupling fidelity (P_Cpl) of I_Ca. P_Cpl is defined as the ‘the ratio of the number of L-type channel openings to the number of Ca²⁺ sparks activated’ (Zhou et al. 1999). We could also define this quantity as the probability that I_Ca will activate one or a cluster of RyRs in a couplon. For any given channel there will be a distribution of P_Cpl that will depend upon, among other things, the open time distribution of the channel. Recently Polakova et al. (2008) have demonstrated that in rats this quantity must be small, for example around 0.15 at −80 mV (see their Fig. 5). This quantity tends to become even smaller at positive potentials (Sobie & Ramay, 2009). We assume without evidence to the contrary that this is also true in rabbits. To explain the high probability of couplon activation in rats after repolarization to negative potentials, Polakova et al. (2008) deduced that in rats there would have to be a substantial number of LCCs available for triggering. We will present an alternative to this to explain the high probability of couplon activation in rabbits although this does not rule out the possibility that numerous DHPRs must be available for couplon activation in rabbit.

We can write the following equation derived by Polakova et al. (2008):

(1)

where P_s is the probability that any DHPR opening activates a spark, n_DHPR is the total number of DHPRs in a couplon and P_o is the DHPR open probability. This equation does not for example describe the way that P_Cpl could depend on voltage, or take into consideration the effect of sequential openings of DHPRs on couplon activation. However, it is useful to illustrate a possible mechanism that can explain our result. During an AP in rabbits, couplons are activated with a probability of 1. If the maximum transient that we observe represents a P_s of roughly 1, then removing I_Na reduces P_s to 0.7. It is unlikely that removing I_Na will reduce n_DHPR or P_o. We need to account for an approximately 30% change of SR release flux and hence of P_s simply as a result of removing I_Na and presumably reverse exchange. Therefore we propose that P_Cpl is modified. How can this occur? If priming the dyadic cleft with Ca²⁺ increases P_Cpl this would explain why removing the priming by elimination of reverse exchange could reduce P_Cpl and therefore P_s and the SR release flux as observed. Neuronal Na⁺ channels which, as our results suggest seem to be involved in ECC in rabbit, activate rapidly (Fozzard & Hanck, 1996). With a sufficient number of these we hypothesize that early during an AP Na⁺ is significantly elevated in the junctional region, which is very small. We suggest that this takes place before I_Ca is activated. We assume, as have others, that this Na⁺ causes reversal of NCX, which must also elevate Ca²⁺ significantly in the junction prior to activation of one or a number of available DHPRs. At least in rat, 27% of NCX molecules co-localized with RyRs (Jayasinghe et al. 2009). Even with fewer NCX molecules per couplon in rabbits, this might be sufficient to prime the diadic cleft prior to activation of available DHPRs. We know that the relationship between RyR open probability (P_RyR) and Ca²⁺ in the cleft is sigmoid (Copello et al. 1997). This could be one source of the non-linearity that is required to increase P_Cpl so that NCX sums in a non-linear fashion with I_Ca to produce triggering. The early priming of the junction with Ca²⁺ may for example move Ca²⁺ concentration along the foot of this sigmoid curve without significantly increasing P_RyR. Any subsequent I_Ca will then further increase junctional Ca²⁺ in a concentration range where it is steeply related to P_RyR. If the junction had not first been primed with Ca²⁺, the I_Ca would be required to supply additional Ca²⁺ to bring its value in the cleft to a point on the sigmoid curve where it is steeply related P_RyR. This will increase the likelihood that a DHPR opening will activate a RyR, i.e. it will increase P_Cpl. Thus junctional priming by reverse NCX could in principle increase P_Cpl sufficiently to account for at least 30% of the release flux. At this stage it is only possible to propose this as a plausible mechanism in which NCX sums in a non-linear fashion with I_Ca to modulate P_s. We require extensive modelling and additional information on the number of neuronal Na⁺ channels, DHPRs and NCX molecules in a couplon, and their localization within the couplon to extend our investigations.

There have been contradictory reports on whether or not neuronal Na⁺ currents are involved in triggering by reverse NCX (Maier et al. 2002; Brette & Orchard, 2006; Lines et al. 2006). We therefore tested the effect of 100 nm TTX on SR Ca²⁺ release in rabbit ventricular cells. TTX at 100 nm blocks neuronal Na⁺ channels without affecting cardiac Na⁺ channels. The IC₅₀ for neuronal Na⁺ channels is around 10 nm depending on species and isoform, while for Na_v1.5 (cardiac Na⁺ channel) it is in the micromolar range (Cribbs et al. 1990; Goldin, 2001; Catterall et al. 2005). We applied 100 nm TTX to a ventricular cell in which we evoked an AP while the cell was dialysed with a pipette containing no Na⁺. The cell interior was therefore nominally Na⁺ free. However, since Na_v1.5 was not blocked Na⁺ continued to enter the cell with each AP. We therefore do not know exactly what the internal [Na⁺] was, but we speculate that is probably quite low. The effect of applying 100 nm TTX was a decline of approximately 35% in release flux. This value is similar to the effect of inactivating I_Na. We conclude that the decline in release flux is attributable to the blockade of a neuronal TTX-sensitive Na⁺ channels. The result is consistent with work by Maier et al. on Guinea pigs, and Lines et al. who worked on rats but is inconsistent with the results of Brette and Orchard who also worked on rats (Brette & Orchard, 2006). We have no explanation for this discrepancy. However we infer that at least in rabbit a neuronal I_Na activates a small reverse exchange, which produces a large effect on release. Our experiments with AP voltage clamp only inform us that a I_Na is involved in some manner in triggering SR Ca²⁺ release. Our result with TTX suggests that one of these currents is one or several isoforms of the neuronal I_Na. However, a large part of the cardiac I_Na is mediated by Na_v1.5. Although we have argued that neuronal Na⁺ channels are important in triggering SR Ca²⁺ release, this does not exclude the possibility that there is a significant contribution from Na_v1.5 to this process. There are a number of mechanisms by which Na_v1.5 could influence triggering, e.g. setting the level of cytosolic Na⁺. Our results (see Fig. 1B) suggest the concentration of dialysing Na⁺ influences the extent to which inactivation of I_Na reduces SR Ca²⁺ release. Before we can understand the contribution of various types of I_Na on triggering we need more information on, for example, the cellular distribution of Na⁺ channels.

The response of the Ca²⁺ transient to TTX is biphasic. There is a secondary increase in release flux after application of TTX (Fig. 5C) which can be explained in terms of theories of Ca²⁺ homeostasis developed by others (Diaz et al. 2005). If the initial effect of TTX is to reduce release flux then NCX will extrude less Ca²⁺ from the cell by forward NCX. However, because reverse NCX is likely to be small Ca²⁺ influx is slightly affected. Thus, after TTX treatment we expect influx into to exceed efflux out of the cell. This means that the Ca²⁺ content of the cell will increase, which will tend to increase SR content until the release flux is sufficient to produce a forward exchange that results in efflux balancing influx. This can explain the biphasic effect of TTX.

In summary, we proposed that a rapidly activated neuronal Na⁺ channel causes immediate reverse of NCX, which primes the junctional cleft with Ca²⁺ and increases the P_Cpl of the I_Ca. We believe that an attractive feature of our proposal is that during the diastolic interval coupling fidelity is low. Any spontaneous openings of LCC are unlikely to activate couplons at times when this might be deleterious. The coupling fidelity of I_Ca would only be elevated early during the AP. Our results, together with those of Larbig et al. (2010), support the idea that in at least two species (mouse and rabbit) a I_Na is an essential part of the trigger mechanism in ECC.

Glossary

Abbreviations

AP

action potential

DHPR

dihydropyridine receptor

ECC

excitation–contraction coupling

I_Ca

L-type Ca²⁺ current

I_Na

Na⁺ current

LCC

L-type Ca²⁺ channel

dSEVC

discontinuous single electrode voltage clamp

NCX

Na⁺–Ca²⁺ exchanger

n_DHPR

number of DHPRs

P_Cpl

coupling fidelity

P_o

open probability

P_RyR

RyR open probability

P_s

probability of spark occurrence

RyR

ryanodine receptor

SR

sarcoplasmic reticulum

Author contribution

N.T., J.B.: conception and design of experiments; N.T., A.R., R.L., J.G., J.B.: collection, analysis, and interpretation of data; N.T., A.R., J.G., J.B.: drafting the article or revising it critically for important intellectual content. All authors approved the final manuscript. The experiments were conducted at the University of Utah.