Calcium and Mitochondrial Reactive Oxygen Species Generation: How to Read the Facts

Abstract

A number of recent discoveries indicate that abnormal Ca²⁺ signaling, oxidative stress, and mitochondrial dysfunction are involved in the neuronal damage in Alzheimer’s disease. However, the literature on the interactions between these factors is controversial especially in the interpretation of the cause-effect relationship between mitochondrial damage induced by Ca²⁺ overload and the production of reactive oxygen species (ROS). In this review, we survey the experimental observations on the Ca²⁺-induced mitochondrial ROS production, explain the sources of controversy in interpreting these results, and discuss the different molecular mechanisms underlying the effect of Ca²⁺ on the ROS emission by brain mitochondria.

INTRODUCTION

Mitochondrial dysfunction has been recognized to play an important role in the pathogenesis of neurodegenerative disease. In the case of Alzheimer’s disease (AD), oxidative stress together with mitochondrial dysfunction appear at the early stage of the pathology [1–4] (for review, see [5]), but the underlying mechanism is unknown. Abnormal metabolism of reactive oxygen species (ROS) as an aggravating or primary factor in numerous pathologies, neurodegenerative diseases including AD, and senescence is firmly established, widely recognized, and extensively reviewed elsewhere [6–14].

Oxidative damage is readily detected in AD postmortem tissue [15]; it is the earliest event [3] that could be detected in the brain prior to amyloid-β (Aβ) plaque formation [3] and onset of symptoms of AD [3,16,17]. Oxidative damage to lipids also occurs before Aβ deposition in AD transgenic mice [18]. The Aβ deposits in transgenic mouse models are associated with evidence of oxidative stress as assessed by elevated expression of Cu/Zn superoxide dismutase and heme oxygenase-1, and increased markers of lipid peroxidation [19–21]. It has also been reported that fibrillar deposits of Aβ protein are associated with oxidative damage [22] and Aβ binding alcohol dehydrogenase (ABAD) has been suggested as a molecular mechanism of Aβ mitochondrial toxicity [23]. Oxidative damage may elevate Aβ_1–42 levels by stimulating β-secretase [24]. For the pathology of sporadic AD, a mitochondrial cascade concept was proposed suggesting that mitochondrial dysfunction is the primary event causing Aβ deposition [25]. Indications of oxidative damage in AD were found not only in brain regions but also in peripheral tissues [26–29].

Considering that ROS-induced oxidative stress is a critical factor in injury and that mitochondria are most likely the major source of these ROS, it is clear that studies of ROS metabolism in AD brain mitochondria could provide critical knowledge needed for successful pharmacological intervention strategies to reduce neural cell death.

Elevated intracellular Ca²⁺ and abnormal Ca²⁺ signaling have long being recognized as marker features in AD which led to the formulation of the “Ca²⁺ hypothesis” of brain aging and AD by Khachaturian [30]. This hypothesis postulated that abnormal Ca²⁺ homeostasis mediates or directly causes most manifestations of Aβ peptide-induced neuronal damage in AD. Recently, this idea has experienced a powerful boost by a string of findings [31,32] (reviewed in [33,34]) that demonstrate the molecular mechanisms of Aβ-mediated disruption of Ca²⁺ homeostasis. To date, a number of studies have amply documented various feedback loops linking mitochondrial dysfunctions, oxidative stress, amyloid-β protein precursor processing, Aβ toxic effects, and Ca²⁺ homeostasis in AD. It is beyond our scope to review all this literature here. Neither does it seem necessary as it has been extensively, comprehensively, and recently reviewed elsewhere (e.g., see [35–37]). Instead, we focus on an important aspect of this research field; the relation between mitochondrial ROS production and Ca²⁺ overload.

Ca²⁺ is a key element in physiological signal transductions and also equally important in pathological processes [38–40]. The primary beneficial role of Ca²⁺ in mitochondria by far is the promotion of ATP synthesis, which results from stimulation of the Krebs cycle enzymes and oxidative phosphorylation [41–43]. This effect is achieved by physiological Ca²⁺ signals and enables the adjustment of ATP production to cellular demand. The mechanisms of the harmful effect of Ca²⁺ on mitochondria is less well characterized, but is generally assumed to involve high Ca²⁺ loads and excessive ROS generation (for review, see [44,45]). However, data available from in vitro studies are very contradictory as to the effect of Ca²⁺ on mitochondrial ROS generation ranging from a significant decrease [46–48] to a substantial stimulation [49–52]. For the understanding of the pathology in neurodegenerative conditions, it is crucial to have a clear conception of the factors and conditions which determine the mitochondrial ROS generation in response to a Ca²⁺ challenge. In this review we discuss and summarize the possible mechanisms which could contribute to a Ca²⁺-dependent ROS generation in brain mitochondria.

BRIEF OVERVIEW OF MITOCHONDRIAL ROS GENERATING AND ELIMINATING SYSTEMS

Several decades of research have firmly established that ROS production is inherent to mitochondrial oxidative metabolism and revealed numerous sources of ROS in mitochondria (Fig. 1). This literature has been extensively reviewed by us and others elsewhere (e.g., [13, 53–63]).

Fig. 1.

Mitochondrial ROS production and scavenging. Abbreviations: mGPDH, mitochondrial alpha-glycerophosphate dehydrogenase located at the outer surface of inner mitochondrial membrane which is pictured as a solid line; C1, CIII, COX, respiratory chain complexe I and III, and cytochrome oxidase, respectively; α-KGDHC, alpha-ketoglutarate dehydrogenase complex; ACO, aconitase, PDHC, pyruvate dehydrogenase complex; MAO, monoamine oxidases located in the outer membrane of mitochondria; GR, glutathione reductase; GPX1, glutathione perodixase I, RDS, other enzymes of mitochondrial ROS defense system including (not pictured) manganese SOD, peroxiredoxins 3 and 5, glutaredoxin 2, thioredoxin 2 and thioredoxin reductase, glutathione S-transferase, catalase, Cu, Zn superoxide dismutase, and phospholipid hydroperoxide glutathione peroxidase 4 (see [68] for a review); GSH and GSSG, reduced and oxidized glutathione; CYPD, cyclophilin D. PTP, mitochondrial permeability transition pore. Stars indicate reactive oxygen species (ROS); enzymes labeled with stars are ROS sources.

Mitochondria oxidize various substrates generated inside and outside mitochondria. In the brain, the Krebs cycle mainly generates NADH and FADH₂, which in turn are oxidized in reactions catalyzed by several enzyme complexes located in the inner membrane of mitochondria. The flux of electrons from substrates through various redox carriers and centers in these enzymes is ultimately terminated in a 4-electron reduction of molecular oxygen to water, catalyzed by cytochrome c oxidase. However, some proportion of electrons is diverted from the flow and participates in a single-electron reduction of oxygen, thereby converting it into superoxide, a primary ROS, which quickly dismutates to form H₂O₂. The latter is membrane permeable and diffuses out of mitochondria. The highest ROS producing capacity in brain mitochondria has been demonstrated for complex I and complex III of the respiratory chain [64–67] and the enzyme dihydrolipoamide dehydrogenase [68–70]. In intact mitochondria, the activities of various ROS sources are linked to each other through the common pools of intermediates such as NADH and CoQ, thus it is not possible to determine under physiological conditions, when the complexes are not inhibited, which one of the possible sites is the major ROS generator. Vast amount of studies have documented that inhibitors of complex I and complex III induce robust ROS release from isolated brain mitochondria [71–75]. However, only complex I inhibition appears to be physiologically important in light of the observation made on in situ synaptic mitochondria, that ~16% inhibition of complex I is already accompanied by an enhanced ROS formation, whereas complex III needs to be inhibited by > 70% for an increase in ROS generation [76]. This amount of complex III inhibition is unlikely to be an underlying in vivo mechanism of pathological ROS generation.

ROS production capacity of mitochondria is controlled by factors affecting and reflecting the metabolic state of intact mitochondria. It has been found that the chemical nature of the substrates fuelling the respiratory chain, the amplitude of the membrane potential in mitochondria (ΔΨm), the pH of the matrix, and the oxygen tension in their surrounding [55,68] are the most important factors controlling the ROS production in mitochondria. Out of these, the importance of substrates and ΔΨm are discussed below in association with the effect of Ca²⁺ on ROS generation.

The ‘ROS defense system’ (RDS) comprises several enzymes specialized for removal of superoxide, H₂O₂, and organic hydroperoxides. Most of these enzymes are ubiquitously present in all mammalian mitochondria; the expression level of these enzymes exhibits tissue and species specificity. A unique feature of RDS is that almost all of its enzymes rely on NADPH as a source of reducing equivalents needed for their activity. The NADPH reduction is carried out by three intramitochondrial enzymes; isocitrate dehydrogenase (NADPH linked), malic enzyme, and transhydrogenase [77]. To note, the intramitochondrial pools of NADPH and reduced glutathione (GSH) are rather large (ca. 3–5 mM NADPH [78,79] and 2–14 mM GSH [80–83]), therefore, transient changes in the activity of enzymes would not immediately affect the RDS and its ability to extinguish short bursts in ROS concentration. However, a prolonged activity of RDS, its ‘endurance’, ultimately depends on the supply of NADPH and GSH, thus depending on the ability of enzymes to regenerate these compounds.

It is clear that elevated mitochondrial ROS emission may be determined by both a true increase in ROS generation from a mitochondrial site and a failure of mitochondrial RDS. However, for the purpose of designing an intervention to prevent oxidative stress and tissue damage by mitochondrial ROS, in-depth knowledge of the mechanism of ROS emission is crucial. To the best of our knowledge, this issue has not yet been addressed in details for AD brain mitochondria.

VARIABLE EFFECTS OF Ca²⁺ ON MITOCHONDRIAL ROS GENERATION

In general, an effect of Ca²⁺ on mitochondrial ROS formation requires the influx of Ca²⁺ into the matrix. In brain mitochondria, the primary mechanism of Ca²⁺ uptake is via a highly selective ion channel, termed uniporter, driven by the electrochemical gradient across the mitochondrial inner membrane. This channel exhibits remarkable low affinity for Ca²⁺ [84]. While the kinetic and pharmacological nature of this channel is well-characterized, the molecular entity of the channel remains to be identified. The electrophoretic Ca²⁺ entry involves a net charge movement, therefore lowering ΔΨm. For liver and heart mitochondria, an additional “rapid mode” uptake has been described [85,86] allowing fast changes in matrix Ca²⁺ concentration and rapid stimulation of Ca²⁺-dependent processes. In cardiac mitochondria, a ryanodine receptor has been identified, which might also mediate Ca²⁺ entry into the matrix [87].

Rapid efflux of Ca²⁺ from the matrix requires Na⁺/Ca²⁺ exchange coupled to H⁺/Na⁺ exchange via the inner membrane of mitochondria (see [88]). A possible efflux pathway for Ca²⁺ is the mitochondrial permeability transition pore (PTP), which is a large conductance channel formed by proteins in the inner and outer membrane of mitochondria allowing the release of solutes < 1.5 kDa including GSH and pyridine nucleotides from the matrix and leading to loss of ΔΨm, osmotic swelling, and rupture of the outer mitochondrial membrane [89–92].

There is a general conception that Ca²⁺ overload leads to stimulated ROS generation in mitochondria. However, data are available in the literature both demonstrating this and indicating the opposite. Mitochondrial Ca²⁺ accumulation has been shown to promote [50,51,93,94], to be without an effect on [95–97], or to decrease ROS generation [46–48]. These studies, even only those performed with brain mitochondria, are difficult to compare due to the great variance in the conditions at which the Ca²⁺ challenge is imposed and mitochondrial ROS generation is measured. In order to understand the mechanism by which Ca²⁺ stimulates or decreases ROS generation, it is crucial to consider key factors which possibly determine the response of mitochondria to a Ca²⁺ challenge. Since isolated mitochondria are used in the vast majority of these studies, the choice of the substrate(s) fuelling the respiratory chain is an obvious variable; whether electrons are donated to complex I or to coenzyme Q (CoQ). An important factor is the metabolic state of mitochondria before and under the Ca²⁺-load which determines whether Ca²⁺-induced changes in ROS generation would be dependent on ΔΨm. A sharp distinction is made by the fact whether mitochondria undergo Ca²⁺-dependent PTP opening or could handle the Ca²⁺-load without a major inner membrane permeability increase. These conditions should be carefully scrutinized for the interpretation of the effect of Ca²⁺ on mitochondrial ROS emission.

SUBSTRATE-DEPENDENCE OF THE EFFECT OF Ca²⁺ ON ROS GENERATION

Mitochondrial respiration in vitro can be supported either by substrates linked to NAD⁺ reduction (glutamate, malate, α-ketoglutarate, pyruvate) donating the electrons to FMN cofactor of complex I, or by FAD-linked substrates (succinate, α-glycerophosphate), which reduce the more distal CoQ pool in the respiratory chain. In the case of NAD⁺ linked substrates, complex I generates superoxide with electrons from the fully reduced FMN [60,98]. The reduction state of FMN is set by the NADH/NAD⁺ ratio, therefore anything that increases this ratio, either inhibition of the respiratory chain or a low ATP demand, will increase ROS generation [60,99,100].

In the presence of FAD-reducing substrates and when ΔΨm is high, electrons can flow back to complex I (reverse electron transport; RET) and reduce NAD⁺ to NADH [101,102]. Superoxide is generated with high rate under this condition [71,99,103], possibly at the same site; the FMN coenzyme of complex I [60]. RET is favored in mitochondria supported by FADH₂-dependent substrates, such as succinate or α-glycerophosphate but only when ubiquinone is highly reduced at a high ΔΨm. This is typically a condition easily created in vitro with isolated mitochondria, but in vivo the dominance of electron input from NAD⁺ linked substrates makes this process unlikely, though not impossible. It has been postulated [13] that during hypoxia succinate concentration could rise to a sufficiently high level to generate high ΔΨm and ROS generation at complex I during reoxygenation. Supportive for this is the finding [104] that succinate stimulates ROS formation even in the presence of NAD⁺-linked substrates without preventing their oxidation. It has been shown that 5 min of ischemia decreased the concentrations of glycolytic intermediates and mitochondrial NAD-linked oxidative substrates, but increased succinate concentration by ~300% to the millimolar range in rat brain [105,106]. Another interesting finding is that hypoxia significantly (> 60%) activated succinate and glutamate oxidation by isolated rat brain mitochondria [107,108].

However, since RET is highly sensitive to ΔΨm and a small decrease in ΔΨm inhibits succinate-dependent ROS generation [109], it is unlikely that Ca²⁺ uptake, which decreases ΔΨm, could stimulate ROS generation via RET. On the contrary, ROS release from succinate-supported brain mitochondria is immediately and almost completely inhibited by a Ca²⁺ overload [48]. Similarly, ROS release from well-coupled brain mitochondria respiring on α-glycerophosphate is reduced by high Ca²⁺ loads (L. Tretter, unpublished observation) consistent with the ΔΨm-dependent character of the RET-related ROS generation in these [110] and other mitochondria [111]. It is of note that Ca²⁺-induced decrease in RET-related ROS generation, which is due to a drop in ΔΨm, is observed only when conditions are unfavorable for PTP opening (in the presence of ADP or ATP or other PTP inhibitors). Without that, Ca²⁺-induced PTP dominates the response of mitochondria and determines the changes in ROS generation.

It has to be noted here that submicromolar concentrations of Ca²⁺ stimulate ROS production by mitochondrial α-glycerophosphate dehydrogenase [112] likely due to stimulation of the enzyme described earlier in liver mitochondria [113], which might be significant in vivo in deenergized mitochondria, which are unable to take up Ca²⁺ and to control the cytosolic Ca²⁺ concentration.

In mitochondria supported by NAD⁺-linked substrates in vitro or respiring on endogenous substrates in vivo, NADH/NAD⁺ ratio is critical for ROS generation not only by complex I but also by the Krebs cycle enzyme, α-KGDHC. A common feature of neurodegeneration in various diseases is a decline in the activity of mitochondrial enzymes, of which reduction in α-KGDHC activity in AD is well documented [114–116] (for review see [117]). Aβ, which is deposited in the mitochondria of AD patients and transgenic mice before substantial accumulation extracellularly [23,118], has been shown in isolated brain mitochondria to inhibit α-KGDHC [119]. α-KGDHC is a key enzyme in the Krebs cycle providing NADH for complex I in the respiratory chain. This enzyme is sensitive to inhibition by different ROS including H₂O₂ [120–122], peroxynitrite [123], or intrinsic radical species [124]. More intriguing with relation to ROS generation is the ability of α-KGDHC to generate ROS during its normal catalytic function attributable to the dihydrolipoamide dehydrogenase component of the enzyme [69,70,125] (for review, see [126]). The latter enzyme is the common component of pyruvate, α-KGDHC, and branched-chain ketoacid dehydrogenase complexes, and also participates in glycine cleavage system [127]. ROS generation by α-KGDHC is regulated by the NADH/NAD⁺ ratio; an increase in this ratio, while inhibiting the physiological catalytic activity, promotes ROS generation by the enzyme [70]. Ca²⁺ is a well known regulator of α-KGDHC by activating the enzyme in low μM concentrations [42,128]. It was demonstrated with isolated α-KGDHC that parallel with the activation of the enzyme, α-KGDHC produces higher amount of H₂O₂ in the presence of Ca²⁺ [70]. It is not possible yet to unambiguously determine to what extent α-KGDHC contributes to mitochondrial ROS production in situ, let alone in vivo, but some in vitro data allow the cautious assumption that it might be significant under certain physiological or pathological conditions. Isolated brain mitochondrial produce the highest amount of ROS when supplied with α-ketoglutarate as compared to other substrates [69]. Recently it has been reported that specific inhibitors of α-KGDHC administered together with glutamate in cultured neurons, inhibited the glutamate-induced ROS production by about 20% suggesting that α-KGDHC might be a source of ROS under glutamate stimulation [129]; the latter is known to involve accumulation of Ca²⁺ in mitochondria, to enhance ROS generation, and to induce cell death [130–137] (for a recent review, see [39]).

Given the activation of this enzyme by Ca²⁺ [42, 128], one might expect an enhanced ROS production by α-KGDHC under high Ca²⁺ loads. However, stimulation of the enzyme is only observed with ≤ 20 μM Ca²⁺ concentration [42,128] and it was demonstrated with brain mitochondria that the effect of Ca²⁺ is biphasic, activating the enzyme in low μM concentration but inducing progressive inhibition in ≥ 100 μM concentrations [138]. In agreement with this, we found a stimulated ROS generation by isolated α-KGDHC only in Ca²⁺ concentrations up to 20 μM [70]. On the other hand, when respiration and oxidative phosphorylation is inhibited by high Ca²⁺ (see below), the NADH/NAD⁺ ratio increases, favoring an accelerated ROS generation by the enzyme. Due to the common pyridine nucleotide pool, it is not possible to establish the relative contribution of complex I and that of α-KGDHC to the increased mitochondrial ROS emission promoted by an increased NADH/NAD⁺ ratio.

DEPENDENCE ON ΔΨ OF THE EFFECT OF Ca²⁺ ON ROS GENERATION

Given the fact that under certain conditions mitochondrial ROS generation is dependent on ΔΨ, it has to be considered whether depolarization associated with Ca²⁺ uptake could be a factor in the effect of Ca²⁺ on ROS formation.

The first evidence for the ΔΨ-dependent nature of ROS generation is that in isolated succinate-supported heart mitochondria, uncouplers decreased the rate of ROS emission [109,139]. Importantly, this effect was evident only in a narrow ΔΨ range and only in well-coupled, highly polarized mitochondria, where decrease in ΔΨ by only 10 mV resulted in 80% decrease in the rate of ROS generation [109]. The ‘turbo’ mode of isolated mitochondria exhibiting high ΔΨ and high rate of ROS formation is achieved only when bovine serum albumine (BSA) is present during the isolation or incubation to eliminate the uncoupling effect of contaminating free fatty acids, succinate is used as a respiratory substrate, and the experiments are performed in the absence of ADP [140]. The high protonmotive force in these mitochondria drives the back flow of electrons via complex I and RET is responsible for the high rate of ROS generation characteristic for these mitochondria (see above). With a few mV decrease in ΔΨ, the protonmotive force is no longer sufficient to maintain RET explaining the decrease in the rate of ROS generation. The ΔΨ-dependence of ROS generation with NAD⁺-linked substrates is also evident in the −150 and −180 mV ΔΨ range [48,141] though the depolarization-induced decrease in ROS generation is less dramatic [46,142].

Ca²⁺ load depolarizes mitochondria due to the electrophoretic Ca²⁺ uptake [38], which may be transient or sustained depending on the amount of the Ca²⁺ load [143]. The substantial decrease in ROS emission by Ca²⁺ from succinate-supported mitochondria [46, 48,95] is evidently due to the elimination of RET. Reduction in the ROS release by Ca²⁺ was also observed in mitochondria respiring on NAD⁺-linked substrates [46,48,95,143]. The mechanism by which depolarization reduces mitochondrial H₂O₂ formation probably involves the oxidation of redox centers, which mediate the generation of superoxide. Again, for this effect of Ca²⁺, incubation conditions of mitochondria have to be favorable for high ΔΨ and unfavorable for PTP induction (for example, presence of ATP but not ADP alone, to inhibit PTP induction or presence of ADP plus oligomycin to prevent PTP and ATP synthesis). Under these conditions Ca²⁺ load that causes sustained depolarization of the highly polarized mitochondria will decrease ROS emission [48,143].

An opposite effect of Ca²⁺, e.g., stimulation of ROS release that is unrelated to PTP induction, is observed in mitochondria studied in the presence of ADP. Under physiological conditions, ADP (and ATP) is continuously present in the mitochondrial matrix controlling the rate of respiration and ATP synthesis, so it is highly adequate to include adenine nucleotides in the incubation medium for isolated mitochondria. ADP, on one hand, is an inhibitor of PTP induction [144–147], and, on the other, stimulates respiration and ATP synthesis, therefore decreases ΔΨ (state 3). In these mitochondria, ΔΨ is below the range in which ROS generation is dependent on ΔΨ, therefore Ca²⁺-induced PTP-independent depolarization is no longer expected to decrease ROS formation. In this case, the response of mitochondria depends on the amount of Ca²⁺ load; high Ca²⁺ concentrations, in our case 100–300 μM, cause sustained depolarization but no alteration in ROS release from mitochondria [143], clearly showing the lack of effect of high Ca²⁺ load per se on the mitochondrial ROS producing machinery. However, in lower, but still pathological concentrations (10–100 μM), Ca²⁺ depolarizes mitochondria only transiently; thereafter ΔΨ recovers to a higher (more negative) value than that before the Ca²⁺ challenge. This ‘after-hyperpolarization’ parallels a significant increase in the ROS release from mitochondria [143]. The mechanism of the relative hyperpolarization following depolarization by moderate Ca²⁺ load is yet to be clarified but could be related to inhibition of the adenylate translocase [148] or F₀F₁-ATPase by Ca²⁺ [149,150]. Nonetheless, the stimulated ROS generation under this condition is most likely due to a shift of ΔΨ towards higher values, into the range where ROS generation is sensitive to changes in ΔΨ. This phenomenon highlights another important variable that determines the changes of ROS emission from mitochondria in response to a Ca²⁺ challenge; the amount of Ca²⁺ load.

In summary, in mitochondria actively synthesizing ATP (in the presence of ADP), therefore being depolarized, the effect of Ca²⁺ is dependent on the amount of Ca²⁺ load. In the lower range of Ca²⁺ load, transient Ca²⁺-induced depolarization is followed by a recovery to a relative hyperpolarized state and, due to the latter, ROS generation is stimulated, whereas large Ca²⁺ concentrations dissipate ΔΨ without the tendency of recovery and fail to influence ROS emission from mitochondria. In highly polarized mitochondria not synthesizing ATP (in the presence of ATP and/or oligomycin) but exhibiting high rate of basal ROS generation, the drop in ΔΨ due to Ca²⁺ uptake is not followed by recovery to a hyperpolarized state at any Ca²⁺ concentration and is associated with a decreased ROS generation. In these effects of Ca²⁺, clearly the ΔΨ-dependent feature of ROS generation is reflected.

PERMEABILITY TRANSITION, Ca²⁺, AND ROS GENERATION

Several reports demonstrate that the opening of PTP correlates with an increase in ROS production in vitro in isolated mitochondria [151,152] and in situ in rat CA1 pyramidal neurons in organotypical slices [153]. However, to the best of our knowledge, no study has yet been published detailing the molecular mechanism of this phenomenon. This is a controversial issue as PTP induction by Ca²⁺ is not associated with an increased ROS production in other studies. We even have observed a slight decrease in ROS emission from mitochondria experiencing a Ca²⁺-induced permeability increase of the inner membrane, which is consistent with a net loss of pyridine nucleotides from the matrix of Ca²⁺-loaded mitochondria exhibiting PTP [47]. Nevertheless, we can make an educated guess about how Ca²⁺ overloading of mitochondria and/or PTP opening could significantly induce their ROS production (Fig. 2).

Fig. 2.

Enhancement of ROS production in mitochondria that underwent Ca²⁺-induced permeability transition. Abbreviations are the same as in Fig. 1. Overloading of mitochondria with Ca²⁺ results in PTP opening, which in turn, results in leakage of pyridine nucleotides (NAD⁺, NADP) and Krebs cycle substrates such as oxaloacetate, alpha-ketoglutarate, and malate from the mitochondrial matrix. This renders both the Krebs cycle and RDS inoperable and results in sever oxidation of mitochondrial glutathione. Leakage of NAD⁺ also stimulates ROS production by α-KGDHC and PDHC. In addition to that, elevated Ca²⁺ inhibits GPX1, and directly stimulates ROS production by MAO and mGPDH. Elevated intramitochondrial ROS may further damage aconitase and complex I thereby turning them into ROS sources. Both the inability of permeabilized mitochondria to efficiently scavenge ROS (due to the damage of the RDS) and the elevated primary ROS production contribute to the enhanced ROS emission from mitochondria.

As illustrated in Figs 1 and 2, brain mitochondria contain several potential ROS sources such as monoamine oxidase (MAO), complex I, α-KGDHC, PDHC, and α-glycerophosphate dehydrogenase, and also quite efficient ROS defense system. The latter is “fed” by NADPH that is used by glutathione reductase (GR) to regenerate oxidized glutathione (GSSG) to GSH, which is further used by glutathione peroxidase 1 (GPx1) to detoxify H₂O₂ and by thioredoxin reductase that regenerates oxidized thioredoxin, peroxiredoxins, and glutaredoxin. For generation of NADPH in brain mitochondria, malic enzyme and NADP-linked isocitrate dehydrogenase (reviewed in [55,68]) use the metabolites generated in the Krebs cycle. However, the Krebs cycle cannot operate in permeabilized mitochondria, because its metabolites and perhaps more important, pyridine nucleotides are released from the matrix of mitochondria through the PTP. Indeed, loss of mitochondrial matrix pyridine nucleotides is a prominent consequence of the PTP opening as demonstrated both in vitro in isolated mitochondria and in vivo in perfused rat heart [154]. Moreover, net loss of NAD⁺ was also observed in cell cytosol due likely to the activation of NAD⁺-glycohydrolase, which is associated primarily with the cytosolic surface of mitochondria [154].

In addition to that, a mitochondrion undergoing PTP opening is de-energized, cannot accumulate Ca²⁺, and therefore is incapable of controlling the Ca²⁺ concentration in its vicinity. This would further increase constitutive ROS production because at least two of the mitochondrial ROS sources are stimulated by elevated Ca²⁺; α-glycerophosphate dehydrogenase [112] and MAO-A as demonstrated in primary hippocampal cell cultures and in HT-22 cells [155].

On the other hand, the constitutive scavenging of ROS is expected to be diminished in permeabilized mitochondria because of the loss of GSH from the matrix space. Although the intramitochondrial GSH pool is about the same as in cytosol, ca. 2–14 mM [80–83] so no dilution would occur, it would no longer be regenerated inside mitochondria due to the loss of NADPH. The net loss of GSH/GSSG was demonstrated in mitochondria isolated from brain subjected to ischemia and reperfusion in vivo [156], a treatment that is firmly associated with the PTP opening. To make things worse, Ca²⁺ overloading can directly diminish mitochondrial H₂O₂ scavenging capacity by inhibiting glutathione reductase/peroxidase system [157].

Summarizing, it seems quite feasible that in vivo, PTP opening could stimulate mitochondrial ROS production both due to a net increase in ROS emission and a failure in ROS scavenging. In line with this conclusion, a recent study by Wang and colleagues [158] has demonstrated that PTP opening in mitochondria in intact cells generates superoxide. This study is remarkable and innovative in many aspects; in particular, it is fascinating that the authors apparently have managed to prove two fundamentally important phenomena, namely the possibility and physiological relevance of spontaneous PTP opening at the level of individual mitochondria in a living cell and the association of PTP opening with a burst of superoxide production. In these experiments, a circularly permuted yellow fluorescent protein (cpYFP) sensitive to superoxide was modified with a mitochondria-targeting sequence and transfected into cultured adult cardiomyocytes, where it localized to mitochondria. The authors observed random spontaneous bursts of cpYFP fluorescence arising from a single or a pair of functionally intact mitochondria. These flashes were observed in a number of cell types including hippocampal neurons and primary cultures of cardiomyocytes isolated from cpYFP transgenic mice. Most remarkable, these flashes were associated with a temporary drop in ΔΨm and leakage of matrix-entrapped indicator from mitochondria, were inhibited by well-known PTP inhibitors such as cyclosporin A and bongkrekic acid, and were diminished by knockdown of cyclophilin D, a PTP enhancer. Thus, the authors concluded that the observed phenomenon of spontaneous cpYFP fluorescence was in fact, a result of sporadic PTP opening and closure that was associated with bursts in superoxide production [158].

A growing body of evidence indicates that PTP is involved in the pathology of AD. It has been demonstrated that Aβ peptides exacerbated the PTP inducing effect of Ca²⁺ in both liver and brain mitochondria; brain mitochondria being more resistant to the potentiation by Aβ of Ca²⁺-induced PTP [159,160]. It is remarkable that Aβ has been shown to induce swelling and cytochrome c release from isolated brain mitochondria sensitive to PTP inhibition by cyclosporine A [161]. It is interesting that in vitro, the PTP induction by Aβ_25–35 does not require massive amounts of exogenous Ca²⁺; the amount present in mitochondria endogenously is sufficient to facilitate Aβ_25–35-induced swelling and accumulation of lipid peroxides [162]. More recently, it was found that Aβ_25–35 and Aβ_1–42 oligomers, but not fibrils, caused massive influx of Ca²⁺ into cerebellar granular cells and in situ mitochondrial Ca²⁺ overload resulting in an increased intracellular ROS production, cyclosporine A-inhibitable permeabilization of mitochondria and cytochrome c release [163]. This is remarkable as soluble oligomers are thought to mostly contribute to the AD pathological changes in the brain [164]. Perhaps, the strongest evidence that in vivo mitochondrial PTP is directly linked to neuronal damage in AD pathogenesis has been obtained recently by Du et al. [165]. These authors took advantage of mouse genetically ablated of cyclophilin D (CYPD), a mitochondrial protein that regulates the Ca²⁺ threshold of PTP opening. CYPD deficient mitochondria open PTP at higher Ca²⁺ load than wild type mitochondria. [165] have demonstrated that CYPD deficient cortical mitochondria are resistant to Aβ and Ca²⁺-induced swelling and PTP opening, exhibit higher Ca²⁺ buffering capacity, and produce less ROS. The neurons from CYPD knockout mice are also less prone to die when challenged with Aβ or oxidative stress. Furthermore, CYPD deficiency greatly improves cognitive functions in an AD mouse model/CYPD knockout cross [165].

OTHER FACTORS

In addition to PTP opening, there are other well documented malfunctions of the mitochondrial machinery caused by over-accumulation of Ca²⁺ that result in metabolic insufficiency of mitochondria and, therefore, can stimulate mitochondrial ROS production. Inhibition of mitochondrial enzymes, respiration, and oxidative phosphorylation by high Ca²⁺ is well documented [138,166–168]; among them are several major dehydrogenases of the Krebs cycle, including PDHC [166] and α-KGDHC [138]. Accumulated Ca²⁺ may also decrease the intramitochondrial pool of ADP, thus reducing the exchangeable pool and the amount of ADP available to the F₁-ATPase [167]. Accumulation of ~40 nmol Ca²⁺ by mitochondria strongly inhibited the ATP/ADP translocase activity [148]. Progressive accumulation of large amounts of Ca²⁺ and Pi results in Ca²⁺-Pi precipitate formation in the mitochondrial matrix [169,170]; mitochondria from neural tissues can accumulate so much Ca²⁺ (2000–4000 nmol/mg protein [171] that the precipitate may literally fill up the mitochondrial matrix water space, thus creating diffusion limitations for substrate delivery to primary dehydrogenases. All and any of these events are expected to significantly limit the ability of mitochondria to scavenge ROS and/or to increase net mitochondrial ROS emission [55,68].

CONCLUSION

Convincing evidence is lacking to support the general notion often stated in the literature that Ca²⁺ accumulation by in situ mitochondria results directly in oxidative stress. There is no known target or mechanism that would uniformly determine the effect of Ca²⁺ on ROS emission from mitochondria. ROS generation in response to a Ca²⁺ challenge depends on many variables. In mitochondria not experiencing PTP, the metabolic state is crucial by setting the membrane potential either to a high range of values (no ATP synthesis), where Ca²⁺ uptake results in a decreased ROS generation, or to a depolarized range (ATP synthesis) in which ROS generation is stimulated or not influenced by Ca²⁺ depending on the amount of the Ca²⁺-load. ‘Pathological Ca²⁺ load’ covers wide range of Ca²⁺ concentrations, but effects exerted by Ca²⁺ in different concentrations within this range are not uniform either. Conditions favoring PTP induction or opposite, delaying Ca²⁺-induced pore formation are also crucial for alterations in the mitochondrial ROS emission by Ca²⁺; however, the mechanism underlying the PTP-related changes in ROS release from mitochondria is yet to be elucidated.

Finally, it is important to emphasize that while isolated mitochondria are extremely useful for studies on Ca²⁺-induced changes in ROS generation allowing to choose the most/least favorable conditions for the dissection of a particular aspect of the Ca²⁺ action, the extrapolation of results to the in vivo function needs extreme caution since the intracellular environment for in situ mitochondria are far more complex and the respiratory sate is highly dynamic.