3.1.1. The CBL-B and FBXO40 E3 Ubiquitin Ligases Target IRS1 to Degradation in Skeletal Muscle

One strategy to fight against atrophy may be to stimulate the anabolic pathways leading to skeletal muscle hypertrophy. Insulin-like growth factor 1 (IGF1) induces skeletal muscle hypertrophy by activating the IGF1R/PI3K/AKT pathway, a critical mediator and checkpoint being IRS1. Indeed, the effect of IGF1 is time-limited by the phosphorylation of IRS1 by IGF1R and its subsequent ubiquitination and proteasome-mediated degradation.

Different E3 ligases can target IRS1 in different tissues. For example, in embryonic fibroblasts, the CUL7 E3 ligase, containing FBXW8, has been shown to target IRS1 for ubiquitin-dependent degradation [170]. In skeletal muscle, Casitas B-lineage lymphoma-b (CBL-B), a RING E3 ligase, targets IRS1 for degradation and thus impairs muscular trophic signals in response to unloading conditions [171,172,173], which inhibits downstream IGF1 signaling [173] (Figure 2 and Table 1). Accordingly, mice deficient for CBL-B were partly resistant to unloading-induced skeletal muscle atrophy and dysfunction [173]. These results highlight the importance of CBL-B in the process of muscle atrophy in response to unloading.

DN, Dominant Negative mutation; HM, Hypomorphic Mutation; I/R, Ischemia/Reperfusion; KD, knock-down mutant; KO, Knock-out mutant; m.KO, skeletal muscle–specific KO mice; OX, overexpressing mutant; PTEN, Phosphatase and tensin homologue.

FBXO40 is a muscle-specific F-box protein [174], component of an SCF (Skp1-Cullin1-F-box protein) E3 ligase complex. Following IRS1 activation, IGF1R phosphorylates IRS1 leading to its ubiquitination by FBXO40 and its degradation by the 26S proteasome, in cultured myotube and in mice [22,175]. FBXO40 expression is decreased in muscles from Limb-girdle muscular dystrophy (LGMD) patients, and up-regulated in mice skeletal muscle following denervation and in chronic kidney disease (CKD) mice model, but not during starvation [174,175]. Accordingly, the knock-down of Fbxo40 resulted in thicker myotubes (20% to 50% increase in diameter) [22] and its deletion in mice also induced muscle hypertrophy during the growth phase, a phase associated with high IGF1 levels [22] (Figure 2 and Table 1).

IRS1 is thus an important checkpoint of the IGF1/PI3K/AKT pathway controlled by at least 2 E3 ligases (CBL-B and FBXO40). Although being an attractive target for fighting against muscle atrophy, the multiple ways for degrading IRS1 may complicate the development of drugs.

3.1.2. NEDD4-1 E3 Ubiquitin Ligase, Friend or Foe?

In muscles undergoing atrophy, NEDD4-1 mRNA levels are elevated upon severe sepsis [191], denervation or unloading [178,192,193]. On the one hand, NEDD4-1 E3 Ub ligase targets phosphatase and tensin homologue (PTEN). PTEN is a redox sensitive phosphatase that negatively regulates the PI3K-AKT signaling pathway, thereby affecting metabolic and cell survival processes. The deletion of PTEN improves muscle mass and function in a mouse model of Duchenne muscular dystrophy [194]. PTEN inhibition may thus also represent a potential therapeutic strategy to maintain muscle function during catabolic situations. The over-expression of NEDD4-1 is sufficient for activating the PI3K/AKT signaling in cardiac muscle, following myocardial ischemia/reperfusion (I/R) [176]. However, the negative regulation of PTEN by NEDD4-1 remains to be confirmed in skeletal muscle, especially since NEDD4-1 has also been shown to promote skeletal muscle atrophy in a denervation model. Indeed, NEDD4-1-KO mice exhibited increased weights and type II muscle fiber cross-sectional areas in denervated gastrocnemius muscle [178]. Moreover, NEDD4-1 also negatively regulates the hypertrophic BMP signaling (Figure 1 and Figure 2). Indeed, NEDD4-1 ubiquitinates phosphorylated-SMAD1 leading to its proteasomal degradation, thereby silencing BMP signaling in C2C12 myoblasts, and conversely the knock-down of Nedd4-1 potentiates BMP signal through upregulation of phospho-SMAD1 [195]. Altogether, the exact function of NEDD4-1 in skeletal muscle is still obscure and needs more work.

3.2.1. Regulating the Canonical NF-κB Pathway via the Manipulation of cIAP and TRAF6 E3 Ligases

Among the E3s involved in the regulation of the NF-κB pathway, two promising candidates may be manipulated to limit muscle atrophy, namely cIAP and TRAF6 (Figure 1 and Figure 2). cIAP1 is up-regulated in denervated gastrocnemius muscle, paralleling the upregulation of MAFbx/atrogin-1 and MuRF1/Trim63 mRNA [19]. Mice with genetic ablation of cIAP1 (cIAP1-KO mice) displayed limited denervation-induced atrophy in TA, gastrocnemius and EDL muscles. This was correlated with the blunting of the denervation-induced upregulation of MAFbx/Atrogin-1 and MuRF1/Trim63 [19]. The authors further demonstrated that cIAP1 induced atrophy through the up-regulation of the canonical NF-κB signaling. Conversely, cIAP1 overexpression in myotubes induced atrophy and the strong up-regulation of MAFbx/Atrogin-1 and MuRF1/Trim63 protein expression [19]. The E3 Ub ligase cIAP1 represents thus a potential therapeutic target at least for fighting against denervation-induced muscle atrophy.

TRAF6 is a RING-type Ub ligase that plays an important role during skeletal muscle atrophy. TRAF6 expression is enhanced during starvation or within aged-induced muscle atrophy [179,196,197]. Traf6-KO mice are resistant to skeletal muscle loss (rescue of myofibril degradation, preservation of myofiber size and strength) induced by denervation, cancer cachexia, starvation or Dex and a concomitant suppression of the expression of key regulators of muscle atrophy was observed, including MAFBx/Atrogin-1, MuRF1/TRIM63, p62, Lc3b, Beclin1, Atg12, and Fn14 [179,180,196,197,198]. Moreover, inhibition of Traf6 expression through miR-351 administration in C2C12 myotubes or in denervated mice attenuated Dex-induced muscle atrophy and concomitantly decreased the expression of MAFBx/Atrogin-1 and MuRF1/Trim63 [199,200]. Overexpression of miR-125b targeted Traf6 for degradation and protected skeletal muscle samples from atrophy in starved myotubes or in denervated rat tibialis muscle [201]. The implicated mechanisms involved both direct and indirect effects of TRAF6 on protein breakdown with TRAF6-mediated ubiquitination being required for the optimal activation of JNK, AMPK, FOXO3, and NF-κB catabolic pathway in muscle [202].

In human, gastric cancer patients suffering from cachexia exhibited an upregulation of TRAF6 associated with an upregulation of ubiquitination in the rectus abdominis muscle [203]. Altogether, this highlights the importance for targeting TRAF6 inhibition to counteract muscle atrophy.

3.2.2. WWP1 in the Regulation of Muscle Atrophy

WWP1 is a HECT E3 ligase that is involved in chicken muscular dystrophy. Indeed, a missense mutation in the gene coding WWP1 was identified as the most promising candidate responsible for chicken muscular dystrophy (MD), potentially affecting the E3 function of WWP1 protein [204]. WWP1 was also shown to target the transcription factor KLF15 [181]. In response to glucocorticoids, KLF15 is up-regulated at the mRNA levels [205]. This induction leads to the up-regulation of the E3 ligases MAFbx/Atrogin-1 and MuRF1/Trim63 expression, likely in cooperation with a FOXO transcription factor, while inhibiting the anabolic mTORC1 [205]. Likewise, exogenous KLF15 expression in myotubes and in TA muscle leads to myofiber atrophy [205]. It has recently been shown that KLF15 protein expression was upregulated in skeletal muscle of diabetic mice, without any change in its mRNA expression [181]. This increase correlated with an increase in MAFbx/Atrogin-1, Murf1/Trim63 and Foxo3 genes expression and accordingly, the muscle-specific deletion of Klf15 in this model prevented from diabetes-induced muscle atrophy [181]. The authors identified WWP1 as an E3 ligase targeting KLF15 and showed that knocking-down WWP1 in both C2C12 myotubes and in tibialis anterior muscles increased MuRF1/Trim63 and MAFbx/Atrogin-1 expression and induced atrophy [181] (Figure 2). WWP1 E3 ligase is indeed induced by high glucose conditions in myotubes [206]. Conversely, in high glucose conditions, WWP1 has also been implicated in the down-regulation of AMPKα2 protein levels [206]. The authors have shown that WWP1 interacted with AMPKα2 leading to a proteasome-dependent decrease of AMPKα2 in myotubes; however, direct ubiquitination was not addressed [206]. WWP1 may thus control muscle mass through a direct action on AMPK, a known modulator of FOXO3a, MuRF1/TRIM63 and MAFbx/Atrogin-1 [88].

3.2.3. TRIM32 in the Regulation of Autophagy

TRIM32 is a RING E3 Ub ligase whose mutation is responsible for the development of limb girdle dystrophy 2H (LGMD2H) [207]. Several substrates have been identified for TRIM32 in non-muscle cells, including cell cycle regulators (c-Myc, MYCN, p53), the cell growth and transformation factor ABI2 and PIASY (a SUMO E3 ligase). TRIM32 is also involved in the targeting of factors influencing myogenesis (NDRG2 and TRIM72) that regulate muscle satellite cells renewal and differentiation [208]. While initially postulated to promote muscle atrophy, TRIM32 is in fact a master regulator of myogenesis during recovery situations [208]. Indeed, the dystrophic phenotype of TRIM32 mutations appeared to be largely due to impaired myogenesis [208,209,210].

More recently, TRIM32 was implicated in the early events leading to autophagy. Indeed, TRIM32 targets ULK1, a Ser/Thr protein kinase (Figure 1 and Figure 2). ULK1 is an upstream regulator of autophagy rapidly activated to ensure a rapid response to stress conditions [211]. The authors showed that TRIM32 deficiency was directly responsible for autophagy defects both in cultured cells and in mice treated with Dex. The mechanisms by which TRIM32 controls the activation of autophagy through ULK1 involves its binding to AMBRA1, a positive regulator of autophagy [211]. AMBRA1 is a pivotal factor able to bind several E3 ligases during the course of the autophagy process. In presence of AMBRA1, TRIM32 binds to ULK1, synthesizes unanchored K63 Ub chains that activate ULK1 kinase activity, thus promoting autophagy. The role of TRIM32 during the autophagy process is not limited to ULK1 as p62, an important autophagy receptor [212], is also a TRIM32 substrate. p62 activity is modulated by multi mono-Ub catalyzed by TRIM32 and loss of function of TRIM32 largely abolished autophagy [213]. Altogether, TRIM32 appears as a master regulator of muscle renewal through the initiation of autophagy.

3.2.4. FOXO Transcription Factors Are Regulated by MDM2 and SKP2 E3 Ubiquitin Ligases

Alternatively to phosphorylation, FOXO can be regulated by acetylation/deacetylation, methylation and ubiquitination to modulate its activity, localization as well as degradation [214,215,216].

Ubiquitination modulate FOXO activity by either mono- or polyubiquitination through MDM2 and SKP2 E3 Ub ligases (Figure 1 and Figure 2). MDM2 is the enzyme responsible of a single addition of an ubiquitin moiety to FOXOs, specifically to FOXO4, thus allowing its nuclear localization and transcriptional activation [217,218]. Mono-Ub of FOXO4 is observed under oxidative stress conditions and can be counteracted by deubiquitinating enzymes such as ubiquitin-specific protease (USP7). Importantly, ubiquitination mediated by MDM2 is context specific and upon growth factor stimulation can induce FOXO1 and 3 degradation [217]. In addition, interaction between FOXOs and SKP2, a subunit of the SKP/cullin 1/F-box protein E3 ligase leads to proteasomal degradation of FOXO1 in the cytosol [218].

Combined with the other posttranslational modifications, ubiquitination allows FOXOs to integrate information arising from insulin, growth factors, cytokines, and oxidative stress and to control downstream signaling. Interestingly, FOXO TFs have systematically been envisioned as crucial drivers of catabolic pathways during muscle wasting. Nonetheless, recent work showed that FOXO1 and 3a participate to skeletal muscle adaptation upon exercise thus adding a new of FOXOs in the control of muscle cell homeostasis [219,220,221,222].

3.3.1. MuRF1/TRIM63

Muscle-specific RING finger protein 1 (MuRF1), also named TRIM63, is a RING-type E3 ligase and a founding member of the so-called “atrogenes” (see [6] for a recent review). MuRF1/TRIM63 is a master regulator of skeletal muscle atrophy development occurring in numerous catabolic conditions and MuRF1/Trim63 mRNA appeared to be upregulated in more than 25 atrophying situations [6] (Figure 1 and Figure 2). Mice deleted for MuRF1/TRIM63 (MuRF1-KO mice) were partially resistant (preservation of muscle mass and structure) to skeletal muscle atrophy induced by denervation [4], hindlimb suspension [4,223], glucocorticoid [224], amino acid deprivation [225], and acute lung injury [226]. MuRF1/TRIM63 is responsible for the coordinated breakdown of both thick and thin filaments occurring during catabolic states in skeletal muscle, targeting to degradation the main proteins of the contractile apparatus: myosin heavy chains (MHC) [227], alpha-actin [228], troponin I [229], TCAP/telethonin [230]. During denervation and starvation, MuRF1/TRIM63 has also been involved in the degradation of acetylcholine receptor (CHRN), the major postsynaptic ion channel of the neuromuscular junction. This degradation is mediated by the activation of selective autophagy and degradation of CHRN, likely via the degradation of BIF-1 (Bax interacting factor 1)/EndoB1 (EndophilinB1) and/or SQTM1/p62 (sequestosome-1) [231,232].

While numerous studies have promoted a major role of MuRF1/TRIM63 in the development of skeletal muscle atrophy during catabolic states, in the heart, the analyses of MuRF1 mutants have highlighted a beneficial cardioprotective role [233]. These opposites roles in both kind of muscles imply the development of skeletal muscle-specific drugs to inhibit MuRF1/TRIM63. Moreover, one should also take into account that MuRF1/TRIM63 has two homologs, MuRF2 and MuRF3 that share some redundant functions and could replace its role [12].

3.3.2. MAFbx/Atrogin-1/FBXO32

The multimeric E3 ligase MAFbx/atrogin-1/FBXO32 is another founding member of the atrogene family ([6] for a recent review) crucial for the development of muscle atrophy. Interestingly, nearly all catabolic situations induce an overexpression of both MAFbx/Atrogin-1 and MuRF1/TRIM63, which are controlled by the same TFs (FOXO1/FOXO3a, NF-κB, C/EBP β, Smad 3, etc.) and the same signaling pathways [234] (Figure 1 and Figure 2).

In contrast with MuRF1/TRIM63 that targets directly the contractile proteins for their degradation (α-actin, MHC, etc. [227,228,229,230], MAFbx appeared to target pro-anabolic factors like MyoD, myogenin or eIF3f [235,236,237]. MyoD is a muscle-specific transcription factor that plays crucial roles during cell cycle and muscle differentiation [238]. The eukaryotic initiation factor 3 subunit f (eIF3f) is a pivotal element of protein synthesis and its control by MAFbx allows the latter to master the anabolic processes [235]. While a putative role of MAFbx/Atrogin-1 on sarcomeric proteins was hypothesized using an indirect approach, this has never been confirmed [239]. By contrast, the authors found that desmin, a main component of the intermediate filaments, physically interacted with MAFbx and was degraded in myostatin-treated cultured C2C12 myotubes.

As MAFbx/Atrogin-1 and MuRF1/TRIM63 are controlled by similar signaling pathways, the strategies for the upstream control of MuRF1/Trim63 expression are generally also valid for MAFbx/Atrogin-1 (Table 2). By contrast with MuRF1/TRIM63, no direct inhibitor of MAFbx/Atrogin-1 has been described so far but general strategies, like targeting the interface responsible for substrate recognition or impeding the assembly of the F-box (i.e., the subunit recognizing the substrates) into the SCF complex, may prove to be efficient.

Altogether, controlling concomitantly MAFbx/Atrogin-1 and MuRF1/TRIM63 E3 ligases allows skeletal muscle cells to both increase the degradation of the contractile apparatus and to depress the protein synthesis machinery, which allows a tight regulation of protein homeostasis.

3.3.3. PARKIN Controls Muscle Mass through the Maintenance of Mitochondrial Homeostasis

PARKIN is an E3 ubiquitin ligase implicated in the regulation of mitophagy, a quality control process in which defective mitochondria are degraded. Mitochondrial quality control through both mitochondria turnover and dynamic plays an essential role in the maintenance of muscle mass (see [240] for a review). During mitophagy, PARKIN ubiquitinates several outer mitochondrial membrane proteins leading to subsequent autophagosomal engulfment and lysosomal degradation (Figure 1 and Figure 2).

This role of PARKIN has been emphasized in rodent models or in humans where a deregulation of PARKIN mRNA and/or protein expression prevailed in response to catabolic or anabolic situations. An accumulation of PARKIN protein prevailed during: (i) muscle wasting situations such as chronic kidney disease [241], chronic obstructive pulmonary disease (COPD) [242], physical inactivity [243,244] and (ii) upon exercise training [245,246]. Conversely, PARKIN mRNA or protein levels decreases in skeletal muscles from some elderly populations, perhaps related to the loss of muscle mass and poor physical function, e.g., physically inactive frail older women [247,248] or gastric cancer patients with cachexia [249].

In the last two years many studies using loss/gain of function models have provided insight on the role of PARKIN in skeletal muscle. Loss of function mouse models pointed out the essential role of PARKIN in basal conditions for the maintenance of (i) mitochondrial function [250,251] and (ii) skeletal muscle mass and normal contractile function [184,251]. Such studies also reported that PARKIN helps to resist to some drug-induced muscle damages [252] and is required for exercise-induced mitophagy flux and for the accumulation of functional mitochondria following muscle adaptations to training [250]. In addition, these loss-of-function studies also highlighted that PARKIN-mediated mitochondrial clearance contributes to proteasome activation during denervation in atrophied slow-twitch muscles [253]. On the flip side, gain-of-function studies showed that PARKIN overexpression in mice: (i) attenuates the ageing-related and the sepsis-induced muscle wasting and causes hypertrophy in adult skeletal muscle, (ii) increases mitochondrial content and enzymatic activities and (iii) protects from ageing-related increases of oxidative stress markers, fibrosis and apoptosis [185,186]. It is very likely that this role of PARKIN in controlling muscle mass has been evolutionary conserved. Indeed, similar observations were also reported in the fruit fly model: Parkin deficiency in Drosophila leads to severe degeneration of the flight muscles with accumulation of swollen mitochondria [254] whereas Parkin overexpression promotes mitophagy in older muscles and extend lifespan.

Together, these studies clearly indicate that PARKIN is an important player in the control of muscle mass through its role in the maintenance of mitochondrial homeostasis. This makes it a potential therapeutic target of interest for preserving muscle mass or fighting against atrophy. Nevertheless, the regulation of PARKIN can be very different according to the physiological or pathological situation or during ageing. Further investigations should enable defining how this actor could be a target of interest according to the population considered.

3.3.4. MUSA1/FBXO30

FBXO30, also called muscle ubiquitin ligase of the SCF complex in atrophy-1 (MUSA1), is a FBOX protein forming an SCF complex with SKP1, Cullin1 and ROC1 [77]. Proteins targeted by MUSA1 remain undefined, but its inhibition in denervated muscles reduces remarkably muscle atrophy, and reverts almost completely the strong atrophic phenotype of Smad4-KO mice [77] (Figure 1 and Figure 2). In muscle, Musa1 expression is upregulated in atrophic mice muscle undergoing CKD [255] or sepsis [256].

3.3.5. FBXL21

Very recently, a new E3 ubiquitin ligase involved in muscle function control has emerged, FBXL21 [188]. FBXL21 forms an SCF E3 ligase complex and was first identified as clock-controlled E3 ligase modulating circadian periodicity via subcellular cryptochrome degradation [257]. Accordingly, in mice, the Psttm mutation, corresponding to a hypomorphic mutation of FBXL21 with reduced FBXL21 activity, caused circadian period shortening [257]. Further studies of these mice revealed that they also displayed skeletal muscle deficiencies with a decrease in fiber CSA (gastrocnemius) and impaired exercise tolerance and grip strength for both forelimbs and hindlimbs [188]. The authors nicely demonstrated the circadian degradation of the cytosolic TCAP/Telethonin by FBXL21 (Figure 2), under the control of GSK-3β. They reported that GSK-3β phosphorylated both FBXL21 and TCAP leading to FBXL21-CULLIN1 complex formation and phosphodegron-dependent TCAP turnover.

3.3.6. Ubiquitin Ring-Type E3 Ligases (UBR)

Ubiquitin Ring-type (UBR, also referred to as E3α) proteins are RING finger E3 ligases that compose a 7-member family and that mainly recognize their substrate through the N-end rule pathway [258]. A first member, UBR2/E3alpha-II, has been shown to be significantly induced in skeletal muscle, in two different animal models of cancer cachexia, at the onset and during the progression of muscle wasting [259]. However, its exact function and importance in skeletal muscle maintain during catabolic states have not been further studied. UBR4 is overexpressed in the skeletal from fasted mice and genetic ablation of UBR4 preserves muscle mass in tumor-bearing mice [189] (Table 1). Intriguingly, the protection of UBR4 knockout against tumor-induced atrophy was limited to type IIA fibers. In contrast, UBR5 has been implicated in muscle hypertrophy [260] and reported to be at least partially associated to the proteasome [261]. Recently several members of the UPS have been described as UBR5 substrates, which included an E2 (UBE2B, an abundant muscle E2), several E3 ligases, proteins involved in chromatin remodeling, etc. [189]. As the main UBR4 targets are positive regulators of muscle growth, the authors concluded that UBR4 acts as a negative regulator of muscle hypertrophy.

3.3.7. FBXO21/SMART

FBXO21/SMART forms an SCF complex with Skp1, Cullin1 and Roc1, in skeletal muscle and has been shown to promote atrophy during denervation [187]. Indeed, the authors showed that FBXO21/SMART upregulation was required for atrophy while, knock-down in TA muscle protected denervated muscles from atrophy (Table 1), probably due to a global reduction of protein ubiquitination [187]. FBXO21/SMART might therefore be a new critical E3 to target to limit skeletal muscle atrophy. Further work should determine whether this E3 is crucial for the development of atrophy in other catabolic conditions and what are the mechanisms involved.