1. IntroductionAmyotrophic lateral sclerosis (ALS) or Lou Gehrig’s disease is a devastating neuromuscular disorder. Clinically, ALS is characterized by progressive skeletal muscle weakness and atrophy, leading eventually to respiratory failure and death typically within 2–5 years of diagnosis [1]. Pathologically, ALS is characterized by a progressive degeneration of motor neurons in the motor cortex, brainstem and spinal cord and corresponding retraction of motor axons away from the neuromuscular junctions [2,3]. Although the precise mechanisms underlying neurodegeneration and disease progression are unknown, evidence indicates that mitochondrial dysfunction and mitochondrial oxidative stress (MOS) are significant contributing factors in the death of motor neurons [4]. Mitochondria are primarily responsible for ATP production and are metabolically dynamic organelles that generate significant amounts of reactive oxygen species (ROS) as a by-product of electron transport. Dysfunctional mitochondria and elevated levels of MOS have been implicated in the pathogenesis of several neurodegenerative disorders including ALS [5]. The accumulation of oxidative stress in mitochondria leads to further damage of mitochondrial components resulting in a vicious feed forward cycle of dysfunction.Mutations in the antioxidant enzyme, copper/zinc superoxide dismutase (SOD1), have been implicated in mitochondrial dysfunction and are a major cause of familial ALS [6,7,8]. The G93A mutant human Cu, Zn-superoxide dismutase (G93A mutant hSOD1) mouse model of familial ALS develops an aggressive form of motor neuron disease reminiscent of human ALS [9,10]. These mice have been shown to develop several clinical and histopathological features of ALS including muscle atrophy, hind limb weakness and paralysis, loss of spinal cord motor neurons, and markers of neuroinflammation [9,10]. Several studies in the G93A mutant hSOD1 mouse have provided evidence for MOS and mitochondrial dysfunction. For instance, antioxidants and free radical scavengers targeted to the mitochondria mitigate the degeneration of motor neurons in these mice [11,12]. Glutathione (GSH) is an abundant endogenous tripeptide antioxidant with distinct cytosolic and mitochondrial pools. Depletion of GSH has been observed in whole spinal cord tissue and spinal cord mitochondria of end-stage G93A mutant hSOD1 mice; however, the mechanism underlying mitochondrial GSH depletion has not been investigated [13,14,15].Enzymes required for the synthesis of GSH are not found in the mitochondria, therefore GSH is exclusively synthesized in the cytosol and transported into the mitochondria [16]. GSH is transported into the mitochondria via a facilitated transport process involving mitochondrial inner membrane anion transporters such as dicarboxylate (DIC), 2-oxoglutarate (OGC) and tricarboxylate carriers [17,18,19]. Transport of GSH into the mitochondria from the cytosol is critical to maintaining mitochondrial function. Overexpression of OGC in a rat renal proximal tubular cell line (NRK-52E) enhanced mitochondrial GSH transport and showed significant protective effects against chemically induced apoptosis [20]. Previously, we have shown in primary cerebellar granule neurons that mitochondrial GSH transport is critical for cell survival and inhibition of DIC sensitizes these neurons to oxidative and nitrosative stress [21]. We have also shown that stable overexpression of the OGC transporter in NSC34 motor neuron-like cells was sufficient to increase mitochondrial GSH levels and provide protection from oxidative and nitrosative stress [22]. Furthermore, studies have shown the anti-apoptotic protein Bcl-2 to regulate cellular GSH levels [23]. Overexpression of Bcl-2 conveys an increase in cellular GSH levels, while conversely Bcl-2 knockout mice show a significant reduction in GSH levels [24,25]. Previously, we have shown that GSH transport through OGC is facilitated by a synergistic interaction with Bcl-2 resulting in enhanced transport of GSH into the mitochondria. A direct interaction between Bcl-2 and GSH facilitates GSH transport through OGC and provides protection from oxidative stress [22,26,27].

Here, we show that the Bcl-2/GSH interaction and GSH uptake in isolated rat brain mitochondria are both enhanced by wild type (WT) SOD1 but perturbed in the presence of G93A mutant SOD1 recombinant proteins in vitro. Disruption of the Bcl-2/GSH interaction results in a reduced capacity to transport GSH into the mitochondria. Using isolated mitochondria from the spinal cords of end-stage G93A mutant hSOD1 mice, we show a reduction in the mitochondrial GSH levels corresponding to a disruption of mitochondrial GSH uptake. Furthermore, we show that the OGC is heavily S-nitrosylated in mitochondria isolated from the spinal cords of end-stage G93A mutant hSOD1 mice, further disrupting transport of GSH into the mitochondria.

4. DiscussionMitochondria lack the enzymes necessary to synthesize GSH and therefore must rely on transporters, such as OGC, to import GSH from the cytosol [16]. Transport of GSH from the cytosol into mitochondria is essential for cell survival. Based on our previously published work [21,22,26,27] and the findings presented here, we conclude that OGC-dependent mitochondrial GSH transport is compromised in mitochondria of G93A mutant hSOD1 mice. The Bcl-2/GSH interaction and mitochondrial uptake of GSH is disrupted in the presence of G93A mutant SOD1 recombinant protein in vitro. Furthermore, mitochondria isolated from the spinal cord of end-stage G93A mutant hSOD1 mice show reduced GSH levels and a diminished capacity to take up GSH in vitro, effects that are associated with a marked S-nitrosylation of the OGC transporter.Previous studies have shown the ability of G93A mutant SOD1 to interact with Bcl-2 at mitochondria [28,29]. Indeed, Pedrini et al. showed that G93A mutant SOD1, but not WT SOD1, induces a conformational change in Bcl-2 that exposes its pro-apoptotic BH3 domain [29]. We have shown that GSH-binding by Bcl-2 facilitates an association with OGC to enhance OGC-dependent mitochondrial GSH transport [22,26,27]. With this in mind, we aimed to determine the effect of WT SOD1 and mutant SOD1 on the Bcl-2/GSH interaction. We observed that recombinant G93A mutant SOD1 diminishes the interaction between Bcl-2 and GSH, while recombinant WT SOD1 slightly enhances the Bcl-2/GSH interaction. Furthermore, incubation of rat brain mitochondria with WT SOD1 enhances GSH transport into the mitochondria when compared to controls; however, this effect is lost in mitochondria incubated with G93A mutant SOD1. These data suggest a novel function of WT SOD1 to enhance the Bcl-2/GSH interaction and facilitate mitochondrial GSH transport. In contrast, G93A mutant SOD1 appears to interfere with GSH-binding by Bcl-2 and consequently, decreases mitochondrial GSH transport. The precise mechanism underlying these differential effects of WT and G93A mutant SOD1 is currently unknown. Mutant G93A SOD1 retains its dismutase activity but is known to generate more hydroxyl radicals than WT SOD1 via a gain-of-function mechanism [30]. It is possible that hydroxyl radical production in the presence of G93A mutant SOD1 could reduce the capacity of isolated mitochondria to transport GSH. However, we favor an alternative explanation for this effect in that G93A mutant SOD1, but not WT SOD1, induces a conformational change in Bcl-2 as described previously by Pedrini et al. [29], and that this conformational change reduces the capacity of Bcl-2 to interact with GSH and facilitate OGC-dependent mitochondrial GSH transport.We next examined the mitochondrial pool of GSH within the spinal cord of G93A mutant hSOD1 mice at end-stage compared to NonTg littermate controls. Our results are consistent with previous reports [13,14,15] and show that total mitochondrial GSH is significantly depleted in the lumbar spinal cord of end-stage G93A mutant hSOD1 mice. This reduction in the mitochondrial pool of GSH may be a direct result of deficient mitochondrial GSH transport as mitochondria isolated from spinal cord of end-stage G93A mutant hSOD1 mice displayed a deficit in this function in vitro. We further found that spinal cord mitochondria required OGC for GSH transport and that the OGC transporter is heavily S-nitrosylated in spinal cord from end-stage G93A mutant hSOD1 mice.We hypothesize that the deficit in mitochondrial GSH transport observed in the spinal cord of end-stage G93A mutant hSOD1 mice may arise from multiple effects. First, mutant SOD1 disrupts the Bcl-2/GSH interaction resulting in reduced GSH uptake through the OGC transporter. Second, it has been previously reported that Bcl-2 expression is reduced, and/or Bcl-2 function is compromised, within the spinal cord of ALS patients and G93A mutant hSOD1 mice [31,32]. Given our data establishing an important regulatory role for Bcl-2 in OGC-dependent mitochondrial GSH transport [22,26], loss of Bcl-2 function could play a role in diminishing mitochondrial GSH transport in the mutant SOD1 mouse model of ALS. Third, oxidative modifications to OGC could directly impair mitochondrial GSH transport. In particular, OGC has three cysteine residues which play critical roles in its structure and function making this transporter a prime target for inactivation by S-nitrosylation [17].Our findings implicate the deficiency in spinal cord mitochondrial GSH transport observed in the G93A mutant hSOD1 mouse model as a potential contributing factor in ALS pathogenesis by either directly promoting motor neuron cell death or facilitating the neuroinflammatory phenotype of glial cells. In future studies, it would be interesting to evaluate mitochondrial GSH levels and mitochondrial GSH transport in skeletal muscle which is equally involved in ALS and in which oxidative stress also plays a detrimental role [33,34]. Furthermore, we predict that the deficiency in spinal cord mitochondrial GSH transport limits the therapeutic actions of known mitochondrial “protective” compounds and that rescue of OGC-dependent mitochondrial GSH transport could act synergistically with mitochondrial protective agents to significantly delay disease progression and prolong survival of G93A mutant hSOD1 mice. Identification of OGC-dependent mitochondrial GSH transport as a key deficiency in diverse models of ALS would reveal a novel site for therapeutic intervention. Moreover, strategies aimed at preserving OGC function and sustaining the mitochondrial GSH pool may represent new effective approaches to protect motor neurons in ALS.