Suppressive activity of Nrp-1^lo iT reg cells in vitro and in vivo. (a) CFSE-labeled CD4⁺ T conv cells from 1B3.Thy1.1 mice were cultured in the presence of irradiated APCs and 1B3 peptide with or without different ratios of T reg cells from 1B3.RAG-sufficient or 1B3.RAG^−/− mice. 3 d later, cells were analyzed for CFSE dilution by flow cytometry. (b) T reg cells were sorted based on CD4, CD25, and GITR expression to select total T reg cells followed by an additional separation based on Nrp-1^hi (nT reg cells) and Nrp-1^lo (iT reg cells). 5 × 10⁴ 1B3.Thy1.2 CD4⁺CD25⁻ T conv cells (Teff) were co-transferred with or without similar number of Nrp-1^hi or Nrp-1^lo or total T reg cells from 1B3.Thy1.1 mice in the RAG^−/− recipients and monitored for EAE development. The data are presented as the mean EAE score, with error bars showing SD, and are representative of four independent experiments. n = 12 for Teff+nT reg; n = 13 for Teff+iT reg; n = 11 for Teff; and n = 10 for Teff+Total T reg cells. *, P < 0.05; **, P < 0.001 for Teff+nT reg versus Teff+iT reg. (c) Foxp3 staining in the T reg compartment in different transfer groups at the end of the disease (bars show means). (d) Prediabetic NOD.CD28^−/− mice were injected with 5 × 10⁴ Nrp-1^hi nT reg (n = 8) or Nrp-1^lo iT reg cells (n = 8) from BDC.2.5-TCR-Tg mice or left untreated (n = 12 each). The development of diabetes was monitored by measuring blood glucose of individual mice. Mice with two consecutive readings of blood glucose >250 mg/dl were labeled diabetic.