Reactive oxygen/nitrogen species and contractile function in skeletal muscle during fatigue and recovery

Abstract

The production of reactive oxygen/nitrogen species (ROS/RNS) is generally considered to increase during physical exercise. Nevertheless, direct measurements of ROS/RNS often show modest increases in ROS/RNS in muscle fibres even during intensive fatiguing stimulation, and the major source(s) of ROS/RNS during exercise is still being debated. In rested muscle fibres, mild and acute exposure to exogenous ROS/RNS generally increases myofibrillar submaximal force, whereas stronger or prolonged exposure has the opposite effect. Endogenous production of ROS/RNS seems to preferentially decrease submaximal force and positive effects of antioxidants are mainly observed during fatigue induced by submaximal contractions. Fatigued muscle fibres frequently enter a prolonged state of reduced submaximal force, which is caused by a ROS/RNS‐dependent decrease in sarcoplasmic reticulum Ca²⁺ release and/or myofibrillar Ca²⁺ sensitivity. Increased ROS/RNS production during exercise can also be beneficial and recent human and animal studies show that antioxidant supplementation can hamper the beneficial effects of endurance training. In conclusion, increased ROS/RNS production have both beneficial and detrimental effects on skeletal muscle function and the outcome depends on a combination of factors: the type of ROS/RNS; the magnitude, duration and location of ROS/RNS production; and the defence systems, including both endogenous and exogenous antioxidants.

Abbreviations

[Ca²⁺]_i

free cytosolic Ca²⁺ concentration

DTT

dithiothreitol

eNOS

endothelial NO^• synthase

FDB

flexor digitorum brevis

GSH

reduced glutathione

iNOS

inducible NO^• synthase

NAC

N‐acetylcysteine

NOS

NO^• synthase

nNOS

neuronal NO^• synthase

NOX

NADPH oxidase

PLFFD

prolonged low‐frequency force depression

RNS

reactive nitrogen species

roGFP

redox‐sensitive green fluorescent protein

ROS

reactive oxygen species

RyR

ryanodine receptor

SOD

superoxide dismutase

SR

sarcoplasmic reticulum

t‐BOOH

tert‐butyl hydroperoxide

Tn

troponin

Introduction

In a classical study, Reid et al. showed that the general antioxidant N‐acetylcysteine (NAC) mitigated the force decline when human subjects performed repeated submaximal contractions (Reid et al. 1994). These results imply that the production of reactive oxygen/nitrogen species (ROS/RNS) increases in skeletal muscle during physical exercise and ameliorating the resulting ‘oxidative stress’ with antioxidants lessens the force decrease. Moreover, several studies show increases in ROS/RNS in conditions with skeletal muscle dysfunction and muscle weakness, such as rheumatoid arthritis (Yamada et al. 2015 a,b), Duchenne muscle dystrophy (Khairallah et al. 2012), malignant hyperthermia (Lanner et al. 2012), and in normal ageing (Andersson et al. 2011). On the other hand, submaximal force in intact and skinned fast‐twitch skeletal muscle fibres has been shown to increase with acute exposure to the ROS hydrogen peroxide (H₂O₂) (Andrade et al. 1998 a, 2001; Murphy et al. 2008; Mollica et al. 2012). Furthermore, recent human and animal studies show that treatment with antioxidants hampers the beneficial effects of endurance training (Gomez‐Cabrera et al. 2008; Ristow et al. 2009; Paulsen et al. 2014). Thus, increases in ROS/RNS can have both beneficial and detrimental effects on skeletal muscle contractile function and fitness, and the outcome probably depends on a combination of factors: the type of ROS/RNS, the size of ROS/RNS increase, the duration of ROS/RNS elevation (e.g. milliseconds vs. hours), and the localization of ROS/RNS production and accumulation (Droge, 2002; Westerblad & Allen, 2011; Ristow, 2014). In the first part of this review, we will discuss the metabolism and sources of ROS/RNS that are likely to increase during exercise and ways to measure them. In the latter part, we will discuss ROS/RNS effects on muscle fibre contractility during exercise (i.e. during induction of fatigue) and in the subsequent recovery phase.

ROS/RNS in contracting muscle: metabolism, sources and methods to measure

A free radical is an atom, molecule or ion with unpaired valency electron(s), generally making them highly unstable and reactive (Halliwell & Gutteridge, 1984). The dominant ROS in cells are superoxide (O₂ ^•−) and its downstream derivatives, such as H₂O₂. Similarly, the central RNS in cells are nitric oxide (NO^•) along with its downstream derivatives, such as peroxynitrite (ONOO^•−). ROS/RNS production in skeletal muscle is generally considered to increase during physical exercise (Powers & Jackson, 2008). However, the direct mechanisms and sources of the increased ROS/RNS production during exercise remain uncertain and they are likely to differ depending on the type of activity, e.g. endurance vs. resistance training or short‐term high‐intensity vs. prolonged low‐intensity exercise. Moreover, direct measurements of ROS/RNS during exercise are relatively rare and the increases detected with conventional fluorescent indicators are often rather modest (e.g. Reid et al. 1992; Pye et al. 2007; Sakellariou et al. 2013; Cheng et al. 2015), which implies that such measurements are methodologically problematic.

Superoxide

O₂ ^•− is generated through either incomplete reduction of oxygen in the mitochondrial electron transport chain or as a specific product of enzymatic reactions. O₂ ^•− is negatively charged and hence relatively impermeable to cell membranes. It has a relatively long half‐life (∼1 μs), which permits diffusion within the muscle cell and allows interaction with several cellular targets (Winkler et al. 1999). Calculations in endothelial cells indicate that the steady‐state cellular concentration of O₂ ^•− is in the pico‐ to nanomolar range (Carballal et al. 2014).

Complexes I and III are the two major sites of O₂ ^•− production in the mitochondria (Cadenas et al. 1977; Turrens & Boveris, 1980; Murphy, 2009; Tahara et al. 2009; Quinlan et al. 2012). Early reports suggested that 2–5% of the total oxygen consumed by mitochondria was reduced to O₂ ^•− (Boveris & Chance, 1973; Loschen et al. 1974). This measure of O₂ ^•− production seems far too high in contracting skeletal muscle fibres, because oxygen consumption increases massively during intense exercise and prolonged exercise would then result in dangerously high ROS levels, severe oxidative stress, and muscle damage (Westerblad & Allen, 2011). Accordingly, more recent data indicate that only ∼0.1–0.2% of the oxygen consumed by the mitochondria forms O₂ ^•− (St‐Pierre et al. 2002; Murphy, 2009; Tahara et al. 2009; Brand, 2010). The extent of mitochondrial O₂ ^•− production in muscle also critically depends on other factors, such as sub‐ vs. supramaximal exercise intensity or sufficient vs. limited O₂ delivery. For instance, mitochondrial ROS production appears to be higher during state 4 (basal) than during state 3 (maximal ADP‐stimulated) respiration, and the latter dominates during aerobic exercise (Powers & Jackson, 2008). Thus, only a very small fraction of the oxygen consumed by the mitochondria during exercise is reduced to O₂ ^•− and there is no fixed relation between the rates of mitochondrial oxygen consumption and O₂ ^•− production.

Enzymes that produce O₂ ^•− in skeletal muscle include NADPH oxidase (NOX) (Pal et al. 2013), phospholipase A₂ (Nethery et al. 1999), xanthine oxidase (Gomez‐Cabrera et al. 2010) and uncoupled NO^• synthase (NOS) (Stuehr et al. 2001). Of these enzymes, NOX has received most recent attention and is proposed to be a major contributor to O₂ ^•− production in skeletal muscle both at rest and during contractile activities (Michaelson et al. 2010; Pal et al. 2013; Sakellariou et al. 2013). A skeletal muscle fibre expresses two NOX isoforms, NOX2 and NOX4, and of these NOX2 has received most attention (Sakellariou et al. 2014). NOX2 has several subunits and is localized in the sarcolemma, either at the surface or in the t‐tubular system (Javeshghani et al. 2002; Hidalgo et al. 2006; Sakellariou et al. 2014). NOX4 is less studied in skeletal muscle. It has been proposed to be localized in the sarcoplasmic reticulum (SR) where it may affect the Ca²⁺ release channel (ryanodine receptor 1; RyR1) (Sun et al. 2011; Sakellariou et al. 2014).

For decades, mitochondria have been considered as the major site for O₂ ^•− production in contracting skeletal muscle. Accordingly, a recent study using mouse flexor digitorum brevis (FDB) fibres transfected with a novel mitochondrial‐targeted superoxide biosensor (mt‐cpYFP) shows strictly localized mitochondrial O₂ ^•− production during repetitive contractions (Wei et al. 2011). Moreover, pretreatment with a mitochondrial‐targeted antioxidant (SS‐31) decreased ROS production, as measured with the fluorescent ROS indicator MitoSOX Red, in isolated mouse muscle fibres during repeated tetanic contractions (Cheng et al. 2015). Conversely, recent studies propose NOX2 as the main producer of ROS in contracting skeletal muscles, because the contraction‐mediated increase in cytosolic ROS was prevented by pharmacological inhibition or genetic knockdown of NOX2 (Michaelson et al. 2010; Pal et al. 2013; Sakellariou et al. 2013). Thus, there are conflicting results regarding the importance of different sources of O₂ ^•− in contracting muscle and further studies are required to resolve this issue.

Hydrogen peroxide

Dismutation of O₂ ^•−, both spontaneous and catalysed by superoxide dismutase (SOD), constitutes the major source of H₂O₂ in muscle cells (2O₂ ^•− + 2H⁺ → H₂O₂ + O₂). Two out of three SOD isoforms are highly abundant (∼10–20 μm) within the skeletal muscle fibres: SOD1 requires copper–zinc as a cofactor and is located in the cytosol and in the mitochondrial intermembrane space; SOD2 uses manganese as a cofactor and is located in the mitochondrial matrix (Powers & Jackson, 2008). Of the total SOD activity in skeletal muscle fibres, ∼15–35% is in the mitochondria and the remaining 65–85% in the cytosol. The highest SOD activities are present in oxidative slow‐twitch muscle fibres and endurance exercise increases SOD activity in both slow‐ and fast‐twitch fibres (Higuchi et al. 1985; Powers et al. 1994; Oh‐ishi et al. 1997; Bruton et al. 2008). H₂O₂ is cell permeable and relatively stable with a half‐life of seconds to minutes. The concentration of H₂O₂ in skeletal muscle has been calculated to be ∼10–100 nm at rest and to increase to ∼100–200 nm during contractions (Vasilaki et al. 2006). In comparison to other ROS, H₂O₂ is a relatively weak oxidizing agent, but in the presence of Fe²⁺ it can be converted into the highly reactive and cytotoxic hydroxyl radical (OH^•) (Halliwell & Gutteridge, 1984; Powers & Jackson, 2008). In skeletal muscle, H₂O₂ is metabolized to H₂O via three major antioxidant enzymatic systems: glutathione peroxidases (H₂O₂ + 2GSH → 2H₂O + GSSG), where GSH is reduced glutathione and GSSG is oxidized glutathione; catalase (2H₂O₂ → 2H₂O + O₂); and peroxiredoxins (PRX(reduced) + H₂O₂ → PRX(oxidized) + 2H₂O) (Sakellariou et al. 2014).

Methods to measure ROS

Quantitative measurements of ROS in skeletal muscle are rare and this is to a large extent due to the fact that methods available to measure O₂ ^•− and H₂O₂ production have severe shortcomings. Many studies of mitochondrial ROS production are performed with in vitro measurements on isolated mitochondria. This approach is problematic because the techniques used to isolate mitochondria from muscle alter mitochondrial structure and function by, for instance, causing fragmentation, loss of soluble matrix enzymes and altered respiratory rates and O₂ ^•−/H₂O₂ production (Schwerzmann et al. 1989; Picard et al. 2011 a,b). Fluorescent ROS indicators (e.g. dichlorofluorescein, dihydroethidium, MitoSOX Red) commonly used to assess changes in ROS in intact muscle fibres have limitations in that they, for instance, are not specific to a certain ROS (Kalyanaraman et al. 2012). Moreover, increases in the overall fluorescence of these indicators in response to repeated contractions are relatively small (e.g. Reid et al. 1992; Pye et al. 2007; Sakellariou et al. 2013; Cheng et al. 2015) and therefore it is difficult to assess contraction‐induced ROS changes in a quantitative and spatially/temporally confined manner. To deal with the limited specificity of these fluorescent ROS indicators, it has been suggested that they should be combined with HPLC‐based methods (Kalyanaraman et al. 2012), but such methods are cumbersome when following real‐time changes in ROS, e.g. during repeated contractions and in the following recovery period.

Promising new tools for dynamic and site‐specific assessment of ROS have recently been developed and these will hopefully lead to a better understanding of the role(s) of ROS in muscle biology. For instance, genetically engineered fusion of redox‐sensitive green fluorescent protein (roGFP) and the peroxide Orp1 has been used to measure the concentration of H₂O₂ (Gutscher et al. 2009). Furthermore, Rodney and co‐workers recently used roGFP fused to the regulatory subunit p47^phox of NOX2 for real‐time ROS measurements in contracting mouse skeletal muscle fibres (Pal et al. 2013). Furthermore, roGFP‐based ROS indicators are relatively sensitive, reversible, ratiometric (i.e. independent of changes in indicator concentration) and pH insensitive (Hanson et al. 2004; Gutscher et al. 2009; Meyer & Dick, 2010; Pal et al. 2013). Thus, genetically engineered fusion proteins constitute a novel group of ROS indicators that allow more specific detection of one ROS and can be targeted to organelles or proteins (e.g. mitochondria or NOX2).

Nitric oxide

NO^• is a versatile biological signalling molecule that is generated via enzymatic reactions of nitric oxide synthase (NOS) and the production increases in muscle fibres during repeated contractions (Pye et al. 2007; Cheng et al. 2015). NO^• can also be formed from the inorganic anions nitrate (NO₃ ⁻) and nitrite (NO₂ ⁻) (Weitzberg et al. 2010). Skeletal muscle constitutively expresses neuronal and endothelial NOS (nNOS and eNOS, respectively), whereas inducible NOS (iNOS) is upregulated in response to acute inflammatory insults. Despite its name, nNOS is expressed at higher levels in human skeletal muscle than in human brain (Nakane et al. 1993). NO^• is synthesized by NOS from l‐arginine, NADPH and O₂. nNOS and eNOS are activated by increases in the free cytosolic Ca²⁺ concentration ([Ca²⁺]_i) (Forstermann et al. 1994). The K _m of [Ca²⁺]_i for the activation of nNOS is ∼200 nm and [Ca²⁺]_i reaches higher levels during contractions (>1 μm), hence nNOS will become highly active during contractile activities (Bredt & Snyder, 1990). Under normal conditions, nNOS is mostly compartmentalized to submembrane scaffolds that are part of the dystrophin glycoprotein complex (Brenman et al. 1995). In addition, minor fractions of nNOS are detectable in association with the SR and with mitochondria (Buchwalow et al. 2005). In line with this, data from our group and others show that some nNOS co‐localize with the RyR1 in skeletal muscle from mouse and human subjects (Salanova et al. 2008; Yamada et al. 2015 b). Interestingly, the amount of nNOS co‐localized with RyR1 was markedly increased in muscles from mice with collagen‐induced arthritis (Yamada et al. 2015 b).

NO^• on its own is relatively stable but it reacts rapidly with numerous cellular targets, which results in a biological half‐life time of ∼0.3 s in skeletal muscle under physiological conditions (Thomas et al. 2001). Nevertheless, NO^• is small, uncharged and freely diffusible through membranes and is therefore considered to exert effects over distances even exceeding 100 μm (Thomas et al. 2001). The concentration of NO^• has been calculated to be ∼20 nm in resting rat diaphragm muscle fibres (Boczkowski et al. 1999).

Peroxynitrite

ONOO^•−, formed when NO^• reacts with O₂ ^•−, is a potent oxidizing and nitrating agent able to react with a wide range of biomolecules. The biological half‐life under physiological conditions is ∼10 ms and it is estimated to influence cellular targets within ∼5–20 μm (Radi, 1998; Romero et al. 1999; Szabó et al. 2007; Carballal et al. 2014). Calculations in endothelial cells indicate that the steady‐state level of ONOO^•− is ∼1 nm (Carballal et al. 2014).

The rate constant of ONOO^•− formation has been estimated to be within the range of (4–16) × 10⁹ m ⁻¹ s⁻¹ (Goldstein & Czapski, 1995; Botti et al. 2010), which is higher than the rate of SOD conversion of O₂ ^•− to H₂O₂ ((1–2) × 10⁹ m ⁻¹ s⁻¹) (Klug‐Roth et al. 1973; Hsu et al. 1996). Thus, when NO^• is produced at a high rate, it will rapidly react with O₂ ^•− to produce significant amounts of ONOO^•− even in the presence of the high physiological concentrations of SOD (∼10–20 μm). In fact, the formation of ONOO^•− from NO^• and O₂ ^•− can occur about six times faster than the rate at which SOD can convert O₂ ^•− to H₂O₂ (Beckman & Koppenol, 1996). Peroxiredoxins represent an efficient detoxification system for ONOO^•− and CO₂ accounts for another fraction of the ONOO^•− consumption by forming carbonate radicals and nitrogen dioxide (Carballal et al. 2014). On the other hand, GSH (present at mm concentration in muscles) does not react sufficiently fast in vivo to directly scavenge ONOO^•−; instead, glutathione may counteract ONOO^•−‐dependent processes via reactions with secondary radicals (Ferrer‐Sueta & Radi, 2009).

Acute effects of ROS/RNS on myofibrillar function in skeletal muscle

The intracellular events leading to contraction of skeletal muscle fibres start with RyR1‐mediated Ca²⁺ release from SR. Ca²⁺ then binds to the troponin (Tn) complex consisting of TnC, TnI and TnT. Ca²⁺ binding to the Tn–Ca²⁺ complex moves the position of the tropomyosin filaments, hence turning on cross‐bridge cycling and contraction by uncovering active sites of actin for myosin binding (Gordon et al. 2000). In addition, the degree of myofibrillar activation depends on the kinetics of cross‐bridge attachment and detachment to actin, because an actin‐bound cross‐bridge exerts mechanical impact on tropomyosin that facilitates the binding of neighbouring cross‐bridges (Brenner, 1988; Gordon et al. 2000). The result of these interacting processes is the steep, sigmoidal force–[Ca²⁺]_i relationship. Changes in the Tn–Ca²⁺ interaction and/or cross‐bridge kinetics can affect the steepness of the force–[Ca²⁺]_i relationship as well as the position on the [Ca²⁺]_i axis; for simplicity, we will refer to alterations in the latter parameter as changes in myofibrillar Ca²⁺ sensitivity. Generally, acute ROS/RNS‐mediated changes in force generation are more marked with activation at submaximal than at maximal frequencies (Lamb & Westerblad, 2011). Submaximal contractions occur on the steep part of the force–[Ca²⁺]_i relationship, which means that even small changes in myofibrillar Ca²⁺ sensitivity, or SR Ca²⁺ release, have large effects on the generated force (Fig. 1). In this context it is worth noting that everyday activities generally require low to moderate forces and the firing frequencies of motor units are therefore set to produce submaximal contractions (Marsden et al. 1971; Grimby & Hannerz, 1977)

Figure 1. Changes in myofibrillar Ca²⁺ sensitivity have a much larger effect on submaximal than on maximal force .

Schematic representation of the effect of increased (blue lines) and decreased (red lines) myofibrillar Ca²⁺ sensitivity. (1) Stimulation at frequencies giving unfused tetani results in forces on the steep part of the force–[Ca²⁺]_i relationship and changes in sensitivity have a large effect (∼40% in the example) on force output. (2) Conversely, the same changes in sensitivity have little effect (here ∼10%) at higher stimulation frequencies and fused tetani. Similar changes occur with changes in tetanic [Ca²⁺]_i, i.e. larger effects in unfused contractions.

Intact fast‐twitch mouse muscle fibres acutely exposed to the reducing agent dithiothreitol (DTT) showed decreased submaximal (30–60 Hz) force. Acute exposure to H₂O₂, or its non‐metabolizable analogue tert‐butyl hydroperoxide (t‐BOOH), has the opposite effect, that is, it increased submaximal force (Andrade et al. 1998 a, 2001; Cheng et al. 2015). The effect on [Ca²⁺]_i during contractions was small both with DTT and H₂O₂/t‐BOOH exposure, hence the changes in force production were explained by altered myofibrillar Ca²⁺ sensitivity. The force‐potentiating effect of H₂O₂/t‐BOOH exposure was transient and prolonged (more than 5 min) exposures resulted in a marked decrease in submaximal force, again accompanied by only minor changes in tetanic [Ca²⁺]_i (Andrade et al. 1998 a, 2001; Cheng et al. 2015) (Fig. 2 A). Intriguingly, the depression of submaximal force induced by prolonged exposure to H₂O₂ was reversed by exposure to DTT and, vice versa, the depression induced by initial exposure to DTT was reversed by H₂O₂ (Andrade et al. 1998 a) (Fig. 2 B). To sum up, myofibrillar Ca²⁺ sensitivity is highly susceptible to acute exogenous exposure to oxidizing and reducing agents, the effect is readily reversed by the opposite redox challenge, and rested muscle fibres appear to be in a suboptimally reduced state (Andrade et al. 1998 a; Lamb & Westerblad, 2011; Powers et al. 2011).

Figure 2. Transient and reversible effects of H₂O₂ on myofibrillar contractile function .

Original [Ca²⁺]_i and force records from submaximal (50 Hz) contractions of single FDB muscle fibres exposed to H₂O₂ (300 μm) for up to 8 min (A) and H₂O₂ for 6 min followed by exposure to the reducing agent DTT (1 mm) for 10 min (B). Note that H₂O₂ causes major force changes whereas [Ca²⁺]_i is little affected, which means that H₂O₂ mainly acts at the myofibrillar level. The effects on force are time dependent with brief H₂O₂ exposure resulting in increased submaximal force, which is followed by a progressive force decline (A). Furthermore, the force depression caused by prolonged H₂O₂ exposure is reversed by reduction with DTT (B). Conversely, prolonged exposure to DTT results in depressed submaximal force that can be reversed by application of H₂O₂ (not shown). Figure adapted from Andrade et al. (1998).

The effect of H₂O₂ application on the myofibrillar Ca²⁺ sensitivity of skinned fast‐twitch fibres differs markedly depending on the presence or absence of myoglobin and glutathione, which are normally present in intact skeletal muscle fibres (Murphy et al. 2008; Lamb & Westerblad, 2011). Thus, application of H₂O₂ on its own had little effect in skinned rat fast‐twitch fibres, whereas myofibrillar Ca²⁺ sensitivity was severely decreased in the presence of myoglobin. The proposed mechanism was that H₂O₂, through the Fenton reaction, interacts with Fe²⁺ on myoglobin to produce the highly reactive OH^• (Murphy et al. 2008). On the other hand, application of H₂O₂ to skinned fast‐twitch fibres in the presence of myoglobin and glutathione resulted in an initial increase in myofibrillar Ca²⁺ sensitivity followed by a decrease (Murphy et al. 2008), i.e. a pattern very similar to that observed when intact fast‐twitch fibres are exposed to H₂O₂ or t‐BOOH (see Fig. 2). Subsequent experiments on skinned fast‐twitch rat and human muscle fibres revealed a likely mechanism for the initial increase in myofibrillar Ca²⁺ sensitivity: GSH reacts with oxidized cysteine residues on the fast isoform of TnI (TnI_f; probably Cys133) and the resulting S‐glutathionylation increases myofibrillar Ca²⁺ sensitivity (Mollica et al. 2012). Accordingly, human muscle showed a marked increase in TnI_f S‐glutathionylation following 40 min of low‐intensity cycling at ∼60% peak oxygen consumption (Mollica et al. 2012), and a recent study shows that repeated stimulation can increase myofibrillar Ca²⁺ sensitivity by S‐glutathionylation of TnI_f in in situ experiments performed on rat gastrocnemius muscles (Watanabe et al. 2015). Notably, no corresponding increase in myofibrillar Ca²⁺ sensitivity was observed in mammalian slow‐twitch fibres, or in chicken and toad fibres, which have TnI isoforms that lack the equivalent of Cys133 (Mollica et al. 2012).

Intact fast‐twitch fibres showed several changes (albeit relatively minor, ∼5%) in cross‐bridge function in response to acute exposure to either H₂O₂ or t‐BOOH: decreased maximum shortening velocity, increased maximum force production (i.e. force at saturating [Ca²⁺]_i), and increased rate of force redevelopment after shortening (Andrade et al. 2001). The same treatment resulted in increased myofibrillar Ca²⁺ sensitivity and the faster force redevelopment might contribute to this increase (Brenner, 1988). The concentration of peroxide used by Andrade et al. (1 μm) would be within or close to the physiological range and hence endogenous increases in H₂O₂ occurring during, for instance, high‐intensity exercise are likely to affect cross‐bridge function (Andrade et al. 2001).

What about the effects of acute exposure to ROS/RNS other than H₂O₂ on myofibrillar function? Myofibrillar Ca²⁺ sensitivity appears not to be affected by O₂ ^•− (Bruton et al. 2008; Murphy et al. 2008). Acute exposure to NO^• donors leads to decreased myofibrillar Ca²⁺ sensitivity (Andrade et al. 1998 b; Dutka et al. 2011). Studies on intact muscle fibres exposed to NO^• donors indicate that the NO^•‐induced decrease in Ca²⁺ sensitivity is not caused by altered cross‐bridge kinetics (Morrison et al. 1996; Andrade et al. 1998 b) and therefore it would be due to impaired Ca²⁺–Tn interaction. ONOO^•− is known to decrease both myofibrillar Ca²⁺ sensitivity and maximum force in slow‐ and fast‐twitch muscles (Dutka et al. 2011).

Effects of ROS/RNS on contractile function during fatigue and recovery

Studies of human exercise show a clear positive effect of decreased ROS/RNS on endurance when fatigue is induced by submaximal contractions, whereas the effect is small or absent with maximal contractions (Reid et al. 1994; Powers et al. 2011). ‘Classical’ fatigue‐causing factors may dominate during the latter type of exercise and potential effects of ROS/RNS would therefore be difficult to discern (Allen et al. 2008). Such ‘classical’ factors that contribute to the force decrease in acute fatigue include accumulation of inorganic phosphate ions due breakdown of creatine phosphate (Dahlstedt et al. 2003), depletion of inter‐ and intramyofibrillar glycogen (Ørtenblad et al. 2013; Nielsen et al. 2014), and impaired action potential propagation (Pedersen et al. 2004; de Paoli et al. 2013). In a general sense, the positive effects of reducing ROS/RNS during fatigue with submaximal contractions fit with the fact that acute ROS/RNS effects are most marked on the steep part of the force–Ca²⁺ relationship (see above). However, acute exogenous application of H₂O₂ results in a transient increase in myofibrillar Ca²⁺ sensitivity (see Fig. 2) and some skinned fibre experiments show increased rather than decreased myofibrillar Ca²⁺ sensitivity after fatiguing contractions (Gejl et al. 2016; Watanabe et al. 2015). Thus, these results suggest that reducing ROS/RNS during fatigue would impair rather than improve performance. One tentative explanation for this apparent conflict is that exogenously applied and endogenously produced H₂O₂ have different effects. Another tentative explanation is that the effect of increases in ROS/RNS other than H₂O₂ dominates; for instance, both NO^• and ONOO^•− have been shown to decrease myofibrillar Ca²⁺ sensitivity (Andrade et al. 1998 b; Dutka et al. 2011). Furthermore, deleterious effects of oxidants may overpower any potentiating effects when the physical activity is prolonged and/or highly intense (Lamb & Westerblad, 2011).

In contrast to the situation during the actual induction of acute fatigue, obvious effects of increased ROS/RNS are seen during the subsequent recovery period (Westerblad & Allen, 2011). For instance, isolated mouse soleus fibres did not fatigue prematurely when exposed to severe oxidative stress (fatigued at 43°C in the presence of 10 μm H₂O₂ or t‐BOOH), but contractures developed and the fibres died ∼10 min after the end of stimulation (Place et al. 2009). Under less extreme conditions, fatigued muscle fibres frequently enter a prolonged state of severely depressed submaximal force, i.e. prolonged low‐frequency force depression (PLFFD) (Allen et al. 2008). At the muscle fibre level, depressed submaximal force can be due to decreased SR Ca²⁺ release and/or reduced myofibrillar Ca²⁺ sensitivity. Acutely fatigued fast‐twitch mouse FDB fibres displayed a marked PLFFD that is mainly caused by decreased SR Ca²⁺ release (Westerblad et al. 1993). Intriguingly, the cause of PLFFD changes towards reduced myofibrillar Ca²⁺ sensitivity in genetically modified mouse FDB fibres overexpressing the mitochondrial matrix redox enzyme SOD2 (Bruton et al. 2008), in rat FDB fibres that endogenously express more SOD2 (Bruton et al. 2008), and in mouse FDB fibres treated with the mitochondria‐targeted antioxidant SS‐31 or the NOS inhibitor l‐NAME (Cheng et al. 2015). In a recent study we show that PLFFD is accompanied by RyR1 fragmentation in vastus lateralis muscles of recreationally active human subjects after one session of high‐intensity interval training (6 × 30 s all‐out cycling) (Place et al. 2015). Conversely, when elite endurance athletes performed the same exercise, a similar PLFFD was observed but the RyR1 remained intact. The elite endurance athletes had higher levels of the antioxidant enzymes SOD2 and catalase in their muscles than recreationally active subjects. Moreover, a similar RyR1 fragmentation could be induced by high‐intensity stimulation of isolated mouse FDB muscles and this fragmentation was blocked by the antioxidant NAC (Place et al. 2015). Collectively, these results indicate that accumulation of mitochondrially generated O₂ ^•−, or ONOO^•−, preferentially affects SR Ca²⁺ release, probably via redox modifications of RyR1 (Bellinger et al. 2008 a,b; Andersson et al. 2011; Lanner et al. 2012). Conversely, when O₂ ^•− is more efficiently metabolized, the resulting increase in H₂O₂, or downstream products, preferentially leads to changes in myofibrillar Ca²⁺ sensitivity. In other words, antioxidants do not prevent PLFFD, but they can change the underlying mechanism from impaired SR Ca²⁺ release to reduced myofibrillar Ca²⁺ sensitivity. The question is then: Does this matter? Our answer is: Yes, it probably does.

Impaired SR Ca²⁺ release caused by redox modifications of RyR1 is associated with increased SR Ca²⁺ leak at rest, and the resulting increase in resting [Ca²⁺]_i may stimulate mitochondrial biogenesis and thereby improve muscle endurance (Wright et al. 2007; Bruton et al. 2010). On the other hand, major adaptations are unlikely to be triggered when PLFFD is caused by decreased myofibrillar Ca²⁺ sensitivity. Accordingly, prolonged changes in gene expression were recently shown after one session of high‐intensity interval training that induced PLFFD accompanied by ROS/RNS‐dependent RyR1 fragmentation in recreationally active subjects. Conversely, prolonged changes in mRNA levels were not observed in elite endurance athletes where PLFFD occurred while RyR1 remained intact (Place et al. 2015). Thus, ROS/RNS‐induced changes in RyR1 structure and function provide a mechanism as to why treatment with antioxidants hamper the beneficial effects of endurance training (Gomez‐Cabrera et al. 2008; Ristow et al. 2009; Paulsen et al. 2014).

A recent study from our laboratory highlights the complexity of fatigue‐induced redox effects (Cheng et al. 2015). Neither exposure to DTT nor t‐BOOH had any clear‐cut effect on the impaired SR Ca²⁺ release during PLFFD. Unexpectedly, application of DTT as well as t‐BOOH resulted in a major, but transient, increase in myofibrillar Ca²⁺ sensitivity during PLFFD. The effect of DTT during PLFFD was opposite to that in the rested state, where DTT decreased force production, which implies that redox‐sensitive sites were suboptimally reduced under resting conditions, became too oxidized during fatigue, and this was reversed by DTT. The effect of t‐BOOH was more intriguing since the results showed a transiently improved myofibrillar force production both in the unfatigued state and during PLFFD. Thus, some functionally important sites on myofibrillar proteins appeared to remain in a suboptimally reduced state during fatigue and the function of these was then temporarily improved by exogenous peroxide application (Fig. 3). While these intriguing findings illustrate the complexity of redox effects, further studies are clearly required before they can be understood at a more precise mechanistic level.

Figure 3. The prolonged low‐frequency force depression (PLFFD) after fatiguing stimulation is the result of complex ROS/RNS effects on SR Ca²⁺ release and myofibrillar Ca²⁺ sensitivity .

A, [Ca²⁺]_i (upper panel) and force (lower panel; mean data ± SEM) in 30 Hz contractions produced during PLFFD (red circles) initially in standard Tyrode solution, followed by addition of DTT (1 mm) or the non‐metabolizable analogue of H₂O₂, t‐BOOH (10 μm). The effect of the same exposures on unfatigued fibres are also shown (green circles). B, simplified model of ROS/RNS effects on SR Ca²⁺ release and myofibrillar Ca²⁺ sensitivity. Key proteins for SR Ca²⁺ release are the t‐tubular voltage sensors, the dihydropyridine receptors (blue boxes), and the SR Ca²⁺ release channels, RyR1 (green boxes). These proteins appear to be in an optimal redox state at rest and become overly oxidized during fatiguing stimulation resulting in decreased [Ca²⁺]_i, which is not affected by application of either t‐BOOH or DTT (see A). In the rested state, myofibrillar proteins are in a suboptimal reduced state. Some myofibrillar proteins become overly oxidized during induction of fatigue and the resulting force decrease is transiently counteracted by application of DTT. Intriguingly, other myofibrillar proteins apparently remain reduced during fatigue since application of the oxidizing agent t‐BOOH temporarily improves force generation, i.e. similar to the effect in the rested state (see A). Figure adapted from Cheng et al. (2015).

Conclusions

In this review we discuss intricate ROS/RNS effects on myofibrillar contractile function. While mechanisms underlying some changes in myofibrillar function are fairly well established, others are more uncertain. For instance, it is fairly well established that S‐glutathionylation of TnI_f can cause an initial increase in myofibrillar Ca²⁺ sensitivity during exogenous H₂O₂ exposure, whereas the mechanism(s) behind the decreased sensitivity during prolonged exposure remains uncertain. Due to such uncertainties we find it premature to discuss complex changes in contractile function, e.g. mechanisms underlying PLFFD, in terms of one specific ROS/RNS acting on one specific molecular target. This type of knowledge would, for instance, require improved methods to measure changes in a specific ROS/RNS with good temporal and spatial resolution. Moreover, an obvious risk with reductionistic approaches to assess the complex ROS/RNS effects is that simplistic cellular or subcellular experiments are prioritized to get at the molecular mechanism. Our opinion is that such experiments often provide the correct answer to the wrong question.

In studies of ROS/RNS‐induced changes in myofibrillar function, we find it essential to measure both force and [Ca²⁺]_i, since both are likely to be affected. Experiments on enzymatically isolated muscle fibres are technically much easier to perform than experiments on dissected fibres. However, enzymatically dissociated fibres lack tendons and force measurements become cumbersome. Intact whole muscles are also easier to use than dissected single muscle fibres, but [Ca²⁺]_i is then difficult to measure, especially in deeper parts of the muscle. The mechanism of PLFFD shifts from decreased tetanic [Ca²⁺]_i in wild‐type muscles to reduced myofibrillar Ca²⁺ sensitivity in SOD2‐overexpressing muscles (Bruton et al. 2008). Experiments with enzymatically dissociated fibres would detect decreased tetanic [Ca²⁺]_i during recovery of wild‐type fibres, whereas [Ca²⁺]_i returned to the pre‐fatigue level with SOD2 overexpression. Thus, the conclusion would be that increased SOD2 activity prevents PLFFD. Conversely, experiments with intact whole muscles would detect a similar force depression during recovery of wild‐type and SOD2‐overexpressing muscle, and the conclusion would be that SOD2 overexpression has no effect on PLFFD. Thus, the correct conclusion requires experiments with simultaneous measurements of force and [Ca²⁺]_i in fully intact fibres, or the combination of experiments with enzymatically dissociated fibres and whole muscles.

A wide range of antioxidants, including gp91ds‐tat and SS‐31, could not prevent the loss of force observed after acute fatiguing stimulation of healthy mouse muscle fibres (Cheng et al. 2015). Moreover, improved antioxidant capacity in muscles of endurance athletes did not prevent the development of PLFFD after a session of high‐intensity interval training (Place et al. 2015). On the other hand, prolonged administration of antioxidants (e.g. vitamin C and E) has been shown to hamper the beneficial effects of endurance training in humans and rodents (Gomez‐Cabrera et al. 2008; Ristow et al. 2009; Paulsen et al. 2014). Thus, the overall impact of antioxidant supplementation in association with endurance training is negative: antioxidants do not prevent the exercise‐induced prolonged ROS/RNS‐dependent decline in contractile function, whereas they can hamper the beneficial adaptations that come with endurance exercise.

Additional information

Competing interests

None declared.

Author contributions

All authors have approved the final version of the manuscript and agree to be accountable for all aspects of the work. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.

Funding

The authors acknowledge support from the Swedish Research Council and the Swedish National Centre for Sports Research.

Biographies

Arthur J. Cheng is a researcher at Karolinska Institutet. His research is focused on skeletal muscle fatigue and recovery and the role of ROS/RNS in these processes.

Johanna T. Lanner is assistant professor and leads a research group at Karolinska Institutet. She has discovered novel ways to regulate the ryanodine receptor and her recent research focuses on molecular mechanisms in skeletal muscle weakness.