FOXP3 promoter demethylation reveals the committed Treg population in humans

Abstract

Background: Naturally occurring thymus derived regulatory T cells (Tregs) are central in the maintenance of self-tolerance. The transcription factor FOXP3 is crucial for the suppressive activity of Tregs and is considered the most specific marker for this population. However, human non regulatory T cells upregulate FOXP3 transiently upon activation which calls for other means to identify the Treg population. Since epigenetic mechanisms are involved in the establishment of stable gene expression patterns during cell differentiation, we hypothesized that the methylation profile of the FOXP3 promoter would allow the distinction of truly committed Tregs.

Methodology/principal findings: Human CD4(+)CD25(hi) Tregs displayed a demethylated FOXP3 promoter (1.4%+/-0.95% SEM methylated) in contrast to CD4(+)CD25(lo) T cells which were partially methylated (27.9%+/-7.1%). Furthermore, stimulated CD4(+)CD25(lo) T cells transiently expressed FOXP3 but remained partially methylated, suggesting promoter methylation as a mechanism for regulation of stable FOXP3 expression and Treg commitment. In addition, transient FOXP3 expressing cells exhibited suppressive abilities that correlate to the methylation status of the FOXP3 promoter. As an alternative to bisulphite sequencing, we present a restriction enzyme based screening method for the identification of committed Tregs and apply this method to evaluate the effect of various culturing conditions. We show that a partial demethylation occurs in long-term cultures after activation, whereas the addition of TGF-beta and/or IL-10 does not induce any additional change in methylation level.

Conclusions/significance: The unique FOXP3 promoter methylation profile in Tregs suggests that a demethylated pattern is a prerequisite for stable FOXP3 expression and suppressive phenotype. Presently, FOXP3 is used to identify Tregs in several human diseases and there are future implications for adoptive Treg transfer in immunotherapy. In these settings there is a need to distinguish true Tregs from transiently FOXP3(+) activated T cells. The screening method we present allows this distinction and enables the identification of cells suitable for in vitro expansions and clinical use.

Figure 1. FOXP3 expression in sorted cell populations and confirmation of Treg function.

(A) The sorted CD4⁺CD25^lo, CD4⁺CD25^hi and CD19⁺ cells were analyzed by flow cytometry. (B) FOXP3 expression in the sorted populations as determined by intracellular flow cytometry. Plots showing results from one representative donor out of four analyzed, except for plot showing FOXP3 expression in CD19⁺ cells where one single donor was analyzed (C) Sorted CD4⁺CD25⁺ cells suppress the proliferation of CD4⁺CD25⁻ cells. CD4⁺CD25⁻ cells were activated with CD3 and CD28 antibodies together with CD4⁻ feeder cells and increasing numbers of CD4⁺CD25⁺ cells in triplicate samples. Proliferation was measured as incorporation of [³H]Thymidine for 18 hours, here illustrated as counts per minute (cpm) on the y-axis. The CD4⁺CD25⁺ to CD25⁻ cell ratio is displayed on the x-axis. Squares indicate coculture of CD25⁻ and CD25⁺ cells. Triangles indicate control samples with only CD4⁺CD25⁻ cells. (D) FOXP3 mRNA expression of sorted cell populations. FOXP3 mRNA was measured by real-time PCR in FACS sorted CD4⁺CD25^hi, CD4⁺CD25^lo and CD19 cells. Data was normalized to the expression in CD4⁺CD25^lo cells using the 2^−ΔΔCt method and RPII as housekeeping gene. Data represent mean of triplicate samples from one single donor.

Figure 2. Identification of cross-species conserved FOXP3 promoter elements.

(A) FOXP3 transcript and m-Vista alignment showing conservation between human and mouse genomic sequences. Dark blue boxes display exons, outlined boxes are UTR's. Conserved regions are in red and conserved regions corresponding to exons are in light blue. (B) Schematic view of the conserved region upstream of the transcription start site indicating the location of promoter elements and CpG dinucleotides. (C) Clustal W alignment of human, mouse and rat genomic sequences showing a detailed view of the conservation at the FOXP3 core promoter. The broken arrow shows the position of the human transcription start site. The location of the TATA box is indicated by a black box. Red boxes indicate the CpG dinucleotides analyzed in this study. Transcription factor binding sites are marked with grey boxes.

Figure 3. Tregs are unmethylated at the FOXP3 promoter region.

Bars in (A) to (C) indicate the percentage of methylated (black) and unmethylated (white) cells at each CpG position (specified on x-axis labels as relative the transcription start site) in (A) CD4⁺CD25^hi, (B) CD4⁺CD25^lo and (C) CD19⁺ cell populations isolated by FACS from one male donor. (D) to (F) Methylation grade of the most discriminating CpG position (−77). Percentage methylated (filled) versus unmethylated (open) CpGs in the (D) CD4⁺CD25^hi, (E) CD4⁺CD25^lo and (F) CD19⁺ cells.

Figure 4. Restriction enzyme based analysis of the methylation status at the FOXP3 −77 reporter position.

PCR product from bisulphite treated DNA was digested with the restriction enzyme Aci1, which only cleaves DNA at position −77 in the FOXP3 promoter when DNA originally was methylated. Digestion was followed by GeneScan analysis. (A) Electropherograms showing peaks from the Aci1 digestion product. Amounts of digested and undigested product were calculated as area under the curve. Shown are results from one representative donor out of three analyzed, except for electropherogram obtained from CD19⁺ cells where one single donor was analyzed. Digestion product from methylated and unmethylated BAC was analyzed three times with similar results. (B) Filled bars indicate percentage methylated, and open bars percentage unmethylated −77 sites. Methylase treated and unmodified BACs were used as controls.

Figure 5. Transient expression of FOXP3 and CD25 in stimulated CD4⁺CD25^lo cells from male donors.

Plots showing results from one representative donor out of four analyzed. (A) FACS analysis of CD4⁺CD25^lo cells stimulated at day 0 with CD3/CD28 Dynabeads (proportion cell∶beads 1∶1) in presence of 180 U IL-2. Stimuli removed after 48 h and cells followed for an additional 2 weeks with regards to their expression of CD25 (left) and FOXP3 (middle). Left column demonstrating the relationship between CD25 (y-axis) and FOXP3 (x-axis) expression. (B) Expression of CD25 (grey filled bars) and FOXP3 (open bars) in stimulated cells as described above, displayed as mean fluorescent intensity (MFI).

Figure 6. Activated CD25^lo T cells demethylate during proliferation and exert suppressive activity in vitro.

The methylation level of activated cells derived from the CD25^lo population was monitored with the COBRA based analysis described in Figure 4. (A) Cultures from three separate donors were setup and stimulated as in Figure 5, after which the methylation status in the CD25^hi and CD25^lo fraction was monitored for 16 days. Cultures of T regulatory cells from two donors were kept in parallel and likewise analyzed with respect to methylation at day 0 as well as indicated time points. Shown are mean values±SEM. (B) FOXP3 expression in the CD25^hi and CD25^lo cells isolated at day 0 was determined by intracellular flow cytometry (left graph). During activation, FOXP3 expression was also determined in the sorted populations derived from stimulated CD25^lo cells (right graph). Data represent mean values from three separate donors±SEM. (C) CD25^hi and CD25^lo cells from three donors were isolated on day 5 post-activation. To examine the suppressive capacity of these activated CD25^lo derived cells, they were put into co-culture with autologous CFSE-stained non-activated CD4⁺ T cells in the presence of anti-CD3 and anti-CD28 antibodies, and autologous CD4⁻ cells as feeder cells. A control sample was included for each donor where only responder and feeder cells were included in the culture. Suppressive activity was evaluated on day 3 of co-culture as the mean fluorescence intensity ratio of CFSE⁺ responder cells in each sample relative the control sample.

Figure 7. Activation results in a partial demethylation at long-term follow-up.

CD25 and FOXP3 expression was monitored in CD4⁺CD25^lo cells during activation with (A) CD3/CD28 dynabeads as described in Figure 5 or repeated stimulation with anti-CD3 (5 µg/mL) and anti-CD28 (1 µg/mL) antibodies every 7 days. CD4⁺CD25^hi Tregs were also monitored as they were stimulated continuously with CD3/CD28 dynabeads. (B) CD4⁺CD25^lo cells stimulated with high (10 µg/mL), medium (5 µg/mL) or low (0.5 µg/mL) anti-CD3 antibody together with 1 µg/mL anti-CD28 antibody. (C) CD4⁺CD25^lo cells stimulated with anti-CD3 (5 µg/mL) and anti-CD28 (1 µg/mL) antibodies together with TGF-β (5 ng/mL), TGF-β (5 ng/mL) and IL-10 (10 ng/mL) or IL-10 only (10 ng/mL). Stimuli were removed on day 2 after stimulation and FACS analysis performed on days 0, 2, 5, 7, 10, 14 and 21. Data represent mean values from three separate donors±SEM, except for expanded Tregs where data represent mean values from two separate donors. Methylation status of the −77 reporter position was monitored during conditions described above (D–F). Evaluation of methylation status was performed on DNA from one donor at days 0, 7, 14 and 21 for each separate population. Methylation of expanded Tregs at day 7, 12 and 16 represent mean values from two separate donors.