Background

Inclusion body myositis (IBM) is an inflammatory myopathy occurring after middle age, with autoimmune and degenerative mechanisms [1, 2]. Other idiopathic inflammatory myopathies (IIMs) are dermatomyositis (DM), polymyositis (PM), overlap syndromes including anti-synthetase syndrome and necrotizing pauci-immune myositis [3]. The distinction between IBM, PM, and PM with mitochondrial pathology is not neat, raising the question whether IBM is a variant of PM occurring in the older age, related to immunosenescence [4]. IBM pathogenesis centrally involves cytotoxic, senescent CD8+ T cells, defects of autophagy and ubiquitin–proteasome system (UPS) resulting in proteostasis impairment and abnormal sarcoplasmic protein aggregation, along with endoplasmic reticulum and mitochondrial alterations, and antibodies to the cytosolic 5′-nucleotidase 1A (anti-cN1A) [1, 5]. The driving mechanisms of this pathology, however, are still evasive.

IBM belongs to a group of neurological disorders, the TDP-43 proteinopathies, which pathogenically involve TDP-43 [TAR-DNA-binding protein 43 (transactive response DNA-binding protein of 43 kDa)] [6]. TDP-43, encoded by the TARDBP gene, an RNA- and DNA-binding nuclear regulatory protein, member of the heterogeneous nuclear ribonucleoprotein (hnRNP) family [7, 8]. In skeletal muscles, TDP-43 is involved in transcription regulation, RNA splicing, mRNA stability, RNA transport, and quality control and undergoes post-translational modifications with functional consequences [9]. TDP-43 functions in muscles are complex, including myoregeneration (Table 1). In neurodegeneration, the mechanisms of TDP-43 involvement include cytotoxic aggregations, nuclear loss, alteration of cellular functions, and others [6, 10].

Table 1 Roles of TDP-43 in muscles
Full size table

IBM muscle biopsies reveal cytoplasmic aggregation of TDP-43 and TDP-43 nuclear loss [10]. Even an 1% amount of myofibers staining for TDP-43 in a muscle biopsy was highly sensitive and specific for IBM [11].

TDP-43 may have an emerging intriguing role in viral infections [12]. TDP-43 is involved in controlling IFN responses triggered by endogenous RNA, but the TDP-43 role as an RNA-binding protein in viral infections is rarely investigated [13, 14]. Loss of TDP-43 results in dsRNA intracellular accumulation and interferon (IFN) triggering [13]. The TDP-43 ortholog of Caenorhabditis elegans called TDP1 limits dsRNA accumulation [21]. Also, knockdown of TARDBP increases viral replication in macrophages [14] and TDP-43 knockdown amplifies enterovirus infections, suggesting an antiviral effect of TDP-43 [22]. Moreover, TDP-43 binding is protective against HIV-1 by sterically hindering a HIV-1 promoter [23]. Also, after TDP-43 knockdown in mouse brain, the type I IFN-inducible genes, including the mouse orthologs of the intracellular sensor molecules RIG-I and MDA-5 which detect viral RNA, are the most overexpressed [21, 57]. IFN-ƴ induces ER stress and aggregation of TDP-43 and other proteins [1, 5, 54]. IFN-I may also be induced by anti-Ro52, present in some IBM patients [50, 58]. Ro52 or TRIM21 (tripartite motif proteins), is an IFN–inducible E3 ligase involved in IFN type I downregulation [59]. Other infection-related factors may intervene in IBM, such as activation of NLRP3 inflammasome, heat shock proteins (HSP), ribosomal proteins, or molecular mimicry with a mycobacterial protein guanylate-binding protein 2 (GBP2) with antiviral and anti-tuberculous functions [5, 60]. GBP2 is involved in the control of mRNA splicing [5] with possible relevance in TDP-43 dysfunction when mRNA splicing is altered.

Also, the glycogen synthase kinase 3 (GSK3), a serine/threonine kinase with 2 isoforms (α and β), is activated in IBM [33]. GSK3, involved in many cellular processes, is an immunomodulator in IBM [33]. GSK3 delays and decreases IFN-1 production, enhances IFNγ signaling, but also increases and delays pro-inflammatory cytokines production [33]. Moreover, GSK3β is one of the protein kinases involved in the TDP-43 phosphorylation [34]. TDP-43 expression activates GSK3, and GSK inhibition decreases TDP-43 aggregation [35].

Also, activation of autophagy is part of the innate immune response, and autophagy receptors may become viral targets [15]. Amongst these autophagy receptors, NBR1 (neighbor of BRCA1), a ubiquitin-binding scaffold protein, increases in viral infections [61], and NBR1 accumulates and is abnormal in IBM muscle [62].

In IBM, the dysregulation of a deubiquitinase called cylindromatosis (CYLD) reduces the autophagic clearance of protein aggregates [63]. CYLD is expressed with phosphorylated TDP-43 in the sIBM myofibers [63]. CYLD, required for antiviral host defense, is involved in the STING cleavage [64] and negatively regulates NF-kB [63].

IFN-ƴ and low RNA amounts in cytoplasm also stimulate aggregation of TDP-43 and other RBPs with “prion-like” low-complexity (LC) domains, favored by proteins misfolding in aging [1, 65].

Potential links between IBM and viral infections

The IBM occurrence may reflect various pathogenic associations, including viral infections [4]. In general, chronic IIM may be triggered by viruses such as Coxsackie B, enterovirus, parvovirus, HTLV-1, or HIV [66]. Mechanisms of viral-induced myositis hypothetically include direct invasion of myocytes by the virus, molecular mimicry, exposure of cryptic epitopes after conformational alterations, myotoxic cytokines such as IFNs and autoimmune reactions [66,67,68]. Latent viral infection, viral-induced denaturation of self-structures or homologies with various viral proteins could result in a prolonged immune response [66]. For instance, enterovirus 71 (EV71) may upregulate TRIM21 (Ro52), which degrades SAMHD1, a host antiviral molecule [59]. Also, during a viral infection, many ribonucleoproteins including TDP-43, are hijacked [12]. Coxsackie virus B3 protease 3C causes TDP-43 cytoplasmic redistribution and aggregation [12, 22].

Also, the aging cellular environment may make the myofiber susceptible to a newly invading virus, or may allow cytopathic manifestation of a virus, or a vertically transmitted genomic endogenous virus such as a retrovirus dormant for years, such as HTLV1, may start to be transcribed due to the age-modified milieu [16]. Endogenous retroviruses (ERVs, genomic remnants of ancient viral infections, most inactive and non-infectious) are mutually reinforcing with TDP-43 proteinopathies regarding neurodegeneration [17, 26]. Moreover, aging may favor both ERVs expression and TDP-43 proteinopathy [26].

However, no definite evidence for a viral etiology of IBM has been established [27]. Mumps virus was described as a potential IBM cause, later questioned in immunohistochemical studies [28]. IBM patients have an increased prevalence of hepatitis C virus (HCV) or human lymphotropic T virus-1 (HTLV1) [69,70,71]. The relationship between HCV and TDP-43 is yet to be clarified. TDP-43 binds YB (Y-box-binding protein-1), a host factor involved in HCV capsids assembling, and TDP-43 knockdown significantly decreased HCV replication [19]. The persistent HCV-related IFN upregulation and lymphocyte exhaustion may in fact contribute to the chronic myopathy in HCV [4]. TDP-43 facilitates HBV gene expression stimulating its transcription and assembly of protein complexes [12]. Furthermore, the clinical picture of IBM patients with HCV is different from the one of patients with IBM and HIV; therefore, no unique mechanism links a chronic viral infection to IBM [20].

Most of the HIV-positive patients with myositis had overlap** features of PM and IBM, which clinically progress to IBM, and most of them have anti-c1NA antibodies and rimmed vacuoles [20]. TDP-43 suppresses HIV-1 transcription by binding HIV-1 long terminal repeat [72]. Knocking down TDP-43 with siRNAs in cell cultures reactivates HIV-1 by reversing its latency [23]. Notwithstanding, HIV-1 can replicate in human immune cells independent of TDP-43 [73]. In viral-associated IBM in HIV and HTLV-1, the viral antigen is not present in myofibers but in the T cells and macrophages instead [1]. HIV infection can induce T cells immune senescence [74]. Thus, it is more conceivable that the virally induced senescent, IFN-ƴ producing cytotoxic CD8+ T cells lead to IBM.

IBM has been reported to be induced by Covid-19 in a 54-year female patient with diabetes mellitus and hyperlipidemia on statins [75]. Also, an axial paraspinal myopathy was reported in Covid-19 [76], and paraspinal myositis may be a feature of IBM [27]. However, long-term consequences of SARS-CoV2 infection, including muscular involvement, are starting to be recognized [77]. After COVID-19, the prevalence of myositis-specific antibodies and myositis-associated antibodies increases [78]. Possible mechanisms include type I IFN pathways, NLRP3 inflammasome activation, or a previous exposure to common coronaviruses [79]. SARS-CoV-2 impairs the stress granules (SGs) disassembly, and the SARS CoV-2 nucleocapsid N protein binds the SG-related amyloid proteins, favoring aggregation [89]. The HLA-DRB1*03 allele, as a component of the ancestral HLA 8.1 haplotype, is a susceptibility factor for IIMs and many other autoimmune diseases [90]. An arginine in position 74 of the DRβ1 chain confers the allelic risk for IBM [89]. HLA DRB1*01 is also associated with rheumatoid arthritis and hematologic malignancies, all overrepresented in IBM and associating age-related stochastic accumulation of CD8+ CD28- T cells [1, 86]. HLA-DRB1 alleles expression also impacts durable control of viral replication, HLA DR B1*03:01 being associated with high HIV viremia [91, 92], while HLA DRB1*01 was associated with spontaneous viral clearance of hepatitis C [92].

HLA DRB1*13 is common for IBM susceptibility and for protection against infection with several viruses, including HIV, HCV, HBV [87]. In IBM, the HLA DRB1*13:01 was associated with the highest age of onset and the lower strength [88]. Nevertheless, intriguingly, HLA DRB1*13 was protective against autoimmune diseases such as systemic lupus erythematosus, psoriasis, systemic sclerosis, and others [93]. However, HLA DRB1*13 is also associated with a slow progression of HIV [94]. HLA DRB1 *13 is associated with the clearance of hepatitis B as well [95]. Surprisingly, HLA DRB1*13 is neuroprotective, along with apoE, against age-related brain changes [96]. HLA-C*14:02:01 allele was higher in IBM patients with high LGL T cell expression [84]. HLA-C*14:02 allele was also associated with a T cell response in HIV-1 infection, which was nevertheless non-protective for the viral infection [97].

HLA-F, found in IBM and Sjogren’s syndrome, also elicits antiviral responses through activation of the KIR3DS1+ NK cells [20]. Therefore, HIV testing is advisable mostly in PM/IBM overlaps [20]. Also, pan-JAK inhibitors in aged mice alleviated the senescence—associated secretory phenotype but may also reactivate latent viruses [55]. Trials of immunosuppressive therapies in IBM have been recently nicely reviewed [105]. Immunosuppression is not routinely advised unless IBM is rapidly progressive or associated with other autoimmune diseases [105].

Followed both inflammatory and myodegenerative pathways presumed to be involved in IBM pathogenesis [105]. Most studies addressed inflammation or the involvement of T cells. Alemtuzumab (against CD52), natalizumab, anti-TNF alpha such as infliximab or etanercept, or IL-1 inhibitors as anakinra and canakinumab showed modest or no improvement [105], Rapamycin (sirolimus) targets mTOR important in IL-2 immune responses and protein metabolism (NCT04789070) [105]. Novel therapeutic avenues involve anti-KLRG1 antibodies, targeting a surface marker of the highly differentiated CD8T cells (NCT04659031) [84, 105]. Moreover, in HIV, the KLRG1 expression on NK cells correlates with HIV transcription, and targeting KLRG1 on NK cells potentially aids in elimination of HIV-infected cells [106]. Therapies against myodegeneration have recently become targets in clinical trials (arimoclomol, bimagrumab, follistatin, oxandrolone, rapamycin) [105].

Possible future directions may address other pathways. The attempts to reduce TDP-43 level led to muscle weakness and defective regeneration in myopathy models [6]. However, in neurological disorders such as ALS and other TDP-43-associated diseases, affecting skeletal and cardiac muscles besides neurons, there are several TDP-43 directed therapies [107, 108]. In ALS inhibition or deletion of cGAS and STING prevents TDP-43-induced upregulation of NF-kB and IFN type I [107]. Nevertheless, the neurological and muscular effects are not completely superposable [6].

Research including new therapies and repurposing for IBM some drugs used with other indications could serve as directions for the future [109]. Future therapeutic approaches could include inhibition of TDP-43 aggregation, the TDP-43-mitochondria association, proteasomal degradation of cytoplasmic TDP-43, or reducing TDP-43 aggregation-induced cell stress [37, 38, 110, 111]. Drugs stimulating the proteasome, such as chlorpromazine and other phenothiazines, methylene blue as a structural analogue of chlorpromazine and pyrazolones may target proteotoxic disorders [112]. The efficacy of zetomipzomib (KZR-616), a selective inhibitor of the immunoproteasome, is being studied in a phase 2 controlled multicenter study for active PM and DM [113, 114]. GSK3 inhibition decreases TDP-43 aggregation [34]. Lithium inhibits GSK-3 and induces autophagy, which may be relevant for IBM [115]. Also, lithium protected synapses from HIV-1 Tat-induced neuronal loss, in cultures and may be neuroprotective in HIV [116, 117]. Some other GSK3 inhibitors (including famotidine, naproxen, olanzapine, curcumin-all sterically hindering the enzyme binding pocket) may be tested for repurposing in IBM [33]. Also, regulating CYLD could be tested as a possible a therapeutic strategy in IBM [63].

The connection between a chronic viral infection and IBM deserves to be investigated further. There are questions waiting to be answered. Which factors are involved in transforming acute viral myositis into chronic inflammatory idiopathic myopathy? And moreover, why do some aged patients develop after a viral infection an IIM, for instance an anti-synthetase syndrome, and others an IBM? For instance, in HIV infection, what conditionate the switch from a PM phenotype to an IBM one? [4]. Serial studies in patients with chronic viral infections and signs of myopathy and/or sarcopenia would probably shed light on this progression, also regarding the progression to immunosenescence, mitochondrial dysfunction and proteinopathy, and the role of TDP-43 in this setting.

Conclusions

TDP-43 is important in preventing the dsRNA-induced IFN responses [13]. Viral infections may disrupt TDP-43 solubility and function, leading to its accumulation and lack of splicing regulation. The phenotypic differences between several IBM subtypes may be conditioned, besides genetic predisposing factors and age, also by environmental triggers such as certain viruses, and by epigenetic regulators [65]. Malat1 upregulation in certain viral infections may contribute to a protracted immune response [80].

Finding early disease markers and untangling mechanisms after a viral injury could inform whether there is a window of opportunity for the anti-inflammatory therapy, hopefully stop** or slowing the plethora of accompanying proteostasis, mitochondrial, and metabolic defects. Certain viruses, high viremia, coinfections, reactivation of latent viruses, and post-acute expansion of cytotoxic T cells may all contribute to IBM, mainly in an age-shaped immune landscape, with CD8+ T cells with IFN-ƴ production. In most such cases, the virally induced senescent, IFN-ƴ producing cytotoxic CD8+ T cells are the ones involved in IBM, in a genetically predisposed host. Immunophenoty** IBM patients to identify elevated CD8+ CD57+ populations may help stratify patients with prognostic and possibly therapeutic implications [84]. Identifying pathogenic mechanisms may lead to the identification of potential new treatments or to drug repurposing to improve the outcome in this debilitating disease.