The ability to accurately assay a broad range of well-defined autoantibody specificities in myositis patients is imperative for clinical phenotyping and patient care. In myositis research, the systematic serotyping of patient cohorts is often performed using various research assays, and understanding the relationship between research testing and clinical testing is paramount to advancing the field of myositis. An increasingly popular research multiplex antibody assay platform was developed by Euroimmun (EUROLINE Autoimmune Inflammatory Myopathies 16 Ag IgG, Lubeck, Germany) for myositis-specific and associated autoantibodies. Prior research has suggested that the inter-rater reliability of Euroimmun compared to in-house immunoprecipitation methods (considered the gold standard) is reasonable, with the possible exception of TIF1γ [1–4]. In this communication, we share our experience comparing serological results obtained for clinical purposes with a commercial myositis autoantibody panel (the traditional Oklahoma Medical Research Foundation Myositis panel, “OMRF”) and serologies obtained for research purposes using the Euroimmun Myositis panel.
The OMRF myositis autoantibody panel utilizes several different assays to read out antibodies. These include immunodiffusion, indirect immunofluorescence (ANA), immunoprecipitation of 35S-methionine-labeled proteins from cell extracts and RNA-immunoprecipitation. The assay is performed at the Clinical Immunology Laboratory in Oklahoma City. It tests for autoantibodies recognizing Jo-1, Mi2, SRP, PM/Scl, PL-7, PL-12, Ku, EJ, and OJ. Results are provided as negative, positive, or weak positive/indeterminate. Weak positive and indeterminate results were excluded from this analysis (10 values). The Euroimmun Autoimmune Inflammatory Myopathies 16 Ag platform tests for Mi2α, Mi2β, PM/Scl75, PM/Scl100, Ku, Jo-1, SRP, PL-7, PL-12, EJ, OJ, TIF1γ, MDA-5, NXP-2, SAE, and Ro52. Based on the Euroimmun manufacturer package insert, different thresholds correspond with borderline (8–14), low positive (≥15, or +), moderately positive (≥36, or ++) and strongly positive (≥71, or +++) results. Borderline Euroimmun results were considered negative for this analysis. Intra-assay reproducibility was previously established both by Euroimmun as well as by our own lab with excellent agreement. Sensitivity and specificity were subsequently calculated for the OMRF and Euroimmun platforms along with the mean ± standard deviation (SD) of Cohen’s kappa statistic in order to measure inter-rater agreement using Stata version 14 (College Station, Texas).
We performed Euroimmun testing on serum samples from all consented patients in The Johns Hopkins Myositis Cohort. The subset of patients (dermatomyositis, polymyositis, or immune-mediated necrotizing myopathy, n = 281) was subsequently identified with clinical OMRF myositis autoantibody testing performed on a serum sample taken on the same day as that used for Euroimmun testing. The importance of matching the bleed date has become increasingly appreciated with studies reporting a change in autoantibody titer over time [5–7]. At the time of antibody testing, patients had a mean age of 52 (±14), and the majority of patients were female (71%). The racial composition of the cohort was 72% Caucasian, 13% African American, 4% Asian, and 11% unknown. The mean disease duration from symptom onset to antibody testing was 3.9±5.1 years, and the majority of patients were on immunosuppressive treatment at the time of antibody testing.
A total of 154 patients (55% of samples tested) had at least one antibody specificity by Euroimmun using the standard positive cutoff of ≥15. Of note, this low prevalence of seropositive patients is likely explained by the fact that several other prominent myositis antibody specificities such as HMGCR, TIF1γ, SAE, NXP-2, and MDA-5 were not included in this study (see below). Using the cutoff of ≥15 on Euroimmun, a total of 15 patients had more than one autoantibody (excluding co-positivity of Mi2α-Mi2β and PM/Scl75-PM/Scl100). The number of patients tested, the number of positive results by each assay, and the corresponding sensitivities, specificities, and kappa statistics are presented in Table 1. The inter-assay agreement using different cutoffs for the Euroimmun assay was calculated using kappa statistics and was found to be highest overall for the moderate cutoff (0.73±0.18, 0.78±0.13, and 0.71±0.27 for thresholds of ≥15/+, ≥36/++, and ≥71/+++, respectively). The lowest threshold (≥15/+) resulted in the best kappa statistic for anti-Jo-1, -Mi2α, -Mi2β, -PM/Scl100, and -Ku autoantibodies. The moderate positive threshold of ≥36/++ resulted in the best kappa statistic for anti-SRP -EJ, and -PL-12. The highest cut-off ≥71/+++ had the best kappa statistic for anti-PM/Scl75 and -PL-7.
Table 1.
Autoantibody (EUROIMMUN vs OMRF) | #of patients tested | #of OMRF positive | #of positives on EBO | Cut-off (>=15)/+ | #of positives on EBO | Cut-off (>=36)/++ | #of positives on EBO | Cut-off (>=71)/+++ | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Sensitivity | Specificity | Kappa | Sensitivity | Specificity | Kappa | Sensitivity | Specificity | Kappa | ||||||
Jo1 | 281 | 29 | 31 | 96.6 | 98.6 | 0.93 | 27 | 89.7 | 99.6 | 0.92 | 24 | 79.3 | 99.6 | 0.85 |
Mi2 | 280 | 20 | ||||||||||||
Mi2a | 280 | 20 | 94.7 | 99.6 | 0.94 | 11 | 55.0 | 100.0 | 0.69 | 2 | 10.5 | 100.0 | 0.18 | |
Mi2b | 280 | 30 | 85.0 | 95.0 | 0.65 | 18 | 60.0 | 97.7 | 0.60 | 8 | 40.0 | 100.0 | 0.55 | |
Mi2a or Mi2b | 280 | 33 | 100.0 | 94.9 | 0.73 | 22 | 80.0 | 97.6 | 0.74 | 8 | 40.0 | 100.0 | 0.55 | |
M12a and Mi2b | 280 | 17 | 80.0 | 99.6 | 0.85 | 7 | 35.0 | 100.0 | 0.50 | 2 | 10.0 | 100.0 | 0.17 | |
SRP | 280 | 18 | 27 | 94.4 | 96.2 | 0.74 | 20 | 94.4 | 98.6 | 0.89 | 13 | 70.6 | 99.6 | 0.79 |
PM-Scl | 279 | 12 | ||||||||||||
PM75 | 279 | 26 | 83.3 | 94.0 | 0.50 | 13 | 75.0 | 98.5 | 0.71 | 7 | 58.3 | 100.0 | 0.73 | |
PM100 | 279 | 15 | 83.3 | 98.1 | 0.73 | 10 | 66.7 | 99.2 | 0.72 | 4 | 25.0 | 99.6 | 0.36 | |
PM75 or PM100 | 279 | 33 | 100.0 | 92.1 | 0.50 | 16 | 83.3 | 97.7 | 0.70 | 8 | 58.3 | 99.6 | 0.69 | |
PM75 and PM100 | 279 | 8 | 66.7 | 100.0 | 0.79 | 7 | 58.3 | 100.0 | 0.73 | 3 | 25.0 | 100.0 | 0.39 | |
PL7 | 280 | 5 | 15 | 100.0 | 96.4 | 0.49 | 9 | 100.0 | 98.5 | 0.71 | 5 | 100.0 | 100.0 | 1.00 |
PL12 | 281 | 3 | 7 | 100.0 | 98.6 | 0.59 | 4 | 100.0 | 99.6 | 0.86 | 4 | 100.0 | 99.6 | 0.86 |
Ku | 279 | 3 | 5 | 100.0 | 99.3 | 0.75 | 5 | 100.0 | 99.3 | 0.75 | 5 | 100.0 | 99.3 | 0.75 |
EJ | 281 | 2 | 3 | 100.0 | 99.6 | 0.80 | 2 | 100.0 | 100.0 | 1.00 | 2 | 100.0 | 100.0 | 1.00 |
OJ | 281 | 0 | 2 | NA | NA | NA | 0 | NA | NA | NA | 0 | NA | NA | NA |
To better understand discordant results (patients who were positive by Euroimmun but negative by OMRF), we describe the clinical phenotype of 19 patients who had antisynthetase antibodies in Table 2. Among these 19 Euroimmun positive patients, only 5 (26%) had a clinical picture consistent with the antisynthetase syndrome, and those that did were often positive for other antibody specificities. Interestingly, many of these patients were amyopathic DM patients.
Table 2.
Patient positive by EBO≥15 but negative by OMRF | Inflammation on muscle biopsy | Fever (>100.4F) | Interstitial lung disease | Mechanic’s hands | Raynaud’s phenomenon | Inflammatory arthritis | Additional antibody specificities observed on EBO |
---|---|---|---|---|---|---|---|
N, (%) | N, (%) | N, (%) | N, (%) | N, (%) | N, (%) | ||
PL7 (N=10) | 5 (50) | 1 (10) | 3 (30) | 3 (30) | 6 (60) | 2 (20) | Jo1 (3), PL12 (1), Ku (1) |
PL12 (N=4) | 2 (50) | 0 (0) | 2 (50) | 1 (25) | 2 (50) | 2 (50) | Jo1 (2), PL7 (1) |
OJ (N=2) | 0 | 0 | 0 | 0 | 0 | 0 | SRP (1) |
Jo1 (N=2) | 0 | 0 | 0 | 0 | 0 | 0 | SRP (1), Ku (1) |
EJ (N=1) | 1 (100) | 0 | 0 | 0 | 0 | 1 (100) | SRP (1) |
Our findings will enable researchers to correctly interpret serologies acquired by research testing using the Euroimmun platform. Our study highlights the importance of carefully defined cutoffs for assigning positive antibody status. That is, different autoantibody specificities may require different thresholds to define a positive result. The sensitivity of the Euroimmun assay readout appears to depend on the defined cut-off value. It is also possible that the increased sensitivity may come from measuring different epitopes that are not tested for by OMRF. However, data in Table 2 suggest that this increased sensitivity does not translate to clinically-meaningful results. One important distinction between the OMRF and Euroimmun assays is that Euroimmun reports antibody specificity to individual subunits for Mi-2 (α/β) and PM/Scl (75/100), while OMRF does not. OMRF utilizes immunoprecipitation of 35S-methionine-labeled lysate proteins for both anti-Mi-2 and anti-PM/Scl. Analysis of these immunoprecipitates by fluorography shows a series of multiple, distinctive protein bands that are readily recognizable to an experienced investigator. The presence of multiple bands with these autoantibodies (since both antigens are multi-protein complexes) allows definitive identification of the autoantibody (Personal Communication from Dr. Ira Targoff).
A limitation of this data is the inability to assess some of the more common myositis autoantibodies that were not consistently ordered at our center as part of the comprehensive OMRF panel (anti-TIF1γ, -NXP-2, and –MDA-5). Furthermore, this study included too few patients with some autoantibodies (e.g. anti-PL-7, -PL-12, -EJ and -OJ) to make any robust conclusions about the performance of the assays. Based on these results, we conclude that the same threshold may not be appropriate for determining a positive antibody result for all specificities tested using the Euroimmun panel and that sensitivity analyses should be conducted with different titer cut-offs.
Acknowledgement:
We gratefully acknowledge Dr. Ira Targoff for his assistance in describing OMRF laboratory methodology.
Funding: CM and ET are supported in part by the Jerome L Greene Foundation. This research was supported by the Intramural Research Program of the National Institute of Arthritis and Musculoskeletal and Skin Diseases of the National Institutes of Health. LCS, SKD and the Myositis Research Database are supported by the Huayi and Siuling Zhang Discovery Fund. IPF’s research was supported by a Fellowship from the Myositis Association. ET is also supported by the Rheumatology Research Foundation. JJP is supported by the K Bridge Award from the Rheumatology Research Foundation. The Johns Hopkins Rheumatic Diseases Research Core Center, where some of the assays were performed, is supported by NIH P30-AR-070254.
REFERENCES
- 1.Ronnelid J, Helmers SB, Storfors H, Grip K, Ronnblom L, Franck-Larsson K et al. Use of a commercial line blot assay as a screening test for autoantibodies in inflammatory myopathies. Autoimmun Rev 2009. September;9(1):58–61. [DOI] [PubMed] [Google Scholar]
- 2.Ghirardello A, Rampudda M, Okholm L, Bassi N, Tarricone E, Zampieri S et al. Diagnostic performance and validation of autoantibody testing in myositis by a commercial line blot assay. Rheumatology (Oxford) 2010. December;49(12):2370–4. [DOI] [PubMed] [Google Scholar]
- 3.Cavazzana I, Fredi M, Ceribelli A, Mordenti C, Ferrari F, Carabellese N et al. Testing for myositis specific autoantibodies: Comparison between line blot and immunoprecipitation assays in 57 myositis sera. J Immunol Methods 2016;433: 1–5. [DOI] [PubMed] [Google Scholar]
- 4.Espinosa-Ortega F, Holmqvist M, Alexanderson H, Storfors H, Mimori T, Lundberg IE et al. Comparison of autoantibody specificities tested by a line blot assay and immunoprecipitation-based algorithm in patients with idiopathic inflammatory myopathies. Ann Rheum Dis 2019. February 13 pii: annrheumdis-2018-214690. doi: 10.1136/annrheumdis-2018-214690. [Epub ahead of print] [DOI] [PubMed] [Google Scholar]
- 5.Matsushita T, Mizumaki K, Kano M, Yagi N, Tennichi M, Takeuchi A, et al. Antimelanoma differentiation-associated protein 5 antibody level is a novel tool for monitoring disease activity in rapidly progressive interstitial lung disease with dermatomyositis. Br J Dermatol 2017; 176:294–295. [DOI] [PubMed] [Google Scholar]
- 6.Abe Y, Matsushita M, Tada K, Yamaji K, Takasaki Y, Tamura N. Clinical characteristics and change in the antibody titres of patients with anti-MDA5 antibody-positive inflammatory myositis. Rheumatology (Oxford) 2017. September 1;56(9):1492–1497. [DOI] [PubMed] [Google Scholar]
- 7.Aggarwal R, Oddis C, Goudeau D, Koontz D, Qi Z, Reed A, et al. Autoantibody levels in myositis patients correlate with clinical response during B cell depletion with rituximab. Rheumatology 2016:55:991–999. [DOI] [PMC free article] [PubMed] [Google Scholar]