In polymyositis and dermatomyositis (DM), identified autoantibodies occur in <50% of adult patients and in a smaller proportion of children. This study was undertaken as part of a larger effort to define novel autoantibodies that assist in the clinical evaluation of myositis.
Sera from children and adults satisfying criteria for idiopathic inflammatory myopathies and from patients with other connective tissue diseases (CTDs), patients with noninflammatory myopathies, and healthy individuals were tested for autoantibodies by immunoprecipitation (IP). A previously unrecognized autoantibody that immunoprecipitated a 155-kd protein along with a weaker 140-kd protein was seen. When the presence of this anti-p155 autoantibody in test sera was suggested based on IP results, it was confirmed by immunoblotting of immunoprecipitates.
Sera from 51 of 244 myositis patients (21%), including 30 with juvenile DM (29%), 5 with juvenile CTD–associated myositis (33%), 8 with adult DM (21%), 6 with cancer-associated DM (75%), and 2 with adult CTD–associated myositis (15%), were found to have anti-p155 autoantibody. One of 49 patients with lupus, and none of 89 others without myositis, had anti-p155. Caucasian patients with anti-p155 had a unique HLA risk factor, DQA1*0301 (odds ratio 5.4, corrected P = 0.004). In adults with anti-p155, of several clinical features assessed only the frequency of V-sign rash was increased, but patients with this antibody were clinically distinct from those with autoantibodies to aminoacyl–transfer RNA synthetases.
A newly recognized autoantibody, anti-p155, is associated with DM and cancer-associated DM, and is one of the most common autoantibodies in this condition, occurring as frequently in children as in adults. The clinical features and immunogenetics associated with anti-p155 differ from those associated with antisynthetases.
Patients with polymyositis (PM) and dermatomyositis (DM) frequently have autoantibodies to intracellular antigens (1, 2). In 60–80% of these patients, indirect immunofluorescence (IF) testing shows the presence of antinuclear or anticytoplasmic antibodies, and some of the remaining patients have other evidence of such autoantibodies (3, 4). Several autoantibodies that occur in PM and DM have been defined (2), but generally, each occurs in a small proportion of the patients with these conditions. Anti–Jo-1 (anti–histidyl–transfer RNA [anti–histidyl–tRNA]) synthetase, usually considered to be the most common myositis autoantibody, occurs in only ∼20% of adult patients with PM or DM and in a smaller percentage of children with DM. Together, defined autoantibodies occur in ∼50–60% of patients with PM or DM (5).
Despite low sensitivity, some of these autoantibodies have high specificity for myositis. Anti–Jo-1 and other autoantibodies to aminoacyl-tRNA synthetases, anti–signal recognition particle (anti-SRP) autoantibodies, and anti–Mi-2 autoantibodies (2) usually occur in association with myositis. The autoantibodies also tend to be mutually exclusive, with occasional exceptions, and are often associated with certain clinical, genetic, and prognostic features, allowing them to define patient subsets that may not correspond to clinically defined groups (2, 3).
The autoantibody with the strongest established association with DM is anti–Mi-2, directed against a nuclear protein associated with the nucleosome remodeling and histone deacetylase complex. Investigators using immunodiffusion (ID) or immunoprecipitation (IP) have found anti–Mi-2 almost exclusively in DM (3, 6, 7). Even within this group, anti–Mi-2 shows relatively low sensitivity; it is present in <20% of US patients with DM (7, 8). Other autoantibodies that occur in DM may also occur in PM (such as anti–Jo-1) or in overlap syndromes (such as anti–PM-Scl). Most myositis autoantibodies have been found less commonly in juvenile DM (9, 10) than in adult DM; they are usually found in <20% of juvenile DM patients (9, 10). In cancer-associated myositis, which is more often DM than PM, the frequency of myositis autoantibodies is also lower (3). Therefore, autoantibodies that could be used as diagnostic markers for additional myositis patients, particularly in juvenile DM and cancer-associated myositis, could have significant clinical utility.
In an effort to better characterize the autoantibodies in idiopathic inflammatory myopathies (IIMs), and to define their clinical and prognostic usefulness, sera from a large group of patients with adult or juvenile IIM have been examined. In this report, we describe a newly identified autoantibody specificity and its association with DM.
PATIENTS AND METHODS
Two hundred twenty-eight patients with IIM who met Bohan and Peter criteria for probable or definite DM or PM (11, 12) were research subjects at the National Institutes of Health (NIH) or at the Center for Biologics Evaluation and Research at the Food and Drug Administration, as part of institutional review board–approved protocols. Sera from these patients were first collected and tested for myositis autoantibodies during the period from 1997 to 2000. Sixteen additional PM samples from the serum bank at the Oklahoma Medical Research Foundation were used. See Appendix A for a list of the members of the Childhood Myositis Heterogeneity and International Myositis Collaborative Study Groups who contributed to this study.
Classification as DM, PM, cancer-associated myositis, connective tissue disease (CTD)–associated myositis, or inclusion body myositis was done using previously described criteria (3). Patients were classified as having juvenile IIM if they were <18 years of age at myositis onset. Sera from normal subjects or from those with other diseases, including systemic lupus erythematosus (SLE), systemic sclerosis, rheumatoid arthritis, psoriatic arthritis, Sjögren's syndrome, genetic muscle diseases, or primary dystrophies, who were evaluated at the same centers, were also studied. Serum of patient p155-1, which was anti-p155 positive by immunoblotting, was used as the main reference serum for this study. Serum p155-2, another positive sample, was used as a reference in some experiments. Serum p155-3 was the original prototype serum. Adult IIM patients with anti-p155 antibodies were assessed for 16 clinical features, as described previously (3), and the results were compared with the findings in a group of 181 PM/DM patients previously studied at the NIH (3).
Serum samples were screened for the presence of autoantibodies by IP for protein and nucleic acid analysis, as well as by indirect IF on HEp-2 cells and ID against calf thymus extract, as previously described (7, 13, 14). The earliest available serum sample was screened. HeLa cell extract was used as the antigen source for screening IP experiments, and K562 cell extract was used as the antigen source for immunoblotting experiments, except where indicated. For protein analysis, 1 mg or 5 μl protein A–Sepharose was mixed with 10–15 μl serum, washed, and incubated with extract of HeLa or K562 cells labeled with 35S-methionine. Immunoprecipitates were analyzed by 8–10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and developed by autoradiography. For nucleic acid analysis, 3–4 mg protein A–Sepharose was mixed with 20 μl patient serum, washed, and incubated with unlabeled HeLa extract. Immunoprecipitates were treated with phenol–chloroform extraction and ethanol precipitation and were analyzed by urea-PAGE, with silver stain development. Previously described myositis autoantibodies, except for anti–56 kd, anti–Ro 52, and anti-PMS1, have been detected with these methods.
Immunoblotting was performed as previously described (13). In most experiments, immunoprecipitates from K562 cell extract, rather than whole cell extract, were used as antigen (IP blotting). Unlabeled immunoprecipitates were prepared as described above for nucleic acid IP, but in this case a crosslinking step was included to reduce the amount of immunoglobulin. Fifty to one hundred microliters of serum and 10–20 mg protein A–Sepharose were used per sample. Protein A–Sepharose was incubated with patient serum and washed, then treated with dimethyl pimelimidate as previously described (15), and then incubated with cell extract. Samples were subjected to 7.5% SDS-PAGE, then transferred to nitrocellulose. Nitrocellulose strips were then probed with the test serum samples, followed by alkaline phosphatase–conjugated anti-human IgG and substrate steps. In some experiments, a reference serum was used to prepare the immunoprecipitates, and the test serum was used for blotting (IP blotting), while in other experiments, the test serum was used to prepare the immunoprecipitates, and the reference serum was used for blotting (reverse IP blotting).
Definition of a positive p155 antibody test result.
IP was used to screen for the p155 antibody in sera from patients with myositis. If IP results were consistent with the presence of anti-p155 (a 155-kd band similar to that obtained with the reference serum, often accompanied by a 140-kd band), IP blotting or reverse IP blotting was used to confirm the presence of the antibody. All samples from the control groups were tested by reverse IP blotting. Sera that were positive by either IP blotting or reverse IP blotting were considered to be confirmed positive for anti-p155 autoantibodies. Sera that were negative by IP blotting but positive by reverse IP blotting were still considered positive. Sera that were negative by reverse IP blotting were considered negative for anti-p155 antibodies regardless of IP results. Sera that had consistent IP results but could not be tested by immunoblotting, and sera that were negative by IP blotting but could not be tested by reverse IP blotting, were excluded from the analysis.
HLA–DQA1 and DRB1 typing.
Genomic DNA was purified from peripheral blood mononuclear cells using the QIAamp DNA blood isolation kit (Qiagen, Valencia, CA). High-resolution HLA–DRB1 and DQA1 genetic typing was performed using commercially prepared reagents for sequence-specific primer–polymerase chain reaction, according to the recommendations of the manufacturers (GenoVision, West Chester, PA and Dynal Biotech, Lafayette Hill, PA). The HLA loci and their specific alleles were determined by standardized and validated genetic typing techniques, as previously described (16, 17).
The frequencies of clinical features or HLA alleles among groups were compared by chi-square test or Fisher's exact test using GraphPad InStat (GraphPad Software, San Diego, CA). Corrections for multiple comparisons were made using the Bonferroni adjustment (3). Corrected P (Pcorr) values less than 0.05 were considered significant.
During IP testing of sera from patients with IIM, a distinct protein of 155 kd was detected in the sera of several patients with DM (Figure 1). An accompanying, weaker protein of 140 kd could be detected with positive samples on most tests (Figure 1). In contrast to antisynthetases and anti-SRP, no nucleic acid could be detected by IP with these sera. A nuclear speckled pattern was seen by indirect IF on HEp-2 cells (results not shown), but the results of ID against calf thymus extract were negative. These tests did not reveal other known myositis autoantibodies in sera that had the 155-kd protein, but IP did show other unidentified protein bands with some sera.
Confirmation of the antibody.
In order to determine whether different sera with this IP result were reacting with the same protein, IP blotting was performed. When serum p155-1 (Figure 1) was used both for preparation of the immunoprecipitates (the antigen for IP blotting) and for the immunoblot, blotting of a 155-kd protein was seen. Serum p155-1 also showed blotting of unidentified proteins of lower molecular weight (Figure 2). Thirty other IP-positive sera (sera showing 155-kd bands by IP consistent with those obtained with the reference serum) were then tested by IP blotting against immunoprecipitates of serum p155-1. Nine (30%) of 30 showed definite positive staining by IP blotting, 5 (16.7%) showed equivocal staining, and 16 (53.3%) showed definite negative results. A control normal serum and an IP-negative myositis serum were negative by IP blotting.
These data supported our hypothesis that the 9 sera with positive IP blotting results had antibodies to the same p155 antigen as did the reference serum. However, the data did not exclude the possibility that the other 21 sera with positive IP results also had the antibody; these sera could have been unable to react by IP blotting despite IP reaction with the same 155-kd protein if, for example, they exclusively recognized conformational epitopes. A similar finding was previously noted for anti-OJ autoantibodies (13).
To address this issue, IP-positive sera were further studied by reverse IP blotting in which the test serum was used to prepare the immunoprecipitates and reference serum p155-1, shown above to have immunoblot reactivity, was used for immunoblotting. As shown in Figure 2, when serum p155-2 was used for the IP step, both serum p155-1 and serum p155-2 stained a similar 155-kd protein. This was consistent with our supposition that both serum p155-1 and serum p155-2 reacted with the same protein. In contrast, when serum p155-3, another IP-positive serum, was used in the IP step, this serum did not show significant reaction with the 155-kd protein by immunoblot. However, serum p155-1 did react with a 155-kd protein in p155-3 immunoprecipitates. We concluded that serum p155-3 could recognize p155 protein by IP, but not by immunoblotting. This suggested exclusive reactivity with conformational epitopes. Thus, reverse IP blotting could be used to confirm the presence of anti-p155 in IP-positive sera, even if the sera were unable to react with the antigen by immunoblotting.
Screening of patient populations for anti-p155 antibody.
Of 244 total myositis patient sera screened by IP, 54 were positive for protein bands consistent with anti-p155. Of these 54 sera, 51 were further tested by reverse IP blotting, and 3 were not available for testing by this method. Two of these 3 had previously been tested by IP blotting and were confirmed positive by that method, but the third serum was not subjected to confirmatory testing. Of the 51 sera subjected to reverse IP blotting, 49 were confirmed positive, and 2 were negative. Thus, a total of 51 sera from patients with myositis were confirmed positive for anti-p155 (Table 1). The 2 sera that were negative and the sample that was not subjected to confirmatory testing were not studied further.
Table 1. Frequency of anti-p155 autoantibodies in patient clinical groups*
Clinical group (n)
No. (%) of subjects positive for anti-p155
All samples were screened for anti-p155 antibodies by protein immunoprecipitation. Samples with characteristic bands at 155 kd were confirmed to be positive by immunoprecipitation blotting or reverse immunoprecipitation blotting, as described in Patients and Methods. IIM = idiopathic inflammatory myopathy; DM = dermatomyositis; PM = polymyositis; CTD = connective tissue disease; SSc = systemic sclerosis.
Two of the 5 patients with juvenile CTD–associated myositis who were positive for anti-p155 had juvenile DM. Of the 10 patients with juvenile CTD–associated myositis who were negative for anti-p155, 7 had juvenile DM and 3 had juvenile PM.
The 2 patients with CTD-associated myositis who were positive for anti-p155 had DM. Of the 11 patients with CTD-associated myositis who were negative for anti-p155, 4 had DM, 6 had PM, and 1 had inclusion body myositis and systemic lupus erythematosus (SLE).
The 6 patients with cancer-associated myositis who were positive for anti-p155 had DM, and the 2 patients with cancer-associated myositis who were negative for anti-p155 had PM.
The 51 positive sera included 30 of 103 in the juvenile DM group (29% of those tested) and 8 of 39 in the adult DM group (21% of those tested) (Table 1). Of interest, the antibody was found in samples from 6 of the 8 adult patients with cancer-associated myositis, all 6 of whom had a diagnosis of DM. Thus, malignancy was present in 37.5% of anti-p155–-positive adult myositis patients in this study (6 of 16), compared with 11.8% of adult DM patients overall (6 of 51). Malignancy was not seen in anti-p155–negative DM patients. The antibody was not found in patients with adult PM (0 of 48) or juvenile PM (0 of 9) but was found in 5 of 15 children and 2 of 13 adults with CTD-associated myositis, including 2 children and 2 adults with DM, and 3 children with PM. Thus, including children and adults, the antibody was much more common among all patients diagnosed as having DM (48 of 163 [29.4%]) than among those diagnosed as having PM (3 of 71 [4.2%]). The antibody was found in only 1 of 138 sera from patients with other CTDs or nonmyositis myopathies or healthy controls. The 1 positive serum in this group was from a patient with SLE (Table 1); however, anti-p155 was more frequent in patients with IIM than in patients with SLE (P < 0.001).
Immunogenetic associations of anti-p155 autoantibody.
In Caucasian patients with anti-p155 autoantibody, DQA1*0301 was found to be an immunogenetic risk factor, present in 30%, compared with 7% of 555 healthy Caucasian controls (odds ratio 5.4, 95% confidence interval 2.3–12.5, Pcorr = 0.004). Of note, DQA1*0501 and DRB1*0301, alleles known to be risk factors for IIM in Caucasian patients (7), were not risk factors for presence of anti-p155 autoantibody. DQA1*0501 was present in 33% of patients with anti-p155 autoantibody versus 42% of controls, and DRB1*0301 was present in 33% of patients with anti-p155 antibody versus 20% of controls.
Clinical features of adult IIM patients with anti-p155 autoantibody.
Compared with a historical group of patients with IIM, the 16 adult patients with anti-p155 antibody had a higher frequency of the V-sign rash, consistent with the association of anti-p155 with DM (Table 2). There was also a higher frequency of the V-sign in anti-p155–positive patients with cancer-associated myositis than in the historical group of patients with cancer-associated myositis. Anti-p155–positive patients also had a high frequency of heliotrope rash (8 of 13 [62%]) and Gottron's papules (13 of 14 [93%]). Data on the frequencies of these features in the historical group were not available for comparison. There were no differences in the clinical features between anti-p155–positive patients and the overall DM group, although there was a lower frequency of interstitial lung disease (ILD) in patients with anti-p155 (0%, versus 26% in the overall DM group).
Table 2. Selected clinical features in anti-p155–positive adult myositis patients compared with adult patients with IIM overall and with antisynthetase-positive adult myositis patients*
The most striking differences were found when the anti-p155–positive patients were compared with the groups defined by other myositis autoantibodies (Table 2). Antisynthetase-positive patients had a significantly higher frequency of several extramuscular features commonly associated with the antisynthetase syndrome, including fevers, Raynaud's phenomenon, arthritis, dyspnea, ILD, and mechanic's hands. In contrast, anti-p155–positive patients had a higher frequency of the V-sign than patients with antisynthetase autoantibodies (77% versus 15%; Pcorr = 0.0016). As expected, anti-p155–positive patients appeared to have the V-sign more often than patients with anti-SRP (77% versus 0%; Pcorr = 0.05 [data not shown]), although only a small number of anti-SRP–positive patients were included. The only significant difference in patients with anti-p155 compared with those with anti–Mi-2 was a higher frequency of cuticular overgrowth in patients with anti–Mi-2 (Pcorr = 0.0032 [data not shown]).
This report describes a newly recognized autoantibody, anti-p155, which was strongly associated with myositis (present in 22% of all patients with PM/DM versus <1% of controls), DM (present in 29.4%, versus 4.2% of patients with PM), and cancer-associated DM (present in 75%). The frequency of anti-p155 was relatively high compared with other myositis autoantibodies, and was higher than that usually reported for anti–Mi-2 in Caucasian patients with adult and juvenile DM (18). Anti-p155, present in 29% of those tested, appears to be more common than other defined autoantibodies in juvenile DM (9, 10).
Since other myositis autoantibodies were infrequent in patients with anti-p155, this autoantibody expands the proportion of myositis patients for whom a marker autoantibody is detectable. This is particularly true for patients with juvenile DM, who have a low frequency of other myositis autoantibodies. This also means that anti-p155 could potentially define a new myositis subgroup. Thus far, it has not been possible to significantly distinguish this subgroup from the overall population of patients with DM, but further studies of the clinical and prognostic associations of anti-p155 are needed.
Anti-p155 was not completely specific for IIM, since it was found in 1 lupus patient without myositis, although it has not been detected in normal subjects or those with other autoimmune diseases or muscular dystrophy. Other autoantibodies that are considered to be “myositis-specific” have also occasionally been observed in patients who had not developed myositis after significant periods of observation. Anti-SRP was recently found in patients with systemic sclerosis, ILD (19), and rheumatoid arthritis (20). Antisynthetases can also rarely occur without myositis (21), usually in patients with other manifestations of the antisynthetase syndrome. The presence of these antibodies could still be of potential value for diagnosis when considered in conjunction with the patient's clinical features.
An autoantibody to a 56-kd protein was reported to be found in up to 90% of patients with juvenile DM (22–24). However, unlike anti-p155, anti–56 kd was frequently seen in PM patients (23, 24). Several reports of studies of myositis antibodies included no reference to anti–56 kd (5, 7, 25), and it was not detected in this study. Its relationship to anti-p155 antibody requires further study.
The identification of anti-p155 antibody in both adult and juvenile DM supports previous reports that myositis autoantibodies can occur in juvenile DM, and usually have clinical associations similar to those in adults (26). Unlike most myositis autoantibodies that are less frequent in juvenile DM, anti-p155 is at least as frequent in juvenile DM as in adult DM.
Immunoblotting was used to confirm that all IP-positive sera were reacting with the same antigen. However, a reverse IP blotting strategy was necessary because two-thirds of sera did not react by immunoblot, suggesting exclusive reaction with conformational epitopes. Reverse IP blotting confirmed that patients shared an autoantibody with reference serum p155-1, but the possibility exists that serum p155-1 has >1 antibody to a 155-kd protein. However, the combination of a strong band at 155 kd accompanied by a weaker 140-kd band makes this less likely. Further clarification will be possible when the nature of the antigen is better understood.
It is notable that a relatively high proportion (75%) of patients with cancer-associated myositis had anti-p155, including all 6 with DM, representing 37.5% of anti-p155–positive adult patients with DM. Confirmation of a definite association of anti-p155 with cancer-associated myositis will require further study; however, this could be clinically useful in identifying patients with DM who should be examined more intensively for cancer. Also, it might be of value for studying the mechanism of the association of cancer with DM.
The antibody was found in all DM clinical groups (adult, juvenile, cancer-associated myositis, and CTD-associated myositis), suggesting a link among a portion of patients with different forms of DM. However, the relatively high frequency of the antibody in both cancer-associated myositis and juvenile DM is of interest. Understanding the relationship between these groups would require a better understanding of the etiology of DM and the reason for the association of DM with these autoantibodies and with malignancy. It is noteworthy, however, that both patients with juvenile myositis and patients with cancer-associated myositis predominantly have DM rather than PM, and that both have a low frequency of the previously identified myositis autoantibodies.
An autoantibody associated with clinically amyopathic DM (anti–CADM-140), reacting with a 140-kd protein, has recently been identified using IP (27). Anti–CADM-140 was associated with a cytoplasmic pattern by indirect IF, in contrast to the nuclear pattern seen with anti-p155. Another autoantibody, anti-MJ, has been described in patients with juvenile DM in preliminary reports (28, 29). A reference serum for anti-MJ (kindly provided by Dr. Chester Oddis, Pittsburgh, PA) was negative for anti-p155 by direct and reverse IP blotting. Thus, 3 new autoantibodies have been associated with DM and may define separate DM subgroups. At least 2 of them occur relatively frequently in juvenile DM. Previously, the only autoantibody associated predominantly with DM was anti–Mi-2. Although it can occur in cancer-associated myositis and juvenile DM, the frequency of anti–Mi-2 in these DM subgroups appears to be lower than the frequency in DM overall, and it is usually not seen in amyopathic DM.
The presence of anti-p155 in typical juvenile DM is of interest with regard to disease pathogenesis. Antigens targeted by myositis autoantibodies may be up-regulated in the affected muscles, particularly regenerating myoblasts, of patients with IIM, with specificity for the particular clinical subset of disease (30). This raises the possibility of a role of these autoantigens in the pathogenesis of myositis. Current studies are focused on further characterization of the autoantibodies in IIM, their mechanisms of induction, and their possible roles in pathogenesis.
A recent article described a similar autoantibody that also reacts with 155- and 140-kd proteins, associated with DM with an increased frequency in patients with cancer-associated myositis (31).
We thank Drs. Steven Bauer, Kurt Brorson, and Paul Plotz for helpful comments on the manuscript, and Dr. Plotz for contribution of serum samples and support of this work.
MEMBERS OF THE CHILDHOOD MYOSITIS HETEROGENEITY AND INTERNATIONAL MYOSITIS COLLABORATIVE STUDY GROUPS
The following members of the Childhood Myositis Heterogeneity and International Myositis Collaborative Study Groups contributed to this study: Carol Lindsley, Hilary Haftel, Carol Wallace, Donald Goldsmith, Maria Perez, Luis Javier Jara, David Sherry, Chihiro Morishima, Alexander Lawton, Richard Leff, John Varga, Harry Gewanter, Michael Knee, Alan Lichtbraun, Ignacio Garcia de la Torre, Elif Oral, Murray Passo, Donald Person, Peter Merkel, Michael Grisanti, Gail Cawkwell, Gloria Higgins, Rafael Rivas-Chacon, Edward Snell, Walter Chase, Terri Finkel, Lawrence Zemel, Ildy Katona, Ilona Szer, Andrew Eichenfeld, Stephen George, Eugene Arthur, Michael Henrickson, Alan Martin, Nancy Sadler, Stephen Plotnick, Peter Jongen, Thomas Bunch, Yeong Wook Song, Balu Athreya.