Mitochondrial-Mediated Suppression of ROS Production Upon Exposure of Neurons to Lethal Stress: Mitochondrial Targeted Preconditioning

Abstract

Preconditioning represents the condition where transient exposure of cells to an initiating event leads to protection against subsequent, potentially lethal stimuli. Recent studies have established that mitochondrial-centered mechanisms are important mediators in promoting development of the preconditioning response. However, many details concerning these mechanisms are unclear. The purpose of this review is to describe the initiating and subsequent intracellular events involving mitochondria which can lead to neuronal preconditioning. These mitochondrial specific targets include: 1) potassium channels located on the inner mitochondrial membrane; 2) respiratory chain enzymes; and 3) oxidative phosphorylation. Following activation of mitochondrial ATP-sensitive potassium (mitoK_ATP) channels and/or increased production of reactive oxygen species (ROS) following disruption of the respiratory chain or during energy substrate deprivation, morphological changes or signaling events involving protein kinases confer immediate or delayed preconditioning on neurons that will allow them to survive otherwise lethal insults. While the mechanisms involved are not known with certainty, the results of preconditioning are the enhanced viability, the attenuated influx of intracellular calcium, the reduced availability of ROS, suppression of apoptosis, and the maintenance of ATP levels during and following stress.

1. Introduction

Mitochondria are important regulators of neuronal function in addition to their role in energy production. There is extensive evidence that transient exposure of mitochondria to physiological or pathological stimuli, intracellular events, or pharmacological agents, can initiate and promote sustained changes in mitochondria that ultimately protect neurons against a number of potentially lethal stresses [1,2]. The establishment of a sustained, protective response against otherwise lethal stresses following transient exposure of cells to initiating stimuli has been termed “preconditioning.” Two forms of preconditioning have been described. Immediate preconditioning occurs within minutes of the initiating stimulus and lasts for several hours before disappearing, while delayed preconditioning takes several hours to develop and is present for several days [3,4]. Given the temporal aspects of these two types of preconditioning, it seems likely that immediate preconditioning primarily involves cellular changes relating to the activity or function of enzymes, second messengers, and ion channels already present while delayed preconditioning principally is due to de novo protein synthesis. While the mechanisms involved are not fully understood, the result of preconditioning is that the neurons are able to limit the influx of calcium and the availability of reactive oxygen species (ROS) during stress [5,6]. While low level ROS production occurs normally and maintains proper cellular function, excess levels of ROS can overwhelm anti-oxidant systems, especially in metabolically compromised cells, and cause damage and death of neurons. The mitochondrial specific targets of stimuli which induce neuronal preconditioning include: 1) potassium channels located on the inner mitochondrial membrane; 2) respiratory chain enzymes; and 3) oxidative phosphorylation. The purpose of this review is to describe the initiating and subsequent intracellular events involving mitochondria which can lead to neuronal preconditioning.

2. Mitochondrial potassium channels

Several different potassium channels have been identified in the inner mitochondrial membrane and their activation may initiate neuronal preconditioning [7,8] (Figure 1). Activation of these channels allows potassium ions to flow into mitochondria and results in depolarization. The two most likely targets of preconditioning are the ATP-sensitive potassium (K_ATP) and the large conductance calcium activated potassium (BK_Ca) channels [9,10,11]. While there is extensive evidence for the existence and importance of the mitochondrial (mito) K_ATP channels in neuronal preconditioning [12,13,14], there is speculation that the mitoBK_Ca channels, if present, are not involved in neuronal preconditioning [15].

Figure 1.

Schematic illustration showing signaling events which occur following opening of mitoK_ATP channels or liberation of ROS from the protein complexes which form the electron transport chain. These two initiating events, which can occur separately or together, lead to the activation of intramitochondrial and intracellular signaling mechanisms. The final result of preconditioning is the protection of neurons during and following exposure to potentially lethal stresses, due to the sustained depolarization of mitochondria, the attenuation of intracellular calcium influx, the elimination of the ROS surge, the preservation of ATP levels, the prevention of apoptosis, and the maintenance of normal mitochondrial morphology. Abbreviations: ΔΨm, mitochondrial membrane potential; O₂⁻, superoxide anion; H₂O₂, hydrogen peroxide; Ca²⁺, calcium; ADP, adenosine di-phosophate; ATP, adenosine tri-phosphate; PKC, protein kinase C; Gsk3β, phospho-glycogen synthase kinase 3 beta; PI3K, phosphoinositide 3-kinase; Bad, Bcl-2 associated death promoter, Akt; protein kinase B.

2.1 ATP-sensitive potassium channels

The structure of mitoK_ATP channels is not known with certainty but may be inferred from what is known about the better understood plasmalemmal K_ATP channels. These K_ATP channels are typically composed of four pore forming inwardly rectifying potassium channel (Kir) subunits and four modulatory sulfonylurea receptor (SUR) subunits [16]. We have shown that the Kir 6.1-immunopositive subunits are predominant in brain mitochondria, and that these subunits are localized to the inner mitochondrial membrane using immunogold electron microscopy [17]. The Kir subunits are more concentrated in mitochondria compared to whole brain tissue [17], thereby emphasizing the functional importance of mitoK_ATP channels to neurons. The identification of SUR subunits has been more problematic and their exact nature is unclear [17]. Although there have been recent reports [18] that the mitoK_ATP channel lacks Kir subunits, or that the channel doesn’t exist in a form similar to the plasmalemmal K_ATP channel [19], the vast majority of published papers support the presence of Kir pore-forming subunits, as does our finding that the correct targeting sequences are present on the Kir subunits to direct them into the appropriate location on the inner membrane of mitochondria [17]. Furthermore, the bulk of the evidence indicates that isolated mitochondria or mitochondria in cultured cells or tissue slices depolarize in a dose-dependent manner to well-characterized mitoK_ATP channel openers such as diazoxide and BMS-191095 and are responsive to other factors such as endogenously produced peroxynitrite [5,6,17]. Other drugs have been used to activate mitoK_ATP channels, but suffer from non-specific effects or other limitations. In addition to peroxynitrite, ATP and ADP are natural regulators of mitoK_ATP channels [16]. Nonetheless, there is a critical need for more information concerning the exact structure of the mitoK_ATP channel as well as on the normally occurring, regulatory factors.

Diazoxide, a drug used against acute hypertension or hypoglycemia in people, is the most commonly used mitoK_ATP channel opener [5,6,12,13,17]. Diazoxide is more potent than the non-specific K_ATP channel activator cromakalim by a factor of 1000 against mitoK_ATP channels [19]. Another effect of diazoxide at high doses is the inhibition of succinate dehydrogenase (SDH; complex II of the electron transport chain), which we believe causes the liberation of ROS independent of mitoK_ATP channel opening [5]. This view is supported by examination of the effects of the specific inhibitor of SDH, 3-nitropropionic acid (3-NPA), which increases ROS production by mitochondria [20].

It is our opinion that diazoxide is the most potent inducer of preconditioning due to its combined effects on mitochondrial membrane depolarization and ROS production. First, ROS production due to SDH inhibition is able to enhance mitoK_ATP channel opening [17,20] so that the maximal mitochondrial membrane depolarization is larger than that seen with the more selective BMS-191095. Second, the effects of diazoxide probably represent activation of separate preconditioning pathways due to mitochondrial membrane depolarization and enhanced ROS production through SDH inhibition. Nonetheless, it is apparent that at relatively low doses diazoxide is very selective for mitoK_ATP channels but even at relatively high doses diazoxide is selective for mitochondria [5,17,20,21].

Another agent, BMS-191095, is very selective for mitoK_ATP channels and does not inhibit SDH or augment basal ROS production at the doses used [6,20]. Similarly, we showed that BMS-191095 has no effects on the plasma membrane potential of cultured neurons [6,14] or freshly dissociated cerebral vascular smooth muscle cells [14]. While there has been no documentation of non-specific effects of BMS-191095, only a small number of studies have been conducted and more research is needed to rule out potential targets distinct from mitoK_ATP channels.

The mitoK_ATP channel effects of diazoxide and BMS-191095 are blocked by 5-hydroxydecanoate (5-HD) or glibenclamide [20,22–24]. A potential concern about 5-HD is that it may not be a direct antagonist of mitoK_ATP channels but rather needs to be metabolized into a form which is antagonistic to the actions of diazoxide and BMS-191095 [25,26]. Therefore, 5-HD may not be effective in some experimental models or under some protocols because of lack of penetration to the mitochondria, failure for the conversion to its active form, or other yet unknown, competing aspects of the experimental procedures. Nonetheless, 5-HD is specific for mitochondria while glibenclamide inhibits K_ATP channels located both in mitochondria and in the cell membrane.

In the first study specifically targeting mitoK_ATP channels, we examined whether diazoxide treatment was able to protect the brain against ischemic stress in piglets [22]. Our results indicated that 5–10 µM diazoxide (a concentration insufficient to inhibit SDH or to have vasoactive effects), given 20 minutes prior to ischemia (immediate preconditioning), was able to completely protect neuronal function in piglets. These protective effects of diazoxide were blocked by co-administration of 5-HD. Extending these studies, we found that immediate preconditioning with diazoxide in piglets against ischemic stress was able to prevent mitochondrial swelling and limit calcium influx in brain cells [27]. Similar immediate neuroprotective results were obtained in neonatal rats with hypoxia-ischemic injury [24] and in adult rats with middle cerebral artery occlusion (MCAO) [23,28], as well as in cultured neurons [5,21,] and astroglia [29] exposed to anoxic and chemical challenges. Diazoxide application also protects cerebral resistance blood vessels [30], the microcirculation, and the blood-brain barrier [31] against ischemic stress. In cultured neurons and astroglia, diazoxide treatment also induced delayed preconditioning against oxygen-glucose deprivation and/or glutamate excitotoxicity [5,21,29]. The major effect of delayed preconditioning with diazoxide appears to be the suppression of the increase in ROS upon exposure to lethal stresses such as glutamate [5,21]. In these in vivo models, immediate preconditioning reduced the impairment of neurological function or the extent of infarcted tissue following ischemia, while in the in vitro models preconditioning increased the survival of neurons and astroglia.

We have shown that BMS-191095 also is able to confer both immediate and delayed preconditioning in cultured rat neurons [6,32]. Similar to diazoxide, BMS-191095 reduces the ROS availability upon exposure to potentially lethal stimuli [6,32]. In contrast to diazoxide, BMS-191095 does not increase mitochondrial ROS levels upon its application to either isolated mitochondria or cultured neurons [6,20]. Thus, mitoK_ATP channel opening can be dissociated from mitochondrial ROS production and mitoK_ATP channel opening alone is sufficient for inducing preconditioning.

BMS-191095 was also able to protect the adult rat brain in vivo and these effects are blocked by co-treatment with 5-HD [14]. Nonetheless, it seems likely that mitoK_ATP channel opening and ROS production are synergistic in inducing preconditioning and that in the case of diazoxide, it may not be possible to separate these dual effects.

2.2 Large conductance calcium activated potassium channels

The typical plasmalemmal BK_Ca channel is composed of pore-forming α subunits (BK_Caα) and modulatory β subunits (BK_Caβ) [33]. Following the report that BK_Ca channels are also located in the inner membrane of mitochondria, several studies have reported immediate and delayed preconditioning in the heart via the activation of BK_Ca channels [34,35]. In these studies, preconditioning usually was induced by administration of the benzimidazole derivative NS1619 (1,3-dihydro-1-[2-hydroxy-5- (trifluoromethyl)phenyl]-5-(trifluoromethyl)-2H-benzimidazol-2-one), which is the most commonly used agonist against BK_Ca channels.

In the brain, we made the first observation that NS1619 could protect against ischemic stress [39]. However, the in vivo experimental model did not allow direct determination of the cellular site of action of NS1619. The protective effect of the activation of BK_Ca channels was also shown in the brain when NS1619 was given after the initiation of injury [37,38]. A major limitation of NS1619 is its relative non-selectivity. In addition to activating BK_Ca channels, NS1619 inhibits Complex I of the mitochondrial respiratory chain [40], blocks L-type Ca²⁺ and K_V channels [41], and releases Ca²⁺ from intracellular pools [42]. Therefore, in our view, the concept that neuronal preconditioning can be induced by targeting BK_Ca channels has not been critically evaluated. For example, there is no evidence that NS1619 is specific for BK_Ca channels in any cellular/subcellular location. A concern from an early paper is that the molecular weight of the putative mitochondrial BK_Caα subunit identified was much less than expected [34], and the authors did not provide a reasonable explanation for this discrepancy. A more recent study demonstrated the presence of a BK_Caα subunit of appropriate molecular weight (125 kDa), and also provided evidence that it was located on the inner membrane of brain mitochondria [36]. In our study, we could confirm the presence of the pore forming BK_Caα subunit in brain homogenates using commercially available antibodies but could not detect either the 125 kDa or the putative 55 kDa proteins in isolated mitochondria or in mitochondrial enriched samples [15]. Thus, the existence and functional importance of BK_Ca channels remain unclear. On the other hand, our recent work supports the view that the preconditioning effects of NS1619 are de to disruptive effects on the electron transport chain. While we could elicit ROS generation and depolarization in brain mitochondria, and induce preconditioning in neurons with NS1619, none of these effects could be antagonized with any of the traditional BK_Ca channel blockers (paxilline, 4–aminopyridine, charybdotoxin, iberiotoxin) [15]. Conversely, these antagonists blocked the plasma membrane hyperpolarization following NS1619 application in neurons as expected [15]. Support for a mechanisms not involving mitoBK_Ca channels comes from experiments in which co-application of ROS scavengers with NS1619 abolished preconditioning in both neuronal and heart preparations [15,35]. As with diazoxide and BMS-191095, the development of preconditioning with NS1619 is due to increased resistance to oxidative stress.

3.0 Respiratory chain enzymes

Although a very efficient system, superoxide anion is continuously released at different locations along the electron transport chain under normal conditions (Figure 1). Superoxide anion can either act within the mitochondria as a signaling molecule or following conversion to hydrogen peroxide, is able to influence events in the cytosole. Unlike superoxide anion, which is unable to leave the matrix, hydrogen peroxide is able to traverse the inner mitochondrial membrane through aquaporin-like channels.

In a recent review, Zhang and Gutterman [43] describe possible intra-mitochondrial sources of ROS. Within the four protein complexes associated with the respiratory chain, the primary sites of ROS production and release are Complex I, Complex II, and Complex III. Complex I (NADH-ubiquinone oxidoreductase) and Complex II (succinate dehydrogenase, SDH) are where electrons are accepted from NADH+H⁺ and FADH_2,, respectively, and transferred to Complex III (ubiquinol-cyctochrome c oxidoreductase) and finally to Complex IV (cytochrome c oxidase), where the final electron acceptor is oxygen and the final product is water (Figure 1).

To examine mitochondrial sites of ROS production, many studies have relied on inhibitors of the electron transport chain. The most commonly used agents include rotenone (Complex I inhibitor), thenoyltrifluoroacetone or 3-NPA (Complex II inhibitors), and myxothiazol, carbonylcyanide m-chlorophenyl hydrazone, and stigmatellin (Complex III inhibitors). A potential problem with these agents is that the interruption of mitochondrial function may alter other aspects of cell metabolism such as ATP synthesis, and thus complicate the interpretation of results. Another concern is that agents such as rotenone may either increase or decrease ROS production, depending upon the cell type investigated and the metabolic status of the cells [44]. Nonetheless, when used in conjunction with other approaches and appropriate control techniques, these inhibitors can provide specific information concerning intramitochondrial sources of ROS.

Inhibition of SDH with 3-NPA is able to induce delayed preconditioning in rats when administered 3 days prior to temporary occlusion of the MCAO [45]. Thus, treatment with 3-NPA reduced infarct volume by about 20%. On the other hand, 3-NPA did not induce immediate preconditioning in this same model (unpublished observations). We are unaware of similar studies targeting protein complexes of the electron transport chain in the brain. A surprising finding was that 3-NPA induced alterations of mitochondrial membrane potential in cultured neurons [45]. Since 3-NPA does not depolarize isolated brain mitochondria [20], it appears that cytosolic mechanisms are necessary for mitochondrial ROS to activate mitoK_ATP channels in intact neurons. Based upon our earlier studies [17], peroxynitrite, which represents the chemical union of superoxide anion and nitric oxide (NO), may be a more potent activator of mitoK_ATP channels than superoxide ion. These results strongly suggest that secondary opening of mitoK_ATP channels plays a key role as the trigger in the development of 3-NPA-induced tolerance in the brain against ischemic stress. Moreover, respiratory chain inhibition by any means results in decreased ATP generation by the disruption of H⁺ translocation into the intermembrane space and a consequent reduction of proton motive force. A resultant increase in the ADP/ATP ratio is expected to open the mitoK_ATP channels and activate cytoprotective mechanisms.

Our recent studies using NS1619 indicate that this agent induced neuronal preconditioning by increased ROS production and mitochondrial depolarization independent of activation of a putative mitoK_Ca channel [15]. Similar to 3-NPA, these effects appear to be secondary to respiratory chain inhibition [40,46].

While most studies have focused on production of ROS by mitochondria, recent work from our laboratory has shown that reactive nitrogen species (RNS) are also produced by mitochondria [17,47,48]. These RNS are produced via an arginine-independent pathway in mitochondria, and are degraded by catalysts of peroxynitrite decomposition. The basal level of RNS production is abolished with rotenone and restored with Complex II substrates such as succinate. In contrast, 3-NPA, which produces ROS, does not increase levels of RNS in isolated mitochondria [47,48]. On the other hand, NO diffusing into mitochondria from the cytosole is able to combine with superoxide anion to form peroxynitrite. We and others have not been able to document the existence of a mitochondrial NO synthase [47,49,50]. The contribution of mitochondrial RNS to preconditioning has never been directly investigated. However, peroxynitrite is an endogenous inducer of opening of the mitoK_ATP channel [17] and thus may lead to preconditioning through this mechanism or via a direct oxidative mechanism.

4.0 Substrate limitation

Glucose is the major energy source of the central nervous system. Due to the relatively large energy requirements of the brain, the inability of the brain to store glucose as glycogen in significant amounts, and transport restrictions of the blood-brain barrier, even brief periods of reduced CBF or glucose deprivation rapidly depletes brain levels of ATP. While severe hypoglycemic episodes are known to cause neurophysiologic and intellectual deficits, coma, and even death in humans [51], mild metabolic stress, such as that caused by dietary restriction or fasting can protect the brain against ischemic stress [52]. Thus, reduced caloric intake, moderate hypoglycemia, and brief periods of ischemia or anoxia can reduce ischemic brain damage, decrease the risk of neurodegenerative diseases, and/or increase the lifespan of laboratory animals [53,54].

Brief periods of ischemia or anoxia were the original stimuli found to lead to the development of immediate and delayed preconditioning and thus lead to the descriptive terms of “ischemic preconditioning” or “ischemic tolerance” [55,56]. The global nature of these stimuli leads to the activation of multiple signaling pathways and it is unclear to what extent factors such as anoxia and energy substrate depletion are contributing to preconditioning. In our recent experiments, we determined whether transient withdrawal of the major energy substrate, glucose, would induce neuroprotection in primary neuronal cultures [57]. Energy deprivation was achieved by replacing the regular cell culture medium with a medium lacking both glucose and amino acids. In order to adjust the osmolarity we used mannitol, which neurons cannot metabolize, to substitute for the osmotic activity of the eliminated glucose and amino acids. The incubation of neurons in this solution did not provide acute protection, but resulted in delayed tolerance against various insults such as oxygen-glucose deprivation, glutamate excitotoxicity and exogenous hydrogen peroxide toxicity. The acute effect of energy deprivation was mitochondrial membrane depolarization and reduced ATP production. To exclude the possibility that the neuroprotection was caused by some unknown effect of mannitol on neurons, we performed experiments with another non-metabolizable sugar, sucrose. The protection afforded by the sucrose-supplemented medium was similar to that of the mannitol-containing solution. In contrast, supplying the medium with either glucose or lactate instead of mannitol completely eliminated the protection. These results suggest that the neuroprotection was induced by the lack of energy and not by the non-metabolizable sugars. One of the major findings of these experiments was that replacing the mannitol-containing medium with the regular cell culture medium resulted in the restoration of ATP content within one day while mitochondria remained depolarized for more than 24 hours. We speculate that the latter effect might be the consequence of secondary activation of mitoK_ATP channels due to reduced levels of ATP and a subsequent K⁺ influx into the mitochondrial matrix. As with direct activation of mitoK_ATP channels and mitochondrial ROS generation, energy depletion also reduces intracellular calcium increases and ROS production upon exposure of neurons to glutamate.

5.0 Mechanisms of cell death and protection

The sequence of events affecting the brain following asphyxia or blood flow insufficiency is widely understood. Metabolic failure due to inadequate oxygen supply results in anoxic depolarization that induces a subsequent Ca²⁺ inflow from the extracellular space into the cytosol and mitochondria of neurons [58]. In particular, mitochondrial Ca²⁺ appears to be the most important link between elevated intracellular Ca²⁺ levels and neuronal and astroglial cell death. In addition to being the major site of ATP production and numerous metabolic processes, the mitochondria also are involved in regulating intracellular Ca²⁺ levels; respiring polarized mitochondria rapidly accumulate Ca²⁺ from the cytosol [59]. However, excessive mitochondrial Ca²⁺ accumulation increases generation of ROS [60] which disrupts the electron transport chain, resulting in decreased ATP production and immediate cell death. Mitochondrial Ca²⁺ overload also leads to the opening of the mitochondrial permeability transition pore and the release of apoptosis-inducing factors [61,62]. Thus, anoxia or ischemia leads to necrosis and/or apoptosis in the central nervous system.

There are several excellent in vitro models that mimic one or more of the pathological features of ischemic brain injury. For example, glutamate application leads to necrotic cell death in cultured neurons due to NMDA receptor activation, calcium influx, and excess ROS production [6]. In a model described as perhaps the best for mirroring the in vivo situation, oxygen-glucose deprivation followed by recovery with oxygen and glucose replacement also leads to necrotic death of neurons, but through mechanisms less dependent upon NMDA receptor activation but still due to enhance ROS production [6]. In contrast, hydrogen peroxide exposure leads to neuronal cell death through combined apoptosis and necrosis depending upon dose and exposure time [6], due to direct effects of the applied ROS.

6.0 Mechanisms of Preconditioning

Preconditioning has been reported to occur in almost all cell types studied and is undoubtedly a normal physiological mechanism which is initiated in response to biological, nutritional, behavioral and pharmacological stimuli. In addition to protecting against cell death, “preconditioning-like” responses also represent beneficial adaptations to changing but non-lethal conditions so that cells maintain appropriate functioning for situations such as altered nutritional or dietary status or exercise load. Preconditioning stimuli are diverse and can target cytosolic as well as mitochondrial sites with the subsequent activation of signaling pathways which can be independent, interacting, and/or duplicative of each other (Figure 1). Nonetheless, the final result of these initiating events and signaling pathways is to prepare the cells to withstand stress.

To illustrate different mechanisms, we will present two examples of neuronal preconditioning which represent recent work by our laboratory. These examples represent delayed preconditioning in cultured, primary cortical neurons following: 1) selective activation of mitoK_ATP channels with BMS-191095 [6]; and 2) enhanced production of ROS by the electron transport chain following application of NS1619 [15].

6.1 Selective mitoKATP channel opening with BMS-191095 application

BMS-191095 is one of the most selective mitoK_ATP channel opener available and no major, non-specific effects of this drug have been reported. BMS-191095 induces both immediate and delayed preconditioning in cultured neurons and brain [6,14,32], as well as in heart and skeletal muscle [64,65]. Several signaling events occur immediately following activation of mitoK_ATP channels with BMS-191095 in neurons, such as mitochondrial depolarization, a transient increase in the level of free cytosolic calcium and the phosphorylation of Akt, Gsk3β, PKC, but not ERK 1/2, JNK 1/2, or p38 [6,32]. These changes probably also explain immediate preconditioning with BMS-191095 although we have not examined precise mechanisms involved. Nonetheless, our results using diazoxide suggest that attenuation of calcium influx and prevention of mitochondrial swelling are important components of immediate preconditioning [27]. Furthermore, these initial signaling events also will lead to delayed preconditioning such that the neurons are protected against lethal stresses such as glutamate excitotoxicity for several days following the application of BMS-191095[6,32]. The principal events associated with protection against neuronal cell death include sustained mitochondrial depolarization, attenuation of the increase in the ROS surge, and maintained ATP levels [6]. The lack of an increase in ROS production by neurons or isolated brain mitochondria following BMS-191095 application and the failure of a ROS scavenger to block preconditioning when co-applied with BMS-191095 support the concept that only mitoK_ATP channel opening and not an increase in ROS levels is the primary, initiating event associated with effects of this drug [6]. While we do not know all the mechanisms involved, augmented cellular levels of catalase, an important ROS scavenger, appears to be a major factor in delayed neuroprotection. BMS-191095 application leads to enhanced levels of mRNA for catalase, increased protein levels of catalase, and augmented activity of catalase in neurons [6]. The link between BMS-191095 application and catalase up regulation probably involves the phosphoinositide 3 kinase (PI3K) system. Inhibition of PI3K with wortmannin did not block BMS-191095 effects on mitochondrial depolarization, but prevented the increased in catalase levels and blocked protection of neurons against glutamate excitotoxicity. Additionally, inhibition of catalase with 3-aminotriazole in BMS-191095 treated neurons dose-dependently increased glutamate-mediated neuronal cell death [6]. Thus, sustained depolarization and ATP homeostasis together with the activation of the PI3K–Akt–Gsk3beta axis and upregulation of catalase, represent multiple effects of a single initiating event, namely mitoK_ATP channel opening, which combine to protect neurons against otherwise lethal stress [6].

6.2 Electron transport chain inhibition with NS1619

Although investigators have referred to NS1619 as selective for BK_Ca channels in general or more recently as a selective mitoK_Ca channel opener, our evidence and that of other authors indicate that this is not the case (see previous section) [6,40,41,42,46,65]. Our results support the view that the primary mitochondrial effect of NS1619 is inhibition of Complex I of the electron transport chain rather than activation of a putative mitoK_Ca channel. On the other hand, irrespective of the site of action, NS1619 is an effective agent for causing immediate (unpublished observations) and delayed preconditioning and protects neurons against oxygen-glucose deprivation, H₂O₂, or glutamate excitotoxicity [15]. Application of NS1619 to neurons has multiple immediate effects, including increased ROS generation, depolarization of isolated mitochondria, decreased ATP levels, hyperpolarization of the neuronal cell membrane, and activation of the PI3K signaling cascade. We believe that mitochondrial depolarization is secondary to increased ROS levels or to alterations of the ADP/ATP ratio. To investigate the relative importance of these events in the development of delayed preconditioning, we co-applied a ROS scavenger, blockers of BK_Ca channels, and an inhibitor of the PI3K system. While blockers of BK_Ca channels prevented hyperpolarization of the neuronal cell membrane, they were completely ineffective against delayed preconditioning when the neurons were challenged by prolonged exposure to exogenous hydrogen peroxide. Conversely, a superoxide dismutase mimetic or wortmannin completely blocked NS1619-induced delayed preconditioning in this model. In contrast to BMS-191095 [6], we could not detect increases in levels or activity of any ROS scavengers including catalase after NS1619 treatment. This finding for NS1619 is similar to those we obtained with diazoxide [5,21], where ROS levels upon glutamate exposure following preconditioning were blunted and neurons were protected against both OGD and glutamate but no obvious changes in anti-oxidant protein levels were found. On the other hand, similar to BMS-191095-induced preconditioning we demonstrated the activation of several downstream targets of PI3K such as Akt and Gsk3β. Furthermore, we found that NS1619 dose-dependently inhibited caspase-3/7. Thus, besides protecting against necrotic cell death as seen in the cases of OGD and glutamate excitotoxicity, preconditioning with NS1619 has the potential to prevent apoptosis induced by hydrogen peroxide. In summary, NS1619 is a potent inducer of delayed neuronal PC. However, the neuroprotective effect seems to be independent of plasmalemmal and mitochondrial BK_Ca channel opening. Thus, rather than involving mitoK_Ca channel activation, preconditioning by NS1619 appears to be the consequence of ROS generation through inhibition of the electron transport chain protein complexes, activation of the PI3K signaling pathway, inhibition of caspases, and development of resistance to oxidative stress.

7. Disruption of preconditioning mechanisms

A number of pathophysiological stressors such as ischemia/reperfusion are able to disrupt normal functioning of plasmalemmal K_ATP channels in the neurovascular unit [66–68] and therefore it is not surprising that mitoK_ATP channels would similarly be affected. For example, nutritional and genetic models of insulin resistance reduce function of several potassium channels in cerebral vascular smooth muscle, including K_ATP channels, through mechanisms involving enhanced baseline vascular levels of ROS [66,67]. Similarly, ischemia and reperfusion are able to reduce K_ATP channel-dependent dilation in cerebral arteries [68]. Our recent work has indicated that immediate preconditioning is abolished in hearts from insulin resistant rats [69]. Thus, ischemic- as well as diazoxide-induced preconditioning fail to protect hearts from ischemia/reperfusion in Zucker obese rats while both of these approaches limit infarct size in hearts from non-insulin resistant Zucker lean rats. In these animals, a substantial of number of the mitochondria in the heart were swollen or had disruption of the normal pattern for cristae. An additional finding from this study was that infarct size was enhanced in the hearts from Zucker obese compared to lean rats, which is similar to what we [70] and others [71] have found in brains of insulin resistant, obese mice following MCAO. These results may indicate that normal protective mechanisms initiated at the level of the mitochondria are impaired in common disease states and thus the brain and other organs are more at risk during ischemic episodes. The potential for the uncoupling of the tight relationship between metabolic need and blood flow in the brain, due to cerebral vascular dysfunction associated with insulin resistance [66,67], as well as the elimination of normal, protective mechanisms involving mitochondria such as preconditioning, may account for the increased risk and severity of neurological diseases and strokes in patients suffering from the metabolic syndrome. Clearly, more work is needed to determine the mechanisms by which preconditioning mechanisms in the brain can be restored in sick patients.

Summary and Conclusions

Targeting mitochondrial-centered mechanisms is an effective means of protecting neurons against potentially lethal stimuli. The two primary initiating events, which can occur separately or together, are the opening of mitoK_ATP channels and the production of ROS from the electron transport chain. Following these early events in the mitochondria, intramitochondrial and intracellular signaling mechanisms lead to the protection of neurons during and following exposure to potentially lethal stresses by attenuating intracellular calcium influx, limiting the ROS surge, preserving ATP levels, preventing apoptosis, and maintaining normal mitochondrial morphology.

Abbreviations

Ca²⁺

calcium

ADP

adenosine di-phosophate

ATP

adenosine tri-phosphate

PKC

protein kinase C

Gsk3β

phospho-glycogen synthase kinase 3 beta

PI3K

phosphoinositide 3-kinase

Bad

Bcl-2 associated death promoter

NO

nitric oxide.