The Human IL-13 Locus in Neonatal CD4+ T Cells Is Refractory to the Acquisition of a Repressive Chromatin Architecture*

The Th2 cytokine IL-13 is a major effector molecule in human allergic inflammation. Notably, IL-13 expression at birth correlates with subsequent susceptibility to atopic disease. In order to characterize the chromatin-based mechanisms that regulate IL-13 expression in human neonatal CD4+ T cells, we analyzed patterns of DNase I hypersensitivity and epigenetic modifications within the IL-13 locus in cord blood CD4+ T cells, naive or differentiated in vitro under Th1- or Th2-polarizing conditions. In naive CD4+ T cells, hypersensitivity associated with DNA hypomethylation was limited to the distal promoter. Unexpectedly, during both Th1 and Th2 differentiation, the locus was extensively remodeled, as revealed by the formation of numerous HS sites and decreased DNA methylation. Obvious differences in chromatin architecture were limited to the proximal promoter, where strong hypersensitivity, hypomethylation, and permissive histone modifications were found selectively in Th2 cells. In addition to revealing the locations of putative cis-regulatory elements that may be required to control IL-13 expression in neonatal CD4+ T cells, our results suggest that differential IL-13 expression may depend on the acquisition of a permissive chromatin architecture at the proximal promoter in Th2 cells rather than the formation of locus-wide repressive chromatin in Th1 cells.

The Th2 cytokine IL-13 is a major effector molecule in human allergic inflammation. Notably, IL-13 expression at birth correlates with subsequent susceptibility to atopic disease. In order to characterize the chromatin-based mechanisms that regulate IL-13 expression in human neonatal CD4 ؉ T cells, we analyzed patterns of DNase I hypersensitivity and epigenetic modifications within the IL-13 locus in cord blood CD4 ؉ T cells, naive or differentiated in vitro under Th1-or Th2-polarizing conditions. In naive CD4 ؉ T cells, hypersensitivity associated with DNA hypomethylation was limited to the distal promoter. Unexpectedly, during both Th1 and Th2 differentiation, the locus was extensively remodeled, as revealed by the formation of numerous HS sites and decreased DNA methylation. Obvious differences in chromatin architecture were limited to the proximal promoter, where strong hypersensitivity, hypomethylation, and permissive histone modifications were found selectively in Th2 cells. In addition to revealing the locations of putative cis-regulatory elements that may be required to control IL-13 expression in neonatal CD4 ؉ T cells, our results suggest that differential IL-13 expression may depend on the acquisition of a permissive chromatin architecture at the proximal promoter in Th2 cells rather than the formation of locus-wide repressive chromatin in Th1 cells.
The cytokine IL-13 has received considerable attention in recent years because of its critical role as an effector molecule of Th2-mediated disease. IL-13 appears to be necessary and sufficient to induce bronchial hyperresponsiveness, airway eosinophilia, epithelial cell damage, and goblet cell hyperplasia with mucus hyperproduction, key signatures of allergic inflammation in experimental asthma (1)(2)(3). Furthermore, IL-13 instigates chronic airway alterations through its ability to induce fibrosis, parenchymal and vascular remodeling, and accumulation of macrophages in the lung (4 -7). Mounting evidence from animal models also implicates IL-13 in the pathogenesis of other Th2-associated disorders, such as atopic dermatitis (8,9) and matrix metalloproteinase-and cathepsin-dependent emphysema (10). Of note, numerous studies in humans have documented increased expression of both IL-13 and its receptors at sites of allergic inflammation (11) and strong associations between genetic variation in IL-13 and increased risk of allergic disease (12).
Analysis of chromatin structure has provided considerable insights into the role played by epigenetic mechanisms in the regulation of gene expression throughout the Th2 cytokine cluster that includes IL-5, IL-13, and IL-4 (13). Although focused mostly on IL-4, studies using adult murine CD4 ϩ T cells showed that during Th2 differentiation, the IL- 13 locus acquires an open chromatin configuration characterized by increased DNase I hypersensitivity (14 -16), extended demethylation (17), and permissive histone modifications (18,19). In contrast, the chromatin within the IL-13 locus in Th1 cells was found to be in an inaccessible state similar to that found in naive CD4 ϩ T cells, with the exception of a single constitutive DNase I hypersensitive (HS) 2 site in the IL-13/IL-4 intergenic region. Information about IL-13 locus remodeling during differentiation of human adult T cells is comparatively limited (20,21) but essentially consistent with that obtained from mouse models (13).
We chose to analyze the chromatin-based mechanisms regulating human IL-13 gene expression in neonatal CD4 ϩ T cells. Indeed, immunological events in early life are essential determinants of Th2 allergic inflammation in adults (22,23), and robust correlations exist between IL-13 expression at birth or within the first year of life and subsequent susceptibility to allergic disease (24 -27).
In an initial effort to characterize the molecular mechanisms controlling IL-13 expression in the neonatal period, we analyzed the patterns of DNase I hypersensitivity and CpG methylation across the locus in human cord blood CD4 ϩ T cells, naive or differentiated in vitro under Th1-or Th2-polarizing conditions. We show herein that the IL-13 chromatin in neonatal CD4 ϩ T cells undergoes extensive remodeling during differentiation into a polarized T helper phenotype. Surprisingly, overall similar patterns of chromatin accessibility and methylation were observed in Th2 and Th1 populations, although IL-13 was not expressed by Th1 cells. Substantial differences in accessibility were limited to the proximal promoter, where a strong DNase I HS site and a permissive epigenetic state were generated selectively in Th2 cells. Our results suggest that the IL-13 locus in neonatal CD4 ϩ T cells is refractory to the acquisition of a repressive chromatin structure. Control of proximal promoter accessibility may therefore be a critical determinant of IL-13 expression.
Real Time Reverse Transcriptase-PCR-After lysis of the cell pellets, total RNA was extracted using TRIzol Reagent (Gibco) and reverse transcribed using Omniscript (Qiagen) and oligo(dT) primers (Gibco). Real time PCR was performed using predeveloped primer and probe sets for IL-13, IFN-␥, and glyceraldehyde-3-phosphate dehydrogenase and universal PCR master mix (Applied Biosystems) in a model 7900HT real time PCR machine (Applied Biosystems). Cytokine cDNA copy number was normalized to glyceraldehyde-3-phosphate dehydrogenase cDNA copy number.
DNase I HS Site Mapping-DNase I hypersensitivity was assessed using methods adapted from Elder et al. (28) and Burch and Weintraub (29). All procedures were performed on ice unless otherwise indicated. Freshly isolated naive or in vitro differentiated CD4 ϩ T cells were washed once in PBS and once in a 1:1 mix of PBS and RSB buffer (10 mM Tris, pH 7.4, 10 mM NaCl, 3 mM MgCl 2 ). Cells were then resuspended in 100% RSB buffer at 20 -40 ϫ 10 6 /ml. While slowly vortexing, a 10% solution of Nonidet P-40 was gradually added to the cell suspension until a final concentration of 0.23% was reached. The nuclei thus released were pelleted and washed once with RSB. DNA concentration was estimated (A 260 ), and nuclei were resuspended in RSB buffer containing Ca 2ϩ (100 M) at a DNA concentration of 0.3 mg/ml. Aliquots of nuclei were incubated with increasing concentrations of DNase I (5-120 units/mg DNA; Gibco) for 10 min at 37°C. An equal amount of nuclei lysis buffer (20 mM Tris, pH 8.0, 1% SDS, 0.25 M EDTA, 20 g/ml RNase) was added to stop the reaction. Following a 1-h incubation at 37°C, Proteinase K (Sigma) was added (100 ng/ml, final concentration), and samples were incubated at 50°C for 3 h with occasional mixing. Four serial phenol/chloroform extractions were then performed, followed by dialysis against two changes of TE buffer. DNA samples (15 g each) were digested overnight at room temperature with restriction enzymes (100 -200 units; New England Biolabs) and ethanol-precipitated. DNA was reconstituted in TE, electrophoresed in a 0.8 -1.2% agarose gel in TBE buffer, and transferred to nylon. Target DNA was visualized by indirect end labeling with a radiolabeled probe annealing to one end of the restriction fragment. Probe templates (200 -500 bp) were radiolabeled by random priming as described (30).
DNA Methylation Analysis-Genomic DNA was isolated from either freshly isolated or in vitro differentiated T cells as described for DNase I HS site mapping. Following digestion with EcoRI and ethanol precipitation, DNA (2 g) was resuspended in water, bisulfite-treated as per the manufacturer's instructions (EZ DNA Methylation Kit; Zymo Research), and resuspended in nuclease-free water (Ambion). Primers were designed using Primer3 software (31) following conversion of sequence obtained from GenBank TM (accession number AC004039). PCRs for each primer set were optimized for Mg 2ϩ concentration and annealing temperature. DNA template (50 ng) was amplified with primers complementary to bisulfiteconverted DNA sequence (see supplemental material) and Taq polymerase (Invitrogen) for 40 cycles. Amplified DNA was gelpurified using spin columns (Corning-Costar) and cloned using the TOPO TA cloning kit (Invitrogen). For each PCR amplification, 48 colonies were picked, transferred to 96-well plates, and lysed using the Colony Fast Screen kit (PCR Screen; Epicenter). To rule out any bias related to bisulfite conversion and PCR, the DNA samples analyzed in Fig. 7 were subjected to two bisulfite conversions, followed by two PCR amplifications and cloning events per conversion. Small aliquots (3.0 l of a 1:10 dilution) were removed from each lysate and PCR-amplified with the M13 forward and reverse primers (1.0 pmol) using 1.5 mM MgCl 2 , 1ϫ Taq buffer, and 0.05 units of platinum Taq at an annealing and extension temperature of 62.8°C for 40 cycles. PCR products were run on 1.5% agarose gels to assess insert size and yield. Products were then diluted 1:2 to 1:6 with water. Cycle sequencing reactions consisted of diluted PCR product (2 l), Big Dye 3.0 nucleotide mix (0.4 l; ABI) and diluted (1:20) 5ϫ sequencing buffer (1.8 l), M13 primers (7 pmol; 0.17 l), and water (5.6 l). Cycle sequencing products were purified using the Agencourt bead system and an FX robot (Biomeck). Samples were run on a 3730XL DNA Analyzer (ABI), using a 36-cm capillary system. Data were assessed for initial quality using the SQPR program (32) and assembled using CONSED software (33). Results are presented as percentage methylation at each CpG position: ((number of protected cytosines)/(number of unprotected ϩ number of protected cytosines) ϫ 100). Between 25 and 45 colonies (typically above 40) from each PCR amplification provided readable sequence. Supplemental Fig. 2 shows the CpG site-by-CpG site analysis of methylation at the whole IL-13 locus in Th2 cells and demonstrates the absence of PCR and clonal bias.
Chromatin Immunoprecipitation-The chromatin immunoprecipitation procedure was adapted from Litt et al. (34). Essentially, cells were fixed by adding 37% formaldehyde directly to the culture media and incubating for 10 min at room temperature. Fixation was stopped by adding one-tenth volume of 1.25 M glycine and incubating at room temperature for 5 min. Cells were collected, washed twice with ice-cold PBS, and resuspended at 20 -40 ϫ 10 6 cells/ml in SDS lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris (pH 8.0). Cells in a volume of 700 l in an ice bath were sonicated (one-eighth inch tip; Misonix) at 14 watts for six 10-s pulses over a 6-min period. Cellular debris was pelleted, and chromatin preparations were frozen in liquid nitrogen. Aliquots of chromatin (2 ϫ 10 6 cell equivalents) were diluted in chilled ChIP buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 17.7 mM Tris (8.1), 167 mM NaCl) to a volume of 1 ml, precleared with Protein G-agarose (50 l; Upstate Biotechnology, Inc.) for 2 h, and incubated overnight with 10 g of anti-acetylated H3 or H4 antibodies, anti-histone H3, or normal rabbit IgG (Upstate Biotechnology). Following a 2-h incubation with Protein G-agarose beads, the beads were washed five times with cold ChIP buffer. Antibody-chromatin complexes were eluted first with 250 l of ChIP buffer plus 3.0% SDS and then with 250 l of ChIP buffer plus 1.0% SDS. Fixation was reversed by incubating overnight at 65°C. DNA was purified by treating sequentially with RNase (New England Biolabs) and proteinase K (New England Biolabs). Following phenol/chloroform and a chloroform extraction, DNA was precipitated with 0.2 M NaCl 2 , 30 g of glycogen, and EtOH (70%). DNA yield was determined using Picogreen (Invitrogen). Enrichment for target templates was assessed using Sybr Green real time PCR (Applied Biosystems) with 2 ng of sample DNA and a 0.5 M concentration of each primer (see supplemental material) in a volume of 20 l. Reactions were run, and data were collected using a model 7900HT real time PCR machine (Applied Biosystems). Copy number was calculated using the following formula: 10 [(CtϪ40)/Ϫ3.33] Results were normalized to the copy number of isotype immunoprecipitated samples to correct for differences in primer efficiency and then normalized to a RAG2-negative control copy number within each sample.
Comparative Sequence Analysis-Alignments between the sequence of the human and mouse IL-13 locus were performed, and the extent of DNA sequence homology was computed using the World Wide Web-based program Genome Vista (available on the World Wide Web at pipeline.lbl.gov/cgi-bin/ GenomeVista). Regions with a length of at least 50 bp, which showed at least 75% sequence identity at each segment of the alignment between successive gaps, were identified as conserved noncoding sequences (CNS).

IL-13
Chromatin Is Accessible in Neonatal Naive CD4 ϩ T Cells-Nuclease HS sites are believed to reflect the DNA binding activity of sequence-specific trans-acting factors that induce destabilization or displacement of local nucleosomes (35). This chromatin-modifying activity can result in the exposure of additional binding sites, potentially increasing the transcriptional competence of a promoter or the activity of a distal regulatory element (36,37). The mapping of HS sites therefore identifies the locations of putative cis-regulatory elements and determines their state of accessibility to trans-acting factors at specific developmental stages.
As a first step toward characterizing IL-13 regulation in neonatal T cells, we mapped the locations of DNase I HS sites across the IL-13 locus in resting human naive (CD45RA ϩ CD45RO Ϫ ) CD4 ϩ T cells freshly isolated from umbilical cord blood (Fig. 1A). In a 24.1-kb KpnI restriction fragment spanning the IL-13 transcription unit and RAD50/ IL-13 intergenic region, two HS sites were detected within the IL-13 distal promoter (HS4 and HS5; Fig. 2A, NAIVE). HS4 is just proximal to a CpG island and a conserved CGRE element (18), whereas HS5 is located ϳ1.2 kb upstream of the ATG. In contrast, no HS sites were detected within the proximal promoter, transcription unit, or 6.5 kb of 3Ј-flanking chromatin (Fig. 2, A and B, NAIVE). The IL-13 locus chromatin is therefore largely inaccessible to DNase I in neonatal naive CD4 ϩ T cells, except for the distal promoter.
Covalent modifications of chromatin constituents can also have regulatory significance. Cytosine methylation at CpG dinucleotides can increase the repressive nature of chromatin by providing docking sites for methyl-binding proteins, which in turn recruit complexes with histone deacetylase and histone methyltransferase activity (38). In contrast, hypomethylated regions frequently correlate with an open or permissive chromatin conformation, typically thought to result from competitive inhibition of maintenance methylation activity due to transcription factor binding during S-phase of the cell cycle, a process called passive demethylation (39). We therefore examined the levels of CpG methylation within the IL-13 locus using genomic DNA purified from freshly isolated neonatal naive CD4 ϩ T cells. Following bisulfite treatment, overlapping amplicons spanning ϳ2.0 kb of the 5Ј-flanking region were cloned and sequenced. Many of the CpG dinucleotides located within this region were close to 100% methylated. In contrast, CpG dinucleotides, which co-localize with HS4 and HS5, and the proximal end of the CpG island, were found to be markedly hypomethylated (Fig. 3, Naive). The co-localization of nuclease hypersensitivity and hypomethylation at HS4 and HS5 provides strong independent support for the presence of occupied cisregulatory elements in neonatal naive CD4 ϩ T cells.
We also analyzed ϳ3.9 kb of 3Ј-flanking sequence and found two discrete regions of hypomethylation (Fig. 4, Naive). CpG dinucleotides located at the 3Ј end of amplicon 08 (positions ϩ3104, ϩ3156, ϩ3253, and ϩ3302 relative to the IL-13 ATG; Fig. 4, asterisks) were partially unmethylated, and most CpG dinucleotides within a 340-bp region in amplicons 21 and 11 (ϩ4775 to ϩ5113) were hypomethylated. Notably, in contrast to the hypomethylated sites that co-localize with HS4 and HS5, those located in the 3Ј-flanking sequence did not correlate with DNase I hypersensitivity, suggesting that a permissive epigenetic state may coexist with chromatin otherwise resistant to nuclease activity.
Near Equivalent States of IL-13 Locus Accessibility in Neonatal Th1 and Th2 Cells-According to the paradigm proposed for adult murine and human CD4 ϩ T cells, silencing of Th2 cytokine gene transcription in Th1 cells relies on the development of a repressive chromatin structure during differentiation and its maintenance in effector cells (14,15,40). Whether the same mechanism silences IL-13 transcription in human neonatal Th1 cells is not known. We therefore analyzed the patterns of DNase I hypersensitivity in neonatal naive CD4 ؉ T cells differentiated in vitro under Th1-and Th2-polarizing conditions. After 1-2 weeks of culture under Th1 conditions, significant levels of IFN-␥, but little or no IL-4 and IL-13, were detected ( Fig. 1, B and C, Th1). Conversely, after 2 weeks of culture under Th2 conditions, high IL-13 expression was observed, in the virtual absence of IFN-␥ (Fig. 1, B and C; Th2). IL-4 expression in Th2 cells was typically modest, consistent with the low propensity of neonatal CD4 ϩ T cells to express this cytokine (41). Differentiation therefore induced polarized Th1 and Th2 cytokine expression patterns in neonatal CD4 ϩ T cells.
An examination of nuclease accessibility in activated Th1 and Th2 cells revealed the persistent presence of HS4 and HS5 ( Fig. 2A), suggesting that these sites are constitutively occupied by transcription factors. In contrast, the proximal promoter was found to contain an HS site (HS6) considerably stronger in Th2 cells, a result consistent with preferential IL-13 promoter activity in these cells. A single, weak HS site (HS7) was induced in the first intron preferentially in Th2 cells. Immediately downstream of the 3Ј-untranslated region, a cluster of three novel HS sites was induced in both Th1 and Th2 cells (HS8 to -10; Fig. 2B). Remarkably, the position of HS8 was predicted by hypomethylation of a CpG site first observed in naive cells (Fig. 4, Naive), raising the possibility that trans-acting factors may reside at this location in naive cells in preparation for differentiation and/or activation (42). In addition, two closely spaced HS sites (HS11 and HS12; Fig. 2B) were detected in the distal 3Ј-flanking chromatin. Again, the pattern of CpG hypomethylation in naive cells predicted the location where HS11 formed. Further downstream, the CNS-1 enhancer, an element critical for high level Th2 cytokine expression (43,44), became hypersensitive in both Th1 and Th2 cells (HS13; Fig. 2B).
The differential expression of a subset of HS sites became more readily apparent when visualized at higher resolution. The most obvious difference was the preferential expression of HS6 in Th2 cells (Fig. 5A). HS9, which resolves into a doublet at this resolution (Fig. 5B), also appeared to be more intense in Th2 cells. Most of the CNS-1 enhancer element was accessible to nuclease (Fig. 5C), but obvious differences between Th1 and Th2 cells were not detectable. A Permissive Epigenetic State Exists at the Proximal Promoter in both Resting and Activated Th2 but not Th1 Cells-We also analyzed CpG methylation profiles in Th1 and Th2 cells, resting or activated with immobilized anti-CD3 for 24 h. Figs. 3 and 4 show that, overall, the patterns of CpG methylation detected throughout the IL-13 locus remained surprisingly stable in these cells. The hypomethylated sites in the distal 5Ј promoter were essentially invariant in all cell types (Fig. 3), supporting the conclusion HS4 and HS5 are constitutively occupied by nucleoprotein complexes in the neonatal CD4 ϩ T cell lineage. The hypomethylated site co-localizing with HS8 was present in all cell types, but the CpG located at ϩ3253 relative to the IL-13 ATG (Fig. 4, asterisk) was further demethylated in differentiated cells regardless of activation state, possibly reflecting changes linked to Th cell differentiation. The hypomethylated region co-localizing with HS11 expanded in the 3Ј direction in differentiated cells to include CpG dinucleotides co-localizing with HS12 (positions ϩ5113/ϩ5435). Further 3Ј, the CNS-1 enhancer, which contains a limited number of CpG dinucleotides in the distal half, became moderately demethylated in all differentiated cell types (HS13; Fig. 4).
In contrast, analysis of proximal promoter CpG methylation revealed substantial differences between CD4 ϩ T cells polarized under opposing conditions. Indeed, the CpG methylation levels at positions Ϫ67, Ϫ102, Ϫ192, and Ϫ280 were significantly reduced in Th2 cells compared with Th1 cells, both resting and activated (Fig. 6A). Immediately downstream of the transcription start site (positioned at Ϫ56), CpG dinucleotides located at Ϫ34, ϩ6, and ϩ21 were also hypomethylated in Th2 compared with Th1 cells, particularly upon activation. Moreover, anti-CD3-mediated cross-linking led to further hypomethylation of several proximal promoter CpG sites in Th2 cells but had negligible effects in Th1 cells (Fig. 6B). These data suggest that a permissive epigenetic state exists at the IL-13 proximal promoter selectively in Th2 cells.
To further support this conclusion, we examined whether histones at that location bore post-translational modifications typically associated with active genes. Fig. 6D shows that histone H4 acetylation levels at the IL-13 proximal promoter were considerably higher in Th2 cells. An opposite pattern was observed for the IFN-␥ promoter. These results suggest that covalent histone modifications that favor rapid nucleosome displacement and subsequent recruitment of transcription factors in response to activation signals are established at the proximal promoter in neonatal Th2 but not Th1 cells.

DISCUSSION
The Th2 cytokine IL-13 orchestrates multiple facets of human allergic inflammation through processes typically occurring in early life. Since a large amount of evidence supports a regulatory role for chromatin in controlling profiles of Th2 cytokine gene expression elicited by immune challenges (45), we analyzed the state of IL-13 DNase I hypersensitivity and CpG methylation during the in vitro differentiation of human

IL-13 Locus Accessibility in Neonatal CD4 ؉ T Cells
neonatal naive CD4 ϩ T cells under Th1-or Th2-polarizing conditions. Except for the distal 5Ј promoter region, which was found to be constitutively accessible and hypomethylated, the locus was extensively remodeled during T helper differentiation, as revealed by the formation of multiple HS sites and decreased levels of CpG methylation. Although IL-13 expression was restricted to the Th2 cell population, surprisingly, the locus acquired nearly equivalent patterns of hypersensitivity and CpG methylation during Th1 and Th2 differentiation throughout the locus but not at the proximal promoter, which became preferentially hypersensitive and hypomethylated in Th2 cells. These results suggest that regulation of IL-13 transcription in neonatal CD4 ϩ T helper cells may rely less on the formation of locus-wide repressive chromatin than on mechanisms controlling proximal promoter accessibility.
A Permissive Epigenetic State Exists at the IL-13 Locus in Neonatal Naive CD4 ϩ T Cells-The state of accessibility at the IL-13 locus in human neonatal naive CD4 ϩ T cells contrasts with that observed in the mouse locus. Examination of freshly isolated murine naive CD4 ϩ T cells revealed an IL-13 chromatin architecture devoid of HS sites except for a single constitu-tive HS site mapping within the IL-13/IL-4 intergenic region (14). The absence of HS sites was taken to imply the existence of a repressive chromatin architecture acting to suppress cytokine expression in naive CD4 ϩ T cells (45). However, human neonatal naive CD4 ϩ T cells express significant amounts of IL-13 in response to T cell receptor cross-linking (41), and the presence of HS4 and HS5 in these cells may represent a mechanism for rapid activation of IL-13 production in neonates.
Complex Relationships between Nuclease Hypersensitivity and DNA Methylation at the IL-13 Locus-Locus-wide high resolution mapping of cytosine methylation in neonatal naive CD4 ϩ T cells revealed two regions in the distal promoter that were constitutively hypersensitive and constitutively hypomethylated (HS4 and HS5) and two discrete hypomethylated sites in the 3Ј-flanking chromatin that lacked detectable DNase I hypersensitivity. The latter locations eventually acquired HS sites (HS8 and HS11) during differentiation. Pioneer transcription factors, which can bind to their cognate sites without disrupting local nucleosomes (42), could inhibit maintenance methylation (39) and create a pattern of localized hypomethylation. As differentiation proceeds, pioneer factors recruit chromatin-remodeling complexes to prepare the locus for gene activation, resulting in the generation of nuclease HS sites. The 3Ј-flanking chromatin in naive T cells could thus be maintained in a state poised for rapid remodeling by resident pioneer factors.
At other locations, Th differentiation induced demethylation of CpG dinucleotides co-localizing with HS8, HS12, and HS13, whereas Th2, but not Th1, differentiation induced demethylation at the proximal promoter (HS6). Only three HS sites were not found to correlate with hypomethylation: HS9, which appears to be preferentially expressed in Th2 cells when examined at high resolution; HS10, which mapped to a region lacking CpG dinucleotides; and HS14. Further analysis will be required to elucidate the basis for the apparent dissociation between the state of nuclease accessibility and DNA methylation at these HS sites.
In the Absence of a Locus-wide Repressive Chromatin Configuration, Control of Proximal Promoter Accessibility May Be Rate-limiting-The results of our analysis of the state of IL-13 chromatin in differentiated neonatal Th cells contrast with those obtained from adult mice (14 -19) and humans (21), where the Th1/Th2 paradigm was reflected in the chromatin architecture. In those studies, IL-13 expression in Th2 cells correlated with an open chromatin structure, whereas silencing in Th1 cells correlated with an inaccessible state across the locus. Rather, our data are reminiscent of those recently reported for the murine IL-10 gene, where differences in chromatin architecture between Th1 and Th2 cell clones were most apparent at the proximal promoter (46). Our results likewise suggest the IL-13 locus is maintained in an accessible state in neonatal Th cells differentiated under Th1 conditions and may be a reflection of the propensity of these cells to produce IL-13 (41).
Differential accessibility was limited to the proximal promoter in polarized neonatal Th cells, suggesting that control of IL-13 expression depends on regulating transcription factor access at this location. For example, activationinduced remodeling of the IL-2 and IL-12 promoters results in displacement of a local nucleosome, which renders additional essential transactivator sites accessible (47,48). Apart from the global repression conferred by a higher order chromatin structure, gene repression can also result from an active process. A formal model of gene repression mediated by trans-regulators emerged from studies examining the mechanisms limiting the activity of the even-skipped stripe 2 enhancer involved in specifying the anteroposterior axis during Drosophila neurogenesis (49). In this model, short range repression was mediated by the binding of repressor complexes close to important activator sites, either in distal enhancers or near the transcription start site. Whereas binding of a short range repressor within a distal enhancer acted locally, thus permitting other activators and enhancers located at a distance to function (50,51), binding of a repressor near the transcription start site had a dominant repressive effect, inhibiting the initiation of gene transcription by multiple distal positive elements (51,52). Whether similar processes inhibit the activity of critical Th2 cytokine gene promoter trans-activators, such as GATA3 (53) and NFAT (54), needs to be determined through functional dissection of   Fig. 2 were digested with combinations of restriction enzymes to limit the size of the region analyzed. A, a 4.8-kb EcoRI/BamHI restriction fragment within the 5Ј-flanking sequence and transcription unit and encompassing the positions of HS4 to HS7 was visualized with probe C. B, a 2.1-kb EcoRI/BamHI restriction fragment encompassing 3Ј-proximal flanking sequence and the positions of HS8 to HS10 was visualized with probe D. C, a 2.2-kb SpeI/PvuII restriction fragment located in the distal 3Ј-flanking sequence and encompassing CNS-1 and part of the HS11 and HS12 position was visualized with probe E. The Th2 blot was exposed longer than the Th1 blot. D, schematic representation of HS sites mapped in the IL-13 locus (downward arrows). Restriction sites were EcoRI (E), KpnI (K), BamHI (B), PvuII (P), and SpeI (S) (PvuII and SpeI sites are not shown exhaustively). The positions of probe templates are marked by letters and short horizontal arrows below the diagram.
both distal elements and the proximal promoter. Much of the empirical data to date has focused on the role of distal elements in controlling gene activity. In mice, the IL-13/IL-4 intergenic constitutive HS site, HS3, has recently been implicated in mediating the spread of a repressive histone modification across the locus during Th1 differentiation (55). Whether the syntenic region (HS11) in human CD4 ϩ T cells likewise acts to repress IL-13 expression is unknown, but our data provided no evidence of a barrier to nucleases at HS11 or most of the locus. High level Th2 cytokine expression in murine T cells requires the CNS-1 enhancer (43,44), but we failed to find examples in the literature suggesting that enhancers act to modify proximal promoter DNA methylation in a dominant fashion. Instead, an elegant study revealed that promoter and enhancer DNA methylation states can be regulated independently (56). CNS-1 may therefore be required for high level cytokine expression but not for inducing a stable permissive epigenetic state at the proximal promoter. Another class of distal elements, locus control regions (LCRs), act to inhibit position effects at sites of transgene integration. The recently characterized Th2 cytokine LCR is located within the 3Ј end of the RAD50 gene, within a 25-kb region positioned 15-20 kb upstream of the murine IL-13 promoter (57,58). Three of the LCR HS sites were required for copy number-dependent expression of Th2 cytokines in transgenic mice (59). One of these HS sites (RHS7) was required for Th2 cytokine expression and long range physical interactions between Th2 cytokine gene promoters and the LCR (60). Th2-specific interactions between the IL-13 promoter and LCR HS sites were limited to RHS4, whereas RHS7 maintained physical interactions with the IL-13 promoter in naive CD4 ϩ T cells and Th1 cells (58). Stable long range interaction with the LCR could help to maintain a permissive epigenetic state at the proximal promoter. However, a more direct role of proximal promoter elements cannot be excluded, because the chromatin opening activity of the  4). Brackets with asterisks mark CpG positions at which statistically significant differences in methylation were found between resting Th1 and Th2 cells or stimulated Th1 and Th2 cells (Mann-Whitney U Test, one-tailed, p Ͻ 0.05). The percentage of methylation is indicated by the scale to the left, and the position of CpG dinucleotides relative to the IL-13 ATG is indicated at the bottom of each graph. D, analysis of histone acetylation levels at the IL-13 and IFN-␥ promoters in neonatal Th1 and Th2 cells. Results shown are from two independent chromatin immunoprecipitations from each of two donors. In vitro differentiated neonatal Th1 and Th2 cells were fixed with formaldehyde after 4 days of rest, sonicated, and immunoprecipitated with anti-acetylated histone H3, anti-acetylated histone H4, or control antibodies. Enrichment for target template was assessed using Sybrா Green-mediated real time PCR. Values represent relative copy number after correcting for primer efficiency and normalizing to the negative control RAG2. FIGURE 7. Comparison between accessibility, sequence conservation, and DNA methylation profiles across the IL-13 locus. Top, schematic representation of the IL-13 locus, indicating the locations of constitutive (black ovals), major Th2-specific (gray ovals), and inducible (white ovals) HS sites. The right pointing arrow indicates the translation start site. Middle, sequence comparison between the human and mouse IL-13 locus, generated by VISTA analysis. Dark blue, protein coding sequence; light blue, the 3Ј-untranslated region. Regions with a length of at least 50 bp, which have at least 75% sequence identity at each segment of the alignment between successive gaps, were identified as CNS and are shown in red. Bottom, profiles of DNA methylation detected in neonatal CD4 ϩ Th2 cell populations.

IL-13 Locus Accessibility in Neonatal CD4 ؉ T Cells
␤-globin LCR was compromised when transgene promoters were deleted (61), suggesting that LCRs, enhancers, and promoters work in concert to control accessibility.
Evidence for a Phylogenetically Divergent IL-13 Cis-regulatory Architecture-The tools of comparative genomics and bioinformatics have provided insights into the IL-4/IL-13 regulatory architecture. In several cases, the results of sequence comparisons have been used successfully to predict the locations of functional cis-regulatory elements and guide their genetic manipulation (40,43,62). Of the 11 HS sites mapped within the IL-13 locus, five co-localize with peaks of human/ mouse sequence conservation (Fig. 7): HS5, which may be homologous to murine HSI (14); HS6, which maps within the proximal promoter (53,54,63); HS7, which may be homologous to the intronic murine mast cell-specific HS site (64); HS8, which has no known murine counterpart; and HS13, which corresponds to the CNS-1 enhancer. The combination of phylogenetic sequence comparisons, analysis of the native chromatin structure in primary CD4 ϩ T cells, and functional studies provides strong independent evidence that HS5, HS6, and HS13 represent conserved cis-regulatory elements controlling IL-13 expression in neonatal CD4 ϩ T cells. The role of HS7 and HS8 remains to be determined.
Among the other HS sites, only HS11 may have a murine counterpart, the constitutive HS3 (14). HS4, HS9, HS10, HS12, and HS14 do not co-localize with peaks of strong sequence conservation and may therefore represent evolutionarily more recent, species-specific cis-regulatory elements. Our results therefore suggest that additional elements not identifiable through phylogenetic sequence comparisons may be required for proper regulation of IL-13 expression in human neonatal CD4 ϩ T cells.