Inflammation-driven reprogramming of CD4+ Foxp3+ regulatory T cells into pathogenic Th1/Th17 T effectors is abrogated by mTOR inhibition in vivo

Abstract

While natural CD4(+)Foxp3(+) regulatory T (nT(REG)) cells have long been viewed as a stable and distinct lineage that is committed to suppressive functions in vivo, recent evidence supporting this notion remains highly controversial. We sought to determine whether Foxp3 expression and the nT(REG) cell phenotype are stable in vivo and modulated by the inflammatory microenvironment. Here, we show that Foxp3(+) nT(REG) cells from thymic or peripheral lymphoid organs reveal extensive functional plasticity in vivo. We show that nT(REG) cells readily lose Foxp3 expression, destabilizing their phenotype, in turn, enabling them to reprogram into Th1 and Th17 effector cells. nT(REG) cell reprogramming is a characteristic of the entire Foxp3(+) nT(REG) population and the stable Foxp3(NEG) T(REG) cell phenotype is associated with a methylated foxp3 promoter. The extent of nT(REG) cell reprogramming is modulated by the presence of effector T cell-mediated signals, and occurs independently of variation in IL-2 production in vivo. Moreover, the gut microenvironment or parasitic infection favours the reprogramming of Foxp3(+) T(REG) cells into effector T cells and promotes host immunity. IL-17 is predominantly produced by reprogrammed Foxp3(+) nT(REG) cells, and precedes Foxp3 down-regulation, a process accentuated in mesenteric sites. Lastly, mTOR inhibition with the immunosuppressive drug, rapamycin, stabilizes Foxp3 expression in T(REG) cells and strongly inhibits IL-17 but not RORγt expression in reprogrammed Foxp3(-) T(REG) cells. Overall, inflammatory signals modulate mTOR signalling and influence the stability of the Foxp3(+) nT(REG) cell phenotype.

Figure 1. Loss of Foxp3 expression in thymic or peripheral T_REG cells in lymphopenic hosts is modulated by the frequency of T_EFF cells, not T_REG cells.

(A–C) TCRβ^−/− mice received GFPtg CD4⁺CD25⁺ T_REG cells (0.3×10⁶), and 7, 14 and 21 days post transfer, donor GFP⁺ T cells from mesLN were examined for Foxp3 expression. Representative histograms of Foxp3 expression and percentage of Foxp3⁺ cells within GFP⁺CD4⁺ T_REG cells (A) and proportion of cycling cells (Ki-67 expression) within Foxp3⁺ or Foxp3⁻ donor GFP⁺CD4⁺ T cells (C) at various time points are shown. (B) TCRβ^−/− mice received either thymus- or LN- derived GFP⁺CD4⁺CD25⁺ T_REG cells. The percentage of Foxp3^+/− cells within GFP⁺CD4⁺ T cells is shown 14 days post adoptive transfer. (D–E) Recipients received GFPtg CD4⁺CD25⁺ T_REG cells (0.3×10⁶) either alone or in combination with indicated numbers of CD4⁺CD25⁻ T_EFF, CD4⁺CD25⁺ T_REG or total CD4⁺ T cells. 14 days post T cell transfer cells, mesLN were analyzed for Foxp3 by flow cytometry. The percentage of Foxp3⁺ or Foxp3⁻ cells within donor GFP⁺CD4⁺ T cells is shown. Results are representative of 2 to 4 independent experiments (n = 3–4) are shown as mean ± SEM.

Figure 2. The Foxp3^+→− T_REG cell phenotype is stable and correlates with a methylated Foxp3 promoter.

TCRβ^−/− mice received CD4⁺GFP⁺ T cells (0.5×10⁶) from Ly5.1⁺ congenic Foxp3^GFP reporter mice. Ly5.1⁺CD4⁺GFP⁺ (Foxp3⁺) or GFP⁻ (Foxp3⁻) cells were sorted from lymphoid tissues of recipient mice 14 days post transfer. (A) Sorted populations were activated in vitro with plate-bound anti-CD3 for 5 days or (B–C) re-introduced into secondary TCRβ^−/− recipient mice. Freshly-isolated CD4⁺GFP^+/− T cells from Ly5.1⁺Foxp3^GFP mice were used as controls. 14 days post secondary transfer, donor Ly5.1⁺CD4⁺ T cells from lymphoid tissues of secondary recipient mice were (B) re-analyzed for Foxp3 expression or (C) sorted according to GFP expression and total genomic DNA was subjected to methylation analysis of Foxp3 promoter region. The percentage of Foxp3⁺ or Foxp3⁻ cells within indicated donor T cell populations is shown in (B) as mean ± SEM (n = 3). (C) For each donor T cell population the percentage of methylated CpG motifs within Foxp3 promoter region was examined at eight different sites and averaged.

Figure 3. Foxp3^+→− T cells lose their T_REG cell phenotype and reprogram into Th1 and Th17 effector cells in lymphopenic hosts.

(A–B) TCRβ^−/− mice received CD4⁺GFP⁺ T_REG cells (0.5×10⁶) isolated from Foxp3^GFP reporter mice, and 14 days later, donor CD4⁺GFP⁺(Foxp3⁺)/GFP⁻(Foxp3⁻) cells were sorted from recipient mice and their proliferation (A) and suppressive activity (B) were assessed following in vitro activation. Freshly isolated T_REG and T_EFF cells from Foxp3^GFP mice were used as controls. Data from one of three independent experiments is presented as mean ± s.d. of triplicate wells. (C) TCRβ^−/− mice received GFPtgCD4⁺CD25⁺ T_REG cells (0.3×10⁶), and 14 days post transfer, GFP⁺ donor T cells were examined for the production of various cytokines relative to Foxp3 expression. Frequencies and mean fluorescent intensity (MFI) (C) of cytokines produced by Foxp3^+/− cells are shown as mean ± SEM from one out of 4 independent experiments (n = 4).

Figure 4. Intestinal inflammation or parasitic infection favours the reprogramming of Foxp3⁺ T_REG cells into effector T cells and promotes host immunity.

(A–B) TCRβ^−/− mice received GFPtg CD4⁺CD25⁺ T cells (0.3×10⁶). 14 days later Foxp3 expression within donor GFP⁺CD4⁺ T cells (A) and the frequency of total IFN-γ or IL-17 producing cells (B) in indicated tissues of recipient mice from one of 3 representative experiments (n>3) is shown. (C–G) TCRβ^−/− mice were infected or not (NI) with 5×10⁶ promastigotes of WT or GP63^−/− (KO) L. major into the right footpad 2 weeks prior reconstitution with GFPtgCD4⁺CD25⁺ T_REG cells (0.3×10⁶). (C) 4 weeks later, GFP⁺CD4⁺ T cells from draining (infected) and non-draining popliteal LN, perLN and mesLN were analyzed for Foxp3 expression. Footpad swelling (D), absolute number of infiltrated lymphocytes (E), and frequencies of IFN-γ producing CD4⁺ T cells (F) are shown in infected and non-infected sites. (G) The loss of Foxp3 expression by T_REG cells was compared between mice infected with WT or GP63^−/− L. major strains. Results are representative of 2 independent experiments with n = 4–5.

Figure 5. T_REG cell reprogramming occurs independently of variations in IL-2 production in vivo.

(A–B) TCRβ^−/− mice were treated every other day with 5 or 50 ng of rhIL-2, starting 2 days prior transfer of GFPtg CD4⁺CD25⁺ T cells (0.3×10⁶). MesLN were analyzed for Foxp3 expression 14 days later. The percentage of Foxp3^+/− cells (A) and the frequency of CD25⁺ cells and MFI of CD25 expression (B) within donor GFP⁺CD4⁺ T cells are shown. (C–D) TCRβ^−/− mice received GFPtg CD4⁺CD25⁺ T cells (0.3×10⁶). Donor T cells from perLN and mesLN were analyzed for Foxp3 expression and IL-2 production. (C) The frequency of IL-2 secreting GFP⁺CD4⁺Foxp3⁻ T cells and (D) correlation between frequency of IL-2⁺ and Foxp3⁻ cells within donor GFP⁺CD4⁺ T cells is shown. Results are representative of 2 independent experiments (n = 3–4).

Figure 6. Predominant IL-17 secretion precedes Foxp3 down-regulation in reprogramming Foxp3⁺ T_REG cells, a process accentuated in mesenteric sites.

(A–E) TCRβ^−/− mice received CFSE-labelled CD4⁺CD25⁺ T cells (0.3×10⁶) isolated from congenic Ly5.1⁺ mice, and, donor T cells were examined for Foxp3 expression at the indicated timepoints post transfer. (A) Representative FACS plots of Foxp3 expression relative to CFSE dilution in donor Ly5.1⁺CD4⁺ T cells are shown at various time points. (B–E) Representative FACS profiles (B,C) and proportion (D,E) of IL-17/IFN-γ-secreting donor Ly5.1⁺CD4⁺Foxp3⁺ (B,D) and Foxp3⁻ (C,E) T cells undergoing expansion at various time points are shown. Results are representative of 2 independent experiments (n = 3). (F) TCRβ^−/− mice received CD4⁺GFP⁺ T cells (0.5×10⁶) isolated from Ly5.1⁺ congenic Foxp3^GFP reporter mice, 14 days post transfer Ly5.1⁺CD4⁺GFP⁻ cells (Foxp3+→−) were sorted from lymphoid tissues of recipient mice, and reintroduced into secondary TCRβ^−/− recipient mice. Freshly-isolated CD4⁺GFP⁻ T cells (Foxp3-) from Ly5.1⁺Foxp3^GFP mice were used as a control. 14 days post secondary transfer, donor Ly5.1⁺CD4⁺ T cells from mesLN of secondary recipients were analyzed for IL-17/IFN-γ secretion relative to Foxp3 expression. Proportion of cytokine producing Foxp3⁻Ly5.1⁺CD4⁺ T cells is shown as mean ± SEM from one out of 2 independent experiments (n = 4).

Figure 7. mTOR inhibition stabilizes Foxp3 expression in T_REG cells and strongly inhibits IL-17 but not RORγt expression.

TCRβ^−/− mice were transferred with GFPtg CD4⁺CD25⁺ T_REG cells (0.3×10⁶), and then treated every second day with rapamycin (R) (18 mg/kg) or control vehicle (formulation without rapamycin) (V) as of day 0. 7 and 14 days later, perLN and mesLN of recipient mice were analyzed by FACS. (A) Foxp3 expression within donor GFP⁺CD4⁺ T cells, (B) proportion of IL-17/IFN-γ producing Foxp3^+/− donor T cells and (C) RORγt expression within Foxp3^+/− donor T cells are shown as mean ± SD from one out of 2 independent experiments (n = 4).