A number of muscular disorders are hallmarked by the aggregation of misfolded proteins within muscle fibers. A specialized form of macroautophagy, termed aggrephagy, is designated to remove and degrade protein aggregates. Abnormalities in this pathway are highlighted in a specific muscular disorders.
The accumulation of misfolded proteins is a common pathological characteristic of degenerative diseases of the central nervous system, but it is also present in muscular disorders where the inhibition of autophagy leads to protein aggregates and/or vacuoles formation, causing muscle fiber degeneration and myopathy 
Autophagic vacuolar myopathies (AVMs) and protein aggregate myopathies (PAMs) represent a series of clinically heterogeneous myopathies characterized by lysosomal or extralysosomal accumulation of proteins or other substances, with varied ages of onset, progressive disease course and variable degree of severity 
AVMs, characterized by the presence of autophagic vacuoles, can be classified into three groups based on specific pathomechanical and morphological findings.
Specifically, the first group is characterized by the deficiency of the lysosomal α-1,4-glucosidase enzyme, causing Pompe disease, and the second by abnormal and functionally impaired lysosomes as in X-linked myopathy with excessive autophagy (XMEA) and Danon disease. The third group is defined by secondary lysosomal dysfunction leading to rimmed vacuoles, popcorn-like clear membrane-bound organelles with a densely blue rim by hematoxylin and eosin staining. This group comprises inclusion body myopathy with Paget disease and frontotemporal dementia (IBMPFD), hereditary and sporadic inclusion body myopathy (hIBM and sIBM) and oculopharyngeal muscular dystrophy (OPMD), among others 
PAMs are mainly identified by the presence of extralysosomal protein aggregates, and it has been hypothesized that abnormal protein synthesis, impaired extralysosomal degradation system or the integration of proteins in the intracellular structured constituents could contribute to the development of protein aggregation 
. One type of PAMs could be classified as “catabolic”, where the aggregation is likely due to extralysosomal protein degradation failure, as observed in but not limited to myofibrillar myopathies (MFMs), cores diseases, sIBM and hIBM, reducing body myopathy and oculopharyngeal muscular dystrophy. In another subgroup, defined as “anabolic PAM” marked by actin filament and granular myosin aggregates, the accumulation is ascribable to synthetic/developmental defects of actin and myosin filaments, notably associated with onset often in early childhood. Nemaline myopathies and myosinopathies belong to this group 
2. Protein Quality Control (PQC) System: Protein Misfolding and Aggrephagy
2.1. Protein Folding and Misfolding and Molecular Helpers
The folding of newly synthesized proteins is a highly demanding process, especially in the crowded cellular environment where protein content might reach concentrations of 300–400 g/L 
. Folding is ruled by kinetic (the vectorial nature of translation) and thermodynamic (the need for energy minimization) factors. It can occur co-translationally, during ribosomal-assisted synthesis either on free ribosomes or on ribosomes that are bound to the sarcoplasmic/endoplasmic reticulum (SR/ER) 
, as well as post-translationally in the cytoplasm or in confined organelles (mitochondria and endoplasmic reticulum) 
The newly synthesized polypeptides, while reaching their native structures, fluctuate between unfolded and folded states, depicting a funnel-shaped energy landscape 
. Moreover, while the energy landscape is smoother for small proteins (<100 amino acids in length) that can fold very rapidly within microseconds, larger proteins (>100 amino acids), which account for ~90% of the proteome in a cell, have a rougher path because of a higher tendency to collapse in the aqueous solvent 
It is no surprise then that proteostasis (the maintenance of protein homeostasis) involves numerous pathways composed of hundreds of proteins such as specialized chaperones and folding enzymes, facilitating folding and refolding, and degradation components like the ubiquitin-proteasome system (UPS) and autophagy for the removal of the permanently misfolded and aggregated proteins 
Heat shock proteins are a special family of chaperones that are upregulated upon thermal stress or other proteotoxic stresses that elevate the concentrations of aggregation-prone intermediates. They are classified based on their molecular weight (HSP40, HSP60, HSP70, HSP90, HSP100 and the small HSPs) 
The HSP70 family includes 47 proteins encoded by 17 genes in the human genome that can be stress-inducible, such as HSP70s, or constitutively expressed as the heat shock cognate HSC70 
All HSP70 members share common structural features, having a 45 kDa N-terminal ATPase nucleotide binding domain (NTD) and a 25 kDa C-terminal substrate binding domain (SBD) that is able to recognize co-chaperones through its conserved motif EEVD 
. The mechanism of action of HSP70 chaperones is based on their ability to cycle through conformational changes in an ATP-dependent manner under the regulation of the chaperones of the HSP40 (DnaJ) family and nucleotide exchange factors (NEF).
A member of the HSP70 family, well-conserved among eukaryotes, is BiP, which, along with lectin chaperones, is the master of ER quality control (ERQC). BiP recognizes nascent polypeptide chains of both glycosylated and non-glycosylated proteins and can bind to unfolded substrates in an ATP-dependent manner. This process of binding and release is assisted by co-factors such as the ER-localized DnaJ-like proteins (ERdjs) and two NEF named GRP170 and Sil1 
HSP40 proteins, also known as JDPs (J domain proteins), are encoded by 50 genes in the human genome 
. They are all characterized by the presence of a ~70 amino acid J domain highly conserved and have been divided into three classes, A, B and C (previously named class I, II and III). JDPs from class A and B also share a glycine-phenilalanine-rich region (G/F) and two sandwich domains at the C-terminal, while only class A present a Zn-binding domain. The class C proteins have little homology with one another, only sharing the J domain, which could be located at any place in the protein sequence 
HSP90/HSC family proteins are encoded by six genes in the human genomes and function as a homodimer. Each HSP90 monomer is composed of three highly conserved domains, an N-terminal domain (NTD) implicated in the binding of ATP, a middle domain (MD) necessary for the ATP hydrolysis and a C-terminal domain (CTD) with a MEEVD motif for co-chaperone coupling. Upon binding and hydrolysis of ATP, HSP90 cycle through different conformational states with different affinity for HSP70, the adaptor protein Hop/Sti1 and the protein client to fold 
Chaperones of the HSP110 family are encoded in the human genome by four genes and are highly homologous to the HSP70 family. Three of these proteins (HSPH1-3 according to a revised nomenclature 
) are both nuclear and cytosolic, while the fourth HSPH4, also known as GRP170 or HYOU1, is located in the SR/ER where it covers a fundamental role in the ERQC, also acting as NEF for BiP 
2.2. Alternative PQC: Autophagy and Aggrephagy
If the misfolded state of proteins becomes irreversible, then the ubiquitin-proteasome and the autophagy systems intervene for their degradation.
The initial step for both mechanisms is the ubiquitination of the cargoes that have to be recognized and targeted for degradation. Ubiquitin is a small regulatory protein (76 amino acids) 
that was first described in 1975 by Goldstein and colleagues 
. It is part of a family of proteins with diverse amino acid sequences but sharing similar folding structures 
. Ubiquitin is found in all eukaryotes but not in prokaryotes, and its sequence is remarkably conserved between humans and yeast, with a difference of only three amino acids.
The cargoes’ ubiquitination occurs at lysine residues via an isopeptide bound with the C-terminal Gly of ubiquitin. Furthermore, ubiquitin itself has seven lysine residues (K6, K11, K27, K29, K33, K48 and K63) and an N-terminal methionine (M1) which in turn can be ubiquitinated. Moreover, ubiquitin can undergo ubiquitin-like modifications such as SUMOylation and NEDDylation, thus generating a wide range of signaling codes. Conjugation of ubiquitin occurs in a three-step cascade sequentially involving three classes of enzymes: the ubiquitin-activating enzyme E1, the ubiquitin-conjugating enzyme E2 and a substrate-specific ubiquitin-protein ligase E3.
Before proteasome degradation could proceed, misfolded proteins should be unfolded and fit through the narrow pore of the 20S subunit. If this cannot occur, ubiquitinated proteins are targeted for autophagy degradation. In particular, the ubiquitination of Lys68 seems to be preferentially a signal for degradation through this pathway 
It is nowadays clear that UPS and autophagy are interconnected, sharing some of their components, and the perturbation of one system affects the other. For instance, proteasome inhibition induces activation of autophagy as a compensatory system in cell lines 
as well as in in vivo experiments in genetically modified Drosophila melanogaster 
The term autophagy was first conceived by biochemist Christian de Duve, a Nobel Prize laureate in Physiology or Medicine of 1974, for his discovery of lysosome. In 2016, another Nobel Prize in Physiology or Medicine was awarded to the cell biologist Yoshinori Ohsumi for his in-depth studies on the autophagy machinery and the discovery of Autophagy-related genes (Atg) 
. In mammalian cells, three types of autophagy are characterized, chaperone-mediated autophagy (CMA), microautophagy and macroautophagy.
Autophagy is a tightly regulated process necessary for the maintenance of proper cellular homeostasis by bulk degradation of cytoplasmic content during nutrient starvation or the removal of specific elements such as damaged organelles, lipids, pathogens and, more importantly, protein aggregates.
Activation of autophagy involves the formation of an isolation membrane, the phagophore, which expands into a double-membrane vesicle, the autophagosome, to engulf what needs to be recycled. The fusion with a lysosome generates an autolysosome in which the lytic enzymes can carry out degradation 
. In brief, initiation of autophagy is achieved by the ULK complex comprised of, in mammals, ULK1or ULK2 as well as the mammalian homologs ATG13, ATG101 and RB1 inducible coiled-coil 1 (RB1CC1, also known as FIP200). The following nucleation step is controlled by the ATG14-containing class III phosphatidylinositol 3-kinase (PtdIns3K) complex, consisting of PIK3C3/VPS34, PIK3R4/VPS15, BECN1, the nuclear receptor binding factor 2 (NRBF2) and the membrane curvature sensor ATG14. The phagophore expansion step then can proceed with regulation from the ATG12-ATG5-ATG16L1 complex and the Atg8/LC3 complex composed of different subfamilies and their isoforms (LC3 family including LC3A, LC3B, LC3B2 and LC3C and GABARAP family including GABARAP, GABARAPL1 and GABARAPL2). Closure of the membrane and fusion with a lysosome will form a mature autolysosome 
Molecular regulation of autophagy involves a complex yet beautifully harmonized cascade of events that is initiated by inhibition of the mechanistic target or rapamycin mTOR, which is composed of two complexes, mTOR complex1 and mTOR complex2 (mTORC1 and mTORC2). Likewise, the energy-sensing AMP-activated protein kinase (AMPK) is also involved in the initiation of autophagy by the inhibition of mTOR and activation of ULK1 through its phosphorylation. Many detailed studies illustrate in depth these mechanisms 
2.3. Key Proteins in Aggrephagy
One of the first studied proteins involved in the removal of aggregates is the cytoplasmic histone deacetylase 6 (HDAC6), a microtubule-associate deacetylase 
containing a ubiquitin-binding domain (BUZ) and two catalytic domains (DD1 and DD2). Kawaguchi and colleagues 
showed that, in an in vitro system, HDAC6 interacts with polyubiquitin-positive aggresomes via its BUZ domain. It can also associate with the motor protein dynein through its dynein motor binding domain (DMB). Moreover, it was demonstrated that HDAC6 mediates the recruitment of lysosomes to the MTOC, thus allowing the removal of aggregated Huntingtin in vitro 
Sequestosome 1, also known as p62, is the first selective autophagy receptor that was described to be capable of binding polyubiquitinated proteins 
as well as binding directly to LC3 and other human homologues of Atg8 
. Mutations in the p62 gene have been linked to diverse human diseases such as amyotrophic lateral sclerosis (ALS), frontotemporal dementia (FD), neurodegeneration with ataxia and also distal myopathy with rimmed vacuoles and Paget disease of the bone 
. P62 is ubiquitously expressed and is able to shuttle between the nucleus and cytoplasm 
and is characterized by specific domains that shape its functions. It contains a Phox1 and Bem1p (PB1) domain necessary for homo or hetero-oligomerization, thus mediating the formation of protein aggregates, a ZZ-type zinc finger (ZZ) domain, two nuclear localization signal (NLS) sequences and a nuclear export signal (NES) sequence, an LC3-interacting region (LIR) and a C-terminal ubiquitin-associated (UBA) domain. Homodimerization via the UBA domain keeps p62 inactive 
, while phosphorylation of the Ser407 is able to reverse it to an active monomeric form 
. Thus, through its PB1 domain, p62 is able to bind ubiquitinated proteins, forming long helical filaments 
that are next targeted for autophagy degradation via the binding to LC3 with its LIR domain 
2.3.3. NBR1 (Neighbor of BRCA1 Gene 1)
NBR1 is a soluble selective autophagy receptor that was discovered soon after p62 for its ability to bind to LC3 and the presence of a similar structural organization 
. It is evolutionarily conserved and precedes the existence of p62 since most non-metazoans only contain NBR1 but not p62 
. Its structure contains a PB1 domain which is able to bind to p62, a ZZ domain, an LIR motif for the interaction with LC3 and a C-terminal UBA domain 
. Moreover, unlike p62, it contains four tryptophan (FW) domains allowing its binding to microtubule-associated protein MAP1B and TAX1BP1 
, as well as an amphipathic helix (HA) domain that, together with the UBA domain, is necessary for co-localization with LAMP2 
2.3.4. WDFY3 (Alfy)
The link between Alfy (autophagy-linked FYVE protein) and protein aggregates’ removal was first demonstrated in 2004 
. Alfy is a 400 kDa protein, evolutionarily conserved and expressed in all tissues. It contains five WD40 repeats and a PH-BEACH domain assemblage, and a C terminal phosphatidylinositol-3-phosphate (PI3P)-binding FYVE domain; therefore, it is also known as WDFY3. Alfy was found to co-localize with the p62-positive aggregates, and, in vitro, this interaction was found to be mediated by the region comprising amino acids 170–206. It is required for aggrephagy but not for starvation-induced bulk autophagy 
2.3.5. TOLLIP (TOLL-Interacting Protein)
Tollip is the mammalian homolog of the yeast Cue5. It is a member of the CUET family of proteins and therefore is characterized by a ubiquitin-binding CUE domain and an LC3-interacting region (LIR). It was demonstrated that Cue5 and Tollip are able to promote the clearance of the aggregate of human huntingtin mutant and that Tollip is more efficient than p62 in the clearance of aggregates in HeLa cells 
Optineurin is a cytosolic protein ubiquitously expressed and with numerous cellular functions such as vesicle trafficking, cell division control, and autophagy. It is linked to numerous diseases, like amyotrophic lateral sclerosis and primary open angle glaucoma (POAG) 
. Optineurin is constituted by a NEMO-like domain, an LC3-interacting (LIR) domain, an LZ domain, a ubiquitin-binding domain (UBD), coil-coiled (CC) domains, and a zinc-finger-like (ZnF) domain. It was first demonstrated 
that lack of optineurin in HeLa cells increases aggregation of SOD1 G93C and huntingtin mutant, revealing its involvement in the aggregates’ clearance.
2.3.7. TAX1BP1 (Tax1-Binding Protein 1)
TAX1BP1 involvement in the clearance of aggregates was first discovered in 2012 by Newman and colleagues 
, and it was more recently demonstrated that loss of TAX1BP1 induces the formation of proteotoxic aggregates in the brain 
. TAX1BP1 has an N-terminal SKIP carboxyl homology (SKICH) domain, a central oligomerization domain containing three coiled-coil (CC) regions, and a C-terminal ubiquitin binding domain (UBD) containing two zinc fingers 
. Its direct interaction with NBR1 allows its recruitment to the aggregates and, in turn, TAX1BP1 is able to recruit the autophagy factor FIP200 (an ULK-interacting protein also known as RB1-inducible coiled-coil protein 1, RBCC1) via the Claw domain, allowing autophagosome formation 
2.3.8. CCT2 (Chaperonin Containing TCP1 Subunit 2)
Very recently, Ma and colleagues 
identified a new function of the chaperonin subunit CCT2 in the elimination of aggregates. The chaperonin TCP-1 ring complex (TRiC) is constituted by eight subunits (CCT1-8) and each one contains an equatorial domain responsible for the binding of ATP as well as an apical domain responsible for substrate binding.
2.4. Aggrephagy Players as Therapeutic Targets
Defective autophagy is known to be linked to various human diseases, such as neurodegenerative, cardiovascular, metabolic, autoimmune diseases and cancer, as well as aging 
. Various drugs targeting macroautophagy exist, and although their pharmacological effects are non-specific, they have been approved in the clinical practice for diseases other than myopathies. Their mechanism of action and use in clinical settings or in preclinical studies have been recently reviewed 
. Among the activators of autophagy, rapamycin (known as sirolimus) and its analogs (rapalogs), such as everolimus, tacrolimus and temsirolimus, to name some, act by allosterically inhibiting the kinase activity of mTORC1.
Despite many compounds being known to modulate autophagy and, therefore, could be potential drugs for those diseases in which autophagy is impaired, more studies are needed to fully clarify their exact mechanism of action and, moreover, to reduce the tissue’s toxicity, especially in chronic administration.
3. Aggrephagy Involvement in Muscle Diseases
3.1. Myotilinopathies and Desminopathies
Myofibrillar myopathies (MFMs) represent a broad hereditary group of muscle disorders characterized by myofibrillar alterations, Z-disk disorganization, sarcoplasmic buildup of myofibrillar degradation products, often with the presence of rimmed vacuoles, and accumulation of multiple proteins in the muscle tissue. Therefore, MFMs diagnosis is based on histological features since the clinical spectrum is relatively wide.
Recently, due to the use of next-generation sequencing, patients with MFM phenotypes have been found to carry mutations in new genes, such as SVIL 
and UNC45B 
. Additionally, known genes related to different disorders, such as FHL1
and p62 + TIA1
, have now been associated with MFM phenotypes 
Protein aggregates in MFMs have been found to contain a great number of different proteins, including phosphorylated tau (p-tau) and β-amyloid, and have also been associated with increased immunoreactivity for the proteasome 19S and 20S subunits 
. Prompted by these observations, Olivé and colleagues 
examined and first reported in 2008 the presence of p62 and UBB + 1, a mutant form of ubiquitin B known to accumulate in neurodegenerative disorders such as Alzheimer’s and Huntington’s disease, in a cohort of myotilinopathy and desminopathy patients. In their analyzed samples, twelve myotilin and five desmin-mutated patients, the muscle biopsies showed the presence of protein aggregates within the cytoplasm or in the subsarcolemmal region of the muscle fibers as revealed by the modified Gomori trichrome staining.
3.2. Sporadic Inclusion Body Myositis
sIBM is the most recurrent sporadic muscle disease in older populations, with the slow progression of muscle impairment accompanied by muscular atrophy. It is still categorized as idiopathic inflammatory myopathies, similar to polymyositis and dermatomyositis, which are characterized by endomysial inflammation associated with various degenerative changes of the muscle fibers and heterogeneous muscle impairment 
sIBM generally shows CD8+ T cells invasion of non-necrotic muscle fibers, incremented expression of MHC class I, rimmed vacuoles, and protein aggregates which could contain, among others, beta amyloid protein, p62 and TDP-43, like those found in Alzheimer’s disease, in amyotrophic lateral sclerosis and frontotemporal dementia 
Prof. Valerie Askanas’s group investigated the expression of proteins involved in the selective removal of aggregates in muscle biopsies of sIBM patients defined by the typical clinical features 
. Their interest arose from the observation of a certain degree of phenotypic similarity between the sIBM muscle tissue and the brain of Alzheimer’s and Parkinson’s disease patients. The group showed that in sIBM but not in control samples (including polymyositis, dermatomyositis, amyotrophic lateral sclerosis, peripheral neuropathy patients and normal controls), there is increased positivity for p62, LC3 and NBR1. Specifically, in 80% of the vacuolated and 20–25% of the non-vacuolated fibers, p62 strongly accumulates within the aggregates, co-localizing with ubiquitin and p-tau 
3.3. Rigid Spine Syndrome with FHL1 Mutation
Reducing body myopathy due to FHL1 mutations is an X-linked disorder, marked by the presence of intracytoplasmic inclusion bodies that contain a multitude of proteins, among which mutant FHL1 protein is the most represented.
In patients affected by rigid spine syndrome (RSS) with a mutation in the FHL1
gene, Bonaldo’s group described for the first time the activation of the pathway involved in the selective removal of aggresomes 
. In about 20% of the muscle fibers of the proband and her brother’s biopsies, the scholars identified reducing bodies (RBs) that were Congo-red positive, thus indicating an amyloid-like nature and also revealed positivity for the FHL1 antibody.
3.4. Oculopharyngeal Muscular Dystrophy
Oculopharingeal muscular dystrophy has mainly autosomal dominant inheritance with progressive late onset of proximal limb, facial and extraocular muscle weakness due to a poly-alanine expansion in the Poly(A) Binding Protein Nuclear 1 (PABPN1) 
It has been demonstrated that PABPN1 plays a critical role in mRNA biogenesis and myogenesis. In fact, PAPBN1 is implicated in post-transcriptional RNA processing, including long non-coding RNAs (lncRNAs) 
. Despite being ubiquitous, its alteration results in a muscle-specific disease, with defective myoblast proliferation and differentiation correlated with shorter mRNA poly(A) tails and subsequent nuclear accumulation of poly(A) RNA and misfolded PABPN1. Thus, it has been postulated that the pathogenetic mechanism involves a reduced bioavailability of PABPN1 due to its aggregation [184
In a cohort of 21 Chinese patients clinically affected by OPMD and genetically characterized by five different pathological genotypes, Lin and colleagues investigated the involvement of the autophagy pathway, focusing on the physical association between PABPN1 and p62, NBR1, LC3 and the ubiquitinated protein marker FK2 [186
3.5. PLIN4-Related Myopathy
Recently, in an Italian family affected by a late onset myopathy, with distal or less frequently limb-girdle muscle involvement and autophagic features within the muscle, it was identified by multi-omic approach a potentially causative large coding expansion in the PLIN4
gene encoding for perilipin 4 
. Perilipin 4 is part of a family of proteins involved in the metabolism of lipids, protecting them from lipases by coating the lipid droplets (LDs). Its structure is more divergent compared to the other perilipins, containing a large amphipathic helix responsible for the interaction with the lipids, which is organized in a 3–11 alpha helix structurally similar to α-synuclein and apolipoproteins. The identified genetic mutation causes a 99 nt coding region to be repeated nine more times compared to the wild-type, thus adding 297 extra amino acids to the encoded protein.
The pathogenic mechanism involved is indicated to be the aggregation of the mutated protein with consequent activation of the aggrephagy pathway for the aggregates’ removal, as demonstrated by the co-localization between perilipin 4 signal and FK2, p62 and NBR1 as well as by the increased signal of WDFY3. The aggregates are present at the subsarcolemmal region of the muscle fibers and within the vacuoles, and positivity for these antibodies is directly correlated with the severity of the phenotype, thus indicating a progressive blockage and inefficiency of the aggrephagy mechanism.