Brain mitochondria are primed by moderate Ca2+ rise upon hypoxia/reoxygenation for functional breakdown and morphological disintegration.

In animal models, brain ischemia causes changes in respiratory capacity, mitochondrial morphology, and cytochrome c release from mitochondria as well as a rise in cytosolic Ca2+ concentration. However, the causal relationship of the cellular processes leading to mitochondrial deterioration in brain has not yet been clarified. Here, by applying various techniques, we used isolated rat brain mitochondria to investigate how hypoxia/reoxygenation and nonphysiological Ca2+ concentrations in the low micromolar range affect active (state 3) respiration, membrane permeability, swelling, and morphology of mitochondria. Either transient hypoxia or a micromolar rise in extramitochondrial Ca2+ concentration, given as a single insult alone, slightly decreased active respiration. However, the combination of both insults caused devastating effects. These implied almost complete loss of active respiration, release of both NADH and cytochrome c, and rupture of mitochondria, as shown by electron microscopy. Mitochondrial respiration deteriorated even in the presence of cyclosporin A, documenting that membrane permeabilization occurred independent of mitochondrial permeability transition pore. Ca2+ has to enter the mitochondrial matrix in order to mediate this mitochondrial injury, because blockade of the mitochondrial Ca2+-transport system by ruthenium red in combination with CGP37157 completely prevented damage. Furthermore, protection of respiration from Ca2+-mediated damage by the adenine nucleotide ADP, but not by AMP, during hypoxia/reoxygenation is consistent with the delayed susceptibility of brain mitochondria to prolonged hypoxia, which is observed in vivo.

Stroke commonly results in tissue infarction characterized by necrotic cell death of all cell types within the infarcted area of the brain. From investigations using animal models of stroke (1,2), it has become evident that mitochondria are injured during cerebral ischemia and post-ischemic reperfusion. Ische-mia induces a decrease in the mitochondrial capacity for respiratory activity (1,2).
Brain mitochondria undergo ultrastructural changes after transient and during permanent cerebral ischemia. After transient focal ischemia, cortical neuronal mitochondria become injured, which is manifested by condensation, increased matrix density, and deposits of electron-dense material, finally resulting in disintegration. In contrast, permanent ischemia causes increasing loss of matrix density, associated with mitochondrial swelling, which disappears after 24 h of focal ischemia (3).
A further response to cerebral ischemia is the permeabilization of at least the outer mitochondrial membrane. This results in the liberation of proapoptotic proteins such as cytochrome c, caspase 9, and second mitochondria-derived activator of caspases (4,5). Elevation of cytosolic Ca 2ϩ concentration during ischemia/reperfusion might be a signal for permeabilization of the mitochondrial membrane. In rat CA1 pyramidal neurons of organotypic slices subjected to a hypoxic-hypoglycemic treatment, an increase in cellular calcium was shown (6).
Elevated cytosolic calcium concentrations favor the opening of the mitochondrial permeability transition pore (7,8). In fact, it could be demonstrated that cyclosporin A as well as the non-immunosuppressive analogue N-methyl-Val-4-cyclosporin A, both of which are known to prevent opening of the mitochondrial permeability transition pore, diminished the infarct size, but they could not completely prevent tissue necrosis (9 -11). Thus, other mechanisms besides the opening of the mitochondrial permeability transition pore seem to be involved in tissue damage during ischemia/reperfusion. From in vitro studies using isolated brain mitochondria, it has become clear that the release of cytochrome c does not necessarily require the opening of the mitochondrial permeability transition pore (12,13). Cytochrome c has been identified as an important modulator of death and survival in cells (14).
The outcome for tissue survival after transient ischemia in in vivo models of stroke depends upon the interplay between a variety of factors, such as increased cytosolic Ca 2ϩ concentration, reactive oxygen species, nitric oxide, and substrate supply. Therefore, in vitro models are required in order to delineate the effect of the distinct factors separately from their subsequent interaction. In this context, investigations on isolated brain mitochondria have revealed that Ca 2ϩ accumulation, a hallmark of ischemia/reperfusion, can either inhibit or stimulate H 2 O 2 production (15). Moreover, experiments with isolated mitochondria demonstrated that the BH3 domain of Bax induces the release of cytochrome c from mitochondria (16).
The purpose of the present in vitro study was to use isolated brain mitochondria to elucidate the role played by hypoxia/ reoxygenation and elevated extramitochondrial Ca 2ϩ concentration as signals for permeabilization of the mitochondrial membrane for matrix and intermembrane space proteins. Therefore, we subjected isolated rat brain mitochondria to hypoxia/reoxygenation and/or elevated extramitochondrial Ca 2ϩ concentration and determined respiration, mitochondrial morphology, cytochrome c release, and membrane permeability by applying various complementary techniques.
We found that in hypoxia/reoxygenation an elevated extramitochondrial Ca 2ϩ concentration dramatically enhanced inhibition of active respiration. Moreover, Ca 2ϩ had to enter the mitochondrial matrix to mediate this effect. Under these conditions there was a complete loss of mitochondrial integrity. The presence of ADP, but not AMP, during hypoxia/reoxygenation completely prevented mitochondrial damage. Permeabilization of the mitochondrial membrane did not depend on cyclosporin A, a compound known to keep the mitochondrial permeability transition pore in the closed state.

EXPERIMENTAL PROCEDURES
Materials-Cyclosporin A was purchased from Sigma, cytochrome c from Roche Diagnostics, mouse monoclonal cytochrome c antibody from Pharmingen, and anti-mouse Ig and horseradish peroxidase from Roche Diagnostics. CGP37157 was from Tocris (Cologne, Germany). All other chemicals were of analytical grade.
Preparation of Brain Mitochondria-Mitochondria were prepared from the brains of 220 -240-g male Wistar rats in ice-cold medium containing 250 mM mannitol, 20 mM Tris, 1 mM EGTA, 1 mM EDTA, and 0.3% (w/v) bovine serum albumin at pH 7.4 (isolation medium) using a standard procedure (17). After the initial isolation, Percoll was used for purification of mitochondria from a fraction containing some endoplasmic reticulum, Golgi apparatus, and plasma membranes. The mitochondria were well coupled, as indicated by a respiratory control index greater than 4 with glutamate plus malate as substrates. Protein content was measured according to the method of Bradford (18) using bovine serum albumin as the standard.
Incubation of Mitochondria-Mitochondria (0.5-1.0 mg of protein/ ml) were incubated in a medium containing 10 mM KH 2 PO 4 , 0.5 mM EGTA, 60 mM KCl, 60 mM Tris, 110 mM mannitol, and 1 mM free Mg 2ϩ at pH 7.4 and 30°C. Extramitochondrial calcium was adjusted by using Ca 2ϩ -EGTA buffers. For calculating the concentration of free calcium, we used the complexing constants according to Fabiato and Fabiato (19).
Hypoxia was produced by bubbling 2 ml of the incubation medium with N 2 until oxygen was not detectable any more by means of a Clark-type electrode. Afterward, the mitochondria added to the medium further decreased the oxygen concentration via the respiratory chain. The final oxygen concentration was less than 2 nmol/ml. As the K m value of the mitochondrial cytochrome oxidase is about 0.3 M, the limiting effect of oxygen was illustrated independently by the collapse of the membrane potential measured by means of a TPP ϩ -sensitive electrode in separate experiments (data not shown). A 2-ml volume of air-saturated incubation medium was added to achieve reoxygenation.
Measurement of Respiration-Oxygen uptake of the mitochondria was measured at 30°C in a thermostat-controlled chamber equipped with a Clark-type electrode. For the calibration of the oxygen electrode, the oxygen content of the air-saturated incubation medium was taken to be 217 nmol/ml (20). Immunoblotting for Detection of Cytochrome c-After 10 min of incubation, 2-ml samples of the incubation mixture were centrifuged at 12,000 ϫ g for 10 min at 4°C, and the resulting supernatants were centrifuged at 100,000 ϫ g for 15 min at 4°C. The supernatants were used for Western blot analysis as described by Ghafourifar et al. (21).
Electron Microscopy-For electron microscopy, three independent mitochondrial preparations were used for each incubation strategy. After sedimentation at 320 ϫ g at 4°C, the mitochondrial pellet was fixed with a mixture of 4% formaldehyde (freshly prepared from paraformaldehyde) and 0.4% glutaraldehyde for 1 h at 4°C. Thereafter, the pellet was rinsed thoroughly with phosphate-buffered saline (pH 7.4), postfixed in 1% osmium tetroxide for 1 h at 4°C, dehydrated in a graded series of ethanol, en bloc contrasted with 1% uranyl acetate in 70% ethanol, and flat-embedded between two polyethylene sheets in Curcupan (Fluka/Sigma, Deisenhofen, Germany). Each washing and incubation step was followed by sedimentation at 320 ϫ g at 4°C to collect the mitochondria. Ultrathin sections (50 -70 nm) were prepared with a Leica Ultracut UCT (Bensheim, Germany), mounted on Formvar-coated slot grids, and examined with a Zeiss transmission electron microscope 900 (Oberkochen, Germany).

RESULTS
Influence of Extramitochondrial Ca 2ϩ and Hypoxia/Reoxygenation on Mitochondrial Respiration-To investigate whether elevated extramitochondrial Ca 2ϩ concentration plays a role in the impairment of mitochondrial function during ischemia/reperfusion, respiration of isolated rat brain mitochondria was analyzed. The influence of extramitochondrial Ca 2ϩ alone and of the combination of a rise in Ca 2ϩ with hypoxia/reoxygenation was tested (Fig. 1). First, at three different extramitochondrial Ca 2ϩ concentrations (0, 1.5, and 3.5 M) isolated rat brain mitochondria were subjected to 10 min of hypoxia and 5 min of reoxygenation in the absence of substrates. In general, during hypoxia/reoxygenation no substrates were present. After the addition of 5 mM glutamate plus 5 mM malate, active respiration (state 3) was induced by the addition of 200 M ADP. At 3.5 M extramitochondrial Ca 2ϩ , a decrease of about 30% in comparison with Ca 2ϩ -free incubation, was measured ( Fig. 1A, black bars). The rate of active respiration of freshly isolated mitochondria was determined to be 71.0 Ϯ 4.9 (n ϭ 18) nmol of O 2 ϫ min Ϫ1 ϫ mg Ϫ1 of mitochondrial protein. In Ca 2ϩ -free incubation, 10 min of hypoxia followed by 5 min of reoxygenation caused a significant decrease in active respiration amounting to about 50%. The Ca 2ϩ rise enhanced the inhibition of respiration by hypoxia/ reoxygenation in a dose-dependent manner. In the case of 3.5 M extramitochondrial Ca 2ϩ , active respiration finally decreased down to about 15% of active respiration of freshly isolated mitochondria (Fig. 1A, gray bars).
In the next series of experiments we investigated the influence of the duration of hypoxia on active respiration (state 3) of isolated rat brain mitochondria. Therefore, the time of hypoxia was varied in the absence of extramitochondrial Ca 2ϩ or in the presence of 3.5 M Ca 2ϩ , and then active respiration was measured immediately after reoxygenation. In Ca 2ϩ -free incubation, active respiration decreased within a hypoxic period of 15 min to 58% of initial state 3 respiration (Fig. 1B, black bars). At any hypoxic period, 3.5 M extramitochondrial Ca 2ϩ caused a substantial additional reduction of respiration (Fig. 1B, gray bars).
To investigate whether the deleterious effect on active respiration found after hypoxia/reoxygenation by Ca 2ϩ is exerted from the inside or outside of mitochondria, we repeated the experiments in the presence of ruthenium red in combination with CGP37157. Under these conditions both Ca 2ϩ uptake by the electrogenic uniporter and Ca 2ϩ efflux by the Na ϩ -Ca 2ϩexchanger is blocked. In Fig. 2, oxygen traces after 10 min of hypoxia and 5 min of reoxygenation (trace a), after 10 min of hypoxia and 5 min of reoxygenation in the presence of 3.5 M Ca 2ϩ (trace b), and in the additional presence of 2 M ruthenium red and 25 M CGP37157 (trace c) are presented. 3.5 M extramitochondrial Ca 2ϩ caused a decrease in active respiration (14.6 Ϯ 3.4 (n ϭ 4) versus 21.9 Ϯ 2.8 (n ϭ 4) nmol of O 2 /min/mg of mitochondrial protein). The block of the mitochondrial Ca 2ϩ transport system not only prevented Ca 2ϩinduced decrease in active respiration but also partially protected mitochondria from hypoxia/reoxygenation-mediated damage of active respiration (32.0 Ϯ 4.4 (n ϭ 4) versus 21.9 Ϯ 2.3 (n ϭ 4) nmol of O 2 /min/mg of mitochondrial protein). Thus, protection could be partial because small amounts of endogenous Ca 2ϩ are present during exposure of mitochondria to hypoxia/reoxygenation. In conclusion, Ca 2ϩ exerts its deleterious effect in the mitochondrial matrix.
Adenine nucleotides are present commonly within mammalian cells. In the ischemic phase, ATP is converted first into ADP and subsequently into AMP to maintain energy-consuming processes for as long as possible. In a separate series of experiments, we tested whether extramitochondrial ADP or extramitochondrial AMP affects the decrease in active respiration by hypoxia/reoxygenation. Fig. 3 shows the oxygen traces of a normoxic control (trace a), of mitochondria after 10 min of hypoxia followed by 5 min of reoxygenation in the presence of 3.5 M extramitochondrial Ca 2ϩ (trace b) and in the additional presence of 5 mM ADP (trace c) or 5 mM AMP (trace d). The presence of ADP during hypoxia/reoxygenation almost completely protected brain mitochondria from decrease in active respiration (62.0 Ϯ 5.7 (n ϭ 5) versus 19.2 Ϯ 3.6 (n ϭ 5) and 73.0 Ϯ 4.4 (n ϭ 5) nmol of O 2 /min/mg of mitochondrial protein). Under this condition, the addition of 200 M ADP did not result in any further stimulation of respiration because the ADP added previously was not completely consumed by the mitochondria. In contrast, AMP had no protective effect (11.7 versus 19.2 nmol of O 2 /min/mg).
Mitochondrial Morphology Is Influenced by Extramitochondrial Ca 2ϩ and Hypoxia/Reoxygenation-To elucidate the role of hypoxia/reoxygenation and elevated extramitochondrial Ca 2ϩ concentration on mitochondrial morphology, we performed electron microscopy analyses. The electron micrographs were inspected by two independent investigators blinded to the treatment group. Typical structures corresponding to distinct incubation conditions are presented at a magnification of 1:20,000 in Fig. 4. In each image in Fig. 4 (A-D) an inset illustrates detailed structures at the magnification of 1:30,000. Increasing the extramitochondrial Ca 2ϩ concentration from zero up to 3.5 M did not significantly modify the mitochondrial structure (Fig. 4, B versus A). We were not able to detect changes in the cristae structure. Moreover, intact outer membrane structures were clearly maintained.
Hypoxia/reoxygenation by itself (Fig. 4C) led to several changes in mitochondrial morphology. Besides a population of mitochondria with normal morphology, mitochondria with dented (bleb-like) outer membranes and another population of mitochondria with loss of cristae were found. The combination of 3.5 M extramitochondrial Ca 2ϩ with hypoxia/reoxygenation FIG. 1. Influence of Ca 2؉ rise and hypoxia/reoxygenation on active respiration (state 3) of rat brain mitochondria. Rat brain mitochondria (0.5 mg of protein/ml) were incubated at 30°C in incubation medium. The substrates (5 mM glutamate plus 5 mM malate) were added before oxygen consumption of mitochondria was measured. Control measurements and hypoxia/reoxygenation were performed as described under "Experimental Procedures." Final Ca 2ϩ concentrations were adjusted using Ca 2ϩ -EGTA buffers. Active (state 3) respiration was induced by the addition of 200 M ADP. A, extramitochondrial Ca 2ϩ concentration was varied, and additionally mitochondria were subjected to 10 min of hypoxia and 5 min of reoxygenation. B, zero or 3.5 M Ca 2ϩ was used at various durations of hypoxia. In this case respiration was measured after 1 min of reoxygenation. The respiration of untreated rat brain mitochondria (100%) corresponds to 71 nmol of O 2 min Ϫ1 mg Ϫ1 . Data represent mean values Ϯ S.E. from five preparations of mitochondria. caused dramatic changes in mitochondrial morphology. Mitochondrial integrity nearly completely disappeared (Fig. 4D). Instead, parts of the mitochondrial structures and conglomerates were found (mitochondrial debris). The loss of mitochondrial structure caused by hypoxia/reoxygenation seems to illustrate an irreversible and uncontrolled breakdown of mitochondria.
Mitochondrial disruption was also studied by light scattering experiments. The protocol of typical traces is shown in Fig. 5A. Isolated rat brain mitochondria, which were preincubated at 3.5 M Ca 2ϩ , could be forced to swell by an extramitochondrial Ca 2ϩ stimulus only when Ca 2ϩ reached very high concentrations, such as 200 M (trace 3). This low sensitivity to Ca 2ϩ has already been reported previously (12). In trace 1, the presence of 2 M cyclosporin A prevented mitochondrial swelling after stimulation with 200 M Ca 2ϩ . Cyclosporin A keeps the mitochondrial permeability transition pore in the closed state. Thus, Ca 2ϩ concentrations in the high micromolar range induce opening of the mitochondrial permeability transition pore in intact brain mitochondria. However, in mitochondria that previously had been exposed to 10 min of hypoxia and 5 min of reoxygenation in the continuous presence of 3.5 M Ca 2ϩ , the basal value of absorbance was already increased, most likely because of the condensation seen in the electron micrographs of mitochondria (see Fig. 4). Moreover, we could not find any decrease in absorbance even after adding a stimulus of 200 M Ca 2ϩ (trace 2). This observation confirms our conclusion that mitochondria lose their morphological integrity during hypoxia/reoxygenation in the presence of 3.5 M Ca 2ϩ , because intact brain mitochondria would be able to swell.
Furthermore, the involvement of the permeability transition pore in mitochondrial rupture because of hypoxia/reoxygenation in the presence of Ca 2ϩ was studied by respiration experiments (Fig. 5B). Cyclosporin A, applied at a concentration which was sufficient to prevent intact mitochondria from Ca 2ϩinduced swelling, could not prevent decrease in active respiration. Respiration measurements (Fig. 5B) show that for active respiration of brain mitochondria subjected to substrate-free 10 min of hypoxia and 5 min of reoxygenation in the presence of 3.5 M Ca 2ϩ , no significant difference was found between the values either with or without 2 M cyclosporin A. Thus, we conclude that mitochondrial rupture did not require the opening of the mitochondrial permeability transition pore.
A further series of experiments was performed to test whether mitochondrial constituents are lost during hypoxia/ reoxygenation in the presence of low micromolar Ca 2ϩ , to substantiate our conclusion of mitochondrial disintegration under these conditions. It is well known that the active respiration of intact brain mitochondria is insensitive to extramitochondrial cytochrome c and NADH, because neither of these compounds can permeate through the mitochondrial membrane system. Therefore, we analyzed whether active respiration after 10 min of hypoxia and 5 min of reoxygenation in the presence of 3.5 M Ca 2ϩ becomes sensitive to extramitochondrial cytochrome c and NADH. This approach was used to demonstrate membrane permeabilization. Respiration of intact mitochondria is shown in Fig. 6A. The oxygen consumption after hypoxia/reoxygenation is depicted in Fig. 6B. In intact mitochondria, ADP caused a 4.6-fold increase in respiration (stimulation of active respiration) to 87 Ϯ 5.8 (n ϭ 5) nmol of O 2 /min/mg of mitochondrial protein, whereas after hypoxia/reoxygenation the rate of active respiration was as low as 15.2 Ϯ 1.9 (n ϭ 5) nmol of O 2 /min/mg of mitochondrial protein. The application of 30 M cytochrome c after hypoxia/reoxygenation in the presence of 3.5 M Ca 2ϩ only very moderately accelerated state 3 respiration (19.6 Ϯ 2.2 (n ϭ 5) versus 15.2 Ϯ 1.9 (n ϭ 5) nmol of O 2 /min/mg of mitochondrial protein). In contrast, a nearly 4-fold increase in the rate of oxygen consumption was found in the presence of 5 mM NADH in the incubation medium (55.3 Ϯ 3.3 (n ϭ 5) versus 15.2 Ϯ 1.9 (n ϭ 5) nmol of O 2 /min/mg of mitochondrial protein). This oxygen consumption was not associated with ATP synthesis, because oligomycin was without influence. In intact mitochondria, however, the inhibition of ATP synthesis by oligomycin slowed down the rate of active respiration to resting (state 4) level (18.7 Ϯ 2.3 (n ϭ 5) versus 87 Ϯ 5.8 (n ϭ 5) nmol of O 2 /min/mg of mitochondrial protein).
Stimulation of oxygen consumption by extramitochondrial NADH clearly demonstrates that the inner mitochondrial membrane becomes permeabilized, because under physiological conditions the mitochondrial membrane is impermeable to NADH. The moderate increase in respiration after cytochrome c addition proves that most of the cytochrome c was still associated with parts of the respiratory chain. In fact, Western blot analysis revealed that only 15% of the cytochrome c pool was released into the incubation medium during 10 min of hypoxia and 5 min of reoxygenation in the presence of 3.5 M Ca 2ϩ (data not shown).

Impairment of Mitochondria upon Ischemia/Reperfusion-
Animal models of stroke have revealed that mitochondria are impaired upon ischemia/reperfusion in brain tissue. Decrease in respiratory capacity (1), change in mitochondrial morphology (3), and permeabilization of the mitochondrial membrane system have been reported (4). A further hallmark of ischemic neuronal insults is the disturbance of cellular Ca 2ϩ homeostasis, characterized by increased cytosolic Ca 2ϩ concentration. Elevation of the cytosolic Ca 2ϩ concentration is a well known trigger of mitochondrial damage (6).
Hypoxia/reoxygenation caused impairment of mitochondrial function in cultured astrocytes (22) and isolated mitochondria (23). Here we provide evidence that hypoxia/reoxygenation induces decrease in active respiration and changes in mitochondrial structure, which is characterized by a subpopulation of mitochondria with dented (bleb-like) outer membranes and another subpopulation of mitochondria with diminished number of cristae.
Mitochondria isolated from brain display high resistance to relatively high extramitochondrial Ca 2ϩ concentrations in comparison with mitochondria from other tissues such as liver (13). Also, recently published data from our own studies (12) have shown that Ca 2ϩ in the micromolar range induces only a moderate decrease in active respiration and the release of a small amount of cytochrome c through the outer membrane from brain mitochondria with still intact morphology. Ca 2ϩ has to enter the mitochondrial matrix via the mitochondrial Ca 2ϩ transport system in order to affect the permeability of the mitochondrial membrane. It is known that alterations in the Ca 2ϩ concentration can modify the number of contact sites between the inner and outer membrane, possibly by unmasking nonspecific channels of the mitochondrial outer membrane such as the voltage dependent anion channel (24 -26). This might explain the Ca 2ϩ -mediated change in the permeability of the mitochondrial membrane system. At the cellular level, an increase in cytosolic Ca 2ϩ concentration has been identified as the major cause for mitochondrial impairment, such as in striatal neurons (27) and in hippocampal astrocytes (28).
Here we demonstrate that the application of hypoxia/reoxygenation sensitizes mitochondria to moderately elevated extramitochondrial Ca 2ϩ concentration. Then, even low micromolar Ca 2ϩ causes dramatic functional and morphological changes in isolated brain mitochondria. These include permeabilization and breakdown of the mitochondrial membrane. Again, Ca 2ϩ has to enter the mitochondrial matrix to exert the deleterious FIG. 4. Influence of Ca 2؉ and (or) hypoxia/reoxygenation (hypox/reox) on mitochondrial morphology. Rat brain mitochondria (about 0.5 mg of protein/ml) were incubated at 30°C in the incubation medium without Ca 2ϩ (A), in the presence of 3.5 M Ca 2ϩ (B), exposed to 10 min of hypoxia/5 min of reoxygenation (C), or exposed to 10 min of hypoxia and 5 min of reoxygenation in the presence of 3.5 M Ca 2ϩ (D) and were then fixed and transferred to electron micrograph analysis. Magnification was 1:20,000 and 1:30,000 for each overview and inset photomicrograph, respectively. The electron micrographs were inspected by two independent investigators blinded to the treatment group. The experiment shown is typical for three preparations of brain mitochondria. effect, because blocking the Ca 2ϩ transport system completely protected mitochondria from Ca 2ϩ -mediated damage. Transient hypoxia in combination with Ca 2ϩ elevated into the low micromolar range most appropriately mimics the in vivo situation during ischemia/reperfusion. Interestingly, in our experiments, the impairment of mitochondria was not sensitive to cyclosporin A. In vivo studies, however, using animal models of stroke indicate the involvement of the mitochondrial permeability transition pore (10). Remarkably, the latter seems to be true only for a part of the brain cells localized within the infarct area, because cyclosporin A only diminished the size of the infarct but could not completely prevent necrosis (10,29,30). It still remains unclear which Ca 2ϩ -mediated mechanism is responsible for the cyclosporin A-independent rupture of the mitochondrial membrane.
Investigations on isolated mitochondria reveal that elevated Ca 2ϩ concentration, hypoxia/reoxygenation, or the combination of both treatments clearly induces distinct effects. These observations suggest that, depending on the degree of hypoxia and the level of the cytosolic Ca 2ϩ , mechanisms other than the opening of the mitochondrial permeability transition pore may be involved in the process of mitochondrial damage during ischemia/reperfusion. Impact of Mitochondrial Deterioration after Ischemia/Reperfusion on Brain Cell Fate-A complete breakdown of mitochondrial ATP production is a prerequisite of necrotic cell death in brain (31). ATP is required for the maintenance of cellular morphology, ion homeostasis, protein synthesis, and many other cellular functions. Moreover, ATP is even necessary to perform the apoptotic cell death program (32). Another factor in inducing neuronal demise is the change of the permeability FIG. 5. Effect of cyclosporin A, extramitochondrial Ca 2؉ , and hypoxia/reoxygenation (hypox/reox) on mitochondrial volume (A) and on active respiration (B). About 0.5 mg of protein/ml of rat brain mitochondria was incubated at 30°C in the incubation medium. A, absorbance was followed at 546 nm. CaCl 2 (200 M) was added, indicated by the arrow, to the three samples as a stimulus to induce mitochondrial swelling. Sample 1 was incubated with 2 M cyclosporin at 3.5 M Ca 2ϩ , sample 2 was exposed to 10 min of hypoxia/5 min of reoxygenation in the presence of 3.5 M Ca 2ϩ , and sample 3 was incubated in the presence of 3.5 M Ca 2ϩ . The experiment shown is typical for five preparations of brain mitochondria. B, active respiration was determined in the presence of 5 mM glutamate plus 5 mM malate upon the addition of 200 M ADP. 10 min of hypoxia/5 min of reoxygenation (H/R) was performed as described under "Experimental Procedures." The additions were: Ca 2ϩ , Ca 2ϩ -EGTA buffer with 3.5 M Ca 2ϩ final concentration; CSA, 2 M cyclosporin A. Data of respiration are given in % active respiration of freshly isolated rat brain mitochondria (100% correspond to 71 nmol of O 2 min Ϫ1 mg Ϫ1 ). Mean values of active respiration Ϯ S.E. of five separate preparations of rat brain mitochondria are presented. of the mitochondrial membrane. Depending on the mode of activation, permeabilization can cause either apoptosis or necrosis. Distinct permeabilization of the mitochondrial outer membrane can be achieved in brain mitochondria by increased cytosolic Ca 2ϩ concentrations or by members of the Bcl-2 family in cooperation with cardiolipin (33). This type of permeabilization causes a partial release of proapoptotic factors such as cytochrome c (12,13). The members of the Bcl-2 family also interact with the mitochondrial permeability transition pore (34). Reversible permeabilization of the mitochondrial membrane by opening of the permeability transition pore also causes the release of proapoptotic factors from mitochondria. If sufficient ATP is available within the cell, apoptosis is initiated. In fact, signs of apoptosis and cyclosporin A sensitivity in brain injury have been demonstrated in animal models of stroke (35). In contrast, the disruption of the mitochondrial membrane causes necrotic cell death (36), which can be induced by permanent opening of the permeability transition pore by Ca 2ϩ overload (37,38) or by massive lipid peroxidation (39). Our experiments demonstrate that extramitochondrial Ca 2ϩ concentration in the low micromolar range leads to cyclosporin A-insensitive (e.g. permeability transition pore-independent) disruption of the mitochondrial membrane.
Isolated mitochondria can be maintained under conditions of almost complete anoxia (23), whereas in vivo, some oxygen may diffuse from the environment into the infarct area. Both the in vivo studies of stroke and the cell culture investigation on hypoxia/reoxygenation require a relatively long period of hypoxia to reach significant injury (2). However, in isolated brain mitochondria only a few minutes of hypoxia are sufficient to cause dramatic damage. Differences in local oxygen concentration may be the reason for this apparent discrepancy between in vivo and isolated mitochondria experiments in the time required to reach injury. We have found that at elevated extramitochondrial Ca 2ϩ concentrations, ADP at physiological concentration (5 mM) protects mitochondria from hypoxia/ reoxygenation-induced damage. Only when all of the ADP is converted into AMP, mitochondrial damage occurs. This finding may contribute further to the fact that longer periods of ischemia are required to achieve tissue damage in comparison with isolated mitochondria that have to be exposed only for a short period of time to hypoxia in order to induce mitochondrial damage.
Pathological Consequences in Brain Mitochondria Caused by Ischemia/Reperfusion-From both the present data and the observations obtained with animal models of stroke, we conclude that three qualitatively different situations of stroke injury have to be distinguished. (i) Increased cytosolic Ca 2ϩ concentrations into the low micromolar range induce permeabilization of the mitochondrial outer membrane and release of proapoptotic factors such as cytochrome c, resulting in the induction of apoptosis. (ii) Short periods of hypoxia or moderate elevation of the extramitochondrial Ca 2ϩ concentration cause reversible opening of the mitochondrial permeability transition pore followed by the release of proapoptotic factors from morphologically intact mitochondria, also resulting in the induction of apoptosis. (iii) Long-lasting hypoxia, followed by reoxygenation at moderately elevated cytosolic Ca 2ϩ concentration, leads to permanent opening of the permeability transition pore or the permeability transition pore-independent disruption of the mitochondrial membrane. The resulting failure of mitochondrial energy metabolism causes necrotic cell death.
In conclusion, we provide evidence that isolated rat brain mitochondria are highly vulnerable to hypoxia/reoxygenation applied in combination with the rise of the extramitochondrial Ca 2ϩ concentration into the low micromolar range. However, each stimulus, Ca 2ϩ or hypoxia/reoxygenation, applied individually can be well tolerated by brain mitochondria. Thus, our data explain the apparent discrepancy between in vivo and in vitro data concerning the harm exerted by cytosolic Ca 2ϩ concentrations in the low micromolar range. When observed in vitro, high resistance of mitochondria is seen, whereas in vivo, upon ischemia/reperfusion cell death occurs. Therefore, in cells of the nervous tissue therapeutic concepts aimed at preventing neural damage after ischemia (stroke) should focus on the prevention of pathological elevations of cytosolic Ca 2ϩ .