Unsupervised clustering reveals a unique Treg profile in slow progressors to type 1 diabetes

Objective To profile CD4 + regulatory T cells (Tregs) in a well-characterised cohort of slow progressors to type 1 diabetes, individuals positive for multiple islet autoantibodies who remain diabetes-free for at least 10 years. Research Design and Methods Peripheral blood samples were obtained from extreme slow progressor individuals (n=8), with up to 32 years follow-up, and age and gender-matched to healthy donors. One participant in this study was identified with a raised HbA1c at the time of assessment, and was individually evaluated in the data analysis. PBMCs were isolated, from donors, and to assess frequency, phenotype and function of Tregs, multi-parameter flow cytometry and T cell suppression assays were performed. Unsupervised clustering analysis, FlowSOM and CITRUS, was used to evaluate Treg phenotypes. Results Treg mediated suppression of CD4 + effector T cells, from slow progressors was significantly impaired, compared to healthy donors (P<0.05). Effector CD4 T cells, from slow progressors, were more responsive to Treg suppression, compared to healthy donors, demonstrated by increased suppression of CD25 expression on effector CD4 T cells (P<0.05). Unsupervised clustering on memory CD4 T cells, from slow progressors, showed an increased frequency of activated-memory CD4 Tregs associated with increased expression of GITR, compared to healthy donors (P<0.05). The participant with a raised HbA1c had a different Treg profile, compared to slow progressors and the matched controls. Conclusions CD4 + Tregs from slow progressor individuals have a unique Treg signature. This report highlights the need for further study of Treg heterogeneity in individuals at-risk of developing type 1 diabetes.