Rate-dependent changes of twitch force duration in rat cardiac trabeculae: a property of the contractile system

Abstract

1. We examined the mechanisms for rate-dependent changes in twitch force duration by simultaneously measuring force and [Ca²⁺]_i in rat cardiac trabeculae.

2. Peak force decreased when the rate of stimulation was increased from 0.2 to 0.5 Hz, whilst it increased from 1 to 2 Hz. Over the same range of frequencies, peak [Ca²⁺]_i transients increased monotonically, whilst both force and [Ca²⁺]_i transient duration were abbreviated.

3. Changes in peak force or peak [Ca²⁺]_i transients were not responsible for the changes in force or [Ca²⁺]_i transient duration.

4. The changes in twitch force and [Ca²⁺]_i transient duration were completed roughly within one beat following an abrupt change in the rate of stimulation.

5. Rate-dependent changes resembled those observed with isoproterenol (isoprenaline) application. However, kinase inhibitors (i.e. K252-a, H-89, KN-62 and KN-93) had no effect on the rate-dependent changes of twitch force and [Ca²⁺]_i transient kinetics, suggesting that protein kinase A (PKA), protein kinase G (PKG) and Ca²⁺-calmodulin-dependent protein kinase II (CaM/kinase II) were not responsible for these kinetic changes.

6. Despite the changes in twitch force and [Ca²⁺]_i transient kinetics, the rate-limiting step for the rate-dependent force relaxation still resides at the level of the contractile proteins.

7. Our results suggest that rate-dependent changes in force and [Ca²⁺]_i transients do not depend on PKA or CaM/kinase II activity but might result from intrinsic features of the contractile and/or Ca²⁺-handling proteins.

Increases in cardiac output following sympathetic stimulation via β-adrenergic receptor (β-AR) activation are associated with elevated heart rate and enhanced myocardial contractility (Armour, 1997). Simultaneous with these changes, there is also a reduction in the duration of systole, which ensures complete relaxation between contractions thus allowing adequate time for diastolic filling (Boudoulas, 1991). In isolated cardiac muscle, activation of β-ARs or elevations in intracellular cAMP level increase peak force and [Ca²⁺]_i transients, and also accelerate their kinetics (Li & Rouleau, 1995; Hussain et al. 1997). Interestingly, elevating the heart rate by pacing, without β-AR activation, mimics many features of β-AR stimulation (Covell et al. 1967). For example, increased stimulation rates in normal myocardium can elevate contractility, producing the Treppe phenomenon or the Bowditch effect (Ross et al. 1995). By contrast, increases in stimulation rate decrease contractility in heart failure patients and animal models of heart disease, which is accompanied by incomplete relaxation and elevated diastolic tone (Pieske et al. 1995; Licata et al. 1997). These disruptions in the force-frequency relationship in diseased hearts have been associated with disrupted excitation-contraction coupling with reduced Ca²⁺ cycling at higher rates of stimulation (Pieske et al. 1995), reduced cAMP levels and reduced sarcoplasmic reticulum (SR) Ca²⁺-ATPase activity suggesting that phospholamban is involved in this phenomenon (Hasenfuss et al. 1996; Maier et al. 1998).

Changes in the duration of contraction with pacing rates have been reported previously in isolated muscle preparations (Schouten, 1990; Borzak et al. 1991; Tang et al. 1996), but few studies have examined the underlying mechanisms for these kinetic changes. Accelerated decline of [Ca²⁺]_i transients in isolated myocytes occurs simultaneously with a reduced duration of myocyte shortening patterns, consistent with relaxation of the cell shortening pattern being controlled by the [Ca²⁺]_i (Capogrossi et al. 1991; Bassani et al. 1994; Hussain et al. 1997). However, the rate-limiting step of twitch force and pressure relaxation in load-bearing cardiac muscle occurs at the level of the contractile proteins (Kurihara & Allen, 1982; Burkhoff, 1994; Backx et al. 1995) suggesting that altered cross-bridge cycling might also contribute (see below). Furthermore, unlike in isolated myocytes, twitch force duration in cardiac muscle is also strongly modulated by the level of force development and sarcomere length in addition to the velocity and extent of shortening, the degree of phosphorylation of the contractile proteins and the [Ca²⁺]_i transient profile (Kurihara & Allen, 1982; Spurgeon et al. 1992; Burkhoff, 1994; Backx et al. 1995).

The time-averaged [Ca²⁺]_i depends strongly on the stimulation rate (Bountra et al. 1988; Kaye et al. 1996), therefore Ca²⁺-dependent kinases and cell signalling pathways appear to be logical candidates for producing the frequency-dependent kinetic alterations (Xu et al. 1997). Accordingly, in isolated rat ventricular myocytes, inhibitors of both the Ca²⁺-calmodulin-dependent protein kinase II (CaM/kinase II) and PKA have been reported to reduce peak shortening and [Ca²⁺]_i transient amplitude as well as their relaxation rates (Bassani et al. 1995; Li et al. 1997), suggesting that phosphorylation of target proteins by CaM/kinase II is responsible for this stimulation-dependent acceleration of the kinetics. However, contradictory studies have also been reported. In rat myocytes, the calmodulin inhibitor W7 reduced peak shortening and [Ca²⁺]_i transient amplitude, but had minor effects on their kinetics (Frampton & Orchard, 1992). Furthermore, neither a chemical inhibitor of CaM/kinase II (KN-63) nor a peptide inhibitor of this enzyme had a significant effect on the rate-dependent abbreviation of the [Ca²⁺]_i transient in rat myocytes (Hussain et al. 1997).

Few studies have previously examined the rate-dependent changes in twitch force kinetics simultaneously with [Ca²⁺]_i transients. This is rather surprising given that defects in rate-dependent changes of peak twitch force development and relaxation occur in heart disease (Schmidt et al. 1995; Pieske et al. 1995; Maier et al. 1998). Indeed, alterations in rate-dependent relaxation properties may be important contributors to the mechanical dysfunction observed in diseased hearts. In this study we examined the rate-dependent changes in twitch force duration of cardiac trabeculae and explored whether these changes were linked to altered kinetics of [Ca²⁺]_i transients. We also investigated the role of Ca²⁺-activated intracellular kinases (e.g. CaM/kinase II, PKA and PKG), sarcomere length and action potential duration in these rate-dependent changes. Increasing stimulation rate abbreviated both twitch force and [Ca²⁺]_i transient duration, a phenomenon which we found to be independent of intracellular kinase activity, sarcomere length and rate-dependent alterations of action potential duration.

METHODS

Muscle isolation and force measurements

Long, thin trabeculae were dissected from the right ventricle of 250–300 g male rats (LBN-F1 strain, Harlan, Indianapolis, IN, USA) as described by Backx et al. (1995). Briefly, rats were anaesthetized by intraperitoneal (i.p) injection of sodium pentabarbital (45 mg kg⁻¹). Hearts were rapidly excised (in less than 10 s), the aorta of the freshly excised heart was cannulated and the heart was arrested by retrograde perfusion with a high-K⁺ modified Krebs-Henseleit (K-H) solution (see below). Dissected trabeculae were mounted between a force transducer (AE801L, SensoNor, Horton, Norway) and a length manipulator in a perfusion bath located on the stage of an inverted microscope (Backx & ter Keurs, 1993). The trabeculae were perfused with the modified K-H solution and electrically stimulated via platinum electrodes running down either side of the bath (A-M systems Inc., Evertt, WA, USA). Twitch force was normalized by the cross-sectional area of the trabeculae. The sarcomere length of the trabeculae was measured by illuminating the central region of the muscle with 683 nm laser light. The diastolic sarcomere length was set to 2.2–2.3 μm unless otherwise stated. All the experiments were performed in accordance with the guidelines of the Animal Care and Use Committee of the Toronto Hospital.

Solutions

Dissection and perfusion solutions consisted of a modified K-H solution containing (mM): 120 NaCl, 5 KCl, 1.2 NaH₂PO₄, 19 NaHCO₃, 1 CaCl₂ and 10 glucose. The dissection solution contained an additional 20 mM KCl. All solutions were equilibrated with 95 % O₂-5 % CO₂ to obtain a pH of 7.4. Isoproterenol (isoprenaline; Sigma Chemical Co.), a non-selective β-AR agonist, was dissolved in deionized water and was added to oxygenated K-H solution to a final concentration of 100 nM. K252-a (a non-selective kinase inhibitor), KN-93 and KN-63 (specific CaM/kinase II inhibitors), H-89 (a specific PKA inhibitor), noradrenaline (norepinephrine; an α₁- and β₁-adrenergic receptor agonist) and phentolamine mesylate (a specific α₁-AR antagonist) were purchased from Research Biochemicals International (Natick, MA, USA). KN-93 was also purchased from Seikagaku America Inc. (MD, USA). K252-a was dissolved in dimethylsulfoxide (DMSO) and added to oxygenated K-H solution to a final concentration of 2.8 μM. KN-93, KN-63 and H-89 were dissolved in de-ionized water and were added to the perfusion solution to final concentrations of 1–5 μM, 5–50 μM and 20 μM, respectively. Phentolamine mesylate and noradrenaline were dissolved in de-ionized water and added to the perfusion solution to a final concentration of 1 μM and 100 nM, respectively. Cyclopiazonic acid (CPA) stock solution was made by dissolving CPA powder in DMSO to 60 mM; 125 μl of this stock solution was then diluted in 75 ml oxygenated K-H solution to a final concentration of 100 μM. All the experiments were performed at room temperature (20–22°C).

[Ca²⁺]_i transient measurement

[Ca²⁺]_i transients were measured using fura-2 salt, which was loaded into trabeculae by a microinjection technique (Backx & ter Keurs, 1993). Briefly, fura-2 salt was electrophoretically injected using a microelectrode containing 1 mM fura-2 potassium salt solution back-filled with 140 mM KCl solution. After the injection period, the preparations were stimulated at 1 Hz for 30–45 min to enhance the rate of spread of fura-2 from the site of injection, thereby producing a uniform distribution of the dye within the muscle. Excitation ultraviolet (UV) light from a 75 W xenon lamp (Oriel Corp, Stratford, CT, USA) was passed through bandpass filters (Omega Optical, Brattleboro, VT, USA) centred at 340 and 380 nm and projected onto the muscle via a × 10 objective (× 10 Fluor, Nikon, Tokyo, Japan) in the inverted microscope (Olympus CK-2). The emitted fluorescent light was collected by the objective and transmitted through a bandpass filter at 510 nm to a photomultiplier (R2368, Hamanatsu, Bridgeport, NJ, USA). The photomultiplier output was filtered at 100 Hz (3 dB), recorded using an A/D data acquisition board (2801A, Data Translation, Marlboro, MA, USA) sampling at 1 kHz, and stored on the computer for later analysis. The preparations were only illuminated with the excitation light for short periods to reduce the effects of photobleaching, which could otherwise interfere with the accurate calibration of the fluorescence signals. In all the preparations studied, the autofluorescence was recorded before loading with fura-2 and again at the end of the experiment after applying 20–50 μM ionomycin and 0.5 mM MnCl₂ in a nominally Ca²⁺-free K-H solution in order to quench the fura-2 fluorescence. Only when the autofluorescence levels at the beginning and the end of the experiment differed by < 15 % were the results of experiments included in the study.

Statistics and curve fits

The parameters over a range of stimulation frequencies and a range of extracellular Ca²⁺ concentrations ([Ca²⁺]_o) were compared by an ANOVA multiple comparison test. Student's paired t test was used to compare the parameters before and after application of a drug. Linear regression analysis was employed to examine the correlation between experimental parameters in this study. Goodness of fit was determined by the correlation coefficient (r). The values given are means ±s.e.m., and statistical significance was determined at P < 0.05.

RESULTS

Rate-dependent alterations in force and [Ca²⁺]_i

Typical force and [Ca²⁺]_i transient traces recorded from the right ventricular trabecula of rat heart at 0.2 and 2 Hz are shown in Fig. 1A. This figure shows that peak twitch force and peak [Ca²⁺]_i transients increased following the elevation of the stimulation frequency while their duration was abbreviated. A summary of the rate-dependent changes in peak force and [Ca²⁺]_i transients from nine experiments is presented in Fig. 1B, which shows a biphasic force-frequency relationship. Peak force decreased from 27.2 ± 2.6 to 23.9 ± 2.8 mN mm⁻² when the frequency of stimulation was increased from 0.2 to 0.5 Hz, whilst it increased to 28.4 ± 2.9 and 33.6 ± 4.8 mN mm⁻² when the frequency was further increased to 1 and 2 Hz, respectively. On the other hand, peak [Ca²⁺]_i transients increased continuously and significantly (P < 0.05) as the frequency of stimulation was increased from 0.2 to 2 Hz (0.54 ± 0.02 to 0.83 ± 0.05 μM; n = 9).

Figure 1. Frequency-dependent changes in amplitude and kinetics of twitch force and [Ca²⁺]_i transients in rat cardiac trabeculae.

A, typical twitch force (smooth line) and [Ca²⁺]_i transient traces (noisy line) from rat cardiac trabecula stimulated at 0.2 Hz (left) and 2 Hz (right). B, peak twitch force (▪, continuous line) and peak [Ca²⁺]_i transient (○, dashed line) are plotted as functions of stimulation rate. The trabeculae were stimulated at four different frequencies: 0.2, 0.5, 1 and 2 Hz. Time-averaged [Ca²⁺]_i (integration of [Ca²⁺]_i transient per interpulse interval, *) is plotted on the same graph as a function of the stimulation rate. The data points representing time-averaged [Ca²⁺]_i were fitted by a linear regression model (dash and dot line; r = 0.99, P < 0.05). C, twitch force duration (▪) and [Ca²⁺]_i transient duration (○) are plotted as functions of the stimulation rate. Data points in B and C represent means ±s.e.m., n = 9.

In this study twitch force duration and [Ca²⁺]_i transient duration were measured as the difference between the times to reach 50 % of the peak magnitude during the rising and declining phases of the twitch. Despite the biphasic change in the peak force with stimulation rate, twitch duration decreased monotonically and significantly (P < 0.05) from 329.1 ± 16.7 ms at 0.2 Hz to 199.8 ± 12.6 ms at 2 Hz while [Ca²⁺]_i transient duration also decreased continuously and significantly (P < 0.05) from 194.9 ± 17.9 ms at 0.2 Hz to 85.9 ± 6.7 ms at 2 Hz (Fig. 1C).

The results in Fig. 1 show that changes in stimulation frequency were accompanied by alterations in the amplitude of twitch force and [Ca²⁺]_i transients which can themselves cause changes in kinetics (Bers & Berlin, 1995; Janssen & Hunter, 1995). We therefore asked whether changes in peak force with frequency could be responsible for the kinetic alterations. Figure 2A shows typical results of experiments in which [Ca²⁺]_o was changed, thereby changing the peak twitch force and [Ca²⁺]_i transient. Increasing the [Ca²⁺]_o from 0.3 to 1 mM resulted in a 2-fold elevation of the peak [Ca²⁺]_i transient and a corresponding enhancement of peak force. Figure 2B shows that twitch duration correlated positively with peak force as [Ca²⁺]_o was increased from 0.3 to 3 mM (r = 0.99, P < 0.05). By contrast, Fig. 2C shows that twitch duration decreased as force increased for stimulation rates above 0.5 Hz, resulting in a negative correlation between these parameters (r = −0.86). These results suggest that changes in peak force cannot be solely responsible for the changes in twitch duration observed when stimulation rate is altered.

Figure 2. Correlation of twitch duration with peak force at different extracellular Ca²⁺ concentrations and different stimulation rates.

A, typical raw traces of twitch force and [Ca²⁺]_i transient at [Ca²⁺]_o of 0.3, 0.5, 1 and 3 mM (from left to right). Peak twitch force (smooth line) and [Ca²⁺]_i transient (noisy line) were elevated as [Ca²⁺]_o was increased. B, twitch duration is plotted as a function of peak force when [Ca²⁺]_o was increased from 0.3 to 0.5, 1 and 3 mM, at 0.5 Hz (▪). The data points were fitted to a linear regression model (dashed line) (r = 0.99, P < 0.05). C, twitch duration is plotted as a function of peak force when the stimulation rate was increased from 0.2 to 0.5, 1 and 2 Hz in the presence of 1 mM extracellular Ca²⁺ (○). The data points were fitted to a linear regression model (dotted line; r = −0.86). Data points represent means ±s.e.m., n = 4.

To further test the possible relationship between twitch duration and peak force we examined the transient response of muscles to abrupt alterations in the stimulation frequency. Five trabeculae were examined for this purpose. Figure 3A and B shows typical force amplitude and kinetic responses to abrupt changes in frequency from 0.2 to 2 Hz and back to 0.2 Hz at a sarcomere length of 2.2 μm. It is clear in Fig. 3A that peak force decreased rapidly following a change in stimulation rate from 0.2 to 2 Hz. Over the next 10–15 beats the force rose to a new steady-state level above that observed at 0.2 Hz stimulation. Unlike peak force, twitch duration changes were largely completed after one to two beats following abrupt changes in stimulation rate (Fig. 3B). The lack of correlation between twitch duration and peak twitch force following changes in stimulation frequency is illustrated in Fig. 3C in which the twitch duration is plotted as a function of peak force for sequential beats (indicated by continuous arrows). These results further support the conclusion that factors other than peak force contribute to changes in twitch force kinetics.

Figure 3. Transient response of rat cardiac trabecula to an abrupt change in the frequency of stimulation at long and short sarcomere lengths.

A, peak twitch force of sequential beats (□) following an abrupt increase (from 0.2 to 2 Hz) and decrease (from 2 to 0.2 Hz) in stimulation rate at a sarcomere length of 2.2 μm is plotted as a function of time. B, twitch duration (○) of corresponding beats (as in A) is plotted as a function of time following an abrupt increase and decrease in the stimulation rate (from 0.2 to 2 Hz and back to 0.2 Hz). The vertical dotted lines indicate where the stimulation frequency was changed. C, duration of every twitch is plotted as a function of the corresponding peak force at sarcomere lengths of 2.2 μm (▵) and 1.9 μm (⋄) following an increase in the stimulation rate (0.2 to 2 Hz). The continuous arrows show the direction of temporal changes in twitch duration and peak force following the change in the stimulation rate. The dashed arrows show where the frequency of stimulation was increased from 0.2 to 2 Hz.

Force of contraction is also strongly influenced by sarcomere length (Schouten, 1990). Decreasing the sarcomere length results in decreased twitch force and twitch duration (Kentish & Wrzosek, 1998; Backx & ter Keurs, 1993). Therefore, we compared the pattern of rate-dependent changes in peak force and twitch duration at a sarcomere length of 1.9 μm to that at 2.2 μm (Fig. 3C). At a sarcomere length of 1.9 μm, peak force was significantly smaller (8.5 ± 1.2 mN mm⁻² compared to 21.8 ± 2.6 mN mm⁻²) and twitch duration was significantly shorter (191.6 ± 21.3 ms compared to 296.7 ± 19.4 ms, n = 4) compared to that at a sarcomere length of 2.2 μm. Figure 3C demonstrates that despite changes in sarcomere length, the twitch duration-peak force relationship was similar for muscles held at 1.9 versus 2.2 μm when the stimulation frequency changed from 0.2 to 2 Hz, suggesting that rate-dependent changes in twitch kinetics are not affected by sarcomere length.

β-AR and phosphorylation mechanisms

The alterations in twitch force and [Ca²⁺]_i transients with increasing stimulation frequencies resemble the changes observed with β-adrenergic receptor stimulation. Indeed, Fig. 4A shows that, at 0.2 Hz, 100 nM isoproterenol significantly increased the peak twitch force of contraction (from 24.6 ± 2.6 to 35.2 ± 0.8 mN mm⁻²) and the peak [Ca²⁺]_i transient (from 0.54 ± 0.02 to 0.79 ± 0.12 μM), while decreasing twitch duration (from 286.7 ± 19.2 to 203.0 ± 13.9 ms) and [Ca²⁺]_i transient duration (from 204.0 ± 21.9 to 121.7 ± 8.2 ms) (n = 7, P < 0.05; Fig. 4). Figure 4B shows superimposed twitch force (left) and [Ca²⁺]_i transients (right) for the same trabecula illustrated in Fig. 4A, stimulated at 0.2 Hz and 2 Hz in the absence of isoproterenol, and at 0.2 Hz in the presence of 100 nM isoproterenol. It is clear that the twitch force and [Ca²⁺]_i transients recorded with isoproterenol closely resemble those recorded without drug at 2 Hz stimulation suggesting possible shared pathways responsible for these kinetic changes. The possibility of common pathways is supported by two additional observations. First, the declining phases of the [Ca²⁺]_i transients measured while the muscle was stimulated at 2 Hz or when stimulated at 0.2 Hz in the presence of isoproterenol cross over the [Ca²⁺]_i transient traces recorded at 0.2 Hz without drug. This cross-over unquestionably establishes that the rate of removal of Ca²⁺ from the cytosol is increased by isoproterenol and at 2 Hz stimulation in comparison to the slower rate of stimulation without drug. Second, the twitch duration-frequency and the [Ca²⁺]_i transient duration-frequency curves in the presence and absence of isoproterenol (Fig. 4C) appeared to converge at higher frequencies of stimulation.

Figure 4. Effects of β-AR activation and increased frequency of stimulation on twitch force and [Ca²⁺]_i transient.

A, typical raw traces of twitch force (smooth line) and [Ca²⁺]_i transient (noisy line) from a rat cardiac trabecula stimulated at 0.2 Hz under control conditions in the absence (left) and presence of 100 nM isoproterenol (right). B, superimposed twitch force traces (left) and [Ca²⁺]_i transient traces (right) under control conditions at 0.2 Hz (continuous line) and 2 Hz (dashed line), and with 100 nM isoproterenol at 0.2 Hz (dash and dot line). C, twitch duration (left) and [Ca²⁺]_i transient duration (right) at different frequencies of stimulation in the absence (dashed line; control data from Fig. 1C) and presence (▪) of 100 nM isoproterenol. Data points represent means ±s.e.m., n = 7.

Ca²⁺-activated intracellular kinases

The results above suggested to us a possible role for cellular kinases in the frequency-dependent changes in twitch and [Ca²⁺]_i kinetics. One possible mechanism for kinase modulation is via changes in the time-averaged [Ca²⁺]_i with stimulation rate which could then activate various cellular kinases (Kaye et al. 1996). As shown in Fig. 1B, the time-averaged [Ca²⁺]_i did, indeed, correlate positively with stimulation rate (from 0.094 ± 0.023 μM at 0.2 Hz to 0.250 ± 0.030 μM at 2 Hz; r = 0.98, P < 0.05).

To examine the possible role of intracellular kinases we used K252-a at a concentration that inhibits PKA, PKC, PKG and CaM/kinase II (information from Calbiochem Biochemicals catalogue, 1999). Inhibition of PKA by 2.8 μM K252-a was confirmed by examining the effects of 100 nM noradrenaline on the twitch force and [Ca²⁺]_i transient magnitude and kinetics (Fig. 5). Figure 5A shows that twitch force amplitude increased significantly following β-AR stimulation under control conditions (left) but not in the presence of K252-a (right). Noradrenaline-mediated α₁-AR activation was blocked in these experiments by 1 μM phentolamine mesylate, a specific α₁-AR antagonist. Raw traces of force and [Ca²⁺]_i transient under control conditions (left) and in the presence of K252-a (right) are shown before and after β-AR stimulation in Fig. 5B and C, respectively. Application of 100 nM noradrenaline increased peak force and [Ca²⁺]_i transient and abbreviated their kinetics under control conditions (Fig. 5A–C, left panels) but not, as expected, in the presence of 2.8 μM K252-a (Fig. 5A–C, right panels). Despite these effects of K252-a, no alterations were observed in twitch duration (275.5 ± 25.5 ms compared to 301.0 ± 66.0 ms) or [Ca²⁺]_i transient duration (191.6 ± 29.5 ms compared to 215.1 ± 15.3 ms) at 0.2 Hz (Fig. 5D). Following an increase in the stimulation rate in the presence of K252-a (from 0.2 to 2 Hz), twitch duration and [Ca²⁺]_i transient duration decreased significantly (to 197.5 ± 10.5 and 78.3 ± 12.8 ms, respectively; P < 0.05, n = 5), much like the effect observed under control conditions (Fig. 5D).

Figure 5. Inhibition of intracellular kinases by K252-a.

A, left, force traces recorded by a chart recorder showing an inotropic response following β-AR stimulation by 100 nM noradrenaline (NA) in the presence of 1 μM phentolamine mesylate (PhM) under control conditions. The lack of such an inotropic response is shown in the presence of 2.8 μM K252-a (right). B, typical raw traces of twitch force (smooth line) and [Ca²⁺]_i transient (noisy line) in the absence (left) and presence (right) of K252-a. C, raw traces of twitch force and [Ca²⁺]_i transient following β-AR stimulation in the absence (left) and presence (right) of K252-a. D, left, twitch duration before (▪) and after (○) kinase inhibition as the frequency of stimulation was increased from 0.2 to 2 Hz. Right, frequency-dependent changes in [Ca²⁺]_i transient duration before (▪) and after (○) kinase inhibition. Symbols represent means ±s.e.m., n = 5.

The fact that the concentration of K252-a used in the experiments described above was sufficient to inhibit not only PKA, but also PKC, PKG and CaM/kinase II suggests that these kinases are not involved in the kinetic changes of force and [Ca²⁺]_i transients observed when stimulation frequency is changed. This conclusion is supported by additional studies using inhibitors at sufficient concentrations to specifically inhibit PKA (20 μM H-89), CaM/kinase II (1–5 μM KN-93 and 5–50 μM KN-62) and PKG (via blockade of nitric oxide production; 100 μM L-NOARG) (data not shown). None of these inhibitors had significant effects on the rate-dependent changes in twitch or [Ca²⁺]_i transient kinetics. When the frequency of stimulation was increased from 0.2 to 2 Hz, twitch duration decreased by 35, 39 and 30 % despite inhibition of PKA, CaM/kinase II and PKG, respectively (compared to 40 % decrease under control conditions), and [Ca²⁺]_i transient duration decreased by 57, 49 and 45 %, respectively (compared to 50 % decrease under control conditions). We also found that application of 10 μM propranolol did not affect the rate dependence of the contraction kinetics. Taken together, our results suggest that these kinases do not contribute measurably to the kinetic alterations associated with stimulation rate changes, although other kinases not affected by these inhibitors might obviously play a role.

Possible role of the contractile proteins in the relaxation phase of a twitch force

The time course of the force of contraction is determined by the interaction of intracellular Ca²⁺ handling with contractile proteins. To evaluate the relative role of contractile proteins versus[Ca²⁺]_i in modulating twitch duration following changes in stimulation frequency we used phase plots (i.e. force-[Ca²⁺]_i relationship during twitches) (Backx et al. 1995) at different stimulation rates. Figure 6 shows phase plots when a muscle was stimulated at 0.2 Hz (Fig. 6A) and 2 Hz (Fig. 6B). The short arrows show the direction of temporal changes during the twitch. The steady-state force-[Ca²⁺]_i relationship was measured later in the same muscle in the presence of 100 μM CPA, a potent inhibitor of the SR Ca²⁺-ATPase pump (Backx et al. 1995). At both frequencies of stimulation, the force-[Ca²⁺]_i relationships during the relaxation phase were located to the left of the steady-state relationship indicating that the intrinsic rate of relaxation of contractile proteins, and not the rate of Ca²⁺ removal from the cytosol, controls the rate of force relaxation despite marked differences in twitch kinetics (Gao et al. 1995). The effect of stimulation rate on the relationship between force and [Ca²⁺]_i during twitch relaxation was quantified by measuring the difference in [Ca²⁺]_i (Δ[Ca²⁺]_i) between phase plots and the steady-state curves at the point where force had relaxed to 50 % of its peak value. Figure 6C shows that Δ[Ca²⁺]_i was increased slightly, but not significantly, at 2 Hz compared to 0.2 Hz (0.32 ± 0.02 μM compared to 0.26 ± 0.02 μM, respectively; n = 5). This leftward shift of the relaxation phase indicates that contractile protein kinetics tend to become even more rate limiting relative to [Ca²⁺]_i kinetics at the high stimulation rates and suggests that the [Ca²⁺]_i kinetics are affected to a greater extent by changes in stimulation rate than the contractile proteins.

Figure 6. Role of the contractile proteins in the relaxation phase of twitch force.

The loops in A and B represent normalized twitch force (F/F_max) as a function of [Ca²⁺]_i transient during a cardiac cycle under control conditions ([Ca²⁺]_o, 1 mM; sarcomere length, 2.2 μm) at 0.2 and 2 Hz, respectively. The dotted lines represent the steady-state force-[Ca²⁺]_i relationship measured in the presence of 100 μM CPA in the same muscle. The short arrows show the direction of twitch force development during one cardiac cycle. F_max is the force developed by the trabecula tetanized in the presence of 100 μM CPA and 8 mM external Ca²⁺ at 10 Hz. Δ[Ca²⁺]_i was measured as the difference between the [Ca²⁺]_i quantified at 50 % force relaxation and the [Ca²⁺]_i at steady state for the same force. C, mean value of Δ[Ca²⁺]_i at 0.2 Hz (□) compared to that at 2 Hz (). Data are means ±s.e.m. (not significant, n = 5).

DISCUSSION

In this study we examined the underlying mechanism for twitch duration shortening following increases of stimulation rate in thin rat cardiac trabeculae. The use of thin trabeculae allowed dynamic measurement of sarcomere length and [Ca²⁺]_i, while minimizing diffusion distances which have previously been shown to influence muscle properties, especially at high stimulation rates (Schouten & ter Keurs, 1986). For the majority of our studies, the diastolic sarcomere length was fixed at 2.2–2.3 μm. Ensuring that the sarcomere length was set at pre-defined values was critical for these studies since twitch amplitude and duration depend strongly on sarcomere length (Allen & Kentish, 1985; Backx & ter Keurs, 1993; Janssen & Hunter, 1995).

[Ca²⁺]_i transient amplitude was found to increase monotonically with stimulation frequency as reported previously (Capogrossi et al. 1991). This could reflect augmented Ca²⁺ loading into the SR resulting from increased time-averaged Ca²⁺ entry through L-type Ca²⁺ channels and possibly reduced net Ca²⁺ extrusion by Na⁺-Ca²⁺ exchange following elevations in stimulation rate (Capogrossi et al. 1991). Increased stimulation frequencies can also elevate [Na⁺]_i (Bountra et al. 1988) thereby loading Ca²⁺ into the SR via reduced forward-mode or enhanced reverse-mode Na⁺-Ca²⁺ exchange activity (Bers, 1992). In contrast to [Ca²⁺]_i transients, twitch force changed in a biphasic manner with stimulation frequency, being negative from 0.2 to 0.5 Hz and positive from 0.5 to 2 Hz (Fig. 1B), consistent with previous observations in rat trabeculae (Tang et al. 1996) and isolated myocytes (Borzak et al. 1991). These results establish that the peak twitch force and peak [Ca²⁺]_i transient amplitudes are modulated differently by the stimulation rate. This lack of concordance between peak force and [Ca²⁺]_i transients is not unexpected since a kinetic delay exists between these parameters (Kurihara & Allen, 1982; Burkhoff, 1994; Backx et al. 1995), and, as discussed more fully below, the kinetics of both force and [Ca²⁺]_i transients are simultaneously affected by stimulation rate.

Abbreviation of [Ca²⁺]_i transient duration with increased stimulation frequencies arises almost exclusively from an accelerated rate of the [Ca²⁺]_i declining phase. In fact, time-to-peak [Ca²⁺]_i did not decrease significantly at 2 Hz compared to 0.2 Hz (60.9 ± 6.0 ms compared to 46.5 ± 5.5 ms, respectively). These changes could simply result from the increases in [Ca²⁺]_i transient amplitude since the SR Ca²⁺-ATPase activity depends on the [Ca²⁺]_i (Bers & Berlin, 1995). However, the results in Fig. 4B (right panel) demonstrate that other factors must also contribute to these changes. In particular, the slope of the declining phase of the [Ca²⁺]_i relaxation (d[Ca²⁺]_i/dt) was greater at 2 versus 0.2 Hz at all [Ca²⁺]_i, causing the [Ca²⁺]_i transients to cross over, thereby establishing that the intrinsic rates of Ca²⁺ removal differed at higher frequencies of stimulation. The abbreviation of [Ca²⁺]_i transient duration could originate from a number of possible mechanisms. For example, changes in stimulation rate can influence action potential duration which, in turn, can affect the dynamics of Ca²⁺ cycling (Clark et al. 1996). However, the action potential duration measured in single isolated rat ventricular myocytes actually increased slightly with elevated stimulation rates (the time to 50 % repolarization was 10.8 ± 1.5 ms at 0.2 Hz versus 15.6 ± 3.6 ms at 2 Hz, data not shown) consistent with previous voltage-clamp studies (Hussain et al. 1997).

Sarcomere length and force amplitude have been shown to be major determinants of twitch force duration (Janssen & Hunter, 1995). However, we found that the twitch duration-peak force relationship was strongly influenced by stimulation rate. For example, at different [Ca²⁺]_o, twitch force duration and force amplitude were positively correlated, yet plots at different frequencies of stimulation showed a negative correlation between these parameters (Fig. 2). Similar conclusions were made when twitch duration and peak force were varied by changing the sarcomere length (Fig. 3). These results show that the absolute relationship between twitch duration and peak force is strongly influenced by stimulation rate.

Role of intracellular kinases

The rate-dependent alterations in the peak values and kinetics of twitch force and [Ca²⁺]_i transients were remarkably similar to changes observed with β-AR stimulation. This observation raised the intriguing possibility that, under physiological conditions, the enhanced relaxation with increased heart rate following sympathetic activation of the sino-atrial node utilizes an intrinsic myocardial mechanism. Furthermore, in diseased myocardium the normal rate-dependent modulation of contractile performance is impaired (Pieske et al. 1995; Licata et al. 1997). Previous studies have linked the delayed relaxation and disrupted diastolic function in diseased heart to downregulation of the adrenergic receptors and cAMP-dependent pathways (Schmidt et al. 1995; Mulieri et al. 1997) but modification of this intrinsic system might also occur. We initially postulated that rate-dependent abbreviation of twitch force and [Ca²⁺]_i transient duration involved various intracellular kinases. However, blockade of PKA, PKC, PKG and CaM/kinase II, either collectively with K252-a (Fig. 5) or individually with H-89 (PKA inhibitor), KN-93 and KN-62 (CaM/kinase II inhibitors), and L-NOARG (a nitric oxide synthase inhibitor), did not affect the rate-dependent changes in twitch kinetics. These observations are consistent with previous results in guinea-pig cardiomyocytes (Money-Kyrle et al. 1998) using PKA inhibitors and in rat myocytes using CaM/kinase inhibitors (Frampton & Orchard, 1992; Hussain et al. 1997). Our findings appear to contradict those of Bassani et al. (1995) and Li et al. (1997) who reported that rate-dependent acceleration of twitch duration in rat cardiac myocytes is mediated by CaM/kinase II. These discrepancies might be related to differences in the experimental preparations or protocols used. In studies done by Bassani et al. (1995) and Li et al. (1997) the rate of relaxation of the steady-state twitches was compared to that of post-rest twitches following a rest period, which is not equivalent to examining the steady-state properties at a number of different frequencies (as in the present study). In addition, cell shortening in single myocytes might not be equivalent to the force of contraction in load-bearing preparations, which could be argued to be a more physiologically relevant preparation.

Previous studies suggested that nitric oxide (NO) production and subsequent activation of PKG could play a role in the force-frequency relationship (Kaye et al. 1995). In our study, the application of K252-a and L-NOARG did not affect the frequency-dependent changes in twitch and [Ca²⁺]_i transient kinetics (data not shown). These observations establish that neither NO nor PKG activation underlies the frequency-dependent twitch shortening in rat cardiac trabeculae, consistent with previous reports in paced adult rat ventricular myocytes (Balligand et al. 1993) and in rat papillary muscles (Weyrich et al. 1994).

Alterations in twitch duration following abrupt changes in stimulation frequency occurred almost instantaneously, as observed previously (Schouten, 1990). This rapid onset of twitch duration changes might further argue against the role of kinases. By contrast, peak twitch force required 10–15 beats before reaching steady state following abrupt changes in stimulation rate, as predicted by the two-compartment model of the SR used previously to predict mechanical restitution curves (Schouten et al. 1987). As a result, no unique temporal relationship between twitch duration and peak twitch force exists during the transient response of rat trabeculae to changes in stimulation rate as illustrated in Fig. 3C.

Role of the contractile proteins

The duration of twitch force and [Ca²⁺]_i transients decreased inversely with the stimulation rate between 0.2 and 2 Hz. The reduced twitch force duration with increasing frequency could be primarily determined by enhanced rates of Ca²⁺ transient decline as described in isolated myocytes (Spurgeon et al. 1992). Phase plot analysis demonstrated that the rate-limiting step for force relaxation resided at the level of the contractile proteins at all the stimulation rates studied (Fig. 6). The kinetic interaction between force and [Ca²⁺]_i during a twitch is complex, therefore making definitive connections between changes in force and [Ca²⁺]_i transient kinetics difficult. We did observe that the force-[Ca²⁺]_i relationship during the relaxation phase of the cardiac cycle measured at 2 Hz, while being similar to that at 0.2 Hz, was shifted leftward (presented as a larger Δ[Ca²⁺]_i at 2 Hz, Fig. 6C). These results are consistent with a differentially larger effect of stimulation rate on [Ca²⁺]_i cycling compared to contractile protein kinetics. Indeed, [Ca²⁺]_i transient duration decreased nearly 2-fold versus a more modest 1.5-fold reduction in twitch force duration when the stimulation frequency was increased from 0.2 to 2 Hz. It is conceivable, therefore, that the frequency-dependent abbreviation of force duration is controlled primarily by changes in [Ca²⁺]_i transient kinetics although modulation of contractile protein kinetic properties might also contribute. Regardless, [Ca²⁺]_i transient abbreviation appears to be independent of kinases in cardiac muscle that are known to affect function. One alternative mechanism to explain our observations is the possibility of direct effects of Ca²⁺ on the SR Ca²⁺ pump (SERCA2a) activity. Elevated [Ca²⁺]_i has been shown to enhance phospholamban (PLB)-SERCA2a heterodimer dissociation, thereby increasing SR Ca²⁺ pump activity (Kimura et al. 1997). Indeed, if the degree of association between PLB and SERCA2a depends on the time-averaged [Ca²⁺]_i (Fig. 1), then the SR Ca²⁺ pump activity would be predicted to increase as observed in our studies. This mechanism might also explain the rapid time course for the observed changes in twitch force duration with stimulation rate.

Summary

In this study we examined the factors that could underlie the rate-dependent changes in twitch force and [Ca²⁺]_i transient kinetics in rat cardiac trabeculae. We found that these changes do not appear to depend on the activation of intracellular kinases or on the frequency-dependent alterations in action potential duration, but are an intrinsic property of rat myocardium possibly involving direct modulation of PLB or PLB-SERCA2a interaction by Ca²⁺. It is clear that the kinetic changes of twitch force and [Ca²⁺]_i transients occur simultaneously and in a manner such that the contractile proteins still control relaxation. These intrinsic properties of cardiac muscle might play an important role physiologically and alterations in these properties may be of fundamental importance in the disrupted relaxation observed in heart disease.