TGF-β mediates the induction of Foxp3 in Ag-specific Th2 memory cells with sequential addition of ATRA and rapamycin in the presence of IL-4 and IFN-γ neutralization. (A) Purified OVA-specific Th2 memory cells were stimulated with spleen DCs and OVA_323–339 peptides under indicated conditions. TGF-β (20 ng/mL), ATRA (100 nM), and rapamycin (Rapa; 20 nM) were added at the indicated concentrations. At day 4 of culture, Foxp3 expression levels in gated CD4⁺KJ1-26⁺ cells were analyzed by flow cytometry. (B) Purified OVA-specific Th2 memory cells were stimulated with plate-bound anti-CD3 (2 μg/mL) and soluble anti-CD28 (1 μg/mL) Abs under the indicated conditions. The final dose of TGF-β was 5 ng/mL. (C) The mean ± SEM percentage of Foxp3⁺ cells in each group in B are plotted (**P < 0.005). Data are representative of at least three independent experiments.

Ag-specific Th2 memory-derived Foxp3⁺ cells are genuine regulatory T cells. (A) Reactivated Th2 memory cells or Th2 memory-derived Tregs were analyzed by flow cytometry for surface expression of CD25, CD39, FR4, GITR, PD-1, and ICOS, together with Foxp3. (B and C) OVA-specific Th2 memory-derived Treg-mediated suppression of proliferation (B) and cytokine production (C) of OVA-specific Th2 memory cells (T effectors; Teff) were examined as described in SI Materials and Methods. Numbers inside the histograms indicate the percentage of cells dividing more than three times in each group. Data are representative of at least three independent experiments.

Ag-specific Th2 memory-derived Foxp3⁺ cells lose their Th2-type effector function. (A) Purified OVA-specific Th2 memory cells were stimulated with spleen DCs and OVA_323–339 peptides under the indicated conditions for 4 days, and the correlation of Foxp3 and GATA-3 expression on gated CD4⁺KJ1-26⁺ cells was analyzed by flow cytometry. (B) OVA-specific Th2 memory cells were cultured as indicated in the presence of anti-CD3 and anti-CD28 Abs for 5 days. The resultant cells were stained for Foxp3 and either IL-4 or IL-13 after restimulation. (C) Purified OVA-specific Th2 memory, reactivated Th2 memory (“None” in B), and Th2 memory-derived Tregs (TGF-β/αIL-4/αIFN-γ/ATRA/Rapa in B) were restimulated for anti-CD3 Ab (2 μg/mL) for 24 h. Culture supernatants from triplicate samples were analyzed for cytokine production by ELISA. (D) OVA-specific Th2 memory cells were cultured as indicated in the presence of anti-CD3 and anti-CD28 Abs for 5 days. The resultant cells were stained for Foxp3 and IL-9 after restimulation. Data are representative of at least three independent experiments.

Th2 memory-derived Tregs can modulate Th2 memory-induced airway hyperreactivity and eosinophilia. (A–E) Th2 memory-mediated airway inflammation was induced as described in Fig. S9A. (A) One day after the last challenge, AHR was analyzed, and the means ± SEM of the mice in each group are plotted. Values for enhanced pause (Penh) are given (*P < 0.05; **P < 0.005). (B) Twenty-four hours after AHR analysis, BALF and serum samples were collected. The number of total BALF cells, eosinophils, monocytes, and lymphocytes were analyzed. (C) Serum levels of OVA-specific IgE were detected by ELISA. (D) MdLN cells in each group were restimulated with OVA_323–339 peptides for 6 h. IL-4, IL-5, or IL-13 production, together with Foxp3 expression on gated CD4⁺KJ1-26⁺ cells, was analyzed by flow cytometry. (E) Foxp3 and GATA-3 expression on gated CD4⁺KJ1-26⁺ cells from MdLN, BALF, lung, and spleen were analyzed by flow cytometry. (F) CCR4 and CCR7 expression on gated CD4⁺KJ1-26⁺Foxp3⁺ cells in BALF from Th2 memory-derived Treg-transferred mice were analyzed by flow cytometry. (G–J) Airway inflammation was induced as described in Fig. S9D. AHR (G), BALF eosinophils (H), and OVA-specific IgE in sera (I) were analyzed. (J) Foxp3 expression of CD4⁺KJ1-26⁺ cells in MdLN and BALF was determined by flow cytometry. These experiments included five to six mice per group. Data are representative of at least two independent experiments.