Ca2+-induced Cyt c release from isolated CNS mitochondria
Addition of 1.4 µmol Ca2+/mg protein to CNS mitochondria incubated in 125 mm KCl-based medium caused a decrease of light scattering reflecting mitochondrial swelling (Fig. 1a) and complete sustained depolarization (not shown; Brustovetsky and Dubinsky 2000a). This Ca2+ concentration was used to ensure full activation of the mPT (Brustovetsky and Dubinsky 2000a). Higher Ca2+ concentrations did not produce larger amplitude swelling (not shown). Alamethicin, an artificial channel-former, produced an additional decrease of light scattering indicating complete unfolding of mitochondrial matrices.
Figure 1. In 125 mm KCl-based medium, Ca2+ caused mitochondrial swelling which was not sufficient to rupture the outer membrane and release Cyt c. (a) 1.4 µmol Ca2+/mg protein produced a decrease in light scattering representative of mitochondrial swelling. Alamethicin (30 µg/mL, Al) was added to promote maximal swelling. Data represent 13 replicates from five different mitochondrial preparations. (b–e) Electron micrographs of CNS mitochondria incubated in 125 mm KCl. (b) Prior to Ca2+ addition mitochondria were in a condensed state; (c,e) Incubation with 1.4 µmol Ca2+/mg protein for 15 min induced mitochondrial swelling that did not disrupt the outer mitochondrial membrane; (d) Alamethicin (30 µg/mL) applied to CNS mitochondria produced prominent swelling, rupture and loss of the outer membrane. Scale bar in (d) also applies to (b) and (c). (f) Despite swelling, CNS mitochondria did not release Cyt c in response to 1.4 or 2.33 µmol Ca2+/mg prot. Mitochondria were incubated with Ca2+ for 15 min. Alamethicin (30 µg/mL, Al) caused maximal Cyt c release from mitochondria incubated in 125 mm KCl. In mannitol-sucrose medium (MS), alamethicin-induced Cyt c release was greatly diminished. Data are means ± SEM of 3–18 replicates from seven different mitochondrial preparations. *p < 0.001 versus control.
Download figure to PowerPoint
TEM of CNS mitochondria confirmed mitochondrial swelling. Initially, mitochondria were in a condensed state with an appreciably enlarged intermembrane space (Fig. 1b). Ca2+ exposure increased the matrix size and decreased the intermembrane space (Fig. 1c). Under higher magnification it was evident that mitochondria retained intact outer membranes (Fig. 1e). After treatment with alamethicin all mitochondria appeared maximally swollen with completely unfolded inner membranes (Fig. 1d). The outer membranes could not be detected in these conditions even at higher magnification.
Incubation of CNS mitochondria in 125 mm KCl-based medium resulted in a minor, basal release of Cyt c (Fig. 1f, without Ca2+) representing about 9% of the maximal release induced in the presence of alamethicin. Neither 0.5 mm EGTA nor 1 µm CsA affected basal Cyt c release (not shown). Surprisingly, when CNS mitochondria were challenged by Ca2+, Cyt c release was not detected (Fig. 1f). Similar results were obtained with pyruvate plus malate as substrates. Alamethicin produced the maximal Cyt c release corresponding to the maximal mitochondrial swelling (Fig. 1f). In low ionic strength mannitol-sucrose medium, alamethicin-induced Cyt c release was greatly diminished (Fig. 1f) due to electrostatic retention of Cyt c by mitochondria (Jacobs and Sanadi 1960; Gogvadze et al. 2001). Thus, the absence of Cyt c release correlated with an inact outer membrane following elevated Ca2+. Direct interaction of elevated Ca2+ with the outer membrane did not cause Cyt c release.
We hypothesized that after isolation in mannitol-sucrose medium, CNS mitochondria might partially lose K+. Loss of K+ could result in an osmotic difference across the inner mitochondrial membrane and condensation of matrices during incubation in 125 mm KCl. From such a condensed initial state, Ca2+-induced swelling of the mitochondrial matrix might not be sufficient to rupture the outer membrane precluding Cyt c release. To balance the osmolarity across the membrane we used 75 mm KCl-based medium, as established in early studies on brain mitochondria (Nicholls and Scott 1980). In these conditions light scattering changes after Ca2+(Fig. 2a) and alamethicin (not shown) were nearly identical to those found in 125 mm KCl. TEM revealed that in 75 mm KCl-based medium, many of the mitochondrial matrices entirely filled out the outer membrane dimensions (compare Fig. 2c 75 mm KCl with Fig. 1b 125 mm KCl). Ca2+ exposure in the 75 mm KCl-based media produced mitochondrial swelling sufficient to cause outer membrane rupture and the appearance of long stretches without outer membranes (Figs 2d and f, arrows). In the extensive regions lacking outer membrane (Fig. 2f), mitochondrial shape was often irregular or distorted and lacked matrix cristae. The cristae of the matrices, although less dense, were better preserved in those portions of mitochondria that retained both inner and outer membranes. These features are characteristic of mitochondria with the ruptured outer membrane (C. Mannella, personal communication). CsA (0.5–5 µm) or ADP (100 µm), inhibitors of the mPT (Hunter and Haworth 1979; Fournier et al. 1987; Crompton et al. 1988), added individually could prevent swelling and sustained depolarization induced by 0.47–0.94 but not with 1.4 µmol Ca2+/mg mitochondrial protein (not shown; Brustovetsky and Dubinsky 2000a). In all experiments ADP was applied in combination with 1 µm oligomycin to prevent ADP phosphorylation. Only the combination of CsA and ADP prevented mitochondrial depolarization and swelling in response to high Ca2+ (Fig. 2b) consistent with their synergistic antagonism of the mPT in liver mitochondria (Novgorodov et al. 1992). TEM confirmed that CsA plus ADP prevented Ca2+-induced swelling of most mitochondria (Fig. 2e). However, some mitochondria were swollen with a ruptured outer membrane even in the presence of mPT inhibitors (Figs 2e and g, arrows). Occasional swollen mitochondria were encountered (∼ 10.9% with mPT inhibitors and Ca2+ vs. 3.3% in control). Most mitochondria appeared to be more condensed than in control conditions (compare Fig. 2c to Fig. 2e), perhaps accounting for the observed increase of light scattering (Fig. 2b).
Figure 2. In 75 mm KCl-based medium Ca2+ induced depolarization, mitochondrial swelling, rupture of the outer membrane and release of Cyt c sensitive to inhibitors of the Ca2+ uniporter and the mPT. (a) In 75 mm KCl-based media, Ca2+ induced a decrease of light scattering (mitochondrial swelling, thick traces), TPP+ efflux (decrease of Δψ, thin traces), and Cyt c release (h). (b,h) The combination of mPT inhibitors, 1 µm CsA plus 100 µm ADP in the presence of 1 µm oligomycin prevented these changes. Cyt c release was also inhibited by 1 µm Ru360. *p < 0.05, **p < 0.01, ***p < 0.001 versus control. Data are means ± SEM of 3–18 replicates from seven different mitochondrial preparations. (c–g) Electron micrographs of CNS mitochondria incubated in 75 mm KCl-based medium. (c) CNS mitochondria prior to Ca2+ addition. After 15 min of addition of 1.4 µmol Ca2+/mg protein (d,f), mitochondrial matrixes were expanded unless 1 µm CsA and 100 µm ADP with 1 µm oligomycin were present (e). Note the absence of long segments of outer membranes (arrowheads) in 75 mm (d,f) compared to 125 mm KCl (Figs 1c and e). Such swollen mitochondria were only rarely encountered in the presence of mPT inhibitors (e,g). Scale bar in (c) also applies to (d) and (e); that in (f) applies to (g).
Download figure to PowerPoint
In line with the observed rupture of the outer membrane, in 75 mm KCl-based medium Ca2+ induced substantial Cyt c release (Fig. 2h). Higher Ca2+ concentrations did not increase Cyt c release. Similar results were obtained in 75 mm KCl-based medium in the presence of 0.1% BSA (not shown) that preserves metabolic integrity of CNS mitochondria purified with density gradients (Lai and Clark 1989). The Ca2+-induced depolarization that accompanied swelling was not a cause of Cyt c release. Depolarization during 30-min incubation of mitochondria with 1 µm FCCP did not result in Cyt c release (0.89 ± 0.06 ng Cyt c/µg protein in␣control vs. 0.94 ± 0.04 after FCCP, n= 3, p > 0.05, two-tailed unpaired t-test). Corresponding to their suppression of mitochondrial swelling, CsA plus ADP inhibited Cyt␣c␣release (Fig. 2h). Residual Cyt c release in the presence of the inhibitors (1.23 ± 0.04 with mPT inhibitors and Ca2+ vs. 0.92 ± 0.04 ng Cyt c/µg prot in control, n= 9–17, p < 0.05, two-tailed unpaired t-test) could be due to either a swelling-independent mechanism of Ca2+-induced Cyt c release (Andreyev and Fiskum 1999; Schild et al. 2001) or damage of the small fraction of mitochondria exposed to the concentrated Ca2+ aliquot as it was added. Ru360, the active fraction of ruthenium red, an inhibitor of the mitochondrial Ca2+ uniporter, also inhibited the mPT (not shown) and Cyt c release (Fig. 2h). These findings demonstrated that the main fraction of Ca2+-induced Cyt c release from CNS mitochondria was mediated by the mPT. L-NMMA (10 mm), a nitric oxide synthase (NOS) inhibitor, reported to prevent Ca2+-induced Cyt c release from isolated rat liver mitochondria (Ghafourifar et al. 1999), did not inhibit Ca2+-induced Cyt c release from CNS mitochondria (not shown).
Thus, Cyt c release in 75 mm KCl correlated with Ca2+-induced, mPT-mediated excessive swelling of isolated CNS mitochondria resulting in rupture of the outer membrane. Inhibition of the mPT prevented mitochondrial swelling and suppressed Cyt c release. Despite the lack of a decrease of light scattering in the presence of mPT inhibitors, TEM revealed a small fraction of swollen mitochondria with irregular distorted shapes and ruptured outer membranes.
Dextran, an inhibitor of mitochondrial porin (VDAC), has been reported to inhibit Ca2+-induced Cyt c release from isolated brain mitochondria (Schild et al. 2001). The proposed mechanism posited that dextran enhanced contact site formation, thus recruiting VDAC to these sites and preventing Cyt c release through unrecruited VDACs (Schild et al. 2001). However, Ca2+ increases contact site formation as well (Bakker et al. 1994) yet activates Cyt c release. In liver mitochondria 10% dextran T70 substantially increased the number of contact sites between the inner and outer membranes (Wicker et al. 1993) and enhanced the voltage sensitivity of VDAC without influencing mitochondrial respiration (Gellerich et al. 1993). We tested dextran T70 and Koenig's polyanion (KPa), a potent VDAC inhibitor (Konig et al. 1982), for their ability to prevent Ca2+-induced Cyt c release. In 75 mm KCl dextran T70 slightly decreased both basal and Ca2+-induced Cyt c release (Table 1). However, Ca2+ continued to induce the same proportional Cyt c release in the presence of dextran T70.
Table 1. Dextran T70 did not prevent Ca2+-induced Cyt c release from CNS mitochondria
| Experimental conditions||Cytochrome c release, ng/µg protein |
|– Ca2+||+ Ca2+||Increase|
|Control||0.91 ± 0.04||1.38 ± 0.04b||+ 52%|
|Control + dextran T70||0.75 ± 0.03||1.15 ± 0.15a||+ 53%|
KPa (20 µg/mL) inhibited respiration and slightly depolarized mitochondria (Figs 3a and b), presumably due to suppression of substrate and ADP translocation through VDAC (Benz et al. 1988). In the parallel experiments, the same amount of KPa enhanced Ca2+-induced swelling and amplified Cyt c release (Figs 3c and d). The basal Cyt c release was not affected by KPa (Fig. 3d). Increased swelling in the presence of KPa could be due to facilitation of the mPT following mitochondrial depolarization (Bernardi 1992). The amplified Cyt c release accompanying enhanced swelling substantiated the hypothesis that Ca2+-induced Cyt c release requires swelling of the mitochondrial matrix beyond the limits of the outer membrane presumably to cause its rupture. Thus, inhibition of VDAC by KPa did not prevent Ca2+-induced Cyt c release.
Figure 3. Koenig's polyanion (KPa) inhibited respiration (thick traces) and slightly depolarized CNS mitochondria (thin traces) (a,b). Numbers near respiration traces represent rates of respiration, nmol O2/min/mg protein. In the presence of 20 µg/mL of KPa, 1.4 µmol Ca2+/mg protein produced a larger amplitude of swelling (c) and increased Cyt c release (d) compared to controls. Data are means ± SEM from three replicate experiments. *p < 0.01, **p < 0.001 versus control.
Download figure to PowerPoint
Measurements of Cyt c release and visualization of mitochondria by TEM clearly indicated a correlation between the rupture of the outer membrane and Cyt c leakage. Cyt c is electrostatically associated with the inner membrane and can readily be displaced from the membrane in high ionic strength medium (Jacobs and Sanadi 1960; Gogvadze et al. 2001). In low ionic-strength mannitol-sucrose medium, maximal swelling of alamethicin-treated mitochondria produced about 10% of maximal Cyt c release obtained in 125 mm KCl (Fig. 1f). However, the substantial Cyt c release observed in 75 mm KCl compared to 125 mm KCl indicated that rupture of the outer membrane may be more important than the ionic strength of the medium. To examine this possibility we compared Cyt c release in two osmotically equivalent media; 25 mm KCl plus 140 mm PEG (MW 300, PEG300, osmolarity 268.3 ± 0.3 mOsm, n= 3) and 125 mm KCl (266.1 ± 1.0 mOsm, n= 7). PEG300 readily entered mitochondria through the Ca2+-activated mPT pore driven by its concentration gradient (Brustovetsky and Dubinsky 2000a) producing larger amplitude swelling than observed in KCl alone (Fig. 4a). Significant Cyt c release was detected (Fig. 4b) despite the decreased ionic strength of the medium. The Cyt c release was almost completely inhibited by CsA plus ADP. Thus, Ca2+-induced, mPT-mediated Cyt c release depended on the extent of mitochondrial swelling and presumed rupture of the outer membrane rather than on the osmolarity of the medium.
Figure 4. In KCl plus PEG300 medium, osmotically identical to 125 mm KCl, 1.4 µmol Ca2+/mg protein produced large amplitude swelling (a)␣and appreciable Cyt c release from CNS mitochondria (b). CsA (1 μm) plus 100 µm ADP in the presence of 1 µm oligomycin inhibited Cyt c release (b). Data in (b) represent means ± SEM of triplicate measurements. *p < 0.05, **p < 0.001 versus control.
Download figure to PowerPoint
Respiratory responses of CNS mitochondria to elevated Ca2+
Ca2+-induced Cyt c release may cause inhibition of mitochondrial respiration (Borutaite et al. 1999; Schild et al. 2001). In this case, addition of Cyt c to the incubation medium could prevent or restore respiratory activity. To test this possibility, respiration rates were measured prior to and after Ca2+ in the absence and presence of exogenous Cyt c(Fig. 5). In the absence of Ca2+ in 75 mm KCl-based medium with succinate plus glutamate as substrates, CNS mitochondria respired at a stable, linear rate until oxygen depletion. Addition of 1.4 µmol Ca2+/mg protein, known to induce Cyt c release (Fig. 2h), increased the rate of respiration (Fig. 5a). Exogenous Cyt c did not affect basal respiration with succinate plus glutamate: 45.6 ± 3.5 versus 46.1 ± 3.2 nmol O2/min/mg prot with Cyt c (p > 0.05, n= 3, two-tailed unpaired t-test) or respiration in the presence of Ca2+: 197.4 ± 10.0 versus 195.7 ± 10.3 nmol O2/min/mg prot with Cyt c (p > 0.05, n= 3, two-tailed unpaired t-test). Similar results were obtained in 125 mm KCl. With pyruvate plus malate (Fig. 5b) or glutamate plus malate (not shown) as substrates the same amount of Ca2+ inhibited respiration. Again, exogenous Cyt c neither influenced basal respiration: 14.3 ± 0.9 versus 13.8 ± 0.9 nmol O2/min/mg prot with Cyt c (p > 0.05, n= 3, two-tailed unpaired t-test) nor prevented the inhibition of respiration after Ca2+: 3.9 ± 0.8␣versus 3.0 ± 0.4 nmol O2/min/mg prot with Cyt c (p > 0.05, n= 3, two-tailed unpaired t-test). Succinate, added after Ca2+, activated respiration (Fig. 5b). As a positive control, ADP-activated respiration with pyruvate plus malate is shown (Fig. 5b). Thus Ca2+-induced alterations of respiration were substrate-dependent and were not sensitive to exogenous Cyt c.
Figure 5. Exogenous Cyt c (30 µg/mL) did not alter changes in respiration caused by Ca2+. CNS mitochondria were incubated in 75 mm KCl-based medium in the presence of 3 mm succinate plus 3 mm glutamate (a) or 3 mm pyruvate plus 1 mm malate (b). The indicated additions of 1.4 µmol Ca2+/mg protein also caused depolarization (not shown). Basal mitochondrial respiration (left trace, a) was linear in the absence of Ca2+. Succinate (3 mm), added after Ca2+, activated respiration in pyruvate plus malate (b). Sequential additions of ADP (b) were used to demonstrate the respiratory capacity of CNS mitochondria oxidizing pyruvate plus malate. Numbers near respiration traces represent rates of respiration, nmol O2/min/mg protein. Incubation medium was supplemented with 0.1% BSA (essentially free fatty acid-free). The respiration traces are representatives from three independent experiments.
Download figure to PowerPoint
Release of Cyt c in cultured neurons upon neurotoxic insult
Isolated CNS mitochondria are a convenient model to study mitochondrial responses to various stimuli. However, the results obtained with isolated mitochondria should be substantiated in intact cells. Therefore, the studies of Cyt c translocation from isolated CNS mitochondria were extended to cultured neurons by examining the effects of the excitotoxin, glutamate. Preliminary experiments with isolated mitochondria first established the effects of preparing neuronal cytosolic extracts upon mitochondrial Cyt c release. Replacement of a high ionic-strength 125 mm KCl-based medium with a low ionic-strength mannitol-sucrose medium minimized Cyt c release during trituration of isolated mitochondria with a 25-gauge syringe needle (0.25 ± 0.4 ng Cyt c/µg protein for untreated mitochondria vs. 0.31 ± 0.3 ng Cyt c/µg protein for mitochondria challenged with 1.4 µmol Ca2+/mg protein, p > 0.05, n= 3, two-tailed unpaired t-test). The total alamethicin-releasable amount of Cyt c in these experiments was 10.0 ± 0.5 ng Cyt c/µg protein (n = 3). Thus, an increased amount of Cyt c in the cytosolic extracts of glutamate-treated neurons prepared by the same way will indicate induction of Cyt c release into the cytosol following glutamate exposure rather than a non-specific leakage due to damage of the outer membrane during preparation of the extracts. If the outer mitochondrial membrane would be damaged during preparation of neuronal cytosolic extracts in mannitol-sucrose medium, Cyt c leakage due to this damage should be minimal and similar in control and glutamate-treated neurons.
Cultured cortical neurons responded to 5 min of 500 µm glutamate by a sustained increase of cytosolic Ca2+ measured with fura-2 (Fig. 6a). Concomitantly, a transient mitochondrial depolarization was observed with rhodamine 123 (Fig. 6c). The initial depolarization was followed by a sustained, secondary depolarization (Fig. 6c) that could indicate induction of the mPT (Vergun et al. 1999; Brustovetsky and Dubinsky 2000b). Indeed, 5 µm CsA prevented the sustained increase of cytosolic Ca2+ and the secondary depolarization (Figs 6b and d). At the end of the glutamate treatment, the relative rhodamine 123 intensity (F/F0), reflecting the extent of depolarization, was lower in the presence of CsA (1.17 ± 0.01, n= 91 vs. 1.44 ± 0.03 without CsA, n= 72, p < 0.0001, two-tailed unpaired t-test). This action of CsA substantiated mPT activation following excitotoxic insult.
Figure 6. Cyclosporin A prevented secondary glutamate-induced elevations in cytosolic Ca2+ and mitochondrial depolarization in cultured cortical neurons. Application of 500 µm glutamate for 5 min resulted in longlasting increases in cytosolic Ca2+ and mitochondrial depolarization as measured simultaneously by Fura 2 ratios and R123 fluorescence normalized to initial values (Control). FCCP 200 nm was added at the end of the traces to fully depolarize mitochondria. The absence of any FCCP associated increases in cytosolic Ca2+ indicated that the depolarized mitochondria had failed to accumulate appreciable Ca2+ (Brustovetsky and Dubinsky 2000b). When 5 µm CsA was present for 1 h prior to and during glutamate addition, mitochondrial depolarization was kept to a minimum and cytosolic Ca2+ recovered. Data are from eight of 72 neurons sampled in six control experiments and eight of 91 neurons sampled in five CsA experiments.
Download figure to PowerPoint
Cyt c release was measured in cytosolic extracts of untreated and glutamate-challenged cultured cortical neurons. Glutamate caused a 2.3-fold increase of Cyt c in the cytosol 30 min after glutamate was removed from the cells (Fig. 7). This represented only about 25% of the maximum Cyt c released by alamethicin in NaCl-based extracellular medium. For comparison, cortical neurons incubated with staurosporine for 6 h released a similar amount of Cyt c (Fig. 7c). Increasing the recovery time after glutamate removal did not increase Cyt c release. Equivalent glutamate treatment of cultured astrocytes did not produce any detectable Cyt c release above the minimum sensitivity of the assay (0.1 ng/µg protein). Glutamate-induced Cyt c release was inhibited by 20 µm MK801, an inhibitor of the NMDA-type glutamate receptors, and by CsA, an inhibitor of the mPT, in a concentration-dependent manner. Compared to the experiments with isolated CNS mitochondria in vitro, in␣situ CsA was effective added alone. The endogenous adenine nucleotides presumably facilitated CsA inhibition of the mPT in situ. Relatively high concentration of CsA (5 µm) were required to suppress Cyt c release. This is in line with the postulate that CsA binding to cytosolic cyclophilins might reduce the actual concentration of CsA reaching mitochondria (Bernardi and Petronilli 1996). FK506, a calcineurin inhibitor, did not inhibit Cyt c release (Fig. 6c). Thus overstimulation of neuronal glutamate receptors and subsequent increases in cytosolic Ca2+ resulted in a rapid partial release of Cyt c from neuronal mitochondria. Inhibition by CsA links this Cyt c release to the neuronal mPT.
Figure 7. Cyt c release paralleled secondary glutamate-induced elevations in cytosolic Ca2+ and mitochondrial depolarization in cultured cortical neurons. Glutamate induced Cyt c release in the presence of the indicated inhibitors, 30 min following glutamate removal. Prior to glutamate exposure, indicated cultures were pre-treated with 20 µm MK801, or with 0.5–5 µm CsA, or 5 µm FK506. For comparison, Cyt c was released in response to 1 µm staurosporine (STS) for 6 h or 30 µg/mL of alamethicin (Al). Data in C are means ± SEM for 3–19 replicates from three different platings. *p < 0.05, **p < 0.01 versus control.
Download figure to PowerPoint
To determine whether the amount of Cyt c released into the cytosol of glutamate-treated neurons was sufficient for caspase activation, caspase-3 activity in the cytosolic extracts was measured 30 min after glutamate removal caspase-3 activity was increased about fivefold: 3.5 ± 0.2 in cytosol of control neurons versus 16.9 ± 0.6 arbitrary fluorescence units in the cytosol of glutamate-treated neurons (n = 3–5, p < 0.0001, two-tailed unpaired t-test). Thus, partial Cyt c release from neuronal mitochondria was sufficient to trigger caspase cascade.