Elsevier

Free Radical Biology and Medicine

Volume 98, September 2016, Pages 113-122
Free Radical Biology and Medicine

Exercise-induced hormesis and skeletal muscle health

https://doi.org/10.1016/j.freeradbiomed.2016.02.025Get rights and content

Abstract

Hormesis refers to the phenomenon that an exposure or repeated exposures of a toxin can elicit adaptive changes within the organism to resist to higher doses of toxin with reduced harm. Skeletal muscle shows considerable plasticity and adaptions in response to a single bout of acute exercise or chronic training, especially in antioxidant defense capacity and metabolic functions mainly due to remodeling of mitochondria. It has thus been hypothesized that contraction-induced production of reactive oxygen species (ROS) may stimulate the hormesis-like adaptations. Furthermore, there has been considerable evidence that select ROS such as hydrogen peroxide and nitric oxide, or even oxidatively degraded macromolecules, may serve as signaling molecules to stimulate such hermetic adaptations due to the activation of redox-sensitive signaling pathways. Recent research has highlighted the important role of nuclear factor (NF) κB, mitogen-activated protein kinase (MAPK), and peroxisome proliferator-activated receptor γ co-activator 1α (PGC-1α), along with other newly discovered signaling pathways, in some of the most vital biological functions such as mitochondrial biogenesis, antioxidant defense, inflammation, protein turnover, apoptosis, and autophagy. The inability of the cell to maintain proper redox signaling underlies mechanisms of biological aging, during which inflammatory and catabolic pathways prevail. Research evidence and mechanisms connecting exercise-induced hormesis and redox signaling are reviewed.

Introduction

Hormesis is a biological concept which states that exposure to a low dose of a noxious or toxic agent can bring about results deemed beneficial to the long-term welfare of the organisms [1]. According to the recent literature, a biological phenomenon can be called hormesis if it fulfills the following conditions: (1) it shows a biphasic dose-relationship in which the response to low dose is opposite to the response to a high dose; (2) the concentration and effects of the low dose are measurable, i.e., are not due to placebo [2]; and (3) the factors acting on the biological system are present in natural environment [2]. Among the various hormetic agents are hypoxia, heat, starvation, pro-oxidants, and other types of stress such as pain, sleeplessness, noise, and cold [3], [4]. Although exercise itself is not a specific hormetic stimulus, numerous biochemical and physiological changes take place during exercise at the cell, organ and circulatory levels that have been shown to elicit hormetic responses. Thus, exercise has been suggested to have hormesis-like benefits [5], [6]. It is interesting to note that studies on the efficacy and mechanism of exercise-induced hormesis are increasing in recent years. Among the various best-known hormetic effects studied to date are upregulation of antioxidant network, mitochondrial adaptation, cardiac protection against ischemia-reperfusion, heat tolerance, adaptation to low energy substrates (especially blood sugar), and muscle hypertrophy in response to blood flow restriction [7].

Redox signaling induced by intrinsic generation of reactive oxygen species (ROS) and reactive nitrogen species (RNS) is closely related to exercise-induced hormesis [8], [9]. This is because mild oxidative stress, resulting from the imbalance between ROS and RNS generated during muscular contraction and the endogenous antioxidant defense system, can activate specific cellular pathways that lead to various adaptations including, but not limited to, post-translational enzyme activation/inhibition, modulation of transcription factors (TF) and cofactors (via covalent modification and association/dissociation), up-or downregulation of gene transcription, and altered potential epigenetic mechanisms [10], [11]. Hydrogen peroxide (H2O2) and nitric oxide (NO) serve as the most important signaling molecules due to their mild chemical reactivity, relative stability, and diffusiability [12], [13], [14], [15]. An important paradigm of redox signaling is based on reversible modification of cysteine residues on specific enzymes which subsequently control downstream enzymes and pathways [9]. However, concurrent definition of redox signaling is not strictly limited to sulfhydryl modification but includes modification of protein function due to the electron transfer process. Phosphorylation/dephosphorylation via kinases and phosphatases, acetylation/deacetylation, methylation and sulfoxidation, are also potential covalent modulations in the process of redox signaling.

For the past several decades, the Free Radical Theory of Aging first proposed by Harman [16] took the center stage of aging research. It was later modified to Mitochondrial Theory of Aging pointing to a critical role of ROS of mitochondrial origin in the progression of aging. Over the past two decades, however, evidence has continued to merge that increased ROS generation and accumulation of oxidatively damaged macromolecules cannot explain biological aging as previously thought [17]. Thus, there is some confusion as to how acute and long-term physical exercise, which increases ROS generation and renders macromolecule to oxidative modification, would impact on the health and longevity of the aging population, who are encouraged to be more physically active. Furthermore, since aging directly affects metabolic rate, mitochondrial structure and turnover, ROS generation, and antioxidant defense capacity, it is conceivable that cell redox signaling patterns and intensity among the older individuals would be different from those of the young ones. Finally, aged individuals may perceive “low dose” of toxins and stress differently than the young individuals, i.e., aging may alter the shape of the inverted U curve typical of a hormetic effect, as well as the “breaking point”, i.e., the point where the curve dramatically changes its direction. This in turn can impact on physiological function and performance of the elderly people. Obviously, the topic to be reviewed is quite broad and complex. In order to keep in line with the purpose of his special Olympic Issue of FRBM focusing on Human Performance and Redox Signaling in Health and Disease, the current review will focus on those exercise-induced hormetic effects for which redox signaling is experimentally proven to be the underlying mechanism. Also, primary attention will be given to the skeletal muscle, an organ vital for physical function and performance. For more complete review, the readers are referred to several recent articles on hormesis [7], [17], [18].

At the very early stage of free radical research, it was reported that wild animals and birds display higher antioxidant enzyme activities than their domestic counterparts [19]. It was speculated that frequent muscular activity associate with the life pattern of the former was responsible for the observed difference in antioxidant defense, and this hypothesis was later experimentally confirmed with animals involved in exercise training. Numerous studies in the 1980–90's showed that exercise-training promoted an increase in key antioxidant enzymes in skeletal muscle [20], [21], [22], [23], [24]. Furthermore, trained animals or humans were found to have lower level of oxidative damage such as lipid peroxidation than their sedentary counterparts when subjected to similar related workload (%VO2max) [25]. Interestingly, exercise intensity was found to determine the magnitude of adaptation of a particular antioxidant enzyme such as superoxide dismutase (SOD), such that when daily running time and treadmill speed and grade were increased, so did SOD activity in the muscle [26]. However, as exercise intensity reached maximal level, SOD activity leveled off. Besides SOD, glutathione (GSH) antioxidant system were also altered by training [27], [28]. Training adaptation of GPX was found to be muscle fiber specific, with the type 2a fibers which have moderate antioxidant defense to be more responsive than type 1 fibers, which have higher antioxidant capacity [29]. To examine the biological consequence of this training adaptation, respiratory function of isolated muscle mitochondria from endurance trained and sedentary rats were examined after being exposed to equal dose of H2O2 [30]. “Trained” mitochondria demonstrated a higher rate of state 3 respiration and respiratory control index (RCR) compared to those from sedentary rats, indicating a greater resistance to ROS insult. In addition to SOD and GPX, the rate-limiting enzyme for GSH synthesis, γ-glutamylcysteine synthetase (GCS), was reportedly induced by training in muscle and liver [31], [32]. As a result, muscle and hepatic GSH contents, as well as GSH export under cardiac ischemic insult, were increased in the trained state.

Inducible form of nitric oxide synthase (iNOS) has shown clear adaptation in response to endurance training [33], [34]. Increased NO production can have dual consequences: NO can react with superoxide radicals to form peroxynitrite thereby inflicting strong oxidative damage, whereas NO is also known to exert vasodilative effect to increase blood flow to muscle thus indirectly enhance muscle antioxidant defense during exercise [35], [36].

The definition of antioxidant enzyme may be broadened to including those which indirectly protect the organisms from oxidative damage, such as enzymes removing modified lipids (phospholipase A2), repairing oxidatively damaged DNA (8-oxoG DNA glycosylase-1, OGG), and selectively degrading damaged proteins (proteases). Exercise training has been shown to induce some of these enzymes and thus reducing the impact of exercise-induced ROS to these macromolecules. However, the discussion of these enzymes to acute and exercise is beyond the scope of the current review.

Although muscular contraction during exercise was found to increase ROS generation in as early as 1980s [37], [38], the idea that exercise-generated ROS can directly induce antioxidant enzyme adaptation remained a hypothesis [37]. An important breakthrough was reported by Hollander et al. [39] showing that exercise training could increase the protein content of mitochondrial SOD (SOD2), but not cytosolic SOD. The data indicated that training induced antioxidant enzyme activity is caused by accumulation of enzyme protein, and related to mitochondrial generation of ROS. Oh-ishi et al. [40] and Gore et al. [41] were the first to report that an acute bout of exhaustive exercise could alter mRNA level of antioxidant enzymes. Shortly after, Hollander et al. [42] demonstrated that a single bout of running significantly increased mRNA of SOD2, the first evidence that exercise can change the gene expression of an antioxidant enzyme. Furthermore, Roberts et al. (1999) showed that iNOS mRNA level could be increased after an acute bout of exercise [43]. Later, a studied by Gomez-Cabrera confirmed the above findings showing that mRNA levels of both MnSOD and iNOS were elevated after acute exercise [44]. Despite these insightful discoveries, how exercise can activate the gene expression of SOD2 and other antioxidant enzymes remains unclear in the 1990s.

In 2001, Zhou et al. [45] reported a significant finding, demonstrating in C2C12 muscle cells that nuclear factor (NF) κB plays a critical role in the response of SOD, catalase and GPX to H2O2 treatment. While H2O2 induced a dose-responsive elevation of NFκB binding, NFκB-driven luciferease activity and mRNA level of GPX, deletion and mRNA silencing of κB binding sites on the DNA promoter abolished the H2O2 effects. A few years after, Ji et al. [46] reported that an acute bout of exercise activated NFκB binding in rat skeletal muscle, with a peak activation at 2–4 h after exercise. The experiments also showed that both IκB and IκB kinase (IKK) were phosphorylated after exercise, whereas P50/P65 was transported into the nucleus. It is now clear that SOD2 promoter region contains NFκB binding sites occupancy of which is essential for the transactivation of the gene [47]. IκB phosphorylation is catalyzed by IKK whose upstream activators include protein kinase C (PKC) and NFκB-inducing kinase (NIK, a family of MAPK kinase kinase). Besides SOD2, the best-known proteins and enzymes that require consensus binding of p65 in their promoters are GCS, iNOS, TNF-α, IL-6, cyclooxygenase (COX)−2, vascular cell adhesion molecule-1 (VCAM-1), and muscle ring-finger protein-1 (MuRF1). These genes are involved in a wide variety of biological functions such as antioxidant defense, inflammation, immunity and proteolysis (Fig. 1).

In addition to NFκB, the broad biological functions regulated by MAPK family enzymes caught the attention of the exercise scientists and the first report on exercise activation of MAPK was published by Goodyear et al. in 1996 [48]. Extracellular stimulators of MAPK pathway include growth factors, inflammatory cytokines, and phorbol esters, whereas ROS are the primary intracellular activators. Gomez-Cabrera et al. [44] reported that ERK1/2 and p38 were activated after an acute bout of treadmill running in rats and the activation coincided with elevated gene expression of MnSOD and iNOS. Importantly, when allopurinol, a xanthine oxidase (XO) inhibitor was used to partially block ROS generation, exercise-induced MnSOD and iNOS mRNA expression was severely hampered, along with decreased activity of ERK and p38, suggesting that MAPK enzymes play a critical role in the signaling of antioxidant enzymes.

Lipid peroxides, a unique group of byproducts of lipid peroxidation of polyunsaturated fatty acids (PUFA) and phospholipids, have recently received attention as an important source of ROS [11], [49]. Exercise-induced lipid peroxidation, shown by elevated plasma and muscle malondialdehyde (MDA), conjugated diene, 4-hydroxynunenal (HNE), and 8-isoprostane, as well as pentane from expiratory gas, has been documented extensively during the past three decades [10]. Carbonyl stress arising from lipid peroxides has been shown to activate genes involved in phase I and II antioxidant/detoxification due to activation of NF E2-related factor-2 (Nrf2). Nrf2 is normally sequestrated with Keap1 protein in the cytosol. Once activated, Nrf2 is liberated from Keap 1, translocated into the nucleus and bind to the antioxidant response elements (AREs) in the promoter region of the target genes involved in phase II detoxification and GSH synthesis and conjugation [50]. This mechanism is widely known in adaptation of heart and smooth muscle to toxins and xenobiotics, whereas whether or not exercise produced HNE and other lipid peroxidation products play a role in antioxidant adaptation in skeletal muscle is rarely documented and requires further investigation.

Exercise-induced activation and/or gene expression have been observed in other antioxidant systems such as heat shock proteins (HSP72, HSP25) [51], [52] and mitochondrial uncoupling proteins (UCPs). Limited by the focus of the current review, only a brief review will be provided to the latter. Mitochondrial production of ROS is partly dependent on the cross-membrane proton motive force (Δψm). Thus, increasing mitochondrial membrane permeability with uncouplers (such as DNP, FCCP) to reduce Δψm is viewed as a classic way to reduce ROS production. Located in the mitochondrial inner membrane, UCPs are a heterogeneous family of proteins that play an important role in partially dissipating the proton electrochemical gradient Δψm [53]. The best characterized UCP1 is expressed exclusively in the brown adipose tissue of rodents with a key function of adaptive thermogenesis. UCP2 is expressed ubiquitously with a physiological role yet to be fully understood [54], whereas UCP3, expressed primarily in the skeletal muscle, is regarded a plausible regulator of trans-membrane proton potential and hence efficiency of oxidative phosphorylation [55].

An acute bout of exercise or contractile activity has been shown increase UCP3 express in mammalian skeletal muscle [55], [56], [57]. However, Goglia et al. [58] was the first to link exercise-induced UCP3 to antioxidant defense by postulating that, by translocating fatty acid peroxides from inner to the outer membrane leaflet, UCP may fulfill a role in antioxidant defense of the mitochondria. According to this promise, an upregulation of UCP3 as observed by the previous studies might have provided a proton shunt back to the matrix to decrease Δψ, reducing superoxide radical production from the electron transport chain. Jiang et al. [59] reported that in response to a prolonged bout of exercise, UCP3 mRNA and protein expression were remarkably elevated in rat skeletal muscle. While ROS production was increased progressively with increased exercise duration, it plunged at 150 min when UCP3 protein rose to the highest level. State 4 respiration rate increased at the peak of 120 min, but returned to resting rate at 150 min. These data demonstrate that UCP3 upregulation during prolonged exercise might reduce ROS production thus serving as a means of antioxidant defense to protect mitochondria from oxidative stress. Similar findings were observed in rat heart, wherein exercise-induced UCP2 coincided with a reduction of ROS production in a time-coordinated manner [60]. Interestingly, increased UCP expression was associated with a decrease of state 3 respiration and ATP production rate. This is not surprising, as decreased Δψm associated with proton shunt controlled by UCP is known to reduce the efficiency of oxidative phosphorylation.

The cellular mechanism by which muscle contraction increases UCP expression in the mitochondria is still unclear. Anderson et al. [61] demonstrated that UCP3 gene expression was increased by an acute bout of exercise in mouse gastrocnemius muscle and this upregulation was dependent on mitochondrial H2O2 production, indicative of redox sensitivity. In the UCP3−/ (knockout) mice, exercise failed to increase mitochondrial uncoupling respiration, whereas H2O2 production was greater compared to that in the wild type (WT) mice. St. Pierre et al. [62] showed that while H2O2 stimulated UCP3 and UCP2 expression in WT muscle cells, PGC-1α knockout (KO) abolished these effects. Thus, PGC-1α could be an important mediator in the upregulation of UCP3.

Inflammation constitutes an important mechanism in the pathogenesis of muscle disuse atrophy and oxidative damage induced by lengthening contraction, during which the production of several pro-inflammatory cytokines are elevated [63]. TNF-α, the most potent activator of NFκB pathway, is well-documented to increase its expression in association with muscle waste, cancer and aging [64], [65], [66]. Recent observations indicate that endurance exercise, by upregulating PGC-1α-controlled mitochondrial biogenesis and antioxidant enzymes, can play an important role in anti-inflammatory effects [67], [68]. PGC-1α KO mice showed higher basal mRNA expression of TNF-α and IL-6 in skeletal muscle, as well as higher serum IL-6 level than WT mice [69], whereas PGC-1α overexpressing mice had lower expression of TNF-α and IL-6 mRNA in skeletal muscle, and reduced serum TNF-α and IL-6 levels [70]. Moreover, a single exercise bout elicited a significant increase in skeletal muscle TNF-α mRNA and serum TNF-α content in PGC-1α KO mice, but not in WT mice, indicating PGC-1α blocks inflammatory gene expression [69]. Recently, we reported that ROS-induced NFκB activation and TNF-α, IL-1β and IL-6 expressions were elevated in mouse muscle shortly after a two-week immobilization, but these inflammatory markers were profoundly reduced by PGC-1α over-expression via electroporation [71].

The inhibitory effects of PGC-1α on inflammation may be attributed to several important mechanisms. First, intact PGC-1α signaling is required for antioxidant enzyme gene expression especially SOD2 and GPX in the mitochondria [62]. Second, PGC-1α is known to be a strong inhibitor of NFκB, the major pathway that promotes expression of several inflammatory cytokines, such as TNF-α, IL-1β, IL-6 and COX-2, as well as VCAM-1 [64]. PGC-1α has been shown to directly interact with P65 at the protein level, attenuating its binding to DNA [72]. Third, PGC-1α was reported to suppress nuclear retention of FoxO3 by AICAR-induced AMPK activation [73]. It is well known that FoxO is a major activator of ubiquitin-proteolysis and autophagy-lysosome pathways [74], whereas high levels of PGC-1α mitigates the FoxO-induced catabolic processes [71]. Increased proteolysis is regarded one of the main contributors to inflammatory responses in skeletal muscle due to systemic and local immune responses mediated by blood-borne polymorphoneutrophils and macrophages.

In summary, endurance exercise can increase generation of ROS from the mitochondria, whereas muscle contraction at high intensity can activate XO to generate superoxide radical [10]. In addition, NADPH oxidase, NOX-2 and COX-2 are also important ROS sources within the cell, activated by exercise-induced cellular disturbances [75]. A small surplus of ROS appears to promote cellular antioxidant defense capacity due to increased gene expression of antioxidant enzymes, upregulation of UCPs and suppression of inflammatory and proteolytic pathways. Muscles pre-exposed to contraction mediated ROS appear more resistant to higher level of oxidative challenge, thus exhibiting a hormetic effect. Fig. 1 summarizes the main signaling pathways involved in exercise-induced hormesis effects on antioxidant system.

Muscle contraction can lead to a wide range of intracellular and circulatory adaptations many of which may be considered as hormetic effects [7]. However, not all “hormetins” are related to redox signaling mechanism. The following review will focus on muscle mitochondrial adaptation which has been shown to depend on redox signaling.

When animals are stressed with low oxygen environment (hypoxia), treated with thyroid hormones, under hypothermia, or engage in long-term physical work with high oxygen consumption, there is a substantial increase in mitochondrial volume, density and oxidative enzyme activity in the skeletal muscle, and to some extent in the myocardium [76]. Increased mitochondrial population not only facilitates utilizing additional oxygen to metabolize fuels to provide ATP for muscle contraction, but also shift fuels from carbohydrate to fat as a more efficient energy source. In addition, proliferation of mitochondria helps distribute oxygen consumption among increased electron transport chains and thus reduces the production of ROS [37]. Thus, working muscle is alleviated of both metabolic and oxidative stress resulting from heavy workload. Research up to date suggest that these adaptations may be accomplished by the following mechanisms: (1) de novo synthesis of new mitochondria (biogenesis); (2) mitochondrial morphology change through fusion and fission dynamics; and (3) selective degradation of “old” and damaged mitochondria (mitophagy). Furthermore, ROS-induced redox signaling is strongly implicated in all of the above mechanisms [13], [14], [77], [78], [79].

Mitochondrial biogenesis is regulated by complex signaling pathways involving synthesis, import, and incorporation of proteins and lipids into the existing mitochondrial reticulum, as well as replication of the mitochondrial DNA (mtDNA) [80], [81], [82], [83]. Although muscle mitochondrial adaptation to aerobic exercise has been confirmed with a wide range of experimental models and in numerous animal and human studies, the underlying mechanism was not elucidated until later in the past century, when PGC-1α emerged as a master regulator of not only the biogenesis of mitochondria, but also antioxidant defense, inflammatory response, and fiber transformation in the skeletal muscle [84].

PGC-1α expression has been shown to increase in rat skeletal muscle following both an acute exercise challenge [85], [86], [87], [88] and following long-term exercise training [89], [90], [91], [92], whereas muscle inactivity results in decreased PGC-1α expression [93], [94], [95]. Along with PGC-1α upregulation are increased NRF-1 and Tfam levels, although the data are not consistent due to the different exercise protocols used with variable intensities. A cross-section study showed that endurance-trained human subjects had seven times higher PGC-1α, five times higher Tfam and more than two-fold higher NRF-1 protein contents in the vastus lateralis muscle than their sedentary counterparts [96]. There is little doubt that the training adaptations were dependent on intact PGC-1α signaling, at least in animals, because PGC-1α KO mice undergoing endurance training showed virtually no change in mitochondrial markers [97].

Cellular signals known to activate PGC-1α pathway during muscle contraction are numerous and complicated involving multiple cell compartments. PGC-1α expression is linked to Ca2+ /calmodulin-dependent protein kinase IV (CamKIV), which together with calcineurin A, are activated through calcium ion dynamics within the muscle in response to exercise [86]. The increased calcium signaling during muscle contraction activates several important transcription factors such as cAMP-response element binding protein (CREB), a target of CamKIV, and myocyte enhancer factors (MEF) 2 [98]. Another factor that regulates PGC-1α expression during exercise involves p38 MAPK, which activates MEF2 and activating transcription factor (ATF) 2 [99]. Interactions of ATF2-CREB appeared to be an early event in PGC-1α mediated signaling processes [100]. P38 MAPK also stimulates PGC-1α by its phosphorylation in response to cytokine stimulation in muscle cells [101]. As a metabolic energy deprivation sensor, AMPK is activated by heavy muscle contraction due to increased AMP/ATP ratio and Ca2+ flux, enhancing PGC-1α transcription as well as activity. It was demonstrated that activation of p38 MAPK-mediated phosphorylation of CREB and its subsequent binding to PGC-1α promoter played a key role in activating PGC-1α expression in response to increased muscle activity [89].

There is evidence that PGC-1α signaling pathway is redox-sensitive. In cultured myocytes H2O2 has been shown to promote PGC-1α expression along with all the gene products controlled by PGC-1α [62]. Rats subjected to repeated sprinting exercise were found to have six times higher PGC-1α content than rested controls, but rats injected with allopurinol and subjected to the same exercise protocol showed only half of the increment, indicating ROS produced by XO were involved in the cellular signals for the observed responses [95]. While endurance training increased PGC-1α, cytochrome c oxidase subunit IV (COXIV) and phospho-CREB protein contents in rat skeletal muscle, injection of pyrrolidinedithiocarbamate (PDTC), a NFκB inhibitor and antioxidant, daily during training attenuated the observed adaptations [92].

Recent studies suggest that PGC-1α also has a regulatory role in the expression of endogenous antioxidant proteins. PGC-1α KO fibroblasts exhibit a decrease in SOD2, catalase and GPX mRNA contents relative to WT fibroblasts, whereas PGC-1α KO mice were more vulnerable to oxidative stress [62]. Reduced mRNA levels of SOD1 (CuZnSOD), SOD2 and/or GPX1 [102], as well as SOD2 protein content [103], [104], were observed in skeletal muscle from PGC-1α KO mice compared to WT, while PGC-1α over-expressing mice showed an upregulation of SOD2 protein content in skeletal muscle [70]. In addition, PGC-1α has been shown to regulate the mRNA expression of UCP 2 and UCP3 in cell culture [105], suggesting that PGC-1α may also increase the uncoupling capacity thereby reducing mitochondrial ROS production. Furthermore, it has also been shown that PGC-1α promotes sirtuin (SIRT) 3 gene expression, which is mediated by an estrogen related receptor (ERR) α binding element mapped to the SIRT3 promoter region [106]. SIRT3 deacetylates and activates mitochondrial enzymes including SOD2 through a post-translational mechanism [107], [108]. Taken together, PGC-1α seems to reduce ROS damage by upregulating antioxidant gene expression and activity, although the implication of this role in exercise-induced hormetic effects is unclear.

There is strong evidence that alterations of mitochondrial morphological changes due to dynamics protein expressions could affect cellular energy metabolism [109]. For example, mitofusin (Mfn) 2 repression in L6E9 muscle cells can lead to decreased rates of pyruvate or glucose oxidation, reduction of mitochondrial membrane potential (Δψm) and a dramatic discontinuity of the mitochondrial network, while mitochondrial mass is unaltered [110]. Accordingly, cells with low Mfn2 activity rely mainly on the use of anaerobic glycolysis to generate energy [111]. Inversely, changes in cellular energy demand and metabolic rate are known to impact on fusion and fission protein expression through the various signal transduction pathways [82]. During heavy physical exercise muscle ATP production increases dramatically and so does ROS generation. There changes are expected to profoundly change both the morphology and gene expression of mitochondrial fusion and fission proteins, and induce mitochondrial functional changes facilitating the new physiological state.

The effect of exercise on mitochondrial fusion and fission has been studied only sparsely and the available data suggest that an acute bout of exercise and chronic training may have differential effects. Bo et al. [112] reported that during an acute bout of prolonged exercise with incremental duration, there was an increased fission 1 protein (Fis1) expression but decreased Mfn1/2 expression, and the magnitude of these alterations depended on exercise duration. Mitochondrial fusion and fission protein expression also seemed to be associated with increased ROS generation and state 4 respiration, but with decreased state 3 respiration and attenuated ATP synthase. These findings suggest that heavy exercise may suppress mitochondrial fusion while promoting fission, resulting in impairment in oxidative phosphorylation and energy production. Chronic training, on the other hand, seems to lead to an expansion of the mitochondrial reticulum network through enhanced fusion, as increased mRNA levels of Mnf2 and Drp1 were observed in the trained human along with an increase in maximal mitochondrial respiratory capacity [113]. It was reported that mRNA levels of Mfn1/2 and Fis1 were elevated significantly above the resting levels 24 h after an acute bout of exercise in rats, suggesting increased gene expression might occur in the recovery period [114]. Similar findings were also obtained in cyclists at 24 h post-exercise when Mfn1 and Mfn2 mRNA levels were increased in muscle biopsies [115]. These results suggest that establishment of higher level balance of mitochondrial fusion and fission may be an important mechanism behind the morphological and functional adaptions to endurance training. Recent literature clearly indicates that mitochondrial dynamics is closely linked to mitochondrial degradation via autophagy (mitophagy), which is regulated by ROS and redox signaling [116]. Fragmented mitochondrial due to overexpression of fission protein may stimulate the initiation of mitophagy, whereas PGC-1α induces Mfn2 that inhibits Fis1 and mitochondrial fission. The relationship between PGC-1α-controlled mitochondrial biogenesis, mitochondrial dynamics and mitophage is illustrated in Fig. 1. Future studies are required to elucidate the “Goldilocks Zone” [11], within which exercise induces sufficient ROS as signaling factors to stimulate mitochondrial biogenesis while not inflicting detrimental effects to elicit excessive mitochondrial fission and subsequent decline of mitochondrial population.

Aging is a degenerative process which affects all aspects of cell life. Because of the extensive role redox signaling plays, it is expected to undergo substantial changes in aging skeletal muscle and thus impact on a wide range of physiological and pathogenic conditions associated with aging. Two questions arise when we consider the hormetic effects of exercise among the aged population: (1) does aging attenuate the capacity of redox signaling and the associated gene expression in the skeletal muscle; and (2) how does exercise alter the patterns and intensity of redox signaling in the aging muscle.

As skeletal muscle ages, there is a progressive loss of muscle mass, contractile force and other functions termed sarcopenia. The most widespread report on senescent muscle is a decline of aerobic capacity caused by both the loss of mitochondrial population density and capacity of oxidative phosphorylation, reflected by lower respiratory rate and ATP production [117]. These observations of mitochondrial enzyme profiles are related to decreased mtDNA and mitochondrial protein expression with aging.

Recent data suggest that age-related downregulation of PGC-1α may play an important role in the decline of mitochondrial biogenesis and turnover, and contribute to the etiology of sarcopenia [118], [119]. Several recent studies indicated that both PGC-1α mRNA and protein levels were significantly decreased in senescent muscle in rats and the reduction was greater (up to >50%) in oxidative muscle fibers [70], [117]. Decreased cytochrome c, and mtDNA/nDNA ratio were also reported, but the effect of aging on Tfam, the major controller of mtDNA replication and mitochondrial proliferation, has been inconsistent. In agreement with the rodent studies, older human subjects (>65 years) showed lower PGC-1α mRNA level in their leg muscle compared to young controls, along with decreased NRF-1 and Tfam protein contents [120]. Direct evidence supporting a crucial role of PGC-1α in aging muscle was provide in a transgenic mouse study wherein muscle-specific overexpression of PGC-1α ameliorated a wide range of age-related physiological and cellular deteriorations, such as insulin resistance, body fat accumulation, neuromuscular junction degradation, and systemic inflammation at old age [70]. Lean body mass was also increased comparing PGC-1α vs. WT mice, indicating PGC-1α has a direct role in preserving muscle mass. However, some studies showed no age difference in the protein levels of PGC-1α, NRF-1 or Tfam comparing old and young human subjects thus shedding some doubt to the true function of PGC-1α signaling in age-associated muscle deterioration [96]. The exact mechanism for age-related decline of PGC-1α signaling is still unclear. However, aging is known to alter several important upstream enzymes and TFs that control PGC-1α expression and post-translational modification. For example, CREB binding to PGC-1α promoter has been shown to decline in aged rat muscle [121]. AMPK protein level and phosphorylation (activation) were shown to decrease with aging [121], [122]. Age effect on p38 has been controversial with both increase and no change being reported [123], [124].

Whether age-associated decline of PGC-1α signaling and mitochondrial biogenesis is due to a cell program that diminishes the gene expression or due to negative control inserted by other signaling pathways is currently unclear. It is well known that the concentration of H2O2, a common signaling molecule that impacts on multiple signaling pathways, is increased in aged skeletal muscle [125]. Increased ROS are known to activate NFκB and FoxO, two redox-sensitive signaling pathways that can downregulate PGC-1α [13], [74]. NFκB induces pro-inflammatory cytokines such as TNF-α, IL-1 and IL-6, whereas FoxO3 is known to be a strong activator of atrogen-1 leading to protein degradation, and also to downregulate antioxidant enzymes and increase ROS generation, subsequently activating NFκB [69], [74]. Activated NFκB and FoxO pathways along with upregulation of inflammatory cytokine expression have been shown to promote muscle protein degradation during aging, largely due to enhanced ubiquitin-proteolysis [65]. On the contrary, PGC-1α has demonstrated inhibitory effects on FoxO3, thus suppressing inflammatory cytokine expression, evidenced by the report that PGC-1α KO mice had higher TNF-α and IL-6 levels in skeletal muscle than WT mice [69], whereas PGC-1α overexpression suppressed age-associated elevation of TNF-α and IL-6 in the muscle and blood [70].

Another potentially important player in regulating mitochondrial homeostasis during aging is sirtuin (SIRT1-7). NAD+-dependent deacetylase SIRT1 can be a potential regulator of PGC-1α transcriptional activity [126]. SIRT1 has been reported to directly increase PGC-1α transcriptional activity by physical interaction with and deacetylating PGC-1α both in vitro and in vivo [127]. During aging, NADH level increases whereas NAD+ level decreases, which decreases SIRT1 activity in mouse skeletal muscle [128]. Recent research reveals that NAD+ -dependent decline of SIRT1 activity renders PGC-1α largely in the acetylated and inactivate state in aged mouse muscle, whereas increasing intracellular NAD+ concentration (by boosting NMN adenylyltransferase, NMNAT) reversed age-associated reduction of mitochondrial coded protein contents [129]. Furthermore, AMPK may serve as a metabolic switch by phosphorylating PGC-1α and allowing it to be deacetylated by SIRT1 under energy deficient state. Interestingly, SIRT1 has also been shown to inhibit NFκB activity due to the deacetylation of the p50 subunit [130]. Aging has also been shown to decrease the expression of SIRT3, a mitochondrial analog of SIRT1, in human skeletal muscle and this was interpreted as a potential mechanism for the age downregulation of mitochondrial biogenesis [96]. The interactions between PGC-1α, FoxO and SIRT1 are indeed a complex one partly displayed in Fig. 1. The SIRT-PGC-1α crosstalk seem to play a critical role in maintaining muscle aerobic capacity and mass during aging, whereas a disruption of the interplay may in part explain the mechanisms for sarcopenia [131].

In skeletal muscle antioxidant enzyme activities are increased with old age, whereas protein and mRNA levels of CuZnSOD, MnSOD and GPX were found to be either decreased or unaltered in the aged muscles [132]. Also, aged muscle exhibits lesser an extent of antioxidant enzyme adaptation to training compared to young muscle [132]. These findings have raised the possibility that redox signaling pathways that control antioxidant gene expression, such as NFκB and MAPK family enzymes may be decreased during aging. However, data in this area have been far from consistent. Several authors have reported decreased NFκB binding capacity and MAPK (ERK1/2 and p70S6K) activities in the plantaris and tibialis anterior muscles of aged rats at rest and in response to stimulated contraction (see 128 for a review). However, no difference in p38, p70S6K and JNK activities was found in the extensor digitorum longus (EDL) muscle between young and old rats [123]. Further, Williamson et al. [122] reported higher resting activities of several MAPK enzymes in the leg muscle of old men compared to young men, but the amount of protein in the MAPK pathway was unaltered with age. The above discrepancies derived from different muscle types and species are not surprising as muscle antioxidant signaling is highly fiber specific reflecting ROS generation patterns and intrinsic antioxidant capacity. Another confounding factor in studying antioxidant signaling, especially NFκB signaling, is chronic inflammation due to minor injury and/or diseases often seen in senescent animals [133]. NFκB is believed to be constitutively activated at old age, which leads to the higher basal expression of pro-inflammatory cytokines, chemokines, and ROS-generating enzymes such as iNOS and COX-2. Chronic activation of NFκB has been identified as a major reason for aged-related muscle wasting and sarcopenia [118]. It was reported that that 4-hydoxyhenxenal, a lipid peroxidation product often found in aged muscle, could activate NFκB by activating NIK/IKK signaling cascade due to ERK and p38 activation. Since NFκB activation often leads to increased pro-inflammatory cytokine expression, this vicious cycle was hypothesized as the basis for the inflammation theory of aging [133].

It is widely perceived that aging can attenuate the magnitude of training adaptation seen at a younger age, but a clear explanation of this attenuation is still lacking. For example, Derbre et al. [97] showed that muscle PGC-1α, NRF-1 and cytochrome c contents from aged rats did not respond to endurance training as the young rats did, and that the lack of training response was notably identical to PGC-1α KO mice. Furthermore, mitochondrial biogenic markers in aged muscle did not respond to cold exposure or thyroid hormone (T3), the classic PGC-1α stimulators. These findings raised the possibility that loss of sensitivity to exercise-induced redox change might underlie the mechanism for sarcopenia. However, many other studies to date do no support such a view and point out that aged muscle still maintains the ability of exercise-induced redox signaling seen in the younger muscles [120], [121]: (for a thorough review see [134]). For example, Kang et al. [121] showed that while PGC-1α mRNA and protein levels were 35% and 80% lower in the soleus muscle of 24 month vs. 3 month old rats, respectively, 12 weeks of endurance training resulted in a 2.7-fold higher PGC-1α content along with increased Tfam, cytochrome c and mtDNA contents, as well as higher CREB phosphorylation and DNA binding capacity, in old rats. These studies support a view that the critical regulators of mitochondrial biogenesis were functional in spite of old age.

Because the magnitude of training adaptation in older skeletal muscle reported in the literature displays a wide range of difference, one might speculate that aged muscle is less responsive to exercise stimulus compared to young muscle, i.e., sensitivity to training may be reduced with advanced age. However, a recent study directly challenged this perception. Iversen et al. [135] compared biopsies obtained from leg muscle of endurance trained vs. untrained elderly subjects (71 years old) in response to an acute bout of bicycle exercise at 75% of their matched VO2max. While both trained and untrained subjects increased PGC-1α mRNA expression 2 h after exercise, a surprise finding was that untrained subjects displayed twice as high PGC-1α response (i.e., 12- vs. 6-fold increase) as the trained subjects. Both groups showed remarkable increases in the phosphorylation level of AMPK and p38, the two major upstream enzymes that activate PGC-1α expression. It was also shown that while older subjects had lower basal muscle PGC-1α mRNA, NRF-1 and Tfam protein contents than young subjects, training increased PGC-1α mRNA by 2-fold and NRF-1 content by 1.5-fold [120]. These findings clearly indicate that skeletal muscle of elderly subjects maintains the ability of responding to acute exercise and that aging does not abolish muscle plasticity at least as far as mitochondrial adaptation is concerned. However, despite decades of investigation, controversy still exists in the literature as to whether or not training adaptation is attenuated with age and what role redox signaling plays in contributing to this controversy. A clear consensus on these issues is yet to be obtained with more research.

If exercise truly has a hormetic effect, then the “inverted U-shape” concept of hormesis would predict that a high dose of ROS should inhibit the hormetic effects and even cause damage to the biological systems of interest [2]. However, there is no substantial evidence in the literature to show that high intensity of exercise decreases antioxidant function among either animal or human studies. To explain the seemingly perplexing controversy, one should consider the following biological factors. (1) The exercise intensity at which a body can sustain is mainly determined by the cardiovascular system, which determines the amount of blood that can be transported to the exercising muscle. Maximal oxygen flux to the mitochondria is rarely reached before cardiac output is peaked and exercise ceases. (2) Mitochondrial coupling of oxidative phosphorylation is increased during the state 3 respiration, reducing electron “spill”, the major source of ROS during aerobic exercise [10]. (3) Increased UCP expression can result in proton shunt back to the matrix, diminishing cross-membrane proton gradient and tendency of superoxide anion formation [133]. The above observations predict that exercising muscles rarely experience high levels of oxidative stress even during maximal exercise intensity. Finally, the kinetic property of antioxidant enzymes plays a major role in protecting the cell from accumulating ROS at a high steady-state concentration. SOD and catalase do not display the Michaelis–Menten saturation kinetics (Vmax), instead, they increase catalytic activities in response to increasing substrate concentration, i.e., O2●− and H2O2, respectively [136]. Mammalian GPX also has a high Km for H2O2 (1 μM) and can reduce a wide range of physiologically generated peroxides. Thus, although antioxidant enzymes display dose-dependent adaptation to exercise intensity [137], downregulation of antioxidant system due to “over exercise” is rare. The above intrinsic mechanisms can partially explain why supplementation of antioxidants during exercise among healthy humans with proper diet is neither necessary nor healthy. A highly reduced intracellular environment does not provide extra protection to homeostasis, but only disturbs redox signaling that requires a slightly oxidative cellular environment. Several animal and human studies have clearly demonstrated that supplementation of high dose of antioxidants can attenuated or abolish exercise-induced antioxidant and metabolic adaptations, or even cause adverse effects [44], [138], [139], [140], [141], [142]. This caution is especially relevant to the elderly population 60% of which routinely takes dietary supplementation of antioxidants, vitamins and phytochemicals. The subject of antioxidant supplementation has been widely reviewed in the literature and in this special issue.

It is noteworthy that a high rate of ROS production during rigorous muscular work does interfere with redox signaling that mediates increased force production [36], heat shock response, muscle hypertrophy and metabolic adaptations [7], [11]. Highly oxidative environment is inductive to catabolic pathways such as inflammation, proteolysis, autophagy/mitophagy and apoptosis, which also use redox signaling as a mechanism for enzyme modulation and gene expression [17]. Aging can markedly change the vulnerability and dose-responsiveness of catabolic pathways to ROS, making the exercise-induced hormetic benefits less predictable [18].

Finally, a remark goes to the word “exercise” or “physical exercise”, which has been used non-discriminatively in the literature as if it represents a specific metabolic stimulus, or even a form of oxidative stress, to the body. While muscular contraction is required during exercise, the resulting effect of a particular form of muscle work can be quite different. During long-term (one to several hours) exercise at moderate intensity (50–75%VO2max), the primary source of ROS is from the mitochondria, whereas the amount of ROS produced is small (less than 1% of oxygen consumed) and mostly removed by antioxidant system [10]. During high-intensity (80–100%VO2 max), short-term (several minutes to an hour) exercise, XO activation is a main source of ROS generation. Thus, it is not surprising that during an acute bout of exercise administration of allopurinol was able to inhibit muscle adaptations resulting from redox signaling [44], [88], whereas allopurinol did not prevent mitochondrial adaptations in response to endurance training [143]. Recently, it was demonstrated in single muscle fibers that cytosolic ROS preceded mitochondrial generation of O2●−, suggesting that XO and NADPH oxidase (NOX)-2 could be important sources of contraction-mediated oxidants [75]. Muscle lengthening contraction (LC, also known as eccentric exercise) is a powerful stimulus to NOX-2 activation to produce O2●−, mainly due to the infiltration and respiratory burst of phagocytes (mainly neutrophils and macrophages) and sub-sarcolemma fraction of NOX-2 [144]. During a rigorous exercise bout in which LC is the predominate form of muscle contraction, the ROS produced (besides O2●−, also H2O2 and hypochlorous acid) are mainly from NOX-2 that stimulate inflammatory pathways controlled by NFκB and FoxO [74]. Thus, ROS produced during LC promote catabolic processes such as proteolysis and inflammation instead of mitochondrial biogenesis. Furthermore, NFκB and FoxO are known to inhibit PGC-1α DNA binding and training-induced upregulation of PGC-1α [72], [92]. Therefore, in designing exercise protocols for the purpose of promoting metabolic adaptation, LC component should be minimized. This proposition would be particularly relevant to aged people whose muscles are more prone to injury due to stretch, tear or simply overload, which inflicts inflammatory response that result in not only oxidative stress but also immobilization that escalates catabolic processes [65].

Section snippets

Conclusion

Despite a highly efficient antioxidant defense system in the cell, a small surplus of relatively stable ROS (H2O2 and NO) produced during muscle contraction may serve as signaling molecules to stimulate cellular adaptations to reach new homeostasis due to the activation of redox-sensitive signaling pathways. NFκB, MAPK, and PGC-1α have been identified as some of the most important signaling pathways and, through their crosstalk, participate in the control of several critical cellular events,

References (144)

  • S. Ghosh et al.

    Missing pieces in the NF-κB puzzle

    Cell

    (2002)
  • H. Motohashi et al.

    Nrf2–Keap1 defines a physiologically important stress response mechanism

    Trends Mol. Med.

    (2004)
  • C.D. Touchberry et al.

    Acute heat stress prior to downhill running may enhance skeletal muscle remodeling

    Cell Stress Chaperones

    (2012)
  • M. Klingenberg et al.

    Structure and function of the uncoupling protein from brown adipose tissue

    Biochim. Biophys. Acta (BBA)-Biomembranes

    (1999)
  • N. Jiang et al.

    Upregulation of uncoupling protein-3 in skeletal muscle during exercise: a potential antioxidant function

    Free Rad. Biol. Med.

    (2009)
  • H. Bo et al.

    Regulation of mitochondrial uncoupling respiration during exercise in rat heart: role of reactive oxygen species (ROS) and uncoupling protein 2

    Free Rad. Biol. Med.

    (2008)
  • E.J. Anderson et al.

    Induction of endogenous uncoupling protein 3 suppresses mitochondrial oxidant emission during fatty acid-supported respiration

    J. Biol. Chem.

    (2007)
  • J. St-Pierre et al.

    Suppression of reactive oxygen species and neurodegeneration by the PGC-1 transcriptional coactivators

    Cell

    (2006)
  • H. Chen et al.

    Mitochondrial fusion is required for mtDNA stability in skeletal muscle and tolerance of mtDNA mutations

    Cell

    (2010)
  • R.B. Hamanaka et al.

    Mitochondrial reactive oxygen species regulate cellular signaling and dictate biological outcomes

    Trends Biochem. Sci.

    (2010)
  • I. Manoli et al.

    Mitochondria as key components of the stress response

    Trends Endocrinol. Metab.

    (2007)
  • Z. Wu et al.

    Mechanisms controlling mitochondrial biogenesis and respiration through the thermogenic coactivator PGC-1

    Cell

    (1999)
  • S. Terada et al.

    Effects of low-intensity prolonged exercise on PGC-1 mRNA expression in rat epitrochlearis muscle

    Biochem. Biophys. Res. Commun.

    (2002)
  • C. Kang et al.

    Exercise activation of muscle peroxisome proliferator-activated receptor-γ coactivator-1α signaling is redox sensitive

    Free Radic. Biol. Med.

    (2009)
  • T. Akimoto et al.

    Exercise stimulates pgc-1α transcription in skeletal muscle through activation of the p38 MAPK pathway

    J. Biol. Chem.

    (2005)
  • M. Suwa et al.

    Endurance exercise increases the SIRT1 and peroxisome proliferator-activated receptor γ coactivator-1 α protein expressions in rat skeletal muscle

    Metab. Clin. Exp.

    (2008)
  • J.A. Timmons et al.

    Expression profiling following local muscle inactivity in humans provides new perspective on diabetes-related genes

    Genomics

    (2006)
  • E.J. Calabrese et al.

    Hormesis: its impact on medicine and health

    Hum. Exp. Toxicol.

    (2013)
  • S.V. Jargin

    Hormesis and homeopathy: the artificial twins

    J. Intercult. Ethnopharmacol.

    (2015)
  • F.Z. Marques et al.

    Hormesis as a pro-healthy aging intervention in human beings?

    Dose-Response

    (2009)
  • Z. Radak et al.

    Exercise and hormesis: oxidative stress-related adaptation for successful aging

    Biogeronto

    (2005)
  • L.L. Ji et al.

    Exercise and hormesis: activation of cellular antioxidant signaling pathways

    Ann. N. Y. Acad. Sci.

    (2006)
  • J.M. Peake et al.

    Modulating exercise-induced hormesis: does less equal more?

    J. Appl. Physiol.

    (2015)
  • S.K. Powers et al.

    Exercise-induced oxidative stress: cellular mechanisms and impact on muscle force production

    Physiol. Rev.

    (2008)
  • R.J. Alleman et al.

    The “Goldilocks zone” from a redox perspective—adaptive vs. deleterious responses to oxidative stress in striated muscle

    Front. Physiol.

    (2014)
  • B. D’Autréaux et al.

    ROS as signalling molecules: mechanisms that generate specificity in ROS homeostasis

    Nat. Rev. Mol. Cell Biol.

    (2007)
  • J. Pourova et al.

    Reactive oxygen and nitrogen species in normal physiological processes

    Acta Physiol.

    (2010)
  • Y. Collins et al.

    Mitochondrial redox signalling at a glance

    J. Cell. Sci.

    (2012)
  • D. Harman

    Aging: a theory based on free radical and radiation chemistry

    J. Gerontol.

    (1955)
  • M.V. Skulachev et al.

    Aging as an evolvability-increasing program which can be switched off by organism to mobilize additional resources for survival

    Curr. Aging Sci.

    (2015)
  • R.R. Jenkins

    Free radical chemistry

    Sports Med.

    (1988)
  • A. Quintanilha et al.

    Vitamin E, physical exercise and tissue oxidative damage

    Ciba Found Symp.

    (1983)
  • M. Higuchi et al.

    Superoxide dismutase and catalase in skeletal muscle: adaptive response to exercise

    J. Gerontol.

    (1985)
  • M.H. Laughlin et al.

    Skeletal muscle oxidative capacity, antioxidant enzymes, and exercise training

    J. Appl. Physiol.

    (1990)
  • M.M. Kanter et al.

    Effect of exercise training on antioxidant enzymes and cardiotoxicity of doxorubicin

    J. Appl. Physiol.

    (1985)
  • S.K. Powers et al.

    Rigorous exercise training increases superoxide dismutase activity in ventricular myocardium

    Am. J. Physiol.

    (1993)
  • J.M. Lawler et al.

    Acute exercise and skeletal muscle antioxidant and metabolic enzymes: effects of fiber type and age

    Am. J. Physiol.

    (1993)
  • C. Leeuwenburgh et al.

    Aging and exercise training in skeletal muscle: responses of glutathione and antioxidant enzyme systems

    Am. J. Physiol.

    (1994)
  • C. Leeuwenburgh et al.

    Adaptations of glutathione antioxidant system to endurance training are tissue and muscle fiber specific

    Am. J. Physiol.

    (1997)
  • R. Chandwaney et al.

    Oxidative stress and mitochondrial function in skeletal muscle: effects of aging and exercise training

    Age

    (1998)
  • Cited by (95)

    • Effect of change of interleukin-6 over time on gait speed response: Results from the lifestyle interventions and independence for elders study

      2023, Mechanisms of Ageing and Development
      Citation Excerpt :

      Therefore, inflammation which is physiologically a protective response of human body to cope with endogenous and environmental stressors including exercise (Franceschi and Campisi, 2014), may reflect a detrimental process when becomes dysregulated in amplitude and duration, as during aging (Franceschi and Campisi, 2014). Inflammation might represent a type of hormetic response also to exercise, in which repeated, transient and mild-intensity stressors may generate beneficial effects (Ji et al., 2016; Santoro et al., 2020). IL-6 is also one of the first identified myokines, and emerged as one of the main signaling molecule released during aerobic exercise, such that it has been recognized as a muscle-derived exerkine (Piccirillo, 2019).

    View all citing articles on Scopus

    This work is supported in part by the National Science Foundation of China major International Collaboration Project No. 31110103919.

    View full text