Intracellular Ca2+ Increases the Mitochondrial NADH Concentration During Elevated Work in Intact Cardiac Muscle
It is not clear how mitochondrial energy production is regulated in intact tissue when energy consumption suddenly changes. Whereas mitochondrial [NADH] ([NADH]m) may regulate cellular respiration rate and energetic state, it is not clear how [NADH]m itself is controlled during increased work in vivo. We have varied work and [Ca2+] in intact cardiac muscle while assessing [NADH]m using fluorescence spectroscopy. When increased work was accompanied by increasing average [Ca2+]c (by increasing [Ca2+]o or pacing frequency), [NADH]m initially fell and subsequently recovered to a new steady state level. Upon reduction of work, [NADH]m overshot and then returned to control levels. In contrast, when work was increased without increasing average [Ca2+]c (by increasing sarcomere length), [NADH]m fell similarly, but no recovery or overshoot was observed. This Ca2+-dependent recovery and overshoot may be attributed to Ca2+-dependent stimulation of mitochondrial dehydrogenases. We conclude that the immediate initial increase in respiration rate upon elevation of work is not activated by increased [NADH]m (since [NADH]m rapidly fell) or by [Ca2+]c (since work could also be increased at constant [Ca2+]c). However, during sustained high work, a Ca2+-dependent mechanism causes slow recovery of [NADH]m toward control values. This demonstrates a Ca2+-dependent feed-forward control mechanism of cellular energetics in cardiac muscle during increased work.
The regulation of oxidative phosphorylation has been the subject of extensive research.1 2 By use of isolated mitochondria, external conditions can easily be manipulated to study the effects of various parameters on the oxidative phosphorylation rate.1 A disadvantage with this approach, however, is that it is not possible to accurately simulate in vivo conditions. In particular, during dynamic changes of the ATP consumption rate (eg, during increasing work), intracellular conditions are not at steady state, and these time-dependent changes may be critical for energetic regulation. To circumvent this problem, we have used intact cardiac muscle to evaluate the dynamic effects of changing work on the mitochondrial energy state during conditions of nonlimiting O2 supply.
Although various parameters have been proposed to control flux through the respiratory chain, matching ATP production to energy demand, no general model has been agreed on.1 2 Nevertheless, most investigators believe that [O2] is not limiting in vivo,1 and in this case, most current theories and experimental results indicate that respiration rate (and therefore ATP production rate) depends on three cytosolic parameters (see Fig 1⇓), [ADP], [ATP], and [Pi]3 (or combinations such as [ATP]/[ADP][Pi]4 or [ATP]/[ADP]5 ), and on [NADH]m2 3 6 (or [NADH]m/[NAD+]m4 ). In particular, during conditions of nonlimiting [ADP], it has been suggested that [NADH]m may exert strong control of the oxidative phosphorylation rate and cellular energetic state.2 3 7
During increased work, either increased [ADP]/[ATP], [NAD+]m/[NADH]m, or [Ca2+]m may cause increased NADH production rate8 and, consequently, [NADH]m (provided that the [NADH]m consumption rate is not correspondingly increased; see Fig 1⇑). [Ca2+]m has been implicated for two reasons. First, increases in [Ca2+]c often accompany increased energy consumption (eg, during muscle contraction, hormone secretion, or neuronal activity), and increased [Ca2+]c has been shown to increase [Ca2+]m.9 Second, there is clear evidence from isolated mitochondria and isolated enzymes that Ca2+ can stimulate three key mitochondrial enzymes: pyruvate dehydrogenase, NAD-linked isocitrate dehydrogenase, and 2-oxoglutarate dehydrogenase.8 10 11 Thus, as outlined in Fig 1⇑, when work is associated with increased [Ca2+]c and, subsequently, [Ca2+]m, NADH production may be stimulated.
We have previously shown that with nonlimiting O2 supply in intact heart muscle, increased work caused decreased [NADH]m and that prolonged stimulation caused a partial recovery of [NADH]m.12 However, we had no evidence regarding the mechanism of this recovery. We hypothesized that the recovery may have been caused by a work-dependent increase of [ADP]/ATP], [NAD+]m/[NADH]m, or [Ca2+]m. Furthermore, during increased work in vivo, there is currently no direct evidence demonstrating a Ca2+-dependent stimulation of [NADH]m or the role of Ca2+ in regulation of the oxidative phosphorylation rate. The main goal of the present study was therefore to explicitly test, for the first time, whether the recovery of [NADH]m during sustained increases in work depended on Ca2+. This was examined by increasing work by Ca2+-dependent and Ca2+-independent means.
Materials and Methods
Trabeculae Preparation and Solutions
Thin multicellular trabeculae (≈200-μm diameter) were isolated from rat right ventricle as described previously.12 Briefly, brown male LBN-F1 rats (370 to 510 g) were deeply anesthetized and anticoagulated by injecting 65 mg pentobarbital and 1000 U heparin intraperitoneally. The hearts were excised and then perfused at ≈28°C using the Langendorff method before removing the trabeculae. The perfusion solution contained (mmol/L) NaCl 108, KCl 21, MgCl2 1.2, CaCl2 0.5, NaHCO3 24, glucose 4, and sodium pyruvate 10, along with insulin (20 U/L), and was equilibrated with a 95% O2/5% CO2 gas mixture to produce pH 7.40 (at 24°C).
After dissection, the muscle was mounted in a muscle chamber, paced at 0.5 Hz, and superfused at 15 mL/min with the solution described above, except that 6 mmol/L KCl, instead of 21 mmol/L, was used and the temperature was 24°C. This solution was used for all experiments except for the [CaCl2], which depended on protocol (see below). Muscle force was measured during isometric contractions, and sarcomere length was varied by adjusting muscle stretch. Sarcomere length was assessed from laser light (Uniphase 102-4) diffraction patterns, created by transmitting the light through the muscle.13
[NADH]m was assessed using methods described previously.12 Briefly, the trabeculae were excited by light at 350 nm, and fluorescence was detected at 385 and 456 nm. The fluorescence signal at 456 nm predominantly arises from mitochondrial NADH14 15 and motion artifact, whereas the reference signal at 385 nm mainly reflects motion artifact.12 We used a previously developed method to eliminate the motion artifact from the fluorescence signal at 456 nm by dividing it with a modified reference signal.12 [NADH]m was calibrated to obtain the NADH ratio (NADH/NAD+) by using sodium cyanide to maximally reduce NAD+ and carbonyl cyanide p-(trifluoromethoxy)phenylhydrazone (FCCP) to maximally oxidize it; see Brandes and Bers12 for further technical details.
Assessment of ATP Hydrolysis Rate and Data Analysis
Time-averaged force was used as an index of work-related ATP hydrolysis rate.12 This indirect and simplified measure assumes that the ATP hydrolysis rate is related to active force during the whole (isometric) contraction cycle.16 A 0.2-Hz low-pass filter was applied to the force signal to obtain average force by using software digital filtering (MicroCal Software Inc). The signal-to-noise ratio of the fluorescence ratio was similarly improved by digital low-pass filtering. Results were reported as mean±SE. Statistical analysis was performed using Student's t test (paired where applicable), and differences were considered significant at P<.05.
To separate effects of Ca2+-dependent and Ca2+-independent mechanisms regulating [NADH]m during altered work, three different types of protocol were used. In the first, various standard perfusion solutions with increasing [Ca2+]o (0.3, 0.6, and 2.0 mmol/L) were used at constant pacing frequency (1 Hz) and maximum sarcomere length. This is expected to increase the intracellular Ca2+ transient amplitude and time-averaged [Ca2+]c, as well as time-averaged force (work). In the second protocol, the pacing frequency was varied from 0.25 to 3 Hz at constant [Ca2+]o (2 mmol/L) and maximum sarcomere length. This is also expected to increase the time-averaged [Ca2+]c and work, although the individual Ca2+ transients may be less affected.17 In the third protocol, sarcomere length (and consequently myofilament overlap) was varied from a length (at peak of contraction) of ≈1.7 μm to maximum (≈2.1 to 2.2 μm) at constant pacing frequency (1 Hz) and [Ca2+]o (2 mmol/L). This is expected to increase work without significantly altering the Ca2+ transient amplitudes.13
Ca2+-Dependent Changes of Work
Increasing [Ca2+]o is a direct way to increase both [Ca2+]c transient amplitudes and work (energy consumption rate)* in cardiac muscle.18 19 20 Fig 2A⇓ indeed shows that when [Ca2+]o was increased from 0.3 mmol/L to either 0.6 or 2 mmol/L at a constant stimulation frequency (1 Hz), the amplitude of the muscle force transients increased substantially. Because work was varied in different ways in the present study, time-averaged force (middle trace) was calculated and used as an index of muscle work or energy consumption rate (ATP hydrolysis rate).12 Direct determination of ATP hydrolysis or O2 consumption rates are not practical because of the small muscle size and kinetic resolution required. Furthermore, because the contractions are nearly isometric, the time-varying elastance model21 is not appropriate for estimating O2 consumption. Both Cooper16 and Hasenfuss et al22 have shown that there is a linear relationship between energy consumption and the force-time integral, making time-averaged force a reasonable index of ATP hydrolysis rate and work. Furthermore, increased [Ca2+]o, pacing frequency, and sarcomere length23 are all known to increase the O2 consumption rate, and this was reflected here as increased time-averaged force (see below for frequency and sarcomere length protocols).
Fig 2A⇑ demonstrates that the increase in time-averaged force, or work, caused [NADH]m to rapidly decrease to MIN. However, during sustained work elevation, [NADH]m partially recovered to a new steady state level (SS) slightly below the original control value, creating an apparent undershoot. When [Ca2+]o was reduced back to 0.3 mmol/L, there was a transient overshoot of [NADH]m, causing it to reach MAX before it eventually returned to the original control level. These dynamic changes in [NADH]m were more pronounced when increasing [Ca2+]o from 0.3 to 2 mmol/L rather than 0.6.
Increased pacing frequency is a second direct way to increase both time-averaged [Ca2+]c and work. Fig 2B⇑ shows that in contrast to increased [Ca2+]o, increased pacing frequency did not dramatically alter the amplitudes of individual contractions.17 The slight negative force-frequency relationship typically observed in rat ventricle was also observed here and may be associated with a small reduction in individual Ca2+ transients.17 However, our large increase in the number of Ca2+ transients per unit time (twofold to eightfold) more than offset this effect, such that time-averaged [Ca2+]c and force would be considerably elevated, as shown in Fig 2B⇑. Fig 2B⇑ further demonstrates an undershoot and overshoot of [NADH]m qualitatively similar to that seen for altered [Ca2+]o (Fig 2A⇑). Again, larger increments in work (average force) produced larger initial decreases in [NADH]m (MIN) and larger overshoots (MAX) upon the abrupt reduction in work and ATP consumption rate.
The rapid initial decrease in [NADH]m, upon a sudden increase in work using either protocol, can be explained by a rapid increase of NADH consumption via an increased oxidative phosphorylation rate, which is not matched by a corresponding stimulation of the NADH production rate (via the tricarboxylic acid cycle). That is, decreased [NADH]m is a consequence of stimulation of oxidative phosphorylation by a parameter other than NADH. Furthermore, since [NADH]m is lowered under these conditions, the initial increase in the oxidative phosphorylation rate cannot be controlled by increased [NADH]m (or redox potential).
The recovery of [NADH]m from the initial nadir, however, could reflect an additional mechanism that causes stimulation of NADH production, thereby opposing the initial decline of [NADH]m. If this mechanism was also slow to deactivate, that would explain the overshoot as well (ie, maintained stimulation of NADH production after consumption had declined). We hypothesize that the [NADH]m undershoot and overshoot are a consequence of slowly increasing and decreasing [Ca2+]m, which in turn would be related to the time-averaged value of [Ca2+]c.8 9 To test whether the initial decline and recovery of [NADH]m are Ca2+ dependent, work alone was next altered by a protocol that does not alter time-averaged [Ca2+]c.
Ca2+-Independent Changes of Work
To increase work without changing time-averaged [Ca2+]c, the sarcomere length was abruptly increased, because this has previously been shown to result in unaltered [Ca2+]c.13 24 Fig 2C⇑ shows the effects of increasing work by increasing sarcomere length from a slack control length to various lengths, up to that generating maximum force, using constant frequency (1 Hz) and [Ca2+]o (2 mmol/L). At each length, the force transient amplitudes remained constant, suggesting that the Ca2+-transient amplitudes were also constant. The immediate response to the stretch is similar to that obtained by Allen and Kurihara,24 who reported that force quickly increased but that [Ca2+]c did not change significantly after a quick stretch from slack to optimum (maximum force). However, in contrast to our results, they also observed slowly increasing force and [Ca2+]c after the initial changes. This discrepancy may be explained by their use of a slack force that was only ≈5% of maximum compared with our shortest length, which produced ≈30% of maximum force. Furthermore, the time constant for the slow increase was ≈4 minutes, whereas our measurement time, at each length, was only ≈1 minute.
Fig 2C⇑ demonstrates that the fall of [NADH]m was similar to that observed initially when increased work was associated with increased time-averaged [Ca2+]c (Fig 2A and 2B⇑⇑). However, in this case, in which time-averaged [Ca2+]c did not increase, there was neither recovery of [NADH]m during sustained work nor overshoot of [NADH]m on return to the initial sarcomere length. In Fig 2C⇑, MIN is equal to SS and therefore corresponds to the MIN values obtained in Fig 2A and 2B⇑⇑. This result supports the hypothesis that the partial recovery of [NADH]m, when work was elevated by increased frequency or [Ca2+]o, is Ca2+ dependent. Furthermore, this result also indicates that feedback by [ADP]/[ATP] or by [NAD+]m/[NADH]m was not responsible for the recovery or overshoot of [NADH]m observed in Fig 2A and 2B⇑⇑.8
Work-Dependent Initial Changes of [NADH]m
Fig 3A⇓ demonstrates that the initial decrease of [NADH]m (MIN) as a function of work was qualitatively similar for all three protocols used to vary work. Thus, both the Ca2+-dependent protocols (changing [Ca2+]o or frequency) and the Ca2+-independent protocol (changing sarcomere length) caused [NADH]m to initially fall with increasing work. The initial decline in [NADH]m, resulting from increased rate of oxidative phosphorylation, was therefore independent of [Ca2+]c but depended on work level. Consequently, we conclude that the oxidative phosphorylation rate is initially not uniquely controlled by [NADH]m3 or [Ca2+]c.2
The Ca2+-independent mechanism of [NADH]m downregulation during the initial stages of increased work is opposed by the Ca2+-dependent mechanism, which caused [NADH]m to recover (and overshoot) when using the frequency or [Ca2+]o protocol (Fig 2A and 2B⇑⇑). Based on Fig 3A⇑, the Ca2+-independent mechanism appears to be faster than the Ca2+-dependent recovery mechanism, because the slopes of the frequency and [Ca2+]o-dependent data were not shallower than the slope of the sarcomere length–dependent data. This is consistent with the results of Fig 2A and 2B⇑⇑ and with the notion that increased [Ca2+]m and activation of dehydrogenases occurs with some time lag after abrupt changes in time-averaged [Ca2+]c.8 9
Work- and Ca2+-Dependent Changes of [NADH]m During Sustained Work
Fig 3B⇑ shows pooled data from experiments in which frequency was altered (as in Fig 2B⇑). It is clear that the Ca2+-dependent recovery mechanism limits the decline of [NADH]m (SS) versus the initial Ca2+-independent fall (MIN). Indeed, at higher work, SS fell only to a level that was within ≈12% of the control value. MAX during the overshoot also increases monotonically with work (measured immediately before the sudden reduction back to control). Pooled data using the protocol in which [Ca2+]o was altered produced qualitatively similar results.
Fig 3C⇑ shows the relationship between the amount of [NADH]m recovery and work for the three protocols. The amount of [NADH]m recovery increases as a function of work when using increasing [Ca2+]o or frequency, whereas increasing sarcomere length shows no recovery of [NADH]m. This once again demonstrates the strict Ca2+ dependence of the [NADH]m recovery during sustained elevated work.
Kinetics of Ca2+-Dependent NADH Recovery
Our results also provide some kinetic information about the Ca2+-dependent activation of NADH production. When work was altered by increasing the pacing frequency from 0.25 to 1 Hz, [NADH]m recovered with τ ≈30 seconds (see Fig 2B⇑). This is similar to the rate of Ca2+ uptake by mitochondria in intact ventricular myocytes when the sarcoplasmic reticulum Ca2+ content was discharged into the cytosol (τ≈41 seconds)25 and the rate of increased [Ca2+]m in intact rat ventricular myocytes in response to increased contraction frequency from 0 to 4 Hz (τ≈40 seconds).9 The overshoot declined with τ≈35 seconds, similar to the kinetics of [Ca2+]m decline or Ca2+ efflux from the mitochondria in intact cells, τ≈21 seconds9 and τ≈45 seconds, respectively.26 Thus, the kinetics of the Ca2+-dependent stimulation of NADH production observed here in intact cardiac muscle likely reflect the kinetics of mitochondrial Ca2+ transport and the time needed for activation or deactivation of the mitochondrial dehydrogenases.
Work and Temperature Regimens
As we have discussed previously,12 a limitation of the results obtained here may be a result of using a nonphysiological temperature (24°C) and low work load. However, Fig 3A⇑ demonstrated that the initial decrease of [NADH]m was related to work over the whole range of work levels studied here, from 25% to 160% of control, approaching the level in vivo.12 This suggests that increasing work further may lower the initial [NADH]m even more.
To evaluate the effects of using a physiological temperature and an even larger range of work levels, the frequency protocol was repeated at 37°C. Fig 4⇓ shows that for the whole range of work levels (>13-fold increase) the results were qualitatively similar to those in Fig 2B⇑ (at 24°C with a 6-fold increase in work levels). This suggests that the phenomena described here are not unique to studies at 24°C but also occur at physiological temperatures and pacing frequencies.
Upon a sudden increase in work and ATP consumption rate, during either increasing [Ca2+]c or constant [Ca2+]c, there was a rapid decrease in [NADH]m. This implies that the initial increase in the oxidative phosphorylation rate is Ca2+ independent (eg, insignificant Ca2+ control of the ATP synthase2 ) and also that it is not controlled by increased [NADH]m. Furthermore, because there was initially a net decrease in [NADH]m, the NADH production rate (via the tricarboxylic acid cycle) is not adequately stimulated to match the respiratory demand. Initial control of the oxidative phosphorylation rate may therefore be controlled by alterations in cytosolic phosphates (eg, [ADP], [Pi], [ATP]/[ADP][Pi], or [ATP]/[ADP]) rather than [NADH]m or [Ca2+]c.
When the increase in work was caused by increased time-averaged [Ca2+]c, [NADH]m slowly recovered. This is most likely a result of slowly increasing [Ca2+]m, which may stimulate mitochondrial dehydrogenases and, consequently, the flux of reducing equivalents through the tricarboxylic acid cycle to raise [NADH]m. The increased [NADH]m may subsequently further stimulate oxidative phosphorylation to better match the rate of ATP production to its consumption. In this scenario, the initially decreasing [NADH]m may correlate with increasing [ADP] and [Pi] (decreasing phosphorylation potential), whereas during [NADH]m recovery, [ADP] and [Pi] may decrease (increasing phosphorylation potential). Such large changes in [Pi] and [ADP] would occur only transiently and might not be observed in a typical (31P NMR) experiment, in which steady state phosphate levels are measured during increased work loads in vivo.3 27
However, using time-resolved 31P NMR, a transient increase in [ADP] was observed in intact perfused hearts within 30 seconds of increased work load by the addition of isoprenaline.28 In a pacing protocol, similar to that used in the present study, Elliott et al29 also demonstrated that [Pi] followed a time dependence that was the mirror image of the time dependence of [NADH]m observed here. Although increased [NADH]m and phosphorylation potential could occur without an increase in O2 consumption rate (or oxidative phosphorylation rate),7 Elliott et al showed that the O2 consumption rate also slowly increased after increased work (pacing frequency). These observations are consistent with the results obtained here and argues for shared control of the oxidative phosphorylation rate by [NADH]m (via [Ca2+]c) and phosphates. Specifically, the recovery of [NADH]m may switch some of the control of oxidative phosphorylation rate away from cytosolic phosphate control over to [NADH]m control.3
It has previously been suggested that oxidative phosphorylation may be exclusively controlled by either ADP (if [ADP] but not [NADH]m is rate limiting) or by NADH (if [NADH]m but not [ADP] is rate limiting).3 Therefore, at low but increasing work loads, [NADH]m may be constant while [ADP] increases and thereby increases the oxidative phosphorylation rate. Conversely, at higher increasing work loads, [ADP] may be constant while [NADH]m increases and thereby increases the oxidative phosphorylation rate. The progressive decline of [NADH]m with work observed in Fig 3A⇑ (and Fig 4⇑), however, suggests that there was no such shift from an ADP- to an NADH-limited domain within the work regimen studied here.
The changes in cytosolic phosphates, driven by changes in [NADH]m, may also have important consequences for regulation of the cytosolic phosphorylation potential. Initially, with increased work, the phosphorylation potential may fall, but this may be opposed by rising [NADH]m, causing the phosphorylation potential to recover.7 This recovery may be necessary for maintaining the cytosolic energetic state needed for processes that hydrolyze ATP (eg, sarcoplasmic reticulum Ca2+ uptake).
Our results are most consistent with the following control mechanisms (see Fig 1⇑): Increased cytosolic [Ca2+] causes increased work and ATP hydrolysis rate (by myofilament and transport ATPases). Increased ATP hydrolysis rate, in turn, may cause increased [ADP] (and/or Pi), which could stimulate the oxidative phosphorylation rate (more than the NADH production rate) and thereby a fall in [NADH]m. On a slower time scale, cytosolic Ca2+ may enter the mitochondria and activate pyruvate dehydrogenase, NAD-linked isocitrate dehydrogenase, or 2-oxoglutarate (α-ketoglutarate) dehydrogenase. (Note that in the present study 10 mmol/L pyruvate was used, and this may limit the amount of Ca2+-dependent activation of pyruvate dehydrogenase.) Activation of the dehydrogenases is expected to increase the NADH production rate and consequently cause [NADH]m to recover. Increased [NADH]m may further increase the oxidative phosphorylation rate and/or cause [ADP] (and the cytosolic phosphorylation potential) to return toward control levels.
Thus, we provide the first clear experimental evidence in intact cardiac muscle to support the idea that [Ca2+]c is an important dynamic control signal for stimulating the reduction of mitochondrial NAD+ to NADH, thereby actively adjusting the cellular energy state during conditions of increased ATP consumption rate. Since an increase in time-averaged [Ca2+]c is a physiological stimulus causing the immediate increase of myofilament ATP consumption rates, the parallel but time-delayed Ca2+-dependent activation of mitochondrial dehydrogenases constitutes part of a feed-forward control of cellular energetic state.
Selected Abbreviations and Acronyms
|τ||=||exponential time constant|
|MAX||=||maximum value of [NADH]m after decreased work|
|MIN||=||minimum value of [NADH]m during increased work|
|[NADH]m, [NAD+]m||=||mitochondrial [NADH] and [NAD+]|
|NMR||=||nuclear magnetic resonance|
|SS||=||steady state [NADH]m during increased work|
This study was supported in part by a Grant-in-Aid from the American Heart Association of Metropolitan Chicago (Dr Brandes) and by National Institute of Health grants HL-30077 and HL-52478 (Dr Bers).
*Although the energy consumption rate is strictly related to work/time (power), the term “work” will be used here rather than work/time.
- Received July 23, 1996.
- Accepted October 15, 1996.
Harris DA, Das AM. Control of mitochondrial ATP synthesis in the heart. Biochem J. 1991;280:561-573. Review.
Koretsky AP, Katz LA, Balaban RS. Determination of pyridine nucleotide fluorescence from the perfused heart using an internal standard. Am J Physiol. 1987;253:H856-H862.
Scholz TD, Laughlin MR, Balaban RS, Kupriyanov VV, Heineman FW. Effect of substrate on mitochondrial NADH, cytosolic redox state, and phosphorylated compounds in isolated hearts. Am J Physiol. 1995;268:H82-H91.
Crompton M. The role of Ca2+ in the function and dysfunction of heart mitochondria. In: Langer GA, ed. Calcium and the Heart. New York, NY: Raven Press Publishers; 1990:167-199.
Miyata H, Silverman HS, Sollott SJ, Lakatta EG, Stern MD, Hansford RG. Measurement of mitochondrial free Ca2+ concentration in living single rat cardiac myocytes. Am J Physiol. 1991;261:H1123-H1134.
McCormack JG, Halestrap AP, Denton RM. Role of calcium ions in regulation of mammalian intramitochondrial metabolism. Physiol Rev. 1990;70:391-425. Review.
Backx PH, Ter Keurs HE. Fluorescent properties of rat cardiac trabeculae microinjected with fura-2 salt. Am J Physiol. 1993;264:H1098-H1110.
Cooper G IV. Myocardial energetics during isometric twitch contractions of cat papillary muscle. Am J Physiol. 1979;236:H244-H253.
Frampton JE, Harrison SM, Boyett MR, Orchard CH. Ca2+ and Na+ in rat myocytes showing different force-frequency relationships. Am J Physiol. 1991;261:C739-C750.
Spurgeon HA, Stern MD, Baartz G, Raffaeli S, Hansford RG, Talo A, Lakatta EG, Capogrossi MC. Simultaneous measurement of Ca2+, contraction, and potential in cardiac myocytes. Am J Physiol. 1990;258:H574-H586.
Yasumura Y, Nozawa T, Futaki S, Tanaka N, Suga H. Time-invariant oxygen cost of mechanical energy in dog left ventricle: consistency and inconsistency of time-varying elastance model with myocardial energetics. Circ Res. 1989;64:764-778.
Hasenfuss G, Mulieri LA, Blanchard EM, Holubarsch C, Leavitt BJ, Ittleman F, Alpert NR. Energetics of isometric force development in control and volume-overload human myocardium: comparison with animal species. Circ Res. 1991;68:836-846.
Unitt JF, McCormack JG, Reid D, MacLachlan LK, England PJ. Direct evidence for a role of intramitochondrial Ca2+ in the regulation of oxidative phosphorylation in the stimulated rat heart: studies using 31P n.m.r. and ruthenium red. Biochem J. 1989;262:293-301.