Activation of reverse Na⁺–Ca²⁺ exchange by the Na⁺ current augments the cardiac Ca²⁺ transient: evidence from NCX knockout mice

Abstract

The hypothesis that Na⁺ influx during the action potential (AP) activates reverse Na⁺–Ca²⁺ exchange (NCX) and subsequent entry of trigger Ca²⁺ is controversial. We tested this hypothesis by monitoring intracellular Ca²⁺ before and after selective inactivation of I_Na prior to a simulated action potential in patch-clamped ventricular myocytes isolated from adult wild-type (WT) and NCX knockout (KO) mice. First, we inactivated I_Na using a ramp prepulse to −45 mV. In WT cells, inactivation of I_Na decreased the Ca²⁺ transient amplitude by 51.1 ± 4.6% (P < 0.001, n= 14) and reduced its maximum release flux by 53.0 ± 4.6% (P < 0.001, n= 14). There was no effect on diastolic Ca²⁺. In striking contrast, Ca²⁺ transients in NCX KO cardiomyocytes were unaffected by the presence or absence of I_Na (n= 8). We obtained similar results when measuring trigger Ca²⁺ influx in myocytes with depleted sarcoplasmic reticulum. In WT cells, inactivation of I_Na decreased trigger Ca²⁺ influx by 37.8 ± 6% and maximum rate of flux by 30.6 ± 7.7% at 2.5 mm external Ca²⁺ (P < 0.001 and P < 0.05, n= 9). This effect was again absent in the KO cells (n= 8). Second, exposure to 10 μm tetrodotoxin to block I_Na also reduced the Ca²⁺ transients in WT myocytes but not in NCX KO myocytes. We conclude that I_Na and reverse NCX modulate Ca²⁺ release in murine WT cardiomyocytes by augmenting the pool of Ca²⁺ that triggers ryanodine receptors. This is an important mechanism for regulation of Ca²⁺ release and contractility in murine heart.

Introduction

In cardiac muscle, Ca²⁺ influx through L-type Ca²⁺ channels (LCCs) located in the transverse-tubules triggers ryanodine receptors (RyRs) on the sarcoplasmic reticulum (SR) to open and release Ca²⁺ into the cytoplasm. In this cascade of events, Ca²⁺ influx through LCCs is widely regarded as the sole trigger of SR Ca²⁺ release (Bers, 2002). Although it is widely accepted that Na⁺–Ca²⁺ exchange (NCX) is the dominant Ca²⁺ efflux mechanism (Blaustein & Lederer, 1999; Philipson & Nicoll, 2000; Bers, 2002), there has been considerable debate regarding the possible contribution of reverse mode NCX to Ca²⁺ influx and Ca²⁺-induced Ca²⁺ release (CICR) in cardiac muscle. Conditions that favour reverse NCX include a positive membrane potential and increased intracellular Na⁺. These are conditions that may exist during the early phases of the cardiac action potential (AP). Several groups have increased Ca²⁺ transient amplitude in patch-clamped myocytes depolarized to positive potentials by including higher concentrations of Na⁺ in the pipette solution. The increased Ca²⁺ transients have been attributed to reverse NCX (Wasserstrom & Vites, 1996; Litwin et al. 1998; Su et al. 2001; Viatchenko-Karpinski et al. 2005). Two studies (Litwin et al. 1998; Sobie et al. 2008) specifically propose that NCX and LCCs act synergistically to prime the diadic cleft and provide trigger Ca²⁺ for CICR. In contrast, other researchers have been unable to detect a significant contribution of Ca²⁺ influx via reverse NCX to contractile events at physiological Na⁺ levels in mammalian cardiomyocytes (Bouchard et al. 1993; Sipido et al. 1997). The reasons for these discrepancies are unclear but may be attributable to differences in experimental conditions. In any case, the role of reverse NCX in cardiac excitation–contraction coupling is unresolved and may vary with species.

In 1990, LeBlanc & Hume (1990) suggested that a local accumulation of Na⁺, caused by Na⁺ influx through Na⁺ channels during the upstroke of the AP, could activate reverse mode NCX to trigger SR Ca²⁺ release during a physiological AP. These investigators showed that the Na⁺ channel blocker tetrodotoxin (TTX) decreases Ca²⁺ transient magnitude in guinea pig cardiomyocytes. Ca²⁺ transients could be elicited in the presence of an LCC blocker and were dependent on external Ca²⁺. In short, it appeared that Na⁺ current (I_Na) activation contributed to the generation of Ca²⁺ transients via reverse NCX. This study garnered much attention and helped popularize the concept of intracellular regions of restricted diffusion. The diadic cleft space, bordered by LCCs and the sarcolemmal membrane on one side and the RyRs and the junctional SR membrane on the apposing side received increased attention. In an accompanying commentary, Lederer et al. (1990) referred to this cleft region as the ‘fuzzy space’. Another report in 1992, however, suggested that NCX could not operate fast enough to have a significant role in the triggering of SR Ca²⁺ release (Sham et al. 1992). This study concluded that only the Ca²⁺ that enters through LCCs is responsible for CICR. Subsequent studies have either supported (Lipp & Niggli, 1994) or refuted (Evans & Cannell, 1997) a possible role for I_Na in triggering reverse NCX. Interpretation of experiments examining this hypothesis has been hampered by possible loss of voltage control and uncertainties regarding subsarcolemmal ion concentrations. Thus, this issue has been unresolved for many years.

We now re-examine the role of I_Na and NCX in CICR using a genetic tool: cardiomyocytes isolated from mice with cardiac-specific knockout of NCX (Henderson et al. 2004). Ablation of NCX has relatively modest effects on cardiac function, and myocytes adapt to the absence of the primary Ca²⁺ efflux mechanism by decreasing Ca²⁺ influx into the cell by up to 80%. The decrease in Ca²⁺ influx is due to the action of autoregulatory mechanisms on LCCs (Pott et al. 2005).

In control myocytes with intact NCX, we find that I_Na enhances the Ca²⁺ transient. This effect of I_Na is completely absent in NCX KO myocytes. The results strongly support the hypothesis that reverse NCX triggered by I_Na during an AP works synergistically with Ca²⁺ entering through LCCs to trigger Ca²⁺ release in cardiac myocytes.

Methods

Cardiac-specific NCX KO mice

We used cardiac-specific NCX KO mice generated using Cre/loxP technology to excise exon 11 of NCX1 as described in detail previously (Henderson et al. 2004; Pott et al. 2005). This results in total deletion of NCX1 activity from 80–90% of ventricular cardiomyocytes. The animals used in this study were between 8 and 10 weeks of age, did not display any gross pathology, and showed no evidence of heart failure. All procedures were in accordance with the guidelines of the University of California at Los Angeles Office for Protection of Research Subjects. We have read the detailed Journal of Physiology policy and relevant UK regulations regarding animal experimentation as described in Drummond (2009) and our procedures are in compliance with the policies and regulations described in that article.

Isolation of ventricular myocytes from adult mouse hearts

We used collagenase/protease digestion to isolate ventricular myocytes from WT and NCX KO mice as reported previously (Mitra & Morad, 1985; Pott et al. 2005). Following isolation, we stored the dissociated cells for up to 6 h at room temperature in modified Tyrode solution, containing (in mm): 136 NaCl, 5.4 KCl, 10 Hepes, 1 MgCl₂, 0.33 NaH₂PO₄, 1 CaCl₂, 10 d-dextrose (pH 7.4 with NaOH). This solution was also used, with modifications described below, as the standard bath solution for electrophysiological recordings. We anaesthetized mice using inhaled 5% isoflurane and removed hearts via thoracotomy. For each experiment, cells were isolated from at least three animals.

Electrophysiology

To apply action potential waveforms or record whole-cell membrane current, we placed the cells in an experimental chamber (0.5 ml) mounted on the stage of a Nikon Diaphot inverted microscope. A room temperature bath solution (20–22°C) continuously perfused the chamber. Patch electrodes were pulled from borosilicate glass (TW150F-3; World Precision Instruments, Sarasota, FL, USA) on a Sutter P-97 horizontal puller (Sutter Instruments, Novato, CA, USA). The fire-polished electrodes had a tip diameter of 2–3 μm and a resistance of 1–2 MΩ when filled with the patch electrode solution, which contained (in mm): 127 KCl, 10 Hepes, 10 NaCl, 0.05 cAMP, 5 phosphocreatine, 5 MgATP, 0.2 K₅fura-2. This solution was used for experiments where we applied an action potential waveform. For experiments in which we measured I_Ca, K⁺ in the bath solution was replaced with Cs⁺, and the pipette solution contained (in mm): CsCl 110, tetraethylammonium chloride 30, NaCl 10, Hepes 10, MgATP 5, cAMP 0.05, K₅fura-2 0.2. We recorded membrane current using an Axopatch 200B patch-clamp amplifier (MDS Analytical Technologies, Toronto, Canada). A Digidata 1322A (MDS Analytical Technologies, Toronto, Canada) digitizer was used for A/D conversion and was controlled by pCLAMP9 software (MDS Analytical Technologies, Toronto, Canada). We applied series resistance compensation to all recordings. A series of five conditioning pulses from −75 to 0 mV with a duration of 100 ms preceded every Ca²⁺ transient recording to normalize the SR Ca²⁺ load. Between each conditioning pulse, membrane potential was held at −75 mV for 1 s.

Fluorescence measurements

We recorded fura-2 fluorescence signals during voltage clamp using a custom-designed photometric epifluorescence detection system. The inverted microscope (Nikon Diaphot) was modified for dual wavelength excitation at 360 and 390 nm from LED light sources using an electronic chopper. We switched excitation wavelengths at 1200 Hz, so that both wavelengths were updated every 800 μs. The fluorescence emission was measured at 510 ± 40 nm as described by Goldhaber et al. (1991). Fluorescence emission measurements, corresponding to these two excitation wavelengths, were simultaneously recorded using the Axopatch 1322A A–D converter (MDS Analytical Technologies). The fluorescence ratio (F₃₆₀/F₃₉₀) calculated by dividing the fluorescence intensity during excitation at 360 nm with the intensity during excitation at 390 nm reflects free intracellular Ca²⁺. For trigger Ca²⁺ measurements in the absence of a SR Ca²⁺ load, the response of the 360 nm wavelength of fura-2 was close to the noise floor, and thus the ratios were extremely noisy. For this reason, we chose to report these results as ‘self-ratios’ (analogous to what is typically done with fluo-3) using the much more sensitive 390 nm wavelength. However, at rest, the 360 nm data was sufficient to verify that the resting Ca²⁺ was the same in both WT and KO (at the same external Ca²⁺). We therefore report these experiments using the inverse of single wavelength data at 390 nm normalized to baseline (ΔF/F₀). All fluorescence data are presented after filtering with a 50 Hz lowpass Gaussian filter (Clampex 9).

Determination of cellular phenotype

Ten to twenty per cent of ventricular myocytes isolated from KO hearts have a WT phenotype (Henderson et al. 2004). Those few cells that retain the NCX protein exhibit an inward NCX current (I_NCX) when patch clamped and exposed to a rapid pulse of caffeine to release SR Ca²⁺. Subsequent to experimental protocols, myocytes isolated from NCX KO mice were tested to ensure KO phenotype by rapid exposure to a 1 s pulse of caffeine. Those cells with a detectable inward exchange current were excluded from analysis. Phenotype detection was not carried out in experiments analysing the trigger Ca²⁺ influx since pretreatment with ryanodine and thapsigargin empties the SR of Ca²⁺ and therefore renders these cells unresponsive to caffeine. For these experiments, we relied instead upon the statistical likelihood that 80–90% of the cells would be KO phenotype, consistent with our caffeine results using KO myocytes with intact SR.

Solution exchange

Miniature solenoid valves (The Lee Co., Westbrook, CT, USA) controlled by pCLAMP digital outputs controlled the bath solution flow through a micromanifold (ALA Scientific Instruments, Westbury, NY, USA). This enabled precise timing of solution exchanges in relation to the voltage-clamp protocol.

Statistical analysis

Data are expressed as means ±s.e.m. We used paired or unpaired Student's t tests for statistical comparisons as appropriate. Statistical analysis was performed using SPSS (Chicago, IL, USA) and EXCEL (Redmond, WA, USA) software.

Results

Inactivation of I_Na decreases the Ca²⁺ transient in WT but not in NCX KO cardiomyocytes

We first measured Ca²⁺ transients evoked by a voltage waveform simulating an AP in patch-clamped ventricular myocytes from WT and NCX KO mice in the presence and absence of I_Na. The shape of the AP clamp waveform was the average of six WT APs (holding potential −70 mV). To eliminate I_Na prior to the AP clamp and to avoid an abrupt rise in subsarcolemmal Na⁺, we used a prepulse that depolarized the cell slowly to −45 mV using a 1.3 s ramp. We then held the voltage at −45 mV for an additional 20 ms followed by a brief repolarization (2 ms) to −70 mV before applying the AP clamp (see Fig. 3B below).

Figure 3. Effects of I_Na inactivation on Ca²⁺ transients in WT and NCX KO myocytes.

A, representative Ca²⁺ transient recordings with (thin line) and without ramp prepulse (thick line) to −45 mV. The prepulse reduces Ca²⁺ transient amplitude and rate of rise only in WT myocytes. B, pulse protocols. Control Ca²⁺ transients were obtained with stimulation by the AP clamp only (No Prepulse). A ramp over 1.3 s to −45 mV with a brief repolarization to −70 mV before a subsequent AP clamp was used to obtain Ca²⁺ transients with inactivation of I_Na. Scale bar =F₃₆₀/F₃₉₀= 0.1.

To ensure that the ramp prepulse inactivated I_Na, we recorded inward current in control myocytes during a voltage clamp from −75 to −40 mV using a square wave pulse (Fig. 1A). A large fast inward I_Na was elicited. When the square wave pulse was preceded by the 1.3 s ramp, there was no detectable inward current indicating inactivation of I_Na (Fig. 1B; note the change in scale). In these same cells, we also recorded fura-2 fluorescence as an indicator of intracellular free Ca²⁺ (not shown). There was no detectable Ca²⁺ rise during the prepulse to −45 mV. We also examined the effects of the duration of the ramp prepulse on I_Na, I_Ca and the Ca²⁺ transient. We varied the ramp duration from 0.6 to 1.8 s. The ramp prepulse was to −40 mV, rather than −45 mV, and the square wave depolarization was to 0 mV to elicit the inward I_Ca (Fig. 2A). In all cases, I_Na was inactivated and only the slower smaller I_Ca is observed (Fig. 2B). Ramp duration had no effect on I_Ca elicited by the pulse to 0 mV. The Ca²⁺ currents shown in Fig. 2B are overlain and shown on an expanded time scale in Fig. 2C. The data demonstrate that I_Ca kinetics are unaffected by the prepulse ramp duration. Occasionally, we found that a ramp prepulse to −40 mV could elicit a small slow increase in cytoplasmic Ca²⁺ near the conclusion of the prepulse. This Ca²⁺ could be attributed to reverse NCX as it was never observed in NCX KO cells. To eliminate this possible confounding factor, we subsequently only used a prepulse to −45 mV. At this voltage, we never detected a Ca²⁺ rise.

Figure 1. Effect of ramp prepulse on I_Na.

A, a rectangular pulse to −40 mV induces a large inward current. Top, voltage command; bottom, induced current. B, a ramp prepulse to −45 mV inactivates I_Na. The spikes in the current trace shown in B are capacitive artifacts. Note the different current and time scales in A and B.

Figure 2. Effects of various prepulse ramp durations on I_Na and I_Ca.

A, command voltage (mV) with ramp durations from 0.6 s to 1.8 s. A rectangular pulse to 0 mV is used to examine I_Ca kinetics after the ramp. B, current traces display I_Ca with no discernible I_Na. C, superimposed current recordings on an expanded time scale. Currents and Ca²⁺ transients were simultaneously recorded. Note that the time scales in A and B are defined by the sloped lines in A of 0.6, 1.2 and 1.8 s duration.

In control WT cells, inactivation of I_Na using the voltage protocol described above decreased the Ca²⁺ transient amplitude by 51.1 ± 4.6% (P < 0.001, n= 14) and reduced Ca²⁺ release flux (measured as the maximum rate of rise of the transient) by 53.0 ± 4.6% (P < 0.001, n= 14). In striking contrast, the prepulse had no effect on Ca²⁺ transients in NCX KO cardiomyocytes. Representative data are shown in Fig. 3 and data are summarized in Fig. 4. There was no significant effect of the prepulse on diastolic Ca²⁺ in either WT or KO cells (not shown). There was a tendency for the amplitude of the normal Ca²⁺ transient to be smaller in KO myocytes though this difference was not statistically significant. These experiments suggest that I_Na and reverse NCX have a major effect on Ca²⁺ release in mouse ventricular myocytes. The effect we observe is larger than has been seen previously (LeBlanc & Hume, 1990) and may be species dependent.

Figure 4. Summary of effects of I_Na inactivation on Ca²⁺ transients.

Plots compare amplitude (upper panels) and rate of rise (lower panels) of Ca²⁺ transients in WT vs. NCX KO cardiomyocytes. PP, prepulse; *P < 0.001.

Inactivation of I_Na decreases trigger Ca²⁺ in WT but not NCX KO cardiomyocytes

To detect the contribution of I_Na and subsequent reverse NCX activity to the Ca²⁺ influx that triggers SR Ca²⁺ release, the experiments described above were repeated after disabling the SR by incubating the cells for 20 min in Tyrode solution containing thapsigargin (0.01 mm) and ryanodine (0.2 μm). With 1 mm Ca²⁺ in the bath solution, eliminating I_Na caused a major reduction in Ca²⁺ influx in control WT cells. Fluorescence amplitude and rate of rise were reduced by 84.6 ± 10.0% and 83.6 ± 10.1%, respectively (P < 0.001, n= 8; Figs 5 and 6). In marked contrast, eliminating I_Na had no effect on Ca²⁺ entry in NCX KO cardiomyocytes (n= 8). These results indicate that trigger Ca²⁺ is remarkably dependent on I_Na in mouse ventricular myocytes.

Figure 5. Effects of I_Na inactivation on trigger Ca²⁺ in WT and NCX KO cardiomyocytes.

A, representative Ca²⁺ transient recordings with (thin line) and without ramp prepulse (thick line) to −45 mV. Cells were incubated with 0.2 mm ryanodine and 0.01 mm thapsigargin. The prepulse reduces Ca²⁺ transient amplitude and rate of rise only in WT myocytes. The reduction is more profound at 1 mm as compared to 2.5 mm external Ca²⁺. Scale bar =F/F₀ ratio of 0.1.

Figure 6. Summary of effects of I_Na inactivation on trigger Ca²⁺.

The prepulse reduces Ca²⁺ transient amplitude and rate of rise in WT cardiomyocytes with depleted SR. A, WT cells at 1 mm external Ca²⁺. B, WT cells at 2.5 mm external Ca²⁺. C, NCX KO cardiomyocytes at 1 mm external Ca²⁺. Histograms comparing amplitude (upper panels) and rate of rise (lower panels) of trigger Ca²⁺ transients. *P < 0.05; **P < 0.001.

We were surprised by the large magnitude of the effect of I_Na on Ca²⁺ influx. With 1 mm Ca²⁺ in the bath solution, we were concerned that we might be overestimating the contribution of reverse NCX to trigger Ca²⁺ flux. To investigate this possibility, we repeated these experiments using 2.5 mm extracellular Ca²⁺. The higher Ca²⁺ would be expected to increase the contribution of I_Ca whereas reverse NCX activity may already be saturated at 1 mm Ca²⁺. With 2.5 mm Ca²⁺, I_Na inactivation induced a smaller relative decrease in the Ca²⁺ influx. Transient amplitude was decreased by 37.8 ± 6% and rate of rise of the transient was reduced by 30.6 ± 7.7% in WT myocytes (P < 0.001 and P < 0.05, n= 9; Figs 5 and 6). The experiment at 2.5 mm Ca²⁺ was not carried out using KO myocytes.

TTX decreases Ca²⁺ transient amplitude in WT but not in NCX KO cardiomyocytes

As an alternative to inactivating I_Na with a depolarizing prepulse, we blocked I_Na using TTX (10 μm). The effects of TTX on the Ca²⁺ transients of WT and NCX KO myocytes were almost identical to the effects of the depolarizing prepulse described above. In WT myocytes with intact SR, exposure to TTX reduced Ca²⁺ transient amplitude and rate of rise by 60.2 ± 4.4% and 59.4 ± 7.9%, respectively (n= 5, P < 0.05; Fig. 7). Strikingly, TTX had no effect on Ca²⁺ transients in NCX KO cardiomyocytes. Furthermore, TTX had no effect on the Ca²⁺ transient if the Na⁺ channel had already been inactivated by a prepulse (middle panels, Fig. 7). These results confirm that elimination of I_Na reduces the Ca²⁺ transient.

Figure 7. Effects of TTX on Ca²⁺ transients in WT and NCX KO myocytes.

Top, representative Ca²⁺ transient recordings in the presence (thin line) or absence of 10 μm TTX (thick line). Measurements were made in the presence or absence of a depolarizing ramp prepulse as designated. Bottom, summary of effects of TTX on Ca²⁺ transients. n= 5; *P < 0.05. Scale bar =F₃₆₀/F₃₉₀= 0.1.

Overexpression of NCX increases the effect of I_Na on Ca²⁺ transient amplitude

We tested whether the effect of I_Na on the Ca²⁺ transient would be more prominent in myocytes overexpressing NCX. Higher exchange activity could possibly lead to a greater contribution of NCX to CICR. Myocytes from our homozygous NCX-overexpressing mice exhibit about a 150% increase in NCX activity (Adachi-Akahane et al. 1997). We carried out experiments, as described above, using myocytes from homozygous NCX-overexpressing mice and their corresponding controls with intact SR at 1 mm extracellular Ca²⁺. Eliminating I_Na reduced the Ca²⁺ transient amplitude in the NCX-overexpressing myocytes by 62.1 ± 8.8% (n= 11) whereas, in age-matched controls, Ca²⁺ transient amplitude reduction was only 43.7 ± 7.7% (n= 9). The differences between these two groups of myocytes did not quite reach statistical significance (P= 0.07).

Discussion

The issue of whether I_Na and reverse NCX contribute to the CICR process in cardiac myocytes has been addressed previously, most notably in the pioneering work of LeBlanc & Hume (1990). However, the contribution of reverse NCX to CICR and the formation of the Ca²⁺ transient in cardiac cells remains controversial (Litwin et al. 1998; LeBlanc & Hume, 1990; Sham et al. 1992; Lipp & Niggli, 1994; Lines et al. 2006; Vites & Wasserstrom, 1996; Evans & Cannell, 1997). The availability of NCX KO mice gives us a unique opportunity to test the hypothesis that I_Na-driven reverse NCX is an important component of CICR. Our strategy, similar to that originally used by LeBlanc & Hume (1990), was to monitor evoked Ca²⁺ transients in the presence and absence of I_Na. The larger Ca²⁺ transient in the presence of I_Na supports the hypothesis that Na⁺ accumulation in the restricted space of the diadic cleft activates reverse NCX, which then adds Ca²⁺ to the pool that triggers RyRs. This is strongly supported by our finding that this effect is absent in myocytes from NCX KO hearts.

We inhibited I_Na by applying a prepulse using a slow ramp depolarization to −45 mV, a voltage at which we find Na⁺ channels are inactivated. The slow ramp prepulse avoided the accumulation of Na⁺ in the restricted space of the diadic cleft, which might have occurred if a square wave prepulse had been used to inactivate I_Na. We used an AP waveform for the test pulse to mimic physiological stimulation and minimize voltage errors associated with rapid depolarization to an arbitrary test potential. Successful use of a prepulse to inactivate fast Na⁺ currents has been reported previously (Stafstrom et al. 1982; Magistretti & Alfonso, 1999; Sobie et al. 2008). We could not test directly whether I_Ca was inhibited by our prepulse protocol to inactivate I_Na, but this seems unlikely for three reasons. First, adjusting the ramp duration had no effect on the subsequent Ca²⁺ transient or I_Ca during a clamp to 0 mV (Fig. 2). Second, we saw no effect of the prepulse on the Ca²⁺ transient in NCX KO myocytes. Had there been a decrease of I_Ca with the prepulse, this should have also resulted in a smaller Ca²⁺ transient in KO cells. Third, a prepulse in the absence of Na⁺ has no effect on I_Ca in rabbit cardiomyocytes (N. Torres and J. H. B. Bridge, unpublished data).

Although we did not directly measure SR Ca²⁺ load, it seems unlikely that alterations in SR Ca²⁺ confound the results of our study. First, we have previously reported that SR Ca²⁺ content is identical in myocytes from WT and NCX KO myocytes (Pott et al. 2005). Second, it is improbable that both a depolarizing prepulse and TTX would reduce SR Ca²⁺ in only WT, but not KO myocytes. Third, NCX activity occurring during a depolarizing prepulse might increase SR load in WT myocytes and would be expected to increase the Ca²⁺ transient. Instead, we found that the prepulse decreased the Ca²⁺ transient in WT myocytes. Fourth, even in the absence of a functional SR, a depolarizing prepulse decreases the Ca²⁺ transients of WT, but not NCX KO, myocytes (Fig. 6). Thus, we see effects of I_Na on Ca²⁺ influx that cannot be accounted for by alterations in SR Ca²⁺.

Another possible source of an artifact would be loss of voltage control in the presence of I_Na, which may occur at negative clamp potentials (Sham et al. 1992). The escape to more positive potentials might alter I_Ca. However, this is unlikely to explain our results for two reasons: First, the peak of our action potential clamp is greater than 0 mV and closer to the Na⁺ reversal potential. Thus, additional depolarization due to voltage escape will be minimal and, in any case, would not cause further activation of I_Ca than already occurs at 0 mV. Second, effects due to loss of voltage control should be similar in Ca²⁺ transients from both WT and NCX KO myocytes.

Ca²⁺ transients recorded in the presence of thapsigargin and ryanodine reflect Ca²⁺ entry across the sarcolemma and reflect the Ca²⁺ that triggers SR Ca²⁺ release under normal conditions. We propose that in normal myocytes this trigger Ca²⁺ is composed of both Ca²⁺ entry through L-type Ca²⁺ channels, as well as Ca²⁺ entry via reverse NCX in response to I_Na. It is interesting that in the absence of a functional SR, eliminating I_Na with a prepulse reduced the Ca²⁺ signal in WT cells to levels well below that observed in NCX KO myocytes (compare Fig. 6A and C). The degree of Ca²⁺ influx in WT myocytes with a prepulse ought to be similar to Ca²⁺ influx in a KO myocyte. That is, under these conditions, in both cases, the Ca²⁺ influx should be carried solely by the Ca²⁺ current. A reason why this might not be reflected by the Ca²⁺ transient shown in Fig. 6A is as follows: In the WT myocytes, NCX is still present and will be poised to efflux Ca²⁺, especially during a prepulse when there is no bolus of Na⁺ from I_Na. This could provide a sufficient Ca²⁺ efflux mechanism to dampen the Ca²⁺ transient, particularly since the amplitude of the Ca²⁺ transient will be small under conditions of SR Ca²⁺ depletion. Also, there may be adaptations that we do not yet fully understand in the NCX KO myocytes. Indeed, we have previously provided evidence that diadic cleft Ca²⁺ is higher in KO myocytes (Pott et al. 2007). Overall, the data indicate that the Ca²⁺ transients are influenced by reverse exchange whether or not the SR is functional.

It is also improbable that Ca²⁺ influx during the ramp partially inactivated LCCs and decreased Ca²⁺ influx since we found no discernible effect of ramp duration on the amplitude or kinetics of I_Ca (Fig. 2). Additionally, we found no detectable increase in resting Ca²⁺ during the prepulse to −45 mV, so any Ca²⁺ influx during this time is small. Also, we have confirmed our results on the effects of I_Na inactivation using TTX as an alternative approach to eliminate I_Na. Potential problems introduced by the effects of a prepulse on NCX would probably not occur with the use of TTX to block I_Na. There may be other possible sources of artifacts that we have not considered but, in general, it is unlikely that artifacts altering I_Ca or SR Ca²⁺ load would affect Ca²⁺ transients in WT but not NCX KO myocytes.

We explain the effects of I_Na on CICR as follows. Na⁺ entry into the diadic cleft during the upstroke of the AP produces favourable conditions for reverse NCX activity (LeBlanc & Hume, 1990; Lederer et al. 1990). The resulting Ca²⁺ entry contributes to CICR. Several investigators have found that Ca²⁺ provided by LCCs is a much more efficient trigger for CICR than Ca²⁺ provided by the exchanger (Sham et al. 1992; Sipido et al. 1997), so we do not propose that NCX is directly triggering Ca²⁺ release. Perhaps, reverse NCX rapidly primes the diadic cleft with Ca²⁺ and together with Ca²⁺ entry through LCCs, triggers Ca²⁺ release from the SR. We suggest that reverse NCX can synergistically enhance the efficiency of the LCCs to trigger CICR when both are present (Lines et al. 2006; Sobie et al. 2008). We find that inactivation of I_Na in WT cells reduces CICR by 50%. When the SR is disabled by thapsigargin and ryanodine, inactivation of I_Na also reduces Ca²⁺ entry in WT cells but has no effect in KO. Thus there is a direct correlation between the extent of Ca²⁺ entry and SR Ca²⁺ release in WT cells that is modified by I_Na. If the reduction in Ca²⁺ entry or Ca²⁺ release were due to voltage errors or inactivation of I_Ca by the prepulse, then we would expect similar results in NCX KO cells. However, KO cells were not affected by inactivation of I_Na.

It is interesting that in myocytes from NCX-overexpressing mice there was a trend towards a greater dependence of Ca²⁺ release on I_Na compared to control cells expressing normal levels of NCX. However, these results did not reach statistical significance.

Conclusion

We conclude that I_Na and NCX are essential components of the Ca²⁺ release trigger in mouse cardiac myocytes. Activation of I_Na during the upstroke of the AP causes rapid Na⁺ influx into the diadic cleft which, together with depolarization to positive potentials, reverses NCX and promotes Ca²⁺ entry via the exchanger into the diadic cleft space. This primes the diadic cleft with Ca²⁺ and acts synergistically with the opening of LCCs. We suggest that a non-linear summation of these two triggers, i.e. reverse NCX and LCCs results in more extensive activation of RyRs by LCC openings.

Glossary

Abbreviations

AP

action potential

CICR

Ca²⁺-induced Ca²⁺ release

I_Ca

L-type Ca²⁺ channel current

I_Na

Na⁺ current

I_NCX

Na⁺–Ca²⁺ exchange current

LCC(s)

L-type Ca²⁺ channel(s)

KO

knockout

NCX

Na⁺–Ca²⁺ exchange

RyR(s)

ryanodine receptor(s)

SR

sarcoplasmic reticulum

TTX

tetrodotoxin

WT

wild-type

Author contributions

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