The role of nNOS and PGC-1α in skeletal muscle cells

Nitric oxide (NO) is a free radical molecule that is physiologically synthesised in a tightly controlled manner by a family of cytochrome-P450-like flavo-haem proteins, the NO synthases (NOSs) (Forstermann and Sessa, 2012). In mammals, three quite distinct NOS isoforms have been described. Neuronal NOS (nNOS or NOS1) was the first isoform to be reported; it is predominantly found in neuronal tissue and is also highly expressed in skeletal muscle. Inducible NOS (iNOS or NOS2) is expressed on demand in a wide range of cells (particularly macrophages) and tissues, whereas endothelial NOS (eNOS or NOS3) was initially described in vascular endothelial cells but has subsequently been shown to also be expressed in other cell types, including cardiomyocytes and adipocytes (Forstermann and Sessa, 2012; Tanaka et al., 2003; Wei et al., 1996). NOSs are the products of distinct genes and have different regulation, catalytic properties and inhibitor sensitivities. The three NOS isoforms function as homodimers and catalyse the two-step oxidation of L-arginine to citrulline and NO. The NOS monomer exhibits a bidomain structure; it has an N-terminal oxygenase domain, which contains the binding sites for haem and L-arginine, and a C-terminal reductase domain, which contains binding sites for the essential cofactors flavin adenine dinucleotide (FAD), flavin mononucleotide (FMN) and nicotinamide adenine dinucleotide phosphate (NADPH). nNOS contains an additional PDZ (postsynaptic density protein 95/discs large/ZO-1 homology domain) domain located at the N-terminus, through which it can directly interact with the PDZ domains of other proteins. These interactions determine the subcellular distribution and activity of nNOS (Fig. 1) (Zhou and Zhu, 2009; Aquilano et al., 2014).

The mechanism by which NO exerts its physiological action is peculiar and considerably different from that of other second messengers. Once produced, NO can exist as reduction and oxidation products (i.e. NO⁻, NO⁺) and, owing to its high diffusibility and permeability, it does not need to be transported through membrane carriers or pores. Furthermore, NO does not need to interact with intracellular or extracellular receptors, as is the case for other second messengers, in order to exert its biological function. NO has been implicated in several processes, including fertilisation, embryogenesis, modulation of neurotransmission, vasodilatation and inflammatory response (Alderton et al., 2001).

In the past two decades, NO has been receiving increasing attention. Besides canonical NO signalling, which consists of the selective activation of soluble guanylate cyclase (sGC) and inhibition of cytochrome c oxidase through the interaction of NO with haem, other signalling modes have been discovered, including the inhibition of mitochondrial respiratory complexes and the Krebs cycle enzyme aconitase (Cooper, 1999; Valdez et al., 2000). Moreover, S-nitrosylation of certain reactive protein cysteines has been recognised as another fundamental aspect of NO-mediated signalling (Stamler et al., 1992). This has opened a new area of research and has tremendously increased the spectrum of biological activities of NO and its targets. For instance, NO has been implicated in mitochondrial biogenesis and lipid metabolism (Aquilano et al., 2014; Schild et al., 2006). A master regulator of both mitochondrial function and lipid catabolism is peroxisome proliferator activated receptor γ coactivator 1α (PGC-1α) (Nisoli et al., 2004; Piantadosi and Suliman, 2012); it acts as a coactivator of a wide array of transcription factors that are involved in the expression of oxidative phosphorylation (OXPHOS) and mitochondrial fatty acid oxidation genes (Aquilano et al., 2013b; Lettieri Barbato et al., 2014). Particularly in skeletal muscle, PGC-1α plays a key role in coordinating the oxidation of intracellular fatty acids with mitochondrial remodelling (Scarpulla et al., 2012) and, for this reason, modulation of NO metabolism and/or targeting of PGC-1α could be a promising strategy to combat several myopathies that are associated with altered mitochondrial metabolism. However, in our opinion, the intersection between nNOS- and PGC-1α-governed pathways in skeletal muscle remains only poorly emphasised at present. Here, we review the most important metabolic routes that are governed by nNOS and PGC-1α, in order to highlight how they cooperatively modulate skeletal muscle contraction and oxidative metabolism.

In skeletal muscle, nNOS and eNOS isoforms are expressed constitutively, whereas iNOS is only expressed during inflammatory responses, such as those occurring in individuals with type 2 diabetes (Torres et al., 2004). Signalling by the nNOS splice variant nNOSμ (Fig. 1) is essential for skeletal muscle health and is commonly reduced in neuromuscular diseases, including Duchennes's dystrophy (Brenman et al., 1995). nNOSμ is thought to be the predominant source of NO in skeletal muscle. However, it has been reported that the nNOSβ variant, lacking the PDZ domain, is also present and has a crucial role in cell signalling (Percival et al., 2010). In skeletal muscle, the different nNOS splicing variants have important roles in the regulation of many muscle functions, including blood flow, contraction and muscle metabolism (Suhr et al., 2013).

The physiological functions of NO have been initially explained by it being a freely diffusible messenger that acts on targets that are distant from its site of synthesis (Lancaster, 1994). However, skeletal muscle displays high concentrations of potent NO scavengers, such as myoglobin or glutathione (GSH), which could considerably limit diffusion-based NO signalling (Flögel et al., 2001). In addition, NO is extremely reactive with a wide range of biomolecules; therefore, the specificity of NO signalling requires that NOSs are located in the vicinity of NO effector protein targets (Sullivan and Pollock, 2003; Zhou and Zhu, 2009). For example, in myotubes, anchoring of nNOSμ at the plasma membrane through specific interaction with the scaffold protein α-syntrophin, which forms part of the dystrophin-associated glycoprotein complex (Fig. 2i), assures the coupling of NO production to Ca²⁺ influx, which is essential for skeletal muscle contraction (Brenman et al., 1996; Thomas et al., 1998). The sarcolemmal nNOSμ has a higher activity in type II (fast twitch) fibres than in type I (slow twitch) fibres. Furthermore, by acting in synergy with eNOS that is expressed in the blood vessels, sarcolemmal NOSμ enhances blood flow and oxygen delivery, thereby more efficiently matching the blood supply to the metabolic demands of active muscle (Thomas et al., 2003). In particular, NO that is derived from sarcolemmal nNOSμ finely attenuates α-adrenergic vasoconstriction and increases blood supply in contracting skeletal muscle (Thomas et al., 1998; Thomas et al., 2003). Indeed, mice with reduced sarcolemmal nNOSμ show a considerable decrease in blood supply in electrically evoked muscle contraction, with consequent functional ischemia and fatigue, as well as reduced exercise capacity (Thomas et al., 2003). nNOSμ has also been shown to be present in a membrane-unbound form in gastrocnemius homogenates (Thomas et al., 2003). However, in the absence of sarcolemmal nNOSμ, the soluble form is not able to compensate for the loss of its sarcolemmal counterpart in contractile performance (Thomas et al., 1998). However, the exact function of cytoplasmic nNOSμ remains to be elucidated fully.

As mentioned above, skeletal muscle cells also express the PDZ-lacking nNOS splice variant nNOSβ, which is mostly soluble but has also been found to localise at the Golgi complex in mice (Percival et al., 2010) (Fig. 2ii). In contrast to muscles that lack only nNOSμ, muscles lacking both nNOSμ and nNOSβ show severe myopathy and exhibit microstructural cellular defects of the cytoskeleton, Golgi complex and mitochondria (Percival et al., 2010). Morphologically, skeletal muscles that lack both nNOSμ and nNOSβ have a reduced mass, are intrinsically weak and highly susceptible to fatigue, and manifest marked post-exercise weakness (Percival et al., 2010), suggesting that nNOSβ also modulates the ability of skeletal muscle to maintain force production during and after exercise.

It has been reported that nNOS also localises in the nucleus (Rothe et al., 1999; Riefler and Firestein, 2001). Work from our group has confirmed this by showing the presence of nNOS in the nucleus of skeletal muscle cells, where it is involved in transcriptional regulation by acting as an inhibitor of the constitutive transcription factor Sp1 (Baldelli et al., 2008). The nuclear expression of nNOS impairs the binding of Sp1 to the promoter of the sod1 gene, which encodes copper, zinc superoxide dismutase (SOD1). However, inhibition of SOD1 expression is not triggered by NO signalling, because the administration of specific nNOS inhibitors did not restore SOD1 expression (Baldelli et al., 2008). We have also demonstrated that the formation of the nNOS–Sp1 complex at the sod1 promoter is promoted by the PDZ domain of nNOS. Indeed, the expression of an nNOS mutant that lacks the PDZ domain (ΔnNOS nNOSβ) dramatically reduces the interaction of nNOS with Sp1. Interestingly, we found that the loss of nNOS–Sp1 complex formation is due to the inability of ΔnNOS nNOSβ to translocate to the nucleus (Baldelli et al., 2011), raising the intriguing question of the importance of nNOS splicing in regulating its nuclear function. More recently, we have demonstrated that nNOS recruitment to the nucleus through its PDZ domain is generally applicable, as nNOS can be targeted to the nucleus in neuroblastoma and HeLa cells and, more importantly, in myocytes (Aquilano et al., 2014).

The regulation of gene expression by NO is well established, and S-nitrosylation of transcription factors has been suggested to be the main mechanism through which their activity is affected. S-nitrosylation of nuclear proteins has been predominantly ascribed to the trans-S-nitrosylation activity of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (Kornberg et al., 2010). However, the exact function of nuclear nNOS has remained elusive even though compelling evidence has shown that many of the factors that are involved in its activity (e.g. sGC subunits, calmodulin and enzymes involved in tetrahydrobiopterin synthesis) are present in the nucleus (Bachs et al., 1992; Elzaouk et al., 2004; Pifarré et al., 2009). Besides regulating transcription by interacting with transcription factors (i.e. by the formation of nNOS–Sp1 complex), we have found that, during myocyte differentiation, nuclear nNOS-derived NO functions by driving mitochondrial biogenesis (Aquilano et al., 2014). In fact, in differentiating myocytes, nNOS expression is induced, which is accompanied by its redistribution to the nucleus and increased S-nitrosylation of nuclear proteins, including cAMP response element-binding protein (CREB). S-nitrosylated CREB (CREB-SNO) more efficiently engages with the promoter of the gene encoding PGC-1α, which contains a CREB consensus sequence. This results in a strong induction of mitochondrial biogenesis, which is crucial for myocyte differentiation and function (Fig. 2iii,iv). Although nNOSβ (lacking the PDZ domain) fully retains the ability to synthesise NO, this construct is unable to induce mitochondrial biogenesis because it fails to be recruited to the nucleus and is thus unable to favour CREB S-nitrosylation and subsequent PGC-1α-mediated mitochondrial biogenesis (Aquilano et al., 2014).

We have also shown that nuclear α-syntrophin is the genuine mediator of nuclear recruitment of nNOS (Aquilano et al., 2014) (Fig. 2). By knocking down α-syntrophin in differentiating myocytes, we have shown that lack of nNOS targeting to the nucleus through α-syntrophin was the underlying cause of impaired mitochondrial biogenesis in these cells.

In conclusion, these findings suggest that an adequate level of expression of both nNOSμ and nNOSβ is important for the regulation of contractile activity and in controlling fatigue. Furthermore, in addition to sarcolemmal anchoring, α-syntrophin-mediated nNOS anchoring to the nuclear envelope also participates in NO signalling that promotes mitochondrial biogenesis and differentiation of skeletal muscle cell precursors. Therefore, the control of nNOS splicing and its subcellular localisation in skeletal muscle could represent a novel therapeutic avenue to prevent or treat myopathies.

NO influences myocyte metabolism by multiple means; it promotes glucose uptake, while inhibiting mitochondrial respiration, glycolysis and phosphocreatine breakdown. Each of the activities of NO is likely to involve a distinct molecular target and has been shown to be reversible (Stamler and Meissner, 2001).

The best-known effect of NO on skeletal muscle metabolism is its capacity to inhibit mitochondrial respiration directly by interfering with electron transport chain complexes in several ways; (1) NO inhibits cytochrome c oxidase activity by competing with O₂, (2) NO inhibits electron transfer between cytochrome b and c (both belonging to Complex III) and increases mitochondrial production of O₂•⁻, and (3) NO inhibits electron transfer and NADH-dehydrogenase function at the level of Complex I (Grozdanovic, 2001; Sarti et al., 2012a).

The decrease in mitochondrial respiration has a crucial role in many pathological states, including conditions that are linked to skeletal muscle abnormalities. In fact, reduced mitochondrial energy production and decreased energy expenditure contribute largely to metabolic dysfunctions that are typical of aging, including insulin resistance and diabetes, which ultimately lead to lipid and glycogen accumulation, in turn further increasing insulin resistance (Derave et al., 2000; Finocchietto et al., 2008).

At low concentrations of O₂, NO binds to the haem of cytochrome c oxidase, thus acting as a competitive inhibitor of O₂. By contrast, at high O₂ concentrations, NO binds to oxidised cytochrome c oxidase through the copper moiety of the binuclear centre, instead of the iron moiety, thus producing nitrite (NO₂⁻) and consuming O₂ (Brown, 1995; Sarti et al., 2012b). At the early phase of NO production, the persistent inhibition of cytochrome c oxidase by NO can promote the release of small amounts of hydrogen peroxide, which itself acts as a signalling molecule and induces adaptive defensive responses. However, at later phases of NO generation, hydrogen peroxide is produced at higher concentrations, resulting in the formation of peroxynitrite that induces degeneration of skeletal muscle mass (Cleeter et al., 1994).

NO inhibits succinate-cytochrome c reductase and NADH-cytochrome c reductase in skeletal muscle (Cadenas, 2004; Poderoso et al., 1996). NO is also able to reversibly reduce cytochrome b, owing to its interaction with the iron–sulphur centre (Boveris et al., 2000). Even though the mechanisms are not completely clear, several studies have shown that prolonged exposure to NO can result in the inhibition of Complex I activity in skeletal muscle (Brown and Borutaite, 2004; Stamler and Meissner, 2001), which might be mediated by tyrosine nitration, S-nitrosylation and damage to Fe–S centres that displaces elemental sulphur or cysteine residues (Brown and Borutaite, 2004).

The importance of NO in regulating mitochondrial activity has been emphasised by the discovery of the presence of a mitochondrial NOS isoform (mtNOS) (Fig. 2v) (Elfering et al., 2002). mtNOS localises to mitochondria of several tissues, including skeletal muscle, where it regulates oxygen consumption (Aguirre et al., 2012; Finocchietto et al., 2008). At least in skeletal muscle, mtNOS appears to be a mitochondrial form of nNOS and can regulate mitochondrial respiration. In particular, it has been demonstrated that, under certain circumstances (i.e. upon hyperinsulinemia), mtNOS can be activated in situ by Akt, leading to decreased mitochondrial respiration in the muscle (Finocchietto et al., 2008). Hence, persistent mtNOS activation could contribute to mitochondrial dysfunction upon insulin resistance and, therefore, to the progression of the metabolic syndromes (Finocchietto et al., 2008).

Another aspect of NO-mediated regulation of skeletal muscle metabolism is at the level of glucose uptake. NO stimulates glucose transport by activating upstream signalling events that result in increased amounts of the glucose transporter GLUT4 at the cell surface (Etgen et al., 1997). GLUT4-mediated glucose uptake is mainly achieved through the activation of cGMP- and 5′-AMP-activated protein kinase (AMPK) (Lira et al., 2007). In addition, inhibition of NOSs limits the uptake of glucose in skeletal muscle, both under basal conditions and during physical activity, thereby impairing anaerobic skeletal muscle ATP production (Balon and Nadler, 1997; Higaki et al., 2001; Ross et al., 2007). Moreover, under resting conditions, NO might regulate carbohydrate metabolism by reversibly inhibiting the glycolytic enzyme GAPDH by mediating its S-nitrosylation (Hara et al., 2005; Hara et al., 2006). Finally, NO inhibits the breakdown of phosphocreatine, which is mediated by creatine kinase, thereby decreasing ATP synthesis through this pathway (Gross et al., 1996; Wolosker et al., 1996).

PGC-1α is a transcriptional coactivator, encoded by the PPARGC1A gene, that was first identified as an interacting partner of the peroxisome proliferator-activated receptor gamma (PPARγ) in adipocytes of brown adipose tissue (Puigserver et al., 1998). Later, PGC-1α was found to be expressed in other tissues that are rich in mitochondria and have high energy demands, such as cardiac and skeletal muscle, kidney, liver and brain (Austin and St-Pierre, 2012). In these tissues, PGC-1α controls the expression of genes that are tailored to the immediate energy demands of the organism and are involved in gluconeogenesis, lipogenesis and peroxisomal and mitochondrial fatty acid oxidation. Undoubtedly, PGC-1α plays a crucial role in maintaining muscle metabolic function and controls numerous genes that impact on muscle morphology and physiological function (Baar, 2004; Pilegaard et al., 2003).

PGC-1α represents the master regulator of mitochondrial biogenesis, as it is an upstream inducer of mitochondrial metabolism, acting to positively influence the activity of some hormone nuclear receptors (PPARs and ERRα) and nuclear transcription factors [nuclear respiratory factor 1 and 2 (NRF-1 and NRF-2)], which enhance the expression of OXPHOS components and mitochondrial transporters and transcription factors [mitochondrial transcription factor A (TFAM), TFBM1 and TFBM2]. For a long time, the coordination of the two genomes (nuclear versus mitochondrial) was considered to be exclusively achieved by the nucleus-encoded proteins TFAM, TFB1M and TFB2M; among these, TFAM was believed to be essential for transcription, replication and maintenance of mitochondrial DNA (mtDNA) (Lu et al., 2013). Importantly, we have demonstrated that PGC-1α also resides in mitochondria and, at the mitochondrial matrix, associates with TFAM and with mtDNA, thereby coactivating the transcription of mtDNA-encoded genes (Pagliei et al., 2013).

When expressed ectopically, PGC-1α induces mitochondrial biogenesis and increases cellular respiration in cell cultures. For instance, Wu and colleagues have shown that ectopic expression of PGC-1α in the skeletal muscle cell lineage induces a significant increase in mtDNA content and, in parallel, a rise in mitochondrial density (Wu et al., 1999). Here, the PGC-1α-mediated mitochondrial biogenesis occurs in parallel with an increase in the basal oxygen consumption and in the expression of OXPHOS genes (i.e. nucleus-encoded cytochrome c oxidase subunit IV and mitochondria-encoded cytochrome c oxidase subunit II) (Wu et al., 1999).

PGC-1α is upregulated by muscle contraction and is involved in most of the metabolic adaptations that occur in muscle (Geng et al., 2010; Lira et al., 2010a; Lira et al., 2010b). A bout of exercise increases the nuclear abundance of PGC-1α, which is responsible for the activation of the entire programme of mitochondrial and metabolic adaptations, including an enhanced capacity for substrate transport and oxidation (Fig. 3) (Little et al., 2011). Accordingly, overexpression of PGC-1α in skeletal muscle recapitulates many aspects of endurance training adaptation (Koves et al., 2005). Moreover, an acute endurance exercise in mice promotes the fast import of PGC-1α into mitochondria and increases the expression of TFAM-induced mitochondrial genes (Safdar et al., 2011). This mitochondrial event has been found to be associated with an increased activity of nuclear PGC-1α, thereby facilitating the concomitant transcription of nucleus-encoded mitochondrial genes (Fig. 3). In skeletal muscle, mitochondria exist as two subcellular populations (Smith et al., 2013) – subsarcolemmal and intermyofibrillar mitochondria – among which the former more efficiently respond to exercise training. Recently, it has been reported that, in humans, as well as in rats, acute exercise increases the levels of both PGC-1α and TFAM exclusively in subsarcolemmal mitochondria (Smith et al., 2013). One of the most important regulators of PGC-1α is AMPK, which serves as an energy sensor under conditions of low energy charge (i.e. increased AMP∶ATP ratio). Thus, conditions that cause significant cellular energy stress, such as exercise, are able to increase AMPK activity and promote mitochondrial biogenesis through AMPK-mediated PGC-1α phospho-activation (Reznick and Shulman, 2006; Steinberg and Kemp, 2009). The use of AMPK agonists or AMPK inhibition in rodents has made it possible to identify the signalling mechanisms that regulate the translocation of PGC-1α to mitochondria during exercise (Smith et al., 2013). In particular, the exposure of rats to the AMPK agonist 5-aminoimidazole-4-carboxamide-1-beta-ribofuranoside recapitulates the effect of exercise in increasing the levels of PGC-1α in only subsarcolemmal mitochondria, indicating that AMPK signalling is likely to be involved. In addition, the inactivation of AMPK prevents exercise-induced PGC-1α translocation to subsarcolemmal mitochondria, further suggesting that AMPK has a pivotal role in modulating PGC-1α translocation (Smith et al., 2013). Thus, taken together, these findings place PGC-1α at centre stage of mitochondrial–nuclear communications that coordinate the efficient expression of mitochondrial genes in order to meet the higher metabolic needs of exercised skeletal muscle (Fig. 3).

The regulation of proteins involved in fat metabolism is another process in which PGC-1α is involved in skeletal muscle. In fact, once overexpressed, PGC-1α enhances the transcription of mRNA encoding enzymes involved in fat oxidation, including carnitine palmitoyltransferase and medium-chain acyl-coenzyme A dehydrogenase (Calvo et al., 2008; Vega et al., 2000).

In response to exercise and other metabolic stresses (e.g. ischemia), PGC-1α controls an angiogenic pathway that is aimed at delivering oxygen and substrates to skeletal muscle for maintaining adequate energy production upon increased energetic needs (Arany et al., 2008). PGC-1α mediates the induction of vascular endothelial growth factor (VEGF), the primary mediator of the angiogenic process. Remarkably, the induction of VEGF by PGC-1α does not involve the canonical hypoxia response pathway or hypoxia inducible factor (HIF). Instead, PGC-1α coactivates the orphan nuclear receptor ERRα, which binds to conserved binding sites located in the promoter of the VEGF gene. Mice lacking PGC-1α show a dramatic energetic failure upon ischemic insult, whereas PGC-1α-overexpressing mice more rapidly reconstitute blood flow in skeletal muscle (Arany et al., 2008).

It is well established that NO donors and cGMP analogues increase mitochondrial biogenesis in muscle cells. Initially, NO was implicated in this process based on work that found a role for eNOS-derived NO in the differentiation of brown adipose tissue (BAT) (Nisoli et al., 2003). Several groups have shown that treatment of muscle cells with NO donors increases mitochondrial markers, demonstrating that NO induces the synthesis of new mitochondria that are able to generate ATP by oxidative phosphorylation (Lira et al., 2010b; McConell et al., 2010). It has been demonstrated that, in cultured skeletal muscle cells, NO donors increase mitochondrial biogenesis and function (Nisoli et al., 2004; Tengan et al., 2007). These effects of NO on skeletal muscle mitochondrial biogenesis appear to be mainly mediated by AMPK and Ca²⁺ (Lira et al., 2010b). However, the exact molecular pathways involved in mitochondrial biogenesis in response to NO production are not entirely understood, but in many tissues and cells, including skeletal muscle cells, they have been ascribed to PGC-1α (Lira et al., 2010b). As discussed above, stimulation of mitochondrial biogenesis by NO also requires the expression of transcription factors, including CREB (Ventura-Clapier et al., 2008). Although the serine/threonine kinase Akt and protein kinase A (PKA) are able to phosphorylate and activate eNOS, thus leading to an increase in NO production, phosphorylation of CREB by PKA results in expression of its target gene PPARGC1A and thus in mitochondrial biogenesis. Furthermore, we have shown that CREB binding to the PPARGC1A promoter is greatly increased when CREB is also S-nitrosylated (Aquilano et al., 2014). In addition, Akt might also induce mitochondrial biogenesis through the phosphorylation of NRF-1 and CREB, thus enabling their nuclear translocation and activation of target genes such as TFAM (Piantadosi and Suliman, 2012).

AMPK is also involved in NO-dependent regulation of PGC-1α and mitochondrial biogenesis. In particular, treatment of myotubes with NO donors results in the inhibition of mitochondrial ATP production and an increase in the AMP∶ATP ratio (Lira et al., 2010b). This event leads to AMPK phospho-activation, which is responsible for PGC-1α phosphorylation and PGC-1α-mediated induction of the expression of mitochondrial genes (Lira et al., 2010b) (Fig. 3). Moreover, Ca²⁺-mediated signalling pathways are also able to induce mitochondrial biogenesis by activating NO synthesis or stimulating Ca²⁺-dependent transcription factors. In fact, exercise in humans induces phosphorylation of nNOSμ by AMPK, leading to nNOSμ activation, increased NO production and mitochondrial biogenesis, and increased glucose uptake (Chen et al., 2003). AMPK is also able to phosphorylate and activate eNOS during contraction of human skeletal muscle (Chen et al., 1999). Taken together, these findings strongly indicate that phosphorylation of both nNOSμ and eNOS by AMPK might be involved in mitochondrial biogenesis. Accordingly, treatment with a NO donor could increase the activation of AMPK and mitochondrial biogenesis, whereas the pharmacological inhibition of NOS attenuates these effects (McConell et al., 2010).

However, it is worth mentioning that upregulation of PGC-1α expression does not always require NO signalling. For example, Wadley et al. (Wadley et al., 2007) have reported that normal skeletal muscle of nNOS- and eNOS-knockout mice harbours baseline levels of PGC-1α expression. Furthermore, normal upregulation of PGC-1α was elicited in these mice in response to acute exercise, indicating that NOS-mediated NO signalling is part of a redundant system of metabolic regulation in skeletal muscle (Wadley et al., 2007).

The production of reactive oxygen species (ROS) and NO-derived reactive species (RNS) is an integral part of skeletal muscle metabolism (Westerblad and Allen, 2011). At low concentrations, ROS and RNS might function as signalling molecules; thus NO can act as an antioxidant by directly scavenging more-damaging species, such as hydroxyl radicals (Aquilano et al., 2007a). Physiologically, the cell counteracts an excess of ROS or RNS by enhancing its antioxidant defence system, which includes superoxide dismutases and the enzymes involved in GSH metabolism (e.g. GSH peroxidase, γ-glutamyl cysteine ligase). An imbalance between ROS or RNS levels and antioxidant defence leads to oxidative stress and might result in cell death (Aquilano et al., 2007b). However, a controlled mild flux of ROS or RNS is a central means of inducing redox-sensitive signalling pathways that act to fine-tune the metabolic adaptive response by enhancing energy request during muscle contraction and regulating the cellular defence against the damaging effects of oxidative stress. Indeed, upregulation of endogenous antioxidant defence systems and complex regulation of repair systems such as those involving heat shock proteins (HSP70, HSP27, HO-1) are seen in response to training and exercise (Fehrenbach and Northoff, 2001). For this reason, it has become evident that physical exercise can ameliorate not only skeletal muscle function but can also have beneficial effects on systemic physiology, subsequently leading to improved health and increased lifespan (Mercken et al., 2012).

Nutrient starvation or caloric restriction is broadly applicable and extends the life of most species through the retardation of aging, thereby exerting beneficial effects on virtually all cells and tissues (Lettieri Barbato et al., 2012). It is well established that caloric restriction improves skeletal muscle function to levels that are comparable to that of physical exercise, thus favouring mitochondrial biogenesis and oxidative metabolism and restraining insulin resistance (Mercken et al., 2012). The beneficial effect of caloric restriction on lifespan has been linked to the increase in NO production. In particular, caloric restriction increases both eNOS and nNOS activity (Nisoli et al., 2005; Fusco et al., 2012). In our laboratory, we have shown that, in skeletal muscle, caloric restriction also increases NO bioavailability independently of NOS upregulation (Aquilano et al., 2013a). In particular, caloric restriction reduces the level of GSH, the main intracellular NO buffer (Aquilano et al., 2011a; Aquilano et al., 2011b; Baldelli et al., 2008). The resulting increased NO bioavailability is the genuine mediator of the upregulation of PGC-1α, which, in turn, favours the expression of antioxidant proteins SOD2 and γ-GCS (also known as glutamate cysteine ligase) (Aquilano et al., 2013a).

It has become increasingly clear that NO is a crucial player in skeletal muscle physiology, as it regulates a variety of highly relevant pathways to maintain both skeletal muscle integrity and proper signalling mechanisms during adaptation to mechanical and metabolic stimulation (i.e. exercise and caloric restriction). The expression and localisation of a specific nNOS isoform is essential for modulating the activity of the electron transport chain, mitochondrial biogenesis, glucose uptake and utilisation, and fatty acid oxidation. Together with these effects, and more strictly related to energetic metabolism, nNOS in skeletal muscle can also regulate the delivery of local oxygen and nutrients by inducing vasodilatation as well as angiogenesis (Huber-Abel et al., 2012). The metabolic regulator PGC-1α is at the crossroads of nNOS-mediated signalling and associated adaptation processes, as NO directly induces PGC-1α expression, thus augmenting mitochondrial biogenesis (Aquilano et al., 2014), while, at the same time, favouring VEGF-mediated vascularisation (Arany et al., 2008; Rowe et al., 2011). With the finding that exercise and caloric restriction exert particularly beneficial outcomes in the activation of the nNOS enzyme and PGC-1α, it will be interesting in future studies to focus on the underlying mechanisms that might determine the adaptive response of skeletal muscle to physical exercise under pathological conditions.

M.R.C. and K.A. contributed equally to this work.