Low frequency of GITR+ T cells in ex vivo and in vitro expanded Treg cells from type 1 diabetic patients.

Reported alterations in T(reg) cells from type 1 diabetes (T1D) patients led us to a revision of their phenotypical features compared with controls. A fine cytometric analysis was designed for their characterization, using a panel of markers including FOXP3, CTLA4, glucocorticoid-induced TNFR family related (GITR) and CD127. The frequency of peripheral CD4(+)CD25(hi) T(reg) cells was similar between samples. However, the yield of sorted T(reg) cells was significantly lower in patients than in controls. When comparing the T(reg)-cell phenotype between samples, the only difference concerned the expression of GITR. A significant decrease of GITR(+) cells and GITR mean fluorescence intensity within the T(reg)-cell population, and to a lesser extent in the effector population, was observed in T1D compared with controls. Moreover, GITR expression was analyzed in several conditions of T-cell activation and differences were only observed in T1D T(reg) cells versus controls when responding to sub-optimal stimulation, that is, soluble anti-CD3 or medium alone but not in the presence of anti-CD3-/anti-CD28-coated beads. However, expanded T1D T(reg)-cell-mediated suppression was as efficient as that mediated by their control counterparts, showing no association between their regulatory capacity and the reduced GITR. Our results show a higher susceptibility to apoptosis in patients' versus controls' T(reg) cells, suggesting that GITR is a T(reg)-cell marker that would be primarily involved in T(reg)-cell survival rather than in their suppressor function.

TCR bias of in vivo expanded T cells in pancreatic islets and spleen at the onset in human type 1 diabetes.

Autoreactive T cells, responsible for the destruction of pancreatic ß cells in type 1 diabetes, are known to have a skewed TCR repertoire in the NOD mouse. To define the autoreactive T cell repertoire in human diabetes, we searched for intraislet monoclonal expansions from a recent onset in human pancreas to then trace them down to the patient's peripheral blood and spleen. Islet infiltration was diverse, but five monoclonal TCR ß-chain variable expansions were detected for Vß1, Vß7, Vß11, Vß17, and Vß22 families. To identify any sequence bias in the TCRs from intrapancreatic T cells, we analyzed 139 different CDR3 sequences. We observed amino acid preferences in the NDN region that suggested a skewed TCR repertoire within infiltrating T cells. The monoclonal expanded TCR sequences contained amino acid combinations that fit the observed bias. Using these CDR3 sequences as a marker, we traced some of these expansions in the spleen. There, we identified a Vß22 monoclonal expansion with identical CDR3 sequence to that found in the islets within a polyclonal TCR ß-chain variable repertoire. The same Vß22 TCR was detected in the patient's PBMCs, making a cross talk between the pancreas and spleen that was reflected in peripheral blood evident. No other pancreatic monoclonal expansions were found in peripheral blood or the spleen, suggesting that the Vß22 clone may have expanded or accumulated in situ by an autoantigen present in both the spleen and pancreas. Thus, the patient's spleen might be contributing to disease perpetuation by expanding or retaining some autoreactive T cells.