Interpretation of an extended autoantibody profile in a well-characterized Australian systemic sclerosis (Scleroderma) cohort using principal components analysis

Kevin Patterson, Peter J Roberts-Thomson, S Lester, Juann Tan, Paul Hakendorf, Maureen Rischmueller, Jane Zochling, Joanne M Sahhar, Peter T Nash, Janet Roddy, Catherine L Hill, Mandana Nikpour, Wendy M R Stevens, Susanna M Proudman, Jennifer G Walker

Research output: Contribution to journalArticleResearchpeer-review

61 Citations (Scopus)


Objective To determine the relationships between systemic sclerosis (SSc)-related autoantibodies, as well as their clinical associations, in a well-characterized Australian patient cohort. Methods Serum from 505 Australian SSc patients were analyzed with a commercial line immunoassay (EuroLine; Euroimmun) for autoantibodies to centromere proteins CENP-A and CENP-B, RNA polymerase III (RNAP III; epitopes 11 and 155), the 90-kd nucleolar protein NOR-90, fibrillarin, Th/To, PM/Scl-75, PM/Scl-100, Ku, topoisomerase I (topo I), tripartite motif-containing protein 21/Ro 52, and platelet-derived growth factor receptor. Patient subgroups were identified by hierarchical clustering of the first 2 dimensions of a principal components analysis of quantitative autoantibody scores. Results were compared with detailed clinical data. Results A total of 449 of the 505 patients were positive for at least 1 autoantibody by immunoblotting. Heatmap visualization of autoantibody scores, along with principal components analysis clustering, demonstrated strong, mutually exclusive relationships between CENP, RNAP III, and topo I. Five patient clusters were identified: CENP, RNAP III strong, RNAP III weak, topo I, and other. Clinical features associated with CENP, RNAP III, and topo I were consistent with previously published reports concerning limited cutaneous and diffuse cutaneous SSc. A novel finding was the statistical separation of RNAP III into 2 clusters. Patients in the RNAP III strong cluster had an increased risk of gastric antral vascular ectasia, but a lower risk of esophageal dysmotility. Patients in the other cluster were more likely to be male and to have a history of smoking and a history of malignancy, but were less likely to have telangiectasia, Raynaud s phenomenon, and joint contractures. Conclusion Five major autoantibody clusters with specific clinical and serologic associations were identified in Australian SSc patients. Subclassification and disease stratification using autoantibodies may have clinical utility, particularly in early disease. ? 2015, American College of Rheumatology.
Original languageEnglish
Pages (from-to)3234 - 3244
Number of pages11
JournalArthritis & Rheumatology
Issue number12
Publication statusPublished - 2015

Cite this