Murine models of idiopathic inflammatory myopathy

Abstract Idiopathic inflammatory myopathies (IIMs) are characterized by inflammation of muscles and other organs. Several myositis-specific autoantibodies (MSAs) have been identified in IIMs and were found to be associated with distinct clinical features. Although MSAs are valuable for the diagnosis of IIMs, the pathogenic roles of these antibodies remain unknown. To investigate the pathogenesis of IIMs, several animal models of experimental myositis have been established. Classical murine models of autoimmune myositis, experimental autoimmune myositis, and C protein-induced myositis are established by immunization with muscle-specific antigens, myosin, and skeletal C protein, respectively. Furthermore, a murine model of experimental myositis was generated by immunization with a murine recombinant histidyl-tRNA synthetase, Jo-1, in which muscle and lung inflammation reflecting anti-synthetase syndrome are induced depending on acquired immunity. Recently, the transfer of human IgGs from patients with immune-mediated necrotizing myopathy, comprising anti-signal recognition particles and anti-3-hydroxy-3-methylglutaryl coenzyme A reductase antibodies, was found to induce complement-mediated myositis in recipient mice. CD8+ T cell-mediated myositis can be established depending on autoimmunity against transcriptional intermediary factor 1γ (TIF1γ), an autoantigen for MSAs induced by recombinant human TIF1γ immunization. These new murine models reflecting MSA-related IIMs are useful tools for accurately understanding the pathological mechanisms underlying IIMs.

DM is distinguished from other IIMs based on its specific cutaneous manifestations, heliotrope rash, and Gottron papules/signs. Perifascicular atrophy is the hallmark histopathological feature of DM, along with the sarcoplasmic expression of myxovirus resistance protein A (MxA) and sarcolemmal membrane attack complex deposition [2]. Conversely, IMNM is characterized by relatively severe proximal weakness and myofiber necrosis with minimal inflammatory cell infiltration on muscle biopsy [3], the hallmark histological feature of IBM.
DM, IMNM, and ASS are characterized by the presence of myositis-specific autoantibodies (MSAs). Several MSAs have been identified in patients with IIMs and are reportedly associated with distinct clinical features [4][5][6]. Almost all MSAs have been identified in patients with DM: anti-Mi-2, anti-melanoma differentiation-associated gene 5 (MDA5), anti-transcriptional intermediary factor 1c (TIF1c), anti-nuclear matrix protein 2 (NXP2), and antismall ubiquitin-like modifier activating enzyme (SAE) antibodies, each of which was found to be associated with characteristic clinical features [6]. Anti-MDA5 antibody exhibits high specificity for clinically amyopathic DM, presenting with rapidly progressive interstitial lung disease (ILD) [7]. Anti-TIF1c antibody has been detected in juvenile and adult patients with DM and is closely correlated with internal malignancies, especially in elderly patients [8]. Previously, we reported that the anti-NXP2 antibody, also detected in juvenile and adult patients with IIMs, is characterized by extensive muscular involvement but occasionally lacks a typical rash [9]. In contrast, anti-SAE antibody-positive patients with DM frequently display extensive rash [10].
Although MSAs are useful for diagnosing and prognostic estimation of IIMs, their pathogenic role remains unknown, and it remains to be determined whether autoantibodies are simple disease biomarkers or true pathogenic effectors.
To comprehensively clarify the pathogenesis of IIMs and develop novel therapeutic strategies, numerous animal models of IIMs have been established, including infectious, genetic, and antigeninduced models [15]. Although previously established murine models of experimental autoimmune myositis depend on immune responses against muscle tissue-specific antigens [16,17], we recently established a murine model of myositis by immunization with TIF1c, an autoantigen for a DM-specific MSA [18]. Herein, we review the history and evolution of murine models of myositis and summarize these models in Table 1.

Experimental autoimmune myositis (EAM)
The classical murine model of autoimmune myositis is known as EAM [16,19]. Using SJL/J mice, EAM can be induced by repeated immunization with muscle homogenate or partially purified myosin of the skeletal muscle with complete Freund's adjuvant (CFA) containing mycobacteria on four occasions, performed at weekly intervals. Myositis is induced in the quadriceps femoris muscles of immunized mice, in which CD4 þ T cells predominantly infiltrate the muscle tissue [20]. Immunoblot analysis of immunoglobulin G (IgG) from EAM mice revealed the presence of antibodies against a wide variety of muscle proteins. The transfer of T cells or IgGs collected from EAM mice was shown to cause myositis in naïve mice [16].
However, SJL/J mice exhibit a defect in the dysferlin gene and develop spontaneous muscle fiber degeneration and inflammatory cell infiltration independent of EAM immunization [21]. Another subsequent study has shown that severe EAM can be induced also in BALB/c mice by a twice immunization with increased administration doses of rat myosin and pertussis toxin (PT) [22].

Genetically modified mice with upregulation of major histocompatibility complex (MHC) classI in the skeletal muscles develop myositis
In the human IIMs, upregulation of MHC class I on the surface of muscle cells is observed in an early phase. Nagaraju et al. [23] have reported that transgenic mice with conditional upregulation of MHC class I expression in the skeletal muscles develop clinically muscle weakness depending on muscle cell damage and muscle inflammation. The mice often produce autoantibodies including anti-histidyl-tRNA synthetase antibodies. The mechanism of the myositis has been described to be depending on activation of the endoplasmic reticulum stress response [24]. Even though the transgenic mice are not widely available, this murine model well reflects human IIMs.

C Protein-induced myositis (CIM)
Subsequently, biochemical studies have suggested that skeletal C protein, a myosin-binding protein, is the major immunogen in the myosin fraction that can induce EAM in rats [25]. Sugihara et al. [17] established modified and improved EAM, i.e., CIM, in C57BL/6 (B6) mice by administering a single subcutaneous injection of CFA-conjugated human or murine C protein fragments, along with an intraperitoneal injection of pertussis toxin (PT), another adjuvant. CIM is an improved murine model of autoimmune myositis when compared with classic EAM, given that CIM can be induced even in B6 mice, the most commonly used mouse strain for gene modification. The C protein-immunized mice develop myositis 2-3 weeks after immunization. Immunohistochemistry studies revealed that CD8 þ T cells localize to the endomysial site of muscle tissues of CIM mice when compared with perimysial and perivascular sites. In addition, Sugihara et al. [26] found that major histocompatibility complex (MHC) class I knockout (KO) mice failed to develop CIM, and perforin deficiency, a cytolytic protein in cytotoxic CD8 þ T cell granules, inhibited muscle fiber injury in CIM. Moreover, the adoptive transfer of CD8 þ T cells collected from CIM mice could induce more severe muscle fiber injury in recipient B6 mice than in CD4 þ T cells. Transferring bone marrow-derived dendritic cells presenting a CD8 epitope peptide from the C protein could induce myositis in recipient mice [27]. These results indicate that CIM is characterized by CD8 þ T cell-mediated muscle fiber injury.

Jo-1-induced ASS model
Another murine model of experimental myositis was generated by immunization with purified epitopic peptides derived from histidyl-tRNA synthetase (HisRS, Jo-1) with CFA [38]. Subcutaneous injections of murine HisRS induced local muscle and lung inflammation in congenic B6 mice (B6.G7). Histological studies of muscle tissues revealed diverse infiltration patterns, with perimysial/epimysial inflammation in a perivascular distribution, endomysial inflammation, and muscle fiber invasion/degeneration. The immunized mice showed a HisRS-specific CD4 þ T cell response and antibody production. However, DO11.10/Rag-2-KO mice with abrogated HisRS-specific B and T cell responses demonstrated significant muscle inflammation after immunization with HisRS [39]. Moreover, C3H/HeJ (TLR4-KO) mice lacking significant anti-HisRS antibody levels exhibited muscle inflammation induced by immunization with HisRS. These findings support a key role for innate immune responses, but not HisRS-specific autoimmunity, in a HisRSinduced model of myositis. However, analysis of bronchoalveolar lavage fluid samples collected from anti-Jo-1 antibody-positive patients with IIM revealed that HisRS-specific CD4 þ T cells might be present within the lungs of patients [40].

Transfer of human IgGs from patients with IMNM induces muscle deficiency
Anti-SRP and anti-HMGCR antibodies are often detected as MSAs specific to IMNM. Bergua et al. [41] found that transferring human IgGs from patients with IMNM, comprising anti-SRP or HMGCR antibodies, provoked muscle deficiency and myofiber necrosis in recipient B6 wild-type or Rag-2-KO mice. Muscle deficiency tends to be less severe in mice receiving IgGs from anti-HMGCR antibody-positive patients than in those receiving IgGs from anti-SRP antibody-positive patients. These phenomena may reflect the features of IMNM [3]. The myopathy severity in IMNM IgG-transferred complement C3-KO mice was reduced, whereas supplementation with human complement could increase severity. This model suggested that patient-derived anti-SRP and anti-HMGCR antibodies are pathogenic toward muscles in vivo through a complement-mediated mechanism. Accordingly, anti-SRP and anti-HMGCR autoantibodies are not only diagnostic biomarkers of IMNM but also direct pathogenic factors.

Tif1c-induced myositis model (TIM)
Anti-TIF1c antibody-positive DM is known to be associated with malignancies, particularly in elderly patients [8], and may also be associated with pregnancy [42]. TIF1c is frequently mutated or overexpressed in tumors [43,44] and is overexpressed in the embryo and mammary epithelial cells during pregnancy [45,46]. Therefore, it can be speculated that cancer and pregnancy trigger autoimmunity against TIF1c and autoimmunity to TIF1c can induce the establishment of myositis. However, elucidating the suggested mechanisms can be challenging.
We established TIM in B6 mice by administering weekly subcutaneous injections of recombinant human TIF1cprotein emulsified in CFA four times, along with an intraperitoneal injection of PT [18]. The immunized mice developed TIF1c-specific T cells and anti-human and murine TIF1c antibodies, resulting in myositis in the hamstrings and quadriceps two weeks after the last immunization. Histological studies revealed atrophy and necrosis of muscle fibers, accompanied by infiltrating mononuclear cells in the perifascicular and endomysial sites of muscle tissues.
Immunohistochemistry studies showed that T cells, particularly CD8 þ T cells, predominantly infiltrated and adhered to the muscle fibers, which upregulated the expression of MHC class I and type I interferon (IFN)-responsive molecule Mx1. Beta 2 microglobulin-KO mice lacking MHC class I expression, perforin-KO mice, and anti-CD8 depleting antibody-treated mice rarely developed TIM. Furthermore, the adoptive transfer of CD8 þ T cells from TIM mice could induce myositis in recipient B6 mice, whereas the transfer of CD4 þ T cells failed to exhibit this effect. Collectively, CD8 þ T cells were identified as pathogenic in TIM. In contrast, lMT mice, which completely lack B-cell lineages, developed myositis, and the adoptive transfer of IgGs collected from TIM mice failed to induce myositis in recipient mice. These results indicate that B cells and autoantibodies are not essential for developing TIM. In other words, the anti-TIF1c antibody detected in patients with DM may be a diagnostic biomarker but not a direct pathogenic factor, unlike anti-SRP/HMGCR antibodies in patients with IMNM. Accordingly, TIM, which is dependent on autoimmunity against TIF1c, was mediated by TIF1c-specific CD8 þ T cells but not TIF1c-specific CD4 þ T cells, B cells, and autoantibodies ( Figure 1).
Type I IFNs are suspected to be involved in the pathogenesis of DM [47,48]. We have reported that IFN-a/b receptor deficiency partially inhibited the development of TIM. In addition, treatment with tofacitinib, a Janus kinase (JAK) inhibitor for various cytokine receptor signaling, including type I IFN signaling, could treat established TIM without suppressing TIF1c-specific T cell response and autoantibody production. Therefore, JAK inhibitors could afford a good treatment strategy for DM that does not involve acquired immunity, which is important for protecting against pathogeninduced infection.
While EAM and CIM completely depend on immune responses specific to muscular antigens, myosin, and C protein [16,17], TIM induction relies on autoimmunity against a ubiquitous intracellular molecule, TIF1c, an autoantigen for MSAs. This new murine model of experimental myositis, which closely reflects the pathological mechanisms of DM, could be a useful tool for developing novel treatment strategies against anti-TIF1c antibody-positive DM.

Conclusion
In recent years, murine models of myositis for each MSA, such as TIM, have been developed (Table 1). These models suggest that each MSA-related IIM exhibits distinct pathogenesis; thus, the establishment of myositis models for other MSAs can be expected.

Disclosure statement
No potential conflict of interest was reported by the author(s).