The Origin of Calcium Overload in Rat Cardiac Myocytes Following Metabolic Inhibition With 2,4-Dinitrophenol

J Mol Cell Cardiol 34, 859ÿ ÿ871 (2002) doi:10.1006/jmcc.2002.2024, available online at http://www.idealibrary.com on

1

The Origin of Calcium Overload in Rat Cardiac Myocytes Following Metabolic Inhibition With 2,4-Dinitrophenol Diane Hudman, Richard D. Rainbow, Chris L. Lawrence and Nicholas B. Standen* Department of Cell Physiology and Pharmacology, University of Leicester, University Road, PO Box 138, Leicester LE1 9HN, UK (Received 29 January 2002, accepted for publication 11 April 2002) D. HUDMAN, R. D. RAINBOW, C. L. LAWRENCE AND N. B. STANDEN. The Origin of Calcium Overload in Rat Cardiac Myocytes Following Metabolic Inhibition With 2,4-Dinitrophenol. Journal of Molecular and Cellular Cardiology (2002) 34, 859ÿ ÿ871. We have investigated the characteristics of the rise in cytoplasmic calcium that occurs when rat isolated cardiac ventricular myocytes are exposed to 2,4-dinitrophenol using conventional and confocal ¯uorescence microscopy and patch clamp. 2,4-dinitrophenol (200 M) caused cytoplasmic calcium to increase in two phases: (1) an initial rise in ¯uo-3 ¯uorescence of 36  2% that was maintained until rigor contraction; (2) a further progressive rise so that ¯uo-3 ¯uorescence had increased by 177  12% 535 s after 2,4-dinitrophenol addition. Both phases were unaffected by removal of external Ca2‡. 2,4-dinitrophenol caused mitochondrial depolarization, measured using tetramethyl rhodamine ethyl ester ¯uorescence. Mitochondrial depolarization was associated with a decrease in intra-mitochondrial calcium measured with rhod-2, and experiments on myocytes loaded with both ¯uo3 and rhod-2 showed that ¯uo-3 ¯uorescence increased as rhod-2 ¯uorescence fell. The correlation of the onset of the second phase of the increase in cytoplasmic calcium with rigor suggested that this phase was consequent on ATP depletion. DNP also caused activation of an ATP-sensitive potassium current. Depletion of sarcoplasmic reticulum calcium stores by pretreatment with ryanodine, thapsigargin and caffeine prior to the addition of 2,4-dinitrophenol did not affect the initial increase in cytoplasmic calcium, but abolished phase 2. Our results suggest that the initial rise in cytoplasmic calcium seen on application of 2,4-dinitrophenol results from release of mitochondrial calcium because of mitochondrial depolarization, while the second phase is caused by progressive release of calcium from the # 2002 Elsevier Science Ltd. All rights reserved. sarcoplasmic reticulum following depletion of intracellular ATP. KEY WORDS: Cardiac muscle; Heart; Ischaemia; Dinitrophenol; Ventricular myocytes; Calcium; Mitochondria; SR.

Introduction Regulation of intracellular Ca2‡ is essential for normal function of the myocardium and disruption of Ca2‡ homeostasis is a common correlate of pathological conditions. In particular, there is good evidence that such disrupted Ca2‡ regulation plays a critical role in the damaging effects of ischaemia or hypoxia followed by reperfusion.1,2 For example,

a rise in cytosolic Ca2‡ may trigger ventricular arrhythmias early in ischaemia,3 while further rises in cytosolic Ca2‡ on reperfusion are linked with postischaemic contractile dysfunction. Prolonged elevation of cytosolic Ca2‡ can ultimately lead to cell death.1,2,4 When ischaemia or hypoxia are followed by reperfusion, the resumption of ATP production in the presence of raised cytosolic Ca2‡ appears to be the trigger for cell damage, and the cytosolic

Please address all correspondence to: N. B. Standen, Department of Cell Physiology and Pharmacology, University of Leicester, PO Box 138, Leicester LE1 9HN, UK. Tel: 00 44 116 252 3302; Fax: 00 44 116 252 5045; E-mail: [email protected]

0022ÿ ÿ2828/02/070859 ‡ 13 $35.00/0

#

2002 Elsevier Science Ltd. All rights reserved.

860

D. Hudman et al.

Ca2‡ concentration immediately before reperfusion determines whether cells ultimately recover ÿ7 or die.5ÿ In isolated cardiac myocytes, ischaemia has been simulated by hypoxia or by metabolic inhibition. Such conditions have generally been reported to lead to an initial small increase in cytosolic Ca2‡ concentration, followed by a progressive rapid rise in Ca2‡ that usually occurs after cell shortening into rigor, and so appears to be associated with the depletion of intracellular ATP.2,5,8 Reperfusion with normoxic solution, or removal of metabolic inhibition, leads to a transient fall in cytosolic Ca2‡ concentration followed by either recovery of Ca2‡ homeostasis or an irreversible rise in cytosolic Ca2‡ concentration associated with cell death.1,2 Much of the reperfusion-induced cytosolic Ca2‡ overload appears to result from Ca2‡ entry through the Na‡/Ca2‡ exchanger working in reverse mode consequent on raised cytoplasmic [Na‡] caused by Na‡ loading during ischaemia. Thus, the postischaemic Ca2‡ overload can be attenuated by block of Na‡/Ca2‡ exchange with Ni2‡ or by removal of extracellular Na‡ during ischaemia.9,10 While there is good evidence for a dominant role of Na‡/Ca2‡ exchange in post-ischaemic Ca2‡ loading, the mechanisms that lead to increases in cytosolic Ca2‡ concentration during ischaemia are less well understood, despite its critical role in determining cell recovery or death. It has been suggested that Na‡/Ca2‡ exchange may also contribute to Ca2‡ loading prior to reperfusion, since intracellular Na‡ has been reported to increase during hypoxia, and both this and increases in cytosolic Ca2‡ concentration were attenuated in Na‡-free solution.11 In contrast, Grif®ths12 found that Na‡ was not necessary for the increase in cytosolic Ca2‡ concentration that occurred in myocytes exposed to metabolic inhibition. In the present study, we have therefore investigated the processes that cause intracellular Ca2‡ to increase during metabolic inhibition in rat isolated ventricular cardiac myocytes. To cause metabolic inhibition we used the inhibitor of oxidative phosphorylation 2,4-dinitrophenol (DNP) which has been used to simulate the effects of hypoxia,13,14 and we have used ¯uorescence methods to monitor cytoplasmic and mitochondrial Ca2‡ and mitochondrial membrane potential. We ®nd that DNP leads to a biphasic increase in cytosolic Ca2‡ concentration that is independent of extracellular Ca2‡. Our results suggest that the initial rise in cytosolic Ca2‡ concentration results from release of mitochondrial Ca2‡ following mitochondrial depolarization, while the second, progressive, phase of the increase in

Ca2‡, which follows rigor contraction, represents Ca2‡ release from the SR.

Materials and Methods Isolation of single ventricular myocytes Ventricular myocytes were isolated as described previously from the hearts of adult male Wistar rats (300ÿ ÿ400 g) killed by cervical dislocation.15 The care and sacri®ce of the animals conformed to the requirements of the UK. Animals (Scienti®c Procedures) Act 1986. Brie¯y, the heart was excised, attached to a Langendorff column, and perfused with Ca2‡-free Tyrode containing collagenase (Type I, Sigma) and protease (Type XIV, Sigma) as described previously.15 Typically, the isolation gave a 70ÿ ÿ90% yield of quiescent, rod-shaped myocytes. Cells were stored at 10 C in normal Tyrode for a maximum of 24 h before use. Drugs and experimental solutions For measurements of contraction and ¯uorescence studies, isolated ventricular myocytes were superfused with normal Tyrode containing (mM): NaCl 135, KCl 5, NaH2PO4 0.33, Na-pyruvate 5, glucose 10, MgCl2 1, CaCl2 2, HEPES 10. The pH of all experimental solutions was titrated to 7.4 with NaOH. Ca2‡-free Tyrode solution was as above with the exception of CaCl2. DNP (200 M) was added to normal Tyrode or Ca2‡-free Tyrode as required. For whole-cell patch clamp recordings from ventricular myocytes, the pipette (internal) solution contained (mM): KCl 125.4, KOH 14.6, MgCl2 1, MgATP 0.3, NaADP 0.1, NaGTP 0.3, BAPTA 10, HEPES 10, pH 7.2. For outside-out patch recordings from HEK 293 cells expressing Kir6.2/SUR2A the pipette solution contained (mM): KCl 100, EDTA 10, KOH 40, HEPES 10, and ATP 0.1 to inhibit around 90% of the current. The bath solution contained (mM): KCl 70, NaCl 70, MgCl2 2, CaCl2 2, and HEPES 10. Fluorescent dyes were purchased from Molecular Probes (USA), dissolved in dimethyl sulphoxide containing 5% pluronic acid (w/v), and stored at ÿ20 C. All other chemicals were purchased from Sigma (UK) dissolved in dimethyl sulphoxide at stock concentration and diluted in normal Tyrode or Ca2‡-free Tyrode solution as appropriate. The ®nal concentration of dimethyl sulphoxide did not exceed 0.025%. All solutions were applied by perfusion into

Dinitrophenol and Calcium Overload

the experimental bath and experiments were done at 30ÿ ÿ32 C. Fluorescence measurements For ¯uorescence measurements using conventional microscopy, myocytes were imaged using a Nikon 200 inverted microscope with a 40 oil-immersion objective. Fluorescence was excited with light from a delta-RAM monochromator (PTI) and emitted light at .510 nm was measured with a photo-multiplier tube or imaged with a CCD camera. An acquisition rate of 1 Hz was used throughout. Data acquisition and analysis were performed using PTI Imagemaster or Felix software. Alternatively, myocytes were imaged using an inverted laser scanning confocal microscope (Perkin-Elmer Ultra-View) with an Olympus 60 oil-immersion objective. Confocal emission ¯uorescence was captured at .580 nm. Images were transferred to a cooled frame transfer CCD camera as full frame images (1392  1040 pixels). An acquisition rate of 1 frame per s with 200 ms exposure, 2  2 binning of pixels with a 4 s lapse between images was used for all experiments. Images were digitized at 12 bits and data extracted off line using Ultra-view software. Cytoplasmic and mitochondrial Ca2‡ To measure cytoplasmic Ca2‡, myocytes were loaded for 20 min at room temperature (20ÿ ÿ22 C) with the acetoxymethyl ester (AM) form of the dye ¯uo-3 (¯uo-3 AM, 6 M) and excited at 488 nm. For mitochondrial Ca2‡, myocytes were loaded with the mitochondrial Ca2‡-sensitive dye rhod-2. The loading procedure was designed to load preferentially and heavily into the mitochondria. Myocytes were incubated with 4 M rhod-2 AM in normal Tyrode for 1 hr at room temperature and then washed 4 times with normal Tyrode and incubated overnight at 15 C. Rhod-2 was excited at 568 nm, and emitted light collected at .580 nm. In some experiments, cells were loaded with both ¯uo-3 and rhod-2 by following the rhod-2 loading procedure with 20 min incubation in 6 M ¯uo-3 AM, and in these experiments the excitation wavelength alternated between 488 and 568 nm. Intracellular pH Cells were loaded with 0.5 M of the AM form of the ratio-metric dye BCECF,16 for 15 min and excited at 490 and 430 nm with ®xed emission at 535 nm. BCECF calibration was carried out as described previously17 using the K‡/H‡ exchange activator

861

nigericin (1 M) to equilibrate intracellular and extracellular pH. At the end of each experiment, normal Tyrode solution was replaced by calibration solution and signals were recorded from four pH standards. Calibration buffers contained (mM): HEPES 4, KCl 120, EGTA (MES or PIPES) 0.5, pyruvate 5, glucose 5.6, K2ATP 1, with 4 M ionomycin and 0.2 M carbonyl cyanide mchlorophenylhydrazone. pH was adjusted with KOH or HCl to values of 8.0, 7.42, 7.23, or 6.62.17 Mitochondrial membrane potential Accumulation of the cationic dye tetramethylrhodamine ethyl ester (TMRE) in mitochondria has been shown to be driven by membrane potential in accordance with the Nernst equation.18 TMRE has been used to image time-dependent changes in mitochondrial membrane potentials.19 Myocytes were incubated with the AM form of TMRE (200 nM) for 20 min and ¯uorescence excited at 488 nm. Electrophysiology Whole-cell currents were measured from ventricular myocytes using conventional patch clamp methods with an Axopatch 200A ampli®er and Digidata 1322A interface. Recording pipettes were prepared from thick walled borosilicate glass using a Narishige micropipette puller. The pipette was heatpolished and ®lled with internal solution. Electrode resistances were 3ÿ ÿ6 M and seals were .1 G . Currents were ®ltered at 2 kHz and analysed on-line (pClamp V8, Axon Instruments). For HEK 293 cells, recording methods were as described for myocytes above, except that pipettes had resistances of 8ÿ ÿ12 M , and were coated with Sylgard. Outsideout patches of membrane were excised from HEK 293 cells stably expressing the cardiac KATP channel subunits Kir6.2 and SUR2A. Image processing and statistical analysis Epi¯uorescent image processing was performed using Imagemaster software (PTI) and confocal image processing was performed using Ultra-view software (v3.0, Perkin Elmer). To represent changes in ¯uorescent intensity over time, we have presented data as line graphs created from sequential images. The imaging systems enabled independent recordings to be made from several cells in the ®eld of view simultaneously, and allowed regions of interest to be set to include the whole of each cell throughout the experiment. Changes in mitochondrial membrane

862

D. Hudman et al.

Results DNP causes a biphasic increase in cytosolic Ca2‡ To investigate the effect of DNP on cytosolic calcium we used the dye ¯uo-3. DNP absorbs ultraviolet light

Phase 2

AB

3

2

A

B

100

1.8 1.6 1.4

50

1.2 380

Phase 1

410 440 470 Time (s)

d 1

250

200 400 600 Time (s)

800

b AB

100

0

200

400

600

Time (s) 2‡

800

200 150 100 50

**

50

***

200µM DNP 0

Length ( %)

c Length (%)

Fluo-3 fluorescence (F/F0)

4

Increase in fluo-3 fluorescence (%)

a

strongly, so that it is very dif®cult to use fura-2 under these conditions. Figure 1(a) shows that application of 200 M DNP to a ¯uo-3 AM loaded cardiac myocyte caused ¯uo-3 ¯uorescence to increase in two distinct phases, indicating a similar change in cytosolic calcium concentration. Following addition of DNP, ¯uo-3 ¯uorescence rose immediately by about 30%, and this level was maintained for about 300 s (phase 1), after which ¯uorescence began to rise progressively with time (phase 2). Since the initial increase in Ca2‡ reaches a relatively steady plateau, we measured phase 1 ¯uo-3 ¯uorescence as the average ¯uorescence between 140 and 180 s after the addition of DNP. In 10 cells, the rise in ¯uo-3 ¯uorescence during phase 1 was 36  2%, and the phase 2 began after 300  31 s, reaching 177  12% 535 s after the addition of DNP [Fig. 1(d)]. Figure 1(b) shows cell length recorded simultaneously from the cell of Figure 1(a) and it can be seen that the transition from the initial to the second phase of the rise in Ca2‡ coincided with cell shortening into rigor contraction. Examination of this transition on an expanded timescale [Fig. 1(c)]

F/F0

potential were calculated from the ¯uorescence intensity pro®le along a chosen line in the image. This line was drawn along the longitudinal axis of the cell through the brightest mitochondria, which were assumed to be in the confocal plane. In all cases ¯uorescence intensity was normalised to its initial value recorded in normal Tyrode (F0) and is expressed as the relative ¯uorescence (F/F0). Changes in cell length were also determined using the confocal system. Statistical signi®cance was assessed using analysis of variance (ANOVA) followed by Dunnet's test, or paired or unpaired t-tests as appropriate, and P , 0.05 was regarded as signi®cant. Data are presented as mean  SEM.

0 Phase 1

Phase 2

Figure 1 The effect of DNP on cytosolic Ca and cell length in an isolated rat ventricular myocyte. (a) Record of ¯uo-3 ¯uorescence from a single myocyte loaded with 6 M ¯uo-3 AM and imaged using confocal microscopy. DNP (200 M) was applied and indicated. (b) Cell length measured from the confocal images. A and B mark the beginning and end of cell shortening. (c) Cell length (continuous line) and ¯uo-3 ¯uorescence (open circles) from a single myocyte shown on an expanded timescale. Cell shortening occurs before the increase in ¯uo-3 ¯uorescence. (d) Mean increase in ¯uo-3 ¯uorescence in 10 cells during phase 1 and phase 2. For each cell, phase 1 ¯uorescence was measured as the average value between 80 and 120 s after DNP application and phase 2 ¯uorescence as the value 535 s after DNP application. ** P , 0.01, *** P , 0.001 compared to pre-DNP value.

Dinitrophenol and Calcium Overload

863

shows that the cell shortening preceded phase 2 of the rise in Ca2‡, being complete before ¯uo-3 ¯uorescence increased signi®cantly. These ®ndings were con®rmed in a further 9 cells. DNP can cause an increase in ¯avoprotein auto¯uorescence excited at 490 nm,20 a wavelength very close to that we used to excite ¯uo-3. We have considered the possibility that such auto¯uoresecence might contribute to the apparent increase in ¯uo-3 ¯uorescence we observe on addition of DNP. We therefore compared the effect of DNP in unloaded cells, and in cells loaded with ¯uo-3, using identical excitation intensities and detection gains. In unloaded cells DNP increased ¯uorescence by 7  4%, compared to 32  1% in ¯uo-3 loaded cells (n ˆ 5 in each case). Thus auto¯uorecence can account for about 20% of the phase 1 increase in ¯uorescence caused by DNP, so that the remaining 80% represents a true increase in ¯uo-3 ¯uorescence, indicating a phase 1 rise in Ca2‡. Similarly, the contribution of auto¯uorescence to the ¯uorescence increase in phase 2 is less than 2%.

It is possible that the insensitivity of the DNPinduced rise in cytosolic Ca2‡ to extracellular Ca2‡ arises in part because the sarcolemmal Na‡/ Ca2‡ exchanger is inhibited by severe ATP depletion induced by DNP. We therefore examined the recovery of cytosolic Ca2‡ after release of Ca2‡ from the sarcoplasmic reticulum by caffeine (10 mM) in the presence and absence of DNP. The solid line in Figure 2(c) shows that in the absence of DNP caffeine caused an increase in ¯uo-3 ¯uorescence, indicating sarcoplasmic reticulum Ca2‡ release, followed by a rapid return to its baseline level which should primarily be due to removal of cytoplasmic Ca2‡ by Na‡/Ca2‡ exchange. In the presence of 200 M DNP (broken line) ¯uo-3 ¯uorescence failed to recover fully after caffeine stimulation, suggesting that Na‡/Ca2‡ exchange was inhibited. Mean values for the ¯uo-3 ¯uorescence intensity measured 60 s after caffeine treatment in control and DNP treated cells are shown in the inset to Figure 2(c) (n ˆ 7 cells in each case).

The rise in cytosolic Ca2‡ is not dependent on extracellular calcium

DNP causes intracellular acidi®cation

To determine whether the DNP-induced rise in cytosolic Ca2‡ concentration was dependent on Ca2‡ entry from the external solution, DNP was applied in the absence of extracellular Ca2‡. In these experiments, ¯uo-3 AM loaded myocytes were superfused with Ca2‡-free Tyrode containing 1 mM EGTA for 2 min before the application of DNP. Figure 2(a) shows the mean ¯uo-3 ¯uorescence measured from cells exposed to DNP in the absence or presence of extracellular Ca2‡ (n ˆ 10 in each case). It can be seen that both the initial and later rise in cytosolic Ca2‡ concentration occurred irrespective of the presence of extracellular Ca2‡. Figure 2(b) shows that the presence and absence of extracellular Ca2‡also did not affect the percentage of cells that had contracted into rigor as a function of time after DNP application. Similarly, there was no signi®cant difference in the time to rigor; halfmaximal shortening occurred after 282  18 s and 270  21 s with and without extracellular Ca2‡, respectively. There was also no difference in the degree of shortening of cells in the presence and absence of extracellular Ca2‡ (to 38.0  3.5% and 37.4  3.2% of initial cell length respectively). These ®ndings suggest that that the rise in cytosolic Ca2‡ caused by DNP does not occur because of sarcolemmal Ca2‡ entry, but rather is likely to result from redistribution of intracellular Ca2‡.

In multi-cellular preparations, ischaemia has been reported to produce intracellular acidosis attributed to a build up of lactic acid by anaerobic glycolysis, preceded by intracellular alkalosis, presumed to be ÿ23 due to the breakdown of phosphocreatine.4,21ÿ Similarly, in isolated myocytes loaded with BCECF to monitor changes in cytosolic pH, we found that 200 M DNP caused a decrease in cytosolic pH preceded by a brief period of alkalosis where pH levels rose by 0.20  0.05 pH units from the control level of 7.31  0.003. Cells then acidi®ed over the following 300ÿ ÿ400 s before a further acidi®cation occurred corresponding to shortening into rigor (n ˆ 6), where pH fell to 6.60  0.06 (not shown). We have considered the possibility that changes in ¯uo-3 ¯uorescence resulting from changes in pH rather than cytosolic Ca2‡ concentration might contribute to the effects we observe with DNP. However, ¯uo-3 ¯uorescence shows very little sensitivity to pH above pH 6.8, while ¯uorescence is reduced slightly at lower pH.24 Thus any effect of pH cannot account for the rises in ¯uo-3 ¯uorescence we observe in response to DNP, which we therefore attribute to rises in cytosolic Ca2‡ concentration. As for ¯uo-3, ¯avoprotein auto¯uorescence could potentially contribute to the BCECF signal. However, we found no signi®cant auto¯uorescence measurable at the intensity settings used to record BCECF ¯uorescence in these experiments, re¯ecting the

864

D. Hudman et al.

Figure 2 Removing extracellular Ca2‡ does not affect the rise in cytosolic Ca2‡ induced by DNP. (a) Mean ¯uo-3 ¯uorescence recorded from cells superfused with normal Tyrode (®lled circles) or Ca2‡-free Tyrode (open triangles, n ˆ 10 in each case). DNP (200 M) was applied as indicated. (b) Percentage of cells in rigor as a function of time after addition of DNP in normal Tyrode (®lled circles) and Ca2‡-free Tyrode (open circles). The number of cells was 148 and 150 respectively). Rigor was de®ned as cessation of cell shortening. (c) The records show ¯uo-3 ¯uorescence from myocytes bathed in normal Tyrode in response to the application of 10 mM caffeine (solid line) and 10 mM caffeine together with 200 M DNP (broken line). The insert shows mean ¯uorescence 60 s after caffeine stimulation from 7 cells in each case. ** P , 0.001 compared to caffeine alone.

high intensity of the emitted ¯uorescence from BCECF. DNP causes mitochondrial depolarization The uncoupling effect of DNP occurs because it is a protonophore and so reduces the potential across the inner mitochondrial membrane. It is possible that part of the DNP-induced rise in cytoplasmic Ca2‡ results from release of mitochondrial Ca2‡ consequent on mitochondrial depolarization. We therefore investigated the ability of 200 M DNP to depolarize the mitochondrial membrane potential in our ventricular myocytes using the mitochondrial membrane potential-sensitive dye TMRE. As TMRE is positively charged it accumulates in mitochondria due to the negative mitochondrial membrane

potential. Mitochondrial depolarization causes a loss of TMRE from mitochondria as the dye redistributes between mitochondria and cytoplasm in a Nernstian fashion.25 Myocytes loaded with TMRE and imaged using the confocal microscope [Fig. 3(a)], showed a pattern of mitochondrial ¯uorescence characteristic of cardiac muscle, with mitochondria arranged in longitudinal chains,26 while Figure 3(b) shows a mitochondrial chain in the confocal plane at higher magni®cation. To assess mitochondrial depolarization we measured TMRE ¯uorescence along a line drawn along a mitochondrial chain in the confocal plane. Under these conditions, addition of the uncoupler FCCP (500 nM), which causes complete dissipation of the mitochondrial membrane potential27 in myocytes bathed in normal Tyrode caused an immediate decrease in TMRE ¯uorescence [Fig. 3(c)]. In the

Dinitrophenol and Calcium Overload

865

Figure 3 DNP causes mitochondrial depolarization. (a) Confocal image of a myocyte loaded with TMRE (200 nM), showing characteristic pattern of mitochondrial ¯uorescence. (b) Enlarged view of the region enclosed by the rectangle in Figure 3(a). The box on the image shows the chain of mitochondria through which an analysis line was drawn and used to measure TMRE ¯uorescence. (c) Effect of FCCP (500 nM) on TMRE ¯uorescence. The points show mean values from 6 myocytes, and FCCP was added as indicated. The decrease in TMRE ¯uorescence indicates mitochondrial depolarization, (d) Mean (  SEM) effect of DNP (200 M) on TMRE ¯uorescence in 6 myocytes. (e) Mean decrease in TMRE ¯uorescence following exposure of myocytes to either FCCP or DNP (n ˆ 6 in each case).

absence of FCCP, continued imaging of cells loaded with TMRE did not cause any decrease in ¯uorescence intensity, indicating that our low-intensity confocal imaging system was not itself inducing mitochondrial depolarisation as has sometimes been reported when using linescan imaging.28 Figure 3(d) shows that 200 M DNP also caused mitochondrial depolarization, though this occurred with a slower time course than that seen with FCCP. Figure 3(e) shows the mean reduction in TMRE ¯uorescence in cells exposed to either 500 nM FCCP or 200 M DNP. Although FCCP appears to reduce TMRE ¯uorescence slightly more, the difference was not signi®cant, consistent with 200 M DNP causing near complete dissipation of the mitochondrial membrane potential. The initial rise in cytosolic Ca2‡ represents mitochondrial Ca2‡ release To investigate possible changes in mitochondrial Ca2‡ in response to DNP we used myocytes loaded

with the mitochondrial calcium sensitive dye rhod-2. We con®rmed that mitochondrial depolarization causes release of mitochondrial Ca2‡ by depolarizing mitochondria with 500 nM FCCP (see above). FCCP caused a rapid reduction in rhod-2 ¯uorescence signal by around 80% [Fig. 4(a)], with a similar time-course to its effect on mitochondrial membrane potential [cf. Fig. 3(c)]. To investigate the effects of DNP on mitochondrial and cytosolic Ca2‡ simultaneously, we used myocytes loaded with both rhod-2 AM and ¯uo-3AM. Figure 4(b) illustrates such an experiment. Addition of DNP caused a rapid reduction in rhod-2 ¯uorescence by 43% from its initial level. This coincided with an initial increase in cytosolic Ca2‡ indicated by a 41% increase in ¯uo-3 ¯uorescence. With prolonged DNP exposure, there was a subsequent progressive increase in ¯uo-3 ¯uorescence corresponding to phase 2 described above. This second phase was not correlated with any further change in rhod-2 ¯uorescence [Fig. 4(b)]. The mean changes in rhod-2 and ¯uo-3 ¯uorescence in 6 cells during the initial and second phase are shown in Figure 4(c), and con®rm that

866

D. Hudman et al.

a

b Rhod-2 and fluo-3 fluorescence (F/F 0)

Rhod-2 fluorescence (F/F0)

1.2 1.0 0.8 0.6 0.4 0.2 500nM FCCP 0

50 100 150 200 250 Time (s) c

Change in fluorescence (%)

0.0

3 Phase 2

2 Phase 1

1

200µM DNP 0

0

200 400 Time (s)

600

200

100

0

-100 Phase 1

Phase 2

Figure 4 The effect of DNP on mitochondrial calcium concentration measured using rhod-2. (a) Mean (SEM) rhod-2 ¯uorescence from 10 myocytes. FCCP (500 nM) was added as indicated to cause mitochondrial depolarization. The decrease in rhod-2 ¯uorescence indicates a decrease in mitochondrial calcium concentration. (b) Recording of ¯uorescence from a single myocyte loaded with both 4 M Rhod-2 AM and 6 M ¯uo-3 AM, imaged using confocal microscopy, and exposed to DNP as indicated by the bar below. Rhod-2 ¯uorescence is shown by the continuous line and ¯uo-3 ¯uorescence by the broken line. (c) Mean changes (SEM) in ¯uo-3 and rhod-2 ¯uorescence (®lled and open bars respectively) from dual-loaded myocytes (n ˆ 6) during phase 1 (averaged between 80 and 120 s after DNP application) and phase 2 (measured 535 s after DNP application).

rhod-2 ¯uorescence does not change signi®cantly between these phases of the rise in cytosolic Ca2‡. These results suggest that the phase 1 rise in cytosolic Ca2‡ seen in response to DNP represents from release of mitochondrial Ca2‡ but that the second phase results from a different source. DNP induces whole-cell KATP current The correlation of the start of the second phase of the DNP-induced increase in cytosolic Ca2‡ with the development of rigor suggest that this rise in Ca2‡ may depend upon exhaustion of cytoplasmic ATP. To obtain information about changes in ATP, we

have used patch clamp to measure whole-cell currents through ATP-sensitive K‡ (KATP) channels, since DNP or other metabolic inhibitors have been shown to lead to KATP channel activation in cardiac muscle.29 Figure 5(a) shows a recording of wholecell current from a myocyte exposed to 200 M DNP and held at 0 mV. Application of DNP induced an outward current that began after 180  10 s and reached a peak after 193  13 s. Figure 5(b) shows current-voltage relationships obtained in normal Tyrode before the application of DNP, at the peak of DNP-induced current, and following washout with normal Tyrode. The DNP-induced current reversed close to the calculated EK of ÿ86 mV. Similar results were obtained in 7 further cells; the mean maximal

Dinitrophenol and Calcium Overload

867

Figure 5 KATP currents in response to DNP. (a) Recording of whole-cell current from a myocyte continuously perfused with normal Tyrode and held at 0 mV. DNP (200 M) was added as indicated, (b) Current-voltage relationships measured using steps to voltages between ÿ100 and ‡60 mV from a myocyte before application of DNP (®lled circles), during DNP (®lled squares), and after wash-off (open triangles). (c) Current recorded from an outside-out patch of membrane from a HEK 293 cell stably expressing KIR6.2/SUR2A in the presence of 100 M ATP in the pipette (intracellular) solution. The points show the mean current measured during the last 50 ms of 100 ms steps to ÿ80 mV from a holding potential of ÿ20 mV. Pinacidil (100 M) and DNP (200 M) were added as indicated. C indicates control solution. (d) Mean (  SEM) current normalised to its value in control solution from 6 patches exposed to DNP and pinacidil as in the experiment of panel C. Pinacidil resulted in a signi®cant activation of KATP current, while DNP was without effect. ** P , 0.01 compared to control.

current amplitude at 0 mV was 30.2  8.6 pA/pF. Washout of DNP restored currents to control levels [Figs 5(a), (b)]. It is clear that the increase in KATP current both begins and reaches a maximal value earlier than the occurrence of rigor and the transition from the ®rst to the second phase of the rise in cytosolic Ca2‡ seen in Figure 1. There are a number of possibilities as to why this might occur. KATP channels could be activated by a fall in subsarcolemmal ATP, possibly in conjunction with a rise in ADP, that occurs before ATP levels around the contractile proteins have fallen suf®ciently to cause rigor. It is also possible

that DNP might cause KATP channel activation directly, as has been suggested by Alekseev et al.13 The latter possibility seems unlikely, since there was still a substantial delay between the application of DNP and the start of KATP channel activation. We further investigated the possibility of direct KATP channel activation by DNP using the molecular constituents of cardiac sarcolemmal KATP channels, Kir6.2 and SUR2A, expressed in HEK 293 cells. Figure 5(c) shows results from an outside-out patch of membrane from a HEK 293 cell expressing Kir6.2/SUR2A in the presence of 100 M intracellular ATP. Pipette (intracellular) [K‡] was 140 mM,

868

D. Hudman et al.

bath [K‡] was 70 mM, and inward K‡ currents were measured during 100 ms pulses to ÿ80 mV from a holding potential of ÿ20 mV, repeated every 5 s. As expected, the KATP channel opener pinacidil (100 M) caused substantial and immediate KATP current activation, con®rming channel expression. However, a subsequent 10 min application of DNP (200 M) was without effect, following washout of DNP however, re-application of pinacidil again caused substantial KATP channel activation. Figure 5(d) illustrates mean results from similar experiments on patches from 6 cells, showing that we were unable to detect any direct effect of DNP in causing cloned KATP channel activation. The second phase of the rise in cytosolic Ca2‡ is abolished by depletion of sarcoplasmic reticulum Ca2‡ Since the second phase of the DNP-induced rise in cytosolic Ca2‡ appears to come from an intracellular

source other than mitochondria, we have investigated the possibility that the phase 2 rise in cytosolic Ca2‡ concentration results from release of Ca2‡ from the sarcoplasmic reticulum (SR). We therefore investigated the effect of depleting SR Ca2‡ on the cytosolic Ca2‡ response to DNP. In these experiments, myocytes were loaded with ¯uo-3AM in normal Tyrode and then superfused for 10 min with Ca2‡-free Tyrode containing ryanodine (10 M) or ryanodine and thapsigargin (10 M and 1 M, respectively). To test for SR Ca2‡ depletion, caffeine (10 mM, 10 s) was applied to release SR Ca2‡. Figure 6(a) shows that caffeine caused a substantial rise in cytosolic Ca2‡ in control cells, and that this was greatly reduced by pretreatment with ryanodine and abolished by pretreatment with both ryanodine and thapsigargin. To assess the effect of SR Ca2‡ depletion on the Ca2‡ response to DNP addition, cells loaded with ¯uo-3 AM were pre-treated with ryanodine and thapsigargin and stimulated with caffeine.

Figure 6 SR Ca2‡ depletion abolishes the phase 2 rise in cytosolic Ca2‡ (a) Mean increases in ¯uo-3 ¯uorescence following stimulation with 10 mM caffeine in control cells (C, n ˆ 6) and in cells pretreated for 10 min with ryanodine (R, 10 M, n ˆ 6) or ryanodine and thapsigargin (R ‡ T, 10 and 1 M respectively, n ˆ 6). (b) Mean ¯uo-3 ¯uorescence following the addition of DNP (200 nM) in untreated cells bathed in Ca2‡-free Tyrode solution (open circles, n ˆ 5) or cells pretreated with ryanodine and thapsigargin and exposed to caffeine for 10 s and bathed in Ca2‡-free Tyrode solution throughout (®lled circles, n ˆ 5). (c) Mean increase in ¯uo-3 ¯uorescence in phase 1 (averaged between 80 and 120 s after DNP application) and phase 2 (535 s after DNP application) in control cells (open bars) and in cells in which SR Ca2‡ was depleted. As described above (black bars). *** P , 0.001 vs phase 1.

Dinitrophenol and Calcium Overload

Figure 6(b) shows mean ¯uo-3 ¯uorescence results from cells treated in this way and from control cells bathed in Ca2‡-free Tyrode (n ˆ 5 in each case). It can be seen that there was little difference in the magnitude of the phase 1 rise in cytosolic Ca2‡, but that phase 2 of the increase in Ca2‡ was abolished in SR-depleted cells [Fig. 6(c)]. These results suggest that the phase 2 rise in cytosolic Ca2‡ in response to DNP arises from release of Ca2‡ from the SR.

Discussion In this study we have investigated the mechanisms by which DNP causes an increased cytosolic Ca2‡ load in isolated quiescent rat ventricular myocytes. Our results show that DNP caused cytosolic Ca2‡ to rise in two phases: an initial rapid rise to a relatively steady level that increased ¯uo-3 ¯uorescence by 30ÿ ÿ40%, followed by a further progressive rise that began after the cell had contracted into rigor. Both phases were unaffected by removal of extracellular Ca2‡, and our ®ndings suggest that phase 1 arises from mitochondrial Ca2‡, while phase 2 results from release of Ca2‡ from the SR. Evidence for a mitochondrial source for the initial rise in cytosolic Ca2‡ concentration comes from experiments on cells in which both cytosolic and mitochondrial calcium were monitored simultaneously using both ¯uo-3 and rhod-2. In such cells, the initial rise in cytosolic Ca2‡ concentration correlates well with a fall in mitochondrial Ca2‡ concentration. It seems likely that the mechanisms underlying this initial rise in Ca2‡ may be relatively speci®c to the effects of DNP. First, the rise in Ca2‡ we observed takes the form of a rapid step to a moderately higher level, while studies using substrate-free hypoxia or metabolic inhibition with cyanide and inhibitors of glycolysis have generally reported a slower small rise in Ca2‡.1,5,8 Secondly, a rapid fall in mitochondrial Ca2‡ has not been reported in response to hypoxia, where mitochondrial Ca2‡concentration remains little changed until rigor, after which it begins to rise steadily, an effect thought to re¯ect increased Ca2‡ in¯ux from the cytoplasm as cytosolic Ca2‡ concentration rises.2,5,7 It is likely that the different pattern we observe occurs because DNP acts as a protonophore. Thus, mitochondrial depolarization begins as soon as DNP is applied [Fig. 3(d)], whereas with substratefree hypoxia or CNÿand iodoacetate depolarization of the mitochondrial membrane potential does not occur until ATP is depleted, as indicated by rigor.2,7,27 The route by which Ca2‡ leaves the mitochondria remains unknown, but the negative

869

mitochondrial membrane potential favours Ca2‡ entry into the mitochondrial matrix via the Ca2‡ uniporter, and modest depolarizations of mitochondrial membrane potential have been reported to cause increases in cytosolic Ca2‡.7 Indeed in our experiments it is clear that the initial rise in cytosolic Ca2‡ concentration and the simultaneous fall in mitochondrial Ca2‡ concentration in response to DNP occur before the mitochondrial membrane potential is fully depolarized. The second, progressive, phase of the increase in cytosolic Ca2‡ concentration that we observe begins just after the myocyte has contracted into rigor [Fig. (1)]. Rigor contraction is associated with complete ATP depletion,30 so it is likely that the phase 2 rise in cytosolic Ca2‡ concentration is also a consequence of the exhaustion of ATP. Rigor, and the start of the phase 2 Ca2‡ rise also occur at the time when the depolarization of the mitochondrial membrane potential in response to DNP reaches completion, and at this point mitochondria are likely to be major ATP consumers, contributing to an abrupt fall in ATP.31 We found that the second phase of the rise in cytosolic Ca2‡ concentration was abolished in cells whose SR calcium stores had been depleted using caffeine and ryanodine to empty the SR and thapsigargin to inhibit the Ca2‡-ATPase and so prevent re®lling. This suggests that phase 2 of the rise in cytosolic Ca2‡ concentration in response to DNP results from release of Ca2‡ from the SR. Since this rise begins after ATP depletion, its key determinant may be the failure of Ca2‡ reuptake into SR by the SR Ca2‡-ATPase, coupled with leakage of Ca2‡ from the SR. SR Ca2‡ ef¯ux through the ryanodine receptor may itself increase to some extent as the cytosolic Ca2‡ concentration increases, though the rise in free [Mg2‡] that would result from ATP depletion and the fall in pH should both reduce ryanodine receptor opening.32 In permeabilized myocytes, removal of intracellular ATP has been reported to lead to an initial increase in SR Ca2‡ content, followed by a progressive decline to baseline levels.33 This suggests that ATP depletion alone does not account for the phase 2 SR Ca2‡ release that our results with DNP imply. It is possible that other actions of DNP, for example mitochondrial depolarization, account for these differences between the effects of DNP and ATP depletion alone. Interestingly, it has recently been shown that opening of the mitochondrial permeability transition pore induced by photodamage in cardiac myocytes is also independent of external Ca2‡ and depends on a rise in Ca2‡ resulting from SR Ca2‡ release.31 Application of DNP to patch clamped cardiac myocytes led to the development of whole-cell KATP

870

D. Hudman et al.

current. This current was initiated 180  10 s after application of DNP and rapidly reached a peak, substantially before rigor. DNP has been reported previously to open cardiac sarcolemmal KATP channels before substantial changes in intracellular nucleotides have occurred,34 and Alekseev et al.13 have reported that DNP can cause direct activation of KATP channels in patches excised from cardiac myocytes at constant ATP levels. To investigate whether such DNP activation might be a property of the constituent subunits of cardiac sarcolemmal KATP, Kir6.2 and SUR2A, we studied KATP channels in HEK 293 cells that stably expressed these subunits. Although such cloned channels could be activated by the KATP channel opener pinacidil, we did not observe activation with DNP. It is possible that the reported DNP-sensitivity of native KATP channels in cardiac myocytes13 involves channelassociated components other than Kir6.2 and SUR2A. We feel that it is most likely, however, that the KATP activation we have observed in ventricular myocytes with DNP results from an earlier fall in subsarcolemmal than bulk ATP levels,34 since we have observed similar early KATP channel activation in the absence of DNP using metabolic inhibition with cyanide and iodoacetic acid (Lawrence et al., unpublished observations). Our overall ®ndings suggest that while DNP, like other metabolic inhibitors, induces rises in cytoplasmic Ca2‡, the mechanisms that increase cytosolic Ca2‡ concentration may differ in some details from those involved in metabolic inhibition by substratefree hypoxia, ischaemia, or metabolic inhibition with agents that are not protonophores. Under such conditions, entry of extracellular Ca2‡ appears important, at least in the later stages of ischaemia and following reperfusion, and much of this entry is thought to occur through the Na‡/Ca2‡exchanger running in reverse mode under conditions were internal Na‡ concentration increases.6,10,35 In contrast, DNP causes a rise in cytosolic Ca2‡ concentration that is dominated by release of Ca2‡ from intracellular stores and appears to rapidly inhibit Na/Ca exchange. It seems likely that these differences relate to the immediate commencement of mitochondrial depolarization that occurs with DNP. DNP has provided a useful means to induce ``chemical hypoxia'',14,36 and pre-treatment with DNP has also been shown to protect against subsequent damage by ischaemia and reperfusion.37 The pattern of changes in intracellular Ca2‡ induced by DNP need to be taken into consideration in interpreting such work, and the effects of DNP may provide information about the processes involved in damage by metabolic inhibition and

about mechanisms that can protect against such damage.

Acknowledgements We thank Dr A. Tinker for the gift of HEK 293 cells stably expressing Kir6.2/SUR2A and the British Heart Foundation and Wellcome Trust for support.

References 1. SILVERMAN HS, STERN MD. Ionic basis of ischaemic cardiac injury: insights from cellular studies. Cardiovasc Res 1994; 28: 581ÿ ÿ597. 2. DI LISA F, BERNARDI P. Mitochondrial function as a determinant of recovery or death in cell response to injury. Mol Cell Biochem 1998; 184: 379ÿ ÿ391. 3. TANG T, DUFFIELD R, HO AK. Effects of Ca2‡ channel blockers on Ca2‡ loading induced by metabolic inhibition and hyperkalemia in cardiomyocytes. Eur J Pharmacol 1998; 360: 205ÿ ÿ211. 4. EISNER DA, NICHOLS CG, O'NEILL SC, SMITH GL, VALDEOLMILLOS M. The effects of metabolic inhibition on intracellular calcium and pH in isolated rat ventricular cells. J Physiol 1989; 411: 393ÿ ÿ418. 5. MIYATA H, LAKATTA EG, STERN MD, SILVERMAN HS. Relation of mitochondrial and cytosolic free calcium to cardiac myocyte recovery after exposure to anoxia. Circ Res 1992; 71: 605ÿ ÿ613. 6. RALENKOTTER L, DALES C, DELCAMP TJ, HADLEY RW. Cytosolic [Ca2‡], [Na‡], and pH in guinea pig ventricular myocytes exposed to anoxia and reoxygenation. Am J Physiol 1997; 272: H2679ÿ ÿH2685. 7. DELCAMP TJ, DALES C, RALENKOTTER L, COLE PS, HADLEY RW. Intramitochondrial [Ca2‡] and membrane potential in ventricular myocytes exposed to anoxia-reoxygenation. Am J Physiol 1998; 275: H484ÿ ÿH494. 8. LADILOV YV, SIEGMUND B, BALSER C, PIPER HM. Simulated ischemia increases the susceptibility of rat cardiomyocytes to hypercontracture. Circ Res 1997; 80: 69ÿ ÿ75. 9. Poole-WILSON PA, HARDING DP, BOURDILLON PD, TONES MA. Calcium out of control. J Mol Cell Cardiol 1984; 16: 175ÿ ÿ187. 10. NISHIDA M, BORZAK S, KRAEMER B et al. Role of cation gradients in hypercontracture of myocytes during simulated ischemia and reperfusion. Am J Physiol 1993; 264: H1896ÿ ÿH1906. 11. HAIGNEY MC, MIYATA H, LAKATTA EG, STERN MD, SILVERMAN HS. Dependence of hypoxic cellular calcium loading on Na‡-Ca2‡ exchange. Circ Res 1992; 71: 547ÿ ÿ557. 12. GRIFFITHS EJ. Reversal of mitochondrial Na/Ca exchange during metabolic inhibition in rat cardiomyocytes. FEBS Lett 1999; 453: 400ÿ ÿ404. 13. ALEKSEEV AE, GOMEZ LA, ALEKSANDROVA LA, BRADY PA, TERZIC A. Opening of cardiac sarcolemmal KATP channels by dinitrophenol separate from metabolic inhibition. J Membr Biol 1997; 157: 203ÿ ÿ214.

Dinitrophenol and Calcium Overload 14. JOVANOVIC N, JOVANOVIC S, JOVANOVIC A, TERZIC A. Gene delivery of Kir6.2/SUR2A in conjunction with pinacidil handles intracellular Ca2‡ homeostasis under metabolic stress. FASEB J 1999; 13: 923ÿ ÿ929. 15. LAWRENCE C, RODRIGO GC. A Na‡-activated K‡ current (IK,Na) is present in guinea-pig but not rat ventricular myocytes. P¯uÈgers Archiv 1999; 437: 831ÿ ÿ838. 16. RINK TJ, TSIEN RY, POZZAN T. Cytoplasmic pH and free ÿ196. Mg2‡ in lymphocytes. J Cell Biol 1982; 95: 189ÿ 17. PI Y, WALKER JW. Role of intracellular Ca2‡ and pH in positive inotropic response of cardiomyocytes to diacylglycerol. Am J Physiol 1998; 275: H1473ÿ ÿH1481. 18. BUDINGER GR, DURANTEAU J, CHANDEL NS, SCHUMACKER PT. Hibernation during hypoxia in cardiomyocytes. Role of mitochondria as the O2 sensor. J Biol Chem 1998; 273: 3320ÿ ÿ3326. 19. LOEW LM, TUFT RA, CARRINGTON W, FAY FS. Imaging in ®ve dimensions: time-dependent membrane potentials in individual mitochondria. Biophys J 1993; 65: 2396ÿ ÿ2407. 20. SATO T, O'ROURKE B, MARBAN E. Modulation of mitochondrial ATP-dependent K‡ channels by protein kinase C. Circ Res 1998; 83: 1110ÿ ÿ1114. 21. ALLEN DG, MORRIS PG, ORCHARD CH, PIROLO JS. A nuclear magnetic resonance study of metabolism in the ferret heart during hypoxia and inhibition of glycolysis. J Physiol 1985; 361: 185ÿ ÿ204. 22. FRY CH, HARDING DP, MOUNSEY JP. The effects of cyanide on intracellular ionic exchange in ferret and rat ventricular myocardium. Proc R Soc Lond B 1987; 230: 53ÿ ÿ75. 23. ELLIS D, NOIREAUD J. Intracellular pH in sheep Purkinje ®bres and ferret papillary muscles during hypoxia and recovery. J Physiol 1987; 383: 125ÿ ÿ141. 24. MINTA A, KAO JP, TSIEN RY. Fluorescent indicators for cytosolic calcium based on rhodamine and ¯uorescein chromophores. J Biol Chem 1989; 264: 8171ÿ ÿ8178. 25. EMAUS RK, GRUNWALD R, LEMASTERS JJ. Rhodamine 123 as a probe of transmembrane potential in isolated rat-liver mitochondria: spectral and metabolic properties. Biochim Biophys Acta 1986; 850: 436ÿ ÿ448. 26. DUCHEN MR, LEYSSENS A, CROMPTON M. Transient mitochondrial depolarizations re¯ect focal sarcoplasmic reticular calcium release in single rat cardiomyocytes. J Cell Biol 1998; 142: 975ÿ ÿ988.

871

27. LAWRENCE CL, RODRIGO GC, BILLUPS B, HUDMAN D, DAVIES NW, STANDEN NB. Diaxoxide protects rat isolated cardiac myocytes against hypercontracture following metabolic inhibition, but does not cause mitochondrial depolarization. J Physiol 2001; 533: 33. 28. ZOROV DB, FILBURN CR, KLOTZ LO, ZWEIER JL, SOLLOTT SJ. Reactive oxygen species (ROS)-induced ROS release: a new phenomenon accompanying induction of the mitochondrial permeability transition in cardiac myocytes. J Exp Med 2000; 192: 1001ÿ ÿ1014. 29. BRADY PA, ZHANG S, LOPEZ JR, JOVANOVIC A, ALEKSEEV AE, TERZIC A. Dual effects of glyburide, an antagonist of KATP channels, on metabolic inhibitioninduced Ca2‡ loading in cardiomyocytes. Eur J Pharmacol 1996; 308: 343ÿ ÿ349. 30. BARRY WH, PEETERS GA, RASMUSSEN CA, CUNNINGHAM MJ. Role of changes in [Ca2‡]i in energy deprivation contracture. Circ Res 1987; 61: 726ÿ ÿ734. 31. DUCHEN MR. Mitochondria and Ca2‡ in cell physiology and pathophysiology. Cell Calcium 2000; 28: 339ÿ ÿ348. 32. XU L, MANN G, MEISSNER G. Regulation of cardiac Ca2‡ release channel (ryanodine receptor) by Ca2‡, H‡, Mg2‡, and adenine nucleotides under normal and simulated ischemic conditions. Circ Res 1996; 79: 1100ÿ ÿ1109. 33. YANG Z, STEELE DS. Effects of cytosolic ATP on Ca2‡ sparks and SR Ca2‡ content in permeabilized cardiac myocytes. Circ Res 2001; 89: 526ÿ ÿ533. 34. WEISS J, HILTBRAND B. Functional compartmentation of glycolytic versus oxidative metabolism in isolated rabbit heart. J Clin Invest 1985; 75: 436ÿ ÿ447. 35. LADILOV YV, BALSER C, PIPER HM. Protection of rat cardiomyocytes against simulated ischemia and reoxygenation by treatment with protein kinase C activator. Circ Res 1998; 82: 451ÿ ÿ457. 36. JOVANOVIC A, JOVANOVIC S, LORENZ E, TERZIC A. Recombinant cardiac ATP-sensitive K‡ channel subunits confer resistance to chemical hypoxia-reoxygenation injury. Circulation1998; 98: 1548ÿ ÿ1555. 37. MINNERS J, van den BOS EJ, YELLON DM, SCHWALB H, OPIE LH, SACK MN. Dinitrophenol, Cyclosporin A, and trimetazidine modulate preconditioning in the isolated rat heart: support for a mitochondrial role in cardioprotection. Cardiovasc Res 2000; 47: 68ÿ ÿ73.