Dysregulated mitochondrial Ca²⁺ and ROS signaling in skeletal muscle of ALS mouse model

Abstract

Amyotrophic lateral sclerosis (ALS) is a devastating neuromuscular disease characterized by motor neuron loss and prominent skeletal muscle wasting. Despite more than one hundred years of research efforts, the pathogenic mechanisms underlying neuromuscular degeneration in ALS remain elusive. While the death of motor neuron is a defining hallmark of ALS, accumulated evidences suggested that in addition to being a victim of motor neuron axonal withdrawal, the intrinsic skeletal muscle degeneration may also actively contribute to ALS disease pathogenesis and progression. Examination of spinal cord and muscle autopsy/biopsy samples of ALS patients revealed similar mitochondrial abnormalities in morphology, quantity and disposition, which are accompanied by defective mitochondrial respiratory chain complex and elevated oxidative stress. Detailing the molecular/cellular mechanisms and the role of mitochondrial dysfunction in ALS relies on ALS animal model studies. This review article discusses the dysregulated mitochondrial Ca²⁺ and reactive oxygen species (ROS) signaling revealed in live skeletal muscle derived from the ALS mouse models, and a potential role of the vicious cycle formed between the dysregulated mitochondrial Ca²⁺ signaling and excessive ROS production in promoting muscle wasting during ALS progression.

Introduction

Amyotrophic lateral sclerosis (ALS) is a rapidly progressive neuromuscular disease characterized by motor neuron death and devastating skeletal muscle wasting. The overall worldwide ALS incidence was reported at 1.75 (1.55–1.96)/100,000 person-years of follow-up [1]. A population-based study in the United Kingdom also showed that lifetime risk of ALS was 1 in 472 for women and 1 in 350 for men [2]. Currently there is no effective treatment for ALS [3]. The expected life span for the majority of ALS patients is usually 3–5 years after disease onset [4]. Despite intensive research effort (to the date of Aug 20, 2018, over 23,500 ALS research articles have been published since 1897), the pathogenic mechanism underlying progressive neuromuscular degeneration in ALS remains obscure [5]. Majority of ALS cases are sporadic (SALS), while about 5~10% cases exhibit an autosomal dominant inheritance, termed familial ALS (FALS) [6]. Although multiple factors could contribute to SALS, both SALS and FALS share similar pathological and clinical phenotypes of neuromuscular degeneration [4, 7], indicating a common mechanism underlying neuromuscular degeneration during ALS disease progression. While the death of motor neurons in the motor cortex of the brain and spinal cord is a defining pathological feature [8], accumulating evidences suggest that defects in other cell types or organs also have critical impact in ALS disease progression [9–21].

As one of the largest tissues in human body, skeletal muscle is responsible for voluntary movements of the entire body and is also essential for maintaining the homeostasis of the whole-body metabolism [22]. ALS patients experience progressive and severe muscle degeneration. Such devastating muscle wasting plays a significant role in the disease progression and life quality of ALS patients [23]. Muscle fibers and the motor neuron communicate at the neuromuscular junction (NMJ). While muscle function is largely controlled by motor neuron, retrograde signals are conducted from muscle back to motor neuron [24, 25]. Although skeletal muscle is a victim of motor neuron axonal withdrawal in the course of ALS progression, studies from us [11, 14, 15, 26] and other research groups [10, 12] support skeletal muscle as a primary recipient of ALS-causing factors. Investigating muscle degeneration at different disease stages shall enhance our understanding of ALS pathophysiology and facilitate the development of therapeutic approaches to sustain muscle function for alleviating the disease progression and improving the life quality of ALS patients.

For a cell, mitochondria are a major source of ATP and reactive oxygen species (ROS, as a byproduct of oxidative phosphorylation) [27]. They are not only essential for energy supply but also determine the survival or death of cells. Because of high demands of energy, neurons are highly dependent on mitochondria for physiological functions. Mitochondrial dysfunction is a major player in neuronal degeneration in ALS. There are multiple recent excellent reviews covering this topic [28–38]. In neurons, mitochondria occupy about 3–8% of the cell volume, and ~50% of the energy is utilized by the Na⁺/K⁺ ATPase in maintaining membrane potential and recovery from action potential [39, 40]. Similar to the nervous system, skeletal muscle is a highly energy-demanding organ with mitochondria occupying about 10–15% of the muscle fiber volume [41]. About 10–25% of ATP in muscle is used for sarcoplasmic reticulum Ca²⁺-ATPase, 65–80% for actomyosin ATPase, and 5–10% for Na⁺/K⁺ ATPase [42]. Molecular and functional studies on mitochondrial defects in skeletal muscle are relatively behind those in neurons, possibly because of the traditional, motor neuron centered view of ALS onset. In recent years, more research efforts shifted the attention to the role of muscle in ALS disease onset and progression, and particularly the contribution of mitochondrial dysfunction in muscle degeneration during ALS progression. Muscle uses Ca²⁺ as a messenger to control cellular events ranging from muscle contraction to cell death. While mitochondrial function in muscle is fine-turned by the intracellular Ca²⁺ signaling [43], mitochondria also have the ability to shape intracellular Ca²⁺ signaling profiles in skeletal muscle [14, 44, 45]. There is always a crosstalk between mitochondrial Ca²⁺ and ROS signaling [46]. It is now agreed that low level mitochondrial ROS serves as an important signaling molecule regulating various physiological process [47], while long lasting and excessive production of ROS put cell under oxidative stress, which is detrimental to cell function [48]. This current review discusses the abnormal mitochondrial Ca²⁺ and ROS signaling revealed in live skeletal muscle derived from ALS mouse models, and a potential role of the vicious loops formed between the dysregulated mitochondrial Ca²⁺ signaling and excessive ROS production in promoting muscle wasting during ALS progression.

1. Abnormal mitochondrial morphology and oxidative stress revealed in skeletal muscle samples derived from ALS patients

The ultrastructural studies of spinal cord and muscle autopsy/biopsy samples derived from both SALS and FALS patients revealed remarkable defects in mitochondrial morphology. Early study on the lumber spinal cord of ALS patients by Sasaki and Iwata observed aggregated dark mitochondria in the anterior horn neurons [49]. In the soma of these neurons, mitochondria could also be swollen, accompanied by markedly increased cristae. The multilayer cristae even stacked to form filamentous structures in some cases [49]. The swollen and vacuolated mitochondria with aggregate of tubules, broken cristae, and inclusion bodies were also evident in motor neurons and intra-muscle nerves in other studies [50–52]. Interestingly, the detailed ultrastructural study on muscle samples of ALS patients happened years earlier than studies on motor neurons. In 1966, Afifi et al gave the first description of abnormal mitochondrial disposition and morphology in atrophic muscles of ALS patients [53]. Specifically, in the subsarcolemmal regions, mitochondrial aggregates caused outpouching of the sarcolemmal membrane. In the sarcomere regions, pairs of large, elongated mitochondria parallel to the Z line were observed at the boundaries of degenerating sarcomeres. Ectopically large mitochondria with disorganized and/or discontinuous cristae were spotted at interfilamentous space [53]. Decreased number of mitochondria was also notable in degenerated muscle fibers from ALS patients by Fidziańska et al [54]. The swollen mitochondria characterized by dilatation and disruption of the cristae were further evident in ALS patients’ muscle by Napoli et al (2011) [55] and Chung et al (2002) [56].

Biochemical studies revealed impaired mitochondrial oxidative metabolism in muscle samples derived from ALS patients. Wiedemann et al (1998) incorporated enzyme activity and respiration measurement to demonstrate that NADH dehydrogenase (respiratory chain complex I) had significantly reduced activity. The reduced activity of NADH dehydrogenase was accompanied by decreased maximal respiration capacity, indicating impairment of mitochondrial function in skeletal muscle of SALS patients [57]. Later studies by Vielhaber et al (1999) and Safa Al-Sarraj et al (2014) also reported that cytochrome c oxidase (COX, or respiratory chain complex IV) had reduced activity in ALS patients’ muscle, and the distribution of mitochondria with respiratory chain defects was very heterogeneous [58, 59]. Retrospective histochemical and biochemical studies on the muscle biopsy derived from both FALS and SALS patients have shown that the deficiency of mitochondrial respiratory activity (COX) is correlated with the progression of ALS [60, 61]. Other studies further confirmed the reduced activity of mitochondrial respiratory chain complexes in muscle samples derived from SALS patients [62].

The impairment of mitochondrial oxidative metabolism in skeletal muscle of ALS patients was found to be accompanied by mitochondrial DNA lesions [63–65]. Mitochondrial DNA is vulnerable to ROS mediated oxidative damages, and defects in mitochondrial respiratory chain may boost ROS production [66, 67]. Similar defects in mitochondrial DNA and respiratory chain function were also identified in the spinal cord of ALS patients [68]. Using microarray technology and gene regulatory network analysis, Bernardini et al (2013) systematically identified mitochondrial network genes whose expressions were significantly altered in muscles of ALS patients. Remarkably, the gene network observed in ALS muscle includes genes whose functions connect the muscle structure definition to mitochondrial oxidative phosphorylation and ATP synthesis [69]. Another study used multigene qRT-PCR to demonstrate that in addition to the respiratory chain components (COXIV), the expression levels of genes regulating mitochondrial biogenesis and dynamics were also downregulated in skeletal muscles of ALS patients [70]. Results of those biochemical and multigene expression analysis suggest a potential role of mitochondrial oxidative stress in the occurrence of multi-faceted mitochondrial defects during ALS progression.

Although most ALS are sporadic cases without identified genetic causes, the spinal cord and muscle autopsy/biopsy samples from both sporadic and familial ALS patients all show analogous defects in morphology and biochemical properties of mitochondria. This indicates abnormal mitochondria as a common player in neuromuscular degeneration despite the etiology. As those studies could be only conducted with autopsy/biopsy samples derived from ALS patients, it is not known whether mitochondrial damage is a cause or a consequence of ALS. Answers to these questions rely on ALS animal model studies.

2. Disease-stage-dependent changes in ROS-related mitochondrial dysfunction in skeletal muscle of ALS mouse models

The rodent models expressing human ALS mutations recapitulate many features of the human disease [71–74], and were widely used for investigating pathogenic mechanisms of ALS and for testing preclinical therapies for ALS [72, 75–78]. Similar to the ultrastructural characteristics observed in the spinal cord of ALS patients, the predominant mitochondrial abnormalities were also seen in the spinal cord ventral horn of the transgenic mice with over expression of various ALS-associated human superoxide dismutase 1 (SOD1) mutations [79–81]. The electron microscopy study also revealed that morphological changes of mitochondria in skeletal muscle of ALS SOD1 mutant mice were similar to those observed in human patients [15, 81, 82]. The altered mitochondrial morphology could be associated with mitochondrial respiratory function defects, which are a major source of ROS production [83–85]. Indeed, the molecular and biochemical studies revealed oxidative stress as an essential feature of ALS muscle pathology [18]. The study of Halter et al (2010) showed that accumulation of ROS in skeletal muscle happens early at asymptomatic stages in mutant SOD1 mice [86]. The biochemical studies also link the oxidative stress to abnormal mitochondrial oxidative phosphorylation in skeletal muscle of ALS mouse models. The maximal oxygen consumption rate and ADP affinity significantly decreased in oxidative muscle of SOD1^G93A mice, implying defects in mitochondrial oxidative phosphorylation [87]. In addition, elevated NADH content and PGC-1α expression were detected in skeletal muscle from MLC/hSOD1G93A mice, implying a mitochondrial metabolism to a more oxidative phenotype [10]. PGC-1α is known to be induced by oxidative stress and helps modulate intracellular ROS levels [88] through what is known as retrograde mitochondria-nucleus signaling [89]. The proteomics study by Capitanio et al (2012) suggests that ROS accumulation in ALS SOD1^G93A skeletal muscle could result from an enhanced mitochondrial oxidative metabolism [90]. As observed in skeletal muscle of ALS patients, multiple signaling pathways could be involved in mitochondria-related oxidative stress in skeletal muscle of ALS mouse models (reviewed in Loeffler et al 2016) [18]. The similar morphological and biochemical changes of mitochondria between muscle and neuron, and between ALS patients and animal models justify the usage of ALS animal model to further understand molecular and cellular mechanisms linking mitochondrial dysfunction to neuromuscular degeneration during ALS progression.

Using chemical ROS indicators (ROS Brite^™ 570 and MitoSOX^™ Red), our group (Xiao et al, 2018) detected enhanced intracellular and mitochondrial ROS levels in live skeletal muscle fibers derived from SOD1^G93A mice before and after ALS disease onset [91]. Due to their non-ratiometric and non-reversible nature, those chemical ROS indicators do not provide quantitative evaluation of the ROS production, do not follow dynamic changes of mitochondrial ROS production in live muscle cells, and do not reveal the correlation between dysregulated mitochondrial ROS production and ALS disease progression. To examine the disease-stage-dependent changes of mitochondrial dysfunction in ALS skeletal muscle, we generated double transgenic mouse model (G93A/cpYFP) that carries human ALS mutation SOD1^G93A and mt-cpYFP transgenes [91], in which mt-cpYFP detects dynamic changes of ROS-related mitoflash events at individual mitochondrion level [84, 85, 92–94]. Through examining live skeletal muscle fibers derived from these double transgenic mice, disease-stage-dependent alterations in mitoflash events were revealed. Before ALS symptom onset, the frequency of mitoflash events is doubled, while other kinetic parameters of the signal remain the same as the control. After disease onset, the duration of mitoflash events is significantly prolonged [91]. It is known that mitoflash activity is dynamically coupled with mitochondrial transition pore (mPTP) opening in both cardiac and skeletal muscle, because the transient increase in mt-cpYFP fluorescence was always coupled with the depolarization of mitochondrial inner membrane potential at the same location [84, 95, 96]. Thus, the elongated duration of mitoflash events suggests potential changes in functional status of mPTP in skeletal muscle during ALS progresses. It has been shown that CypD promotes mPTP opening [97–99] and there is a strong connection between CypD-related mPTP opening and mitoflash signaling [84, 93, 96, 100]. Remarkably, the elongated mitoflash activity is associated with increased expression levels of CypD in muscle mitochondria in the later stage of ALS, linking CypD-related mPTP opening to elevated ROS production in ALS muscle [91]. We speculate that increased CypD level in mitochondria of SOD1^G93A muscle may contribute to the elongated and irreversible opening of mPTP at the later stage of ALS. Remarkably, through overexpression of ALS mutation SOD1^G93A in skeletal muscle of normal mice, we found that ALS mutation directly led to an enhanced mitochondrial ROS production and mPTP-related mitoflash activity in the absence of motor neuron withdrawal [91], demonstrating muscle mitochondria as a primary target of ALS mutation.

Although in recent years mt-cpYFP has been characterized and used as a biomarker of mitochondrial ROS generation [93, 94, 101], debates exist about its sensitivity to ROS and pH [102–104]. cpYFP may report both ROS- and pH-related mitochondrial signaling [95, 105], and is a robust indicator of the dynamic status of mitochondrial function [85, 92, 94, 106, 107]. In the case of studying ALS, the cpYFP mitoflash signal can be used as a quantitative biomarker for evaluating ROS-related mitochondrial dysfunction during ALS disease progression, and the double transgenic G93A/cpYFP mice could provide a unique animal model for testing whether potential therapeutic means could reverse or slow down the mitochondrial dysfunction by examining mitoflash activities.

3. Abnormal mitochondrial dynamics and autophagy activity revealed in live skeletal muscle fibers of ALS mice

Mitochondria are dynamic organelles that are constantly remolded by fusion and fission processes, which define mitochondrial morphology, distribution and function [31, 108, 109]. Abnormal mitochondrial dynamics are implicated in various neurodegenerative disorders [110, 111]. While oxidative stress could lead to mitochondrial fragmentation, mitochondrial dynamics also modulate ROS generation [112–115]. Overexpression of ALS mutant SOD1 in cultured motor neurons led to abnormal mitochondrial dynamics and cell toxicity [116, 117]. Through overexpression of a mitochondria-targeted photoactivatable fluorescent protein (mt-PAGFP) in skeletal muscle of SOD1^G93A mice, our group (Luo et al, 2013) examined the migration rate of the PAGFP inside mitochondria network in live muscle fibers, and revealed less-connected mitochondrial networks with reduced rate of fission/fusion in SOD1^G93A muscle. Those abnormalities occur in asymptomatic ALS stage, indicating abnormal mitochondrial network and dynamics as early pathological events in ALS muscle [11]. Mounting evidence indicates a vicious cycle between abnormal mitochondrial dynamics and excessive ROS generation [118, 119]. Interestingly, in ALS G93A muscle, the increased ROS generation is associated with reduced mitochondrial dynamics and network, indicating that this vicious feedback cycle could play a role in mitochondrial dysfunction in muscle during ALS progression. Remarkably, overexpression of ALS-causing mutation SOD1^G93A in skeletal muscle of normal mice led to similar abnormality in mitochondrial network and dynamics [11], and promoted mitochondrial ROS production [91] in the absence of motor neuron axonal withdrawal, again indicating muscle mitochondria could be directly affected by ALS mutations local to skeletal muscle.

Autophagy, an intracellular process targeting misfolded proteins and damaged organelles for lysosomal degradation, plays crucial roles in mitochondrial quality control [120]. Oxidative stress regulates autophagy via multiple singling pathways in skeletal muscle (see excellent review by Rodney et al 2016) [121]. Dysregulation of autophagy is implicated in various neurodegenerative diseases [122]. Upregulated autophagy activity has been reported in the spinal cord of ALS patients and animal models [123–126]. However, contradictory results were also reported. There was no detectable increase in the expression level of autophagosome marker LC3-II in the spinal cord of ALS G93A mice [127]. The effects of compounds that activate autophagy in ALS mouse models are also contradictory. Some studies show that autophagy activation helped limit ALS progression [128, 129]. Treatment of Rapamycin (an autophagy activator via mTOR inhibition) on ALS animal models had conflicting results [130, 131]. Rapamycin alleviated disease progression in ALS TDP-43 mice [132] but augmented motor neuron degeneration in SOD1 mutant mice [126], with an adverse effect on muscle performance [133]. Contradictory results were also reported when lithium was used to activate autophagy to treat ALS mouse models [134, 135]. The contradictory results could be due to many different reasons. One could be differential autophagy activities in non-neuron cells (such as muscle) that may affect whole body homeostasis and complicate the therapeutic outcomes. It is also worth noticing that beside its inhibitory effect on mTOR, Rapamycin also affects the gating properties of skeletal muscle ryanodine receptors by dissociating FK506-binding protein from them [136, 137]. Using live cell imaging, we quantified the autophagosome and lysosome in muscle fibers derived from G93A mice at three disease stages: 6 weeks (asymptomatic stage without axonal withdrawal reported [138]), 2–3 months (asymptomatic stage but axons have begun to pull away from muscle fibers] and 3–4 months (disease onset and ALS-like phenotype becomes well-established [71]). The formation of autophagosomes/lysosomes in G93A muscle was found to be increased at all stages when the mice were on regular diet [26]. Surprisingly, we found that autophagy flux in G93A muscle was significantly suppressed. The young G93A mice (6-week older) did not have the capacity to form more autophagosomes and lysosomes when challenged by starvation (to promote autophagy) plus intraperitoneal injection of colchicine (to block the fusion of autophagosomes to lysosomes), an established physiological autophagy induction procedure to evaluate the autophagy flux in mammalian skeletal muscle [139, 140]. The muscle derived from older G93A mice (> 2 months) even showed a significant reduction in autophagosome and lysosome formation following the autophagy induction procedure. In contrast, the muscle derived from age-matched wild type mice responded to the procedure with a robust increase in the number of autophagosome puncta as expected [139, 140]. A later biochemical study by Olivan et al (2015) also showed a significant increase in the ratio of LC3-II/LC3-I in muscle derived from G93A mice at asymptomatic and late stage [141]. Olivan et al (2015) also examined the autophagy flux in skeletal muscle derived from late stage G93A mice after the isolated muscles were chopped into small pieces and incubated with 100 μM chloroquine for 6 hours. They found a significant increase in the relative expression level of LC3-II. However, the chloroquine treatment was not performed under physiological conditions, and the chopped small pieces of muscle likely experienced unavoidable damage [141], thus the results are hard to be interpreted. In our experiments, we used an established physiological autophagy induction procedure and discovered a significant reduction in autophagy flux in G93A mice. The suppressed autophagy flux implies possible exhaustion of the reserved capacity to form autophagosome under stressed conditions in skeletal muscle of G93A mice upon disease progression [26]. In line with the dysregulation of mitochondrial dynamics and the excessive ROS generation during ALS progression, the reduced autophagy capacity, which occurs early and becomes severe in the later stage of ALS, could further handicap the clearance of damaged mitochondria and become an important player in the vicious cycle promoting mitochondrial damage.

4. Reduced mitochondrial Ca²⁺ uptake leads to cytosolic Ca²⁺ hyperactivity in G93A muscle fibers at the site of neuromuscular junction

4.1. Depolarized mitochondria at the neuromuscular junction (NMJ) of ALS muscle fibers

The first evaluation of mitochondrial lesion on live skeletal muscle fibers derived from adult ALS SOD1^G93A mouse model was conducted by our group (Zhou et al, 2010) [15]. Oxidative phosphorylation creates a proton gradient across the inner mitochondrial member that establishes the mitochondrial inner membrane potential (ΔΨ_m) and acts as a driving force for ATP synthesis [142]. Using a mitochondrial membrane potential indicator (TMRE), we discovered that a portion of SOD1^G93A muscle fibers derived from SOD1^G93A mice at the age of ALS disease onset displayed localized loss of ΔΨ_m in fiber segments near the site of NMJ [15]. The depolarization of mitochondrial inner membrane potential would suggest reduced driving force for ATP synthesis, which is consistent with reports of reduced ATP level in skeletal muscle of ALS animal models [12, 143]. A major event during ALS progression is motor neuron axonal withdrawal from the NMJ. Those depolarized mitochondria at the site of NMJ could be a sole response of muscle fibers to motor neuron withdrawal (denervation). Interestingly, these types of mitochondrial defects also occur in young ALS mice even prior to the motor neuron axonal withdrawal reported in the same ALS mouse model [144]. Thus, this observed mitochondrial lesion could also result from intrinsic defects in ALS muscle.

4.2. Reduced mitochondrial Ca²⁺ uptake leads to cytosolic hyperactive Ca²⁺ transients in ALS muscle fibers near NMJ

Ca²⁺ is fundamental to normal cellular function. Dysregulated Ca²⁺ signaling plays critical roles in muscle degeneration in aging [145, 146] and muscular dystrophy [147–151]. Remarkably, the collapse of the inner mitochondrial membrane potential ΔΨ_m at NMJ is associated with hyperactive intracellular Ca²⁺ transients [14, 15]. With the driving force generated by the negative inner membrane potential ΔΨ_m, mitochondria in skeletal muscle could take up Ca²⁺ from cytosol during excitation-contraction (EC)-coupling [45, 152]. Since both Ca²⁺ transient and mitochondrial Ca²⁺ uptake in fiber regions with normal or defective mitochondria can be compared in the same ALS muscle fiber, in which other cytosolic Ca²⁺ removal processes likely remain the same, we were able to quantify the contribution of mitochondria-mediated Ca²⁺ uptake in shaping the cytosolic Ca²⁺ transients during EC-coupling. Applying live cell imaging with a mitochondrial targeted Ca²⁺ biosensor and the voltage-clamp technique, we have found that mitochondrial Ca²⁺ uptake decreases by ~14.6% in the fiber segment with depolarized mitochondria, accompanied by a ~15.6% elevation of cytosolic Ca²⁺ in ALS G93A muscle fibers. Our calculation estimates that about 10~18% of cytosolic Ca²⁺ removal at the peak of voltage-induced Ca²⁺ release is attributable to mitochondrial Ca²⁺ uptake in mouse skeletal muscle fiber [14]. Additionally, the elevation of cytosolic Ca²⁺ transient in the fiber region with depolarized mitochondria exhibits similar kinetics to the mitochondrial Ca²⁺ uptake profile in the fiber segment with normal mitochondria. Hence, the reduced Ca²⁺ buffering capacity of these depolarized mitochondria is considered as a major cause of the increased cytosolic Ca²⁺ transients during voltage-induced Ca²⁺ release at NMJ in ALS muscle fiber, which could constitute an important pathophysiological component of neuromuscular degeneration in ALS G93A mouse model [14]. Since NMJ is the primary site of neuron-muscle interactions, this type of Ca²⁺ hyperactivity at the NMJ may also imply a defective crosstalk between muscle and motor neuron in SOD1^G93A mice.

In a separate study by Chin et al (2014), the muscle fibers derived from late stage ALS G93A mice showed a significant elevation of the resting intracellular Ca²⁺ level, and the Ca²⁺ transients had a slower decay to the resting level following 10-Hz electrical stimulation. The elevations in cytosolic Ca²⁺ could be linked to decreased protein level of ATPase SERCA1 and the Ca²⁺ buffer protein parvalbumin observed in skeletal muscle of ALS G93A mice [153]. There are also reports showing the protective effect of parvalbumin on Ca^2+- related excitotoxicity in motor neurons of ALS animal models [154–156]. In addition, there are literatures suggesting a correlation between decreased SERCA efficiency (due to ATP deficiency) and elevated cytosolic Ca²⁺ in motor neuron degeneration [157–159]. The elevated intracellular Ca²⁺ could promote the aggregation of mutant SOD1 inside mitochondria and exacerbate the loss of mitochondria inner membrane potential in cultured motor neurons [160]. The same situation may happen in skeletal muscle. Indeed, overexpression of the GFP-tagged SOD1^G93A in skeletal muscle of normal mice revealed mitochondrial accumulation of the mutant protein [15], and the GFP-SOD1^G93A mutant protein formed aggregates inside mitochondria, leading to depolarized mitochondria and reduced mitochondrial dynamics [11]. It is possible that aggregation of ALS mutant protein triggers mitochondrial membrane depolarization that reduces mitochondrial Ca²⁺ uptake, slows down mitochondrial dynamics and promotes mitochondrial ROS production. Initially, the protein aggregation in mitochondria could instigate the unfolded protein response (UPR) aiming to restore protein homeostasis in mitochondria [161]. However, if mitochondrial damage persists during ALS progression, these events mentioned above could promote each other to further damage mitochondria, which could exhaust mitochondrial clearance pathways including autophagy/mitophagy in G93A skeletal muscle.

5. Loss of physiological Ca²⁺ transients following denervation initiates mPTP-related mitochondrial ROS production.

A major pathological event during ALS disease progression is the motor neuron axonal withdrawal from skeletal muscle. Denervation of skeletal muscle induces a dramatic increase in mitochondrial ROS production [162]. It is known that mitochondrial Ca²⁺ overload is a pathological stimulus of ROS generation [83]. Studies have shown that prolonged muscle denervation leads to an increased steady state resting Ca²⁺ level in cytosol, which could overload mitochondria, stimulating ROS production [163, 164]. However, the initial cause of the mitochondrial ROS production in denervated muscle fibers was elusive [165]. Following denervation, no action potential is initiated in muscle fibers, thus, no physiological Ca²⁺ transients occur in the cytosol and mitochondria. How do mitochondria respond to the absence of physiological Ca²⁺ transients in skeletal muscle? Through examining the mitoflash activity in live muscle fibers 24 hours after denervation, we (Karam et al, 2017) had a surprising finding. Mitochondria respond to this condition with a drastic increase in CypD-dependent mitoflash activity accompanied by an elevated fluorescence intensity of MitoSOX Red, indicating enhanced mitochondrial ROS production induced by the short term denervation. Remarkably, restoring physiological Ca²⁺ transients by a brief train of electric stimulation reduced mitochondrial ROS production, diminished the mitoflash activity, and stopped the repetitive mPTP opening in the denervated muscle fibers [85]. The alleviating effect of electric stimulation on mitochondrial ROS generation and mitoflash activity were diminished by application of RU360, an inhibitor of mitochondrial Ca²⁺ uptake through blocking the mitochondrial uniporter [15, 166]. Thus, loss of physiological Ca²⁺ uptake by mitochondria is likely an initial trigger for mPTP opening and excessive mitochondrial ROS production in muscle fibers following denervation. While mitochondrial Ca²⁺ overload is a pathological stimulus leading to mPTP opening and ROS generation, physiological intracellular Ca²⁺ transients are also critical for maintaining mitochondrial functional integrity in skeletal muscle. Our data support that application of the electric stimulation on skeletal muscle could be beneficial to patients suffering from denervation, including ALS. The molecular composition of mPTP is still not fully understood. It is unclear what are the molecular mechanisms underlying the different responses of mitochondria to a steady-state elevated intracellular Ca²⁺ level and to the absence of physiological Ca²⁺ transients following denervation. We speculate that there may be a biphasic dependence of mPTP on the cytosolic Ca²⁺ level or the dynamics of SR Ca²⁺ release. Future studies are needed to investigate these molecular basis [85].

6. Proposed pathologic sequence of muscle degeneration in the ALS mouse model

Potential causes of muscle mitochondrial damage in ALS muscle include ALS-causing mutation (such as SOD1, TDP-43 mutations), denervation, or other unknown factors leading to SALS. Based on published studies from us and other research groups, we speculated dysfunctional mitochondria-related pathologic sequence of muscle degeneration in the ALS G93A mouse model during the disease progression (Fig 1). The factors initiating the vicious cycle of mitochondrial Ca²⁺ and ROS dysregulation in G93A muscle are SOD1^G93A mutation and denervation, which are highlighted in orange boxes in Fig 1. At the early stage, motor neuron axonal withdrawal from the NMJ has not occurred yet. At the muscle side of the NMJ there are abundant nAChRs, which are known to have elevated Ca²⁺ permeability in adult mammalian muscle [167]. During repetitive muscle contraction, the repetitive opening of nAChR channels could lead to an elevated Ca²⁺ influx near NMJ area (micro-domain). Mitochondria at NMJ face elevated local [Ca²⁺]_C and thereby become more susceptible to the deleterious effects of the ALS mutant aggregates as demonstrated in cultured motor neuron cells [168] and become depolarized [11]. The elevated [Ca²⁺]_C may also reduce mitochondrial motility and promote mitochondrial fragmentation [169, 170], which could dampen the recovering capacity of damaged mitochondria. Both elevated mitochondrial Ca²⁺ overload and membrane depolarization are associated with increased CypD-related mPTP transient opening (revealed as the higher frequency of mitoflash events) and promote mitochondrial ROS production. At the early stage of ALS, this increased transient opening of mPTP may be a survival mechanism to eliminate mitochondrial Ca²⁺ overload [171, 172] in skeletal muscle. Upon the progression of ALS, accumulation of these defects would form a vicious cycle, leading to further mitochondrial damage with reduced mitochondrial Ca²⁺ removal and consequently enhanced cytosolic Ca²⁺ hyperactivity. This situation is exacerbated by the decreased level of SERCA1 and parvalbumin [153], resulting in further Ca²⁺ overload within adjacent “normal” mitochondria. Meanwhile the aggregation of SOD1^G93A inside mitochondria further promotes mitochondrial damage and ROS production. These accumulated mitochondrial defects would also inhibit mitochondrial dynamics and exhaust the reserved capacity of autophagy, crippling mitochondrial quality control. When denervation becomes a major event in ALS skeletal muscle, mitochondria respond to the absence of physiological intracellular Ca²⁺ transients with excessive ROS production [85]. This excessive ROS production is accompanied by prolonged CypD-related mPTP opening (revealed as longer lasting mitoflash events) [91], leading to major consequences of abnormal energy metabolism and muscle cell death [173]. A similar Ca²⁺-mediated mechanism may also apply to the neuronal side of the NMJ. Earlier studies by Siklós et al (1996) showed that motor nerve terminals from ALS specimens contained a significantly higher level of calcium [52]. It has been demonstrated that repetitive action potentials increase cytosolic Ca²⁺ level rapidly, and heavily load local mitochondria in motor terminals of SOD1^G93A mice [174, 175]. These alterations on both sides of NMJ lead to progressive withdrawal of the motor axonal terminal, and this can be a self-reinforcing consequence, including loss of mutual trophic signals between muscle and nerve cells.

Fig 1. Proposed pathologic sequence of muscle degeneration related to mitochondrial dysfunction in the ALS mouse model.

At the early stage, the integrity of NMJ is still intact. Muscle fibers could have elevated cytosolic Ca²⁺ near NMJ area, potentially due to Ca²⁺ permeability of nAChR during repetitive muscle contraction. This elevated cytosolic Ca²⁺ level at NMJ could lead to mitochondria Ca²⁺ overload and accumulation of SOD1^G93A inside mitochondria, which causes mitochondrial membrane depolarization and promotes ROS production. The observed transient opening of mPTP at this stage could be an adaptive response attempting to relieve Ca²⁺ overload in mitochondria. Upon the progression of ALS, accumulation of these defects would form a vicious cycle, leading to further mitochondrial damage with reduced mitochondrial Ca²⁺ removal and consequently enhanced cytosolic Ca²⁺ hyperactivity, which is exacerbated by the decreased level of SERCA1 and parvalbumin. The aggregation of SOD1^G93A inside mitochondria further promotes mitochondrial damage and ROS production, inhibiting mitochondrial dynamics and exhausting autophagy capacity. The axonal withdrawal event also promotes excessive ROS production, which is accompanied by prolonged CypD-related mPTP opening. These changes lead to major consequences on abnormal energy metabolism and muscle cell death.

Skeletal muscle degeneration in ALS is a complex scenario, and various molecular signaling pathways could be involved. Our speculation here could be oversimplified. In this current review, via discussion of current understanding of the dysregulated mitochondrial Ca²⁺ and ROS signaling during ALS muscle degeneration, we intend to attract more research efforts to explore the molecular mechanisms underlying the dysregulation of mitochondrial Ca²⁺ and ROS signaling in ALS muscle, and to develop potential alternative means to alleviate ALS progression through preserving mitochondrial function in skeletal muscle. In addition, motor neuron and skeletal muscle show similar abnormalities in mitochondrial structure and biochemical properties. Although many studies on mitochondrial function have been done on cultured motor neuron cell lines, it is difficult to isolate live motor neurons from the spinal cord of adult ALS rodent models. This limits the investigation of disease-stage-dependent mitochondrial changes in adult motor neurons during ALS disease progression. Studies of mitochondrial function in muscle cells may provide clues for unraveling potential mechanisms underlying motor neuron degeneration.

Abbreviations

ΔΨ_m

mitochondrial inner membrane potential

ADP

adenosine diphosphate

ALS

myotrophic lateral sclerosis

ATP

adenosine triphosphate

COX

cytochrome c oxidase

CypD

cyclophilin D

FALS

familial amyotrophic lateral sclerosis

G93A mice

transgenic mice with overexpression ALS-associated mutation SOD1^G93A

LC3

microtubule associated protein 1A/1B-light chain 3

MLC

myosin light chain

mPTP

mitochondrial permeability transition pore

mt-cpYFP

mitochondrial targeted biosensor

mTOR

mechanistic target of rapamycin kinase

nAChRs

nicotinic acetylcholine receptor

NADH

nicotinamide adenine dinucleotide (NAD) + hydrogen (H)

NMJ

neuromuscular junctures

PGC-1α

peroxisome proliferator-activated receptor gamma, coactivator 1 alpha

ROS

reactive oxygen species

RU360

oxygen-bridged dinuclear ruthenium amine complex selectively inhibit Ca²⁺ uniporters in the mitochondria

SALS

sporadic amyotrophic lateral sclerosis

SERCA

sarco/endoplasmic reticulum Ca²⁺-ATPase

SOD1

superoxide dismutase type I