Reactive oxygen species induce swelling and cytochrome c release but not transmembrane depolarization in isolated rat brain mitochondria

Abstract

1. In this study, we have used isolated brain mitochondria to investigate the effects of superoxide anions (O₂⁻) on mitochondrial parameters related to apoptosis, such as swelling, potential, enzymatic activity, NAD(P)H, cytochrome c release, and caspase activity.

2. Addition of the reactive oxygen species (ROS) generator KO₂ produced brain mitochondrial swelling, which was blocked by cyclosporin A (CSA), and which was Ca²⁺ independent.

3. Calcium induced mitochondrial swelling only at high concentrations and in the presence of succinate. This correlated with the increase in O₂⁻ production detected with hydroethidine in mitochondrial preparations exposed to Ca²⁺ and the fact that ROS were required for Ca²⁺-induced mitochondrial swelling.

4. Superoxide anions, but not Ca²⁺, decreased citrate synthase and dehydrogenase enzymatic activities and dropped total mitochondrial NAD(P)H levels.

5. Calcium, but not O₂⁻, triggered a rapid loss of mitochondrial potential. Calcium-induced Δψ_m dissipation was inhibited by Ruthenium Red, but not by CSA.

6. Calcium- and superoxide-induced mitochondrial swelling released cytochrome c and increased caspase activity from isolated mitochondria in a CS A-sensitive manner.

7. In summary, superoxide potently triggers mitochondrial swelling and the release of proteins involved in activation of postmitochondrial apoptotic pathways in the absence of mitochondrial depolarization.

Introduction

Mitochondrial dysfunctions, because of DNA defects, protein mutations or alterations in electron transport (Koutnikova et al., 1997) have been linked to neurodegenerative diseases (Sheehan et al., 1997; Brown et al., 2000) and aging processes (Wallace, 2001). Mitochondria play a central role in different cell death pathways leading to either apoptosis or necrosis (Green & Reed, 1998; Kroemer & Reed, 2000). The first evidence that mitochondria play an important role in cellular death mechanisms came from the observation that ‘in vitro' models required an organelle fraction enriched in mitochondria to induce apoptosis (Newmeyer et al., 1994). Apoptosis might involve changes in respiration-dependent mitochondrial potential (Δψ_m) and, in some models, formation of a voltage-dependent high conductance multiproteic channel, referred to as mitochondrial permeability transitory pore (MPTP) (Bernardi, 1999). However, there is increasing evidence suggesting that MPTP activation is also important for necrosis activation (Kroemer & Reed, 2000).

The Δψ_m role in apoptosis is still an enigma, as either preservation (Prehn et al., 1996) or dissipation (Stout et al., 1998) has been shown to block neuronal death. Moreover, the relation between MPTP and Δψ_m remains unclear. There are reports suggesting that for some apoptotical stimuli, MPTP formation is independent of mitochondrial transmembrane depolarization (Madesh & Hajnoczky, 2001). As a result of MPTP opening, mitochondria might lose their membrane permeability control, releasing internal proteins (cytochrome c, apoptosis-inducing factor (AIF), SMAC/Diablo, caspase family members, etc.) to the cytoplasm. This release results in the activation of different pathways that lead to cell death. For example, cytochrome c participates in the apoptosome formation, and AIF results in chromatin condensation and DNA fragmentation (Susin et al., 1999). Although the physiological role of MPTP has not yet been determined, its role as mediator of cell injury and death is generally accepted.

Calcium influx and O₂⁻ production within the mitochondria prior to cell death are events shared by cells treated with neurotoxic agents like NMDA, veratridine and β-amyloid (Bindokas et al., 1996; Jordan et al., 2000, 2002; Longo et al., 2000). Overexpression of proteins that regulate Ca²⁺ and O₂⁻ levels protect neuronal cultures from neurotoxic stimuli (Jordan et al., 1995; Prehn et al., 1997). Evidence has been presented suggesting that intracellular O₂⁻ production during apoptosis injury is Ca²⁺ dependent and from mitochondrial origin (Ankarcrona et al., 1995; Bindokas et al., 1996; Jordan et al., 2000) and either stimuli result in MPTP formation (for review, see Bernardi, 1999). In the present study, we have analyzed the plausible effects of O₂⁻ and Ca²⁺ on mitochondrial swelling and Δψ_m dissipation using isolated rat brain mitochondria.

Methods

Mitochondria isolation

Mitochondria were isolated from the cortex of adult Sprague–Dawley rats by conventional centrifugation as previously described (Jordán et al., 2002). Briefly, tissue was manually homogenized with four strokes with a Teflon pestle in solution I (Sol I) containing 230 mM mannitol, 70 mM sucrose, 1 mM EGTA, and 5 mM HEPES, pH 7.4 on ice. After centrifugation at 4°C (106 × g; 80 s), supernatant was layered onto solution II (460 mM mannitol, 14 mM sucrose, 1 mM EGTA, and 10 mM HEPES, pH 7.4), and centrifuged (800 × g; 3 min). The top layer was then centrifuged (2000 × g; 5 min) and the mitocondrial pellet resuspended in solution III (215 mM mannitol, 71 mM sucrose, 10 mM succinate, and 10 mM HEPES, pH 7.4) and kept on ice until swelling determinations. To exclude that the observed effects were becuase of contaminating synaptosome, we isolated brain mitochondria using a Percoll gradient as previously described (Kristal & Dubinsky, 1997). No differences in KO₂-induced swelling were observed between mitochondria isolated by both the methods.

Permeability transition pore activity

Permeability transition pore opening was assayed spectrophotometrically as previously described (Kristal et al., 2000; Galindo et al., 2003). Specifically, mitochondria were suspended to reach a protein concentration of 1 mg ml⁻¹ in 200 μl of solution III. Changes in absorbance at 540 nm (A₅₄₀), indicating mitochondrial swelling because of MPTP opening, were followed, after addition of different compounds, using a microplate reader (BioRad, Hercules, CA, U.S.A.). Initial A₅₄₀ values were ≅0.8, and minor differences in loading were compensated by representing the data as the fraction of the initial absorbance measure remaining at a given time. Mitochondrial protein concentrations were quantified spectophotometrically (Micro BCA Protein Reagent Kit, Pierce, Rockford, IL, U.S.A.), with bovine serum albumin as standard.

Monitoring of the mitochondrial membrane potential

Changes in mitochondrial membrane potential were measured in the presence of 0.1 μg ml⁻¹. rhodamine 123 (Rh 123) as described by Emaus et al. (1996). The excitation and emission wavelengths for Rh 123 were 503 and 525 nm (slit 3 nm), respectively, using a Perkin-Elmer (luminiscence-spectrophotometer LS50B) fluorimeter.

Assay for mitochondrial NAD(P)H levels

NAD(P)H fluorescence in intact mitochondria (1 mg ml⁻¹ at 25°C) was measured spectrofluorimetrically with excitation and emission wavelengths of 340 nm (slit 3 nm) and 460 nm (slit 5 nm), respectively, as previously described (Rover Junior et al., 1998). We therefore refer to NAD(P)H, indicating the signal derived from either NADH or NADPH, or both. Under these conditions, an increase in autofluorescence signal indicates an increase in the reduced state of the pyridine nucleotide, NAD(P)H, and a decrease in autofluorescence signal indicates an increased oxidation to NAD(P)⁺.

Mitochondrial dehydrogenase activity

Mitochondrial dehydrogenase activity was estimated by the reduction of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide thiazolyl) blue (MTT) to its corresponding blue formazan product according to a procedure described elsewhere (Mosmann, 1983). After 30 min incubation of mitochondrial suspension with Ca²⁺ or O₂⁻, mitochondria were further incubated for 15 min with 0.5 mg ml⁻¹ MTT. At the end of the incubation period, absorbances were read on a microplate reader (BioRad, Hercules, CA, U.S.A.) with a test wavelength of 540 nm and a reference wavelength of 630 nm.

Citrate synthase activity

Citrate synthase (EC 4.1. 3.7) was measured at 410 nm using a microplate reader (BioRad, Hercules, CA, U.S.A.) at room temperature as described by Shepherd & Garland (1969). The reaction buffer contains 100 mM Tris-HCl, pH 8, 0.1 mM acetyl-CoA, 0.2 mM DTNB, 0.1 (V V⁻¹) Triton X-100, and 500 μg protein of mitochondrial suspension in a final volume of 0.2 ml. The assay is initiated by the addition of 10 μl of 20 mM oxalacetate and 10 min later absorbance was read on a microplate reader (BioRad, Hercules, CA, U.S.A.).

Immunoblot analysis

Immunoblot analysis was performed on extramitochondrial extracts from control and treated mitochondrial suspensions. In all cases, mitochondria suspensions were treated with resuspension (control), CaCl₂, or KO₂ for 30 min and then centrifuged at 15,000 rpm for 15 min. The supernatants (S100), extramitochondrial fractions, were precipitated by trichloroacetic acid (TCA) (10%, 4°C) overnight and centrifuged (15,000 × g;15 min). Pellets were resuspended in 40 μl loading buffer and boiled for 15 min. Samples were loaded onto a 15% SDS-polyacrylamide gel, separated and transferred to a PVDF membrane, which was incubated with mouse monoclonal anticytochrome c antibody (Santa Cruz Biotechnology Inc., U.S.A.; 1 : 1000 dilution). The secondary antibody used (1 : 5000 dilution) was a peroxidase-labeled anti-mouse (Promega, Madison, WI., U.S.A.). The signal was detected using an enhanced chemiluminescence detection kit (Amersham ECL RPN 2106 Kit). Ruthenium Red (5 μM) or cyclosporin A (CsA) (10 μM) was added 15 min before and maintained during CaCl₂ or KO₂ treatment.

Assay of caspase enzymatic activity

Mitochondrial suspension aliquots were exposed for 30 min to CaCl₂ or KO₂ and then centrifuged at 15,000 × g for 15 min. Supernatants, lacking mitochondria, were collected and incubated at 37°C in a buffer containing 25 mM HEPES (pH 7.5), 10% sucrose, 0.1 CHAPS and 10 mM DTT with the fluorogenic substrate N-acetyl-Asp-Glu-Val-Asp-7-amido-4- trifluoromethylcoumarin (DEVD-AFC) (15 μM in DMSO, Calbiochem System Products) (Jordan et al., 1997). Substrate cleavage emitted a fluorescent signal that was quantified in a fluorometer Perkin-Elmer (luminiscence-spectrophotometer LS50B) (excitation 400 nm, emission 505 nm). Enzymatic activity is expressed as arbitrary fluorescent units (AFU). Ruthenium Red (5 μM) or CsA (10 μM) was added 15 min before and maintained during CaCl₂ or KO₂ treatment.

Statistical analysis

Quantitative data are expressed as means±s.e.m. and significance was determined by Student's t-test. Statistical significance was considered at the P<0.05 level.

Results

Superoxide induces brain mitochondrial swelling in a Ca²⁺-independent manner

Induction of mitochondrial swelling in isolated brain mitochondria was monitored by following 540 nm absorbance (A₅₄₀) decrease (Bernardi et al., 1999). We used KO₂ as a source of O₂⁻ anions and corresponding degraded ROS products. KO₂ addition (5 μmol mg⁻¹ mit) resulted in an extensive mitochondrial swelling (Figure 1a) that was greatly inhibited by the pretreatment and presence of 10 μM CsA, a MPTP blocker, but not by the calcineurin inhibitor FK506 (200 nM), (Figure 1b). Addition of KCl (5 mol mg⁻¹ mit) did not result in mitochondrial swelling, confirming that KO₂-induced mitochondrial swelling was because of reactive reactive oxygen species (ROS) and not to potassium (Figure 1a).

Figure 1.

KO₂ and high Ca²⁺ induce mitochondrial swelling. A–B. KO₂ addition (5 μmol mg⁻¹ protein; arrow), induces mitochondrial swelling in brain-isolated mitochondria in an MPTP sensitive manner, where Ca²⁺ cycling is not involved. Mitochondrial suspensions were incubated during 15 min with CsA (10 μM), FK 506 (200 nM), Ruthenium Red (RR; 5 μM), or EGTA (1 mM), prior to addition of KO₂. Control (c) and the effect of KCl (5 mM) addition on non-treated mitochondria are also shown. C–D. Calcium addition (15 μmol mg⁻¹ mitochondria) induces mitochondrial swelling in brain-isolated mitochondria. Mitochondrial suspensions were incubated during 15 min with CsA (10 μM), Ruthenium Red (RR;5 μM), or EGTA (1 mM) before Ca²⁺ addition. Control non-treated mitochondria (c) are also shown. E–F. Ca²⁺-induced swelling is ROS and succinate dependent. In absence of succinate (0 Succ), Ca²⁺ failed to induce mitochondrial swelling. Mitochondria were preincubated for 15 min with MnTBAP (10 nM), catalase (Cat;1750 U) or SOD (450 U), prior to Ca²⁺ addition (15 μmol mg⁻¹ mit). Control nontreated mitochondria (c) are also shown. Lines represent mean values of one experiment performed by triplicate. Histograms (panels b, d, f) represent mean values±s.e.m. of normalized A₅₄₀ at 2200 s from at least five different mitochondrial preparations. ^***P<0.001;^**P<0.01;^*P<0.05 as compared with KO₂ or Ca²⁺-treated mitochondria in absence of the drug.

Since a variety of chemically different pro-oxidants have been shown to induce Ca²⁺ release from mitochondria and the fact that Ca²⁺ induces mitochondrial swelling in liver mitochondria, we next investigated the existence of a positive feedback between O₂⁻ anions and Ca²⁺ in mitochondrial swelling. Blockade of Ca²⁺ influx, by the presence of the mitochondrial Ca²⁺ uptake blocker, Ruthenium Red, or by addition of the Ca²⁺ chelator EGTA (1 mM) did not affect KO₂-induced mitochondrial swelling, suggesting that Ca²⁺ does not seem to play a significant role in KO₂-induced mitochondrial swelling (Figure 1c). In contrast to noncerebral tissue, only relatively high Ca²⁺ concentrations induced brain mitochondrial swelling and required the presence of succinate in the mitochondrial resuspension medium, because Ca²⁺ failed to induce mitochondrial swelling in its absence (0 Succ, Figure 1c), while KO₂ did not (data not show). Even at these high Ca²⁺ concentrations, the Ca²⁺ Ca²⁺-induced mitochondrial swelling was blocked by Ruthenium Red, EGTA, and CsA (Figure 1d). Lower Ca²⁺ concentrations did not produce significant swelling (data not shown), although we cannot exclude that mitochondrial swelling may even be underestimated, as some heavy microsomal membranes could not be entirely separated from mitochondria.

The succinate requirement for Ca²⁺ to induce mitochondrial swelling suggested a role for coenzyme Q as a significant generator of O₂⁻ (Kowaltowski et al., 1995). To test if O₂⁻ production takes place, in isolated mitochondria, during Ca²⁺ exposures, selective oxidation of hydroethidine, by O₂⁻, to ethidium was measured. Brain mitochondria showed a basal O₂⁻ production rate of 0.0426±0.0072 AFU min⁻¹ (n=16). Calcium (15 μmol mg⁻¹ mit) caused an increase in the rate of O₂⁻ production reaching a value of 0.0675±0.0153 AFU min⁻¹ (P<0.05; n=16). To check whether this O₂⁻, or its derived ROS, produced by Ca²⁺ in mitochondria, participated in Ca²⁺-induced mitochondrial swelling, we tested the effects of different free radical scavengers. As shown in Figure 1e, MnTBAP, a manganese porphyrin with SOD mimic activity, or catalase addition (1750 U ml⁻¹) significantly attenuated Ca²⁺-induced mitochondrial swelling, while addition of SOD (450 U ml⁻¹) did not block significantly Ca²⁺-induced swelling (Figure 1f). Consistently with the utilization of KO₂ as a ROS source, the addition of antioxidant drugs decreased significantly KO₂-induced mitochondrial swelling (data not shown).

KO₂-induced mitochondrial swelling occurs in absence of Δψ_m depolarization

One of the most intriguing questions that remains unanswered is whether mitochondrial swelling produces Δψ_m depolarization or if Δψ_m dissipation precedes mitochondrial swelling. In the next set of experiments, we studied whether Ca²⁺ or O₂⁻ was able to modify Δψ_m. We monitored Δψ_m changes by measuring the release of the cationic membrane-permeant fluorescent probe Rh 123 preloaded into mitochondria. Under these conditions, total fluorescence of mitochondrial suspension will increase if the organelles depolarize. As shown in Figure 2, Ca²⁺, but not KO₂, induced mitochondrial depolarization. Calcium concentrations required to evoke Δψ_m depolarization were 20-fold lower than those necessary to induce mitochondrial swelling. Mitochondrial depolarization was dependent of Ca²⁺ influx, but independent of MPTP opening, since it was blocked by 15 min pretreatment with Ruthenium Red (5 μM) but not by CsA (10 μM) (Figure 2a). Nevertheless, KO₂ additions failed to influence Δψ_m within the same concentration range that caused mitochondrial swelling (Figure 2b).

Figure 2.

Changes in mitochondrial membrane potential as assessed by Rh 123. Fluorescence excited at 503 nm and emitted at 527 nm was measured in isolated mitochondria. (a) Addition of mitochondria (Mit, 500 μg) caused a decrease in fluorescence intensity because of quenching of Rh 123. CaCl₂ (0.75 μmol mg⁻¹ mit) was added under control conditions (trace c). In traces RR and CsA, mitochondria were incubated with Ruthenium Red (5 μM, trace RR) or CsA (10 μM, trace CsA) for 15 min and then added to Rh 123 solution. (b) The protocol was similar, but two KO₂ (2.5 and 5 μmol mg⁻¹ mit, continuous line) sequential pulses were added. Calcium was added to the incubation buffer to ensure that the mitochondria could be depolarized. Data are expressed as mean values obtained from one experiment performed in triplicate. Similar data were found in at least four different experiments. FCCP (0.1 μM) was added to induce mitochondrial depolarization. Dashed line represents control experiments performed by adding mitochondrial resuspension buffer (RB) and KO₂ solutions in absence of brain mitochondria.

Effects of high Ca²⁺ and O₂⁻ on mitochondrial enzymatic activity and NAD(P)H levels

The next set of experiments were performed to analyze the possible effects of Ca²⁺ and O₂⁻ on mitochondrial metabolic enzymatic activity.

Mitochondrial dehydrogenase activity was measured by determining the formation of MTT into formazan. Under our experimental conditions, control mitochondrial suspension (500 μg protein) had an enzymatic activity of 0.165±0.036 (n=4) and was not affected by the treatment with Ca²⁺ (15 μmol mg⁻¹ mit) for 30 min (0.145 × 0.034; n=4). However, 30 min incubation of intact mitochondria with KO₂ (5 μmol mg⁻¹ mit) produced a marked decrease in dehydrogenase activity (0.052±0.016; P<0.001; n=4).

Furthermore, the Krebs cycle enzyme citrate synthase was also affected by the presence of O₂⁻. Mitochondrial citrate synthase was measured by determining the reaction of free CoA with DTNB. Incubation of brain mitochondria (500 μg protein) with KO₂ (5 μmol mg⁻¹ mit) for 30 min, did result in a loss in enzymatic activity (0.175±0.015 (P<0.001; n=5)) compared to control mitochondrial suspension (0.356±0.026; n=5), while Ca²⁺ (15 μmol mg⁻¹ mit; 30 min) did not alter citrate synthase activity (0.3494±0.037; n=5).

Recently, NAD(P)H has been proposed to act as a direct antioxidant agent (Kirsch & De Groot, 2001). Therefore, we were interested in whether the NAD(P)H levels were modified by these treatments when using either O₂⁻ or Ca²⁺. As Figure 3 illustrates, Ca²⁺ (15 μmol mg⁻¹ mit) increases NAD(P)H levels in a Ruthenium Red (5 μM, 15 min)-sensitive manner, while, following KO₂ addition (5 μmol mg⁻¹ mit), an initial rapid decline in mitochondrial NAD(P)H concentration was observed. Rotenone, a mitochondrial respiratory chain complex I inhibitor (10 μM, Rot) and the mitocondrial uncoupler, FCCP (1 μM, FCCP) were added to estimate the maximal and the minimum nucleotide signal, respectively (Figure 3).

Figure 3.

Effect of KO₂ and Ca²⁺ on mitochondrial NAD(P)H levels. NAD(P)H levels were monitored spectrofluorimetrically as described in methods. After 60 s of base line, intact mitochondria (500 μg) were added. Arrow indicates CaCl₂ (15 μmol mg⁻¹ mit; Ca²⁺), KO₂ (5 μmol mg⁻¹ mit; KO₂), rotenone (10 μM, Rot), and FCCP (1 μM, FCCP) additions. Dashed line represents mitochondria preparation preincubated for 15 min in the presence of Ruthenium Red (5 μM; RR) before CaCl₂ addition. One single experiment is shown, but similar data were found in four to six different mitochondrial preparations.

O₂⁻ induces release of cytochrome c and caspase activity from mitochondria

A consequence of mitochondrial swelling is the release of intramitochondrial proteins like cytochrome c to the cytoplasm (Maestre et al., 2003). In order to confirm the occurrence of this intramitochondrial protein release in our model, we exposed isolated brain mitochondria for 30 min to CaCl₂ (15 μmol mg⁻¹ mit) or KO₂ (5 μmol mg⁻¹ mit) and tested for cytochrome c presence and DEVD-like caspase activity in the extramitochondrial medium. As shown in Figure 4a, direct evidence for cytochrome c release from rat brain mitochondria was provided by Western blotting technique of cytochrome c in the S100 supernatant, after treating mitochondria with either Ca²⁺ or O₂⁻. We next examined whether DEVD caspase activity was released from isolated mitochondria following exposure to Ca²⁺ or KO₂. After these treatments, we performed measurements of the cleavage of a fluorometric caspase substrate, DEVD-AFC (Figure 4b). Increases in DEVD-caspase activity were found in S100 after Ca²⁺ and O₂⁻ additions (Figure 4b). It is worthy noting that in both assays, O₂⁻ effect was higher than Ca²⁺, perhaps because of the the fact that Ca²⁺ might inhibit caspase activity, as seems to happen in Ca²⁺-mediated neuronal death (Reimertz et al., 2001). Consistently with the mitochondrial swelling results, this effect was abolished by CsA, whereas Ruthenium Red was only effective against Ca²⁺ treatments (Figure 4b).

Figure 4.

Ca²⁺ and O₂⁻ addition induced release of cytochrome c and caspase activity from mitochondria. (a) Mitochondria were incubated during 30 min with resuspension (control), CaCl₂ (15 μmol mg⁻¹ mit), or KO₂ (5 μmol mg⁻¹ mit) and then centrifuged for 15 min at 15,000 rpm at 4°C (S100; see Methods section). For pharmacological studies, mitochondrial suspensions were pretreated for 15 min with Ruthenium Red (RR, 5 μM) or CsA (10 μM) before and during exposures. (a) Representative gel of S100 protein precipitated overnight with 10% TCA is shown. Pellets were subjected to polyacrylamide gel electrophoresis and immunoblot analysis using an antibody that recognizes cytochrome c. (b) S100 supernatant was assayed for caspase activity for 30 min. Data are represented as mean s.d. (n=4). ^*P<0.05 as compared with respective controls (vehicle).

Discussion

Changes in Ca²⁺ influx and O₂⁻ production within the mitochondria prior to cell death are events shared by cells treated with neurotoxic agents such as NMDA, veratridine and β-amyloid (Bindokas et al., 1996; Jordan et al., 2000; Longo et al., 2000). Evidence has been presented that intracellular O₂⁻ production during apoptosis is Ca²⁺ dependent and of mitochondrial origin (Ankarcrona et al., 1995; Bindokas et al., 1996; Jordan et al., 2000). Herein, we used brain-isolated mitochondria to study the effects of Ca²⁺ and O₂⁻ on organelle parameters such as swelling, Δψ_m, enzymatic activity, NAD(P)H and protein release.

The present data show that superoxide anions and Ca²⁺ produce brain mitochondrial swelling. Moreover, when we analyzed the implication of these two second messengers in mitochondrial swelling, we found that while O₂⁻-induced swelling was Ca²⁺-independent, ROS participated significantly in Ca²⁺-induced mitochondrial swelling. Neither, Ca²⁺ free solution nor the presence of the Ca²⁺ uniporter blocker, Ruthenium Red, modified O₂⁻-induced mitochondrial swelling. In contrast to noncerebral tissue, and in line with previous reports, only high Ca²⁺ concentration induced mitochondrial swelling (Berman et al., 2000; Schild et al., 2001; Brustovetsky et al., 2002). Moreover, Ca²⁺, even at these concentrations, required succinate in the resuspension solution to induce brain mitochondrial swelling, suggesting either its requirement for mitochondrial calcium uptake (Votyakova & Reynolds, 2001) or a role for coenzyme Q as a generator of O₂⁻ (Kowaltowski et al., 1995). This correlates well with the increase in superoxide production detected with the specific oxidation sensitive probe, hydroethidine, in mitochondrial preparations exposed to Ca²⁺. Indeed, Ca²⁺- induced swelling involved a ROS-dependent component, since the superoxide dismutase mimetic, MnTBAP, and catalase markedly decreased Ca²⁺-induced swelling.

The membrane permeable SOD mimetic MnTBAP seemed to be more efficient than the poorly membrane permeable SOD to block Ca²⁺-induced mitochondrial swelling. This difference can be explained, in terms of product compartmentalization by either of the following reasons: (i) O₂⁻ formed inside the mitochondria does not pass through mitochondrial membranes and SOD was less efficient, since it was added to the extramitochondrial compartment; (ii) SOD poorly penetrates into the mitochondria being unable to intercept the generated O₂⁻; and (iii) excess O₂⁻ was degraded to H₂O₂ by MnSOD and exceeded mitochondrial catalase capacity to detoxify it. Consistent with the latter reason, mitochondria treatment with catalase significantly reduced Ca²⁺-induced swelling (Figure 1c).

We have also measured the effects of O₂⁻ and supraphysiological Ca²⁺ on mitochondrial metabolic activity. Superoxide anions, or its related ROS, compromised citrate synthase and dehydrogenase enzymatic activities. These effects seem to be more related to oxidation of the enzymes than to mitochondrial swelling, since Ca²⁺ did not modify them. Our observations might indicate that ROS are the second messengers involved in the decrease of these enzymatic activities showed in several neurodegenerative models, like mice deficient in MnSOD (Melov et al., 1999). Moreover, KO₂ additions dropped total mitochondrial NAD(P)H levels. This is interesting because NAD(P)H, besides acting indirectly as an antioxidant in the reduction of GSSG to GSH, operates as an antioxidant in the matrix space, both by scavenging toxic free radicals and repairing biomolecule-derived radicals (Kirsch & De Groot, 2001). Interestingly, Ca²⁺ increased NADPH autofluorescence. This is consistent with the enhanced activity of Ca²⁺-dependent intramitochondrial enzymes, described in dissociated mouse sensory neurons (Duchen, 1992). This increase was blocked by the addition of Ruthenium Red, suggesting the dependence of the response on mitochondrial Ca²⁺ uptake (Duchen, 1992).

Under our experimental conditions Ca²⁺, but not O₂⁻, produced a rapid loss of Δψ_m that was inhibited by Ruthenium Red, suggesting that Ca²⁺ mediates this mitochondrial effect after its uptake into the mitochondria. Indeed, our study reveals that Ca²⁺-triggered Δψ_m depolarization appeared to be MPTP independent, since CsA did not block this depolarization. These results are in agreement with observations on mitochondrial swelling that postulate that Ca²⁺ acts through a voltage sensor, while ROS-induced swelling results from an oxidative action (Petronilli et al., 1993a, 1993b). It has also been shown that CsA has no effect on Δψ_m depolarization after strong depolarizations (Brustovetsky & Dubinsky, 2000).

The lack of O₂⁻ effects on Δψ_m observed here are consistent with the fact that mitochondrial depolarization is not required for the induction of apoptosis in models where an overproduction of O₂⁻ or the release of cytochrome c takes place (Petronilli et al., 1993a, 1993b; Kluck et al., 1997; Vander Heiden et al., 1997; Krohn et al., 1998; Chiu & Oleinick, 2001) like hippocampal neurons treated with staurosporine (Krohn et al., 1999) or neurons subjected to growth factor withdrawal (Kluck et al., 1997). In other models, protein release happens before Δψ_m depolarization, indicating that disruption of the outer membrane, but not the inner mitochondrial membrane, is the primary structural alteration (Vander Heiden et al., 1997), and that the electron-transport chain can maintain the Δψ_m even after cytochrome c has been released (Goldstein et al., 2000). Moreover, it has been described that O₂⁻-induced cytochrome c release does not produce Δψ_m loss (Madesh & Hajnoczky, 2001) and in other models as RAW 264.7 cells, ROS-induced cytochrome c release is accomplished without relevant changes of Δψ_m (Hortelano et al., 1999).

In agreement with the above data, both Ca²⁺-and O₂⁻-induced swelling resulted in the release of mitochondrial intermembrane proteins. We found that, after mitochondrial swelling, the release of a cytochrome c and the activation of DEVD-caspase took place in isolated mitochondria. This release was sensitive to the presence of CsA suggesting that it was mediated by MPTP. This protein release has been observed in neurodegenerative models, such as focal ischemia and Parkinson's disease (Namura et al., 2001; Viswanath et al., 2001) and has been correlated with the point-of-no-return for cell death pathways (Hirsch et al., 1998). Although the synaptosome contamination can overestimate DVED-like caspase activity, our observations are consistent with previous results from Yuan et al. (2001) which showed that mitochondrial caspases are regulated independently of the cytosolic pool of caspases, and that staurosporine treatment led to preferential autoprocessing of caspase-9 associated with mitochondria. On the other hand, Ca²⁺-induced smaller cytochrome c and DVED-like caspase activity release from mitochondria than KO₂, perhaps throughout calpain-I activation that has been shown to inhibit cytochrome c release and caspase activation in a cell-free system (Lankiewicz et al., 2000).

In summary, our data provide direct information and detailed evidence on how superoxide acts on mitochondria to initiate apoptotic pathways, and that these changes are initiated in the absence of mitochondrial depolarization. However, more work is necessary to identify the mechanism that could explain how MPTP might be activated without loss of mitochondrial potential.

Abbreviations

AFU

Arbitrary fluorescent units

CsA

cyclosporin A

DEVD-AFC

N-acetyl-Asp-Glu-Val-Asp-7-amido-4-trifluoromethylcoumarin

MPTP

mitochondrial permeability transitory pore

MnTBAP

Mn(IIItetrakis(4-benzoic acid)porphyrin chloride)

MTT

3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide thiazolyl blue

Rh 123

rhodamine 123

TCA

trichloroacetic acid

Δψ_m

mitochondrial membrane potential