OP0231 MASS CYTOMETRY DATA RECLASSIFY SYSTEMIC AUTOIMMUNE DISEASE PATIENTS IN PHENOTYPICALLY DISTINCTIVE GROUPS

Abstract

Background: Systemic lupus erythematosus (SLE), rheumatoid arthritis (RA), systemic sclerosis (SSC), Sjögren's syndrome (SJS), mixed connective tissue disease (MCTD), primary antiphospholipid syndrome (PAPS) and undifferentiated connective tissue disease (UCTD) are classifed as systemic autoimmune diseases (SADs). They are diagnosed based on different clinical and laboratory criteria. Due to their high Internal heterogeneity and overlapping symptoms, SADs are difficult to diagnose. Therefore, molecular and cellular-based studies need to be undertaken to precisely classify the patients. Mass cytometry is a single-cell proteomics technology that measures approximately 50 markers per cell, thus it is a suitable tool to perform deep-pheno-typing studies in SADs. Objectives: Explore differences and similarities between SADs and build reclas-sifcation framework using high-dimensional cytometry data. Methods: The whole blood samples collected from 129 individuals, including patients and controls were stained with a 39-plex antibody panel and acquired in 9 batches on a CyTOF (HELIOS) instrument. Data were cleaned, and normalized for batch effects using semi-automated cytof analysis pipeline. Cell frequencies and median signal intensities (MSI) for each population were extracted using FlowSOM for mononuclear cells (PBMC) and Phenograph for granulocytes. Secretion of 44 cytokines and chemokines were analyzed using a multiplexed luminex assay. Diseases were compared by Kruskal-Wallis analysis and hierarchical clustering and reclassifcation was done using unsuper-vised k-means clustering. Cytokine analysis across clusters was performed using Kruskal-Wallis test. Results: Differently expressed features were observed between patient groups, regarding frequency of classical monocytes, B and T cells subpopulations, mature and immature granulocytes and intensities of CD38, HLA-DR and CD95 across various populations. However, none of them were disease specifc. K-means clustering identifed four patient clusters, which were composed by a mixture of different diagnosis. Cluster C1 was characterized by increased levels of circulating cells from PBMC compartment, and lower activation of different populations of the T cell compartment. It presented lower frequency in multiple granulocyte populations and the highest expression of CD95 and CD38. This cluster was also associated with antimalarial and steroid treatment. Clusters C1 and C2 were exactly opposite to each other, cluster C3 was characterized by intermediate features between C1 and C2 and cluster C4 could be considered as undifferentiated, mixed group. Higher production of TNFα, IL-10 and IP-10 were found in patients from C1 compared to C2, suggesting more active pheno-type in C1 and physiological one in C2. The cytokine levels were independent of the treatment. Conclusion: We constructed a patient reclassification framework using cell frequencies and expression levels of functional markers. To our knowledge this is the first time when 7 different SADs were compared using mass cytometry. In agreement with other reports we did not detect any disease-specific cellular markers. Distribution of diagnosis across different clusters confirms diseases heterogeneity. Patients can be classified into phenotypically similar groups, that could potentially benefit from the same line of treatment.