SAT0352 AN UNSUPERVISED ANALYSIS IDENTIFIES A SPECIFIC IMPACT OF BIOLOGICS ON T LYMPHOCYTE PHENOTYPES.

Background: It is currently not known if TNF or IL-17A inhibitors have an impact on immune cell frequencies in axial Spondyloarthritis (AxSpA). This question is important to understand the impact of biologics on the immune system. Data from clinical trials didn’t show significant modification on immune cells and especially on lymphocytes. But regarding the risk of infections linked to these treatments lymphocyte cell subsets are certainly disturbed. Moreover, biologics could affect subsets of cells with an unusual phenotype. Objectives: To identify the phenotype of cell subsets affected by biologics. Methods: We used an “unsupervised approach” to analyze CD4+ T cells and CD8+T cells subsets. Contrary to a “supervised approach”, this strategy takes advantages of the fluorescence emitted by all of the surface markers used to characterize the cells at the same time. The objective was on the one hand to overcome statistical problems related to the number of patients and the repetition of the tests and on the other hand to increase the sensitivity of the analysis by identifying and analyzing new cell populations. The first step was to cluster the cells based on a selection of 12 T cells markers characteristic of the classical cell subsets and the stage of maturation to obtain cell clusters with a phenotype based on the combination of these 12 markers. Then, we were able to describe “a posteriori” the change of frequency of the clusters identified. The second step was to create a visualization of the cells affected to confirm their existence in a classical flow cytometry gate. With this pipeline, we analyzed CD4 and CD8 T cells isolated from a group of AxSpA patients (n=7) before and after 3 months of TNF therapy and a group of patients (n=6) before and after 4 months of IL-17A therapy. Results: We observed that after biologics CD4 and CD8 T cells frequencies did not change but there was a redistribution of the different clusters analyzed. Specifically, we identified for CD4+T cells after anti TNF treatment an increase of 2 clusters (CD4+CD27+CD45RA+Va7.2intCD161int and CD4+CD27-CD45RA-CCR6+CD161int) and a decrease of 3 clusters (CD4+CD27+CD45RA+CRTH2intCD161int, CD4+ CD27+CD45RA+CXCR3+, CD4+CD27+CD45RA+gdint CD161int) and for CD8+T cells a decrease of 1 cluster after treatment (CD8+ CD27+CD45RA+CD161+CXCR3+) and an increase of 1 cluster (CD8+ CD27+CD45RA+). The clusters affected by anti-IL-17A therapy were different. For CD4+T cells, we identified a decrease of 2 clusters (CD4+CD27+CD45RA+CXCR5+ CD161+ and CD4+CD27+CD45RA- CXCR3+CCR6+CD161+) and an increase of 2 clusters (CD4+CD27+CD45RA+gdintCD161+, CD4+CD27+ CRTH2intCCR6+) and for CD8+T cells a decrease of 1 cluster (CD8+CD27+CD45RA+ CXCR3+CRTH2intCD161int) and an increase of 1 cluster (CD8+CD27+CD45RA+CXCR3intCD161-). Conclusion: We identified 5 different clusters in CD4+T cells affected by anti TNF and 4 by anti-IL-17A. We identified 2 clusters in CD8+T cells affected by anti TNF and 2 by anti-IL-17A. The phenotypes of these clusters were unexpected and raised new questions about the effect of biologics in AxSpA. We were also able to create a visualization of these cells affected by biologics in a “classic gating view” which will help us to perform scRNAseq. With this unique approach, we show an impact of biologics on the frequency of very specific subset of CD4+ and CD8+ T cells in AxSpA Disclosure of Interests: None declared