Glycolysis controls the induction of human regulatory T cells by modulating the expression of FOXP3 exon 2 splicing variants

Abstract

Human regulatory T cells (T(reg) cells) that develop from conventional T cells (T(conv) cells) following suboptimal stimulation via the T cell antigen receptor (TCR) (induced T(reg) cells (iT(reg) cells)) express the transcription factor Foxp3, are suppressive, and display an active proliferative and metabolic state. Here we found that the induction and suppressive function of iT(reg) cells tightly depended on glycolysis, which controlled Foxp3 splicing variants containing exon 2 (Foxp3-E2) through the glycolytic enzyme enolase-1. The Foxp3-E2-related suppressive activity of iT(reg) cells was altered in human autoimmune diseases, including multiple sclerosis and type 1 diabetes, and was associated with impaired glycolysis and signaling via interleukin 2. This link between glycolysis and Foxp3-E2 variants via enolase-1 shows a previously unknown mechanism for controlling the induction and function of T(reg) cells in health and in autoimmunity.

Figure 1

Generation of human iT_reg cells from T_conv cells during weak stimulation of the TCR. (a) ECAR, as an indicator of glycolysis (left), and OCR, quantifying mitochondrial respiration (right), of T_conv cells after 24 h of in vitro stimulation with medium alone or with mAb to CD3 plus mAb to CD28 (CD3+CD28) (key), in the presence of (above plots) glucose, the ATP-synthase inhibitor oligomycin and 2DG (left), or oligomycin, the uncoupling reagent FCCP and the respiratory-chain inhibitors antimycin A plus rotenone (right), administered sequentially (dotted downward arrows). (b) OCR, quantifying FAO (left), of T_conv cells stimulated for 24 h in vitro as in a, right, in the presence of the specific FAO substrate palmitate alone (+ palm) or palmitate plus the FAO inhibitor Etx (+ palm + inh), and their FAO index (right), calculated as ratio of the FCCP-stimulated OCR in the presence of palmitate to the FCCP-stimulated OCR in the presence of palmitate plus inhibitor. (c) Flow cytometry of CD4⁺CD25⁺ T cells sorted by flow cytometry on the basis of surface CD25 expression (far left), and expression of Ki67, phosphorylation of S6 and expression of Foxp3 (right) in CD25^hi, CD25^int and CD25^lo cells gated as at left. Numbers in plots indicate percent CD25^hi cells (top), CD25^int cells (middle) or CD25^lo cells (bottom) or mean fluorescence intensity (MFI) (in parentheses) of CD25 in those cells (far left), or percent Ki67⁺ cells (left), cells with phosphorylated S6 (middle) or Foxp3⁺ cells (right), as well as the MFI (in parentheses) of Ki67 (left), phosphorylated (p-) S6 (middle) or Foxp3 (right). (d) Proliferation (left) of CFSE-labeled CD4⁺ T cells stimulated for 96 h in vitro with mAb to CD3 plus mAb to CD28 and cultured alone (CD4⁺ alone) or in the presence of flow cytometry–sorted CD25^hi, CD25^int or CD25^lo T cells (above plots), and % proliferation of CD4⁺ T cells in those conditions (far right). Numbers in plots (left) indicate percent CFSE dilution in CD4⁺ T cells cultured alone (top left); numbers above bracketed lines indicate % CFSE dilution in CD4⁺ T cells cultured with iT_reg cells. *P < 0.05 and **P < 0.0001 (paired two-tailed Student’s t-test (b) or Wilcoxon matched-pairs test (d)). Data are from one experiment representative of 2 with technical duplicates (a; mean ± s.e.m.), 3 experiments with technical duplicates (b; mean ± s.e.m. (left) or mean ± s.e.m. of n = 18 (right)), one experiment representative of 8 (c), one experiment representative of 3 (d, left) or 3 independent experiments with technical triplicates (d, right; mean ± s.e.m. of n = 9 replicates).

Figure 2

Generation of human iT_reg cells from T_conv cells during weak stimulation of the TCR in the presence of inhibitors of either glycolysis or FAO. (a) ECAR (left) and OCR (right) of T_conv cells stimulated for 24 h in vitro with mAb to CD3 plus mAb to CD28 (in conditions as in Fig. 1a; above plots), grown in the absence (control (CTR)) or in the presence of 2DG or Etx (key) (as in Supplementary Fig. 3). (b) OCR, quantifying FAO, of T_conv cells stimulated for 24 h in vitro with mAb to CD3 plus mAb to CD28 (in conditions as in a, right; above plots), grown in the absence of 2DG or Etx (control (CTR)) or in the presence of 2DG (left) or Etx (middle), and in the presence of palmitate alone or palmitate plus inhibitor (as in Fig. 1b), and the FAO index of those cells (right; calculated as in Fig. 1b). (c) Flow cytometry analyzing the expression of CD25 and CD4 (left), and of Foxp3 and CD4 (right), by CD4⁺CD25^hi iT_reg-CTR, iT_reg-2DG and iT_reg-Etx cells sorted by flow cytometry. Numbers in plots (top left corner) indicate percent CD25⁺ cells or MFI (in parentheses) of CD25 (left) or percent Foxp3⁺+ cells or MFI (in parentheses) of Foxp3 (right); numbers adjacent to outlined areas (right) indicate percent Foxp3^hi cells (top) or Foxp3^int cells (bottom). (d) Proliferation of CFSE-labeled CD4⁺ T cells stimulated for 96 h in vitro with mAb to CD3 plus mAb to CD28 and cultured alone (filled gray curves) or in the presence of iT_reg-CTR, iT_reg-2DG or iT_reg-Etx cells (black lines) at various ratios (CFSE⁺CD4⁺ T cells to iT_reg cells, 8:1 to 1:1; above plots). Numbers in plots indicate percent CFSE dilution in CD4⁺ T cells cultured alone (top left); numbers above bracketed lines indicate percent CFSE dilution in CD4⁺ T cells cultured with iT_reg cells. Below, frequency of proliferation of CD4⁺T cells in the conditions above (‘–’ indicates CD4⁺ T cells alone). NS, not significant (P > 0.05); *P < 0.05, **P < 0.001 and ***P < 0.0001, iT_reg-2DG versus iT_reg-CTR or iT_reg-Etx versus iT_reg-CTR in d(paired two-tailed Student’s t-test (b) or Wilcoxon matched-pairs test (d)). Data are from one experiment representative of two experiments with technical duplicates (a; mean ± s.e.m.), three independent experiments with technical duplicates (b; mean ± s.e.m. (left) or mean ± s.e.m. of n = 18 (right)), one experiment representative of eight experiments (c), one experiment representative of three (d, top), or three independent experiments with technical triplicates (d, bottom; mean ± s.e.m. of n = 45 (1:1 ratio, iT_reg-CTR cells from five subjects), n = 36 (1:1 ratio, iT_reg-2DG cells from four subjects), n = 27 (1:1 ratio, iT_reg-Etx cells from three subjects), or n = 18 (all other conditions, cells from two subjects).

Figure 3

Biochemical analysis of human iT_reg cells. Immunoblot analysis (far left) of total and/or phosphorylated (p-) S6, enolase-1, STAT3, STAT5, Foxp3 and Erk1/2 in iT_reg-CTR, iT_reg-2DG and iT_reg-Etx cells (above blots) left unstimulated (−) or stimulated (+) for 1 h with mAb to CD3 plus mAb to CD28 (α-CD3 + α-CD28; above lanes); total Foxp3 was assessed with mAb PCH101, which recognizes all splicing variants of Foxp3 (left margin, molecular sizes). Right, densitometry analysis, with each phosphorylated protein normalized to its total form (S6, STAT3 and STAT5), or total protein normalized to total Erk1/2 (enolase-1 and Foxp3), presented relative to results obtained for unstimulated iT_reg-CTR cells. Each symbol (right) represents an individual data point (n = 9); small horizontal lines indicate the mean. *P < 0.0001 and **P < 0.05 (two-tailed Student’s t-test). Data are from one experiment representative of three independent experiments (far left) or three independent experiments with technical triplicates (right).

Figure 4

Phenotypical characterization and metabolic programs of iT_reg-CTR, iT_reg-2DG and iT_reg-Etx cells. (a) Flow cytometry analyzing the expression of CD25 and CD4 in iT_reg-CTR, iT_reg-2DG and iT_reg-Etx cells (far left) and of T_reg cell–specific markers (horizontal axes) (right). Numbers adjacent to outlined areas (far left) indicate percent CD25^hi cells or MFI (in parentheses) of CD25; numbers in plots (right) indicate percent marker-positive cells or MFI (in parentheses) of the marker. (b) ECAR (left) and OCR (right) of iT_reg-CTR, iT_reg-2DG and iT_reg-Etx cells stimulated for 24 h in vitro with mAb to CD3 plus mAb to CD28 (in conditions as in Fig. 1a; above plots). (c) OCR, quantifying FAO (left), of iT_reg-CTR and iT_reg-2DG cells (left) or of iT_reg-CTR and iT_reg-Etx cells (middle), stimulated for 24 h in vitro (as in Fig. 1a, right), in the presence of palmitate alone or palmitate plus inhibitor (as in Fig. 1b), and FAO index (right; calculated as in Fig. 1b). *P < 0.0001 and **P < 0.005 (paired two-tailed Student’s t-test). Data are from one experiment representative of five experiments (a), one experiment representative of two experiments with technical duplicates (b; mean ± s.e.m.) or three independent experiments with technical duplicates (c; mean ± s.e.m. n = 18 replicates).

Figure 5

Glycolysis controls expression of the Foxp3-E2 variants, which are indispensable for the suppressive function of iT_reg cells. (a) Splicing variants of FOXP3 mRNA: green boxes, exons (light green, 5′ and 3′ untranslated regions (UTR); dark green, coding sequence; blue, poly(A) sites); black lines, spliced regions (Δ2, deletion of exon 2); arrows, positions of PCR primers. (b) Real-time PCR analysis of all FOXP3 mRNA (left) andFOXP3-E2 mRNA (middle) in iT_reg-CTR, iT_reg-2DG and iT_reg-Etx cells, and semiquantitative RT-PCR analysis with primers spanning exons 1–3, followed by quantification of densitometry of the PCR productsFOXP3 and FOXP3 lacking exon 2 (FOXP3-Δ2) after separation by electrophoresis through agarose gels (right). (c) Immunoblot analysis (far left) of Foxp3 (above) and total Erk1/2 (used for normalization) in iT_reg-CTR, iT_reg-2DG and iT_reg-Etx cells, probed with mAb PCH101, to a common epitope of the amino terminus (as in Fig. 3), or mAb 150D/E4, to an epitope encoded by exon 2. Right, densitometry of all Foxp3 forms, normalized to total Erk1/2 and presented relative to results obtained for iT_reg-CTR cells. Each symbol (right) represents an individual data point (n = 15); small horizontal lines indicate the mean. (d) Proliferation (left) of CFSE-labeled CD4⁺ T cells stimulated for 72 h in vitro with mAb to CD3 plus mAb to CD28 and cultured alone or in the presence of iT_reg cells (key) generated in the presence of siRNA-E2, siRNA-E5 or siRNA-Scr (left margin), at a ratio of 1:1 or 2:1 (above plots). Numbers in plots indicate percent CFSE dilution in CD4⁺ T cells alone (top left); numbers above bracketed lines indicate percent CFSE dilution in CD4⁺ T cells cultured with iT_reg cells. Right, percent proliferation of CD4⁺ T cells in the conditions at left. *P < 0.05, **P < 0.005 and ***P < 0.001, siRNA-E2 versus siRNA-Scr or siRNA-E5 versus siRNA-Scr in d (Wilcoxon test (b), two-tailed Student’s t-test (c) or Wilcoxon matched-pairs test (d)). Data are from 3 independent experiments with technical triplicates (b; mean ± s.d. of n= 9 replicates), one experiment representative of 5 experiments (c, far left), 5 experiments with technical triplicates (c, right), one experiment representative of 3 experiments (d, left) or 3 independent experiments with 6 replicates in each (d, right; mean ± s.e.m. of n = 18 replicates).

Figure 6

Recruitment of enolase-1 to FOXP3 regulatory regions controls Foxp3-E2 expression. (a) Immunoblot analysis (far left) of enolase-1 isoforms in nuclear extracts of iT_reg-CTR, iT_reg-2DG and iT_reg-Etx cells, probed with a mAb (H-300) that recognizes an epitope common to all isoforms (48, 37 and 31 kDa), and of the control protein MCM7 (bottom); gels were exposed for 1 min (48 kDa), 5 min (37 kDa) or 20 min (31 kDa) to avoid overexposure of the signal. Right, densitometry of all enolase-1 forms, normalized to MCM7 and presented relative to results obtained for iT_reg-CTR cells. Each symbol (right) represents an individual data point (n = 9); small horizontal lines indicate the mean. (b) Recruitment of enolase-1 to the promoter of FOXP3 (left) or FOXP3 CNS2 (right) in T_conv cells and iT_reg-CTR, iT_reg-2DG and iT_reg-Etx cells (key), assessed by chromatin immunoprecipitation with mAb to enolase-1. (c) Quantitative real-time PCR analysis of all FOXP3 transcripts (left) or FOXP3-E2 mRNA (right) in iT_reg-CTR and iT_reg-2DG cells generated in the presence of siRNA-Scr or siRNA-ENO1 (horizontal axis); results were normalized to those of control 18S rRNA and are presented relative to those of iT_reg-CTR cells treated with siRNA-Scr. *P < 0.05, **P < 0.01, ***P < 0.005, †P < 0.001 and ‡P < 0.0001 (two-tailed Student’s t-test). Data are from three independent experiments with technical triplicates (a, c; mean ± s.e.m. of n = 9 in c) or three independent experiments (b mean ± s.d. of n = 3).

Figure 7

Impaired glycolysis, impaired IL-2–IL-2R–STAT5 signaling and altered Foxp3-E2 expression in iT_reg cells from subjects with RRMS. (a) ECAR of T_conv cells isolated from subjects with RRMS who had not undergone treatment (n = 6) and healthy subjects (CTR) (n = 14), stimulated for 24 h in vitro with mAb to CD3/CD28 (in conditions as in Fig. 1a, left). (b) Proliferation of CFSE-labeled CD4⁺ T cells stimulated for 96 h in vitro with mAb to CD3/CD28 and cultured alone or in the presence of iT_reg cells (key) from subjects as in a (left margin), at a ratio of 1:1 or 1:2 (above plots). Right, percent proliferation of CD4⁺ T cells in those conditions (far right). Numbers in plots (left) indicate percent CFSE dilution in CD4⁺ T cells cultured alone (top left); numbers above bracketed lines indicate percent of CFSE dilution in CD4⁺ T cells cultured with iT_reg cells. (c) Flow cytometry analyzing expression of CD25, Foxp3-E2 and CD4 (left) or various T_reg cell markers (right) in iT_reg cells from subjects as in a. Numbers in top left corner indicate % CD25^hi cells or MFI (in parentheses) of CD25 (left) or % Foxp3-E2^hi cells or MFI (in parentheses) of Foxp3-E2 (middle); and numbers above plots (right) indicate % or MFI (in parentheses) of each marker. (d) Immunoblot analysis of phosphorylated and total STAT5 (top) and of the 44- and 47-kDa forms of Foxp3 (bottom) in T_conv cells obtained from subjects as in a and stimulated for 24 or 36 h (above lanes) in vitro with mAb to CD3 plus mAb to CD28. Bottom, densitometry of Foxp3 normalized to total Erk1/2, or of phosphorylated STAT5 normalized to total STAT5, presented relative to results for T_conv cells from CTR after 24 h of stimulation. Each symbol (right) represents an individual data point; small horizontal lines indicate the mean. (e) Immunoblot analysis (left) of the 44- and 47-kDa forms of Foxp3 in T_conv cells obtained from pairs of monozygotic twins (n = 6) discordant for RRMS and without treatment, then stimulated for 24 or 36 h in vitro with mAb to CD3 plus mAb to CD28 (left). Right, densitometry of Foxp3 normalized to total Erk1/2 and presented relative to results obtained for T_conv cells from the healthy twin, assessed after 24 h of stimulation. Each symbol (right) represents an individual data point (n = 27); small horizontal lines indicate the mean. *P < 0.05, ** P< 0.001 and ***P < 0.0001 (Wilcoxon matched-pairs test (b) or Wilcoxon test (d, e)). Data from 6 (RRMS) or 14 (CTR) (a; mean ± s.e.m.), representative of 3 (b, left), with technical triplicates (b, right; mean ± s.e.m. of n = 9 replicates), one experiment representative of 3 experiments (c), one experiment representative of 4 (d, left), or 4 independent experiments (Foxp3) with technical triplicates from 4 subjects (n = 12) or 3 independent experiments (STAT5) with technical triplicates from three subjects (n = 9) (d, right), one experiment representative of 3 (e, left) or 3 independent experiments with nine technical replicates from 3 sets of twins (e, right; n = 27).

Figure 8

Impaired IL-2–IL-2R–STAT5 signaling and altered Foxp3-E2 expression in iT_reg cells from subjects with T1D. (a) Flow cytometry analyzing the expression of CD25, Foxp3-E2 and CD4 (left) or T_reg cell markers (right) in iT_reg cells from subjects with early-onset T1D or healthy subjects (CTR) (numbers in plots as in Fig. 7c). (b) Immunoblot analysis (far left) of phosphorylated and total STAT5 (top) and the 44- and 47-kDa forms of Foxp3 (below) in T_conv cells obtained from subjects as in a and stimulated for 24 or 36 h in vitro with mAb to CD3 plus mAb to CD28. Right, densitometry of phosphorylated STAT5 normalized to total STAT5 or of all Foxp3 normalized to total Erk1/2, presented relative to results obtained for T_conv cells from healthy subjects, assessed after 24 h of stimulation. Each symbol (right) represents an individual data point; small horizontal lines indicate the mean. (c) Immunoblot analysis of the 44- and 47-kDa forms of Foxp3 in T_conv cells obtained from monozygotic twins (n = 2) discordant for T1D, then stimulated for 24 or 36 h in vitro with mAb to CD3 plus mAb to CD28. Right, densitometry of Foxp3 normalized to total Erk1/2, presented relative to results obtained for T_conv cells from healthy donors, assessed after 24 h of stimulation. Each symbol (right) represents an individual data point; small horizontal lines indicate the mean. *P < 0.005 and **P < 0.001 (Wilcoxon test). Data are from one experiment representative of two (a), one experiment representative of three independent experiments (b, left), three independent experiments with technical triplicates from five healthy subjects (n = 15) or seven subjects with T1D (n = 21) (b, right), one experiment representative of three independent experiments (c, left) or three independent experiments with technical triplicates from one set of twins (n = 9) (c, right).