INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

ALTHOUGH NUMEROUS STUDIES suggest that cellular Ca²⁺ overload is responsible for irreversible myocardial injury in ischemia/reperfusion or hypoxia/reoxygenation, it remains unclear how Ca²⁺ mediates cell injury (34). One frequently proposed mechanism for transition from reversible to irreversible injury implicates mitochondrial Ca²⁺ overload with a subsequent Ca²⁺-dependent increase in mitochondrial inner membrane permeability. This is characterized by loss of mitochondrial membrane potential and equilibration of all solutes under 1,500 Da across the inner membrane (for reviews see Refs. 9 and 16). Mitochondrial matrix Ca²⁺ concentration ([Ca²⁺]) increase and interaction with poorly defined binding sites on the matrix side of the inner membrane is a specific and almost absolute requirement (one known exception is the phenylarsine oxide effect) for this megachannel opening, also termed mitochondrial permeability transition (MPT; see Ref. 16). It seems likely that, in cardiac cells during reperfusion/reoxygenation, MPT is initiated by Ca²⁺ in conjunction with another agent(s) called Ca²⁺-releasing or -inducing agent(s). In vitro experiments have generated a long list of transition-inducing agents, including various oxidizing and sulfhydril reagents (16). An increasing number of publications suggest that MPT is strongly regulated by oxidation/reduction of mitochondrial thiols (5, 19, 37). This type of regulation could be operative in cardiac cells, keeping in mind that reperfusion/reoxygenation is known to produce a burst of reactive oxygen species (9) generated in mitochondria during resumption of mitochondrial respiration upon reperfusion (2, 29). Reactive oxygen species are likely to act via oxidation of mitochondrial sulfhydryl groups (SH groups) responsible for modulation of MPT (5, 19, 37). In addition, depletion of mitochondrial antioxidants such as glutathione and NADPH also favors MPT (10, 33). Additional support in favor of the key role of MPT in reperfusion/reoxygenation injury comes from experiments on intact cells or isolated heart. These experiments have demonstrated that reperfusion/reoxygenation results in net Ca²⁺ uptake by myocardial cells, contractile dysfunction, sarcolemmal (SL) damage and cell death that are reduced or prevented with inhibition of mitochondrial electron transport (13, 36), and mitochondrial Ca²⁺ uptake via the uniporter with ruthenium red (1, 24, 36). Net Ca²⁺ uptake is associated with an increase in mitochondrial total (12) and free [Ca²⁺] (1, 24, 25), but not with additional increase in cytoplasmic [Ca²⁺] as found during ischemia (24) or hypoxia (25, 32). In addition, the time course of translocation of [³H]deoxyglucose into mitochondria suggests that nonspecific pores are closed during ischemia and open upon reperfusion (15). Cyclosporin A (CSA), which is the most potent inhibitor of MPT currently available, has been shown to partially protect the heart during reperfusion by improving postischemic function (14, 15). However, despite a reasonable correlation between the factors required for MPT in isolated mitochondria and those thought to be responsible for ischemia/reperfusion injury, the role of Ca²⁺-dependent MPT in the development of SL damage has not been investigated in cardiac cells. In Ca²⁺-overloaded cardiomyocytes, mitochondria are expected to accumulate and buffer large amounts of Ca²⁺ before matrix free [Ca²⁺] reaches the level required to open nonspecific pores and release Ca²⁺ into the cytoplasm. If the pore opening, as manifested by rapid Ca²⁺ efflux and sensitivity to CSA, is related to SL damage, then an increase or decrease in a subpopulation of mitochondria with open pores should correspondingly accelerate or slow down SL permeability increase. To test this hypothesis, we developed a technique that allows monitoring of mitochondrial Ca²⁺ uptake/efflux and relation of these changes to SL damage in Ca²⁺-overloaded cardiomyocytes. Selective inhibition of Na⁺-K⁺-ATPase allowed us to generate myoplasmic Ca²⁺ overload, secondary to intracellular Na⁺ concentration ([Na⁺]_i) increase via Na⁺/Ca²⁺ exchange, that was significantly buffered by mitochondrial Ca²⁺ accumulation. During the time of Ca²⁺ accumulation, cellular ATP concentration ([ATP]) was stable but started to decrease when MPT was induced in a large subpopulation of Ca²⁺-overloaded mitochondria. MPT, as characterized by rapid Ca²⁺ efflux and its sensitivity to CSA and dithiothreitol (DTT), was followed by SL damage. Manipulations directed to closure of nonspecific pores significantly decreased SL damage.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Cell culture. Hearts were obtained from neonatal Sprague-Dawley rats (1-2 days old), and cells were isolated as described (30). Depending on the experimental protocol, cells were plated on Falcon dishes, on slides cut from Primaria dishes, on coverslips, or on plastic disks containing scintillant. Cells were used after 3-4 days in primary culture.

The standard perfusate was HEPES-based buffer containing (in mM) 5 HEPES, pH 7.3 (at 37°C), 138 NaCl, 3.6 KCl, 0.3 MgCl₂, 1 CaCl₂, and 16 glucose. Ouabain and DTT were directly added and dissolved in superfusion or incubation buffer. pH was corrected with NaOH after DTT addition. Because of the limited solubility of p-chloromercuribenzoate (pCMB) at pH 7.3, it was dissolved at alkaline pH, and pH was adjusted to 7.3 after addition of pCMB to the buffer. CSA and carbonyl cyanide p-trifluoromethoxyphenyl-hydrazone (FCCP) were dissolved in dimethyl sulfoxide (DMSO). When added to the buffer, DMSO concentration was 0.05% (vol/vol). Net Ca²⁺ uptake, Ca²⁺ content in the mitochondrial compartment, Na⁺/Ca²⁺ exchange-mediated Ca²⁺ efflux, and [Na⁺]_i were measured in thermostated boxes at 37°C. Lactate and lactate dehydrogenase (LDH) release from the cells was determined in a shaking water bath at 37°C. Extracellular buffer volume was 2-3 ml in these experiments. Oxygen content of the buffer was sufficient to avoid any significant increase in lactate production during spontaneous beating of cells, as demonstrated by the stable lactate concentration ([lactate]) level in the extracellular buffer. Basal rate of lactate production was 15 ± 8 nmol · mg protein¹ · min¹ (n = 7). However, inhibition of mitochondrial electron transport with 3 mM NaCN increased lactate production >15-fold in 5 min. [Lactate] was determined by NAD reduction in the presence of 15 units LDH after 90 min of incubation at room temperature (4).

Net Ca²⁺ uptake by the cells was monitored by the scintillation flow cell (chamber volume 7 ml) technique described previously in detail (21). The system allows continuous, on-line recording of cellular ⁴⁵Ca²⁺ activity with rapid change in the perfusion medium at any selected time over a period of hours in intact functional cells. The cells, cultured on scintillant-containing disks, were mounted in a flow cell and perfused with HEPES buffer at a rate of 10 ml/min. Next, the cells were asymptotically labeled with ⁴⁵Ca²⁺ (1 mCi/mmol Ca²⁺) and then exposed to ouabain or pCMB. Under those conditions, any change in the ⁴⁵Ca²⁺ signal away from the established baseline is due to net changes in cellular calcium. At the end of the experiment, dry weight of the cells and specific activity of the perfusate were determined. Changes in cell-associated Ca²⁺ with respect to the control asymptotic value were expressed as millimoles per kilogram of dry weight.

Ca²⁺ uptake by the mitochondrial compartment was determined by kinetic analysis, described in detail previously (20), or by subtracting the fraction of Ca²⁺ uptake inhibited or released with FCCP from the net Ca²⁺ uptake in the absence of FCCP. It had been previously documented that the kinetically defined "slow" compartment in these cells represents mitochondrial Ca²⁺ exchange (21). For kinetic analysis, the cells were labeled with ⁴⁵Ca²⁺ (30 µCi/µmol Ca²⁺) in the flow cell for 35 min at 37°C. Washout was performed with HEPES buffer at a flow rate of 26 ml/min (not rate limiting for mitochondrial exchange) for at least 60 min. After the first washout, which served as a control, the same cells were labeled with ⁴⁵Ca²⁺ for 10-15 min. Cells were then exposed to 250 µM pCMB or 1 mM ouabain, and Ca²⁺ uptake was continuously monitored. Washout was started when Ca²⁺ uptake had reached plateau. The monoexponential part of the washout curve between 30 and 60 min was fitted and extrapolated to time 0. The time 0 intercept of this kinetically defined slow compartment has been demonstrated to represent the ⁴⁵Ca²⁺ content in the mitochondrial compartment (20). Measurement of the total net cellular Ca²⁺ uptake compared with that from the prior ⁴⁵Ca²⁺ labeling permitted calculation of the change in mitochondrial uptake associated with the experimental intervention.

Na⁺/Ca²⁺ exchange-mediated ⁴⁵Ca²⁺ efflux was determined as described in detail (39). Cells grown on a slide cut from a Primaria dish were labeled with ⁴⁵Ca²⁺ (30 µCi/µmol Ca²⁺) for 35 min at 37°C. After this incubation, cells were superfused first with 0 Na⁺, 0 Ca²⁺ HEPES buffer (pH 7.3 adjusted with Tris, NaCl replaced with choline chloride) for 40 s. Na⁺ and Ca²⁺ were returned to the perfusate at the 41st s, and perfusion was continued for another 40 s. A microcomputer delivered perfusate at a rate 3 ml/s in separate pulses. These pulses were collected in separate vials for scintillation counting. When Na⁺ and Ca²⁺ were returned and Na⁺/Ca²⁺ exchange "turned on," a clearly defined increase in ⁴⁵Ca²⁺ activity is recorded in the effluent (see Fig. 4). Integration of this transient increase in ⁴⁵Ca²⁺ gives a measurement of Ca²⁺ efflux specifically dependent on Na⁺/Ca²⁺ exchange (39).

Cytosolic free Na⁺ concentration ([Na⁺]) was measured with the use of sodium-binding benzofuran isophthalate (SBFI) as described (11) with some modifications. Cells were loaded with the AM ester of SBFI for 1.5 h at 37°C. Final concentrations for SBFI and Pluronic F-127 in the loading solution were 15 µM and 0.4% correspondingly. Excitation light was provided at 340 and 380 nm, and fluorescence was measured at 510 nm. In vivo calibration of the SBFI fluorescence 340/380 ratio (F₃₄₀/F₃₈₀) was performed by exposing the cells to various [Na⁺] in calibration buffer (Na⁺ + K⁺ = 150 mM, 5 mM HEPES, pH 7.3, 20 µM gramicidin D, 40 µM monensin, 1 µM FCCP, and 1 mM iodoacetic acid). In these experiments, the F₃₄₀/F₃₈₀ ratio increases in parallel with increasing [Na⁺] up to 50 mM.

Tissue [ATP] was determined by the firefly luciferase method (Bioluminescent somatic cell assay kit; Sigma) following the manufacturer's instruction. Cells were washed with ice-cold buffer and treated with 200 µl of cell-releasing reagent on ice. This suspension was quickly diluted with 400 µl H₂O, and 20 µl were used to measure light emission with a liquid scintillation spectrometer.

LDH activity was determined as described (30) by oxidation of NADH in potassium phosphate buffer (pH 7.65) containing sodium pyruvate and NADH. LDH release during metabolic inhibition was evaluated relative to the total cellular LDH content.

Cellular protein content was determined by the method of Lowry et al. (23).

SBFI and Pluronic F-127 20% solution in DMSO were from Molecular Probes (Eugene, OR). All other reagents were from Sigma.

Data are presented as means ± SD in Figs. 1-7 or as representative tracings with indication of the number of experiments. For differences between mean values, Student's t-test was used. A P value <0.05 was considered significant.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Cellular Ca²⁺ overload was generated by Na⁺ pump inhibition using ouabain or pCMB to increase cellular [Na⁺] and activate "reverse" Na⁺/Ca²⁺ exchange to induce cellular Ca²⁺ overload. Ouabain specifically inhibits Na⁺-K⁺-ATPase and Na⁺ transport out of cells that results in Na⁺ accumulation. pCMB, a specific sulfhydryl group reagent (38), also inhibits Na⁺-K⁺-ATPase activity by blocking functionally important SH groups (18). This is demonstrated in Fig. 1 in which cells exposed to 250 µM pCMB rapidly accumulated Na⁺. Inhibition of Na⁺-K⁺-ATPase with 1 mM ouabain (Fig. 2A) or 250 µM pCMB (Fig. 2B) resulted in a significant net Ca²⁺ uptake that was followed by net Ca²⁺ efflux. An increase in pCMB concentration ([pCMB]) accelerated net Ca²⁺ uptake and also subsequent Ca²⁺ efflux, but in the range of 250-600 µM [pCMB] had no significant effect on the total amount of accumulated Ca²⁺ (Fig. 2C). For reasons that are not totally clear, Ca²⁺ uptake in ouabain-treated cells started after a lag period that was very variable (37 ± 16.8 min, n = 15). However, once started, Ca²⁺ uptake was always much higher compared with pCMB-treated cells.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Effect of 250 µM p-chloromercuribenzoate (pCMB), added at time 0, on intracellular Na+ concentration ([Na+]i); n = 6.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   A: effect of 1 mM ouabain () or 2 µM FCCP plus 1 mM ouabain (), added at time 0, on cellular Ca2+ uptake and subsequent efflux. Representative tracings of 5 and 3 different experiments. In this and Figs. 3-7, the increase in cellular Ca2+ level over the control asymptotic value is expressed as mmol/kg dry wt. [Ca], Ca2+ content. B: effect of 250 µM pCMB () or 2 µM FCCP plus 250 µM pCMB () on cellular Ca2+ uptake and subsequent efflux. FCCP (2 µM) was also added during pCMB-stimulated Ca2+ uptake at the point indicated by arrowhead. Representative tracings of 3 different experiments. C: effect of pCMB concentration ([pCMB]) on cellular Ca2+ uptake and subsequent efflux. Cells were exposed to different concentrations of pCMB at the point indicated by arrowhead. [pCMB] in µM is indicated close to the corresponding tracing. Representative tracings of at least 3 different experiments.

We have also evaluated the effect of 0 K⁺ perfusion and membrane-impermeable p-chloromercuriphenylsulfonic acid (pCMPSA), both known to inhibit Na⁺-K⁺-ATPase. Both treatments induced significant cellular Ca²⁺ uptake that was followed by Ca²⁺ efflux (data not presented). Experiments with pCMPSA also showed that at least some of the functionally important SH groups of Na⁺-K⁺-ATPase protein, which spans the cell membrane, are exposed to the extracellular environment. SH group location on the extracellular face of Na⁺-K⁺-ATPase has been demonstrated by recording the effect of selective application of pCMPSA on Na⁺-K⁺ pump current in guinea pig ventricular myocytes (17), and this location also explains the rapid effect of pCMB (Fig. 1).

To determine mitochondrial Ca²⁺ accumulation that, in the case of normal, functional mitochondria is expected to follow and buffer cytoplasmic Ca²⁺ overload, we used FCCP as a probe in addition to kinetic analysis. FCCP is expected to collapse the proton electrochemical gradient across the mitochondrial inner membrane, which is the driving force for Ca²⁺ uptake via the uniporter. This reduces the gradient of mitochondrial Ca²⁺ to cytoplasmic Ca²⁺ to unity. Exposure of cells to FCCP before either ouabain or pCMB were added significantly inhibited net Ca²⁺ uptake and totally eliminated the late Ca²⁺ efflux (Fig. 2, A and B). In the absence of FCCP, net Ca²⁺ efflux in ouabain and pCMB-treated cells continued to the level approximately corresponding to the net Ca²⁺ uptake in the presence of FCCP. However, in ouabain-treated cells, the rate of Ca²⁺ efflux was rapid (~1 mmol · kg¹ · min¹) initially and then decreased, reaching the level recorded in the presence of FCCP after a time period >60 min (Fig. 2A). The finding that net Ca²⁺ efflux can be accelerated by exposing cells to FCCP after cellular net Ca²⁺ uptake has reached plateau (compare Fig. 2, B and C, trace labeled 250 µM) and cellular [Ca²⁺] decreases to the level reached in pCMB-treated cells in the presence of FCCP (Fig. 2B) suggests that ~75% of net Ca²⁺ efflux is due to mitochondrial Ca²⁺ efflux.

Ca²⁺ uptake by respiring mitochondria was also quantified by kinetic analysis. By monitoring the slow component of cellular ⁴⁵Ca²⁺ exchange, demonstrated to reflect mitochondrial [Ca²⁺] (20, 21), we can show that, in ouabain-treated cells, ~80% of the total cellular ⁴⁵Ca²⁺ content was coming out slowly, with a half-time of 28 min, which is characteristic of the Ca²⁺ efflux from the mitochondrial compartment (Fig. 3A). In this experiment, cellular net Ca²⁺ uptake was 73 mmol/kg dry wt. As expected, kinetic analysis revealed less mitochondrial Ca²⁺ accumulation in pCMB-treated cells (Fig. 3B). However, the values (between 50 and 60%) still indicate significant Ca²⁺ accumulation in the mitochondrial compartment.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Effect of 1 mM ouabain (A) or 250 µM pCMB (B) on 45Ca2+ content in mitochondrial compartment. Cells were labeled with 45Ca2+ in the absence () or presence () of ouabain or pCMB. Washout was started at time 0, and mitochondrial Ca2+ content was determined as described in METHODS.

Ca²⁺ released from the mitochondria into the cytoplasm is removed from the cell via Na⁺/Ca²⁺ exchange according to the thermodynamic equilibrium determined by Na⁺ and Ca²⁺ transmembrane gradients and membrane potential. In this connection, it is important to demonstrate that pCMB treatment had no significant depressive effect on SL Na⁺/Ca²⁺ exchanger function. The results presented in Fig. 4 show that this is the case. First, we demonstrate that the amount of ⁴⁵Ca²⁺ removed from the cells via Na⁺/Ca²⁺ exchange during the first and second washout is reproducible under the same experimental conditions (Fig. 4A) and that 5 mM Ni²⁺ almost totally inhibited Na⁺/Ca²⁺ exchange (Fig. 4B). This result affirms that Na⁺-Ca²⁺-dependent ⁴⁵Ca²⁺ efflux can be used as an indicator of Na⁺/Ca²⁺ exchange activity in intact cells under these conditions. Cells exposed to pCMB during 35 min incubation with ⁴⁵Ca²⁺ demonstrated much higher ⁴⁵Ca²⁺ efflux compared with control during 0 Na⁺, 0 Ca²⁺ superfusion (Fig. 4C). This finding also means that Na⁺/Ca²⁺ exchange-dependent ⁴⁵Ca²⁺ efflux started at a significantly higher level, which makes it difficult to compare ⁴⁵Ca²⁺ transients in the control and pCMB-treated cells. In addition, in Ca²⁺-overloaded cells, Na⁺/Ca²⁺ exchange-mediated ⁴⁵Ca²⁺ efflux did not return to the baseline level characteristic of control but remained significantly higher for the duration of the washout (Fig. 4C). This finding indicates that cellular ⁴⁵Ca²⁺ is continuously transported out of the cell via the Na⁺/Ca²⁺ exchanger in increased quantity. The high level of ⁴⁵Ca²⁺ efflux in 0 Na⁺, 0 Ca²⁺ solution (initial 40 s) most likely represents residual operation of the exchanger in the Ca²⁺/Ca²⁺ exchange mode (35) due to the known presence of contaminant Ca²⁺ (~10 µM) in the "0" Ca²⁺ solution. This is supported by the effect of EGTA on this component of the ⁴⁵Ca²⁺ washout (Fig. 4D). In these experiments, the ⁴⁵Ca²⁺ transient in pCMB-treated cells was significantly higher than that in controls, reflecting the functional response of the exchanger to increased intracellular [Ca²⁺]. The average integrated Na⁺/Ca²⁺ exchange-mediated Ca²⁺ transient of 43,531 ± 7,412 counts per minute (cpm) (n = 3) under control conditions is significantly less than the transient in the same cells treated with pCMB (153,323 ± 16,572 cpm, n = 3).

View larger version (28K): [in this window] [in a new window]   Fig. 4.   Effect of 250 µM pCMB on sarcolemmal (SL) Na+/Ca2+ exchange-mediated Ca2+ efflux. A: two successive washouts were performed in the same cells under control conditions. Integrated Na+/Ca2+ exchange-dependent 45Ca2+ transient in the first washout () was 115,291 cpm and in the second washout () 118,051 cpm. B: SL Na+/Ca2+ exchange-mediated 45Ca2+ efflux is inhibited by Ni2+. During the first washout, cells responded to Na+/Ca2+ exchange activation with a significant Ca2+ efflux (). When the same cells were exposed to 5 mM Ni2+ during 35 min of labeling with 45Ca2+, this response was essentially absent (). In this experiment, Ni2+ was also present in the perfusion buffers. Lower Ca2+ efflux during 0 Na+, 0 Ca2+ superfusion is probably due to the blocking effect of nickel ions on Ca2+ channels during previous 45Ca2+ uptake. C: effect of 250 µM pCMB on SL Na+/Ca2+ exchange-mediated 45Ca2+ efflux. First washout under control condition is indicated by . When cells were labeled with 45Ca2+ in the presence of 250 µM pCMB, 45Ca2+ efflux was significantly higher during 0 Na+, 0 Ca2+ superfusion (). Ca2+ overloaded cells responded to Na+/Ca2+ exchange activation with a transient increase in 45Ca2+ efflux, but maintained a significantly higher level of sustained 45Ca2+ efflux compared with control. Five separate experiments gave similar results. D: in these experiments, contaminant-free Ca2+ was decreased with 1 mM EGTA in 0 Na+, 0 Ca2+ buffer. Under those conditions, Ca2+ efflux during 0 Na+, 0 Ca2+ perfusion in pCMB-treated cells () declined close to control () before Na+/Ca2+ exchange was activated. Na+/Ca2+ exchange-mediated integrated 45Ca2+ transient was higher in cells exposed to pCMB (131,183 cpm) compared with control (44,338 cpm). Similar results were obtained with 3 different preparations.

Having established the extent of mitochondrial Ca²⁺ uptake in cardiomyocytes by using FCCP and kinetic analysis, we next examined the route of mitochondrial net Ca²⁺ efflux, which followed Ca²⁺ uptake as shown in Fig. 2. Three mechanisms exist for mitochondrial Ca²⁺ efflux: Na⁺-independent efflux, Na⁺-dependent efflux, and efflux via nonspecific pores (16). The first two require energy, and in heart mitochondria the Na⁺-dependent Ca²⁺ efflux mechanism via Na⁺/Ca²⁺ exchange dominates under physiological conditions, but its maximum velocity is roughly two orders of magnitude lower than Ca²⁺ uptake via the uniporter (16). In our experiments, Ca²⁺ efflux rates were only slightly lower than Ca²⁺ uptake rates via the uniporter. The rate of Ca²⁺ efflux in pCMB (250 µM)-treated cells was 1.34 ± 0.28 and that of Ca²⁺ uptake was 2.6 ± 0.34 mmol · kg dry wt¹ · min¹ (n = 9). In ouabain-treated cells, the initial rate of Ca²⁺ efflux was also high (1.2 ± 0.4 mmol · kg dry wt¹ · min¹, n = 5). In addition, diltiazem (500 µM), which is expected to partially inhibit mitochondrial Na⁺/Ca²⁺ exchange (8), had no effect on Ca²⁺ efflux (data not presented). Ruthenium red (3 µM), added at the peak of pCMB-stimulated Ca²⁺ uptake, didn't inhibit Ca²⁺ efflux, ruling out Ca²⁺ efflux via reverse uniporter (data not presented). Therefore, it is most likely that the rapid Ca²⁺ efflux is due to the opening of nonspecific pores triggered by an increase in matrix [Ca²⁺]. In pCMB-treated cells, sustained, rapid Ca²⁺ efflux suggests that MPT was initiated in a large subpopulation of mitochondria so that released Ca²⁺ reuptake by still impermeable mitochondria was not significant during Ca²⁺ efflux, whereas in ouabain-treated cells this reuptake could play a significant role in changing the Ca²⁺ efflux kinetics (Fig. 2A).

An important question emerging from these observations is the relationship between MPT, as manifested by rapid Ca²⁺ efflux and SL damage resulting in cell death. Results presented in Fig. 5 show that SL damage is not related to the extent of Ca²⁺ uptake but rather to the rate of Ca²⁺ efflux. Clearly, mitochondrial Ca²⁺ uptake in ouabain-treated cells was high, but Ca²⁺ efflux, particularly in its late phase, was relatively slow (Fig. 2A), and SL damage, as demonstrated by LDH release, developed relatively slowly compared with the effect of pCMB (Fig. 5). Furthermore, LDH release was initiated close to the point where Ca²⁺ efflux is completed in pCMB-treated cells (compare Figs. 2C and 5). To investigate whether the pore opening, as monitored by the rapid mitochondrial Ca²⁺ efflux, is related to plasma membrane damage, we used DTT and CSA, which are expected to inhibit MPT via different mechanisms. Addition of DTT (5 mM) to pCMB-treated cells after they had been incubated with 250 µM pCMB for 30 min, i.e., at the time point where Ca²⁺ accumulation was close to maximum (Fig. 2, B and C), drastically inhibited LDH release. However, because the effect of DTT in pCMB-treated cells could be also connected with partial reactivation of Na⁺-K⁺-ATPase, this experiment was repeated with ouabain-treated cells. First, DTT, added at the plateau of ouabain-stimulated Ca²⁺ uptake, slightly stimulated uptake and completely inhibited Ca²⁺ efflux. However, this effect was transient, lasting ~1 h, and was followed by Ca²⁺ efflux with a rate characteristic to pore opening in a large subpopulation of mitochondria (Fig. 6A). In accordance with this finding, LDH release was delayed but not totally avoided by DTT (data not shown). When cells were exposed to CSA instead of DTT, Ca²⁺ efflux was inhibited but not completely avoided. Because high cytoplasmic [Na⁺] and mitochondrial [Ca²⁺] are expected to stimulate mitochondrial Ca²⁺ efflux via Na⁺/Ca²⁺ exchange, diltiazem was used to inhibit this route of efflux. From Fig. 6A, it is evident that 500 µM diltiazem and 5 µM CSA are unable to block slow Ca²⁺ efflux, but fast Ca²⁺ efflux, seen in ouabain plus DTT-treated cells after a lag period, was avoided. Because diltiazem is not a very potent inhibitor of mitochondrial Na⁺/Ca²⁺ exchange (8), some Ca²⁺ efflux via this route may still take place. Most importantly, SL damage can be significantly reduced by adding 5 µM CSA to ouabain-treated cells at the time point when, in the majority of experiments, mitochondrial Ca²⁺ accumulation had reached a level close to maximum (Fig. 6B). Different from these findings, CSA (0.2-5.0 µM) added at the plateau of pCMB-stimulated Ca²⁺ uptake had no significant effect on Ca²⁺ efflux (data not shown).

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Effect of pCMB or ouabain on SL permeability; 250 µM (, n = 5) or 400 µM (, n = 6) of pCMB or 1 mM of ouabain (, n = 6) was added at time 0, and lactate dehydrogenase (LDH) release was recorded. LDH release was significantly increased over the basal level after 60, 80, and 105 min for these groups correspondingly.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   A: effect of dithiothreitol (DTT) or cyclosporin A (CSA) plus diltiazem on ouabain-induced Ca2+ uptake and subsequent efflux. Either 5 mM DTT () or 5 µM CSA plus 500 µM diltiazem () was added after 1 mM ouabain at point indicated by second arrowhead. Representative tracings of 5 experiments. B: protective effect of CSA on ouabain-induced SL damage. Ouabain (1 mM) was added at time 0, and LDH release was recorded (, n = 8). CSA (5 µM) was added 40 min after ouabain, and LDH release was recorded (, n = 9, when not indicated SD is less than symbol). Effect of CSA on LDH release was significant at times >80 min.

The data presented above indicate that Na⁺-K⁺-ATPase inhibition and cellular [Na⁺] increase initiated the following sequence of events: reverse Na⁺/Ca²⁺ exchange leading to myoplasmic [Ca²⁺] increase that is buffered by mitochondrial Ca²⁺ uptake up to the level where Ca²⁺-dependent MPT is induced and mitochondria release the accumulated Ca²⁺. Because SL damage appears to be related to the permeability transition, it was important to test the energy state of the cell before and after mitochondrial Ca²⁺ overload induced pore opening and Ca²⁺ efflux. For that reason, cellular [ATP] was followed in parallel with that of net Ca²⁺ uptake and efflux in pCMB-treated cells (Fig. 7). No significant change in [ATP] was observed during cellular (mitochondrial) Ca²⁺ uptake, suggesting that this uptake is not induced by metabolic changes other than Na⁺ pump inhibition. However, the [ATP] decrease started at about the time when Ca²⁺ efflux was initiated and [ATP] then decreased in parallel with mitochondrial Ca²⁺ efflux.

View larger version (25K): [in this window] [in a new window]   Fig. 7.   Effect of 250 µM pCMB on net cellular Ca2+ uptake and efflux (, n = 9) and ATP concentration ([ATP]; , each point is a mean ± SD for 4-5 different cell cultures). pCMB was added at time 0. [ATP] decrease starting from 50 min is significant.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Recent experiments have indicated that reperfusion/reoxygenation injury correlates well with mitochondrial Ca²⁺ uptake, although there is no correlation with cytoplasmic [Ca²⁺] (24, 25). These results combined with those presented briefly in the introduction are all consistent with Ca²⁺-dependent MPT playing an important role in the damage that occurs during reperfusion. Mitochondrial matrix [Ca²⁺] required to promote MPT is not exactly known but, in the absence of an inductor, concentrations >100 µM are required to trigger the transition in isolated mitochondria (16). By measuring the rate of passive osmotic contraction of isolated heart mitochondria in response to polyethylene glycol (considered to be proportional to the number of open pores), the apparent Michaelis constant for Ca²⁺ was found to be ~15 µM. A Hill's plot suggests the presence of at least two cooperative binding sites for Ca²⁺ (26). This requirement for high [Ca²⁺] is not unexpected, because mitochondria can accumulate total Ca²⁺ several orders of magnitude more than free [Ca²⁺] (~1 µM; see Refs. 24 and 25) with >99% of mitochondrial Ca²⁺ bound to multiple binding sites. Mitochondrial matrix buffers Ca²⁺ quickly and reversibly with an effective binding ratio of 4,000, which is 40 times the Ca²⁺ binding ratio of the cytoplasm (3). For that reason, any mitochondrial Ca²⁺ uptake or release will simply result in a new equilibrium between bound and free Ca²⁺, and actual changes in mitochondrial free [Ca²⁺] are expected to be relatively small. The finding that significant mitochondrial Ca²⁺ accumulation is required to initiate MPT in ouabain-treated cells also indicates a low Ca²⁺ binding affinity of MPT-related sites. This allows the majority of intramitochondrial binding sites to be saturated before matrix free [Ca²⁺] reaches the level required to initiate MPT. However, the requirement for matrix Ca²⁺ can be modulated by several factors, meaning that pore opening is also possible at very low matrix [Ca²⁺] (28). In the first part of our study, we propose a method for monitoring mitochondrial Ca²⁺ uptake and efflux in intact cardiomyocytes. Although we recorded net cellular Ca²⁺ uptake and efflux, an increase in cytoplasmic [Ca²⁺] due to mitochondrial Ca²⁺ efflux is expected to result in Ca²⁺ efflux via the SL Na⁺/Ca²⁺ exchanger. This has been demonstrated to be the case in our experimental model. There seems to be a gradual increase in mitochondrial Ca²⁺ efflux during pCMB-stimulated Ca²⁺ uptake. This would explain 1) significantly higher mito- chondrial Ca²⁺ uptake in ouabain-treated cells; 2) the finding that the [pCMB] increase from 250 to 600 µM enhanced the rate of Ca²⁺ accumulation and efflux, but had no significant effect on the amount of accumulated Ca²⁺ (Fig. 2C); and 3) stimulation of Ca²⁺ uptake with DTT after Ca²⁺ uptake had ceased (results not shown). In general, SH group reagents are known to induce MPT (16), but pCMB penetration through the membranes is expected to be slow due to its poor lipophilicity, as indicated by the octanol/water distribution coefficient of 0.03 (18). That is expected to delay the pCMB effect on mitochondrial SH groups, and therefore mitochondria would be expected to accumulate significant amounts of Ca²⁺ before a preponderance of Ca²⁺-permeable mitochondria produces net Ca²⁺ efflux. It should be noted that experiments with DTT are complicated by the ability of DTT to reactivate Na⁺-K⁺-ATPase, as demonstrated by intracellular Na⁺ decrease (P. Korge, unpublished results). This recovery of transmembrane Na⁺ gradient results in Ca²⁺ efflux via SL Na⁺/Ca²⁺ exchange, which is expected to decrease membrane damage. However, support for the role of SH groups in the regulation of MPT in Ca²⁺-overloaded cells comes from ouabain-treated cells in which DTT is not expected to reactivate Na⁺-K⁺-ATPase. A transient but significant inhibitory effect of DTT on Ca²⁺ efflux and LDH release in ouabain-treated cells could be connected with reduction of SH groups oxidized by reactive oxygen species produced by the respiratory chain. In fact, ouabain-induced Ca²⁺ overload has been demonstrated to generate reactive oxygen species in rat ventricle (27). Furthermore, exogenous catalase (37) or butylhydroxytoluene (7) effectively reduced Ca²⁺-dependent MPT generated in the absence of oxidizing or SH group reagents in isolated respiring mitochondria. In ouabain-treated cells, mitochondrial Ca²⁺ efflux started at a significantly higher level of mitochondrial [Ca²⁺], and the rate of efflux was slower compared with pCMB-treated cells. This finding is not unexpected because experiments with isolated heart mitochondria have demonstrated that, in the absence of oxidant, high [Ca²⁺] (150 µM) was required to initiate MPT, but oxidant (cumen hydroperoxide) drastically accelerated the MPT even in the presence of relatively low matrix [Ca²⁺] (26). Mitochondrial inhomogeneity, i.e., simultaneous existence of permeable and impermeable fractions, together with slow Ca²⁺ efflux via the Na⁺-dependent pathway obviously complicate the interpretation of MPT based on Ca²⁺ efflux. However, the ability of 5 µM CSA to effectively prevent SL damage in ouabain-treated cells strongly suggests that significant mitochondrial Ca²⁺ accumulation resulted in pore opening that directly or indirectly was related to SL damage. The reason why mitochondrial Ca²⁺ efflux was not significantly inhibited with CSA in the presence of pCMB remains to be determined. CSA is a potent inhibitor of MPT, but its inhibitory activity in isolated mitochondria could be transient and is not always observed (6). The factors involved in promoting pore opening when CSA is present are not fully understood, but high matrix [Ca²⁺] (15, 26), loss of matrix ADP, ATP, and Mg²⁺ (15, 26), free fatty acid accumulation, or production of small amounts of amphipathic peptides by proteases in the matrix space (6) could be involved. In fact, these conditions may occur during the initial phase of reperfusion and may well explain why CSA did not inhibit MPT, as evaluated by mitochondrial [³H]deoxyglucose uptake in reperfused hearts (15). As stated above, pore opening is regulated by the redox state of mitochondrial SH groups, and the presence of oxidants or SH group reagents can significantly decrease or abolish the effect of CSA (6, 26).

As an important consequence of pore opening, cytoplasmic ATP can gain access to mitochondrial ATPase. In our experiments cellular [ATP] was depleted essentially in parallel with Ca²⁺ efflux in pCMB-treated cells. To prevent an increase in entropy and cell death, a certain level of ATP regeneration is required, but a correlation between ATP depletion, MPT, and cell death is not always observed. Experimental evidence suggests that cell death could be related to MPT by a mechanism that is independent of mitochondrial membrane potential and ATP depletion (28).

Because this study was directed to investigate the importance of Ca²⁺-dependent MPT in SL damage, it was mandatory to follow mitochondrial Ca²⁺ uptake and efflux in the intact, functional cell. This requirement also puts certain limitations on our study, but the results clearly demonstrate that interaction between mitochondrial Ca²⁺ movements and their effect on SL permeability can be investigated in intact cells. Under these conditions, monitoring of mitochondrial Ca²⁺ uptake and rapid efflux is likely to be the most sensitive indicator of MPT, whereas CSA offers the most specific probe to investigate the involvement of MPT in SL damage. The results demonstrate the sequence of ionic fluxes leading to mitochondrial Ca²⁺ overload during the period when ATP resynthesis was not compromised. These initial events were followed by permeability transition, characterized by rapid and significant Ca²⁺ efflux into the cytoplasm and ATP decrease. Irreversible damage, as manifested by a significant increase in LDH leakage (Fig. 5), started after mitochondrial Ca²⁺ efflux had essentially ceased and [ATP] had reached a level ~10% of the initial value (Fig. 7). Although the cause of Ca²⁺ overload-induced injury is apparently multifactorial, we believe that an important component of SL damage is initiated by mitochondrial Ca²⁺ efflux, which rapidly generates significant cytoplasmic [Ca²⁺] overload. The last then could lead to destabilization of SL phospholipid and onset of enzyme leakage (31).