The lineage stability and suppressive program of regulatory T cells require protein O-GlcNAcylation

Abstract

Regulatory T (Treg) cells control self-tolerance, inflammatory responses and tissue homeostasis. In mature Treg cells, continued expression of FOXP3 maintains lineage identity, while T cell receptor (TCR) signaling and interleukin-2 (IL-2)/STAT5 activation support the suppressive effector function of Treg cells, but how these regulators synergize to control Treg cell homeostasis and function remains unclear. Here we show that TCR-activated posttranslational modification by O-linked N-Acetylglucosamine (O-GlcNAc) stabilizes FOXP3 and activates STAT5, thus integrating these critical signaling pathways. O-GlcNAc-deficient Treg cells develop normally but display modestly reduced FOXP3 expression, strongly impaired lineage stability and effector function, and ultimately fatal autoimmunity in mice. Moreover, deficiency in protein O-GlcNAcylation attenuates IL-2/STAT5 signaling, while overexpression of a constitutively active form of STAT5 partially ameliorates Treg cell dysfunction and systemic inflammation in O-GlcNAc deficient mice. Collectively, our data demonstrate that protein O-GlcNAcylation is essential for lineage stability and effector function in Treg cells.

Fig. 1

O-GlcNAc-cycling enzymes regulate FOXP3 stability in vitro. a, b Mean fluorescence intensity (MFI) of OGT (a) and O-GlcNAcylation (b) in CD4⁺CD25⁻naïve T cells, CD4⁺CD25⁺FOXP3⁺ Treg cells and corresponding Fluorescence Minus One (FMO) negative controls. c Treg cells isolated from wildtype mice were stimulated with or without anti-CD3/CD28 beads for 24 h ex vivo (n = 3). MFI of FOXP3 and O-GlcNAcylation was analyzed in CD4⁺CD25⁺ FOXP3⁺ Treg cells. d, e HEK 293 cells were transfected FOXP3 together with OGT or OGA (d) or treated with inhibitors of OGT (ST045849) or OGA (TMG) (e). FOXP3 O-GlcNAcylation was determined by immunoprecipitation followed by western blotting. f FOXP3 stability was determined by treatment of cycloheximide (CHX) in combination with MG132 with or without OGT overexpression, DMSO was used as a control for MG132. g–i FOXP3 stability was determined in the presence of OGT/OGA overexpression (g), TMG (h), or ST045849 (i). Data are shown as mean ± s.e.m. *p < 0.05 by unpaired student’s t-test

Fig. 2

O-GlcNAcylation stabilizes the FOXP3 protein in Treg cells. a–d Treg cells isolated from Ubc-CreER⁺Ogt^fl/Y mice were treated with 4-OHT for 3⁻day ex vivo, ethanol was used as the control, n = 4 each group. MFI of FOXP3 in CD4⁺FOXP3⁺ Treg cells was analyzed in a and quantified in b. Representative flow cytometry of CD4⁺ FOXP3⁺Treg cells was plotted in c and the frequencies of Treg cells were shown in d. e FOXP3 or FOXP3-5A were transfected in HEK 293 cells with or without OGT. FOXP3 O-GlcNAcylation was determined by immunoprecipitation followed by Western blotting. f FOXP3 and FOXP3-5A stability was determined by treatment of CHX in the presence of OGT overexpression. g, h CD4⁺CD25⁻ naïve T cells isolated from wildtype mice were infected with retroviruses expressing FOXP3 or FOXP3-5A in the presence of anti-CD3/CD28 beads (n = 4). MFI of FOXP3 was analyzed (g) and quantified (h) in CD4⁺FOXP3⁺ Treg cells. Data are shown as mean ± s.e.m. *p < 0.05; **p < 0.01 by unpaired student’s t-test (b, d) and paired student’s t-test (h)

Fig. 3

OGT-deficiency in Treg cells leads to a scurfy phenotype in male mice. a, b Representative images of 4-week-old Foxp3^YFP-Cre/Y Ogt^wt/Y and Foxp3^YFP-Cre/Y Ogt^fl/Y mice (a) and peripheral LNs and spleen (b). c, d Body weight curve (c, n = 5) and survival curve (d, n = 6) of Foxp3^YFP-Cre/Y Ogt^wt/Y and Foxp3^YFP-Cre/Y Ogt^fl/Y mice. e Representative images of H&E staining of the colon, skin, liver, and lung from 4-week-old Foxp3^YFP-Cre/Y Ogt^wt/Y and Foxp3^YFP-Cre/Y Ogt^fl/Y mice. f, g Representative flow cytometry plots showing CD44^hiCD62L^lo effector T cells among CD4⁺ (top) and CD8⁺ (bottom) T cells in the LNs (f) and quantification of the frequencies in the LNs and spleen of 2-week-old Foxp3^YFP-Cre/Y Ogt^wt/Y and Foxp3^YFP-Cre/Y Ogt^fl/Y mice, at least n = 6 each group (g). h–k Representative flow cytometry plots showing T-BET⁺, GATA3⁺ and RORγT⁺ cells among CD4⁺ T cells in the LNs (h) and quantification of the frequencies of T-BET⁺ cells (i), GATA3⁺ cells (j) and RORγT⁺ cells (k) in the LNs and spleen of 2-week-old Foxp3^YFP-Cre/Y Ogt^wt/Y and Foxp3^YFP-Cre/Y Ogt^fl/Y mice were shown, n = 8 each group. l Frequencies of IFNγ⁺, IL-4⁺ and IL-17A⁺ cells in CD4⁺FOXP3^- T cells stimulated with PMA/lonomycin in the LNs of 2-week-old Foxp3^YFP-Cre/Y Ogt^wt/Y and Foxp3^YFP-Cre/Y Ogt^fl/Y mice, at least n = 3 each group. Data are shown as mean ± s.e.m. *p < 0.05; **p < 0.01; ***p < 0.001 by unpaired student’s t-test (c, g, and i–l) and Kaplan-Meier Analysis (d)

Fig. 4

Protein O-GlcNAcylation stabilizes the Treg cell lineage. a, b Flow cytometry of YFP⁺CD25⁺ cells among CD4⁺ T cells in the LNs (a) and frequencies of YFP⁺CD25⁺ cells in the LNs and spleen from 2-week-old Foxp3^YFP-Cre/Y Ogt^wt/Y and Foxp3^YFP-Cre/Y Ogt^fl/Y mice, n = 8 each group (b). c, d Flow cytometry of FOXP3⁺CD25⁺ cells among CD4⁺ T cells (c) and MFI of FOXP3 in FOXP3⁺CD25⁺ Treg cells (d, n = 3–4) in the LNs from 2-week-old Foxp3^YFP-Cre/Y Ogt^wt/Y and Foxp3^YFP-Cre/Y Ogt^fl/Y mice. e MFI of FOXP3 in OGT-sufficient and -deficient Treg cells in the LNs from Foxp3^YFP-Cre/Y Ogt^fl/fl mice, n = 5. f Histogram of FOXP3 expression in CD4⁺CD25⁺YFP⁺ Treg cells in the LNs from 2-week-old Foxp3^YFP-Cre/Y Ogt^wt/Y and Foxp3^YFP-Cre/Y Ogt^fl/Y mice, n = 6 each group. g Flow cytometry and quantification of the frequencies of Td-tomato⁺GITR⁺FOXP3⁻ ex-Treg cells among CD4⁺ T cells in the LNs from Foxp3^{eGFP-Cre-ERT2/Y} Ogt^wt/Y Rosa26^tdTomato/wt and Foxp3 ^{eGFP-Cre-ERT2/Y} Ogt^fl/Y Rosa26^tdTomato/wt mice, n = 4. h, i T-BET⁺, GATA3⁺ and RORγT⁺ cells among ex-Treg cells in the LNs from Foxp3^{eGFP-Cre-ERT2/Y} Ogt^wt/Y Rosa26^tdTomato/wt and Foxp3 ^{eGFP-Cre-ERT2/Y} Ogt^fl/Y Rosa26^tdTomato/wt mice, n = 3–4. Data are shown as mean ± s.e.m. *p < 0.05; **p < 0.01; ***p < 0.001 by unpaired (b, d, g–i) and paired (e) student’s t-test

Fig. 5

O-GlcNAcylation is required for the effector differentiation of Treg cells. a–d Frequencies of CD44^hiCD62L^lo (a), CD44⁺KLRG1⁺ (b), CD44⁺PD-1⁺ (c), and CD44⁺CD73⁺ (d) cells among YFP⁺CD25⁺ Treg cells in the LNs and spleen from 2-week-old Foxp3^YFP-Cre/Y Ogt^wt/Y and Foxp3^YFP-Cre/Y Ogt^fl/Y mice, at least n = 4 each group. e–i Frequencies of indicated cell populations among YFP⁻OGT-sufficient and YFP⁺ OGT-deficient CD4⁺TCRβ⁺CD25⁺GITR⁺ Treg cells in the LNs and spleen from Foxp3^YFP-Cre/wt Ogt^fl/fl mice, n = 3 each group. Data are shown as mean ± s.e.m. *p < 0.05; **p < 0.01; ***p < 0.001 by unpaired student’s t-test

Fig. 6

Attenuated IL-2/STAT5 signaling in OGT-deficient Treg cells. a Hierarchical clustering of top differentially expressed genes (60 genes with FDR q values less than 0.1) between OGT-sufficient and OGT-deficient Treg cells using Morpheus. b, c Enrichment plots for up-regulated (b) and down-regulated (c) genes in Blimp1⁺ eTreg cells from the Gene Set Enrichment Analysis (GSEA) of differentially expressed genes between OGT-sufficient and OGT-deficient Treg cells. d Ingenuity Pathway Analysis (IPA) of predicted upstream regulators for observed changes in gene expression. e–j Foxp3^YFP-Cre/wt Ogf^fl/fl female mice were injected with PBS or the IL-2 immune complex (n = 2–5) for 3 consecutive days and Treg cells were analyzed at day 6. Flow cytometry of CD44⁺KLRG1⁺(e), CD4⁺GZMB⁺ (f) and CD4⁺BLIMP-1⁺ (g) Treg cells among YFP⁻OGT-sufficient and YFP⁺ OGT-deficient Treg cells (CD4⁺CD25⁺GITR⁺) in LNs and spleen and corresponding frequencies in h–j. k Expression of STAT5-target genes in OGT-sufficient and OGT-deficient Treg cells. l–m mRNA levels of Socs1 (l) and Socs3 (m) in OGT-sufficient Treg cells after retrovirus infection as indicated, n = 3 each group. Data are shown as mean ± s.e.m. *p < 0.05; **p < 0.01; ***p < 0.001 by two-way ANOVA (h–j) and one-way ANOVA (l, m)

Fig. 7

Constitutive activation of STAT5 partially rescues Treg cell dysfunction. a, b Representative images of mice (a) and peripheral LNs and spleen (b) from 2-week-old Foxp3^YFP-Cre/Y Ogt^fl/Y and Foxp3^YFP-Cre/YOgt^fl/Y Rosa26^Stat5b-CA/wt mice. c Survival curve of Foxp3^YFP-Cre/Y Ogt^fl/Y and Foxp3^YFP-Cre/YOgt^fl/Y Rosa26^Stat5b-CA/wt mice (n = 6). d MFI of FOXP3 in FOXP3⁺CD25⁺ Treg cells in the LNs from 2-week-old Foxp3^YFP-Cre/Y Ogt^wt/Y, Foxp3^YFP-Cre/Y Ogt^fl/Y and Foxp3^YFP-Cre/YOgt^fl/Y Rosa26^Stat5b-CA/wt mice, at least n = 3 each group. e Frequencies of FOXP3⁺CD25⁺ Treg cells among CD4⁺ T cells, at least n = 9 each group. f–g Frequencies of CD44^hiCD62L^lo (f) and CD44⁺KLRG1⁺ (g) eTreg cells among FOXP3⁺CD25⁺ Treg cells in the LNs from 2-week-old Foxp3^YFP-Cre/Y Ogt^wt/Y, Foxp3^YFP-Cre/Y Ogt^fl/Y and Foxp3^YFP-Cre/YOgt^fl/Y Rosa26^Stat5b-CA/wt mice, at least n = 6 each group. h Histogram of FOXP3 expression in CD4⁺CD25⁺YFP⁺ Treg cells in the LNs from 2-week-old Foxp3^YFP-Cre/Y Ogt^wt/Y, Foxp3^YFP-Cre/Y Ogt^fl/Y and Foxp3^YFP-Cre/YOgt^fl/Y Rosa26^Stat5b-CA/wt mice. i, j Absolute number of CD4⁺ (i) and CD8⁺ (j) T cells in the LNs, at least n = 3 each group. k, l Frequencies of CD44^hiCD62L^lo effector T cells in CD4⁺ (k) and CD8⁺ cells (l), at least n = 7 each group. m–o Frequencies of T-BET⁺ (m), GATA3⁺ (n) and RORγT⁺ (o) populations among CD4⁺FOXP3⁻ cells in the LNs, at least n = 6 each group. p Frequencies of INFγ, IL-4 and IL-17A-producing CD4⁺FOXP3⁻ cells from Foxp3^YFP-Cre/Y Ogt^fl/Y and Foxp3^YFP-Cre/YOgt^fl/Y Rosa26^Stat5b-CA/wt mice, at least n = 3 each group. Data are shown as mean ± s.e.m. *p < 0.05; **p < 0.01; ***p < 0.001 by Kaplan-Meier Analysis (c), unpaired student’s t-test (p) and one-way ANOVA (d–g, i–o)

Fig. 8

Activating O-GlcNAcylation promotes Treg cell suppressive function. a–c MFI of protein O-GlcNAcylation (a), FOXP3 (b, c) in mouse CD4⁺FOXP3⁺ Treg cells treated with TMG, H₂O treatment was used as a control (n = 5). d, e mRNA levels of Socs1 (d) and Socs3 (e) in mouse Treg cells treated with TMG, H₂O treatment was used as a control (n = 3). f Gene expression in iTreg cells treated with TMG, H₂O treatment was used as a control (n = 2). g, h MFI of protein O-GlcNAcylation (g) and FOXP3 (h) in human tTreg cells treated with TMG for 7 days (n = 3), H₂O treatment was used as a control. i mRNA levels of SOCS3 in human tTreg cells treated with TMG, H₂O treatment was used as a control (n = 3). j, k Suppression activity of human tTreg cells treated with TMG for 7 days, H₂O treatment was used as a control. Data are shown as mean ± s.e.m. *p < 0.05 by unpaired student’s t-test