The juvenile idiopathic inflammatory myopathies: pathogenesis, clinical and autoantibody phenotypes, and outcomes

Abstract

The aim of this review is to summarize recent advances in the understanding of the clinical and autoantibody phenotypes, their associated outcomes, and the pathogenesis of the juvenile idiopathic inflammatory myopathies (JIIMs). The major clinical and autoantibody phenotypes in children have many features similar to those in adults, and each has distinct demographic and clinical features and associated outcomes. The most common myositis autoantibodies in JIIM patients are anti-p155/140, anti-MJ, and anti-MDA5. Higher mortality has been associated with overlap myositis as well as with the presence of anti-synthetase and anti-MDA5 autoantibodies; a chronic illness course and lipodystrophy have been associated with anti-p155/140 autoantibodies; and calcinosis has been associated with anti-MJ autoantibodies. Histologic abnormalities of JIIMs detectable on muscle biopsy have also been correlated with myositis-specific autoantibodies; for example, patients with anti-MDA5 show low levels of inflammatory infiltrate and muscle damage on biopsy. The first genome-wide association study of adult and juvenile dermatomyositis revealed three novel genetic associations, BLK, PLCL1, and CCL21, and confirmed that the human leukocyte antigen region is the primary risk region for juvenile dermatomyositis. Here we review the well-established pathogenic processes in JIIMs, including the type 1 interferon and endoplasmic reticulum stress pathways. Several novel JIIM-associated inflammatory mediators, such as the innate immune system proteins, myeloid-related peptide 8/14, galectin 9, and eotaxin, have emerged as promising biomarkers of disease. Advances in our understanding of the phenotypes and pathophysiology of the JIIMs are leading to better tools to help clinicians stratify and treat these heterogeneous disorders.

Introduction

The juvenile idiopathic inflammatory myopathies (JIIMs) are heterogeneous, systemic autoimmune diseases characterized by weakness, chronic inflammation of skeletal muscles, and typical skin rashes (Gottron’s papules or heliotrope rash) with onset during childhood. The criteria established by Bohan and Peter based on these features [1], as well as the presence of elevated serum levels of muscle enzymes or increased electrical activity in the muscle detected by electromyography, have been used to diagnose these disorders. However, new classification criteria have recently been developed and validated [2]. Evidence from several large JIIM registry studies has led to increased understanding of the spectrum of phenotypes associated with JIIMs, based on either clinicopathologic features or the presence of autoantibodies found almost exclusively in patients with myositis (known as myositis autoantibodies) [3–5] (Table 1). This subclassification of JIIMs enables better understanding of the disease in patients who share similar demographic and clinical features, laboratory abnormalities, and outcomes. It is now also recognized that the same clinical and autoantibody subgroups are largely present in both children and adults with idiopathic inflammatory myopathies (IIMs), albeit in differing proportions and with minor variations in their associated features [6–8] (Table 1).

Table 1.

Clinicopathologic and serologic classification of juvenile and adult idiopathic inflammatory myopathies

Clinicopathologic phenotypes	Autoantibody phenotypes
Dermatomyositis	Myositis-specific autoantibodies
Polymyositis	Anti-Jo-1, other anti-synthetases
Myositis with other connective tissue disease	Anti-Mi-2
Inclusion body myositis	Anti-signal recognition particle
Cancer-associated myositis	Anti-p155/140 (TIF-1)
Immune-mediated necrotizing myositis	Anti-MJ (NXP-2)
Focal/nodular myositis	Anti-CADM-140 (MDA5)
Ocular/orbital myositis	Anti-200/100 (HMG-CoA reductase)
Granulomatous myositis	Myositis-associated autoantibodies
Eosinophilic myositis	Anti-U1-, U2-, U3-, and U5-RNP
Macrophagic myofasciitis	Anti-PM-Scl
	Anti-Ku
	Anti-Ro
	Anti-SUMO/SAE
	Anti-43kD (cN1A)*
	Myositis autoantibody negative

^*To date, this autoantibody has not been identified in children with myositis.

TIF-1, transcriptional intermediary factor 1; RNP, ribonucleoprotein.

Clinicopathologic classification of JIIMs

The most common clinical phenotype of myositis in children, with a prevalence of approximately 80% of all patients with JIIMs, is juvenile dermatomyositis (JDM). Patients with JDM are often the youngest among those with JIIMs, with a median age at onset of 7.4 years [8, 9]. In addition to symmetric proximal and axial muscle weakness and characteristic Gottron’s papules or heliotrope rash, JDM patients often have other photosensitive rashes, including malar erythema, linear extensor erythema, and V- or shawl-sign rashes [8]. JDM is also distinguished by small-vessel vasculopathy, evidenced by dilated and tortuous periungual capillaries, and an immune-mediated injury resulting in capillary loss [8, 10–14]. Approximately 20–47% of JDM patients have calcinosis, which consists of subcutaneous deposits of calcium carbonate apatite in the tissues and muscles [11, 15–18] (Fig. 1).

Fig. 1.

Juvenile dermatomyositis (JDM) presents with characteristic rash and symmetric muscle weakness in the extremities. Juvenile polymyositis (JPM) presents with more severe muscle weakness and frequent cardiac involvement but the rashes of JDM are absent. JPM is more common in black female patients. Patients with overlap myositis meet the criteria for myositis as well as another autoimmune disease. Common illness features include interstitial lung disease, Raynaud’s phenomenon, arthritis, and sclerodactyly [8]. Clinically amyopathic JDM presents primarily with characteristic skin rashes, with mild or no muscle weakness [24].

Overlap myositis, in which patients meet the criteria for JIIMs as well as another autoimmune disease, is the second most common clinical phenotype of JIIMs, occurring in 6–11% of JIIM patients with a peak age at onset before adolescence [8, 10]. Patients with overlap myositis are more likely to have JDM than juvenile polymyositis (JPM). The most common overlapping autoimmune conditions are systemic lupus erythematosus, juvenile idiopathic arthritis, systemic sclerosis, and localized scleroderma [8]. In this subgroup of disorders, Raynaud’s phenomenon, interstitial lung disease (ILD), arthritis, and malar rash are frequently observed, and sclerodactyly and dysphagia are also common features [8, 19] (Fig. 1). Patients with overlap myositis are more often not white and generally have myositis-associated autoantibodies (MAAs) [8, 19] (Fig. 2). Organ-specific autoantibodies, including those for autoimmune thyroid disease, hepatitis, type I diabetes, and celiac disease, have been observed in 2–15% of patients with JDM. However, these autoantibodies did not correlate with overt organ-specific overlapping autoimmune diseases in one study [20], although patients with autoantibodies had more severe disease and lower body mass index in a second cohort [21].

Fig. 2.

Distribution of myositis autoantibodies by clinical subgroup in patients with juvenile idiopathic inflammatory myopathies. Data are from Rider and Miller [6] and Shah et al. [8]. JDM, juvenile dermatomyositis; JPM, juvenile polymyositis; JCTM, JDM overlapping with another connective tissue disease.

JPM is observed in 4–8% of JIIM patients, has a peak onset during adolescence, and is characterized by proximal and distal muscle weakness, frequent falling episodes, myalgias, and increased creatine kinase (CK) levels [8, 22] (Fig. 1). JPM has a more severe disease onset than JDM, often with weight loss and Raynaud’s phenomenon; cardiac involvement occurs in 35% of patients [8]. Furthermore, patients with JPM do not have Gottron’s papules or heliotrope rash. The pathologic findings also differ from those of JDM, typically with endomysial infiltrates in affected muscles [8, 10, 12, 22]. Because JPM is often misdiagnosed in patients who actually have other noninflammatory myopathies, particularly muscular dystrophies, a muscle biopsy is required for diagnosis [23].

Clinically amyopathic dermatomyositis (CADM) is occasionally observed in children; skin rashes are present for at least 6 months in patients who have no detectable weakness but who have laboratory evidence of muscle inflammation [e.g. elevated serum muscle enzyme levels or abnormal electromyogram, muscle biopsy, or muscle magnetic resonance imaging (MRI) results]. In a retrospective review of 68 cases of juvenile-onset CADM, encompassing both dermatomyositis (DM) sine myositis and hypomyopathic DM, 26% of subjects subsequently developed classical JDM; disease progression occurred up to several years later [24, 25]. Only 4% of these subjects had an elevated serum CK level at onset. Of those with a normal CK level, only a few had abnormal electromyography, muscle biopsy, or MRI results, suggesting that the cutaneous disease can be treated symptomatically with close clinical monitoring for progression to muscle involvement [24]. The manifestations that are often associated with CADM in adults are rare in children (Fig. 1). Calcinosis was reported in only 4%, and none presented with ulcerations, ILD, or malignancy [24, 26]. The outcome of CADM appears to be good in childhood; for example, of 24 patients with juvenile CADM, 65% recovered without therapy, and five of 10 patients who received systemic therapy entered remission [27].

Serologic classification of JIIMs

JIIMs can also be classified based on the presence of two classes of myositis autoantibodies (Table 1), myositis-specific autoantibodies (MSAs), which are present almost exclusively in patients with myositis, and MAAs, which are present in patients with myositis and in those with other autoimmune diseases. These myositis autoantibodies define more homogeneous groups of patients with similar clinical features, responses to therapy, and prognoses. At least one myositis autoantibody can be identified in approximately 70% of JIIM patients [8]. Of the autoantibodies studied in both children and adults, generally the same myositis autoantibodies are present, but they differ in frequency between children and adults. For example, anti-p155/140 and anti-MJ autoantibodies appear to be more prevalent in children than in adults with DM; conversely, the anti-tRNA synthetase autoantibodies, which are found in only 5% of JIIM patients, are present in 25–40% of adult IIM patients [8]. There are many similarities in terms of clinical and demographic features, laboratory findings, and prognosis in pediatric and adult patients with IIMs who have the same myositis autoantibodies [7]. Fig. 2 shows the distribution of myositis autoantibodies by clinical subgroup of JIIMs.

The most common MSAs in JIIMs are anti-p155/140 and anti-MJ autoantibodies, which are associated primarily with JDM [7]. For patients with anti-p155/140 autoantibodies, 155-kDa and 140-kDa proteins have been identified as transcriptional intermediary factor 1 (TIF-1) γ and TIF-1α, respectively. Adult DM patients also react to both TIF-1γ and TIF-1α, and in a large Japanese population, an associated malignancy was more common in adults who reacted to both proteins [28]. In adult DM patients from the USA, anti-TIF-1γ and anti-MJ autoantibodies were frequently associated with malignancy [29]. An association with malignancy has not been reported in JDM patients with these autoantibodies.

Anti-p155/140 autoantibodies are present in 23–30% of JIIM patients, particularly those with JDM or overlap myositis with JDM, but they are not present in patients with JPM. Most JIIM patients with anti-p155/140 autoantibodies are Caucasian, and the peak age at onset is approximately 7 years. Anti-p155/140 autoantibodies are associated with extensive photosensitive skin rashes, including Gottron’s papules, malar rash, V- and shawl-sign rashes, and linear extensor erythema (Fig. 3). Erythroderma, reported in 15% of patients, is more common in those with anti-p155/140 autoantibodies [7]. In patients in the UK JDM registry, anti-p155/140 autoantibodies were associated with skin ulceration and edema [30]. Anti-p155/140 autoantibodies were also associated with generalized lipodystrophy [31] and a chronic course of disease [32]. A suggestion for rewording this: Patients exposed to the average and highest ultraviolet light index (based on the geographic location of their residence in the USA) were more likely to have anti-p155/140 autoantibodies than either anti-MJ autoantibodies or no identified myositis autoantibody, within 6 months of illness onset. This suggests that ultraviolet light might be an environmental factor with a role in the development of anti-p155/140 autoantibodies [33].

Fig. 3.

The myositis autoantibodies seen most frequently in children with juvenile idiopathic inflammatory myopathies (JIIMs) are anti-p155/140, anti-MJ, and anti-MDA5 autoantibodies. These frequencies differ from those seen most frequently in adults. Anti-p155/140 autoantibodies are present in 23–30% of JIIM patients, particularly in those with juvenile dermatomyositis (JDM) or overlap myositis with JDM. Most patients with anti-p155/140 autoantibodies are white and have extensive photosensitive skin rashes. Anti-p155/140 is also associated with a chronic course of illness and generalized lipodystrophy [7, 30–32]. Anti-MJ autoantibodies are present in 12–23% of JIIM patients and are seen primarily in those with JDM. Most patients are white. Anti-MJ autoantibodies are associated with muscle cramps, muscle atrophy, joint contractures, dysphonia, and an absence of truncal rashes. Patients with anti-MJ autoantibodies tended to be weaker and have decreased physical function [7, 34, 35]. Anti-MDA5 autoantibodies were present in 33% of a Japanese cohort but only 7% of a UK cohort. Interstitial lung disease was more common in both cohorts with this autoantibody, compared to patients without these antibodies. In the UK patients, other common features included oral and cutaneous ulceration, arthritis, and milder muscle disease, which was similar to findings in US and European adult cohorts with MDA5 autoantibodies [36, 37].

Anti-MJ autoantibodies, which target the nuclear matrix protein NXP2, are present in 12–23% of JIIM patients; these autoantibodies are primarily found in patients with JDM but also occasionally in those with JPM or overlap myositis. Most patients with anti-MJ autoantibodies are Caucasian, and the average age at disease onset is 6 years. The clinical features associated with anti-MJ autoantibodies in children include frequent muscle cramps, muscle atrophy, joint contractures, and dysphonia, which are present in 40–60% of patients (Fig. 3). Patients with anti-MJ autoantibodies tend to be weaker than individuals without these autoantibodies, and their physical function is decreased [34]. Although JDM patients with this autoantibody have characteristic Gottron’s papules and malar rash, truncal rashes are absent [35]. Gastrointestinal ulcerations and bleeding are uncommon, reported in only 6–8% of patients, but are more often observed in patients with anti-MJ autoantibodies compared to those in the other autoantibody groups [7]. In the UK registry, 43% of JDM patients with anti-MJ autoantibodies had calcinosis, with an associated odds ratio of 2.1 [34]. However, in a US registry, no association with calcinosis was detected [7]. Illness tends to be more severe in patients with anti-MJ autoantibodies; only 8% of such patients entered remission at 2 years in the UK series [34], and patients with these autoantibodies were hospitalized more often than those with other or no autoantibodies in the US series [7].

Recently, anti-MDA5 autoantibodies were identified in patients with JIIMs. In a series of 54 Japanese JDM patients, 33% had anti-MDA5 autoantibodies [36], whereas these autoantibodies were present in only 7% of patients in the UK JDM registry [37]. There was a high rate of ILD in both cohorts, consistent with results from studies of adult patients with this autoantibody [38, 39]. Among the Japanese patients, anti-MDA5 autoantibodies were found in all eight with rapidly progressive ILD, in 10 of 14 with chronic ILD, but in none of the remaining 32 patients without ILD [36]. In the UK JDM registry, 19% of patients with MDA5 autoantibodies had ILD, typically as nonspecific interstitial pneumonitis or organizing pneumonitis but not rapidly progressive ILD [37] (Fig. 3). In addition, fever and low CK levels were common, but cutaneous ulcers were rare in the Japanese series. Those with rapidly progressive ILD had high serum levels of interleukin (IL)-18, Krebs von den Lungen (KL)-6, and ferritin, and a high mortality rate (88%), with diffuse alveolar damage as the common pathology at autopsy [36]. In the UK registry, other common features in patients with anti-MDA5 autoantibodies included oral and cutaneous ulcerations (71% and 52%, respectively), arthritis (86%), and mild muscle disease; these findings are similar to those reported in US [40] and European adult IIM patients with anti-MDA5 autoantibodies [41]. In the UK group, disease was often inactive after 2 years of follow-up [37].

Of the traditional MSAs [i.e. anti-tRNA synthetases, anti-signal recognition particle (SRP), and anti-Mi-2 autoantibodies], anti-tRNA synthetase autoantibodies are less frequently observed in children than in adults with IIMs. These autoantibodies are found in less than 5% of JDM patients but occur in 9% of patients with JPM and 13% of patients with juvenile overlap myositis, and the peak age at disease onset in this autoantibody subgroup is 14 years [7]. Anti-Jo-1 autoantibodies are the most common of the anti-synthetase autoantibodies. As in adults, children with these autoantibodies frequently have ILD (65%), arthritis (75%), fevers (65%), Raynaud’s phenomenon (32%), and ‘mechanic’s hands’ (32%) [7] (Fig. 4). Among the MSA phenotypes, this group also has the highest mortality rate, with death primarily from ILD [7, 42]. Anti-Mi-2 is a traditional MSA that is associated with JDM and its cutaneous features but, unlike in adults, V- and shawl-sign rashes and cuticular overgrowth are not associated with this autoantibody [7] (Fig. 4).

Fig. 4.

Traditional myositis autoantibodies have similar phenotypes in children and adults with the same autoantibodies. Anti-tRNA synthetase autoantibodies are less common in juvenile than in adult myositis. These autoantibodies are more often seen in patients with juvenile polymyositis or juvenile overlap myositis. Patients with anti-synthetase autoantibodies frequently have interstitial lung disease, arthritis, fever, Raynaud’s phenomenon and mechanic’s hands, and a high mortality [7, 42]. Anti-Mi-2 is associated with juvenile dermatomyositis and its cutaneous features but, unlike in adults, is not associated with V- and shawl-sign rashes or cuticular overgrowth [7]. Anti-signal recognition particle (SRP) autoantibodies are seen primarily in African-American teenage girls with severe immune-mediated necrotizing myopathy who have proximal and distal muscle weakness, frequent falling episodes, Raynaud’s phenomenon, high creatine kinase levels, and a chronic illness course, as well as the need to use a wheelchair. Approximately 50% of pediatric patients with anti-SRP autoantibodies also present with cardiac disease, and, similar to adult patients with that autoantibody, their disease is refractory to many therapies [7, 44]. Although 28% of patients have no identified autoantibodies, they might have currently unrecognized autoantibody phenotypes. These patients have mild disease [7].

Two autoantibodies, anti-SRP and anti-HMG-CoA reductase, are most often associated with an immune-mediated necrotizing myopathy that resembles polymyositis. Here, the predominant pathology is myonecrosis, and there is upregulation of the major histocompatibility class I antigen on non-necrotic muscle fibers. Inflammation may also be present, but it is often limited with a perivascular location [43–45]. Anti-SRP autoantibodies are seen primarily in African-American teenage girls with a severe immune-mediated necrotizing myopathy, who have proximal and distal muscle weakness, frequent falling episodes, Raynaud’s phenomenon, very high CK levels, and a chronic illness course, as well as a need to use a wheelchair [7, 44] (Fig. 4). These patients do not present the characteristic rashes of JDM. Cardiac disease is present in 50% of patients with anti-SRP autoantibodies and, similar to adult IIM patients with these autoantibodies, their disease is refractory to many therapies [44]. Anti-200/100, an MSA that targets HMG-CoA reductase, is associated with an immune-mediated necrotizing myopathy that resembles polymyositis in adults and is more common in patients who have been treated with statins [46]. It has been observed in eight adult patients with disease onset during adolescence, but patients with a history of statin exposure tend to be older [47].

Approximately 28% of patients have no identified MSAs or MAAs. Although disease in this subgroup appears to be mild, these patients likely have several currently unknown autoantibody phenotypes. The strength of the clinical associations and prognostic information from the MSAs suggest that serologic data may be valuable in guiding therapy.

Myositis autoantibodies are most accurately detected by validated protein and RNA immunoprecipitation assays. However, certain autoantibodies, particularly anti-p155/140 and anti-MJ autoantibodies, require additional confirmation by reverse immunoprecipitation–immunoblotting or immunodepletion methods [30, 48, 49]. A line-blot assay is almost as sensitive and specific as protein immunoprecipitation in detecting the traditional myositis autoantibodies [50]. Although enzyme-linked immunosorbent assays detect some myositis autoantibodies, the sensitivity and specificity of several of these assays have not been well defined, and their detection is hampered by the use of varying epitopes, by the low abundance and conformational dependence of epitopes for detection of some of these autoantibodies, and by the possibility that autoantibodies may bind to multiple protein antigens [34, 51–54]. It is notable that MSAs rarely overlap, as each patient shows reactivity to only one myositis autoantigen and may have only one of the myositis autoantibodies. This might be explained by the fact that human leukocyte antigen (HLA) risk associations for one MSA are often protective factors for a different autoantibody [55]. It might also suggest distinct immunopathologic features for each autoantibody subgroup [56].

Associated outcomes

In two US registry studies it was demonstrated that mortality is higher in patients with JIIMs than in the general population [42, 57]. The US pediatric rheumatic disease registry includes approximately 50,000 patients from more than 60 centers who were diagnosed between 1992 and 2001. A total of 662 JDM patients were registered in this registry, with a median follow-up of 7.9 years. Mortality was determined through the social security death index and confirmed by reviewing death certificates; the standardized mortality ratio (SMR) was 2.64 for JDM, with five deaths recorded (0.8%) [57]. In the second US JIIM registry study of 405 patients enrolled from 1989 to 2011 with a mean follow-up of 4.3 years, the SMR for JIIMs was 14.4 with 17 deaths (4.2%), and the SMR for JDM was 8.3 with eight deaths (2.4%); mortality was determined through physician reports, as well as the social security death index with review of selected death certificates [42]. The causes of death included pulmonary (n = 7), gastrointestinal (n = 3), and multisystem failure (n = 3). In multivariable analysis, clinical subgroup (overlap myositis > JPM > JDM) and severity of illness at onset were strong predictors of mortality. Other important predictors at illness onset included the presence of anti-synthetase autoantibodies, ILD, Raynaud’s phenomenon, dysphagia, or weight loss, as well as older age and delay to diagnosis [42].

In several large JDM cohorts, 8–77% of patients had a chronic illness course with active disease requiring immunosuppressive medications for more than 2 years, and 22–60% of patients had a monocyclic course with documented remission within 2 years [17, 58–60]. In a large US registry study of JIIMs, 50–65% of patients had chronic disease, and only 14–25% had a monocyclic disease course, with a trend toward more frequent chronic illness in patients with JPM and overlap myositis compared to those with JDM [32]. Anti-SRP, anti-synthetase, and anti-p155/140 autoantibody subgroups were associated with a chronic illness course [32]. A longer duration of untreated disease [11], higher baseline skin disease activity [14], persistent Gottron’s papules and periungual nailfold capillary changes beyond 3 months after diagnosis [58], the presence of subcutaneous edema on MRI at diagnosis [61], and extensive myopathic and severe arteropathic changes on the initial muscle biopsy [62] were all predictors of a chronic illness course. A shorter delay to diagnosis and lower skin disease activity at diagnosis were associated with a monocyclic course of disease [14]. Additional determinants of chronic illness identified in the above-mentioned recent large US registry study of JIIMs included the presence of anti-p155/140 autoantibodies, or any myositis autoantibodies, severe illness at onset, and a documented infection within 6 months of diagnosis [32].

Calcinosis is defined as the presence of dystrophic calcification in subcutaneous, myofascial, or muscle tissues. It occurs in 12–47% of JDM patients and most often develops several years after illness onset [8, 12, 17, 18]. Diagnosis and treatment delays, greater disease activity (especially in the cardiac or pulmonary systems), a longer duration of active disease, and use of nonsteroidal immunosuppressive agents, perhaps as an indicator of greater disease activity, were risk factors for calcinosis [15, 17, 18, 31, 63]. Demographic factors associated with the development of calcinosis include earlier calendar year of diagnosis (i.e. before 1990), geographic location (a higher rate of calcinosis was observed in patients in South America than in Europe), African-American race, and male gender [11, 18, 64]. A higher rate of calcinosis was observed in patients with disease onset at a younger age in one study [34], whereas older age at onset was found to be a risk factor for calcinosis in another study [11]. Certain myositis autoantibodies (e.g. anti-MJ and anti-PM-Scl) as well as pro-inflammatory cytokine polymorphisms [tumor necrosis factor alpha(TNFα)-308A and IL-1a-889C alleles] have been associated with calcinosis in some investigations [34, 65].

Lipodystrophy, a complication of JDM characterized by progressive loss of subcutaneous fat in a localized, partial (i.e. only in the extremities), or widespread pattern, occurs in up to 10% of JDM patients. Lipodystrophy has been associated with calcinosis, muscle atrophy, joint contractures, and facial rash. It should be recognized early because of its association with metabolic sequelae, such as insulin resistance, diabetes, and hyperlipidemia, with rates proportional to the degree of fat loss [31]. Total lipodystrophy has been associated with anti-p155/140 autoantibodies [31].

Etiopathogenesis of JIIMs

Although the exact etiologic basis for JIIMs remains unknown, a number of environmental and genetic factors have been temporally associated with disease onset. The most common form of JIIM, JDM, has several features typical of an autoimmune disease, including the regular occurrence of autoantibodies, evidence of antigen-driven clonal B and T cell expansion within the inflamed muscle, and a strong genetic association with immune-related genes of the HLA region [66]. Although much of this evidence indicates dysfunction of the adaptive immune system, it is increasingly clear that abnormalities of the innate immune system and nonimmune mechanisms also contribute to the disease pathogenesis [67]. Given that the prevalence of JIIMs is approximately one-fifth that of adult forms of the disease, it is reasonable that some of our knowledge of disease pathogenesis is derived from studies of adult IIM. Large cohort studies of JIIMs with linked biologic material, as well as increased international collaboration [5], are increasing understanding of the specific mechanisms involved in the juvenile forms of IIMs.

Genetics

JIIMs are complex genetic disorders with multiple associated genetic regions or loci, each of which contributes differently to the risk of developing disease. The international myositis genetics consortium MYOGEN recently published the first genome-wide analysis of IIM, which included 473 patients with JDM, confirming that the HLA region was the strongest genetic risk region for both adult and juvenile DM [68]. Although the statistical power of the study was limited at the genome-wide level, the investigators used a candidate gene approach to examine loci associated with other autoimmune diseases and identified three new genetic associations with myositis: CCL21 [chemokine (C-C motif) ligand 21], PLCL1 (phospholipase C-like 1), and BLK (B lymphoid kinase) (Table 2).

Table 2.

New genetic loci associated with idiopathic inflammatory myopathies (IIMs) identified by genome-wide association studies [68]

Gene	Function	Mechanistic link with IIM
CCL21	Homeostatic chemokine expressed by T, B, and dendritic cells; important for recruiting lymphocytes to secondary lymphoid organs	Might influence recruitment of inflammatory cells to ectopic lymphoid structures in muscle or skin
PLCL1	Intermediate in inositol phospholipid-based intracellular signaling cascade	Unknown. Single-nucleotide polymorphism previously linked with allergic disease [88]
BLK	Tyrosine kinase from the Src family; known to influence B cell receptor signaling and B cell proliferation	Polymorphisms associated with autoimmune disease are known to reduce BLK expression. Emerging data suggest that this may lead to expansion of circulating innate-like B cells termed B-1 cells and increased production of autoantibodies [89]

BLK, B lymphoid kinase.

Although JDM is a polygenic disorder, the study of monogenic forms of myositis can offer valuable insights into disease pathogenesis. Patients with chronic atypical neutrophilic dermatosis with lipodystrophy and elevated temperature (CANDLE) syndrome share clinical and immunologic features with JDM patients, including myositis, lipodystrophy, and a prominent interferon (IFN) signature detectable in blood samples [69]. CANDLE syndrome is associated with mutations in genes encoding the immunoproteasome, a complex cellular apparatus that processes and degrades intracellular proteins. Abnormalities of protein folding and trafficking have been associated with DM [67], but it is still not clear how the specific subcellular events associated with protein handling lead to distinct phenotypes of JIIMs and myositis associated with CANDLE syndrome.

Immune mechanisms

One of the best-characterized abnormalities in adult and juvenile DM is excess IFN production in blood and muscle compartments. Plasmacytoid dendritic cells are thought to be activated by viral nucleic acids or self DNA and to release large amounts of type I IFNs (IFNα and IFNβ), leading to immune cell activation and vasculopathy [70]. Plasmacytoid dendritic cells are detectable in JDM muscle and are associated with other immune cells within lymphoid aggregates [71]. Free IFNα is tightly regulated by the immune system and is therefore rarely detected in serum. Instead, expression of genes induced by IFN, termed IFN-responsive elements, is often used as a surrogate measure of excessive IFN signaling in vivo. Bilgic et al. have shown that IFNα increases expression of the cytokine IL-6 and a range of chemokines, including monocyte chemoattractant protein (MCP)-1, MCP-2, and CXCL10 (IP-10), and that this signature correlates with disease activity in JDM [72]. Serum from patients with autoantibodies against the autoantigens Jo1, as well as Ro, La, Smith, and ribonucleoprotein (RNP), induced excessive IFN signaling in a reporter cell line [73, 74]. The data indicating that IFNα has an important pathogenic role in IIM suggest that blocking this pathway may provide a novel therapeutic approach in JDM. In a Phase Ib proof-of-concept study, adult DM and polymyositis patients were treated with sifalimumab, a monoclonal antibody against IFNα, and the IFN gene signature was examined in blood and muscle biopsies before and after treatment [75]. The IFN gene signature was suppressed by a median of 53–66% across three time points (days 28, 56, and 98) in blood and 47% at day 98 in muscle specimens after sifalimumab administration. Patients with >15% improvement in manual muscle testing showed greater IFN gene signature suppression than individuals whose manual muscle testing scores failed to improve. These initial results are promising and suggest that neutralizing IFNα might reduce muscle disease activity in IIM; however, larger studies are needed to confirm the efficacy of this approach.

Several other circulating inflammatory mediators have been implicated in the pathology of JIIMs (Table 3). Sanner and colleagues compared circulating chemokine levels in JDM patients with longstanding disease (median follow-up of 16.8 years) and in age- and gender-matched control subjects. Patients with higher levels of eotaxin and MCP-1 at follow-up had evidence of early organ damage, as assessed by the Myositis Damage Index, 1 year after diagnosis [76]. In addition, there was evidence of cardiac dysfunction detected by echocardiography at long-term follow-up in patients with high serum eotaxin and MCP-1 levels [76, 77]. It is unclear whether these chemokines contribute directly to end-organ damage or reflect general immune activation.

Table 3.

Summary of inflammatory mediators implicated in the pathogenesis of juvenile idiopathic inflammatory myopathies (JIIMs)

Mediator	Physiologic function	Role in autoimmunity/IIM	Reference
Type 1 interferons (IFNα and IFNβ)	Robust anti-viral response, activates T and NK cells, promotes differentiation of plasma B cells	Upregulates muscle MHC class I expression; induces autoantibody production; causes vascular damage	[90]
Eotaxin (CCL-11)	Recruits eosinophils via CCR3	Associated with cardiac dysfunction, mechanism unclear	[77, 91]
Monocyte chemoattractant protein (MCP)-1 (CCL2)	Recruits monocytes and macrophages via CCR2	Expressed on muscle endothelium; attracts monocytes and T cells into muscle	[72, 91–93]
Myeloid-related peptide 8/14	Released by monocytes and neutrophils; promotes cell adhesion to endothelium	Activates myoblasts leading to IL-6 and MCP-1 production and increased apoptosis	[92, 94]
IL-6	Potent pro-inflammatory cytokine that drives systemic inflammatory response; B cell growth factor	Complex actions on muscle; can promote myogenesis, but at high concentrations leads to muscle atrophy; serum levels correlate with disease activity in JDM	[72]
Galectin 9	May play an immunoregulatory role suppressing Th1 responses	Serum levels correlate with muscle disease activity; may be secreted by inflamed endothelium	[78, 79]
sTNFRII	Binds TNFα and TNFβ; may have inhibitory role in acute inflammation as it competes with membrane-bound TNFR	Serum levels correlate with muscle disease activity; elevated levels in active JDM compared to remission; sTNFRII may act as a reservoir for TNFα in chronic inflammation	[78, 95, 96]

MHC, major histocompatibility class; JDM, juvenile dermatomyositis; TNF, tumor necrosis factor; TNFR, TNF receptor; sTNFRII, soluble TNFR 2; NK, natural killer; IIM, idiopathic inflammatory myopathy.

It can be difficult to assess myositis clinically and to distinguish active disease from damage, as well as to differentiate it from drug-induced muscle atrophy or weakness as a result of disuse. Thus, several research groups have considered using the inflammatory mediators as potential biomarkers of disease activity, response to treatment, and outcome. Interferon gamma-inducible protein (IP)-10, IL-6, and MCP-1 are among the cytokines and chemokines most frequently studied. In two recent studies examining more than 40 circulating inflammatory mediators in JDM, the association between IP-10 and active disease was confirmed [78, 79]. The authors of this study also identified galectin 9 and TNF receptor type II as novel biomarkers of disease activity in JDM. These novel markers are currently being investigated in an independent validation cohort and, if their importance is confirmed, could provide a valuable tool to improve the clinical management of patients with JDM.

Muscle pathology in JIIMs

The muscle is the target of inflammation in all forms of JIIMs; however, muscle biopsies from patients with JDM, JPM, and juvenile necrotizing autoimmune myopathies have distinct pathologic features (Table 4 and Fig. 5). As histologic findings can vary between patients with the same phenotype of JIIMs, standardized score tools are needed to interpret muscle biopsies. An international consensus group of pediatric rheumatologists and histopathologists proposed a scoring system for JDM muscle biopsies based on four domains: vascular, inflammatory, muscular, and connective tissue [80]. This tool was retested in an independent cohort, and the inflammatory and muscular domains were found to be the most reliable and most strongly correlated with clinical measures of disease activity, including manual muscle testing, the Childhood Myositis Activity Scale, and Physician Global Activity [81]. The severity of histologic changes on biopsy also closely correlated with specific MSAs; for example, it was found that patients with anti-MDA5 autoantibodies had milder muscle disease and fewer histologic abnormalities [37] (Nistala, Yasin, and Wedderburn, unpublished data). These data support the hypothesis that each MSA reflects a distinct pathologic process and thereby influences the type and severity of muscle inflammation detected on biopsy. Ongoing work will examine the value of the muscle biopsy as a predictor of treatment response and long-term disease outcomes.

Table 4.

Histologic abnormalities in juvenile idiopathic inflammatory myopathies (JIIMs) [43, 80, 81]

JIIM subtype	Classic histologic findingsa	Cellular infiltrate
JDM	Perifascicular muscle fiber atrophy, capillary drop out, and some regenerating fibers; MHC class I overexpression on myofibers; tubuloreticular inclusions in endothelial cells on electron microscopy	Plasmacytoid dendritic cell, macrophage, B cell, and T cell infiltration; in some cases, cellular aggregates resemble ectopic lymphoid structures
JPM	Endomysial mononuclear cell infiltration of non-necrotic muscle fibers; HLA class I overexpression on myofibers	Predominant CD8 T cell infiltrate
Juvenile immune-mediated necrotizing myopathy	Severe muscle necrosis; little or no inflammation; occasional upregulation of MHC class I complex on myofibers	Sparse

MHC, major histocompatibility class; HLA, human leukocyte antigen; JDM, juvenile dermatomyositis; JPM, juvenile polymyositis.

^aThese are the ’classic‘ immunohistochemical patterns in each of these clinical subgroups, but a spectrum of pathologic findings may be seen across subgroups for JDM and JPM.

Fig. 5.

Muscle biopsies from patients with juvenile dermatomyositis (JDM; magnification ×10) and juvenile immune-mediated necrotizing myopathy [anti-signal recognition particle (SRP)] were stained with hematoxylin and eosin (H&E; a and d), major histocompatability class I (MHC I; b and e), and neonatal myosin (c and f). a) Perivascular cellular inflammation (long arrow) and perifascicular atrophy (short arrow). b) Diffuse MHC I staining on the sarcolemma and the sarcoplasm of many fibers. c) Neonatal myosin stains for damaged muscle fibers that are regenerating in the perifascicular zone (arrow). d) Fiber size variation and smaller pale necrotic muscle fibers (large arrows) surrounded by infiltrating macrophages and scattered atrophic regenerating fibers (small arrow; magnification ×40). e) Increased expression of MHC class I on muscle fibers; atypical for anti-SRP myositis (magnification ×20). g) Smaller regenerating fibers staining for neonatal myosin spread throughout the muscle (magnification ×20).

The upregulation of HLA class I on muscle cells was one of the earliest changes detected by light microscopy and was originally considered to be a response to inflammation. However, elegant experiments by Nagaraju and colleagues showed that overexpression of HLA class I in a mouse model was sufficient to drive muscle damage, leukocyte recruitment, and muscle weakness [82]. Misfolding of excessive class I protein in the endoplasmic reticulum (ER) triggers cellular damage responses, the so-called ER stress response, and the unfolded protein response, both of which can initiate immune activation [83]. To investigate how this might be relevant to JIIMs, Li and colleagues induced disease in young mice and found more severe myositis when compared to disease induction in older mice [84]. These data suggest that muscle from younger patients might be more vulnerable to ER stress-associated damage.

The results from two recent studies using the MDX mouse model of Duchenne muscular dystrophy have helped to explain how the immune system regulates muscle inflammation [85, 86]. MDX mice have a deletion of the gene that codes for dystrophin, an important structural protein in muscle. The muscles of MDX mice show signs of inflammation with a lymphocytic infiltrate, including a specialized subset of T cells termed regulatory T cells (Tregs). Depleting these Tregs led to significant worsening of muscle inflammation and greater weakness. It was also shown that Tregs enhanced muscle repair by secreting the muscle growth factor amphiregulin and by promoting the differentiation of macrophages toward a regulatory pathway. These findings are highly relevant to patients with JDM, as Tregs were found to be significantly enriched in JDM muscle compared to control muscle tissue [87]. Muscle Tregs may have an endogenous mechanism by which the immune system counters myositis and thereby limits pathology.

Conclusion

Advances in the fields of immunology and genetics are transforming the understanding of JIIMs. Immune markers, such as myositis autoantibodies, have already been translated into the clinic and provided insights into diagnosis and prognosis of JIIM patients. These myositis autoantibodies not only differentiate unique phenotypes of patients with distinct demographic and clinical features, but they are also associated with long-term outcomes, including mortality, disease course, calcinosis, and lipodystrophy. New immune and nonimmune pathways and corresponding biomarkers of disease activity are also emerging. The current challenge is to use this knowledge of dysregulated immune pathways to develop novel therapeutic agents for patients with JIIMs. The Phase Ib study of IFNα blockade in adult IIM is a promising example of this translational approach. Indeed, if successful, such approaches will herald the arrival of new biologic therapies for patients with myositis and offer an exciting future in which treatment is no longer nonspecific, but is targeted to the etiopathogenesis of individual JIIMs.