201-OR: Characterization of Type 1 Diabetes (T1D) Endotypes

To better understand how T1D heterogeneity relates to disease progression and therapeutic responsiveness, we studied 566 recent onset subjects (mean age at onset: 16.5 +/- 8.2) that had participated in the following TrialNet trials: MMF-DZB, anti-CD20; GAD-alum; CTLA4-Ig; Canakinumab; and ATG-GCSF. Our analysis began with a bioassay where plasma collected pre-intervention and within 100 days of diagnosis was used to induce transcription in a well-controlled peripheral blood mononuclear cell population. Microarray analysis identified 2,854 transcripts exhibiting high variation in at least 5 out of 6 trials. Hierarchical clustering and principal component analysis of these transcripts concordantly divided the participants into 2 major groups: Group 1 and Group 2. Analysis of post-baseline samples found group assignments conserved over time. Proportional analyses of the 406 subjects aged <18 years found Groups 1 and 2 distinct: subjects aged >14years segregated within Group 2 (39.9% vs 29.3%, p=0.03); subjects possessing neutral/low-risk HLA haplotypes segregated within Group 1 (20.5% vs 11.8%, p=0.04); placebo arm participants with the most rapid C-peptide rate of decline (<-0.15) segregated within Group 1 (20.3% vs 7.1%, p=0.04); and the subjects most responsive to anti-CD20 (p=0.01) and ATG-GCSF (p=0.07) segregated within Group 1. A local validation cohort (n=66, mean age at onset: 11.1 +/- 4.0) was analyzed by plasma induced transcription. A Random Forest model, which utilized the TrialNet data as a training set, was used to assign the local subjects to Group 1 or 2 (ROC: AUC=0.95; p=7.7E-9). Consistent with pathway analysis of the TrialNet transcriptomic data, ELISA analysis identified elevated plasma IFNγ, TNFα, IL-1β, IL-6, IL-12p70 levels (p<0.05) in Group 1. Flow cytometry measured higher percentages CD4+ CD45RO+ memory T cells (among total CD4 T cells, p=0.01) in Group 1. These analyses support the existence of two major T1D endotypes that exhibit differing immune phenotypes and therapeutic responsiveness.