Locus-Specific Reversible DNA Methylation Regulates Transient IL-10 Expression in Th1 Cells Free

Interleukin-10 is an anti-inflammatory cytokine and plays important roles in immune homeostasis by limiting hyperimmune responses (1, 2). Dysregulated IL-10 expression has been implicated in various immune disorders, such as infectious diseases and autoimmune and allergic disorders (3–9). IL-10–deficient mice spontaneously develop inflammatory bowel disease and display exaggerated inflammatory responses to microbial challenge (10). IL-10 is produced by diverse immune cell types, such as Th cells, regulatory T cells, B cells, mast cells, macrophages, and dendritic cells (11).

CD4⁺ T cells play important functions in protective immunity against many types of pathogens. Naive CD4⁺ T cells can differentiate into Th1 and Th2 cells when they meet Ags under a IL-12 or IL-4 cytokine milieu, respectively, largely provided by APCs. Th1 cells produce IFN-γ, which protects against intracellular pathogens, such as virus and bacteria. In contrast, Th2 cells produce IL-4, IL-5, and IL-13, which protect against extracellular pathogens while mediating allergic disorders (12). Although IL-10 was first reported as Th2-type cytokine (13), conventional Th1 cells also express marginal levels of IL-10 (14, 15). The differential expression levels between Th1 and Th2 cells might be mediated by diverse mechanisms, including active repression of IL-10 expression in Th1 cells (16–18). However, despite this, when large amounts of IL-12 and IL-27 are available, certain chronic inflammatory conditions may lead to substantial derepression of the Il-10 gene locus in Th1 cells, presumably as a precaution against overt immune responses (19–21). These findings suggest that Th1 cells might have negative- and positive-feedback loops to keep a balance between effective immune responses and immunopathology. The underlying mechanisms of such dynamic Il-10 gene expression in Th1 cells are largely unknown.

Epigenetic mechanisms regulate gene expression without altering the sequence of bases in the DNA. Particularly, DNA methylation at gene promoters and distal regulatory elements can directly inhibit transcription (22). Methylated DNA generally represses gene expression by blocking the binding of transcription factors and RNA polymerase II to the regulatory loci and by recruiting silencing factors, including methyl-CpG binding domain proteins (23–25). The role of differential DNA methylation in cytokine gene expression was also reported in effector Th cells (26, 27). We (18) and other investigators (28, 29) reported that differential DNA methylation status regulates IL-10 expression in cell lines or primary CD4⁺ T cells. However, it is unknown whether certain inflammatory conditions may reversibly modulate the DNA methylation status of the Il-10 gene locus.

In this study, we have analyzed DNA methylation patterns in the major CpG locus of Il-10 regulatory elements in conventional Th1 and Th2 cells, as well as in Th1 cells under defined inflammatory conditions. We found that DNA methylation of the Il-10 gene locus is positively correlated with its repressive state in Th1 cells. However, Th1 cells treated with IL-12/IL-27 cytokines or IL-10–producing CD4⁺ T cells isolated from lymphocytic choriomeningitis virus (LCMV)-infected mice showed a reversible DNA demethylation, specifically at region of interest (ROI) 3, one of the prominent methylated regions of the Il-10 gene locus identified by our assay. We further demonstrate that binding of transcription factors STAT1 and STAT3 to ROI 3 mediates high levels of Il-10 gene expression in an IL-12/IL-27–dependent manner.

C57BL/6 mice were purchased from the Joong-Ang Experimental Animal Center (Seoul, Korea). B6(Cg)-Il-10^tm1.1Karp/J (Vert-X; IL-10^GFP reporter) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). STAT1-knockout (KO) mice and Lck^cre STAT3^fl/fl CD4–specific conditional deletion mice were kindly provided by Dr. M.-R. Song (Gwangju Institute of Science and Technology) and Dr. S.-K. Ye (Seoul National University, Seoul, Korea), respectively. For the LCMV Armstrong infection model, mice were infected (2 × 10⁵ PFU) by i.p. injection (30). All mice were housed in specific pathogen–free barrier facilities and used in accordance with protocols approved by the Pohang University of Science and Technology Institutional Animal Care and Use Committee.

Naive CD4⁺ T cells from the lymph nodes and spleen of 6–8-wk-old mice were purified using an EasySep Mouse Naïve CD4⁺ T Cell Isolation Kit (STEMCELL Technologies). Briefly, total splenocytes and lymphocytes were incubated with a mixture of Abs for 10 min after RBC lysis. The mixture of Abs contains CD8α, CD11b, CD11c, CD19, CD25, CD45R, CD49b, TCRγδ, and TER119, which are conjugated with biotin. Biotin beads were added to Ab-labeled cells, and naive CD4⁺ T cells were obtained by negative selection using an EasySep magnet. Mouse primary cells were cultured in T cell media, which contains RPMI 1640 (Life Technologies) supplemented with 10% FBS (HyClone), 100 U/ml penicillin-streptomycin, 3 mM l-glutamine, 10 mM HEPES, and 0.05 mM 2-ME (all from Sigma-Aldrich). HEK293T cells were obtained from the Korean Cell Line Bank (Seoul National University). The cells were cultured in DMEM (Life Technologies) supplemented with 10% FBS (HyClone) and 100 U/ml penicillin-streptomycin (Sigma-Aldrich).

For in vitro Th1 and Th2 differentiation, purified naive CD4⁺ T cells (1 × 10⁶ per milliliter) were stimulated with plate-bound anti-CD3 (1 μg/ml) and anti-CD28 (2 μg/ml) Abs under Th1-differentiation conditions (10 ng/ml IL-12 Ab, 10 μg/ml anti–IL-4 Ab) or Th2-differentiation conditions (10 ng/ml IL-4 Ab, 10 μg/ml anti–IFN-γ Ab, 10 μg/ml anti–IL-12 Ab) for 3 d in T cell media. A total of 100 U/ml recombinant human IL-2 was added on day 3 after detaching from anti-CD3/anti-CD28 Abs, and the cells were expanded in complete medium containing IL-2 for an additional 3 d. This initial 6-d culture was defined as round 1 (R1). For further differentiation into two or three rounds, R1 or round 2 cells were differentiated in vitro for an additional 6 or 12 d, respectively, under the same Th1- or Th2-differentiation conditions described above. Recombinant human IL-2 and anti–IL-4 (11B11) Ab were provided by the National Cancer Institute, Preclinical Repository. TGF-β, IL-4, IL-6, IL-12, and IL-27 were purchased from R&D Systems. Anti-CD3 (145.2C11), anti-CD28 (37.51), anti–IFN-γ (XMG1.2), and anti–IL-12 (C17.8) Abs were obtained from Bio X Cell.

Total RNA was extracted by TRI Reagent (Molecular Research Center), and cDNA was generated using 1 μg of total RNA, oligo(dT) primer (Promega), and ImProm-II Reverse Transcriptase (Promega) in a total volume of 20 μl. A total of 0.5 μl of cDNA was amplified using SYBR Premix Ex Taq (Takara) and DNA Engine with Rotor-Gene Q (QIAGEN). Mouse HPRT primer was used for quantitative RT-PCR (qRT-PCR) to normalize the amount of cDNA used for each condition. The following primers were used for qRT-PCR: HPRT (forward: 5′-TTATGGACAGGACTGAAAGAC-3′, reverse: 5′-GCTTTAATGTAATCCAGCAGGT-3′) and IL-10 (forward: 5′-ATAACTGCACCCACTTCCCA-3′, reverse: 5′-TCATTTCCGATAAGGCTTGG-3′).

For intracellular staining of cytokines, an Intracellular Fixation & Permeabilization Buffer Set (catalog number 88-8824-00; eBioscience) was used with a minor modification of the manufacturer’s protocol. Briefly, cells were restimulated with 50 ng/ml PMA and 1 μM ionomycin in the presence of GolgiStop (BD Biosciences) for 5 h and harvested. Cells were stained with Fixable Viability Dye (eBioscience) for 30 min, washed, suspended in 100 μl of intracellular fixation buffer for 30 min, washed, and stained with Abs against cytokines (IL-10, IL-4, IFN-γ Abs were purchased from eBioscience, and IL-17A was from BioLegend) diluted in permeabilization buffer for 30 min. Cells were then washed, resuspended in PBS, and analyzed using a FACSCanto II flow cytometer (BD Biosciences). Data were analyzed using FlowJo software (TreeStar). To detect the GFP signal from IL-10 reporter mice, 100 μl of 4% paraformaldehyde was added for 10 min before using intracellular fixation buffer. GFP⁺ cells from IL-10 GFP reporter mice were sorted using a MoFlo XDP cell sorter (Beckman Coulter).

Genomic DNA was isolated using a genomic DNA preparation kit (NucleoSpin Tissue; MACHEREY-NAGEL). Methylated DNA immunoprecipitation (MeDIP) was performed as previously described (31). Briefly, 5 μg of genomic DNA was sonicated to an average length of 0.2–0.5 kb, boiled for 10 min, and incubated on ice for 10 min. DNA (4 μg) from the immunoprecipitated portions was incubated for 2 h at 4°C with a 5-methylcytosine Ab (Diagenode), followed by incubation with mouse IgG Dynabeads magnetic beads (Invitrogen) for 2 h at 4°C. The methylated DNA/Ab complexes were digested with proteinase K, and DNA enriched in 5-methylcytosine was recovered using Chelex 100 resin (Sigma-Aldrich). The immunoprecipitated DNA and input DNA pellets were then dissolved in water. The enrichment of methylated DNA was measured by real-time PCR using SYBR Premix Ex Taq (Takara). The following primers were used for MeDIP: ROI 1 (forward: 5′-AGAAAGTGAAAGGGATGGAGG-3′, reverse: 5′-TGGTAGAACAGGAACTCGGG-3′), ROI 2 (forward: 5′-AGAACAGGAGGTCTACATTTAGAG-3′, reverse: 5′-AGAAAGTCTTCACCTGGCTG-3′), ROI 3 (forward: 5′-CATTGTGGGTTATTAGCTACTC-3′, reverse: 5′-TAGGGATAGCAGTCTTCTGG-3′), H19 (forward: 5′-CTATCTTTGGTGTACCACTTCCC-3′, reverse: 5′-TCATGGCATCGAGAAAGAAACTG-3′), and GAPDH (forward: 5′-AATGAAGCCTGTAACAACGCC-3′, reverse: 5′-TCTCAGCAATCCCTCCTTACC-3′). As a loading control, PCR was performed directly on input DNA purified from chromatin before MeDIP. Data are presented as the amount of DNA recovered relative to the input control.

After differentiation of Th1 and Th2 cells from naive CD4⁺ T cells, genomic DNA was purified using the manufacturer’s protocol (NucleoSpin; MACHEREY-NAGEL). Each genomic DNA was treated with HpaII (R0171) and MspI (R0106; both from NEB). The samples were incubated for 3 h at 37°C and then for 20 min at 65°C for heat inactivation. PCR was carried out in a volume of 20 μl with specific primers corresponding to ROIs 1–3 and with control primers which were used for MeDIP. Standard PCR was performed, and the amplified products were visualized in 1% agarose gel.

Bisulfite modification of genomic DNA was performed using an EZ DNA Methylation-Lightning Kit (Zymo Research), according to the manufacturer’s instructions. A bisulfite conversion reaction was carried out on 500 ng of genomic DNA. The reaction volume was adjusted to 20 μl with sterile water, and 130 μl of CT Conversion Reagent was added. The samples were placed in a thermal cycler (MJ Research), and the following steps were performed: 8 min at 98°C, 60 min at 54°C, and storage at 4°C for up to 20 h. The bisulfite-treated DNA was purified using the reagent contained in an EZ DNA Methylation-Lightning Kit (Zymo Research). The converted samples were added to a Zymo-Spin IC Column containing 600 μl of the binding buffer, mixed by inverting the column several times, and centrifuged at full speed for 30 s to remove the flow-through. The column was washed by adding 200 μl of wash buffer and centrifuged at full speed, and 200 μl of desulfonation buffer was added. After incubation for 20 min, the column was centrifuged at full speed for 30 s, washed by adding 200 μl of wash buffer, and centrifuged at full speed. The converted genomic DNA was eluted by adding 20 μl of elution buffer to the column.

Each primer was designed using Pyrosequencing Assay Design Software v2.0 (QIAGEN). The following primers were used: ROI 1 (forward: 5′-TGTTAGGAGGAGAGGTTAGATT-3′, biotinylated reverse: 5′-AATAAACCCATATAAAAAATACACTCCC, sequencing 1: 5′-AATAAACCCATATAAAAAATACACTCCC-3′, sequencing 2: 5′-GGGTAGGTTTGGAAT-3′), ROI 2 (forward: 5′-ATGAGGATTAGTAGGGGTTAGT-3′, biotinylated reverse: 5′-ACAATTCCATCAAAATACATATTTCAAA-3′, sequencing 1: 5′-GGATTAGTAGGGGTTAGTATA-3′, sequencing 2: 5′-GAGTTATATGTTTTTAGAGTTG-3′), and ROI 3 (forward: 5′-GGGTGAGAAGTTGAAGATTTTTA-3′, biotinylated reverse: 5′-CTCAAAATCACTCCCACACT-3, sequencing 1: 5′-GTTGAAGATTTTTAGGATG-3′, sequencing 2: 5′-ATTAGATAGGAGATTAGGTAAA-3′). PCR was carried out in a volume of 20 μl with 20 ng of converted genomic DNA, PCR premixture (Enzynomics), 1 μl of 10 pmol/μl primer, and 1 μl of 10 pmol/μl biotinylated primer. PCR experiments were performed under the same conditions, as follows: 95°C for 10 min, followed by 45 cycles at 95°C for 30 s, 54°C or 60°C for 30 s, and 72°C for 30 s, and a final extension at 72°C for 5 min. DNA template was prepared from 16 to 18 μl of biotinylated PCR product using Streptavidin Sepharose High Performance beads (Amersham Biosciences), following the PS Q96 sample preparation guide, using multichannel pipets. The respective sequencing primer (15 pmol) was added for analysis. Sequencing was performed on a PyroMark ID system with PyroMark Gold Reagents (QIAGEN), according to the manufacturer’s instructions. The methylation percentage was calculated as the average of the degree of methylation at two or three CpG sites formulated in pyrosequencing.

Two reporter constructs were generated in a combination of ROIs 1+2 and ROIs 1+3 (see Fig. 1D and Supplemental Fig. 1 for the corresponding ROIs and their sequences, respectively). Each insert (500 μg) was methylated in vitro using SssI methylase enzyme (Zymo Research) or was left unmethylated. The methylation status was confirmed by methylation-sensitive restriction enzymes, such as HaeII and HpaII. The methylated and unmethylated regions were religated into pXPG plasmid. Each construct was transfected into the HEK293T cell line with Renilla luciferase plasmid as a control for the transfection efficiency, and relative luciferase activity was measured.

HEK293T cells were transfected using GeneExpresso (Excellgen), according to the manufacturer’s protocol, and plated in a 12-well plate. For STAT1 and STAT3 overexpression, STAT1 or STAT3 expression vectors were cotransfected in a different amount, as indicated. After 18 h, cells were stimulated with 50 ng/ml PMA and 1 μM ionomycin for 6 h, and luciferase activity was measured using a Dual-Luciferase Reporter Assay System (Promega). Data were normalized by the activity of Renilla luciferase, which was used as an internal control for transfection.

Chromatin immunoprecipitation (ChIP) assays were performed, as described previously (32, 33). Briefly, cells were cross-linked with formaldehyde at a final concentration of 1%, lysed, and sonicated to shear DNA. After immunoprecipitation with anti-STAT1 Ab, anti-STAT3 Ab (both from Santa Cruz Biotechnology), or isotype IgG (Vector Laboratories) at 4°C overnight, Ab/DNA complexes were eluted, and cross-linking was reversed. Following reversal of cross-links, the presence of selected DNA sequences was assessed by real-time PCR using the following primers sets: ROI 3-1 to ROI 3-3 (forward: 5′-TCAAGGAGCATTTGAATTCCC-3′, reverse: 5′-GAGACAGTATTTGTTGGGTGG-3′), and ROI 3-4 to ROI 3-6 (forward: 5′-ACACAGACTAGACAGGAGAC-3, reverse: 5′-TGACTCACAAATAACAGGAAGG-3′). Data are presented as the amount of DNA recovered relative to the input control. The result of ChIP with isotype IgG was confirmed as a background value (less than 0.001% of relative signal to input).

CpG-rich regions near the Il-10 gene locus were identified using the European Molecular Biology Laboratory database by EMBOSS CpG plot analysis (34). Prediction of transcription factor binding sites in the Il-10 gene locus was performed using the Web-based ECR Browser (35).

All experiments were performed more than three times independently. Statistical analyses were performed using the Student t test (two tailed, two sample unequal variance) or one-way ANOVA. In bar graphs, bars indicate means, and error bars indicate SEM. The p values < 0.05 were considered significant (*p < 0.05, **p < 0.005, ***p < 0.001).

To investigate the underlying mechanisms contributing to the dynamic changes in IL-10 expression in CD4⁺ T cells at the epigenetic level, we first analyzed the kinetics of IL-10 mRNA in in vitro–differentiated Th1 and Th2 cells over a period of 6 d. In accordance with previous reports (16, 18, 36), fully differentiated Th1 cells expressed much lower levels of IL-10 mRNA compared with Th2 cells (Fig. 1A, 1B). However, Th1 cells displayed a temporal increase in IL-10 mRNA expression during the initial stage of differentiation, between days 0 and 3 (Fig. 1A), which might be mediated by the effect of IL-12 (37). At the later stage of differentiation, more significant differences were observed in IL-10 mRNA expression between Th1 and Th2 cells, regardless of the stimulation conditions (Fig. 1A, 1B). We asked whether DNA methylation status contributes to the observed differences in IL-10 mRNA expression between Th1 and Th2 cells. Treatment with 5-azacytidine, an inhibitor of DNA methylation, during the in vitro differentiation period (between days 3 and 6), led to a 2-fold increase in IL-10 mRNA expression in Th1 cells; however, no significant difference was observed in Th2 cells (Fig. 1C). Based on these results, we hypothesized that the IL-10^low phenotype of Th1 cells could be mediated by increased DNA methylation in the Il-10 gene locus.

Next, we asked whether locus-specific or global DNA methylation in the Il-10 gene locus affects IL-10 expression. EMBOSS CpG plot analysis revealed three CpG sites located in the exon 1 and intron 3 regions. We named them ROI 1 (located in the promoter), ROI 2 (located in exon 1), and ROI 3 (located in intron 3) (Fig. 1D, Supplemental Fig. 1). DNA methylation status in each locus was analyzed using a methylation-sensitive enzymatic cleavage (MSEC) assay and MeDIP. First, we performed an MSEC assay using MspI (methylation-insensitive) and HpaII (methylation-sensitive) enzymes. MspI and HpaII recognize CCGG for cleavage, whereas they show differential sensitivity depending on the methylation status. Genomic DNA isolated from Th1 and Th2 cells was treated with HpaII or MspI and amplified with the locus-specific primer pairs listed in 2Materials and Methods. DNA treated with HpaII showed differential sensitivity between Th1 and Th2 cells. Genomic DNA prepared from Th1 cells was resistant to HpaII digestion and showed amplified bands for ROIs 1–3, whereas genomic DNA prepared from Th2 cells did not show amplified bands (Fig. 1E). Genomic DNA treated with MspI enzyme did not show any PCR amplification, regardless of origin. As a control, we confirmed that the H19 (nonexpressed target in lymphocytes) promoter region showed resistance to HpaII digestion in Th1 and Th2 cells (Fig. 1E, bottom panel). Next, we tested the effect of DNA methylation on IL-10 expression using a methylation-dependent reporter assay. First, we generated two reporter constructs composed of ROIs 1 + 2 and ROIs 1 + 3 that were methylated by treatment with SssI enzyme or left unmethylated. HEK293T cells were transfected with the reporter constructs, either unmethylated or methylated, followed by detection of luciferase activity. Compared with unmethylated constructs, methylated constructs showed a significant reduction in relative luciferase activity (Fig. 1F). These results suggested that DNA methylation status mediates differential IL-10 mRNA expression between Th1 and Th2 cells.

To elucidate the kinetics and stability of DNA methylation in the Il-10 gene locus, a 6-d differentiation cycle was repeated for three rounds (R1 through round 3 [R3]) under Th1- and Th2-inducing conditions (Fig. 2A). IL-10 mRNA levels were progressively reduced in Th1 cells as differentiation continued, whereas Th2 cells showed the opposite phenomenon. Especially, Th1 cells in the round 2 and R3 stages showed a significant reduction in IL-10 mRNA levels compared with the R1 stage (Fig. 2B). The methylation status in each round of in vitro differentiation was analyzed by MeDIP assay and pyrosequencing analysis. Compared with Th2 cells, Th1 cells showed higher enrichment of methylation signals in the CpG regions in each round of differentiation in all ROIs (Fig. 2C). As an experimental control for MeDIP, we confirmed the hypomethylation status in the Gapdh locus (housekeeping gene, negative control) as well as the hypermethylation status in the H19 promoter region (not expressed in lymphocytes, positive control). We further analyzed DNA methylation status by bisulfite pyrosequencing for all three ROIs. Depending on their CpG locations within the ROIs, we assigned them subnumbers, such as ROI 1-1 to ROI 1-5 (Supplemental Fig. 1). The percentage of methylation at the CpG sites in each ROI was determined by pyrosequencing analysis, which was repeated more than three times. As expected, there was a sharp decrease in DNA methylation in all ROIs through the rounds of differentiation under Th2 conditions. In contrast, methylation was gradually increased within ROIs 1 and 2 of Th1 cells as differentiation continues to the R3 stage. Interestingly, ROI 3 under Th1 conditions showed a distinct pattern: the DNA methylation status of all CpG residues largely remained constant or even displayed a subtle reduction during the later rounds of differentiation (Fig. 2D).

It is known that Th1 cells are capable of expressing high levels of IL-10 when they are stimulated with strong TCR activation in the presence of IL-12 and IL-27 cytokines (20, 37–40), which mimics chronic inflammatory conditions in vivo. Therefore, we tested whether upregulation of IL-10 expression by IL-12/IL-27 treatment is mediated by an alteration in the DNA methylation pattern in Th1 cells. First, we tested the effect of IL-12 during in vitro Th1 cell differentiation. As expected, maintaining high levels of IL-12 significantly enhanced IL-10 expression in Th1 cells during all three rounds of differentiation (Fig. 3A). Interestingly, however, no significant difference was observed in DNA methylation status between conventional Th1 cells (without IL-12) and Th1 cells maintained with high levels of IL-12 up to R3 of differentiation (Supplemental Fig. 2A). Next, we examined the cooperative role of IL-12 and IL-27 in inducing IL-10 expression. Compared with individual cytokine treatment, cotreatment with IL-12 and IL-27 significantly enhanced IL-10 expression in ex vivo–isolated CD4⁺ T cells (Fig. 3B, 3C). We also tested the effect of IL-12/IL-27 treatment on IL-10 expression in Th1 cells. Naive CD4⁺ T cells isolated from IL-10–GFP reporter mice were differentiated into Th1 cells for 6 d (R1 stage). Then, Th1 cells were cultured under TCR stimulation in the absence (mock) or presence of IL-12/IL-27 cytokines for an additional 3 d. Subsequently, cells were washed to remove exogenous cytokines, transferred to fresh media, and maintained in the resting stage for 5 d (Supplemental Fig. 2B). Th1 cells stimulated under IL-12/IL-27 cytokines showed a significant increase in IL-10 expression compared with control (mock, no cytokine treatment) Th1 cells (Fig. 3D). Removal of IL-12/IL-27 cytokines (resting stage) significantly reduced IL-10 expression similar to the levels in mock-treated cells (Supplemental Fig. 2B). Next, we tested whether enhanced IL-10 expression by IL-12/IL-27 cotreatment is associated with an alteration in DNA methylation status at the ROIs. Cells were sorted based on their GFP (IL-10) expression, and pyrosequencing was performed. Th1 cells differentiated in vitro showed overall methylation patterns, and no significant difference was observed between GFP⁺ and GFP⁻ populations. In the absence of IL-12/IL-27 treatment (mock), a significant difference was only observed in ROI 1-4 and ROI 1-5 between GFP⁺ and GFP⁻ cells (Fig. 3E, 3F). Interestingly, upon IL-12/IL-27 cytokine treatment, specific DNA demethylation was significantly increased, mainly at ROI 3-4 and ROI 3-6, in the GFP⁺ population. Furthermore, removal of IL-12/IL-27 cytokines (resting stage) reduced the levels of IL-10 expression and DNA demethylation (Supplemental Fig. 2B, 2C). These findings suggest that enhanced IL-10 expression in Th1 cells by IL-12/IL-27 treatment is linked to a locus-specific transient remodeling of DNA methylation status at ROI 3.

We next wanted to understand the physiological relevance of IL-12/IL-17–mediated site-specific DNA demethylation at the ROI 3 locus using an in vivo model of infection. For this purpose, we used an LCMV Armstrong infection model that is known to increase IL-10 levels under Th1-dominant inflammatory conditions (30, 41, 42). IL-10–GFP–reporter mice were infected with LCMV or were mock infected, and CD4⁺ T cells were isolated. Compared with the control group, CD4⁺ T cells isolated from LCMV-infected mice showed higher levels of IL-10 expression (Fig. 4A). Next, DNA methylation status at the different ROIs of CD4⁺GFP⁺ cells sorted from mock- or LCMV-infected mice was analyzed by pyrosequencing. IL-10–expressing CD4⁺ T cells isolated from LCMV-infected mice showed enhanced methylation patterns at ROI 1 (promoter), whereas no change was observed at ROI 2 (Fig. 4B). Interestingly, the most significant difference in the DNA demethylation pattern was observed at ROI 3, especially at ROIs 3-4, 3-5, and 3-6 (Fig. 4B). These results showed a close correlation with the data from Th1 cells treated with IL-12/IL-27 in vitro (Fig. 3E, 3F), suggesting that IL-10 expression under Th1-type inflammation is mediated by a reversible DNA demethylation at ROI 3.

Site-specific CpG demethylation favors the binding of transcription factors to induce the expression of their target genes (43). Bioinformatics analysis using rVista 2.0 (44) revealed putative transcription factor binding sites within ROI 3, especially near ROIs 3-4, 3-5, and 3-6 (Supplemental Fig. 3A). Among the predicted transcription factors, we mainly focused on the roles of STAT1 and STAT3, because they were reported previously to be common downstream regulators for IL-12 and IL-27 cytokine signaling (45–49). First, we tested whether STAT1 and STAT3 could enhance Il-10 reporter activity in vitro. An Il-10 reporter construct that contains promoter (ROI 1) and ROI 3 (see also Fig. 1F) was cotransfected with STAT1 or STAT3, under the indicated combinations, into HEK293T cells, and Il-10 reporter activity was measured. Indeed, overexpression of STAT1 and STAT3 significantly enhanced ROI 3–mediated transactivation of Il-10 promoter reporter activity in a dose-dependent manner (Fig. 5A). We further confirmed whether STAT1 and STAT3 bind to ROI 3 (3-4, 3-5, 3-6) in response to IL-12/IL-27 treatment. Th1 cells were treated with IL-12/IL-27 or were left untreated, and the physical association of STAT1 and STAT3 with ROI 3 (3-4, 3-5, 3-6) was analyzed by ChIP analysis. Indeed, higher enrichment of STAT1 and STAT3 was observed at ROI 3 (3-4, 3-5, 3-6) upon IL-12/IL-27 treatment (Fig. 5B). No significant enrichment of STATs was observed in the heterochromatic region of Mest, regardless of IL-12/IL-27 treatment. Isotype IgG was confirmed as a background value. These results indicate that STAT1 and STAT3 could transactivate Il-10 expression by binding to ROI 3. We further validated the role of STAT1 and STAT3 in IL-10 expression. CD4⁺ T cells isolated from STAT1-KO or STAT3–conditional KO (cKO) (STAT3^fl/fl Lck^cre) mice were stimulated in the presence or absence of IL-12, IL-27, or IL-12/IL-27, and IL-10 levels were measured in whole CD4⁺ T cells (Supplemental Fig. 3B, 3C) and among the IFN-γ⁺ population (Fig. 5C, 5D). Compared with wild-type CD4⁺ T cells, IL-10 expression was significantly reduced in CD4⁺ T cells isolated from STAT1-KO or STAT3-cKO mice, regardless of the type of stimulation (Fig. 5C, 5D, right panels), although the difference in the protein level was modest (Supplemental Fig. 3B, 3C). We could observe a more significant reduction in IL-10 expression among the IFN-γ⁺ population of CD4⁺ T cells from STAT1-KO and STAT3-cKO mice (Fig. 5C, 5D, left panels). These results collectively suggested that binding of STAT1 and STAT3 to the demethylated ROI 3 enhances IL-10 expression in response to IL-12/IL-27, an in vivo Th1-mimicking inflammatory condition.

Epigenetic regulation plays an important role in selective expression of Th subset–specific effector cytokines. CpG DNA methylation is one of the key mechanisms for stable silencing of active gene loci. In this study, we investigated the role of DNA methylation in regulating IL-10 gene expression in Th1 cells under inflammatory conditions. We found that reversible DNA methylation, mainly at ROI 3, correlates with transient IL-10 expression in a STAT1/3-dependent manner.

Previous studies have suggested that cell type–specific distinct DNA demethylation patterns are well correlated with the levels of cytokine expression. For example, DNA demethylation at the promoter regions of the Ifng and Il-4 loci regulates their Th1- or Th2-specific expression, respectively (26, 27). Interestingly, however, IL-10 expression is relatively promiscuous, and various Th subsets express IL-10 (11). Th1 cells also express IL-10 under certain inflammatory conditions as a feedback mechanism of preventing hyperimmune activation (21, 50). Most of the studies were focused on the role of specific transcription factors as a positive regulator of Il-10 gene expression (15, 51–54). It is still unclear whether dynamic or static maintenance of epigenetic memory is involved in Il-10 gene expression, although involvement of a reversible histone deacetylase–responsive state was reported in Th1 cells (16, 18, 55). Unlike histone modification, DNA methylation is regarded as one of the key mechanisms of stable silencing, which is not easily reversed (27). In this study, we tested whether DNA methylation is reversible or an irrepressible process in the dynamic regulation of IL-10 expression in Th1 cells. In line with previous studies (17, 18), we confirmed that IL-10 expression was gradually repressed under repetitive conventional Th1 differentiation accompanied by an increased DNA methylation status (Fig. 2). Interestingly, treatment of Th1 cells with IL-12/IL-27, which mimics an in vivo inflammatory condition, significantly enhanced IL-10 expression through selective demethylation at ROI 3 (Fig. 3). Moreover, IL-10–expressing CD4⁺ T cells isolated from LCMV-infected mice showed enhanced DNA demethylation at ROI 3. However, the dynamics of DNA methylation status at ROI 1 (promoter region) do not correlate between in vitro and in vivo systems. Treatment of Th1 cells with IL-12/IL-27 in vitro enhanced DNA demethylation at ROI 1 (Fig. 3E, 3F). Interestingly, however, a significant increase in DNA methylation at ROI 1 was observed in the IL-10–expressing CD4⁺ T cells isolated from LCMV-infected mice (Fig. 4B). Previous reports also support our notion that, compared with naive CD4⁺ T cells, activated CD4⁺ T cells showed a significant increase in DNA methylation at ROI 1 (promoter) while showing higher IL-10 expression (56). Th0 cells treated with IL-27 in vitro showed higher DNA methylation at ROI 1 (28, 29). These results suggest that DNA methylation at the promoter is responsible for the transcriptional repression program of Il-10 gene expression in Th1 cells, whereas locus-specific DNA demethylation, especially at ROI 3, might be a key mechanism for keeping tight control of transient IL-10 production in pathologic conditions. Because a prolonged inflammatory response might have a detrimental effect on immune homeostasis, Th1 cells may develop a feedback loop mechanism under inflammatory conditions by transiently inducing IL-10 expression through a locus-specific demethylation at ROI 3.

Bioinformatics analysis revealed the possibility that STAT proteins may bind to ROI 3. Indeed, ChIP and an Il-10 reporter assay confirmed that STAT1 and STAT3 bind to demethylated ROI 3 (3-4, 3-5, 3-6), which enhances IL-10 expression in response to IL-12/IL-27 (Fig. 5A, 5B). In addition, STAT1 or STAT3 deficiency led to a significant reduction in IL-10 expression in IFN-γ⁺ CD4⁺ T cells treated with IL-12/IL-27 (Fig. 5C, 5D). These results show good correlation with the previously reported data that STAT1 and STAT3 are key transcription factors for inducing IL-10 expression in response to IL-12 and IL-27 in CD4⁺ T cells (45, 57). In addition to STAT1 and STAT3, other factors play a role in inducing IL-10 expression in Th1 cells. Blimp1 and c-Maf are also reported to be key regulators of IL-10 expression in IFN-γ⁺ Th1 cells (51). Both were reported to associate with the CNS-9 region of the IL-10 locus, which is 9 kb upstream from the transcription start site (51). E4bp4 and ETV5 are also known as positive regulators of IL-10 regulation in IFN-γ–producing Th1 cells (58, 59). It will be interesting to test whether there is any redundancy or synergism among these transcription factors to induce IL-10 expression in Th1 cells in response to IL-12/IL-27 cytokines, but we believe that this type of detailed investigation lies beyond the scope of the current study.

In summary, our data demonstrate that locus-specific reversible DNA methylation might act as a feedback mechanism for preventing hyperinflammation by inducing IL-10 expression in Th1 cells. These findings underlie the importance of DNA methylation as a threshold platform, as well as point toward distinct cell type–specific mechanisms of IL-10 expression. Further unraveling of these cell type–specific anti-inflammatory programs may lead to the development of more targeted therapeutic strategies.

We thank Dr. Mi-Ryoung Song and Dr. Sang-Kyu Ye for providing STAT1-KO and STAT3-cKO (STAT3^fl/fl Lck^cre) mice, respectively.

The authors have no financial conflicts of interest.