Transcription factor-dependent chromatin remodeling of Il18r1 during Th1 and Th2 differentiation ¹

Mice

The generation of C57BL/6 Stat4-/- mice was previously described (17). WT C57BL/6 and BALB/c mice were purchased from Harlan Bioproducts (Indianapolis IN). Balb/c Stat6-/- mice (18) were purchased from Jackson laboratories (Bar Harbor, ME). All experiments used C57BL/6 mice, except for Balb/c WT and Stat6-/- mice as noted in figure legends. Mice were maintained in pathogen-free conditions in barrier facilities in the Laboratory Animal Resource Center (Indiana University School of Medicine). All experiments were performed following approval of the IU Animal Care and Use Committee.

Th differentiation and analysis

CD4 cells were isolated from spleen and lymph nodes of mice using magnetic beads (Miltenyi Biotec). For Th differentiation, CD4 cells (1 × 10⁶ cells/ml) were cultured with plate bound anti-CD3 (4 μg/ml), 0.5 μg/ml soluble anti-CD28, under Th1 (2 ng/ml IL-12 and 10 μg/ml anti-IL-4) or Th2 (10 ng/ml IL-4 and 10 μg/ml anti-IFNγ) skewing conditions and expanded after three days. After 5-7 days of culture, cells were harvested for FACS analysis with anti-IL-18Rα (R&D Systems, Minneapolis, MN), RNA isolation and quantitative PCR (performed as described (19)) or assays described below. Transduction with GATA-3-hCD4 expressing bicistronic retroviruses was performed as previously described (20), except that cells had been differentiated towards the Th1 phenotype. Five days after transduction, hCD4+ cells were analyzed for expression of IL-18Rα or CXCR3. In parallel experiments, hCD4+ cells were purified from control and GATA-3 transduced cells, and RNA was isolated for analysis.

DNase hypersensitivity analysis

Isolation and DNase I digestion of nuclei was performed as described (21). Briefly, CD4+ T cells were washed in cold PBS and resuspended in 0.3 M sucrose. Membranes were disrupted with 0.2% NP-40 and cells were incubated on ice for 10 minutes before Dounce homogenization. The cell suspension was layered onto 1.7 M sucrose cushion and nuclei were purified by centrifugation at 13,000 rpm for 15 minutes. The pellet was resuspended in cold buffer (0.3M sucrose, 5% glycerol). Aliquots of purified nuclei were digested with DNase I (Worthington) in 5 mM CaCl2 and 1 mM MgCl2 for three minutes at room temperature. Reactions were terminated with 5% SDS, 100 mM EDTA, 20 mM Tris pH 7.4 and 2mg/ml Proteinase K. DNA was purified using phenol/chloroform and digested to completion with EcoRI or EcoRV as indicated. DNA was separated through 1% agarose and fragments were detected using Southern analysis. Sequences of primers used to generate probes for Southern blotting are available upon request.

Chromatin immunoprecipitation

ChIP was performed as described (7). Briefly, cells were fixed in formaldehyde and nuclei were isolated after cell lysis and centrifugation. Similar numbers of cells for each condition were used in an experiment to facilitate comparisons among samples. Following nuclear lysis, chromatin was fragmented by sonication. Extracts were pre-cleared with salmon sperm DNA, BSA and protein A agarose slurry. The supernatant was incubated overnight with antibodies to acetyl-H3, aceytl-H4, acetyl-H3K9, H3K4me2, H3K4me3, Stat4, Stat6, T-bet, GATA-3 or control IgG as indicated. Antibodies were purchased from Millipore/Upstate Biotechnology or Santa Cruz Biotechnology. Complexes were precipitated with protein A agarose beads. Supernatants from the control precipitation were used as input material. Precipitates were washed with low- and high-salt buffers, and DNA was eluted from the precipitate. DNA crosslinks were reversed at 65°C and DNA was purified using phenol-chloroform. QPCR was performed with site-specific primer sets using ABI PRISM7500. Results are expressed as percent input (PCR from specific ChIP – PCR from control IgG)/PCR of input DNA ± SD of replicates from representative data of 2-3 experiments. Statistics were performed using a two-tailed student's T test. The sequences of the primer pairs are available upon request.

Chromatin remodeling at the Il18r1 gene

The IL-18R locus on mouse chromosome 1 contains the closely linked Il18r1 and Il18rap genes encoding respectively the IL-18Rα and IL-18Rβ receptor components, and the Il1rl1 gene, encoding the T1/ST2/IL-33R protein, within a 100 kb chromosome segment that is about 70 kb telomeric from other IL-1R family genes (Fig. 1A). The IL-18R genes are expressed in undifferentiated CD4+ T cells and expression is dynamically modulated during Th1 or Th2 differentiation (Fig. 1B and C). Surface expression of IL-18Rα is decreased 24 and 48 hours after activation in both Th1 and Th2 cultures, but while it continues to be expressed on only a small portion of Th2 cells, it is induced during Th1 differentiation and is expressed on the majority of differentiated Th1 cells (Fig. 1B). Similar patterns of expression during differentiation are observed in Balb/c and C57BL/6 cultures (Fig. 1B). Message levels assessed using quantitative PCR demonstrates a similar pattern of expression for both Il18r1 and Il18rap being repressed in Th2 and induced in Th1 C57BL/6 cultures, with similar expression patterns in Balb/c cells (Fig. 1C and data not shown).

To further explore the regulation of the Il18r1 gene, we performed DNase hypersensitivity analysis on regions near the transcriptional start site (TSS)³. A genomic DNA probe (p1) was used to define DNase hypersensitivity sites (DH) of an EcoRV fragment that spanned the promoter, exon 1 and the 5′ end of intron 1 of Il18r1 (Fig. 2A). Th1 or Th2 cultures were differentiated for 1 week and isolated nuclei were treated with increasing doses of DNaseI. One DH site (identified as DHI) was identified in the 5′ end of Il18r1 that was present in undifferentiated CD4 T cells. DHI sensitivity was increased following differentiation and showed more sensitivity to DNaseI in Th1 than in Th2 cells, comparing the ratio of the EcoRV band to the DHI band (Fig. 2B). To test whether chromatin was progressively remodeled with increasing time in culture, we differentiated Th1 or Th2 cells for two weeks and repeated the analysis. DHI appeared to show increased sensitivity in Th1 cells while it was increasingly resistant to DNaseI treatment following two weeks of culture in Th2 conditions (Fig. 2B). As we have previously shown that Il18r1 is a Stat4 target gene (7), we also tested the appearance of DHI in Stat4-/- cultures. Despite the ability of Stat4 to mediate increased locus acetylation, the sensitivity of DHI in Stat4-deficient Th1 cultures was indistinguishable from wild type Th1 cells (Fig. 2B). These results suggest that chromatin remodeling of the 5′ region of the Il18r1 gene progresses independently of IL-12 signaling.

To define alterations in chromatin structure in the central portion of the Il18r1 locus, we performed similar experiments to examine DNase hypersensitivity in a second EcoRV fragment that spanned introns 1-7 (probe p2). We observed a number of DNase hypersensitivity sites in this fragment that we termed DHII-DHV (Fig. 2C). DHII was observed most prominently in Th1 cultures after two weeks of culture but was detected in undifferentiated T cells, and one-week Th1 and Th2 cells (Fig. 2C). DHIII was most apparent in Th2 cultures, particularly after two weeks of culture. DHIV was prominent in undifferentiated T cells, and was similar in Th1 and Th2 cells, though after 2 weeks of culture it was less sensitive to DNaseI in Th2 cultures. DHV was not a strong DH site and was equally seen in Th1 and Th2 cells (Fig. 2C). As was observed in the 5′ end of the locus, the establishment of the Th1 chromatin configuration was independent of Stat4.

We examined DH sites in the 3′ end of the Il18r1 using an EcoRI fragment that spanned exon 8 to the 3′ end of the locus but did not observe any DNase hypersensitive sites in that region (Fig. 2D). We further examined whether acute stimulation with IL-12 would affect the DH pattern. Wild type and Stat4 -/- cells were activated for three days and incubated in the presence or absence of IL-12 for 24 hours. However, IL-12 stimulation did not significantly alter the DH pattern (Fig. 2E). The results of all DH analyses are summarized in Figure 2A.

Il18r1 histone modifications during Th differentiation

Since histone acetylation of the Il18r1 locus is Stat4-dependent, but DH patterns appeared to be independent of Stat4 activity, we next wanted to determine if the differences in chromatin structure among undifferentiated Th, Th1 and Th2 cells were also reflected by altered histone acetylation in that region. We used chromatin immunoprecipitation followed by quantitative PCR with primers spanning the TSS and DHI to assess histone acetylation and methylation of the H3K4 residue, modifications associated with active gene expression (22). While the levels of all modifications were significantly increased following Th1 development, changes between undifferentiated cells and Th2 cells were more modest, showing only slight decreases in acetylated H3 levels and H3K4 methylation (Fig. 3).

To determine if the DH pattern corresponded to conserved non-coding sequences (CNS), we overlayed the identified DH sites with a VISTA plot of Il18r1 highlighting CNS between mouse and human sequences. DHIII, DHIV and DHV corresponded to CNS A, B and C, respectively (Fig. 4A). DHI or DHII did not correspond to CNS, and CNS D and E did not correspond to observed DH sites. To determine if the CNS had functional impact on chromatin structure, we examined histone acetylation in these regions using chromatin immunoprecipitation from undifferentiated Th, Th1 and Th2 cells. As with the DHI region (Fig. 3), there were significant increases in histone acetylation in the CNS A-D regions (including DHIII-V) in Th1 cells compared to undifferentiated cells (Fig. 4B). Similar increases in H3K4 methylation were also observed (Fig. 4B). Modifications of CNS E were present at a lower level than those at other CNS regions. Following Th2 differentiation, there were generally lower levels of histone acetylation and H3K4 methylation compared to undifferentiated cells, though patterns differed slightly among the CNS regions, with acetyl-H4 differing less in the CNS C and D regions, and tri-methyl H3K4 differing less in the CNS B and C regions (Fig. 4B).

We then assessed the ability of cytokine stimulation to directly alter histone acetylation in the across the locus. CD4+ T cells were activated with anti-CD3 and anti-CD28 in the absence of any skewing cytokines for three days. Cells were then washed and incubated for 24 hours with IL-12, IL-4 or in the absence of any cytokine before chromatin immunoprecipitation. While IL-12 only had minor effects at the TSS and DHI sites, IL-12 induced acetylation of H3 and H4 at all CNS sites except CNS E (Fig. 4C). In contrast, IL-4 did not decrease acetylation of H3 or H4 at any of the CNS regions (Fig. 4C).

STAT protein regulation of Il18r1 expression and histone modification

While the expression of Il18r1 is vastly different in Th1 and Th2 cells, Stat4 is required for the induced expression in Th1 cells and Stat6 is required for the repression observed in Th2 cells (Fig. 5A and B). The intermediate level of unactivated Il18r1 expression in Stat4-/- Th1 and Stat6-/- Th2 is observed at the level of mRNA (Fig. 5B) and surface expression of IL-18Rα where these populations have an intermediate staining intensity (Fig. 5A).

To determine if STAT proteins have a role in establishing chromatin modifications in the CNS regions, we differentiated C57BL/6 WT and Stat4-/- Th1 cultures, and Balb/c WT and Stat6-/- Th2 cultures and examined the level of histone acetylation at each of the CNS regions. As expected, levels of acetyl-H3 and acetyl-H4 were decreased in Stat4-deficient Th1 cultures compared to WT Th1 cultures at CNS regions A-D (Fig. 5C). Acetyl-H3 and acetyl-H4 levels were very low in WT Th2 cells but were increased in the absence of Stat6 at CNS regions A-C (Fig. 5C). There was no significant difference between levels of histone acetylation in WT Th2 cells derived from C57BL/6 or Balb/c mice (data not shown).

We then tested whether there were differences in the levels of H3K4me2 and H3K4me3 at the Il18r1 CNS regions. Di-methyl-H3K4 and tri-methyl-H3K4 was detected at CNS regions A-D in wild type Th1 cultures (Fig. 5D). In contrast to the patterns of acetylated histones, H3K4 methylation was not decreased in the absence of Stat4, correlating with the diminished but detectable Il18r1 mRNA in Stat4-/- Th1 cells. Methyl-H3K4 levels were lower in Th2 cells but were increased in Stat6-/- Th2 cultures to levels observed in Stat4-/- Th1 cultures (Fig. 5D). Thus, H3K4 di- and tri-methylation is reduced in Th2 cultures, coincident with the repression of Il18r1 mRNA.

Transcription factor binding in the Il18r1 locus

In previous work, we identified that Stat4 binds to the Il18r1 promoter (7). To test whether factors that activate, or repress Il18r1 expression bind to specific regions of the locus, we performed chromatin immunoprecipitation for Stat4, Stat6 and GATA-3, and performed qPCR for the promoter and CNS regions. Stat4 was observed binding to the promoter region of Th1 cells, but had lower levels of binding at DHI and the CNS regions and minimal levels detected in Th2 cells (Fig. 6A). Conversely, GATA-3 binding was detected primarily in Th2 cells with maximal levels in the CNS A region (Fig. 6A).

To test whether these factors bound in response to cytokines during the differentiation process, CD4 T cells were activated for three days and left unstimulated or stimulated with IL-4 or IL-12 for 24 hours. Stat4 binding to the promoter following IL-12 stimulation was increased 150-fold (Fig. 6B). Stat4 was also detected at the CNS A and CNS B sites where binding was increased 80-200-fold with only modest binding at the DH1 site (Fig. 6B). In contrast to Stat4 binding which was 10-fold over control binding in unstimulated cells, Stat6 binding was increased less than five-fold over control at any of the regions in unstimulated or IL-4 stimulated cells (data not shown). GATA-3 binding was not detected in the promoter but was significantly enriched at the CNS A and CNS B regions although IL-4 had only modest effects on GATA-3 binding (Fig. 6B).

These results suggested that IL-4 signaling and the induction of GATA-3 would repress Il18r1 expression in Th1 cells. Transduction of GATA-3 into differentiating cells would divert them to a Th2 phenotype where IL-18Rα would be low and it would be difficult to discern direct effects of GATA-3 from indirect effects on differentiation of the cells. To avoid this problem and examine Il18r1 regulation in absence of altered differentiation, we first differentiated cells under Th1 conditions for one week before transduction with control-hCD4 or GATA-3-hCD4 expressing bicistronic retroviruses. After five days of culture, hCD4+ cells were analyzed for expression of IL-18Rα. Expression of GATA-3 reduced IL-18Rα on Th1 cells and Il18r1 mRNA levels (Fig. 6C and D). To confirm that cells transduced with GATA-3 did not convert to a Th2 phenotype, we also analyzed expression of CXCR3, a chemokine receptor expressed on Th1 cells. GATA-3 did not have an effect on CXCR3, demonstrating specificity in the effects on IL-18Rα expression (Fig. 6C).