Two functional subsets of FOXP3+ regulatory T cells in human thymus and periphery

Abstract

Previous studies suggest that thymus produces a homogenous population of natural regulatory T (Treg) cells that express a transcriptional factor FOXP3 and control autoimmunity through a cell-contact-dependent mechanism. We found two subsets of FOXP3+ natural Treg cells defined by the expression of the costimulatory molecule ICOS in the human thymus and periphery. Whereas the ICOS+FOXP3+ Treg cells used interleukin-10 to suppress dendritic cell function and transforming growth factor (TGF)-beta to suppress T cell function, the ICOS-FOXP3+ Treg cells used TGF-beta only. The survival and proliferation of the two subsets of Treg cells were differentially regulated by signaling through ICOS or CD28, respectively. We suggest that the selection of natural Treg cells in thymus is coupled with Treg cell differentiation into two subsets imprinted with different cytokine expression potentials and use both cell-contact-dependent and independent mechanisms for immunosuppression in periphery.

Figure 1

Two subsets of Foxp3⁺ T_R defined by ICOS expression in thymus. a, Double staining of Foxp3 (red) and ICOS (blue) in the human thymus showing the Foxp3⁺ICOS⁺ and Foxp3⁺ICOS⁻ two T_R subsets located in the thymic medulla at a low magnification (100×). The scale bar, 50 μm. b, Double immunofluorescence staining of Foxp3 (red) and ICOS (green) performed on the same section of thymus further confirmed the two Foxp3⁺ T-cell populations: nuclear Foxp3⁺ cytoplasmic ICOS⁺ T cells and nuclear Foxp3⁺ cytoplasmic ICOS⁻ T cells. c, Flow cytometric analysis of ICOS and CD25 expression on CD4⁺CD8⁻ thymocytes and Foxp3 expression in sorted two T_R subsets. d, Flow cytometric analysis of intracellular cytokines in each thymic T_R subset after two rounds of culturing with anti-CD3 antibody in the presence of IL-2 and IL-7 on parental L cells or ICOSL-L cells for 5 days. e, Flow cytometric analysis of membrane TGF-β on each thymic T_R subset after culturing with anti-CD3 antibody in the presence of IL-2 and IL-7 on parental L cells or ICOSL-L cells for 5 days. The numbers in the histograms indicate the mean fluorescence intensity, which is calculated by subtraction of mean fluorescence intensity with the isotype control from that with the anti-TGF-β antibody. f. Proliferative response of CD4⁺ CD25⁻ T cells and thymic CD25^high ICOS⁺ CD4⁺ T_R or CD25^high ICOS⁻ CD4⁺ T_R primed with anti-CD3 antibody in the presence of IL-2 and IL-7 on ICOSL-L cells for 5 days and their mixtures in response to T cell-depleted peripheral blood mononuclear cells was assessed by [³H]thymidine incorporation. Ratio of T_R to CD4⁺ CD25⁻ T cells is 1 to 2. Error bars represent the s.e.m. of triplicate wells. Similar results were observed in four (a, b) and three (c–f) experiments, and the results of a representative experiment are shown.

Figure 2

Two subsets of Foxp3⁺ T_R defined by ICOS expression in periphery. a,c, Double staining of Foxp3 (red) and ICOS (blue) in the human tonsil (a) and lymph node (c) showing two Foxp3⁺ T cell populations: Foxp3⁺ICOS⁺ T cells and Foxp3⁺ICOS⁻ T cells at a low magnification (100×). All scale bars, 50 μm. b,d, Immunofluorescence staining of Foxp3 (red) and ICOS (green) performed on the same section of tonsil (b) and lymph node (d) showing two Foxp3⁺ T-cell populations: nuclear Foxp3⁺ cytoplasmic ICOS⁺ T_R and nuclear Foxp3⁺ cytoplasmic ICOS⁻ T_R. e, Human peripheral blood CD25^high CD4⁺ T_R were separated into ICOS⁺ and ICOS⁻ subpopulations. Foxp3 expression in each subset was determined by flow cytometry. f, h, Phenotypical analysis of two subsets of T_R in adult blood (f) and cord blood (h). Intranuclear Foxp3 and intracytoplasmic CTLA4 and the cell surface markers on sorted blood CD25^high ICOS⁺ CD4⁺ T_R and CD25^high ICOS⁻ CD4⁺ T_R were determined by flow cytometry. g, Percentages of two subsets of T_R in 6 thymus and 18 adult blood from healthy donors. Percentages of CD25^high ICOS⁺ CD4⁺ T_R and CD25^high ICOS⁻ CD4⁺ T_R among CD4⁺ T cells were determined by flow cytometry. Bars represent the group means. Statistical significance was determined using paired Student’s t-test. Similar results were observed in three (a–d) and five (e) experiments, and at least three experiments for each marker (f, h), and the results of a representative experiment are shown.

Figure 3

Different capacity of two T_R subsets to produce cytokines. Blood CD25^high ICOS⁺ T_R, CD25^high ICOS⁻ T_R, and whole CD25^high CD4⁺ T_R, as well as CD25⁻CD45RO⁻ naïve and CD25⁻CD45RO⁺ memory T cells were isolated by same donor. a, b, Analysis of IL-10 and IL-2 production by freshly isolated T cell subsets (a) or T cells primed for 5 days with anti-CD3 antibody in the presence of IL-2 on parental L cells (b) by intracellular cytokine staining (left) and ELISA (right). ELISA data are means ± s.e.m. of four independent experiments. c–e, Blood T_R and other T-cell subsets were cultured for 5 days with anti-CD3 antibody in the presence of IL-2 on parental L cells or ICOSL-L cells. Cytokine production of primed blood T_R subsets were analyzed by flow cytometry (c, d) and ELISA (e). ELISA data are means ± s.e.m. of four independent experiments. Statistical significance was determined using paired Student’s t-test. *: p<0.05, **: p<0.01 (e). f, g, Flow cytometric analysis of CD25, ICOS, CTLA4, and Foxp3 expression with CFSE labeling by CD25⁻CD45RO⁺ memory T cells and T_R subsets after culturing for 5 days with anti-CD3 antibody in the presence of IL-2. h, Flow cytometric analysis of IL-10 and Foxp3 expression in T-cell subsets after culturing for 5 days with anti-CD3 antibody in the presence of IL-2 on ICOSL-L cells. Similar flow cytometric results were observed in three experiments (a–d, f, g, h), and the results of a representative experiment are shown.

Figure 4

Foxp3⁺ICOS⁺CD25⁺ T_R have the capacity to produce the highest levels of IL-10. a, Blood CD4⁺ T cells were separated into ICOS⁺CD45RO⁺ memory, ICOS⁻CD45RO⁺ memory, and ICOS⁻CD45RO⁻ naive T cells based on the expression of ICOS and CD45RO. Flow cytometric analysis of intracellular cytokines in each T-cell subset after culturing with anti-CD3 antibody in the presence of IL-2 on parental L cells or ICOSL-L cells for 5 days. b, Flow cytometric analysis of Foxp3 expression by five subsets of CD4⁺CD45RO⁺ memory subpopulations: CD25⁻ICOS⁻CD45RO⁺ memory, CD25⁻ICOS⁺ follicular TH-like, CD25⁺ICOS⁺ T_R, CD25⁺ICOS⁻ T_R, and CD25^lowCD45RO⁺CRTH2⁺ TH2 memory T cells. c, d, Five blood CD4⁺ memory T-cell subpopulations were cultured with anti-CD3 antibody in the presence of IL-2 on parental L cells or ICOSL-L cells for 5 days. IL-10 production by T-cell subsets were analyzed by flow cytometry (c) and ELISA (d). ELISA data are means ± s.e.m. of four independent experiments. Similar flow cytometric results were observed in four experiments (a–c), and the results of a representative experiment are shown.

Figure 5

Different capacity of two subsets of T_R to express TGF-β. a, Flow cytometric analysis of membrane TGF-β on the primed blood T_R subsets after culturing with anti-CD3 antibody in the presence of IL-2 on parental L cells or ICOSL-L cells for 5 days. The numbers in the histograms indicate the mean fluorescence intensity, which is calculated by subtraction of mean fluorescence intensity with the isotype control from that with the anti-TGF-β antibody. Similar results were observed in three experiments, and the results of a representative experiment are shown. b, The real-time quantitative RT-PCR analysis for TGFB mRNA expression in blood CD25⁻CD45RO⁺CD4⁺ memory T cells, and two T_R subsets after culturing with anti-CD3 antibody in the presence of IL-2 for 5 days. Data are means ± s.e.m. of three independent experiments.

Figure 6

Suppressive function of two subsets of T_R. CD4⁺ naive T cells and autologous CD25^high ICOS⁺ CD4⁺ T_R or CD25^high ICOS⁻ CD4⁺ T_R were sorted from human peripheral blood. CD4⁺ naive T cells were used as responder, and primed T_R after culturing with anti-CD3 antibody in the presence of IL-2 on ICOSL-L cells for 5 days were used as suppressor. Mixtures of T-cell populations were cultured for 5 days with allogeneic monocyte-derived DCs as stimulator. The results using primed T_R as suppressor were shown here, and that using freshly isolated T_R were shown in Supplementary Figure 1. a, Ratios of T_R to naive T cells are shown in the figure. Proliferation was assessed by [³H]thymidine incorporation and error bars represent the s.e.m. of triplicate wells. b, c, Ratio of T_R to naive T cells is 1 to 2. Neutralizing anti-IL-10 plus IL-10 receptor antibodies and/or TGF-β receptor type I (TβRI) inhibitor were added into T-cell culture. Proliferation was assessed by [³H]thymidine incorporation and error bars represent the s.e.m. of triplicate wells. Statistical significance was determined using Student’s t-test. **: p<0.01 (b). CD4⁺ naive T cells were labeled with CFSE before culture, and cells after culturing were analyzed by flow cytometry, gated on CD4⁺CFSE⁺ cells (c). d, Indicated T-cell populations were cultured in Transwell at 1:2 ratio of T_R to naive T cells with allogeneic monocyte-derived DCs. Proliferation was assessed by [³H]thymidine incorporation and error bars represent the s.e.m. of triplicate wells. e, Primed T_R, naive T cells, and allogeneic monocyte-derived DCs were cultured at 1:1:1 ratio. After 4 days of culture, CD86 expression on DCs were analyzed by flow cytometry, gated on CD11c⁺ cells. Filled histogram indicates CD86 expression on DCs in coculture with naive T alone, and line histogram indicates CD86 expression on DCs in coculture with naive T and T_R. Dot histogram indicates isotype control on DCs. Similar results were observed in three experiments (a–e), and the results of a representative experiment are shown.

Figure 7

Survival and expansion of two T_R subsets were differentially regulated. a, b, Sorted blood CD25^high ICOS⁺ CD4⁺ T_R and CD25^high ICOS⁻ CD4⁺ T_R were cultured in the absence of IL-2 on the idicated condtions for 5 days. Flow cytometric analysis of annexin V staining on each T_R subset after culturing. The numbers in the histograms indicate the percentage of cells in the gate (a). The number of viable T cells counted by trypan-blue exclusion is shown in the bar graph (b). c, 5 × 10⁴ T_R subsets and CD4⁺ naive T cells were cultured in the presence of IL-2 on the indicated conditions for 5 days. The viable cell number of each T_R subset counted by trypan-blue exclusion is shown. Bars represent means of four experiments. d, 5 × 10⁴ T_R subsets were cultured with blood pDCs or mDCs (DC/T cell ratio of 1:2) for 4 days in the presence of neutralizing anti-ICOSL mAb or neutralizing anti-CD80/CD86 mAbs. The viable cell number of each T_R subset counted by trypan-blue exclusion is shown. Similar results were observed in three (a) and four (b, d) experiments, and the results of a representative experiment are shown.