Initial studies, mostly performed using muscles isolated from animals subjected to induced muscle wasting, led to the conclusion in the early 1990s that increased ATP-dependent proteolysis accounted for a large part of muscle protein loss in several muscle wasting models 12. After the subsequent recognition that the UPS (initially discovered and dissected in rabbit reticulocyte lysate 131415) was responsible for ATP-dependent proteolysis in mammalian cells 16, and the demonstration that its general inhibition by proteasome inhibitors reduces the accelerated proteolysis in atrophying muscles 17, several groups turned to the analysis of expression of individual UPS components in different experimental models of muscle wasting (see Table 1).

In general, the UPS functions in two distinct steps to degrade intracellular proteins. First, the substrate protein is labeled by covalent linkage of polyubiquitin (polyUb) chain(s) to one or several lysine residue(s) by the action of an enzymatic cascade involving three types of factors, referred to as E1, E2s and E3s 18. In this reaction, the E3s act as substrate recognition factors and are thus the main determinants of the specificity of the UPS 19. Consequently, their number is very high, with a rough estimation of several hundred E3s present in the genome, depending on the organism 20. The second step in protein degradation by the UPS is the degradation of the polyubiquitylated substrate by the 26S proteasome, a large compartmentalized protease comprising the 20S proteasome, which contains the proteolytic active sites in its internal catalytic chamber, and the 19S regulatory complex, which is essential for degradation of ubiquitylated proteins 2122.

A common theme in most models of muscle wasting is a significant increase at the mRNA level of many components of the UPS. These include ubiquitin (Ub), E2s, proteasome subunits (20S and 19S), deubiquitylating enzymes and certain E3s 68232425. The accepted interpretation is that this upregulation of UPS mRNAs participates in the overall activation of the UPS observed during muscle wasting, in line with the fact that an increase in (i) the rate of Ub conjugation to proteins, (ii) the intracellular level of ubiquitylated proteins and (iii) proteasome activity has been described in most muscle wasting conditions 68; see also Table 1]. However, it should be noted that, despite the upregulation of their corresponding mRNA, the level of many UPS proteins does not appear significantly altered during muscle wasting. The reasons for this discrepancy are unclear, but an interesting hypothesis is that the increased synthesis of these proteins is compensated by an accelerated turnover under the conditions analyzed.

The first specific UPS pathway described to be involved in muscle wasting was the ‘N-end rule’ pathway (2627; see also section on Substrates of the UPS during muscle wasting), which rapidly degrades proteins with destabilizing N-terminal residues (see 28 for review). In mammals, this pathway involves a specific E2 (E2-14K) and several E3 isoforms 29, among which E3α-II (also known as UBR2) is increased in skeletal muscle during cachexia 27. The specific inhibition of this E3 reduces the rate of ubiquitylation of soluble or model proteins in muscle extracts from tumor-bearing or septic rats, as well as in C2C12 cell extracts 2627.

Transcriptome analyses showed that two other types of E3 or E3 components are upregulated in most muscle wasting models. These are the F-box component MAFbx (also known as Atrogin-1) of the SCF^MAFbx E3 and the MuRF family of RING finger E3s 3031, which interact with and possibly ubiquitylate a variety of myofibrillar proteins 3233. Expression of MAFbx and MuRF-1 mRNAs increases dramatically in catabolic states and mice deficient in either of them are partially resistant to denervation atrophy 31. Therefore, MAFbx and MuRF-1 play a crucial role in the loss of muscle proteins and their mRNAs are now considered to be specific atrophic markers.

The major proteins in muscle are the myofibrillar components; however, although monomeric actin and myosin can be degraded by the UPS, it is assumed that this is not the case in intact myofibers since specific interactions between these proteins protects them from degradation by this system 34. Thus, an early step in muscle wasting must be the dissociation or at least the destabilization of the myofibrils, a process in which non-UPS proteases shown to be involved in muscle wasting (such as calpains 3536, caspase 3 37 or cathepsin L 38) probably play a key role. Though the early involvement of non-UPS proteases in the destabilization of myofibrils has not been formally demonstrated, this hypothesis is likely since it helps to explain the observed role of the UPS N-end rule pathway in muscle wasting. Indeed, this pathway requires the substrate to expose a destabilizing N-terminal residue 28, a feature that is rare for native cellular proteins; however, polypeptides with a destabilizing N-terminal residue can be produced by proteolytic cleavage of native proteins by cellular proteases 39.

Another mechanism that controls myofibril organization involves the sarcomeric scaffold protein titin, which possesses a protein kinase domain essential for sarcomere integrity 4041. Interestingly, the activity of this domain is modulated by mechanical strain 42, allowing titin to convert tension forces into molecular signals. This function controls both the intracellular level and localization of the RING finger E3 MuRF-2 (a protein closely related to MuRF-1), which is associated with titin in the sarcomeres through interaction with the UBD (Ub binding domain 43)-containing proteins Nbr1 and p62, and is translocated to the nucleus upon muscle inactivity 44. Nuclear localization of MuRF-2 in turn leads to nuclear exclusion of SRF (serum response factor) 44, a transcription factor central to muscle response to hypertrophic stimuli. Thus, titin appears to be an important relay in the control of muscle disuse atrophy, since it is able to couple mechanical inactivity with muscle wasting by triggering macromolecular rearrangements of sarcomeres and altering muscle transcriptional programs. The importance of the titin–Nbr1–p62–MuRF-2 interaction is underlined by a human mutation disrupting binding of Nbr1 to titin, which causes an autosomal dominant muscle disease called heredity myopathy with early respiratory failure (HMERF) 44. In muscle biopsies of HMERF patients, the structure of sarcomeres is altered and the localization of Nbr1, p62 and MuRF-2 is abnormal 44.

In addition to these mechanisms, the crucial roles of MAFbx and MuRF-1 during atrophy indicate that these E3s target key negative regulators of muscle wasting for degradation. However, only a few actual or potential substrates have been identified for these E3s and the relevance of these substrates for muscle wasting remains unclear (see Signaling pathways involved in muscle wasting).

Glucocorticoids are known to be important mediators of muscle wasting in various catabolic conditions 5; indeed, dexamethasone triggers upregulation of various UPS components 45. However, the underlying mechanisms of glucocorticoid signaling in muscle wasting are not yet fully understood 23.

Presently, two families of transcription factors have been clearly identified as crucial positive mediators of muscle wasting: the FoxO (forkhead box O) transcription factors and NFκB 9. Of note, these factors are involved in the control of expression of the E3s MuRF-1 and MAFbx 464748.

PI3K is activated by anabolic signals triggered by insulin or IGF-1 upon binding to their receptors, which in turn activates Akt 49 (a kinase central to muscle hypertrophy), as shown in transgenic mice in which an active form of Akt can be inducibly expressed in adult skeletal muscle 50. Besides its function in stimulating protein synthesis via mTOR and GSK3 9, a crucial role for activated Akt is to phosphorylate FoxO 464751, leading to transcriptional inactivation of FoxO due to its release from DNA and 14-3-3 protein-mediated cytoplasmic sequestration (see 52 for review), and to subsequent inhibition of muscle wasting. In many systemic diseases, including diabetes, cancer and renal failure, low circulating levels of insulin and IGF-1, or resistance to their action, contribute to muscle wasting via inactivation of the PI3K/Akt pathway, which leads to the subsequent dephosphorylation and activation of FoxO 53545556. Interestingly, it has recently been shown that myostatin treatment can also trigger the same effect in mice 57. Myostatin is a TGFβ super-family member that acts as a negative regulator of muscle growth and is implicated in several forms of muscle wasting 58. Myostatin signaling reverses the IGF-1/PI3K/Akt pathway by inhibiting phosphorylation of Akt, leading to activation of FoxO and thereby activation of E2-14K, MuRF-1 and MAFbx 57.

In many disease states, stress or inflammatory cytokines such as tumor necrosis factor alpha (TNFα) are produced 59. Stimulation of their receptors activates the IKK kinase, which in turn phosphorylates the NFκB inhibitor IκBα, promoting its degradation by the UPS. This leads to the release of NFκB, which migrates to the nucleus to activate its target genes 60. Demonstration of a role for this pathway in muscle wasting comes from many observations, including the fact that constitutive activation of IKK in transgenic mice triggered both a decrease in total muscle mass and an increase in atrophy markers, which were reversed by expression of IκBα 48. A further observation supporting a role for this pathway in muscle wasting is that direct inhibition of NFκB inhibited cachexia in a mouse tumor model 61. In addition, in cancer cachexia, tumor-derived factors such as proteolysis-inducing factor (PIF) 6263, or elevated levels of endogenous factors such as angiotensin I and II, are thought to contribute to muscle atrophy through activation of the NFκB pathway. Supporting evidence comes from the fact that these factors induce increased expression of UPS components (proteasome subunits, E2-14K) and stimulate protein degradation in murine C2C12 myotubes in a manner requiring IκBα degradation and nuclear translocation of NFκB 6465. Interestingly, although activation of NFκB is also required for muscle disuse atrophy, the NFκB family members that are activated appear distinct from those activated by the TNFα pathway 166.

NFκB and FoxO are thus key players in the activation of muscle wasting. Interestingly, their known targets include MuRF-1 and MAFbx for FoxO 4647, and MuRF-1 for NFκB 48. However, as mentioned in the section on Substrates of the UPS during muscle wasting, the key substrates of these E3s in muscle wasting are unclear. Nevertheless, one hypothesis is that induction of MuRF-1 activity is important for the destabilization of myofibrillar complexes required for the subsequent degradation of myofibrillar proteins. Indeed, the interaction between MuRF-1 and titin has been shown to be important for sarcomere stability 67. Furthermore, in yeast two-hybrid screens, MuRF-1 interacts with several myofibrillar proteins (titin, nebulin, the nebulin-related protein NRAP, troponin-I, troponin-T, myosin light chain 2, myotilin and T-cap) 33, suggesting that this E3 can mediate their ubiquitylation and degradation in muscle. MuRF-1-mediated ubiquitylation and degradation, however, has only been proven for troponin I 68 and possibly for myosin light chain 2 33 in cardiac tissue. The lack of clear results regarding the role of MuRF-1 in the degradation of its other interactors could be due to its functional redundancy with MuRF-2.

The known targets of MAFbx (calcineurin A 32 and MyoD 69)) are unlikely to be crucial for muscle wasting, since calcineurin A does not appear to affect myofiber size 70 and MyoD is involved in muscle cell differentiation. The key substrates of MAFbx in muscle wasting thus remain to be identified. Interestingly, however, TNFα/NFκB signaling reduces both the level of MyoD mRNA 48 and the stability of MyoD 71, and forced expression of MAFbx inhibits the formation of myotubes in an in vitro model of muscle differentiation 69. Thus, the role of MAFbx in the down-regulation of MyoD could in fact be indirectly important for muscle wasting, since the inactivation of MyoD functions necessary for the differentiation of muscle satellite cells possibly inhibits muscle regeneration.

In this brief overview, we have summarized the role of the UPS as an essential mediator of muscle wasting. Even though much remains to be understood before a complete picture of the intricate mechanisms controlling muscle atrophy can emerge, our knowledge to date already suffices to demonstrate that the UPS is an attractive potential target for therapeutic treatments aimed at slowing down muscle atrophy when it becomes harmful to the organism 5. The different levels of possible therapeutic intervention are discussed in the section on Potential targets and drugs for muscle wasting therapy.

Muscle wasting occurs when the overall rate of protein proteolysis exceeds in a sustained manner that of protein synthesis in muscle. Since a variety of physiological and pathological conditions can influence the equilibrium between these two processes, several animal (mostly murine) models have been developed (see Table 1) to analyze the mechanisms of muscle mass loss in these different conditions 72. For example, cancer, sepsis and cirrhosis are inflammatory conditions that induce muscle wasting through the inhibition of anabolic pathways and the secretion of cytokines (IL1, TNFα) or catabolic factors (PIF), while hindlimb suspension, denervation and sarcopenia are conditions in which disuse atrophy is present. In most models, despite variations in the extent of involvement of the UPS relative to other processes (including other proteolytic systems), activation of UPS-dependent proteolysis transpires to be a crucial event 6. Moreover, as described in the section on Roles of the UPS in muscle wasting, biochemical and molecular analyses have determined that a convergent mechanism for this activation is the upregulation of crucial ubiquitylation pathways involving the N-End rule E3s as well as the E3s SCF^MAFbx and MuRF-1 73, even though the substrates of these enzymes during muscle wasting remain unclear at present.

Several mouse strains deficient in components involved in muscle wasting have been constructed and tested for their protection against muscle atrophy (see Table 2). Manipulation of the N-end rule pathway, by deletion of the genes encoding either E2-14K or the E3 UBR1 7475, did not significantly impair muscle wasting, probably because of the functional redundancy between the many components of this pathway 29 and/or because this pathway is only responsible for the breakdown of soluble muscle proteins. By contrast, mice in which either the MAFbx- or MuRF-1-encoding gene has been deleted become partially resistant to atrophy induced by denervation, highlighting the crucial role of these two E3s in the process of muscle wasting. Finally, mice with altered control of the NFκB pathway demonstrate the key function of this pathway in inducing muscle atrophy.

In addition to the measurement of muscle mass after isolation of specific muscles from model animals, protein breakdown in incubated muscle is often quantified by the release of tyrosine, which is neither synthesized nor metabolized in skeletal muscle 76. Furthermore, release of 3-methylhistidine, a modified amino acid found in actin and myosin that can be neither reincorporated into protein nor metabolized, can also be measured 7778.