AB0343 THE CHARACTERISTICS OF T CELLS IN SYSTEMIC LUPUS ERYTHEMATOSUS PATIENTS WITH ANXIETY BASED ON MACHINE LEARNING

Abstract

Background: Systemic lupus erythematosus (SLE) is an autoimmune disease, the immune system of patients to be disordered, especially in T cell subsets1. They were prone to mental diseases, anxiety particularly, which lead to suicide2. The recent study had reported that CD4+ T cells in the peripheral blood played the key role in like anxiety behavior of mice3. Although there showed that the level of serum TNF-alpha in SLE patients with anxiety was higher than without anxiety4, finding the important special mediators especially in T cell subsets was still necessary for the prevention of anxiety in SLE patients. Objective(s): In total, 108 SLE patients, which met the diagnostic criteria of the American Society of rheumatology (v1997), were enrolled in this study from Affilliated Hospital of Nantong University, China. Exclusion criteria included other autoimmune diseases and active infection (including hepatitis B or C virus, Epstein-Barr virus, human immunodeficiency virus or Mycobacterium tuberculosis infection). Method(s): We surveyed the abundance of 74 immune cell subpopulations from 108 SLE patients using flow cytometry, and investigated their differences between patients with and without anxiety (24 versus 84). Moreover, machine learning including Lasso regression, Random forest (RF) and Sparsity partial least squares discriminant analysis (sPLS-DA) was employed to build models and futher selected important features for the classification of SLE patients with anxiety. Result(s): SLE patients with anxiety showed higher body mass index (BMI) and lower quality of life. In their peripheral blood, the proportion of internal cell subsets composition of Th cell and Treg cells changed. By machine learning, we finally found that BMI and PD1-CD28-Treg played important rules to developing lupus anxiety. Conclusion(s): In this study, machine learning was applied to build models to select the most important T cell subset in SLE patients with anxiety. These findings suggested that BMI and imbalance of PD1-CD28-Treg containing effector memory Treg cells and effector Treg cells mostly played important roles in the development of SLE anxiety.