Hypomethylation at the regulatory T cell-specific demethylated region in CD25hi T cells is decoupled from FOXP3 expression at the inflamed site in childhood arthritis

Abstract

The maintenance of FOXP3 expression in CD25(hi) regulatory T cells (Tregs) is crucial to the control of inflammation and essential for successful Treg transfer therapies. Coexpression of CD25 and FOXP3 in combination with a hypomethylated region within the FOXP3 gene, called the Treg-specific demethylated region (TSDR), is considered the hallmark of stable Tregs. The TSDR is an epigenetic motif that is important for stable FOXP3 expression and is used as a biomarker to measure Treg lineage commitment. In this study, we report that, unlike in peripheral blood, CD4(+) T cell expression of CD25 and FOXP3 is frequently dissociated at the inflamed site in patients with juvenile idiopathic arthritis, which led us to question the stability of human Tregs in chronic inflammatory environments. We describe a novel CD4(+)CD127(lo)CD25(hi) human T cell population that exhibits extensive TSDR and promoter demethylation in the absence of stable FOXP3 expression. This population expresses high levels of CTLA-4 and can suppress T conventional cell proliferation in vitro. These data collectively suggest that this population may represent a chronically activated FOXP3(lo) Treg population. We show that these cells have defects in IL-2 signaling and reduced expression of a deubiquitinase important for FOXP3 stability. Clinically, the proportions of these cells within the CD25(hi) T cell subset are increased in patients with the more severe courses of disease. Our study demonstrates, therefore, that hypomethylation at the TSDR can be decoupled from FOXP3 expression in human T cells and that environment-specific breakdown in FOXP3 stability may compromise the resolution of inflammation in juvenile idiopathic arthritis.

FIGURE 1.

Dissociation of CD25 and FOXP3 expression at the inflamed site. Flow cytometry data, gated on CD3⁺CD4⁺ T cells, of paired SFMC and PBMC samples from a JIA patient displaying (A) CD25 versus CD127 expression (left) and FOXP3 versus CD127 expression (right). Because FOXP3^hi and CD25^hi cells are almost uniformly CD127^lo, a CD4⁺CD127^lo gate (B, left) was set to analyze the relationship between FOXP3 and CD25 expression (B, right). (C) Multispectral imaging flow cytometry was performed on CD4⁺ T cells from SFMCs of a JIA patient. Four representative image panels are displayed illustrating the four populations identifiable by the imaging of CD25 and FOXP3 staining. Brightfield (BF), CD25, and FOXP3 images were merged, with BF intensity set to 40% and CD25 and FOXP3 at 100%.

FIGURE 2.

The epigenetic footprint of Tregs is not limited to the CD25^hiFOXP3^hi CD4⁺ T cell compartment. CD4⁺ T cells were isolated from a JIA SFMC sample, stained for CD4, CD25, and CD127, before fixation and staining for FOXP3. Cells were then sorted into four populations based on CD25 and FOXP3 expression. (A) Representative sort displaying the presort staining (far left, gated on CD4⁺CD127^lo T cells) and postsort purities (right four plots). The cellular phenotypes were further analyzed for FOXP3 and CD25 protein expression levels. (B) Summary plots of MFI data (n = 3–4): △, P1 Treg; ●, P2 Treg; □, P3 Treg; ×, Tcon. (C) TSDR methylation rates were determined in the sorted populations from one patient (patient 24) and are presented as heat maps plus region average. (D) Summary of the mean TSDR methylation rates in P1–P3 Tregs and Tcons in the four patients studied (patients 5, 9, 24, and 28). P1 Tregs, n = 3 (amplification of the TSDR from P1 Tregs in patient 9 was unsuccessful); P2 Tregs, n = 4; P3 Tregs, n = 3 (patient 5 was excluded from analysis based on a purity of <50%); and Tcons, n = 4. Bars represent the median values.

FIGURE 3.

P3 T cells display phenotypic characteristics of an activated FOXP3^lo Treg. (A) Histogram displaying CTLA-4 levels in the three populations (P1, gray dashed line; P2, black line; P3, gray line) compared with CD127^hi T cells (black dashed line) in one representative patient. Summary graph (right) displays the MFI of CTLA-4 in the four populations described (n = 9). (B) Representative histogram overlays (left) displaying PD-1 expression levels within the CD4⁺CD127^loCD25^hi subset in healthy control PBMC (black dashed line), JIA PBMC (gray line), and JIA SFMC (black line) samples; right, summary graph of PD-1 MFI levels (adult controls, n = 6; JIA PBMCs, n = 5; JIA SFMCs, n = 20). (C) Representative histogram overlays (left) showing PD-1 expression (numbers display the MFI) within the four subsets (P1, gray dashed line; P2, black line; P3, gray line; CD127^hi T cells, black dashed line); right, summary graphs of (n = 24) PD-1 protein expression level (MFI). Bars represent median values. Statistical analysis by Friedman test with Dunn’s multiple comparisons (A and C) or Kruskal–Wallis test with Dunn’s multiple comparisons (B). **p < 0.01, ***p < 0.001.

FIGURE 4.

PD-1 expression and FSC characteristics partition the CD25^hiCD4⁺ SF T cell compartment into P2 and P3 Treg subsets. Displayed is an SFMC sample stained for CD4, CD127, CD25, FOXP3, and PD-1 and analyzed by flow cytometry. (A) Boolean gating strategy used to obtain the CD127^loCD25^hiCD4⁺ T cell population. (B) Flow cytometry plots gated on CD4⁺CD127^loCD25^hi SF T cells displaying FOXP3 versus FSC (left) and FOXP3 versus PD-1 expression (right). Intrasample correlations were analyzed by the Spearman rank correlation analysis; n = 1856 data pairs. ****p < 0.0001. (C) Flow cytometry data gated on CD4⁺CD127^loCD25^hi SF T cells (left) displaying PD-1 versus FSC expression. Flow cytometry plots (right) display CD25 versus FOXP3 expression on the cells in the gates P2 and P3.

FIGURE 5.

P3 Tregs display a modest functional deficit. (A) Flow cytometry plots displaying CD25 versus FOXP3 expression in unsorted sample from patient 21 (left), P2-enriched Tregs, CD127^loCD25^hiFSC^loPD-1^int (middle), and P3-enriched Tregs, CD127^loCD25^hiFSC^hiPD-1^hi (right). (B) TSDR methylation rates in sorted Treg populations from patient 21. (C) Representative histogram overlays (top) displaying expression levels of PD-1, FOXP3, and CD25 in the sorted Treg populations compared with autologous SF Tcons (P2, black line; P3, gray line; CD127^hi T cells, black dashed line). (D) Summary graphs show MFI data for PD-1, FOXP3, and CD25 in the sorted samples (n = 6). CFSE-labeled SF Tcons were cultured alone (0:1) or at a 1:2, 1:4, or 1:8 ratio with either P2 Tregs or P3 Tregs on anti–CD3/CD28-coated plates for 4–5 d. (E) Histogram overlay displaying Tcon CFSE dilution in the presence of P2 Tregs (black line) and P3 Tregs (gray line) at a 1:2 ratio. Tcon alone division is displayed in black dashed line. (F) Summary data demonstrating percentage of divided Tcons in the presence of P2 (filled bar) or P3 (open bar) Tregs (n = 6) or no Tregs (hatched bar). (G) Percentage suppression of SF Tcons at 1:2 (n = 6), 1:4 (n = 6), and 1:8 (n = 5) Treg/Tcon ratios. (H) Levels of IFN-γ, IL-17A, and TNF-α in coculture supernatants. Tcons plus P2 Tregs (●) and Tcons plus P3 Tregs (□) are shown. Bars represent the median value in (D) and means ± SEM in (F)–(H). Statistical analysis was performed by a Mann–Whitney U test (D), repeated measures ANOVA with a Bonferroni multiple comparisons test (F), or a paired t test (H). *p < 0.05, **p < 0.01.

FIGURE 6.

Decreased FOXP3 mRNA and expression of USP7 in P3 Tregs. RNA was extracted from P2- and P3-enriched Treg populations and converted to cDNA. (A) FOXP3 protein MFIs (●) and mRNA levels of FOXP3 (relative to GAPDH) (□) in P2- and P3-sorted subsets (n = 4). (B) Relative levels of USP7 mRNA were analyzed by RT-PCR in P2-enriched Tregs (n = 6), P3-enriched Tregs (n = 4), or healthy control Tregs (n = 3). (C) Left, Flow cytometry plot showing CD25 versus FOXP3 staining gated on CD4⁺ T cells in a SF sample, with gates for P1, P2, and P3 Tregs. Center, Histogram overlay showing USP7 staining in P1 (gray dashed line), P2 (black line), or P3 Tregs (gray line). Numbers represent MFIs (n = 3). Right, Summary data showing MFI. (D) To examine FOXP3 degradation, P3-enriched SF Tregs were cultured for 24 h in the presence/absence of 5μM proteasome inhibitor MG132. Expression of FOXP3 versus Live/Dead dye was analyzed by flow cytometry. Statistical analysis performed by a Mann–Whitney U Test (A) or a one-way ANOVA with a Bonferroni multiple comparisons test (B and C). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

FIGURE 7.

Decreased IL-2 responsiveness in P3 Tregs. (A) P2- and P3-enriched SF Tregs were isolated as previously described and cultured for 24 h with or without 2 U/ml IL-2. Histograms of Tregs 24 h after culture displaying FOXP3 (A) or CD25 (B) expression are shown. To assess IL-2 sensitivity, total SFMCs from JIA patients were stimulated with 100 U/ml recombinant human IL-2 for 15 min. (C) Flow cytometry plot (left) gated on CD3⁺CD4⁺ T cells, showing CD25 versus FOXP3 expression. Histogram (right) displaying p-STAT5 levels in stimulated P1 (gray dashed line), P2 (black line), and P3 (gray line), or CD25^loFOXP3^lo T cells (black dashed line); unstimulated T cell p-STAT5 levels are represented by the filled histogram. (D) Summary graphs of percentage p-STAT5⁺ (left) or MFI of p-STAT5 protein (right) in the four different T cell populations (n = 10; data are from five independent experiments). (E) Graph of percentage Ki67⁺ cells in the three Treg subsets compared with CD127^hi T cells (n = 19). Bars display median values. Statistical analysis was performed by Friedman test with Dunn’s multiple comparisons test (D and E). *p < 0.05, **p < 0.01, ***p < 0.001.

FIGURE 8.

Increased proportions of FOXP3^lo Tregs in the more severe forms of childhood arthritis. (A) Using the gating strategy determined in Fig. 1, the frequency of P1, P2, and P3 Tregs within the CD4⁺CD127^lo T cell compartment was determined in adult control PBMCs (n = 9), child control PBMCs (n = 4), JIA patient PBMCs (n = 11), and JIA patient SFMCs (n = 45). Bars represent median values. (B) The proportion of FOXP3^lo cells within the CD4⁺CD127^loCD25^hi T cell compartment was analyzed according to disease severity at the time of sample: persistent O-JIA, n = 18; polyarticular JIA course, n = 27. Statistical analysis by Kruskal–Wallis test with a Dunn’s multiple comparisons test (A) or Mann–Whitney U test (B). *p < 0.05, **p < 0.01, ***p < 0.001.