In vertebrates, skeletal muscle is a plastic and dynamic tissue that can be remodeled in response to different stimuli, such as exercise or physical activity, and that contributes to maintain metabolic homeostasis. In response to exercise-induced contractile activity, skeletal muscle shows a remarkable adaptation that results in changes in protein content and metabolic regulation that are evidenced by the induction or repression of exercise-specific signaling pathways. In mammals, a frequent adaptation in skeletal muscle following exercise is fiber hypertrophy (Hawley et al., 2014) resulting from an increase in protein synthesis over a reduction in protein degradation that causes increased muscle mass and improved force generation. The mammalian target of rapamycin complex (mTORC), which integrates many intracellular signals in order to regulate anabolic processes, cell cycle progression and autophagy, is suggested to be the main signaling pathway enhancing protein synthesis and to promote skeletal muscle fiber hypertrophy in response to exercise in mammals (Laplante and Sabatini, 2012; Egan and Zierath, 2013; Hawley et al., 2014). Moreover, exercise has been reported to increase the number of satellite cells, that is, muscle progenitor cells that are considered to be essential during muscle maintenance and regeneration in response to injury (Relaix and Zammit, 2012), both in young and old animals (Shefer et al., 2010, 2013; Leiter et al., 2011; Lee et al., 2012; Smith and Merry, 2012). Thus, the potential ability of exercise to induce satellite cell activation is an important matter of investigation with regard to muscle diseases and sarcopenia. In addition, exercised skeletal muscle develops an increased aerobic capacity that has been associated with elevated mitochondrial biogenesis and oxidative metabolism, changes in muscle fiber type or increased oxygen availability as a result of increased muscle capillarization (Egan and Zierath, 2013; Hawley et al., 2014). One of the key intracellular factors involved in skeletal muscle metabolic responses to exercise is the AMP-activated protein kinase (AMPK). AMPK acts as an “energy sensing enzyme” to maintain energy homeostasis in the cells by stimulating metabolic adaptations to increase ATP generation processes such as glucose uptake and utilization or lipid oxidation (Hardie et al., 2012). In addition, AMPK modulates the activity of several factors such as peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α) and controls the transcription of genes encoding proteins involved in oxidative metabolism and mitochondrial biogenesis such as PPARs, NRF1, MEF2 and HDACs (Hardie et al., 2012). In mammals, AMPK is activated in response to muscle contraction or endurance exercise (Winder and Hardie, 1996; Narkar et al., 2008; Nedachi et al., 2008) in an intensity- and duration-dependent manner (Bergeron et al., 2001; Egan et al., 2010).

It is well known that regular exercise is associated with important benefits for human health that include protection against the development of cardiovascular diseases, stimulation of the immune system through its anti-inflammatory effects and maintenance of glucose homeostasis, among others (Fiuza-Luces et al., 2013; Hawley et al., 2014; Benatti and Pedersen, 2015). However, although a great deal of research has been conducted on the beneficial effects of exercise, whether skeletal muscle plays a central role orchestrating the organismal response to exercise is not completely understood. Exercise-induced molecular mechanisms in skeletal muscle that promote health benefits and adaptations may involve that production of muscle-secreted factors or myokines (Pedersen and Febbraio, 2012). Myokines, such as interleukin-6, are expressed by the contracting skeletal muscle and have been suggested to act locally in the muscle tissue or to function at a systemic level, influencing the metabolism and function of several organs such as adipose tissue, liver and blood vessels (Pedersen and Febbraio, 2012). Therefore, identification of the mechanisms by which exercise-induced contractile activity regulates myokine production can improve our understanding of the potential local and systemic roles of myokines in the adaptive responses to exercise in vertebrates.

In non-mammalian vertebrates, including teleost fish, physical activity linked to locomotion has also been shown to greatly influence the metabolic and contractile characteristics of skeletal muscle and, consequently, the performance of the organism in the face of migratory, predatory, behavioral and other challenges that affect survival. In teleost fish, in particular, the physiological responses to swimming in terms of skeletal muscle remodeling (e.g. muscle growth and metabolic adaptation) appear to be similar to those described in mammals (McClelland, 2012; Rodnick and Planas, 2016). For example, sustained swimming stimulates hypertrophy of fast muscle fibers in various fish species (Bugeon et al., 2003; Martin and Johnston, 2005; Ibarz et al., 2011). Despite the parallel cellular response of skeletal muscle to exercise-induced contractile activity in fish and mammals, the arrangement of fast and slow muscle fibers within the musculature is different between these two vertebrate groups. In fish, skeletal muscle is spatially segregated and consists of deep fast muscle fibers that represent approximately 90% of the total muscle mass (also named white muscle) and that are covered by a thin superficial layer of slow muscle fibers (also named red muscle) (Bone, 1978). In fish, these two types of muscles also show functional differences, given that the fast muscle is known to be a fast-contracting, anaerobic muscle that permits sudden bursts of motion, whereas the slow muscle is a slow-contracting, aerobic muscle that permits sustained locomotion (Bone, 1978). However, as indicated above, fast muscle in fish is able to respond to sustained swimming by increasing fiber size and by adopting a more aerobic phenotype (Rodnick and Planas, 2016) and, consequently, shows plasticity. Therefore, in view of the potentially important contribution of fast muscle to the known growth-promoting effects of sustained swimming in fish (Davison and Herbert, 2013) and to the spatial segregation of fast and slow muscle (Bone, 1978), efforts devoted to investigate the physiological responses of fast muscle to swimming-induced exercise are important for our understanding of growth-regulatory mechanisms in fish and other vertebrates.

Given the evolutionarily conserved plasticity of skeletal muscle in response to contractile activity and given that the zebrafish (Danio rerio) is a recognized teleost model for vertebrate research, this species is becoming an emerging model in exercise physiology. To date, the zebrafish exercise model has been applied in studies on muscle growth and development, metabolism, neurobiology and aging (Pelster et al., 2003; McClelland et al., 2006; LeMoine et al., 2010a; Palstra et al., 2010, 2014; Gabriel et al., 2011; Fiaz et al., 2014; Gilbert et al., 2014; Severi et al., 2014; Hasumura and Meguro, 2016). In a previous study, we established the swimming economy of adult zebrafish and determined its optimal swimming speed (U_opt) (Palstra et al., 2010). Under these sustained, aerobic exercise conditions, we also reported that swimming-induced exercise stimulated somatic growth and increased hypertrophy and capillarization of fast skeletal muscle fibers that was accompanied by the activation of transcriptional programs involved in muscle growth, myogenesis, oxidative metabolism and angiogenic processes (Palstra et al., 2010, 2014). In the present study we have aimed at further characterizing the temporal adaptive responses of fast skeletal muscle to swimming-induced exercise in adult zebrafish by investigating the signaling pathways and cellular responses potentially involved in the stimulatory effects of exercise on fast muscle hypertrophy and the transition toward a more aerobic phenotype. Furthermore, we have also investigated in adult zebrafish fast skeletal muscle the exercise-induced regulated expression of myokines known to be involved in the exercise response in mammals. Our results strongly support the notion that exercise increases skeletal muscle growth and myogenesis, at least in part, through the coordinated activation of the mTOR-MEF2 and AMPK signaling pathways. Our results also suggest that the observed exercise-induced responses in fast muscle may have taken place in response to direct contractile events but also indirectly through the local production and action of myokines.

In teleost fish, skeletal muscle adaptations to swimming-induced exercise include the potentiation of muscle mass and the increase of the oxidative characteristics of muscle fibers (Johnston et al., 2011; McClelland, 2012; Rodnick and Planas, 2016). In adult zebrafish, an emerging physiological exercise model, we previously demonstrated that sustained swimming potentiates somatic growth by increasing fast fiber hypertrophy (Palstra et al., 2014), as confirmed by other studies (Hasumura and Meguro, 2016). Furthermore, by conducting transcriptome profiling of the exercised fast muscle in adult zebrafish we previously identified significant changes in the expression of canonical pathways important for the regulation of protein metabolism (i.e., IGF-1/PI3K/Akt/mTOR signaling pathways) and myofibrilogenesis (Palstra et al., 2014). Activation of mTOR signaling, a key pathway for the stimulation of protein synthesis, occurs concomitantly to exercise-induced skeletal muscle hypertrophy in mammals (Bodine, 2006; Egan and Zierath, 2013). Interestingly, the mechanism by which mTOR controls protein synthesis and muscle mass was recently elucidated using a zebrafish muscle inactivity model whereby muscle inactivity results in decreased abundance of Mef2, a transcription factor essential for normal fiber growth, due to the inability of mTOR to repress the inhibitory effects of the eukaryotic initiation factor 4E binding protein 4EBP3L on mef2ca translation (Yogev et al., 2013). In the present study, using a muscle activity model (i.e., swimming-induced exercise) we demonstrate that sustained swimming results in the activation of mTOR and in an increase in the protein abundance of Mef2, an important down-stream target of mTOR. Therefore, our results strongly suggest that the hypertrophic response of adult zebrafish fast skeletal muscle fibers to swimming-induced exercise may be the result of the observed increase in mTOR activity and Mef2 protein levels and support the notion that activity-induced muscle mass in adult zebrafish is dependent on mTOR activation of Mef2.

In parallel to mTOR-mediated increased protein synthesis and Mef2-induced myofibrilogenesis, the addition of new myonuclei to existing muscle fibers is another process that can contribute to the hypertrophic growth of muscle fibers. A number of studies in mammals suggest that exercise is an endogenous stimulus that causes satellite cell activation and these findings are particularly relevant for investigating the causes and potential treatment of muscle diseases such as sarcopenia (Markert et al., 2011; Shefer et al., 2013). However, to the best of our knowledge, the potential contribution of satellite-like cells in exercise-induced muscle growth or remodeling has not been explored in zebrafish. In the present study, we provide evidence supporting the hypothesis that exercise may have activated satellite cells. Most notably, the protein expression levels of Pax7, a transcriptional regulator that is a well-established marker for satellite cells, increased rapidly (after 1 week) in fast muscle in response to swimming-induced exercise, despite the lack of change in the mRNA expression levels of pax7a and pax7b, the two zebrafish pax7 isoforms, possibly due to post-transcriptional modifications, as reported in mammals (Chen J.-F. et al., 2010). Pax7-expressing satellite-like cells have recently been shown to be activated and to proliferate during muscle repair in adult zebrafish (Berberoglu et al., 2017). In addition, satellite cells in isolated zebrafish muscle fibers are also activated and proliferate in response to mechanical stretch (Zhang and Anderson, 2014). These studies indicate that adult zebrafish satellite cells have the ability to become activated and to proliferate in response to injury or mechanical forces, such as during muscle contraction. In support of the possibility of a satellite cell proliferative response to swimming-induced exercise in the zebrafish fast muscle, our results also show that the protein expression levels of H3-P, a mitotic marker, significantly increased, whereas the protein expression levels of p38-MAPK, a muscle cell differentiation marker, showed a trend toward an early decrease in abundance in fast muscle of exercised adult zebrafish. These results provide support to the notion that swimming-induced exercise may induce proliferation of myonuclei in the zebrafish skeletal muscle, as also supported by previously reported increases in the expression of the proliferation marker PCNA in exercised muscle in zebrafish (van der Meulen et al., 2006). Unfortunately, our results on the in vivo proliferation by EdU labeling did not show significant differences between exercised and non-exercised fast muscle and, therefore, we are not able to demonstrate an increase in proliferating myonuclei in response to a 2-week exercise training period by using an alternate method. We can speculate that the reason for this may be that the time point chosen to evaluate myonuclei proliferation (i.e., at week 2 of training), selected based on the temporal induction of H3-P protein expression, may have been too early by using this method and that later time points (i.e., at week 4 of training) may have been more appropriate for detecting possible increases in the number of myonuclei undergoing DNA synthesis and division in response to swimming-induced exercise. Overall, our results provide only partial support for the hypothesis that fast muscle cells (possibly satellite cells) were induced to proliferate in response to swimming-induced exercise in adult zebrafish and that this process may have contributed to the observed exercise-induced hypertrophy of fast muscle fibers in zebrafish. Obviously, additional studies will be needed to determine whether swimming-induced exercise can affect satellite cell proliferation in fish.

An important result from this study is the demonstration that the activity of AMPK, an important energy sensing molecule and a key regulator of metabolic pathways under energy-demanding conditions (Hardie et al., 2012), was significantly increased in fast muscle of exercised zebrafish. Interestingly, coinciding with the induction of AMPK activity, we observed an increase in the expression of Pgc1α, a known target of AMPK and an important mediator of some of the metabolic actions of AMPK in mammals (Correia et al., 2015). The temporal correlation of AMPK activation and the increased expression of pgc1α as well as of some of its targets, such as pparda and myosins (with smyhc1 being up-regulated and myhz2 being downregulated at week 4), indicative of increased oxidative metabolism and of a fiber-type switch toward a more aerobic phenotype (Luquet, 2003; Wang et al., 2004), suggests that exercise-induced AMPK may have increased Pgc1α expression. In support for the hypothesis that swimming-induced exercise may increase the expression and activity of Pgc1α as a result of AMPK activation, previous studies from our laboratory have shown that AMPK activity in fast and slow muscles from trout subjected to aerobic swimming training increases concomitantly with the mRNA expression levels of pgc1α and of known markers of mitochondrial activity and lipid metabolism (Magnoni et al., 2014). Furthermore, AMPK activation by either pharmacological (e.g., AICAR) or electrical stimulation of cultured trout myotubes results in increased pgc1α mRNA expression levels and in the up-regulation of the uptake, transport and utilization of glucose for ATP generation (Magnoni et al., 2012, 2014). Overall, these observations strongly support the notion that in fish, as in mammals, AMPK may constitute a key player orchestrating some of the metabolic adaptations induced by exercise in skeletal muscle, likely through activation of Pgc1α. In the present study, however, the mRNA expression levels of markers of mitochondrial biogenesis were not affected by exercise, in agreement with other reports claiming that Pgc1α may not be a regulator of mitochondrial biogenesis in fish, in contrast to mammals (LeMoine et al., 2008, 2010a). Some authors have even argued against the possibility that Pgc1α could be directly regulated by AMPK in zebrafish based on (1) the lack of conservation in the zebrafish Pgc1α sequence of one of two residues that are phosphorylated by AMPK in mammals and (2) the inability of mammalian AMPK to phosphorylate a short peptide corresponding to the Pgc1α zebrafish sequence that contains the non-conserved mammalian AMPK-sensitive residue (Thr¹⁷⁷) (LeMoine et al., 2010b; Bremer et al., 2016). Clearly, further studies are required to elucidate whether exercise- or pharmacologically-induced active zebrafish AMPK can phosphorylate and activate zebrafish Pgc1α.

A large body of evidence indicates that that the mammalian skeletal muscle produces and secretes factors (i.e., myokines) in response to exercise that can act on the same tissue in an autocrine, paracrine or endocrine fashion and that are involved in metabolic adaptation (Pedersen and Febbraio, 2012). In fish, no information is available on the ability of skeletal muscle to express or produce myokines in response to swimming-induced muscle contraction. Here, we report the temporal regulation of the expression of genes encoding myokines in fast skeletal muscle of adult zebrafish subjected to swimming-induced exercise. One of the most relevant findings is that swimming resulted in increased mRNA expression levels of IL6, known as the prototype myokine in mammals (Pedersen and Febbraio, 2012), in the adult zebrafish skeletal muscle. Functionally, IL6 is known to participate in the regulation of myogenesis and muscle growth in mammals through the regulation of satellite cell proliferation and myonuclear accretion (Serrano et al., 2008) and its circulating or expression levels in muscle are related with training adaptation in humans (Fischer, 2006). A similar early response to swimming-induced exercise was observed with the increased mRNA expression levels of the genes encoding the myokines sparc and decorin and the growth factor igf1 in fast muscle of adult zebrafish. These three myokines are known to participate in the regulation of muscle function in mammals. SPARC is expressed in satellite cells under exercise or regenerating conditions and is suspected to have a role in the maintenance of muscle function (Mousavi and Jasmin, 2006; Jørgensen et al., 2009; Matthews et al., 2009). Decorin induces muscle hypertrophy by modulating the activity of members of TGF-β family (Yamaguchi et al., 1990; Miura et al., 2006). In addition, IGF-1 is a key growth factor that stimulates protein synthesis and muscle hypertrophy through the mTOR pathway (Schiaffino et al., 2013). Therefore, our results on the parallel early increase in the mRNA expression levels of il6, sparc, decorin and igf1 in exercised zebrafish fast muscle suggest that these myokines, known to be secreted by the contractile muscle in mammals (Pedersen and Febbraio, 2012), may represent extracellular signals that could promote metabolic adaptive mechanisms in zebrafish muscle cells. Another relevant myokine-encoding gene that was temporally up-regulated in fast muscle of exercised zebrafish and that could also contribute to metabolic adaptation is apelin, a myokine that when induced by exercise improves insulin sensitivity in insulin resistant or obese conditions in mammals (Besse-Patin et al., 2014; Yang et al., 2015) and that stimulates muscle mitochondrial biogenesis and vascularization when overexpressed in mice (Yamamoto et al., 2011). In support of the hypothesis that these myokines may have been produced and secreted by fast muscle of exercised zebrafish and acted on fast muscle cells to promote metabolic remodeling and fast muscle fiber hypertrophy, the two key intracellular signaling pathways induced by exercise that lead to adaptive metabolic responses in the adult zebrafish fast muscle, namely the mTOR and AMPK pathways, are pathways activated in mammals by IGF-1 and by IL6, SPARC, apelin and decorin, among other myokines, respectively (Carey et al., 2006; Dray et al., 2008; Glund et al., 2009; Kelly et al., 2009; Matthews et al., 2009; Chen Y. et al., 2010; Song et al., 2010; Attané et al., 2012; Das et al., 2012; Goyal et al., 2014; Huh et al., 2014; Lee H. J. et al., 2015; Lee K. et al., 2015). The presence and observed up-regulation of the mRNA expression levels of genes encoding receptors for il6 (il6r) and apl (aplnra and aplnrb) in fast muscle of exercised zebrafish further supports the hypothesis of local effects for these two myokines. These results clearly show that exercise-induced contractile activity in zebrafish enhances the mRNA expression of genes encoding myokines in skeletal muscle as it occurs in mammals, suggesting a potential link between exercise-induced myokine production and fast muscle remodeling in zebrafish. Future studies are needed to demonstrate the ability of the zebrafish skeletal muscle to secrete and respond to exercise-induced myokines and to decipher their potential role in the metabolic adaptation mechanisms involved in the response to swimming-induced exercise in fish.

In conclusion, the results from the present study strongly suggest that the hypertrophic response of the zebrafish fast muscle to exercise may be due to, at least in part, to an mTOR- and Mef2- mediated increase in protein synthesis and myofibrilogenesis. In addition, our results also suggest that the metabolic adaptation of the zebrafish fast muscle to exercise may be orchestrated by AMPK and Pgc1α. The observed exercise-induced increase in the expression of genes encoding myokines in the zebrafish fast muscle leads us to propose a testable working model in which secreted myokines may contribute to the muscle remodeling response to exercise in adult zebrafish by directly promoting metabolic adaptive responses through activation of the mTOR and AMPK signaling pathways that may underlie the hypertrophic response to swimming-induced exercise.

MR and JP designed the study; MR and GA conducted the experiments; MR and JP wrote the manuscript.