Alterations in action potential profile enhance excitation-contraction coupling in rat cardiac myocytes

Abstract

1. Action potential (AP) prolongation typically occurs in heart disease due to reductions in transient outward potassium currents (I_to), and is associated with increased Ca²⁺ transients. We investigated the underlying mechanisms responsible for enhanced Ca²⁺ transients in normal isolated rat ventricular myocytes in response to the AP changes that occur following myocardial infarction.

2. Normal myocytes stimulated with a train of long post-myocardial infarction (MI) APs showed a 2.2-fold elevation of the peak Ca²⁺ transient and a 2.7-fold augmentation of fractional cell shortening, relative to myocytes stimulated with a short control AP.

3. The steady-state Ca²⁺ load of the sarcoplasmic reticulum (SR) was increased 2.0-fold when myocytes were stimulated with trains of long post-MI APs (111 ± 21.6 μmol l⁻¹) compared with short control APs (56 ± 7.2 μmol l⁻¹).

4. Under conditions of equal SR Ca²⁺ load, long post-MI APs still resulted in a 1.7-fold increase in peak [Ca²⁺]_i and a 3.8-fold increase in fractional cell shortening relative to short control APs, establishing that changes in the triggering of SR Ca²⁺ release are largely responsible for elevated Ca²⁺ transients following AP prolongation.

5. Fractional SR Ca²⁺ release calculated from the measured SR Ca²⁺ load and the integrated SR Ca²⁺ fluxes was 24 ± 3 and 11 ± 2 % following post-MI and control APs, respectively.

6. The fractional release (FR) of Ca²⁺ from the SR divided by the integrated L-type Ca²⁺ flux (FR/∫F_Ca,L) was increased 1.2-fold by post-MI APs compared with control APs. Similar increases in excitation-contraction (E-C) coupling gains were observed establishing enhanced E-C coupling efficiency.

7. Our findings demonstrate that AP prolongation alone can markedly enhance E-C coupling in normal myocytes through increases in the L-type Ca²⁺ current (I_Ca,L) trigger combined with modest enhancements in Ca²⁺ release efficiency. We propose that such changes in AP profile in diseased myocardium may contribute significantly to alterations in E-C coupling independent of other biochemical or genetic changes.

Action potential (AP) prolongation is universally observed in human patients with heart disease as well as in several animal models as a result of a reduction in the expression of potassium channel genes Kv4.2 and Kv4.3 which encode the channels for the transient outward current (I_to) (Beuckelmann et al. 1993; Brooksby et al. 1993; Cerbai et al. 1994; Kaab et al. 1996; Wickenden et al. 1998; Kaprielian et al. 1999). We and others have shown previously that this AP prolongation results in elevated [Ca²⁺]_i (Bouchard et al. 1995; Wickenden et al. 1998; Kaprielian et al. 1999) which enhances contractility of compromised myocardium (Fiset et al. 1997). The increase in Ca²⁺ transients and contractility observed following AP prolongation may be explained by an enhanced triggered sarcoplasmic reticulum (SR) Ca²⁺ release, an elevated SR Ca²⁺ load or both. Previous studies have attributed increases in inotropy following AP prolongation to elevations in SR Ca²⁺ load (Brooksby et al. 1993; Bouchard et al. 1995). Changes in SR Ca²⁺ load have also been shown to influence the efficiency of SR Ca²⁺ release (Han et al. 1994; Janczewski et al. 1995; Santana et al. 1997) although these studies used step depolarizations to trigger Ca²⁺ release rather than APs. On the other hand, elevations in SR Ca²⁺ release have also been linked to increases in L-type Ca²⁺ current (I_Ca,L) trigger. Specifically, prolongation of step depolarization duration, particularly in the range 2-20 ms, enhanced SR Ca²⁺ release independently of SR Ca²⁺ load by increasing the I_Ca,L trigger (Isenberg & Han, 1994). This strong dependence of SR Ca²⁺ release on step duration suggests that a similar mechanism may underlie altered Ca²⁺ release in heart disease since the early repolarization phase of the AP, and the corresponding temporal changes in I_Ca,L, are primarily affected by I_to reductions (Kaprielian et al. 1999).

Recently, changes in E-C coupling efficiency have been reported in different animal models of hypertensive heart disease, with one study reporting enhanced SR Ca²⁺ release (Shorofsky et al. 1999) and another finding impaired release (Gomez et al. 1997). Unfortunately neither of these studies took into account the possible effects of AP prolongation that are known to occur in spontaneous hypertensive rats (Brooksby et al. 1993; Cerbai et al. 1994). This is particularly relevant since AP prolongation in spontaneous hypertensive rats (Brooksby et al. 1993) and in rats following myocardial infarction has been directly linked to increased Ca²⁺ transient amplitudes, although the underlying mechanism for these changes was not thoroughly investigated. In addition, based on previous studies demonstrating I_Ca,L-independent forms of SR Ca²⁺ release such as Na⁺-Ca²⁺ exchange-mediated release (Levesque et al. 1991; Wasserstrom & Vites, 1996; Sipido et al. 1997; Litwin et al. 1998) and voltage-sensitive release (Ferrier et al. 1998; Howlett et al. 1998), one might anticipate that prolonged APs might also alter the efficiency of Ca²⁺ release.

This study was designed to determine the mechanism for the changes in Ca²⁺ transients associated specifically with AP prolongation observed following myocardial infarction in terms of changes in trigger I_Ca,L versus changes in SR Ca²⁺ load. We find that when normal myocytes are stimulated with prolonged action potentials, positive inotropic effects are mediated primarily through an increase in triggered Ca²⁺ release and are secondarily due to an enhancement in SR Ca²⁺ load. This elevated Ca²⁺ release appears to result primarily from increased Ca²⁺ influx mediated by I_Ca,L combined with a modest enhancement in the efficiency of Ca²⁺ release evoked by post-MI APs.

METHODS

Isolation of rat ventricular myocytes

The procedure for isolation of adult rat ventricular myocytes was adapted from our previous studies (Wickenden et al. 1997). Male Sprague-Dawley rats (300-400 g, Charles River) were heparinized and killed by intraperitoneal injection of a lethal dose of anaesthetic (sodium pentobarbital, 200 mg kg⁻¹) in accordance with the Guidelines of the Animal Care and Use Committee of the University Health Network. Hearts were cannulated and retrogradely perfused through the aorta for about 1 min with a Ca²⁺-containing standard Tyrode solution (mm): 140 NaCl, 5.4 KCl, 10 Hepes, 1 MgCl₂, 1 CaCl₂ and 10 d-glucose, adjusted to pH 7.4 with NaOH at 37 °C. The hearts were then perfused with nominally Ca²⁺-free standard Tyrode solution for 5 min prior to digestion with the same solution containing collagenase (Type II, 0.38 mg ml⁻¹, Boehringer-Mannheim) and protease (Type XIV, 0.03 mg ml⁻¹, Sigma) for 8-9 min. The enzyme solution was then washed out by perfusing for 2-3 min with a Kraft-Brühe (high K⁺) solution (mm): 120 potassium glutamate, 20 KCl, 10 Hepes, 1 MgCl₂, 0.3 K-EGTA and 10 d-glucose, pH 7.4. All solutions were pre-bubbled with 100 % O₂ for 5 min. Following the enzyme washout, the atria and blood vessels were removed and ventricles separated. The right and left ventricular free walls were dissected from the remainder of the heart. Myocytes were minced and mechanically dissociated in high K⁺ solution and then filtered through a nylon mesh. The cells were then resuspended in high K⁺ solution containing bovine serum albumin (0.04 % w/v) and gentamycin (0.05 mg ml⁻¹) and used within 12 h after isolation. Only Ca²⁺-tolerent, quiescent, rod-shaped myocytes with clear cross-striations were selected for electrophysiological, intracellular Ca²⁺ and unloaded cell shortening measurements.

Intracellular Ca²⁺ and cell shortening measurements during AP clamps

Freshly isolated rat ventricular myocytes were placed in a bath on the stage of an Olympus IX50 inverted microscope and perfused at room temperature (20-23 °C at approximately 1 ml min⁻¹) with extracellular solution of the following composition (mm): 140 NaCl, 4 KCl, 10 Hepes, 1 MgCl₂, 2 CaCl₂ and 10 d-glucose, adjusted to 7.4 with NaOH. Myocytes were voltage clamped using the whole-cell patch clamp technique (Hamill et al. 1981) with an Axopatch 200A amplifier (Axon Instruments). Microelectrodes were pulled from thin-walled 1.5 mm diameter borosilicate glass (World Precision Instruments) using a Flaming-Brown micropipette puller (Sutter Instruments) and heat polished to a final resistance of 1-2 MΩ when filled with a solution containing (mm): 140 KCl, 10 Hepes, 1 MgCl₂, 10 NaCl, 7 MgATP, 0.060 fura-2 pentapotassium salt, adjusted to pH 7.2 with KOH. Series resistance compensation ranged between 80 and 90 %. Fluorescence measurements were performed using light from a 75 W xenon lamp (Oriel Corp., CT, USA) passed through bandpass filters centred at either 340 or 380 nm (Chromatech, VT, USA). The emitted fluorescence was collected by a × 40 Uapo/340 objective lens (Olympus America, Melville, NY, USA) and passed through a 510 nm filter to a photomultiplier tube (R2693, Hamamatsu, Japan). The photomultiplier output was filtered at 100 Hz, recorded using an A/D data acquisition board (Model PP-50 Warner, CT, USA) and stored in a computer for analysis. The ratio (R) of the background-subtracted fluorescence signal (340/380) was used to estimate [Ca²⁺]_i as described previously (Kaprielian et al. 1999). In our experiments β, which is defined as the ratioof fluorescence measured with 380 nm excitation light in the absence of Ca²⁺ to that measured at saturating levels of Ca²⁺ (10 mm), was 11.3 ± 0.3 (n = 2). The fluorescence ratio in the presence of saturating Ca²⁺ (R_max) was 5.9 ± 0.1 (n = 2) and that in the absence of Ca²⁺ (R_min) was 0.23 ± 0.03 (n = 2). Cells were stimulated with either step depolarizations or AP clamps using waveforms obtained from control and post-MI myocytes (Kaprielian et al. 1999). All [Ca²⁺]_i and cell shortening measurements were made under steady-state conditions following loading trains of these APs or 100 ms steps at a frequency of 0.25 Hz. To obtain a fluorescence measurement from both 340 and 380 nm excitation wavelengths, each protocol was repeated at least once. The reliability of using this approach was validated in several cells by repeated fluorescence measures at 340 and 380 nm excitation following the loading trains and the measurements proved to be stable throughout the duration of the experiment.

Unloaded cell shortening was measured simultaneously with fluorescence using a CCD video camera mounted on the sideport of the microscope. Custom-made relay lenses were installed in the sideport to reduce the image of myocytes to fit into the active area of the camera. Raster lines were positioned over both edges of the myocyte and the focus and detection thresholds were adjusted to maximize the signal-to-noise ratio. Dual-edge motion was monitored by an edge-detection system (Crescent Electronics, Salt Lake City, UT, USA), digitized on-line and stored on a personal computer using pCLAMP software.

The fluorescence signals and cell motion signals were both digitized at 4 kHz. Cell motion was measured at 60 Hz, which is the acquisition rate of the CCD camera used. The fluorescence signals were filtered at 100 Hz.

L-type Ca²⁺ current during AP clamps

Calcium current and Ca²⁺ transients were measured simultaneously during an AP clamp by eliminating Na⁺ and K⁺ currents followed by a Cd²⁺ subtraction. Interfering K⁺ currents were eliminated by replacing almost all intracellular and extracellular K⁺ with Cs⁺. The intracellular solution contained (mm): 140 CsCl, 2 KCl, 10 Hepes, 1 MgCl₂, 10 NaCl, 7 MgATP and 0.060 fura-2 pentapotassium salt, adjusted to pH 7.2 with CsOH. The extracellular solution contained (mm): 140 NaCl, 4 CsCl, 10 Hepes, 1 MgCl₂, 2 CaCl₂, 10 d-glucose and 30 μm TTX, pH 7.4 with NaOH. It was applied using a rapid local superfusion device (BMT Research Services, Calgary, Alberta, Canada). After Ca²⁺ currents and transients were measured the extracellular solution was switched to one containing 500 μm CdCl₂ and the trace acquired was used for subtraction to eliminate all other background currents.

Measurement of SR Ca²⁺ load

SR Ca²⁺ load was estimated following trains of either eight AP waveforms or 100 ms steps to 10 mV applied every 4 s to establish steady-state SR load conditions. After eight loading pulses the myocyte was rapidly superfused with standard Tyrode solution containing 20 mm caffeine. Application of caffeine caused myocyte contraction as a result of Ca²⁺ release from the SR (Varro et al. 1993). Throughout the caffeine application the membrane potential was held at -80 mV to enhance I_Na-Ca and thereby allowed estimation of the SR Ca²⁺ content by integrating I_Na-Ca over the duration of the caffeine application. Cell volume was calculated from the mean cell membrane capacitance (178 ± 7 pF, n = 18) and converted to volume by assuming a surface area to volume ratio of 0.5 μm⁻¹ (Page, 1978) and a specific capacity of 1 μF cm⁻². Calcium efflux via non-electrogenic pathways was corrected for by dividing by 0.87, in accordance with the findings of Bassani et al. (1994) that 87 % of the Ca²⁺ flux during a caffeine contracture occurs through Na⁺-Ca²⁺ exchange.

Computation of SR Ca²⁺ flux, F_rel, and measurement of E-C coupling gain

The SR Ca²⁺ release flux (F_rel) was determined from Ca²⁺ transient recordings using the method of Wier et al. (1994). Calcium influx into the myoplasm via sarcolemmal I_Ca,L is believed to contribute little to the amplitude of the Ca²⁺ transient in the rat (Terracciano & MacLeod, 1997). Indeed, our calculations confirmed that Ca²⁺ influx represents less than 6 % of the total Ca²⁺ released (see below, Fig. 8A). Accordingly, the rate of change of [Ca²⁺]_i is almost entirely determined by the net flux of Ca²⁺ across the SR membrane plus the binding and unbinding to intracellular ligands as given by:

(1)

where F_rel is the SR Ca²⁺ release flux, F_pump is the SR Ca²⁺ uptake flux via the SR ATPase, F_leak is the background Ca²⁺ leak from the SR to the cytoplasm (fixed at 0.0175 mm s⁻¹; Balke et al. 1994), [CaL] is the concentration of ligand-bound Ca²⁺ and d/dt represents the first derivative with respect to time. For the purposes of SR Ca²⁺ flux estimation, a corrected calcium transient ([Ca²⁺]_i,corr) was calculated taking into account the kinetics of binding of Ca²⁺ to fura-2 as described by Sipido & Wier (1991):

(2)

For this calculation the value of k_off was taken to be 23 s⁻¹ (Baylor & Hollingworth, 1988) and k_on was determined to be 1.03 × 10⁸m⁻¹ s⁻¹ assuming a K_D for fura-2 of 224 nm. The time derivative, d[Ca²⁺]_i,corr/dt, is calculated from the corrected calcium transient recording in a three-point, sliding-window, linear model using a least squares estimator. The expressions for F_pump (Balke et al. 1994) and d[CaL]/dt (Sipido & Wier, 1991) were:

(3)

(4)

where V_max is the maximal SR uptake flux defined as 0.2 mm s⁻¹ (Balke et al. 1994), L represents intracellular ligands including calmodulin, troponin and fura-2 and and are binding rate constants for the respective ligands. The affinity constant (K_m) was adjusted for each cell to eliminate baseline SR flux, as done previously (Wier et al. 1994). In nine cells, K_m ranged from 0.09 to 0.31 μm with an average of 0.178 ± 0.006 μm, which is comparable to previously reported values (Balke et al. 1994).

Figure 8. Excitation-contraction coupling gains and fractional SR Ca²⁺ release during short control and long post-MI action potentials.

A, a 1.2-fold enhancement in E-C coupling gain in myocytes stimulated with a long post-MI AP (99.6 ± 11.7, n = 9) relative to a short control AP (81.6 ± 11.5, n = 9) when gain is defined as the estimated amount of SR Ca²⁺ release (∫F_rel) over the amount of trigger Ca²⁺ (∫F_Ca,L). B, the E-C coupling gain as defined by the rate of cell shortening over the amount of trigger Ca²⁺. According to this definition, E-C coupling efficiency is elevated 1.5-fold upon stimulation of myocytes with a long post-MI AP (124.6 ± 28.2 μm s⁻¹ mm⁻¹, n = 8) relative to a short control AP (84.8 ± 24.2 μm s⁻¹ mm⁻¹, n = 8). C, fractional release (FR) is enhanced by the post-MI AP (24 ± 3 %, n = 9) over the control AP (11 ± 2 %, n = 9) and the FR normalized to the amount of trigger Ca²⁺ (∫F_Ca,L) was increased 1.2-fold (D). *P < 0.05.

Using these definitions, the SR Ca²⁺ release flux (F_rel) can be obtained directly from the kinetically corrected Ca²⁺ transient (eqn (2)) and eqn (1). The integral of F_rel provides a measure of the amount of Ca²⁺ released by the SR following an AP. An estimate of Ca²⁺ influx responsible for triggering Ca²⁺ release (i.e. ∫F_Ca,L) was determined from the integrated Ca²⁺ current according to:

where z is the valence of Ca²⁺, F is Faraday's constant (9.648 × 10⁴ C mol⁻¹), the factor 0.5 is the fraction of the total cell volume assumed to be accessible to Ca²⁺ (Wier et al. 1994) and V is the total cell volume calculated from the mean cell membrane capacitance (178 ± 7 pF, n = 18) as described above. The term represents twice the integrated I_Ca,L from the baseline to the peak and reflects the contribution of first openings of L-type Ca²⁺ channels (i.e. first latency distribution) to I_Ca,L (Rose et al. 1992). Previous studies have established that the time course of Ca²⁺ release coincides closely with the first latency of L-type Ca²⁺ channel activation (Isenberg & Han, 1994; Lopez-Lopez et al. 1995; Sham et al. 1998; Cleemann et al. 1998; Collier et al. 1999) and that the Ca²⁺ release correlates well with the cumulative probability density distribution of first latency of L-type Ca²⁺ channels or the activation time course of the whole-cell Ca²⁺ current (Isenberg & Han, 1994). Therefore, should provide an accurate estimate of the Ca²⁺ influx that is responsible for triggering Ca²⁺ release from the SR. It is important to point out that the values obtained using to measure trigger Ca²⁺ are very similar to the integral of I_Ca,L over the first 20 ms (i.e. ∫I_Ca,20msdt) which was previously suggested to provide a better measure of I_Ca,L directly involved in the Ca²⁺ release process (Fabiato, 1985; Janczewski et al. 1995).

Two definitions of E-C coupling gain were used to assess the efficiency of Ca²⁺ release from the SR in our studies:

where ∫F_rel is the integrated SR Ca²⁺ release flux or total SR Ca²⁺ released, and d(CS)/dt is the rate of cell shortening. These estimates of gain are similar to previous definitions of Wier et al. (1994) and Litwin et al. (1998) except that integrated fluxes are used instead of peak fluxes or peak current amplitudes. As a result, the E-C gain was calculated as a ratio of the total SR Ca²⁺ release to the total trigger Ca²⁺ influx. This modification was necessary since peak I_Ca,L amplitude, which has previously been used as a measure of trigger I_Ca,L during step depolarizations (Wier et al. 1994; Santana et al. 1997; Litwin et al. 1998), depends strongly on the AP profile (see below, Fig. 6) (Bouchard et al. 1995; Kaprielian et al. 1999).

Figure 6. L-type Ca²⁺ currents and Ca²⁺ transients evoked by short control and long post-MI action potentials.

A, the voltage protocols: control (left) and post-MI (right) APs were applied to Cs⁺-loaded rat ventricular myocytes following a loading train of eight 100 ms steps to +10 mV. B, Ca²⁺ currents evoked by the control and post-MI APs were measured as Cd²⁺-sensitive currents and differed in peak amplitude, time to peak and integrated current. Table 1 summarizes the characteristics of I_Ca,L during control and post-MI APs. C, representative intracellular Ca²⁺ transients triggered by control and post-MI APs. D, peak systolic Ca²⁺ was not significantly different between control (319.2 ± 16.1 nm, n = 5) and post-MI APs (433.4 ± 48.2 nm, n = 5). E, normalized peak Ca²⁺ transients of the post-MI AP were only 1.4-fold greater than the control AP under these intracellular conditions.

From the calculated SR fluxes, another index of calcium release used in our studies was fractional release (FR), which was estimated as the integrated SR Ca²⁺ flux divided by the measured total SR Ca²⁺ content following a loading train of 100 ms steps (∫F_rel/total SR content).

Data analysis

All data are presented as means ±s.e.m. with the number of cells in parentheses. Comparisons of data for each cell between long post-MI APs and short control APs were performed using Student's two-tailed, paired t test. An unpaired t test was used when comparing data that did not originate from the same myocyte. An experimental P value < 0.05 was considered statistically significant.

RESULTS

Effects of the long post-MI AP profile on [Ca²⁺]_i, cell shortening and SR load

In this study normal, freshly isolated rat cardiomyocytes were stimulated with either short control or long post-MI APs that had been previously recorded in myocytes derived from control or post-MI rats (Fig. 1A). In each case the SR was loaded to a steady-state level by introducing eight conditioning APs (Trafford et al. 1997) after which a test AP was applied. Figure 1B and D (left) shows that peak Ca²⁺ transient magnitudes in response to stimulation with test APs were significantly elevated from 190 ± 37 nm(n = 5) with short control APs to 417 ± 115 nm(n = 5) with long post-MI APs, without affecting the resting diastolic levels. Associated with elevated Ca²⁺ transients there was a 2.7-fold increase in peak fractional cell shortening (P < 0.05) triggered by a long post-MI AP (0.084 ± 0.016, n = 7) compared with a short control AP (0.031 ± 0.010, n = 7), shown in Fig. 1C and D (right). These results establish that AP prolongation, as occurs with a reduction in I_to following myocardial infarction, can significantly enhance contractility of the surviving myocytes.

Figure 1. Prolonged action potentials increase Ca²⁺ transients and unloaded cell shortening in rat cardiac myocytes.

A, trains of eight short control APs (left) and long post-MI APs (right) were used to load the myocyte to steady-state SR Ca²⁺ load. B, Ca²⁺ transients measured from the final 9th control AP (left panel) and post-MI AP (right panel). C, unloaded cell shortening associated with the intracellular Ca²⁺ transient and triggered by the control AP (left) and post-MI AP (right). D, peak intracellular [Ca²⁺] (left) is elevated 2.2-fold upon stimulation with a post-MI AP relative to the control AP while fractional cell shortening (right) is increased 2.7-fold. *P < 0.05.

The mechanism responsible for increased peak systolic Ca²⁺ transients and peak cell shortening following AP prolongation may arise from enhanced SR Ca²⁺ loading, enhanced triggered SR Ca²⁺ release, or both. To examine the contribution of SR Ca²⁺ load to these measured changes in [Ca²⁺]_i we recorded the inward Na⁺-Ca²⁺ exchange current at -80 mV in response to a rapid application of 20 mm caffeine to the cell, and converted the integrated current into SR Ca²⁺ content (Varro et al. 1993), using a correction factor of 0.87 to account for Ca²⁺ extrusion via non-electrogenic means (Bassani et al. 1994). As shown in Fig. 2, the SR Ca²⁺ content after a train of short control APs (56 ± 7.2 μmol l⁻¹, n = 7) was reduced (P < 0.05) to about half of that measured with long post-MI APs (111 ± 21.6 μmol l⁻¹, n = 4) similar to previous studies (Terracciano et al. 1997).

Figure 2. SR Ca²⁺ load after conditioning trains of short control and long post-MI action potentials.

Representative current traces of inward Na⁺-Ca²⁺ exchange current during a 20 mm caffeine spritz after a train of short control APs (A) and long post-MI APs (B). C, SR Ca²⁺ content calculated from the integrated Na⁺-Ca²⁺ exchange current after loading trains of control APs (56 ± 7.2 μmol l⁻¹, n = 7) and post-MI APs (111 ± 21.6 μmol l⁻¹, n = 4). *P < 0.05.

SR Ca²⁺ release triggered by control and post-MI APs at equal, low and high SR Ca²⁺ loads

To address whether the positive inotropic effect of post-MI APs acts via increases in SR Ca²⁺ load or enhanced triggered Ca²⁺ release, we applied control and post-MI APs to normal myocytes loaded to a fixed steady-state level of SR Ca²⁺ using a train of eight 100 ms depolarizing steps to +10 mV (Fig. 3A) (Trafford et al. 1997). Typical intracellular Ca²⁺ transients and fractional cell shortening are shown in Fig. 3B and C, respectively, while Fig. 3D summarizes the Ca²⁺ transients and cell shortening from seven to nine cells. On average Ca²⁺ transients were significantly elevated 1.7-fold (P < 0.005) from 242 ± 22 to 404 ± 46 nm while fractional cell shortening increased 3.8-fold (P < 0.05) from 0.014 ± 0.004 to 0.053 ± 0.015. Thus, under conditions of equal SR Ca²⁺ load, post-MI AP waveforms alone are capable of triggering significantly more Ca²⁺ release and evoking a larger degree of cell shortening than control AP waveforms. A comparison of the data in Figs 1 and 3 further reveals that both peak Ca²⁺ transients and cell shortening triggered by either control or post-MI APs were similar whether myocytes were loaded to different extents with different APs or to the same extent using 100 ms steps. This suggests that changes in the triggering of Ca²⁺ release, rather than SR load, play a dominant role in mediating enhancements in SR Ca²⁺ release following AP prolongation under our experimental conditions.

Figure 3. Calcium transients and cell shortening triggered by action potentials under conditions of equal SR load.

A, the voltage protocols: control and post-MI APs were applied to myocytes after a loading protocol of eight 100 ms steps to +10 mV. B, representative Ca²⁺ transients triggered by the control AP (left) and post-MI AP (right). C, fractional cell shortening associated with the intracellular Ca²⁺ transients triggered by control (left) and post-MI (right) APs. D, peak Ca²⁺ transients (left) are elevated 1.7-fold upon stimulation with a post-MI AP relative to the control AP while fractional cell shortening (right) is increased 3.8-fold. *P < 0.05; **P < 0.005.

Next we investigated the possible contributions of SR load in mediating changes in Ca²⁺ release and cell shortening. Figure 4 shows that, following loading with eight short APs (i.e. low SR load), a post-MI AP triggered significantly (P < 0.05) more Ca²⁺ release (314 ± 64 nm, n = 7) than a control AP (198 ± 28 nm, n = 7) preceded by eight long post-MI APs (i.e. high SR load). Furthermore, fractional cell shortening was also significantly enhanced (P < 0.05) under these conditions (CS_control= 0.042 ± 0.010; CS_post-MI= 0.076 ± 0.016, n = 7). Remarkably, the Ca²⁺ released by a post-MI AP with a low SR load produced a 1.6-fold increase in peak [Ca²⁺]_i transient relative to control APs with a high SR load (Fig. 4D). By comparison, post-MI APs caused a 2.2-fold increase compared with control APs under the opposite loading conditions (Fig. 1). Thus, while the positive inotropic effects of AP prolongation appears to be dominated by enhancements in triggered SR Ca²⁺ release, a measurable, yet small, contribution of SR Ca²⁺ load to Ca²⁺ release also exists under our experimental conditions.

Figure 4. Effect of high and low SR load on Ca²⁺ transients and cell shortening triggered by action potentials.

A, the voltage protocols: trains of eight post-MI APs (left) or control APs (right) were used to achieve a high or low steady-state SR load. Ca²⁺ release was then triggered from the SR with either a control AP (left) or post-MI AP (right). B, representative intracellular Ca²⁺ transients triggered by the high SR load/control AP protocol (left) and low SR load/post-MI AP protocol (right). C, typical fractional cell shortening associated with the Ca²⁺ transients shown in B. D, peak Ca²⁺ transients (left) triggered by post-MI APs under conditions of low SR load are increased 1.6-fold over transients triggered by control APs at a high SR load, while fractional shortening measurements (right) show a 1.8-fold difference between low SR load/post-MI AP and high SR load/control AP protocols. *P < 0.05.

To further quantify the relative contributions of trigger and SR Ca²⁺ load to the amount of Ca²⁺ released from the SR following changes in AP profile, we calculated Δ[Ca²⁺] ([Ca²⁺]_systolic - [Ca²⁺]_diastolic) since this should provide a better measure of the changes in [Ca²⁺]_i resulting from SR Ca²⁺ release than does peak [Ca²⁺]_i (Han et al. 1994; Isenberg & Han, 1994; Janczewski et al. 1995; Santana et al. 1997). To eliminate inter-cell variability in this analysis, the Δ[Ca²⁺]_i released for each protocol was normalized by the response measured during a control AP following loading with a train of control APs. As shown in Fig. 5, the Ca²⁺ released at either low or high SR loads increased ≈2-fold when short control APs were replaced with long post-MI APs as expected if the trigger for Ca²⁺ release is the dominant determinant of release. This is similar to the 1.7-fold enhancement in Ca²⁺ release shown in Fig. 3 between control and post-MI APs under equal SR Ca²⁺ load conditions. On the other hand, the relative increase in Ca²⁺ release due to changes in SR load from low to high was found to be only 1.2-fold when triggered with either a short control or long post-MI AP. Collectively, these data strongly suggest that changes in SR Ca²⁺ load account for at most about 20 % of the differences in contractility observed following AP prolongation.

Figure 5. Effect of control and post-MI action potentials on Ca²⁺ release at low and high SR Ca²⁺ loads.

Relative Ca²⁺ release triggered by control and post-MI APs at low (control AP train) and high (post-MI AP train) SR loads. Data are normalized to the amplitude of Δ[Ca²⁺]_i triggered by the control AP at a low SR load.

Efficiency of SR Ca²⁺ release by control and post-MI APs

Having demonstrated that alterations in the AP waveform are capable of enhancing SR Ca²⁺ release, the question remains whether this arises exclusively from changes in I_Ca,L or whether there are also changes in the efficiency of Ca²⁺ release by I_Ca,L. To quantify the efficiency of SR Ca²⁺ release, I_Ca,L was measured simultaneously with Ca²⁺ transients. L-type Ca²⁺ current was recorded by a near complete replacement of intracellular K⁺ with Cs⁺ to eliminate interfering potassium currents and superfusion of the myocytes with extracellular solution containing 30 μm TTX to eliminate Na⁺ currents. The records in Fig. 6 show the Cd²⁺-sensitive L-type Ca²⁺ currents and associated intracellular Ca²⁺ transients produced by short control (left) and long post-MI APs (right) after loading with eight 100 ms steps. The amplitude and time course of Ca²⁺ influx driven by control and post-MI APs were remarkably different (Fig. 6B, left and right, respectively) and their characteristics are summarized in Table 1. Despite an ≈2-fold reduction in peak I_Ca,L (P < 0.001) with the post-MI AP relative to the control AP (I_Ca,L,control= 13.5 ± 1.9 pA pF⁻¹, I_Ca,L,post-MI= 6.2 ± 1.4 pA pF⁻¹, n = 5), the total integrated I_Ca,L (Q_control= 8.2 ± 1.4 pC, Q_post-MI= 23.8 ± 4.3 pC, n = 5) was actually increased about 3-fold (P < 0.01).

Table 1.

Parameters describing L-type Ca²⁺ flux and SR Ca²⁺ release during control and post-MI APs

AP	ICa,L (pA pF−1)	QCa,L (pC)		∫FCa,L (μm)	∫Frel (μm)	d(CS)/dt (μm s−1)
Control AP	13.5 ± 1.9	8.2 ± 1.4	7.7 ± 1.4	0.39 ± 0.07	31.7 ± 4.5	33 ± 9.4
Post-MI AP	6.2 ± 1.4*	23.8 ± 4.3*	14.1 ± 2.9*	0.71 ± 0.15*	70.9 ± 8.4*	88.8 ± 20*

Calcium influx and SR Ca²⁺ release were triggered by AP clamps after loading with eight 100 ms square waves as shown in Figs 6 and 7. I_Ca,L, peak Ca²⁺ current; Q_Ca,L, total charge influx obtained by integrating the L-type Ca²⁺ current trace (n = 5); indicates the charge influx reflecting the first latency of Ca²⁺ channel openings (n = 5); ∫F_Ca,L, the integrated trans-sarcolemmal Ca²⁺ flux derived from (n = 5); ∫F_rel, the integrated SR Ca²⁺ release flux (n = 9); d(CS)/dt, the rate of cell shortening (n = 8). Both ∫F_Ca,L and ∫F_rel are Ca²⁺ concentrations in the Ca²⁺-accessible cell volume or half the total cell volume (Sipido & Wier, 1991).

^*P < 0 05.

As expected from the differences in integrated I_Ca,L, Fig. 6C and D further establishes that peak Ca²⁺ transients were elevated following stimulation with post-MI relative to control APs, although the effect was less pronounced, possibly due to the replacement of K⁺ with Cs⁺, which impairs SR Ca²⁺ release (Levi et al. 1996; Wasserstrom & Vites, 1996; Litwin et al. 1998) (see Discussion). On the other hand, Ca²⁺ channel gating has previously been shown to be largely unaffected by high intracellular [Cs⁺] (Levi et al. 1996). Therefore, in an attempt to obtain the best estimate of the true E-C coupling gain, we decided to combine the I_Ca,L recordings from the Cs⁺ studies with the Ca²⁺ transient and cell shortening data recorded using high K⁺ pipette solutions (Isenberg & Han, 1994) (shown previously in Fig. 3), although similar results were observed when Ca²⁺ transients are recorded in the presence of Cs⁺ (see Discussion).

Excitation-contraction coupling gains provide estimates of the effectiveness of Ca²⁺ release in response to the I_Ca,L trigger and have been calculated previously in a number of different ways (Wier et al. 1994; Janczewski et al. 1995; Santana et al. 1997; Litwin et al. 1998). We used two different definitions (see Methods for rationale and details): E-C gain₁=∫F_rel/∫F_Ca,L and E-C gain₂= d(CS)/dt/∫F_Ca,L. As stated in Methods, our definitions required integration of the F_rel and F_Ca,L since the kinetic properties of I_Ca,L are strongly influenced by the AP profile. Typical SR release fluxes, F_rel, triggered by long post-MI and short control APs are shown in Fig. 7C, right and left, respectively, with their associated intracellular Ca²⁺ transients plotted in Fig. 7B. Figure 7D shows the corresponding rates of cell shortening. For the purposes of calculating E-C coupling gain, the integrated SR Ca²⁺ flux (∫F_rel, see Table 1), was used to provide an estimate of the total Ca²⁺ released from the SR. Using the first definition of E-C coupling gain, the efficiency of Ca²⁺ release by I_Ca,L during a long post-MI AP was elevated about 1.2-fold (P < 0.05) over the control AP (Fig. 8A). Similarly, the second definition of gain indicated a 1.5-fold (P < 0.05) enhancement in SR Ca²⁺ release efficiency by the long post-MI AP as compared with the short control AP (Fig. 8B). The calcium release from the SR was also quantified by calculating the fractional SR Ca²⁺ release by post-MI and control APs following 100 ms loading steps. The total SR content under these conditions was measured as 148 ± 15 μmol l⁻¹ and the fraction of SR Ca²⁺ released (FR) by the post-MI AP (24 ± 3 %, n = 9) was found to be elevated compared with the control AP (11 ± 2 %, n = 9) (Fig. 8C). Normalizing the FR to the integrated L-type Ca²⁺ flux (FR/∫F_Ca,L; Fig. 8D) revealed a 1.2-fold relative increase between control and post-MI APs. These results suggest that AP prolongation not only enhances triggered Ca²⁺ release by producing a more powerful Ca²⁺ influx trigger, but also modestly increases the efficiency of SR Ca²⁺ release by I_Ca,L.

Figure 7. Parameters reflecting Ca²⁺ release in isolated rat ventricular myocytes.

A, control (left) and post-MI (right) APs were applied to myocytes after a loading protocol of eight 100 ms steps to +10 mV. B, kinetically corrected intracellular Ca²⁺ transients triggered by control and post-MI APs. These data were used to calculate the SR release flux, F_rel, according to the equation: d[Ca²⁺]_i/dt =F_rel+F_pump+F_leak+Σd[CaL]/dt (see Methods). C, SR release flux, F_rel, triggered by short control and long post-MI APs. D, rate of unloaded cell shortening, d(CS)/dt, triggered by control and post-MI APs.

DISCUSSION

Changes in AP morphology and duration occur with heart disease in humans as well as in numerous animal models and are typically associated with reductions in transient outward currents (Beuckelmann et al. 1993; Coulombe et al. 1994; Wettwer et al. 1994; Potreau et al. 1995; Kaab et al. 1996; Wickenden et al. 1998; Kaprielian et al. 1999). We have shown previously that following myocardial infarction Ca²⁺ transient amplitudes are increased as a result of AP prolongation (Kaprielian et al. 1999). In this study, we extended this work by examining the possible mechanisms responsible for these observations.

Measurements of caffeine-induced Na⁺-Ca²⁺ exchange current revealed a doubling of the steady-state SR Ca²⁺ content in myocytes stimulated with post-MI APs versus control APs. However, changes in SR load only accounted for a fraction of the increased Ca²⁺ transient amplitude associated with AP prolongation characteristically observed in diseased myocardium (Kaprielian et al. 1999). Specifically, Δ[Ca²⁺]_i was only increased by about 20 % when the SR load was elevated by loading with long post-MI APs whether the trigger for Ca²⁺ release was a short control or long post-MI AP. By contrast, with equal SR Ca²⁺ loads, the increase in Ca²⁺ release triggered by the post-MI AP was consistently found to be about 2-fold larger than the control AP trigger regardless of the loading protocol. This relationship between Ca²⁺ release, SR load and AP trigger suggests that the effects of SR load on Ca²⁺ release, under the experimental conditions used in the present study (high intracellular K⁺, 23 °C), are small relative to the effects of the AP profile in triggering release. At first glance these findings appear to contradict a previous report concluding that the positive inotropic effects following AP prolongation result from increased SR Ca²⁺ loading (Bouchard et al. 1995). This discrepancy might be readily explained by differences in the experimental conditions. In our study, we compared the effects of long and short APs obtained from post-MI and control myocytes, which may be very different from studies using high doses of 4-aminopyridine (4-AP) to prolong APs. In addition, our experiments were conducted using high intracellular K⁺ solutions while those of Bouchard et al. (1995) used Cs⁺ to replace K⁺. Previous reports have consistently shown significant changes in E-C coupling when K⁺-based intracellular solutions are replaced with Cs⁺ (Levi et al. 1996; Wasserstrom & Vites, 1996; Litwin & Bridge, 1997), apparently as a result of the inhibitory action of Cs⁺ on both SR Ca²⁺ release (Litwin & Bridge, 1997) and the contribution of the Na⁺-Ca²⁺ exchanger to the Ca²⁺ release process (Levi et al. 1996; Wasserstrom & Vites, 1996). In concordance with these studies we found that the enhancement of peak [Ca²⁺]_i by long post-MI APs relative to short control APs was reduced from 1.7-fold in the presence of high K⁺ to only 1.4-fold in the presence of Cs⁺. Regardless, recording in the presence of intracellular K⁺ is unquestionably more physiologically relevant when studying the mechanism by which changes in AP profile affect intracellular Ca²⁺ and inotropy in disease (Isenberg & Han, 1994; Levi et al. 1996; Wasserstrom & Vites, 1996).

It is clear that reductions in I_to and resultant prolongation of the rat cardiac AP cause increased Ca²⁺ release from the SR, independent of the SR Ca²⁺ load. It seems reasonable to postulate that the differences in the release of Ca²⁺ between long post-MI and short control APs are a consequence, at least in part, of differences in Ca²⁺ influx ‘trigger’ (Fig. 6B) (Bassani et al. 1995; Lopez-Lopez et al. 1995; Santana et al. 1996). The Ca²⁺ current from the control AP has a large amplitude but a very short duration compared with post-MI APs, which probably reflects the relative differences in the rate of deactivation, associated with repolarization, versus inactivation of I_Ca,L (Kaprielian et al. 1999). The effects of I_to on the trajectory of early repolarization of the AP and its close temporal association with Ca²⁺ channel gating may account for the ability of changes in I_to to so profoundly modulate Ca²⁺ influx. Indeed, Ca²⁺ release depends steeply on the early portion of the pulse stimulus between 0 and 20 ms, and not on pulses of longer duration (Han et al. 1994; Isenberg & Han, 1994; P. H. Backx & R. Kaprielian, unpublished observations). The differences in integrated L-type Ca²⁺ current between control and post-MI APs may reflect changes in the degree of activation of I_Ca,L. Since the upstroke velocities and peaks of control and post-MI APs are quite similar, the enhanced activation of I_Ca,L by the post-MI AP could result from a slower repolarization rate, thereby allowing more time for Ca²⁺ channel activation. Therefore, the enhanced Ca²⁺ release triggered by post-MI APs may result from the recruitment of more Ca²⁺ release units by a larger, more sustained L-type Ca²⁺ influx. On the other hand, the large but short-lived L-type Ca²⁺ flux occurring during short control APs may reflect the transient opening of smaller numbers of Ca²⁺ channels and thus activate fewer Ca²⁺ release units.

To assess whether the enhanced Ca²⁺ transient amplitudes could be attributed exclusively to larger Ca²⁺ triggers or whether changes in the efficiency of Ca²⁺ release also occur when myocytes are stimulated with a post-MI AP, we calculated the E-C coupling gain. We formulated two definitions of E-C coupling gain, (1) ∫F_rel/∫F_Ca,L and (2) d(CS)/dt/∫F_Ca,L, based in part on previous reports (Wier et al. 1994; Janczewski et al. 1995; Santana et al. 1997; Litwin et al. 1998). Both definitions of E-C coupling gain predicted that the efficiency of Ca²⁺ release increased modestly in comparison to enhancements in either cell shortening or Ca²⁺ transient amplitude when stimulated with long post-MI APs versus short control APs. These findings suggest that it is the increase in trigger Ca²⁺ through L-type Ca²⁺ channels that is primarily responsible for the enhancement in contractility observed with AP prolongation.

Interestingly, the relative increase in E-C coupling gain was larger when cell shortening rate (Fig. 8B) rather than SR Ca²⁺ release (Fig. 8A) was used as a measure of E-C coupling. This finding is consistent with the existence of an inherent amplification factor between intracellular Ca²⁺ and activation of contractile proteins due to the very steep relationship between contractile force and [Ca²⁺]_i arising from the positive cooperativity of Ca²⁺ binding to troponin C (Backx et al. 1995). These results suggest that a distinction should generally be made between contraction and release following myocyte excitation. Alternatively, the difference in gains as assessed by Ca²⁺ release versus cell shortening may also arise from the absence of Na⁺-Ca²⁺ exchange current, I_Na-Ca, in the equation describing F_rel. Under our experimental conditions Na⁺-Ca²⁺ exchange is present but was not included in the estimation of F_rel. Calcium efflux through the exchanger will result in an underestimation of F_rel and this efflux is expected to be greater with long post-MI APs since peak intracellular Ca²⁺ levels are higher.

The mechanism by which a prolonged post-MI AP increases the efficiency of Ca²⁺ release is unclear. One possible explanation may be enhancements of reverse-mode Na⁺-Ca²⁺ exchange since the magnitude and time course of reverse-mode Na⁺-Ca²⁺ exchange depend strongly on the profile of the AP during this time period (Bers, 1992). This possibility is supported by recent studies showing that reverse mode Na⁺-Ca²⁺ exchange activity works synergistically with I_Ca,L in triggering Ca²⁺ release from the SR (Litwin et al. 1998; Cordeiro et al. 2000; Piacentino & Houser, 2000). Studies are currently underway in our laboratory to determine the potential role of the Na⁺-Ca²⁺ exchanger in modulating Ca²⁺ release following changes in action potential profile.

In this study the effects of AP prolongation on E-C coupling following infarction were examined by comparing Ca²⁺ release in healthy rat ventricular myocytes stimulated with APs previously recorded from control and post-MI rat myocytes. The similarity of the changes in intracellular Ca²⁺ measured following AP prolongation in normal myocytes versus post-infarction myocytes substantiates further our previous conclusion that elevated Ca²⁺ transients following infarction are due primarily to AP prolongation (Kaprielian et al. 1999). However, numerous other genetic, biochemical and morphological changes, aside from I_to downregulation, can affect intracellular Ca²⁺ handling, E-C coupling and AP profile in heart disease (Gwathmey et al. 1987; Beuckelmann et al. 1992; Arai et al. 1993; Gomez et al. 1997; Litwin & Bridge, 1997; Balke & Shorofsky, 1998; O'Rourke et al. 1999; Shorofsky et al. 1999). Therefore, the effect of AP prolongation on E-C coupling shown in our studies may be modulated by other changes affecting Ca²⁺ handling in diseased myocytes. Nevertheless, our results show that AP profile-mediated alterations in E-C coupling can significantly influence Ca²⁺ handling in heart disease independent of these other possible contributing effects.

Recently, Shorofsky et al. (1999) and Gomez et al. (1997) published contradictory results regarding changes in E-C coupling in the spontaneous hypertensive rat (SHR). Both studies used square wave voltage clamp waveforms to trigger Ca²⁺ release and thereby ignored the possible influences of AP profile on E-C coupling. Indeed, SHRs have prolonged APs (Brooksby et al. 1993; Cerbai et al. 1994), reductions in I_to (Cerbai et al. 1994) and elevated Ca²⁺ transients similar to those seen following myocardial infarction (Kaprielian et al. 1999). Based on our findings, AP prolongation in other heart disease models should be considered when drawing conclusions about changes in E-C coupling within the context of disease.

Although we studied the effects of AP prolongation following myocardial infarction on E-C coupling in the rat, similar changes in AP profile may also affect E-C coupling in failing human myocytes. Human epicardial myocytes have been shown to have a relatively high transient outward current density, which is believed to be responsible for their short, notched APs (Nabauer et al. 1996). In human heart failure, a reduction in I_to is associated with a slowed rate of early repolarization and a loss of the AP notch (Beuckelmann et al. 1993; Kaab et al. 1996). Based on our findings, changes in the profile of the human AP as occur in heart failure could conceivably also affect trigger I_Ca,L and E-C coupling efficiency thereby influencing contractility in a manner similar to what we have found in rat.

In summary, our results establish that the positive inotropic effects of prolonged APs as observed following myocardial infarction in rat are mediated primarily through increases in triggered Ca²⁺ release rather than enhanced SR loads. This increase in Ca²⁺ release is due to stronger trigger Ca²⁺ influxes through L-type Ca²⁺ channels coupled with modest enhancements of Ca²⁺ release efficiency. The effect of I_to on excitation- contraction coupling via alterations in AP morphology may represent an important compensatory process by which the myocardium increases contractility following an infarction.