Clinical Features and the Differential Diagnosis of Myositis

The study of myositis must begin with a correct diagnosis. Myositis, or, as it is perhaps more usefully called, idiopathic inflammatory myopathy, is a disease of muscle inflammation. Muscle weakness and sometimes pain, often but not always symmetrical and proximal, are limited to the trunk, neck, and limbs. These symptoms are accompanied by signs of muscle damage: the liberation of muscle enzymes, including creatine kinase, aldolase, and other intracellular enzymes (such as aspartate aminotransferase, alanine aminotransferase, and lactate dehydrogenase); characteristic changes on electromyography; and the presence of degenerating and regenerating myocytes with inflammatory cells, especially lymphocytes, in and around muscle cells and sometimes also around vessels. We have also observed a characteristic series of patterns on magnetic resonance images of inflamed muscles (Figure 1).

View larger version (102K): [in this window] [in a new window]   Figure 1. Magnetic resonance image in a patient with myositis. A 42-year-old woman had a 5-year history of polymyositis. She had presented with myalgias, proximal weakness, and a creatine kinase level of 1300 U/L. Findings on electromyography and examination of muscle biopsy specimens were consistent with an inflammatory myopathy. Her disease responded to prednisone therapy but recurred after therapy was discontinued. Physical examination showed synovial thickening of the metacarpophalangeal and proximal interphalangeal joints, nodules on the extensor surface of the forearms, and weakness of the quadriceps. The creatine kinase level was 3884 U/L, the aldolase count was 54, and anti-Jo-1 antibodies were present. Examination of a repeat muscle biopsy specimen showed occasional degenerating myofibers and perivascular inflammation. The fat-suppressed T2 (STIR) image shows patchy bright signals characteristic of the edema that accompanies the inflammation of myositis.

The differential diagnosis of the weak patient includes a broad spectrum of diseases and toxicities. The borders separating myositis from its close relatives among the rheumatologic diseases are not sharply defined. For example, systemic lupus erythematosus shares both clinical and serologic manifestations with myositis. The late-onset, indolent myositis—inclusion body myositis—is difficult to distinguish from a bewildering variety of rare cases that in the literature are usually called dystrophies [4]. Some cases treated as myositis are examples of rare or as-yet unrecognized metabolic disorders. The term metabolic myopathy is used for diseases that feature a recognizable genetic deficiency of an enzyme that is vital to the energy production of a muscle cell; the term dystrophy is used for diseases in which no such deficiency is present but in which muscle-cell degeneration has occurred that is presumably caused by genetic defects of one or more structural proteins. This latter assumption is based on the discovery of dystrophin mutations in Duchenne dystrophy. However, because inflammation may be seen in some cases of dystrophy, and because many cases of both metabolic muscle diseases and dystrophies represent either new mutations or the first manifestations of a recessive disease in a family, the real boundaries of all of these illnesses remain unknown. Therefore, it is important to consider them in every case of suspected myositis. For each new case, a review of the major classes of similar diseases catches many of the masqueraders (Table 1). Clinicians should remember that aminotransferase abnormalities do not always indicate liver disease; that fatigue and weakness may be present in the muscles as well as in the head; that drugs and toxins can cause muscle disease; and that a family history of a similar illness, neurologic signs, asymmetry, cranial nerve involvement, and an onset of symptoms related to exercise or eating militate against a diagnosis of myositis.

We studied myositis because some patients with myositis have autoantibodies directed against the ubiquitous intracellular protein, histidyl-transfer RNA (tRNA) synthetase, the enzyme that joins histidine to its cognate tRNA [5]. These antibodies are not seen in any other disease. This strikingly specific association was made even more tantalizing by the subsequent discovery that a small group of patients developed autoantibodies directed against the synthetases responsible for ligating alanine, glycine, isoleucine, or threonine to their respective cognate tRNAs. Furthermore, other myositis-specific autoantibodies have been described in the past decade [6]. In a startling finding—the result of observations by several groups, including Drs. Lori Love, Fred Miller, and colleagues at NIH—the myositis-specific autoantibodies delineated groups of patients who have a remarkably similar clinical illness and share many other features. The common syndromes are summarized in Table 2. Patients with antisynthetase autoantibodies can have either polymyositis or dermatomyositis; those with anti-signal recognition particle (anti-SRP) autoantibodies always have polymyositis; and those with anti-Mi-2 autoantibodies always have dermatomyositis. In the traditional categories, patients with cancer-associated myositis or inclusion body myositis only rarely have myositis-specific autoantibodies, although there are a few exceptions.

Childhood Myositis: Newly Recognized Diversity

Dr. Lisa G. Rider (Food and Drug Administration, Bethesda, Maryland): The idiopathic inflammatory myopathies of childhood are a diverse group of acquired diseases of unknown cause that are all characterized by chronic inflammation in skeletal muscle. Although juvenile dermatomyositis is the most common of these disorders in children, each clinicopathologic entity described in adults has also been reported in children [7], including polymyositis [8], myositis associated with another connective tissue disease (overlap myositis) [9], cancer-associated myositis [10], focal myositis [11], orbital myositis [12], inclusion body myositis [13], eosinophilic myositis [14], and granulomatous myositis [7]. Although the frequency of these clinical conditions in children and adults probably differs, the clinical features, histopathologic findings, and clinical courses appear to be similar in both populations.

The most common idiopathic inflammatory myopathy in children is juvenile dermatomyositis, which in children has a peak incidence between ages 5 and 14 years. Juvenile dermatomyositis is similar to adult dermatomyositis and is characterized by the classic Gottron and heliotrope rashes, proximal and symmetrical muscle weakness, and perivascular muscle inflammation. Juvenile and adult dermatomyositis also share immunopathogenetic features, including damage to the endothelial cells of the primary muscle capillaries, perivascular infiltration of B lymphocytes that is associated with deposition of immunoglobulins and the terminal C5b-9 membrane attack complex on the intramuscular microvasculature, and infiltration of T lymphocytes [15].

Juvenile dermatomyositis may differ from the adult disease in the following ways: 1) The clinical presentation in children is more frequently insidious and may be dominated by constitutional symptoms of fatigue, malaise, fever, anorexia, and weight loss [16]; 2) children more often have a multisystem vasculitis that may involve the skin, gastrointestinal mucosa, muscle, heart, and retina [16]; 3) calcinosis develops more frequently in children, particularly in children with longstanding, untreated disease, those with generalized cutaneous vasculitis, or those with a chronic and severe disease course [17]; and 4) once remission is achieved, children appear to return to normal strength and function more frequently than adults with dermatomyositis [18].

The association of malignancy with the development of myositis has been well described in adults but only rarely reported in children. No cases of malignancy have been reported in retrospective studies of large cohorts of children with myositis [19]. However, tumors (most commonly leukemia and lymphoma) have been described in 10 patients with juvenile dermatomyositis and 3 children with polymyositis [7, 10]. In many of these children, the myositis showed atypical features, such as the absence of Gottron papules or the presence of unusual rashes, unilateral muscle weakness, or distal rather than proximal muscle weakness. Some children developed adenopathy, splenomegaly, or an abdominal mass that was seen on examination; the latter prompted a diagnosis of cancer. The myositis followed a paraneoplastic course in that muscle strength, skin rash, and muscle enzyme levels improved with treatment of the underlying cancer.

The myositis-specific autoantibodies also develop in children. To date, autoantibodies to threonyl-tRNA synthetase [20], alanyl-tRNA synthetase [21], and histidyl-tRNA synthetase (HRS) (22; unpublished observations) have been found in six children. Anti-signal recognition particle autoantibodies [20] have been detected in one child with polymyositis, and anti-Mi-2 autoantibodies have been found in 10 children with juvenile dermatomyositis [20, 23, 24]. We have found that the clinical features of children with myositis-specific autoantibodies for whom complete clinical data are available are similar to those of adults with the same autoantibodies [20].

Some autoantibodies that are found not only in patients with myositis but also in patients with scleroderma, the overlap syndromes, and other disorders may also have a role in identifying groups of patients that share clinical and genetic features. Anti-U1-ribonucleoprotein autoantibodies in high titer are associated with mixed connective tissue disease, a disease with overlap features in which juvenile dermatomyositis frequently occurs [9]. Anti-PM-Scl autoantibodies, which have been reported in 22 white children, are most frequently found in patients with the scleroderma-myositis overlap syndromes [25, 26]. Anti-Ku autoantibodies have been found in children with the scleroderma-myositis overlap syndromes and in patients with juvenile dermatomyositis, systemic lupus erythematosus, and lupus-myositis overlap conditions (unpublished observation; 27, 28).

Thus, the spectrum of childhood myositis resembles the spectrum of adult myositis, but the degree of similarity is still unknown. We speculate that the lower incidence of these disorders in children may be related to differences between children and adults in their genetic susceptibility alleles, environmental exposures, and responses to infectious organisms.

Humoral Autoimmunity in Myositis: The Myositis-Specific Autoantibodies

Dr. Ira N. Targoff (Oklahoma Medical Research Foundation, Oklahoma City, Oklahoma): Autoantibodies to intracellular antigens are found in about 60% to 80% of patients with polymyositis or dermatomyositis, a clear manifestation of autoimmunity. Approximately half these patients have defined autoantibodies of recognized specificity, including the myositis-specific autoantibodies (which occur almost exclusively in myositis), and others (which often occur in the overlap syndromes and are primarily associated with other diseases) [6]. The antisynthetases, the most common myositis-specific autoantibodies, are directed at aminoacyl-tRNA synthetase enzymes, each of which catalyzes attachment of a particular amino acid to its cognate tRNA [29]. Antibodies to 5 of the 20 aminoacyl-tRNA synthetase enzymes have been described [5, 30-32], but anti-histidyl-tRNA synthetase (Jo-1) is more common than all of the other antisynthetases combined. Although synthetases have general similarities in structure, they differ greatly in size and primary structure. They are immunologically distinct, and autoantibodies do not cross-react between synthetases. Individual patients generally have antibodies to only one synthetase. Synthetases for particular amino acids, on the other hand, are relatively conserved over evolution [29]. For example, the sequence that we recently obtained for the human form of glycyl-tRNA synthetase was 62% identical to that of the silkworm [33].

Antisera to four of the synthetase antigens can recognize independent proteins, but anti-OJ antibodies react with a multienzyme complex that is found in higher eukaryotic cells and contains synthetase activities for nine amino acids [32]. The major antigen is isoleucyl-tRNA synthetase, but some anti-OJ sera also react with other complex components [34]. Those additional reactions differ in several respects from other antisynthetases and may represent spreading of the initial response against isoleucyl-tRNA synthetase to other components of the complex. Although anti-OJ sera react with more than one synthetase, they generally do not develop antibodies to synthetases other than those in the multienzyme complex.

An important property of antisynthetase autoantibodies is that they inhibit the function of their target autoantigen in vitro [35]. This specific inhibition was originally used to help identify these antigens as synthetases. Animals inoculated with purified synthetases develop high titers of antibodies that bind to their antigens but do not inhibit them [36]. Such inhibition, however, is characteristic of myositis-specific autoantibodies, which nonetheless appear to arise by standard immune responses to the autoantigen [36], and probably reflects differences in binding sites on the target synthetase.

A second important property of antisynthetase sera is that they immunoprecipitate tRNAs. Different antisynthetases precipitate different sets of tRNAs that can be distinguished by polyacrylamide gel electrophoresis. The tRNAs that are immunoprecipitated are the cognates for the antigenic synthetase [31, 34]. All antisynthetases react with the enzyme, but most sera do not react directly with the tRNA [30, 32], implying that the antibody spares the tRNA binding site. An exception to this is anti-PL-12 sera, which usually have antibody that reacts directly with tRNA^ala [31].

As previously noted, each antisynthetase is independently associated with a similar syndrome of myositis, interstitial lung disease, and other features [21, 30, 31, 34, 37, 38]. A few patients have a similar syndrome, in which myositis, interstitial lung disease, and autoantibodies react with as-yet uncharacterized cytoplasmic proteins (not synthetases). These proteins are also involved in protein synthesis and translation. The best studied of these is anti-KJ, which reacts with what now appears to be a 120-kd protein whose exact role in translation is still unclear [39]. The anti-KJ antibody does not precipitate RNA but does block translation in vitro.

The anti-SRP antibody, another anticytoplasmic antibody that is specific for myositis, reacts with SRP, which is an RNA-protein complex that binds newly synthesized proteins and directs them to the endoplasmic reticulum for translocation. The anti-SRP antibody precipitates the 7SL RNA that is specific for the antigen but does not bind it directly.

The anti-Mi-2 antibody is seen almost exclusively in patients with dermatomyositis [23]. The target is a nuclear antigen whose function is unknown. It includes as many as eight proteins; the largest of these proteins is 240 kd, the main antigen that reacts with all anti-Mi-2 sera.

The relatively common myositis-scleroderma overlap syndrome is interesting in view of the shared features of vascular injury seen in both dermatomyositis and scleroderma. Several autoantibodies have been associated with this syndrome, including anti-U1RNP, anti-U2RNP, anti-Ku, and others [40]. The anti-PM-Scl antibody is unique in that its primary association is with this overlap syndrome [25, 41]. Almost all patients have frank overlap, myositis, or scleroderma alone, or features of both these diseases. The anti-PM-Scl antibody reacts with a complex of 11 proteins in the nucleolus and, less prominently, in the nucleus, whose cellular function is not known. Production of anti-PM-Scl autoantibodies is strongly associated with human leukocyte antigen (HLA)-DR3, which is present in 75% to 100% of patients [25, 41].

The features of the myositis-specific autoantibodies suggest important links to etiologic mechanisms of myositis. The antisynthetases present the interesting phenomenon in which different patients with the same syndrome have antibodies to different, immunologically distinct proteins that have precisely analogous functions. This suggests that an aspect of the shared function leads to selection of these antigens. One hypothesis is that a group of related viruses might interact with the synthetases so that they lead to presentation to the immune system, with a different virus responsible for each antisynthetase [5, 39]. However, no direct evidence supports a viral cause.

Although it is not believed that antisynthetases injure tissue by penetrating cells and inhibiting protein synthesis, many factors point to involvement of humoral immune mechanisms in dermatomyositis, more so than in polymyositis. In dermatomyositis, vascular manifestations such as nail-fold capillary changes, cutaneous ulcers, and intestinal perforation occur. Endothelial cell damage with tubuloreticular inclusions is also seen. Small blood vessels in the muscle appear to be a primary target, with resulting decreased density in muscle capillaries and ischemic muscle injury [42]. In some studies, deposition of immunoglobulin has been found in muscle blood vessels. This antibody deposition is more common in dermatomyositis, especially juvenile dermatomyositis, but is usually not found in patients with polymyositis or normal persons. Recent immunofluorescent studies have shown deposition of membrane attack complex of complement in the small vessels and capillaries, most commonly in juvenile dermatomyositis and to some degree in adult dermatomyositis, but not in polymyositis [42]. Deposition was seen in biopsy specimens from patients with dermatomyositis whose histologic findings were normal by light microscopy, suggesting that it occurred very early in the process.

Autoimmunity to Histidyl-Transfer RNA Synthetase: A Case Study

Dr. Nina Raben (NIH): The most frequent target autoantigen in patients with myositis is HRS, a cytoplasmic protein that is present in all cells. During the past 10 years, substantial progress has been made in cloning and sequencing the genes that encode aminoacyl-tRNA synthetases. Comparison of the available sequences (mostly bacterial and lower eukaryotic) has allowed enzymes to be partitioned into two classes that are based on conserved sequence motifs. Enzymes from class 1 share two sequences, and those from class 2 share three conserved motifs. The partition into two classes suggests that the enzymes have evolved from two different ancestral proteins.

We have determined the sequence of human HRS and have found in it all three of the motifs that identify it as belonging to the class 2 synthetase group [43]. Comparison of HRS from Escherichia coli, yeast, and human cells showed substantial homology. In addition to the class-specific motifs, the three enzymes have well-preserved signature regions, which define them as HRS. However, there is no homology in the n-terminal part of the proteins. In human HRS, this region, which may be a late evolutionary addition, is also a potential candidate for the antigenic epitope because disease-associated antibodies have been shown to specifically recognize antigen only from higher eukaryotes [44].

Several other features make the n-terminal part of HRS interesting. First, it contains a sequence of 32 amino acids that assumes a coiled-coil configuration. This feature is often found in autoantigens [45]. Second, the sequence motif is found in several other synthetases, both class 1 and class 2; this is the only motif shared by synthetases from different classes [44, 46]. Third, the motif is located entirely within the first two exons of human HRS, suggesting that the exons have moved between class 1 and 2 and possibly to or from other proteins.

Efforts to study the structure of HRS and to explore the mechanisms of the HRS-induced immune response in patients have been hampered by the limited availability of purified enzyme. We sought to express the protein with and without the first two exons and to determine which part of HRS is responsible for antigenicity. The baculovirus expression system proved well suited for producing large quantities of human enzyme that could readily be purified to apparent homogeneity. The recombinant protein was recognized by anti-HRS-positive sera in the immunoprecipitation assay. The molecular mass of the enzymatically active protein was identical to that of HRS purified from HeLa cells. The final yield was high (a 1-L culture of Sf9 cells expressing human HRS yielded 8 to 10 mg of purified protein). We now have an adequate capacity for HRS production, enough protein to do immunologic studies, and enough material for x-ray crystallography. The recombinant HRS is remarkably stable. The sensitivity of the enzyme-linked immunosorbent assay (ELISA) with recombinant cytoplasmic or purified antigen is several orders of magnitude higher than the sensitivity of the currently available assays. The stability of the recombinant protein is such that it remains a useful antigen for many months if stored on an ELISA plate.

Removal of the first two exons, encoding 60 amino acids of the n-terminal domain, resulted in high levels of expression of the truncated HRS with the expected reduced molecular mass. The truncated mutant HRS and various synthetic peptides from the n-terminal region were tested for their ability to bind the antibodies and thus inhibit the activity of an ELISA in which whole HRS was an antigen. The truncated HRS was negative in the inhibition assay. By contrast, two peptides (1-47 and 1-60) inhibited ELISA activity almost completely, suggesting that the major antigenic epitope is located within this region. All nine anti-HRS-positive sera we tested recognize this region on HRS. The peptides can also serve as antigens on ELISA plates, and, as expected, the truncated HRS was completely devoid of antigenic activity in an ELISA [44].

Our data strongly suggest that the major epitope on HRS is located outside the core catalytic domain in a region that may have been acquired during the evolution for as-yet unknown functions.

Cellular Immunity in Myositis

Dr. Terrance P. O'Hanlon (Food and Drug Administration): Cell-mediated immune responses are believed to play a pathogenetic role in myositis [47]. Histopathologic studies examining the types and distribution of lymphocytes found infiltrating muscle, suggesting that distinct immunopathogenetic mechanisms may be operating in different clinical groups of patients with myositis [15]. In these studies, muscle biopsy specimens from patients with dermatomyositis contained increased numbers of helper T (CD4⁺) and B lymphocytes that were primarily localized within perivascular regions of muscle. In addition, the early presence of complement membrane attack complex (C5b-9) within vascular endothelium suggests that a humorally mediated destruction of the microvasculature in muscle may antedate a secondary inflammatory response in patients with dermatomyositis. In contrast, analyses of muscle biopsy specimens from patients with polymyositis and inclusion body myositis have shown large numbers of cytotoxic T cells (CD8⁺) and macrophages that were primarily localized within endomysial regions. Activated cytotoxic T cells are usually observed infiltrating non-necrotic muscle fibers that express increased levels of major histocompatibility complex class I antigens.

Several lines of indirect evidence imply that autoreactive T lymphocytes may act as primary determinants of disease pathogenesis in myositis (Table 3). The most convincing evidence, as outlined above, is the histopathologic detection of peripheral blood mononuclear cells that infiltrate patients' muscle. In the muscle, activated cytotoxic T lymphocytes surround and invade non-necrotic muscle fibers [15]. In addition, increased proportions of lymphocytes expressing activation markers (such as interleukin-2 receptor and TLiSA1) are detectable in the peripheral blood of clinically active patients [48]. These data are corroborated by studies suggesting that patients' peripheral blood mononuclear cells traffic to muscle in vivo [49] and show increased proliferative responses to autologous muscle in vitro [50]. The production of myositis-specific autoantibodies in some patients suggests a role for helper T-cell (CD4⁺) functions in these immune disorders [51]. Moreover, patients with myositis often clinically improve after receiving immunosuppressive therapy [52].

More recently, our laboratory and other laboratories have reported restricted patterns of T-cell receptor (TCR) gene expression in defined clinical and serologic groups of patients with myositis [53-55]. These receptors are encoded by multiple, physically separated variable (V), diversity (D), joining (J), and constant (C) gene segments. During T-cell development in the thymus, individual T cells randomly recombine distinct V, D, J, and C gene segments to form functional TCR genes. The mature TCR is encoded by two such independently rearranged TCR genes, designated and ß. The ß-TCR complex discriminates between self and foreign peptide antigens presented in the context of cell-surface major histocompatibility complex molecules. The use of alternative V, D, and J gene segments or families in addition to junctional sequence variability provides a basis for recognizing a vast array of peptide antigens. It has therefore been proposed that the recognition of an antigenic determinant may be mediated by a only a few specific TCR gene families. In fact, restricted patterns of TCR gene expression have been documented in the recognition of specific antigenic peptides [56] and in the target organs of various human autoimmune diseases, including rheumatoid arthritis [57] and multiple sclerosis [58].

Studies examining the pattern of ß-TCR V gene expression in muscle biopsy specimens of patients with myositis have recently been described [54, 55]. The expression of ß-TCR gene expression collected from muscle specimens from 15 patients with polymyositis was analyzed in one such study [54]. The expression of the V 1, V 5, V ß 1, and V ß 15 gene families was detected in 60% to 100% of the muscle biopsy specimens obtained from patients with polymyositis. Deoxyribonucleic acid sequence analysis of V ß 15 muscle-infiltrating T cells, which were represented in all the muscle biopsy specimens obtained from patients with polymyositis, showed a common pattern of junctional sequence and J ß gene use. Together, these data suggest that a common antigen-driven immune response might account for the restricted pattern of TCR gene expression detected in these patients.

Our laboratory has taken a similar experimental approach in characterizing TCR V gene expression in patients with myositis [55]. We focused our analysis on a group of such patients defined serologically by the presence of anti-Jo-1 autoantibodies (anti-HRS). We screened both muscle samples and corresponding peripheral blood samples for the expression of TCR V and V ß gene families in patients with polymyositis and dermatomyositis positive for anti-Jo-1. As anticipated, all TCR V gene families surveyed were detected in the patients' peripheral blood mononuclear cells. In contrast, a restricted number of V gene families was expressed among muscle-infiltrating T cells. Although each TCR V and V ß gene family was represented in the patient population, certain TCR V gene families were over-represented in our survey. Most notably, TCR V 1 and V ß 6 gene families were detected in 82% and 91% of patients with polymyositis who had anti-Jo-1 antibodies, respectively. In contrast, the overall degree of TCR V gene heterogeneity or polyclonality was more pronounced in patients with dermatomyositis. Consequently, it was more difficult to identify predominant patterns of V gene expression among the surveyed patients with dermatomyositis.

We analyzed junctional sequences (V-D-J) expressed by V ß 6 muscle-infiltrating T cells from patients with polymyositis. Forty independent V ß 6 TCR clones were analyzed from four patients with polymyositis positive for anti-Jo-1 (10 per patient) [55]. As shown in Figure 2, we observed that among these patients, all of whom shared HLA DRB1*0301 and DQA1*0501, a restricted number of V ß 6 clonotypes were detected at high frequencies. These clones showed biased use of structurally conserved J ß gene families (J ß 2.1, 2.3, and 2.7). Moreover, similar amino acid sequences were detected within the V-D-J junctional domain (CDR3), a domain commonly associated with fine antigenic specificity [56]. This pattern of TCR V ß 6 gene expression was unlike that detected among V ß 6 clones isolated from corresponding peripheral blood mononuclear cells, which showed random, nonbiased patterns of V ß 6-J ß gene family expression. Moreover, analyses of muscle-infiltrating V ß 3 T cells, a TCR V ß gene family under-represented in the muscle of patients with polymyositis, and V ß 6 infiltrates isolated from the muscle of patients with inclusion body myositis showed nonrestricted TCR gene expression [55, 59]. The predominant detection of conserved TCR V ß 6-J ß gene expression among patients with polymyositis who have anti-Jo-1 autoantibodies supports the hypothesis that a common antigenic stimulus promotes the inflammatory response in these patients.

View larger version (18K): [in this window] [in a new window]   Figure 2. A conserved motif within T-cell receptor V ß 6 muscle-infiltrating lymphocytes from patients with polymyositis and anti-Jo-1 autoantibodies suggests a common antigenic stimulus. Shown are the amino acid sequence alignments (single-letter amino acid code) of T-cell receptor (TCR) V ß 6 clonotypes isolated from four patients with polymyositis who were positive for anti-Jo-1 (patients 1, 6, 7, and 9). A schematic illustration of a functionally rearranged TCR ß-chain gene is positioned above the alignment. The identity of germline joining (J ß) gene family members and the frequency of detection of each TCR clonotype (that is, the number of identical TCR clonotypes identified per 10 independent recombinants analyzed) are indicated to the right of the alignment. Amino acids invariably (*) or generally (.) conserved among the patients' TCR clonotypes are shown above the alignment. The putative boundaries of the V-D-J junctional domain (CDR3) are designated below the alignment [88]. Germline J ß gene sequences contributing to the CDR3 domain are shown in bold type. Regions of CDR3 amino-acid sequence similarity between individual TCR clonotypes are underlined. Acidic amino acids (E/D) mapping within the CDR3 domain and conserved among all the identified TCR V ß 6 clonotypes are double-underlined. Figure modified with permission from the Journal of Immunology and O'Hanlon and colleagues [55].

Cause and Treatment of Myositis

Cause

Dr. Frederick W. Miller (Food and Drug Administration): By definition, the causes of the idiopathic inflammatory myopathies are unknown. Recent findings imply, however, that environmental agents acting on genetically susceptible persons lead to physiologic responses involving the immunologic activation and subsequent tissue damage that we recognize as the myositis syndromes. Furthermore, new findings suggest that the myositis-specific autoantibodies are useful in delineating the genetic and environmental risk factors and the optimal therapy for patients with myositis.

The role of genetic factors in the development of myositis is suggested by an increased occurrence of myositis in certain families [60]; an increased frequency in certain racial groups (the incidence in black persons is two to three times that of white persons in the United States) [2]; and the strong association of certain HLA genes with myositis [61, 62]. Specific genes that appear to increase the risk for myositis include HLA-DR3 (HLA-DRB1*0301), HLA-DRw52, and HLA-DQA1*0501. Certain genetically encoded structures on the constant regions of the chains of immunoglobulins, which are serologically defined as Gm phenotypes and are not linked to the HLA genes, are also associated with an increased risk for myositis. The Gm phenotype most implicated in this regard is Gm 3; 5 [63]. The myositis-specific autoantibodies are useful in defining groups of patients who have even stronger associations with these genetic markers than patients with myositis as a whole [37, 51]. In addition, recent data suggest that particular Gm and HLA loci, or genes closely linked to them, interact to increase the risk for myositis within certain clinical and serologic groups [63].

The acute onset of myositis as a viral-like syndrome in many patients is one of several reasons to suspect that environmental agents play a role in these increasingly recognized systemic connective tissue diseases. In addition, recent evidence indicates that myositis develops nonrandomly in time and in place; clusterings of new cases have been reported in India [64] and North America [65]. In the United States, the onset of myositis in groups of patients who share certain autoantibodies is seasonal; patients with anti-Jo-1 autoantibodies tend to develop their first weakness in the spring, and those with anti-SRP autoantibodies, in the fall [66]. These findings suggest strong environmental influences in the cause of myositis and also suggest that dividing patients into serologic groups to isolate and define causative agents may be beneficial.

Both infectious and noninfectious agents have been implicated as possible initiators of myositis (Table 4). Infectious environmental agents that may cause myositis include picornaviruses, for which indirect supporting serologic, ultrastructural, and animal model data exist [67]. However, investigators found no evidence for such viruses after extensively screening myositis muscle biopsy specimens for viral genomes using sensitive molecular methods [68]. Nonetheless, one of the best-established links between picornaviruses and myositis is the description of children with X-linked hypogammaglobulinemia who develop dermatomyositis in association with echovirus infection that resolves with intravenous administration of -globulin [67]. Retroviruses, notably the human immunodeficiency virus (HIV) and human T-lymphotrophic virus type 1 [69], may also play a role in some myositis cases. It is important to remember that HIV infection can present as polymyositis [70]. A syndrome closely resembling polymyositis has been increasingly reported in patients with Lyme disease [71].

Persons receiving the antirheumatic drug D-penicillamine can develop a syndrome that is clinically, pathologically, and serologically indistinguishable from polymyositis. The syndrome resolves when drug therapy is discontinued, which suggests that some of the "idiopathic" inflammatory myopathies may originate in unappreciated noninfectious environmental exposures. Agents temporally associated with the onset of reported cases of myositis include the following: foods and dietary supplements such as adulterated rapeseed oil, L-tryptophan, and fish containing ciguatera toxin [72, 73]; drugs such as D-penicillamine and cimetidine [74-76]; occupational exposure to silica [72]; and silicone [77] or collagen implants [78]. Because HLA types of persons who develop myositis associated with D-penicillamine differ from those of persons whose myositis is not associated with D-penicillamine exposure [74, 75] individual susceptibility factors, such as immunogenetic or metabolizer genes, or other cofactors may account for the finding that myositis develops in only a few persons exposed to these agents.

Therapy

The primary therapeutic approach to myositis is to reduce inflammation in muscle and other affected tissues through the use of corticosteroids and other immunomodulatory agents (Table 5). Several factors are important in achieving corticosteroid responses, including the clinical and autoantibody group of the patient, the adequacy of the initial dose (1 to 2 mg/kg body weight daily), continuation of prednisone therapy until or after the serum creatine kinase level returns to normal, and a slow rate of tapering the prednisone dose that averages 10 mg/month (about 25% of the existing dose per month) [52]. The roles of bolus corticosteroids and alternate-day initial therapy remain unclear. It is important to remember that improvement in strength may lag behind improvement in the creatine kinase levels by weeks to months and that increases in the creatine kinase level may herald a flare in clinical disease.

Few data are available on the optimum therapy for corticosteroid-resistant patients, but some progress has been made in defining factors that predict disease course and prognosis in groups of patients. Because the clinical and serologic groups of patients with myositis have different disease severities, responses to therapy, and prognoses Figure 3, this information should be used in making therapeutic decisions and in designing future therapy trials [37, 51, 67, 79]. In addition to clinical or autoantibody group, other prognostic features include delay from disease onset to treatment; the severity of myositis; and significant cardiac, gastrointestinal, or pulmonary involvement [67, 79]. Patients with poor prognostic factors probably benefit from earlier, more aggressive therapy, including corticosteroids combined with cytotoxic drugs.

View larger version (23K): [in this window] [in a new window]   Figure 3. Generalized myositis courses in clinical (top) and serologic (bottom) groups. Patients with myositis in association with other connective tissue diseases (myositis overlap) tend to have milder myositis that responds relatively well to therapy and infrequently returns after therapy is discontinued. Patients who have polymyositis have more severe disease at onset with poorer responses to therapy and frequent exacerbations after therapy has been tapered compared with patients who have dermatomyositis [67, 77, 79]. Patients with inclusion body myositis have slowly progressive disease in which the rate of deterioration may decrease with therapy [83]. A similar difference in myositis severity at onset, response to therapy, and disease exacerbation with discontinuation of therapy is seen in the serologic groups. Patients with anti-Mi-2 and anti-MAS (directed against a cytoplasmic RNA of unknown function) autoantibodies have the best prognosis, followed by the antisynthetase group and finally the anti-SRP antibody group [51, 77, 79].

Rehabilitation is important in maintaining range of joint motion and preventing contractures during active myositis, and exercise probably improves strength and endurance when gradually initiated during disease remission [80]. Oral methotrexate and azathioprine remain the primary therapeutic options for corticosteroid-resistant patients. Recent data suggest that methotrexate may be superior to azathioprine in treating myositis, especially in men and in patients with antisynthetase autoantibodies [79]. However, the risk factors of individual patients generally determine which of these two drugs should be used. Intravenous -globulin [81], cyclosporine, chlorambucil [82], or combinations of cytotoxic drugs [83] may be beneficial in some groups of patients with myositis and warrant further evaluation. Pulse cyclophosphamide may be useful in treating interstitial lung disease associated with myositis, especially in the context of the antisynthetase syndrome [84, 85]. Neither plasma exchange nor leukapheresis appears to benefit steroid-resistant patients not receiving cytotoxic agents, as shown by a recent double-blind, sham-controlled trial [86]. The extramuscular systemic manifestations of myositis can be difficult to treat. With the exception of the use of hydroxychloroquine and sunscreens for the treatment of the rash of dermatomyositis, no specific treatment has been shown to benefit the systemic manifestations of idiopathic inflammatory myopathies.

Some patients with inclusion body myositis, particularly those with a creatine kinase level greater than 1000 U/L and evidence of inflammation on examination of muscle biopsy specimens, may benefit from corticosteroid and cytotoxic therapy. Benefits include improvement in function or the slowing of disease progression [83]. Severe progressive dysphagia that is unresponsive to chemotherapy may respond to cricopharyngeal myotomy [87].

In conclusion, although our knowledge of inflammatory muscle disease remains limited, recent discoveries about myositis autoantibodies and TCRs have improved our understanding of the cause, pathogenesis, and therapy for this group of diseases. Further investigations of the immunologic features of the syndromes characterized by chronic muscle inflammation should lead to improved treatments of the increasing number of persons with myositis.