Abstract
Type 1 diabetes (T1D) is characterized by autoimmune destruction of insulin producing β-cells. The time from onset of islet autoimmunity to manifest clinical disease can vary widely in length, and it is fairly uncharacterized both clinically and immunologically. In the current study, peripheral blood mononuclear cells from autoantibody-positive children with high risk for T1D, and from age-matched healthy individuals, were analyzed by mass cytometry using a panel of 32 antibodies. Surface markers were chosen to identify multiple cell types including T, B, NK, monocytes, and DC, and antibodies specific for identification of differentiation, activation and functional markers were also included in the panel. By applying dimensional reduction and computational unsupervised clustering approaches, we delineated in an unbiased fashion 132 phenotypically distinct subsets within the major immune cell populations. We were able to identify an effector memory Treg subset expressing HLA-DR, CCR4, CCR6, CXCR3, and GATA3 that was increased in the high-risk group. In addition, two subsets of NK cells defined by CD16+ CD8+ CXCR3+ and CD16+ CD8+ CXCR3+ CD11c+ were also higher in the same subjects. High-risk individuals did not show impaired glucose tolerance at the time of sampling, suggesting that the changes observed were not the result of metabolic imbalance, and might be potential biomarkers predictive of T1D.