IGRP-specific T-cells in the PLO increased with age in NOD mice (NOD 6-week-old=48.75±12.75, 15-week-old=3155±3166, p=0.0003). The number of IGRP-specific T-cells in the PLO and blood correlated with insulitis progression (r2=0.66, p<0.0001; r2=0.2773, p=0.0171 respectively), suggesting that the increase is dependent on proliferation in the islets. After 10 weeks of age in the NOD mouse, majority of IGRP-specific T-cells in the PLO, PLN and blood had an effector-memory surface phenotype of CD44hiCD62LloCD69loKLRG1hiCD127hi. These cells persisted in the PLO 6 weeks after diabetes onset, when antigen was depleted, consistent with the memory phenotype. IGRP-specific T-cells in the PLO also had effector function, and could lyse IGRP-coated targeT-cells in-vivo. We transferred naïve IGRP-specific T-cells into NOD mice and tracked them after 5 days, and found that this effector-memory phenotype was only acquired in the inflammatory environment of islets but not in the draining lymph nodes. Infiltrated islets were grafted into NOD.RAG1-/- mice and 4/5 recipients developed diabetes, showing that T-cells could migrate from the islets into PLO and re-invade endogenous islets. In a proportion of T1D patients, IGRP-specific T-cells were detectable at low frequencies in the blood, and in these patients the cells were of effector-memory phenotype.
Our data demonstrates that important changes to IGRP-specific T-cells during the pathogenesis of diabetes occur not in the draining lymph nodes but in the islets, where they expand and differentiate into effector-memory T-cells, and emigrate to the PLO, where they can report progression of islet pathology.