Numerous reports show that IGF-I is a requisite for muscle growth and regeneration (Florini et al., 1991; Vandenburgh et al., 1991; Coleman et al., 1995; Musaro et al., 2001; Rabinovsky et al., 2003; Pelosi et al., 2007), as well as a well-known upstream stimulator of mTOR in skeletal muscle. IGF-I binds to IGF-I receptor (IGFR), a receptor tyrosine kinase, and subsequently recruits insulin receptor substrate-1 (IRS-1). Although IRS-1 activates the Ras-Raf-MEK-ERK pathway, the role of this pathway in skeletal muscle is not clear (Rommel et al., 1999). Instead, Akt /mTOR signaling by IGF-I/IGFR/IRS-1 has been shown to be indispensable in prompting muscle hypertrophy (Glass, 2003). Akt phosphorylates TSC1/2, which inhibits the GTPase-activating protein (GAP) activity of TSC1/2 toward small G protein Rheb. Then, GTP-bound Rheb activates mTORC1, resulting in phosphorylation of S6K1and 4EBP1, which promote protein synthesis by activating ribosomal protein S6 and by releasing the translation initiation factor eIF-4E, respectively. In line with IGF-I-Akt-mTORC1 regulation, IGF-I induces hypertrophy of skeletal myofiber in tissue culture (Vandenburgh et al., 1991). Muscle-specific expression of IGF-I in transgenic mice results in at least a 2-fold increase in muscle hypertrophy (Coleman et al., 1995; Musaro et al., 2001), suggesting that the IGF-I/Akt/mTORC1 pathway is indispensable to muscle hypertrophy. In addition, Akt regulates muscle mass by phosphorylating and deactivating glycogen synthase kinase (GSK) β1, followed by the GSK β1-dependent inhibition of the eukaryotic translation initiation factor 2B (eIF2B) (Manning and Cantley, 2007; Schiaffino and Mammucari, 2011).

However, a recent report showed that IGF-I and its receptor IGFR were not important to the induction of hypertrophy and the activation of Akt/mTOR in mechanical loading (Spangenburg et al., 2008). The expression of dominant negative (DN)-IGF-I receptor specifically in skeletal muscle induced muscle hypertrophy using an increased functional overload model induced by synergistic ablation (Spangenburg et al., 2008). Of interest, DN-IGF-I receptor-expressing muscle showed a similar level of activation of Akt and p70S6K1. These results implied that an unknown upstream mediator beyond IGFR might regulate Akt/mTOR signaling in skeletal muscle hypertrophy.

One of the potential IGFR-independent mTOR regulators in skeletal muscle is phosphatidic acid (PA). Hornberger et al. observed that IGF-I-independent mechanical stretch increases phosphatidic acids (PA), followed by mTOR activation (Hornberger et al., 2006). PA directly binds to the FKBP12-rapamycin binding (FRB) domain in competition with rapamycin, and activates mTOR (Fang et al., 2001). PA is synthesized through several pathways: from phosphatidylcholine (PC) by phospholipase D (PLD), from lysophosphatidic acid (LPA) by lysophosphatidic acid acyltransferases (LPAAT), and from diacylglycerol (DAG) by diacylglycerol kinase (DGK) (Wang et al., 2006; Yoon et al., 2015). Among the several enzymes involved in PA biogenesis, PLD activity was increased by mechanical stretch and followed by mTOR activation (Hornberger et al., 2006). In addition, treatment with 1-butanol, a PLD inhibitor, inhibited the increase in mTOR activity, supporting the role of PLD in mechanical stretch (Hornberger et al., 2006). However, PA level continued to remain high after the elevated PLD activity returned to basal level 15 min after mechanical stretch using [3H] arachidonic acid labeling, suggesting that other enzymes produce PA under mechanical stretch. Hornberger et al. found that DGK ζ produces PA under mechanical stimulation, which is followed by mTOR activation (You et al., 2014). Mechanical stimulation does not induce PLD activity under [3H] myristic acid labeling that preferentially labels PC, and FIPI, a PLD inhibitor, did not inhibit mechanical stimulation. Instead, both DAG and membrane DGK activity, which are critical for mTOR activation, were increased during mechanical stimulation. However, the previous report suggested that PLD1-produced PAs preferentially bind to the FRB domain of mTOR and displace DEPTOR, leading to mTORC1 activation (Yoon et al., 2015). Hence, further investigation into whether PA produced by DGK ζ during muscle stretch binds to the FRB domain or activates mTOR through an FRB-independent mechanism is warranted.

It has been established that mTORC1 translocates to the lysosome through regulation of Ragulator-Rag in amino acid signaling (Sancak et al., 2008, 2010). Lysosomal localization of mTOR does not activate mTOR directly, but rather provides close proximity to Rheb, an essential activator of mTOR (Saxton and Sabatini, 2017). mTORC1 is activated by direct interaction with the GTP-bound form of Rheb (Sancak et al., 2008, 2010), which is regulated by the TSC complex [TSC1, TSC2, and Tre2-Bub2-Cdc16-1 domain family member 7 (TBC1D7) Dibble and Manning, 2013], a GAP of Rheb (Huang and Manning, 2008; Saxton and Sabatini, 2017). The location of Rheb, shown to be on the lysosome, was not changed by either amino acids or insulin (Menon et al., 2014). Nevertheless, the TSC complex activates the intrinsic GTPase activity of Rheb on the surface of the lysosome and localizes to the lysosome, at least partially through its association with Rheb-GDP in the absence of growth factors (Menon et al., 2014). Insulin activates Akt, which subsequently phosphorylates the TSC complex, resulting in the dissociation of the TSC complex from Rheb, followed by Rheb GTP loading and mTORC1 activation (Menon et al., 2014).

The Hornberger group found that mTOR and TSC2 were highly enriched in the lysosome of the muscle in the resting state (Jacobs et al., 2013). Mechanical stimulation induces TSC2 phosphorylation at RxRxxS^*/T^*, which resulted in the dissociation of TSC2 from the lysosome and the subsequent change of Rheb to the active Rheb-GTP state. Furthermore, mechanical stimulation also facilitates the association of mTOR with the lysosome. Accordingly, mTOR potentiates the activation of the lysosome through the interactions with Rheb-GTP or PA, as previously reported (Sancak et al., 2008, 2010; Yoon et al., 2011). However, the kinase for TSC phosphorylation was unclear in this study since mechanical stimulation was previously shown to activate mTOR in PI3K/Akt-independent manner (Hornberger et al., 2004; O'Neil et al., 2009). In addition, Song et al. recently suggested the colocalization of mTOR with eukaryotic translation initiation factor 3 subunit F (eIF3F) in resistance exercise (Song et al., 2017). mTOR is localized on the lysosome in the basal state and the mTOR-LAMP2 complex is translocated to the cell periphery under resistance exercise, which provide close proximity to the capillaries. In support of this, the lysosome is shown to migrate to the cell periphery after nutrient stimulation through two kinetin proteins, K1F1Bβ and KIF2, which are essential to mTORC1 activation (Korolchuk et al., 2011). Concurrently, TSC2 dissociates from Rheb, followed by the reduction of TSC2 on the cell periphery and the subsequent increase of mTORC1 activity (Song et al., 2017). In addition, both the association of mTOR with eIF3F and S6K1 activity are increased in fed conditions after exercise, which provides an explanation for the enhanced muscle protein synthesis (Song et al., 2017). Nevertheless, the association of mTOR with eIF3F was mainly determined by using an immunofluorescent approach in Song's study, which requires further investigation by using a range of techniques.

Myostatin, a transforming growth factor-β (TGF-β) family member, plays a critical role in inhibiting the growth of muscle mass and muscle cell differentiation (McPherron et al., 1997). The deletion of myostatin in mice results in muscle hyperplasia and hypertrophy, and more than doubles skeletal muscle (McPherron et al., 1997). Myostatin regulates the number of muscle fibers during development and the growth of muscle fibers postnatally (Lee, 2007). The binding of myostatin to the type II activin receptor IIb leads to interaction with the type I receptor ALK4 or ALK5, which results in the phosphorylation and activation of the transcription factors Smad2 and Smad3(Sartori et al., 2014). Additionally, myostatin decreases Akt phosphorylation, which is accompanied by the accumulation of dephosphorylated active Forkhead Box-O1 (FOXO1) and FOXO3, followed by upregulation of components of the ubiquitin-proteasome pathway, such as atrogin-1 and the muscle-specific E3 ubiquitin ligase muscle RING-finger1 (MURF1) (McFarlane et al., 2006; Lokireddy et al., 2011). In addition, myostatin blocks differentiation-inducing genes, such as myogenin and myoD (Trendelenburg et al., 2009), suggesting that myostatin regulates muscle differentiation by modulating both the programs of differentiation and atrophy.

mTOR regulation by myostatin has sophisticated the molecular mechanism of myostatin signaling. The overexpression of myostatin decreases Akt and mTORC1 components, such as p70S6K1, S6, and 4EBP1 (Amirouche et al., 2009). Supporting the negative regulation of myostatin in mTORC1 signaling, genetic deletion of myostatin elevates the activities and the expression levels of Akt, p70S6K1, and S6 (Lipina et al., 2010). Additionally, treatment with myostatin reduces myoblast differentiation and myotube size by inhibiting the activity of Akt/mTORC1/p70S6K1 in human skeletal muscle cells (HuSkMC) (Trendelenburg et al., 2009). The depletion of raptor increases myostatin-induced Smad2 phosphorylation, followed by further inhibition of myostatin-induced muscle differentiation. The knockdown of rictor itself inhibits muscle cell differentiation, and does not affect myostatin-induced pSmad2 and muscle differentiation. These results suggested that both mTORC1 and myostatin-Smad2 signaling negatively regulate each other. The overexpression of ActRIIB induces inhibition of myostatin, resulting in skeletal muscle hypertrophy, which is reduced partially by treatment with rapamycin (Sartori et al., 2009). Hence, the studies suggested that myostatin attenuates protein synthesis in muscle by coordinating the crosstalk between myostatin-mediated and mTOR signaling. However, on the other hand, several studies suggested that mTOR signaling and myostatin signaling could separately regulate muscle growth. The injection of a myostatin antibody enhances phosphorylation of p70S6K1 and S6 in muscle, but does not change phosphorylation of Akt and 4EBP1 in the concomitant increase of myofibrillar synthesis (Welle et al., 2009). In this study, treatment with rapamycin does not affect myofibrillar synthesis, while it decreases the phosphorylation of p70S6K1 and S6, implying that mTOR is not involved in myostatin-mediated myofibrillar synthesis (Welle et al., 2009). In addition, follistatin, an inhibitor of myostatin, activates Akt/mTOR/p70S6K1/S6 signaling in muscle growth, which exists independently of myostatin-driven mechanisms (Winbanks et al., 2012), supporting the disconnection between myostatin and mTOR signaling. Hence, myostatin may regulate protein synthesis in both an mTOR-dependent and an mTOR-independent manner; it controls the translation through Akt/mTORC1/p70S6K1/S6 signaling and, at the same time, it directly acts on unknown regulators of translation.

Glucocorticoids are some of the most fundamental regulators of energy homeostasis and adjust the metabolism of carbohydrates, fat, and protein in skeletal muscle (Munck et al., 1984). Glucocorticoids bind to the glucocorticoid receptor (GR), which translocates to the nucleus and binds to the glucocorticoid response element (GRE) in the promoters of target genes (Meijsing et al., 2009). Notably, the circulating levels of glucocorticoids are increased under many pathological conditions which are accompanied by muscle atrophy such as cachexia, starvation, sepsis, metabolic acidosis, and severe insulinopenia (Braun and Marks, 2015). Exogenous administration of glucocorticoids induces muscle atrophy and the blockage of GR; adrenalectomy or treatment with the GR antagonist RU486 diminishes muscle atrophy in sepsis, cachexia, starvation, and severe insulinopenia (Menconi et al., 2007; Schakman et al., 2008). Hence, endogenous glucocorticoids are critical regulators in muscle atrophy.

The crosstalk between GR and mTOR has been reported in muscle cells (Shimizu et al., 2011). Tanaka et al. found that REDD1 and KLF15 inhibit mTOR activation as direct targets of GR. REDD1 has a functional GRE, which sequestrates 14-3-3 from TSC1/2, resulting in activation of mTOR (DeYoung et al., 2008). KlF15 plays a critical role in muscle catabolism through the transcriptional upregulation of atrogen-1, MuRF-1, and branched-chain aminotransferase 2 (BCAT2). BCAT2 catalyzes the first reaction of BCAA catabolism, facilitating BCAA degradation, followed by mTOR inactivation and decrease of myofiber size (Shimizu et al., 2011). On the other hand, mTOR negatively modulates GR-mediated transcription by inhibiting GR recruitment of target genes. Hence, both GR and mTOR control each other exclusively in the regulation of muscle mass.

Glucocorticoids also elicit muscle atrophy via controlling transcription of myostatin, an inhibitory regulator of muscle growth, which we discussed in the previous section. Human myostatin promoter is reported to have a putative GRE and is responsive to dexamethasone and RU-486, an antagonist of GR (Ma et al., 2001). Indeed, the expression of myostatin mRNA and protein are increased in a dose-dependent manner in dexamethasone-treated rats (Ma et al., 2003). Myostatin production is also induced by food deprivation in a glucocorticoid-dependent manner (Allen et al., 2010). A recent report also suggests that glucocorticoids increase phosphorylation of CEBP by decreasing PDE3/4 and activating PKA through inhibiting Akt activity, resulting in increased myostatin expression (Xie et al., 2017). Hence, glucocorticoids may regulate mTOR by modulating the level of both BCAT2 and myostatin to regulate catabolism in skeletal muscle.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.