Global Chromatin State Analysis Reveals Lineage-Specific Enhancers during the Initiation of Human T helper 1 and T helper 2 Cell Polarization.
ABSTRACT Naive CD4(+) T cells can differentiate into specific helper and regulatory T cell lineages in order to combat infection and disease. The correct response to cytokines and a controlled balance of these populations is critical for the immune system and the avoidance of autoimmune disorders. To investigate how early cell-fate commitment is regulated, we generated the first human genome-wide maps of histone modifications that reveal enhancer elements after 72 hr of in vitro polarization toward T helper 1 (Th1) and T helper 2 (Th2) cell lineages. Our analysis indicated that even at this very early time point, cell-specific gene regulation and enhancers were at work directing lineage commitment. Further examination of lineage-specific enhancers identified transcription factors (TFs) with known and unknown T cell roles as putative drivers of lineage-specific gene expression. Lastly, an integrative analysis of immunopathogenic-associated SNPs suggests a role for distal regulatory elements in disease etiology.
- SourceAvailable from: Annette E Neele[Show abstract] [Hide abstract]
ABSTRACT: The first functions of macrophages to be identified by Metchnikoff were phagocytosis and microbial killing. Although these are important features, macrophages are functionally very complex and involved in virtually all aspects of life, from immunity and host defense, to homeostasis, tissue repair and development. To accommodate for this, macrophages adopt a plethora of polarization states. Understanding their transcriptional regulation and phenotypic heterogeneity is vital because macrophages are critical in many diseases and have emerged as attractive targets for therapy. Here, we review how epigenetic mechanisms control macrophage polarization.Current opinion in lipidology. 10/2014; 25(5):367-373.
- [Show abstract] [Hide abstract]
ABSTRACT: Chromatin remodeler complexes exhibit the ability to alter nucleosome composition and positions, with seemingly divergent roles in the regulation of chromatin architecture and gene expression. The outcome is directed by subunit variation and interactions with accessory factors. Recent studies have revealed that subunits of chromatin remodelers display an unexpectedly high mutation rate and/or are inactivated in a number of cancers. Consequently, a repertoire of epigenetic processes are likely to be affected, including interactions with histone modifying factors, as well as the ability to precisely modulate nucleosome positions, DNA methylation patterns and potentially, higher-order genome structure. However, the true significance of chromatin remodeler genetic aberrations in promoting a cascade of epigenetic changes, particularly during initiation and progression of cancer, remains largely unknown.Epigenomics 10/2014; 6(4):397-414. · 5.22 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Although genome-wide association studies (GWAS) have identified hundreds of variants associated with a risk for autoimmune and immune-related disorders (AID), our understanding of the disease mechanisms is still limited. In particular, more than 90% of the risk variants lie in non-coding regions, and almost 10% of these map to long non-coding RNA transcripts (lncRNAs). lncRNAs are known to show more cell-type specificity than protein-coding genes. We aimed to characterize lncRNAs and protein-coding genes located in loci associated with nine AIDs which have been well-defined by Immunochip analysis and by transcriptome analysis across seven populations of peripheral blood leukocytes (granulocytes, monocytes, natural killer (NK) cells, B cells, memory T cells, naive CD4(+) and naive CD8(+) T cells) and four populations of cord blood-derived T-helper cells (precursor, primary, and polarized (Th1, Th2) T-helper cells). We show that lncRNAs mapping to loci shared between AID are significantly enriched in immune cell types compared to lncRNAs from the whole genome (α <0.005). We were not able to prioritize single cell types relevant for specific diseases, but we observed five different cell types enriched (α <0.005) in five AID (NK cells for inflammatory bowel disease, juvenile idiopathic arthritis, primary biliary cirrhosis, and psoriasis; memory T and CD8(+) T cells in juvenile idiopathic arthritis, primary biliary cirrhosis, psoriasis, and rheumatoid arthritis; Th0 and Th2 cells for inflammatory bowel disease, juvenile idiopathic arthritis, primary biliary cirrhosis, psoriasis, and rheumatoid arthritis). Furthermore, we show that co-expression analyses of lncRNAs and protein-coding genes can predict the signaling pathways in which these AID-associated lncRNAs are involved. The observed enrichment of lncRNA transcripts in AID loci implies lncRNAs play an important role in AID etiology and suggests that lncRNA genes should be studied in more detail to interpret GWAS findings correctly. The co-expression results strongly support a model in which the lncRNA and protein-coding genes function together in the same pathways.Genome Medicine 10/2014; 6(10):88. · 4.94 Impact Factor
Global Chromatin State Analysis Reveals
Lineage-Specific Enhancers during the Initiation
of Human T helper 1 and T helper 2 Cell Polarization
R. David Hawkins,1,3,7,* Antti Larjo,2,5,7Subhash K. Tripathi,3,4,7Ulrich Wagner,6Ying Luu,6Tapio Lo ¨nnberg,3
Sunil K. Raghav,3Leonard K. Lee,6Riikka Lund,3Bing Ren,6Harri La ¨hdesma ¨ki,3,5,* and Riitta Lahesmaa3,*
1Department of Medicine, Division of Medical Genetics and Department of Genome Sciences, University of Washington School of Medicine,
Seattle, WA 98195, USA
2Department of Signal Processing, Tampere University of Technology, 33101 Tampere, Finland
3Turku Centre for Biotechnology, University of Turku and A˚bo Akademi University, 20520 Turku, Finland
4National Doctoral Programme in Informational and Structural Biology, 20520 Turku, Finland
5Department of Information and Computer Science, Aalto University, FI-00076 Aalto, Finland
6Ludwig Institute for Cancer Research and Department of Cellular and Molecular Medicine, Moores Cancer Center, and Institute of Genomic
Medicine, University of California, San Diego, La Jolla, California 92093, USA
7These authors contributed equally to this work
*Correspondence: email@example.com (R.D.H.), firstname.lastname@example.org (H.L.), email@example.com (R.L.)
Naive CD4+T cells can differentiate into specific
helper and regulatory T cell lineages in order to
combat infection and disease. The correct response
to cytokines and a controlled balance of these popu-
lations is critical for the immune system and the
avoidance of autoimmune disorders. To investigate
how early cell-fate commitment is regulated, we
generated the first human genome-wide maps of his-
tone modifications that reveal enhancer elements
after 72 hr of in vitro polarization toward T helper 1
(Th1) and T helper 2 (Th2) cell lineages. Our analysis
indicated that even at this very early time point,
cell-specific gene regulation and enhancers were at
work directing lineage commitment. Further exami-
nation of lineage-specific enhancers identified tran-
scription factors (TFs) with known and unknown
T cell roles as putative drivers of lineage-specific
gene expression. Lastly, an integrative analysis of
immunopathogenic-associated SNPs suggests a
Lymphocytes play a pivotal role in the regulation of an immune
response. Upon immunogenic signals from antigen-presenting
cells induced by various pathogenic agents, naive CD4+T cells
can differentiate into functionally distinct subsets including T
helper 1 (Th1), Th2, Th17, and regulatory T (Treg) cells (Murphy
and Stockinger, 2010; Rautajoki et al., 2008; Weaver et al.,
2007). Each subtype is marked by specific cytokine secretion
patterns (Zhou et al., 2009). These effector and regulatory
CD4+T cell lineages defend the host from various infections,
while inappropriate activation and differentiation lead to patho-
genesis of inflammatory and autoimmune diseases (Hirota
et al., 2011; Reiner et al., 2007).
Molecular mechanisms leading to the polarization of CD4+
subsets have been made clearer through studies that defined
unique signaling molecules and transcription factors (TFs) for
each lineage. Interleukin-12 (IL-12) activates the signal trans-
ducer and activator of transcription 4 (STAT4) and initiates the
differentiation of Th1 cells that secrete the signature cytokine
interferon-g (IFN-g) and express the key transcriptional regulator
TBX21 (T-bet) (Afkarian et al., 2002; Kaplan et al., 1996b; Schulz
et al., 2009). IL-4 in turn initiates Th2 cell differentiation by acti-
vating STAT6 and inducing the key TF GATA-binding protein 3
(GATA-3) and produce cytokines IL-4 and IL-13 (Elo et al.,
tant for host defense against intracellular and extracellular path-
ogens, inappropriate execution of Th1 and Th2 cell responses
can lead to pathogenesis of autoimmune and inflammatory dis-
eases (Rautajoki et al., 2008).
Cellular specification requires networks of TFs and epigenetic
mechanisms to mediate changes in gene expression to deter-
mine cell fate (Rothenberg, 2007; Zhang et al., 2012). Previous
studies have demonstrated that lineage specification in Th1
and Th2 cells is represented by the epigenetic state of signature
cytokines loci: The Ifng locus in Th1 cells (Hatton et al., 2006;
Schoenborn etal.,2007)and theIl4locus inTh2cellsaremarked
by epigenetic modifications (Ansel et al., 2006). Epigenetic
changes bring cellular specificity, as well as plasticity. Th1 cell-
specific gene loci Ifng and Il18r1 are associated with the histone
activating mark H3K4me3 in Th1 cells, whereas in Th2 cells
these loci are marked with the repressive H3K27me3 mark
(Wei et al., 2009). Opposing modifications, H3K4me3 and
H3K27me3—termed ‘‘bivalent domains,’’ are colocalized at
numerous promoter regions in T cells (Roh et al., 2006; Wei
et al., 2009). Surprisingly, this includes genes such as Gata3
and Tbx21, suggesting that lineages may retain some degree
of plasticity even after polarization.
specific expression (Bulger and Groudine, 2011; Ong and
Immunity 38, 1271–1284, June 27, 2013 ª2013 Elsevier Inc. 1271
(legend on next page)
Enhancers for Early Th1 and Th2 Polarization
1272 Immunity 38, 1271–1284, June 27, 2013 ª2013 Elsevier Inc.
Corces, 2011). Histone modification patterns provide a distinct
signature for global mapping of enhancer elements (Heintzman
et al., 2009; Heintzman et al., 2007). The use of H3K4me1 has
been used to identify enhancers in human, mouse, zebrafish,
and fly genomes in a predominantly cell-specific manner (Aday
et al., 2011; Creyghton et al., 2010; Ernst et al., 2011; Hawkins
sias et al., 2011; Shen et al., 2012). Given the role of enhancers in
driving cell-fate commitment and gene-expression responses to
various stimuli, it is imperative to understand how chromatin
structure defines enhancers that will contribute to induction
and plasticity of T helper cell differentiation to distinct subsets
as well as their role in autoimmune diseases.
In this study, we aimed to characterize the lineage-specific
tive to activated CD4+T cells, by using naive CD4+cells from
state maps from ChIP-seq (chromatin immunoprecipitation
coupled to massively parallel DNA sequencing) by using anti-
bodies recognizing histone modifications that discriminate distal
elements from proximal promoters (H3K4me1, H3K27ac, and
H3K4me3). We focused on 72 hr after polarization in an attempt
to identify early regulators (enhancers and predicted binders) for
cule sequencing, and thousands of lineage-specific enhancers
often correlate with the gene expression changes. Lastly, under
the principle that improper cell-fate specification can lead to
immunopathogenesis, we found that these lineage-specific
enhancers overlap a great number of SNPs from genome-wide
association studies (GWAS) for various autoimmune disorders,
including type 1 diabetes, rheumatoid arthritis, Crohn’s disease,
and asthma. Several SNPs altered TF binding-site motifs, and a
subset of such SNPs within these predicted sites influenced
TF binding. This provides insight into how SNPs located at
tion and disease pathogenesis and is a valuable basis for addi-
tional investigation on the role of TF binding in human disease.
Global Mapping of Early Enhancers
We identified enhancer elements at an early stage of T cell differ-
entiation. We generated global histone modification maps for
H3K4me1 and H3K4me3. These modifications are known to
distinguish enhancer and promoter elements, respectively
(Heintzman et al., 2007). We compared chromatin maps at
72 hr following activation of naive human CD4+T cells (abbrevi-
ated as Th0 for succinctness) with those 72 hr after polarization
toward Th1 and Th2 cell lineages (Figure 1A). At 72 hr, Th1 cells
were positive for TBX21 protein expression, whereas Th2 cells
were positive for GATA-3 (see Figure S1A available online). At
day 7, the majority of cells expressed key lineage markers as
determined by flow cytometry analysis (Figures S1B and S1C),
indicating the cells polarized to Th1 and Th2 cell-specific line-
ages under these culture conditions.
On the basis of the distinct localization of H3K4me1 relative to
H3K4me3 (Heintzman et al., 2007; see Supplemental Experi-
mental Procedures), we determined 16,507 and 13,466 putative
enhancers for Th1 and Th2 polarized cells, respectively, and
16,552 in the control activated cells (Th0) (corrected p % 0.01;
Table S1A). As expected at this early time point of differentiation,
the chromatin profiles were largely similar between Th1 and Th2
cells and control cells (Figure 1B). Enhancers unique to cells
polarized toward Th1 or Th2 cells could also be determined (Fig-
ures 1C and 1D). To identify lineage-specific enhancers, we
implemented a recent analytical method that searches for sam-
ple-specific nucleosome-free regions (NFR) (He et al., 2010). By
using the histone modification signals, the method first predicts
the nucleosome positions in the samples under consideration,
identifies correctly spaced nucleosome pairs, and then detects
nucleosome depletion in either of the samples. This method
was employed to determine whether peaks that appear similar
in two or more lineages actually represented lineage-specific
enhancer locations due to nucleosome repositioning in active
enhancers at the site of TF binding (Figures 1E and 1F; Figure 2;
Figure S2; Table S1B). By using a stringent selection, we identi-
fied enhancers as lineage-specific if the peaks were called in our
original analysis and contained a NFR relative to both the other
lineage and control cells. That is, to be Th2 cell-specific, the
NFR must be present only in Th2 cell relative to both Th0 and
Th1 cells, indicating a more open chromatin structure in only
Th2 cells. We identified the following number of lineage-specific
enhancers: 2,144 (Th1 cells) and 2,654 (Th2 cells) (Table S2A).
Polarized cells showed a substantial increase in the number of
uniqueenhancers comparedtothe1,636foundinactivated cells
(Th0 cells) (Table S2A).
To understand how lineage-specific enhancers might drive
specific gene expression, we generated single-molecule digital
gene expression data by using the Heliscope platform to gain
an accurate tag count of expressed genes in each subtype.
Genes differentially expressed between Th1 and Th0 cells or
Th2 and Th0 cells were identified by using a stringent corrected
and S5). We identified 121 genes upregulated in Th1 cells and
292 in the Th2 cell culture condition, whereas 116 were
Figure 1. Genome-wide Mapping and Identification of Lineage-Specific Enhancer Elements
(A) Illustration of experimental design for activating the cells (Th0) and for polarizing cells to Th1 and Th2 cell subsets.
(B) Global clustering of all predicted enhancers in Th0, Th1, or Th2 cells based on H3K4me1. H3K4me3 (promoter modification) is shown as a control.
(C and D) ABHD6 and IFNG loci showing Th2 and Th1 cell-specific enhancer peaks identified through H3K4me1 predictions.
(E and F) Example of lineage-specific nucleosome-free regions (NFR) shown in high-density plots for (E) a distal Th1 cell-specific enhancer predicted ?9 kb
downstream of IL-10 and (F) Th2 cell-specific enhancer in an intron of BCAR3. The lineage-specific enhancer is determined by nucleosome spacing and
overlapping nucleosome in the other two cell types (dashed rectangle).
(G) Differential gene expression is shown for Th1 and Th2 cells relative to Th0 cells (total genes: blue + gray). Lineage-specific expression relative to Th0 cells and
the other lineage (Th1 or Th2 cells) is shown in red for Th1 cells and blue bars for Th2 cells.
(H) Lineage-specific differential expression relative to Thp cells to accommodate a three-way comparison.
Enhancers for Early Th1 and Th2 Polarization
Immunity 38, 1271–1284, June 27, 2013 ª2013 Elsevier Inc. 1273
upregulated in the Th0 cell culture condition, for a total of 529
genes with highly uniqueexpression patternsat 72 hr(Figure 1G;
Table S4A; see Table S5 for all genes). Additionally, all subtypes
showed differential expression relative to naive T helper precur-
sor cells (Thp) (Figure 1H). Consistent with our previous results, a
number of genes become differentially expressed during the
early stage of human Th cell differentiation (Elo et al., 2010;
Lund et al., 2005; Lund et al., 2007). These genes are likely to
be very important for the cell to gain the desired phenotype.
Insummary,thesechromatin-based enhancermaps providea
view of lineage-specific gene regulation during early human
T cell specification. These unique enhancers and the bound
TFs may be critical for driving gene expression essential for
each lineage commitment.
Continuum of Enhancer States
A substantial fraction of mammalian enhancers are simulta-
neously marked by H3K27ac (Heintzman et al., 2009), and acet-
ylation islands have enhancer activity (Roh et al., 2007). This is
likely the result of histone acetyltransferases (HATs) that also
act as transcriptional coactivators and bind enhancers, such
as p300 (Heintzman et al., 2009; Heintzman et al., 2007; Shen
Figure 2. Continuum of Enhancer Chro-
(A) Clustering of Th1 cell-specific H3K4me1-
marked enhancers. H3K27ac reveals active en-
hancers versus poised enhancers that lack
H3K27ac at 72 hr. DNase hypersensitivity data
(DHS) reflecting the active (hypersensitive sites)
and nonactive sites in the cells polarized to Th1 or
Th2 direction for 7 days.
(B) Same as in (A), except for Th2 cell-specific
(C and D) Distribution of enhancer states in Th1
and Th2 cell lineages.
et al., 2012; Visel et al., 2009). Yet, a sub-
set of enhancers lacks this modification.
Enhancers marked solely by H3K4me1
‘‘poised enhancers’’ as a result of the
absence of the active chromatin modifi-
cation (Creyghton et al., 2010; Hawkins
et al., 2011; Rada-Iglesias et al., 2011).
This also holds true for the lineage-spe-
cific enhancers identified in this study
(Figure 2). The abundance of poised en-
hancers likely reflects the fact that the
and beginning to commit to distinct
To address the fate of the lineage-spe-
cific enhancers for Th1 and Th2 cells, we
compared our enhancer chromatin maps
DNase hypersensitivity (DHS) data gener-
ated by the ENCODE consortium for Th1
and Th2 cells after 7–10 days of polariza-
tion (ENCODE Project Consortium, 2011;
Maurano et al., 2012). Because DHS reflect an open chromatin
structure, it is indicative of TFs bound to regulatory elements
and presumed active. At this later time point for which DHS
maps were generated, our data suggest the polarized cells are
largely committed based on marker expression (Figures S1B
and S1C). For Th1 and Th2 cell-specific enhancers, about 30%
were active based on histone modifications at 72 hr (Figure 2).
In Th1 cells, 76% of those remained active based on 7–10 day
DHS data, whereas 99% remained active in Th2 cells. Those
that were no longer hypersensitive are likely enhancers used to
drive early lineage specification, but not needed to maintain
the cell-fate commitment. Of the poised enhancers, 21% in
Th1 and 78% in Th2 cells became activated based on DHS
data once the cells are committed to their respective lineages
primed to regulate the future cell-fate commitment.
STAT6 and Putative Enhancer Binders
Enhancer functionality is driven by the binding of TFs, which in
turn facilitates looping to the target gene promoter and cell-spe-
cific gene upregulation. We analyzed enhancer sequences for
known TFBS motifs to predict sequence-specific TFs as
Enhancers for Early Th1 and Th2 Polarization
1274 Immunity 38, 1271–1284, June 27, 2013 ª2013 Elsevier Inc.
potential regulators of lineage-specific gene expression during
T cell polarization. Binding of TFs to enhancers were predicted
with ProbTF (La ¨hdesma ¨ki et al., 2008) by using position-specific
frequency matrices (PSFM) from TRANSFAC (Matys et al., 2006)
and combined with an empirical null model to choose hits above
the background (p < 0.01). Motifs were filtered based on gene
expression to ensure the TFs were expressed in each T cell line-
age. Examination of the identified motifs revealed both new and
known Th cell regulators (Table S2A, bound TFs columns).
Lineage-specific enhancers were enriched for known T cell TF
motifs relative to random genomic background (Table S3;
Experimental Procedures). For Th1 cell-specific enhancers, mo-
tifs included those for key factors STAT1, STAT4, ATF3, and
JUN (Afkarian et al., 2002; Thieu et al., 2008; Filen et al.,
2010). Th2 cell-specific enhancers included motifs for STAT6,
PPARG, BACH, GFI1, NFIL3 and GATA3, which are all upregu-
lated during early Th2 cell differentiation and regulate distinct
genes either independently or in a coordinated fashion. For
example, in the first intron of the GAB2 gene upregulated in
Th2 cells, we identified a Th2 cell-specific enhancer that
harbored a STAT6 motif (Figure 3A). STAT6 is a key regulator
of the Th2 cell lineage by mediating the IL-4 signal (Kaplan
et al., 1996a). Interestingly, GAB2, an adaptor protein, activates
PI3K and Akt, which subsequently regulates IL-4 production
(Frossi et al., 2007). This may provide a key part of the IL-4-
STAT6 regulatory feedback loop. The combination of chro-
matin-based enhancer maps and motif analysis filtered for TFs
expressed in these cells reveals how key TFs are likely to utilize
distal regulatory elements to drive lineage specification. In
another study (Aijo ¨ et al., 2012), we identified only a limited num-
ber of enriched motifs for TF binding sites in the promoters of
genes differentially expressed during the early Th cell differenti-
ation, suggesting substantial contribution of enhancer-driven
gene regulation. We also identified motifs in lineage-specific en-
hancers that corresponded to expressed TFs with unknown
roles in Th1 and Th2 cell differentiation for further studies (Fig-
ures 3B–3D; Table S3).
To validate a subset of the enhancers and motif predictions,
we generated ChIP data for at STAT6 motifs within Th2 cell-
associated enhancers. Over 70% of STAT6 binding sites are
enriched over introns and intergenic regions of genome (Elo
et al., 2010). Our motif analysis showed STAT6 motif to be
frequently enriched over Th2 cell-specific enhancers (Figure 3D;
TableS3).We generated ChIP-qPCR
H3K4me1, and H3K27ac at 4 hr and 72 hr (Figure 4) and vali-
dated six Th2 cell-specific enhancers harboring predicted
STAT6 binding sites. These enhancers were located nearby
genes of known and unexplored function such as RUNX1,
FOXP1, GAB2, IL10RA, SETBP1, and ABHD6, which are upre-
gulated in a Th2 cell-specific manner. We found that H3K4me1
and H3K27ac were enriched at these enhancers in Th2 directed
cells in contrast to naive (Thp) and activated T cells (Th0), which
suggest that these genes are regulated by STAT6 in Th2 cells by
an enhancer-specific manner. STAT6 and H3K4me1 were
largely consistent between 4 and 72 hr, whereas H3K27ac
increased at 72 hr relative to 4 hr. The acquisition of H3K27ac
is likely indicative of later transcriptional events beyond the
initial marking of the enhancer elements (compare Figures 4B
and 4C to Figure 4D).
Regulatory SNPs: Overlap of Enhancers with Disease
Genome-wide association studies (GWASs) have produced
large numbers of disease associated single-nucleotide polymor-
phisms (SNPs). However, many of these studies have failed to
identify causative mutations. This may be, in part, due to many
associated SNPs lying outside of gene coding regions. A recent
assessment of GWASs illustrated that 45% of disease or trait
associated SNPs fell in introns, whereas 43% lie in intergenic re-
gions (Hindorff et al., 2009). Regulatory SNPs (rSNPs) have been
implicated in the altered expression of genes that constitute
expression quantitative traits (eQTLs: for review, see Cheung
and Spielman, 2009). Global methods for determining chromatin
states show that associated SNPs can overlap regulatory re-
gions (Ernst et al., 2011; Gaulton et al., 2010; Maurano et al.,
2012). To determine whether disease-associated SNPs are
potentially rSNPs, we analyzed the vicinities of associated
SNPs from the NHGRI GWAS catalog (Hindorff et al., 2009) for
overlap with our global T helper cell enhancer predictions at
72 hr, including lineage-specific enhancers (see Experimental
Procedures). Several autoimmune disease associated SNP
sets were enriched among all the 460 disease categories (Table
S7), and we decided to conduct a detailed integrated analysis of
0.01 for the SNP associations) from asthma (Moffatt et al., 2010),
Crohn’s disease (Duerr et al., 2006; Rioux et al., 2007), multiple
sclerosis (MS) (Hafler et al., 2007), psoriasis (Cargill et al.,
2007; Helms et al., 2003; Nair et al., 2006), rheumatoid arthritis
(RA) (Stahl et al., 2010), type 1 diabetes (T1D) (Barrett et al.,
2009), and ulcerative colitis (UC) (Anderson et al., 2011).
enhancers detected in all three T cell culture conditions. Signifi-
cant (p < 0.001) overlap was found for most of the aforemen-
tioned autoimmune diseases. In total, we determined that
1,281 SNPs overlap our global enhancer maps (Table 1). RA
and UC exhibited the greatest overlap, >400 SNPs each, while
Crohn’s and MS contributed the least, ?40 SNPs from each.
As a control, we performed the same analysis for a GWAS study
of age-related macular degeneration (AMD) and age-related cat-
aracts from The Age-Related Eye Disease Study (AREDS) (Age-
Related Eye Disease Study Research Group, 1999), a disease
presumably with little or no association with Th cells. Theoverlap
of the AREDS SNPs with enhancers was not significant (p value
0.288). When focusing on SNP overlaps with global enhancers
present in each culture condition (Th1 or Th2 or Th0 cells),
some differences between enrichments of overlaps were re-
vealed, such as increased overlap in Th1 cell-conditioned
enhancers for Crohn’s disease (Table S8A). We examined the
distribution of the 1,281 SNPs overlapping enhancer and
found that ?50% are at intronic enhancers and ?43% are at
distal enhancers (Figure 5A). The disease associated SNPs
were typically within 100–200 kb of the nearest neighboring
gene, even when they were within our mapped enhancers
Next, we focused on the overlap of SNPs with enhancers
selectively detected in Th cells cultured in Th1, Th2, or control
Th0 cell conditions. We identified 76 associated SNPs (Table
S9) from the aforementioned disease GWASs directly overlap-
ping our Th1 cell-specific, Th2 cell-specific, or Th0 cell-specific
Enhancers for Early Th1 and Th2 Polarization
Immunity 38, 1271–1284, June 27, 2013 ª2013 Elsevier Inc. 1275
(legend on next page)
Enhancers for Early Th1 and Th2 Polarization
1276 Immunity 38, 1271–1284, June 27, 2013 ª2013 Elsevier Inc.
enhancer predictions that are defined by a condition-specific
NFR of 105–250 bp. These include 9 for T1D, 1 for Crohn’s dis-
ease, 2 for MS, 23 for ulcerative colitis, 27 for RA, 3 for psoriasis,
and 11 for asthma (see Table S8B for lineage distributions).
These SNPs were often distributed relatively uniformly across
all three T cell types. The presence of associated SNPs in en-
hancers may alter regulatory networks, disrupt the proper bal-
ance in differentiation to various Th cell subsets, and contribute
to the development of disease.
Inferring Functional Significance of rSNPs
Figure 3. Identification of Putative Enhancer Binders through Motif Analysis
(A) Example of a Th2 cell-specific enhancer in a GAB2 gene intron containing a motif for a known Th2 cell regulator, STAT6.
(B) Example of a motif at a Th1 cell-specific enhancer upstream of EED, for a TF not previously implicated in T cell biology.
(C and D) Examples of motifs within lineage-specific enhancers for TFs with known and unknown roles in T cell differentiation.
Figure 4. Enhancer and STAT6 Motif Valida-
tion by ChIP-qPCR
(A) Time course of Th2 cell polarization and Th0
cells as a control. H3K4me1, H3K27ac, and
STAT6 localization were tested by ChIP-qPCR at
six ChIP-seq determined loci of Th2 cell-specific
H3K4me1 marked enhancers containing STAT6
(B) ChIP-qPCR for STAT6 binding at 4 hr and 72 hr
in early stages of Th2 cell polarization. Th0 and
Thp cells serve as controls. Enrichment is a
percent of input.
(C) As in (A), except for H3K4me1.
(D) As in (A), except for H3K27ac. Error bars were
represented as mean ± SEM.
ly abundant in a disease-specific manner
(Figure 5C). For example, ATF3, IRF1,
and STAT1 were often occurring in T1D,
whereas VDR and PPARG were recurrent
and NF-kB were frequently represented
at asthma rSNPs, and ELF1, STAT6, and
STAT5A were often present for psoriasis.
We also found Th cell subtype-specific
enrichment of TFBS motifs overlapping
frequently overlapped SRF, NFAT1, and
MYB motifs, and Th2 cell associated
rSNPs were regularly present at AP1,
BACH1, STAT6, BCL6, and NF-kB motifs
To gain insight on how enhancers
overlapping variants might alter gene
expression, we attempted to predict
enhancer-promoter interactions by ex-
ploiting the following models used to
correlate or predict enhancer gene tar-
gets: (1) assigning the nearest neigh-
boring (NN) expressed or upregulated
gene to the enhancer; (2) assigning enhancers to genes if both
are between two CTCF binding sites (Hawkins et al., 2011), a
known enhancer blocker (Bell et al., 1999); or (3) correlating
changes in enhancer histone modification(s) with nearby
changes in gene expression (Creyghton et al., 2010; Ernst
etal., 2011),such ascorrelating changes in enhanceracetylation
(H3K27ac) with changes in nearby gene expression (Kac).
We tested these three models to predict lineage-specific
enhancer gene targets by using only genes expressed in a
Th1-, Th2-, or Th0 cell-specific manner (adjusted p value 0.01,
fold-change > 1.5) (Table S6; see Figure 5E for overlap of predic-
tions by each method). To achieve our CTCF-block model (CB),
we generated maps for CTCF binding by using ChIP-seq in Th1,
Th2, and Th0 cells. The correlation of expression with enhancer
Enhancers for Early Th1 and Th2 Polarization
Immunity 38, 1271–1284, June 27, 2013 ª2013 Elsevier Inc. 1277
acetylation changes (Kac) was limited to genes within 125 kb
from the enhancer (Table S2B), and only highly correlated
(corr > 0.90) genes were selected. Collectively, the models
determined that during the early stages of human Th cell differ-
entiation 37 and 105 enhancer-target pairs were differentially
activated when the cells are polarized to Th1 and Th2 cell sub-
sets, respectively. Control activated Th cells utilized a distinct
set of 27 enhancer-target pairs (Figure 5E). The lineage speci-
ficity of the enhancers and their predicted targets were in
agreement with the regulatory networks controlling Th cell spec-
ification (Figure S3). This begins to provide a better understand-
ing of lineage-specific gene regulation via enhancer elements.
Next, we intersected enhancers with disease-associated
SNPs and enhancers with predicted target genes. In the case
of ulcerative colitis rSNPs, JAK3, ARRDC2, ATXN1 (see Fig-
ure S4), TNFRSF6B, IL-6, NFkB, GATA3, and IL-10 were pre-
dicted targets. A GATA3 intronic Th2 cell-specific enhancer
overlapped associated SNP rs406103 that altered the PPARG
motif, which is specifically expressed in Th2 polarized cells (Fig-
ures 6A and 6C).
We identified five SNPs associated with T1D in motifs from
different Th1 cell-specific enhancer sites. One such enhancer
was predicted to target CCND2 (Figure 6B), an IL-12 inducible
gene known as a cell-cycle G1-S1 checkpoint regulator and
also associated with aryl hydrocarbon receptor (Ahr) signaling,
IL-8 signaling, p53 signaling pathway and molecular cancer
pathways (Grangeiro de Carvalho et al., 2011; Iwanaga et al.,
2008). The associated SNP rs10774213 lay within the enhancer
region with BACH2 motif (Figures 6B and 6C). The reference
nucleotide at this site is A and the variant is G. The A to G tran-
sition overlaps nucleotide 6 in the BACH2 motif, which is almost
exclusively a G (Figure 6C). The C-to-T change at SNP
rs6043388 was associated with T1D and was present at an in-
tronic enhancer for SIRPG with a CREB motif. CREB plays an
important role in T cell survival and differentiation (Wen et al.,
2010) (Figure 6C).
We found 11 associated SNPs for asthma overlapping our
condition-specific enhancers. Interestingly, one such SNP,
rs2604931, lay within an AHR motif in an activated T cell
enhancer (Th0). This G to A transition alters an exclusively A
nucleotide, at position 5 in the motif, to G. Ahr plays an important
role in setting up Treg versus Th17 cell fates in mice depending
on the ligand (Quintana et al., 2008). It remains to be determined
whether it also plays a role in Th1 or Th2 cell specification. The
enhancer with this SNP is predicted to interact with the
MGST2 gene, which is implicated in asthma (Scoggan et al.,
1997; Sjo ¨stro ¨m et al., 2001; Werz and Steinhilber, 2006). Addi-
tionally, an enhancer within the IL-4R gene contains the associ-
ated SNP rs1805012 that lies within the NFKAPPAB motif
(Figure S4). The reference nucleotide at this site is C and the
variant is T. The C to T transition overlaps nucleotide 10 in the
NFKAPPAB motif, which is exclusively a C (Figure 6C).
dicted gene targets of enhancers harboring variants include
many well-known T cell markers, though some only show a
modestchange in expressionor p-value (see TableS2A). This in-
cludes AHRR, RHOH, and DUSP16 in T1D. In rheumatoid
arthritis, BATF, IL-2RA, HLA-F (Figure S4), and PRKCQ are pre-
dicted targets. Genetic variation at the IL-2RA gene cluster has
been shown to be associated primarily with autoimmune dis-
eases including T1D, MS, and RA. We found a RA-associated
SNP, rs7904311, at an intronic enhancer of the IL-2RA gene,
which lay within a STAT1 motif, potentially regulating T helper
cell proliferation and differentiation. Though not all SNPs at en-
hancers overlap known motifs, our enhancer target predictions
still allow us to infer some biological significance for these rSNPs
(Table S2). Although no known TFBS overlapped rSNPs for
Crohn’s disease, we could still predict target genes for a subset
of these enhancers, such as GALNT2 (Figure 6C; Figure S4).
Enhancer elements are critical for fine-tuning the expression of
their target genes. SNPs in the enhancers of these target genes
could cause gain or loss of TF binding, and thus may alter the
target gene expression, Th cell differentiation, and ultimately
contribute to disease, especially given that these are enhancers
utilized during early T cell specification.
To determine whether disease-associated rSNPs alter TF
binding at enhancers, and therefore have a functional effect,
we performed DNA Affinity Precipitation Assays (DAPA) by using
the hg18 reference sequence across the predicted motif or
single base-pair variants (associated SNP) as bait. We selected
three SNPs that overlap TF binding sites at predicted lineage-
specific enhancers (Figure 6D). These are the UC-associated
SNP rs406103 at the PPARG motif predicted to target GATA3
in Th2 cells; the RA-associated SNP rs7904311 at the STAT1
motif predicted to target IL-2RA in Th0 cells; and the T1D-asso-
ciated SNP rs604388 at the CREB motif predicted to target
SIRPG in Th0 cells. The selection of SNPs and putative binding
TFs for validation was limited to availability of high quality anti-
bodies for TFs to be used in DAPA. The results show that
PPARG, STAT1, and CREB bound to the predicted TFBS by
using the reference genome as bait sequence and that this TF
binding was decreased when mutant oligonucleotides corre-
sponding to disease associated SNPs were used instead (Fig-
ure 6D; see Figure S5 for replicates and immunoblots). Notably,
the disrupted TF binding caused by the single base-pair change
was roughly equivalent to our control experiment disrupting four
conserved nucleotides in the STAT6 motif (Figure 6D; see Exper-
imental Procedures). The results indicated that introduction of a
SNP mutation in the identified TFBS of enhancers alters TF bind-
ing, suggesting a possible mechanism by which SNPs affect
enhancer function and target gene expression.
Table 1. Lead Autoimmune Disease SNPs as Regulatory SNPs
Based on Enhancer Overlap
# SNPs p Value # SNPs # SNPs # SNPs
Type 1 diabetes (T1D) 126<0.001 153
Crohn’s disease430.021 010
Multiple sclerosis (MS)42 0.033 011
Ulcerative colitis (UC) 419<0.001 5711
Rheumatoid arthritis (RA) 449<0.001 6714
Asthma134 0.059 245
Psoriasis 68<0.001 120
AREDS (control)14 0.288 –––
Associated lead SNPs were selected based on a p value < 0.01.
Enhancers for Early Th1 and Th2 Polarization
1278 Immunity 38, 1271–1284, June 27, 2013 ª2013 Elsevier Inc.
(legend on next page)
Enhancers for Early Th1 and Th2 Polarization
Immunity 38, 1271–1284, June 27, 2013 ª2013 Elsevier Inc. 1279
The control of cell-fate decisions results from complex inter-
actions of signaling and transcriptional regulatory networks
and chromatinstructure thatcanactasagatekeeper, controlling
DNA access and demarcation of regulatory elements. We used
an enhancer-specific histone modification to generate the first
maps of these elements during the early cell-fate specification
of human T helper cells. By examining Th cells 72 hr after polar-
ization toward the Th1 and Th2 cell lineages, we identified over
30,000 Th cell enhancers, providing insight into early gene regu-
lation and lineage specification. We identified enhancers unique
for Th1 and Th2 cells, suggesting that enhancers function in
lineage-specific gene regulation during early human T helper
cell specification. Furthermore, at this early stage of lineage
commitment, many enhancers were in a poised state based on
chromatin structure. Comparing this poised state with Dnase I
hypersensitivity at 7–10 days, we discovered that many en-
hancers became active upon further polarization, while another
subset remained active both at an early (H3K27ac+) and late
(DHS+) stage of cell-fate commitment. These categories likely
reflect the regulatory changes that are essential to fully commit
to their exact cell fate, as well as the plasticity of Th cells to
respond to external stimuli even after polarization. Examining
the kinetics of STAT6 binding and histone modifications across
a short time course during Th2 cell specification corroborates
this notion of a changing enhancer state. We found that whereas
STAT6 and H3K4me1 were often enriched at 4 hr, H3K27ac
seemed delayed to 72 hr.
Our findings show that chromatin modifications marking
enhancers reflect lineage specification. These enhancers are
enriched for TF binding site (TFBS) motifs representing TFs
that are often expressed in a lineage- or cell-specific manner
and correlate with specific gene expression. Collectively, this
highlights the importance of enhancers in driving unique expres-
sion profiles. Our TFBS motif analysis at predicted enhancer
sequences demonstrated enrichment for both previously un-
known and known lineage-specific Th cell regulators. For
example, Th1 cell-associated enhancers were enriched with
motifs for STAT1, STAT4, ATF3, ETS, IRF, and JUN, whereas
Th2 cell-associated enhancers motifs were enriched for
STAT6, PPARG, BACH, GFI1, NFIL3, and GATA-3. All these
identified TFs were upregulated during early Th1 or Th2 cell dif-
ferentiation and regulate target genes either independently or in
a coordinated fashion. This suggests a means by which key TFs
are likely to utilize distal regulatory elements to drive lineage
specification. We validated several STAT6 predicted motifs at
Th2 cell-specific enhancers by STAT6 ChIP. The sites were
located nearby genes such as RUNX1, FOXP1, GAB2,
IL-10RA, SETBP1, and ABHD6, which were specifically upregu-
lated in a Th2 cells. This indicates that these genes are likely
regulated by STAT6 in an enhancer-specific manner. Addition-
ally, our recent data reveals that only a fraction of genes that
are differentially expressed during the early stages of Th cell dif-
ferentiation are enriched for TFBS motifs in the promoters (Aijo ¨
et al., 2012). This further highlights the role of enhancers and
their importance in priming gene regulation and cell-fate
commitment during the early Th cell differentiation.
The tight regulatory control for proper T helper cell differentia-
tion suggests that perturbations to the system have important
implications for immunopathogenesis. One possible mechanism
for such systematic error is through improper regulatory control
of gene expression critical for a given cell-fate commitment. The
current collection of immunopathogenic GWAS has identified a
large number of disease-associated SNPs that fail to identify
causative genes. Our integrative analysis of chromatin-mapped
regulatory elements not only provides insight into the regulation
of T cell-fate commitment, but also determined that many asso-
ciated SNPs are regulatory SNPs. In several instances these
variants occurred more frequently in known motifs for TFs previ-
ously shown to be important in T helper cell differentiation. Func-
tional validation of a panel of enhancer motifs indicated that
SNPs could alter TF binding. To further our understanding of
the basis of diseases such as asthma, RA, or T1D, we made
use of predicted target genes for enhancers harboring associ-
ated variants to hypothesize that these genes may be affected
by the rSNPs. Our integrative analysis of chromatin state
maps, TFBS motif analysis, and modeling of enhancer-gene
pairs combined with autoimmune GWASs provides new insight
on the basic biology of early T cell differentiation and provides
new avenues of investigation, such as the roles that altered TF
binding and distal regulatory elements play in the etiology of
several immune disorders.
Detailed description of methods for human cord blood CD4+T cell isolation,
immunoblotting, flow cytometry, ChIP, ChIP-seq, and DAPA can be found in
the Supplemental Experimental Procedures.
Human Cord Blood CD4+T Cell Culturing
All the data included in this manuscript have been acquired under the permis-
sion from the Ethics Committee of the Hospital District of Southwest Finland.
Purified naive CD4+T cells were activated with plate-bound anti-CD3
(2.5 mg/ml for coating) and 500 ng/ml soluble anti-CD28 (Immunotech,
Figure 5. Disease Associated SNPs Directly Overlap with Known TF Motifs
(A) Percent distribution of all disease SNPs at T cell enhancers (Th1, Th2, or Th0 cell) identified in this study. The SNPs are categorized by overlap with exons (ex),
introns (in), promoter proximal (pr-pr), or distal (left in blue), and percent distribution of remaining SNPs categorized similarly (right in red), except direct promoter
(pr) overlap was considered.
(B) Histogram of distance of disease SNPs from nearest neighboring gene for SNPs overlapping enhancers (blue) and remaining SNPs (red). Inset shows zoomed
scale up to 5 3 105.
(C) Heatmap of enrichment of TFBS motifs in lineage-specific enhancers containing disease SNPs clustered by disease. Highly enriched motifs are listed on the
right relative to enriched cluster position that is indicated by the color-coded bar.
(D) Heatmap of enrichment of TFBS motifs in lineage-specific enhancers containing disease SNPs clustered by lineage. Highly enriched motifs are indicated on
the right relative to cluster position that is indicated by the color-coded bar.
(E) Overlap of predicted enhancer target genes for three models: nearest neighboring upregulated gene (NN), upregulated genes and enhancers within same
CTCF block (CB), and enhancer H3K27ac with gene upregulation within 125 kb (Kac). Venn diagrams are shown for Th0, Th1, and Th2 cells.
Enhancers for Early Th1 and Th2 Polarization
1280 Immunity 38, 1271–1284, June 27, 2013 ª2013 Elsevier Inc.
(legend on next page)
Enhancers for Early Th1 and Th2 Polarization
Immunity 38, 1271–1284, June 27, 2013 ª2013 Elsevier Inc. 1281
Marseille, France). Th1 cell polarization was initiated with 2.5 ng/ml IL-12 (R&D
Systems, Minneapolis, MN) and Th2 cell neutralizing antibody anti-IL-4 (1 ug/
ml), and 10 ng/ml IL-4 (R&D Systems) plus Th1 cell neutralizing antibody
anti-IFN-g (1ug/ml) to promote Th2 cell differentiation. For neutral Th0
cells, only neutralizing antibody without polarizing cytokines was added.
We added 17 ng/ml IL-2 (R&D Systems) to the cultures at 48 hr (Elo et al.,
Gene Expression and Analysis
using Trizol reagent (Invitrogen). For single-molecule, digital gene-expression
sequencing, 1 ug of total RNA was used as starting material and was pro-
cessed according to the Helicos Digital Gene Expression (DGE) Sample Prep-
aration guide. The sequencing data were generated on the Heliscope platform
by Helicos BioSciences, Cambridge. The raw data filtering, alignment to
RefSeq hg18 transcripts, and transcript counting were done using Helicos
Helisphere (https://launchpad.net/?helicos/+archive/ppa/+packages). Com-
bination of replicate values and identification of differentially expressed genes
were determined with DESeq (Anders and Huber, 2010).
Sequenced reads were aligned to genome (version hg18) by using Bowtie
(Langmead et al., 2009). Only uniquely mapping and single read per genome
location were retained. Input measurements were used to correct the read
density profiles. CTCF binding sites were identified by using MACS (Zhang
et al., 2008) and further filtered for more high-confidence peaks by taking
only sites where CTCF motif (from TRANSFAC) was found (p < 1e-4 for motif
scanning when comparing to 20,000 random sequences). Enhancer elements
were detected by using the methods from Heintzman et al. (2007) (H3K4me1
and H3K4me3) and He et al. (2010) (H3K4me1).
Transcription Factor Binding Analysis
Prediction of TF binding was done by using ProbTF (La ¨hdesma ¨ki et al., 2008)
and TRANSFAC (Matys et al., 2006) version 2009.3. For calculating p values,
the null distribution was estimated by sampling sequences (303 the number
of predicted enhancers) randomly from the genome. TFs with p value < 0.01
were further filtered for expressed ones (DESeq normalized read count >
10). Enrichments of bound TFs were calculated by using binomial test (using
p = 0.01). For exact locations of TF binding, the enhancer areas were scanned
with the TF motifs, and the location with highest score was selected.
Analysis of Enhancer-SNP Overlaps
Disease associated SNPs were downloaded from dbGaP database. Only
SNPs with association with p value < 0.01 were used in overlaps. Enrichment
null distributions were estimated by counting the overlaps of SNPs with
randomly selected locations of sizes equal to the enhancer sets from the
genome (excluding centromeres and telomeres) and repeating this 1,000
The Sequence Read Archive accession number for all sequences reported in
this paper is SRA082670.
Supplemental Information includes five figures, nine tables, and Supplemental
Experimental Procedures and can be found with this article online at http://dx.
We would like to thank all voluntary blood donors and the personnel of Turku
University Hospital Department of Obstetrics and Gynaecology, Maternity
Ward (Hospital District of Southwest Finland) for the cord blood collection.
We thank Sanna Edelman and Soile Tuomela for suggestions and discussion.
We thank Marjo Hakkarainen, Sarita Heinonen, Pa ¨ivi Junni, Elina Pietila ¨, and
the Finnish Microarray and Sequencing Center for technical assistance. This
study was supported by the Academy of Finland (R.L., H.L.: Centre of Excel-
lence in Molecular Systems Immunology and Physiology Research, 2012-
2017; R.L. grants 140019 and 256355); the European Commission Seventh
Framework grant EC-FP7-SYBILLA-201106 (R.L., H.L.); The Sigrid Juselius
Foundation (R.L., H.L.); Turku University Hospital Grant (R.L.); Biocenter
Finland (R.L., R.D.H.); TISE Graduate School (H.L.); NIH U01 ES017166
(B.R.) and Ludwig Institute for Cancer Research (B.R.); WA State Life Sciences
Discover Fund/(265508)Northwest Institute for Genomic Medicine (R.D.H.).
Received: September 2, 2012
Accepted: March 14, 2013
Published: June 20, 2013
Aday, A.W., Zhu, L.J., Lakshmanan, A., Wang, J., and Lawson, N.D. (2011).
Identification of cis regulatory features in the embryonic zebrafish genome
through large-scale profiling of H3K4me1 and H3K4me3 binding sites. Dev.
Biol. 357, 450–462.
Afkarian, M., Sedy, J.R., Yang, J., Jacobson, N.G., Cereb, N., Yang, S.Y.,
Murphy, T.L., and Murphy, K.M. (2002). T-bet is a STAT1-induced regulator
of IL-12R expression in naı ¨ve CD4+ T cells. Nat. Immunol. 3, 549–557.
Age-Related Eye Disease Study Research Group. (1999). The Age-Related
Eye Disease Study (AREDS): design implications. AREDS report no. 1.
Control. Clin. Trials 20, 573–600.
Aijo ¨, T., Edelman, S.M., Lo ¨nnberg, T., Larjo, A., Kallionpa ¨a ¨, H., Tuomela, S.,
Engstro ¨m, E., Lahesmaa, R., and La ¨hdesma ¨ki, H. (2012). An integrative
computational systems biology approach identifies differentially regulated dy-
namic transcriptome signatures which drive the initiation of human T helper
cell differentiation. BMC Genomics 13, 572.
Figure 6. Potential Regulatory Effects of rSNPs at lineage-specific enhancers
(A and B) Circos plots for histone modifications, disease associated SNPs - including rSNPs (open circles), and predicted gene targets at two loci: GATA3 (A) and
CCND2 (B). The outermost track shows the chromosome bands. Black boxes note enriched H3K4me1 peaks. Lineage-specific enhancers are displayed with
light gray radial lines. The profiles show H3K27ac (orange) and H3K4me1 (blue). Disease SNPs are shown with color-coded circle as indicated. The SNPs
overlapping a lineage-specific enhancer are highlighted as empty circles. Innermost track (with light gray background) shows the genes (upregulated in red, rest
are gray). Arcs connect enhancers to genes based on the target prediction method (green, NN; blue, CB; orange, KAc).
(C) List of a subset of SNPs associated with RA, T1D, ulcerative colitis (UC), and asthma that lie directly within known motifs of TFs (dot over base), that play an
important role in T cell biology. Crohn’s disease (CD) rSNPs did not overlap with any known motifs. Predicted target genes of the affected enhancers are listed to
further elucidate the potential effect of the SNP. References for gene function are given in supplement.
(D) DAPA (DNA Affinity Precipitation Assay) experiments to determine whether disease-associated rSNPs can alter transcription-factor binding to their predicted
binding sites at enhancers. Double-stranded oligonucleotides containing the predicted PPARG, STAT1, and CREB binding sites at enhancers were used as bait.
the SNPs at transcription-factor binding sites are shown in 6C (see Supplemental Information and Supplemental Experimental Procedures for complete probe
sequences). DNA sequence with STAT6 binding site and a negative control DNA sequence (oligonucleotide where STAT6 binding site has been mutated) are
provided as controls for DAPA. TF binding to oligonucleotides was detected by immunoblotting by using antibodies specific to the selected transcription factors
(see Supplemental Experimental Procedures for antibodies used). Data shown is representative for three biological replicates (see Figure S5).
Enhancers for Early Th1 and Th2 Polarization
1282 Immunity 38, 1271–1284, June 27, 2013 ª2013 Elsevier Inc.
count data. Genome Biol. 11, R106.
Anderson, C.A., Boucher, G., Lees, C.W., Franke, A., D’Amato, M., Taylor,
K.D., Lee, J.C., Goyette, P., Imielinski, M., Latiano, A., et al. (2011). Meta-anal-
ysis identifies 29 additional ulcerative colitis risk loci, increasing the number of
confirmed associations to 47. Nat. Genet. 43, 246–252.
Ansel, K.M., Djuretic, I., Tanasa, B., and Rao, A. (2006). Regulation of Th2 dif-
ferentiation and Il4 locus accessibility. Annu. Rev. Immunol. 24, 607–656.
Barrett, J.C., Clayton, D.G., Concannon, P., Akolkar, B., Cooper, J.D., Erlich,
H.A., Julier, C., Morahan, G., Nerup, J., Nierras, C., et al.; Type 1 Diabetes
Genetics Consortium. (2009).Genome-wide association study and meta-anal-
ysis find that over 40 loci affect risk of type 1 diabetes. Nat. Genet. 41,
Bell, A.C., West, A.G., and Felsenfeld, G. (1999). The protein CTCF is required
for the enhancer blocking activity of vertebrate insulators. Cell 98, 387–396.
Bulger, M., and Groudine, M. (2011). Functional and mechanistic diversity of
distal transcription enhancers. Cell 144, 327–339.
Cargill, M., Schrodi, S.J., Chang, M., Garcia, V.E., Brandon, R., Callis, K.P.,
Matsunami, N., Ardlie, K.G., Civello, D., Catanese, J.J., et al. (2007). A large-
scale genetic association study confirms IL12B and leads to the identification
of IL23R as psoriasis-risk genes. Am. J. Hum. Genet. 80, 273–290.
Cheung,V.G., and Spielman, R.S.(2009). Geneticsof human gene expression:
mapping DNA variants that influence gene expression. Nat. Rev. Genet. 10,
Creyghton, M.P., Cheng, A.W., Welstead, G.G., Kooistra, T., Carey, B.W.,
Steine, E.J., Hanna, J., Lodato, M.A., Frampton, G.M., Sharp, P.A., et al.
(2010). Histone H3K27ac separates active from poised enhancers and pre-
dicts developmental state. Proc. Natl. Acad. Sci. USA 107, 21931–21936.
Duerr, R.H., Taylor, K.D., Brant, S.R., Rioux, J.D., Silverberg, M.S., Daly, M.J.,
Steinhart, A.H., Abraham, C., Regueiro, M., Griffiths, A., et al. (2006). A
ease gene. Science 314, 1461–1463.
Elo, L.L., Ja ¨rvenpa ¨a ¨, H., Tuomela, S., Raghav, S., Ahlfors, H., Laurila, K.,
Gupta, B., Lund, R.J., Tahvanainen, J., Hawkins, R.D., et al. (2010).
Genome-wide profiling of interleukin-4 and STAT6 transcription factor regula-
tion of human Th2 cell programming. Immunity 32, 852–862.
ENCODE Project Consortium. (2011). A user’s guide to the encyclopedia of
DNA elements (ENCODE). PLoS Biol. 9, e1001046.
Ernst, J., Kheradpour, P., Mikkelsen, T.S., Shoresh, N., Ward, L.D., Epstein,
C.B., Zhang, X., Wang, L., Issner, R., Coyne, M., et al. (2011). Mapping and
analysis of chromatin state dynamics in nine human cell types. Nature 473,
Filen, S., Ylikoski, E., Tripathi, S., West, A., Bjorkman, M., Nystrom, J., Ahlfors,
H., Coffey, E., Rao, K.V., Rasool, O., and Lahesmaa, R. (2010). Activating tran-
scription factor 3 is a positive regulator of human IFNG gene expression.
J. Immunol. 184, 4990–4999.
Frossi, B., Rivera, J., Hirsch, E., and Pucillo, C. (2007). Selective activation of
Fyn/PI3K and p38 MAPK regulates IL-4 production in BMMC under nontoxic
stress condition. J. Immunol. 178, 2549–2555.
Gaulton, K.J., Nammo, T., Pasquali, L., Simon, J.M., Giresi, P.G., Fogarty,
M.P., Panhuis, T.M., Mieczkowski, P., Secchi, A., Bosco, D., et al. (2010). A
map of open chromatin in human pancreatic islets. Nat. Genet. 42, 255–259.
Grangeiro de Carvalho, E., Bonin, M., Kremsner, P.G., and Kun, J.F. (2011).
Plasmodium falciparum-infected erythrocytes and IL-12/IL-18 induce diverse
transcriptomes in human NK cells: IFN-a/b pathway versus TREM signaling.
PLoS ONE 6, e24963.
Hafler, D.A., Compston, A., Sawcer, S., Lander, E.S., Daly, M.J., De Jager,
P.L., de Bakker, P.I., Gabriel, S.B., Mirel, D.B., Ivinson, A.J., et al.;
International Multiple Sclerosis Genetics Consortium. (2007). Risk alleles for
multiple sclerosis identified by a genomewide study. N. Engl. J. Med. 357,
Hatton, R.D., Harrington, L.E., Luther, R.J., Wakefield, T., Janowski, K.M.,
Oliver, J.R., Lallone, R.L., Murphy, K.M., and Weaver, C.T. (2006). A distal
conserved sequence element controls Ifng gene expression by T cells and
NK cells. Immunity 25, 717–729.
matin states in human ES cells reveal potential regulatory sequences and
genes involved in pluripotency. Cell Res. 21, 1393–1409.
He, H.H., Meyer, C.A., Shin, H., Bailey, S.T., Wei, G., Wang, Q., Zhang, Y., Xu,
K., Ni, M., Lupien, M., et al. (2010). Nucleosome dynamics define transcrip-
tional enhancers. Nat. Genet. 42, 343–347.
Heintzman, N.D., Stuart, R.K., Hon, G., Fu, Y., Ching, C.W., Hawkins, R.D.,
Barrera,L.O.,VanCalcar, S.,Qu,C., Ching,K.A.,etal.(2007).Distinct andpre-
dictive chromatin signatures of transcriptional promoters and enhancers in the
human genome. Nat. Genet. 39, 311–318.
Heintzman, N.D., Hon, G.C., Hawkins, R.D., Kheradpour, P., Stark, A., Harp,
L.F., Ye, Z., Lee, L.K., Stuart, R.K., Ching, C.W., et al. (2009). Histone modifi-
cations at human enhancers reflect global cell-type-specific gene expression.
Nature 459, 108–112.
Helms, C., Cao, L., Krueger, J.G., Wijsman, E.M., Chamian, F., Gordon, D.,
Heffernan, M., Daw, J.A., Robarge, J., Ott, J., et al. (2003). A putative
RUNX1 binding site variant between SLC9A3R1 and NAT9 is associated
with susceptibility to psoriasis. Nat. Genet. 35, 349–356.
Hindorff, L.A., Sethupathy, P., Junkins, H.A., Ramos, E.M., Mehta, J.P.,
Collins, F.S., and Manolio, T.A. (2009). Potential etiologic and functional impli-
cations of genome-wide association loci for human diseases and traits. Proc.
Natl. Acad. Sci. USA 106, 9362–9367.
Hirota,K.,Duarte,J.H.,Veldhoen, M.,Hornsby,E.,Li,Y.,Cua,D.J.,Ahlfors, H.,
Wilhelm, C., Tolaini, M., Menzel, U., et al. (2011). Fate mapping of IL-17-pro-
ducing T cells in inflammatory responses. Nat. Immunol. 12, 255–263.
Horiuchi, S., Onodera, A., Hosokawa, H., Watanabe, Y., Tanaka, T., Sugano,
S.,Suzuki, Y.,andNakayama,T.(2011).Genome-wideanalysisreveals unique
regulation of transcription of Th2-specific genes by GATA3. J. Immunol. 186,
Iwanaga,R.,Ozono, E., Fujisawa, J., Ikeda, M.A.,Okamura,N., Huang,Y., and
Ohtani, K. (2008). Activation of the cyclin D2 and cdk6 genes through NF-
kappaB is critical for cell-cycle progression induced by HTLV-I Tax.
Oncogene 27, 5635–5642.
Kaplan, M.H., Schindler, U., Smiley, S.T., and Grusby, M.J. (1996a). Stat6 is
required for mediating responses to IL-4 and for development of Th2 cells.
Immunity 4, 313–319.
Kaplan, M.H.,Sun,Y.L.,Hoey, T.,andGrusby,M.J.(1996b).Impaired IL-12re-
sponses and enhanced development of Th2 cells in Stat4-deficient mice.
Nature 382, 174–177.
La ¨hdesma ¨ki, H., Rust, A.G., and Shmulevich, I. (2008). Probabilistic inference
of transcription factor binding frommultiple data sources. PLoS ONE3,e1820.
Langmead, B., Trapnell, C., Pop, M., and Salzberg, S.L. (2009). Ultrafast and
memory-efficient alignment of short DNA sequences to the human genome.
Genome Biol. 10, R25.
Lund, R.,Ahlfors, H., Kainonen, E.,Lahesmaa,A.M.,Dixon,C.,and Lahesmaa,
R. (2005). Identification of genes involved in the initiation of human Th1 or Th2
cell commitment. Eur. J. Immunol. 35, 3307–3319.
Lund, R.J., Lo ¨yto ¨ma ¨ki, M., Naumanen, T., Dixon, C., Chen, Z., Ahlfors, H.,
Tuomela, S., Tahvanainen, J., Scheinin, J., Henttinen, T., et al. (2007).
Genome-wide identification of novel genes involved in early Th1 and Th2 cell
differentiation. J. Immunol. 178, 3648–3660.
Matys, V., Kel-Margoulis, O.V., Fricke, E., Liebich, I., Land, S., Barre-Dirrie, A.,
Reuter, I., Chekmenev, D., Krull, M., Hornischer, K., et al. (2006). TRANSFAC
and its module TRANSCompel: transcriptional gene regulation in eukaryotes.
Nucleic Acids Res. 34(Database issue), D108–D110.
Maurano, M.T., Humbert, R., Rynes, E., Thurman, R.E., Haugen, E., Wang, H.,
Reynolds, A.P., Sandstrom, R., Qu, H., Brody, J., et al. (2012). Systematic
localization of common disease-associated variation in regulatory DNA.
Science 337, 1190–1195.
Moffatt, M.F., Gut, I.G., Demenais, F., Strachan, D.P., Bouzigon, E., Heath, S.,
von Mutius, E., Farrall, M., Lathrop, M., and Cookson, W.O.; GABRIEL
Enhancers for Early Th1 and Th2 Polarization
Immunity 38, 1271–1284, June 27, 2013 ª2013 Elsevier Inc. 1283
Consortium. (2010). A large-scale, consortium-based genomewide associa-
tion study of asthma. N. Engl. J. Med. 363, 1211–1221.
Murphy, K.M., and Stockinger, B. (2010). Effector T cell plasticity: flexibility in
the face of changing circumstances. Nat. Immunol. 11, 674–680.
Nair, R.P., Stuart, P.E., Nistor, I., Hiremagalore, R., Chia, N.V., Jenisch, S.,
Weichenthal, M., Abecasis, G.R., Lim, H.W., Christophers, E., et al. (2006).
Sequence and haplotype analysis supports HLA-C as the psoriasis suscepti-
bility 1 gene. Am. J. Hum. Genet. 78, 827–851.
Ne `gre, N., Brown, C.D., Ma, L., Bristow, C.A., Miller, S.W., Wagner, U.,
Kheradpour, P., Eaton, M.L., Loriaux, P., Sealfon, R., et al. (2011). A cis-regu-
latory map of the Drosophila genome. Nature 471, 527–531.
Ong, C.-T., and Corces, V.G. (2011). Enhancer function: new insights into the
regulation of tissue-specific gene expression. Nat. Rev. Genet. 12, 283–293.
Quintana, F.J., Basso, A.S., Iglesias, A.H., Korn, T., Farez, M.F., Bettelli, E.,
Caccamo, M., Oukka, M., and Weiner, H.L. (2008). Control of T(reg) and T(H)
17 cell differentiation by the aryl hydrocarbon receptor. Nature 453, 65–71.
Rada-Iglesias, A., Bajpai, R., Swigut, T., Brugmann, S.A., Flynn, R.A., and
Wysocka, J. (2011). A unique chromatin signature uncovers early develop-
mental enhancers in humans. Nature 470, 279–283.
Rautajoki, K.J., Kylaniemi, M.K., Raghav, S.K., Rao, K., and Lahesmaa, R.
(2008). An insight into molecular mechanisms of human T helper cell differen-
tiation. Ann. Med. 40, 322–335.
Reiner, S.L., Sallusto, F., and Lanzavecchia, A. (2007). Division of labor with a
workforce of one: challenges in specifying effector and memory T cell fate.
Science 317, 622–625.
Rioux, J.D., Xavier, R.J., Taylor, K.D., Silverberg, M.S., Goyette, P., Huett, A.,
associationstudy identifies new susceptibility lociforCrohn disease andimpli-
cates autophagy in disease pathogenesis. Nat. Genet. 39, 596–604.
Roh, T.Y., Cuddapah, S., Cui, K., and Zhao, K. (2006). The genomic landscape
of histone modifications in human T cells. Proc. Natl. Acad. Sci. USA 103,
Roh, T.Y., Wei, G., Farrell, C.M., and Zhao, K. (2007). Genome-wide prediction
of conserved and nonconserved enhancers by histone acetylation patterns.
Genome Res. 17, 74–81.
Rothenberg, E.V. (2007). Negotiation of the T lineage fate decision by tran-
scription-factor interplay and microenvironmental signals. Immunity 26,
Fitzpatrick, D.R., Stamatoyannopoulos, J.A., and Wilson, C.B. (2007).
Comprehensive epigenetic profiling identifies multiple distal regulatory ele-
ments directing transcription of the gene encoding interferon-gamma. Nat.
Immunol. 8, 732–742.
Schulz, E.G., Mariani, L., Radbruch, A., and Ho ¨fer, T. (2009). Sequential polar-
interleukin-12. Immunity 30, 673–683.
Scoggan, K.A., Jakobsson, P.J., and Ford-Hutchinson, A.W. (1997).
Production of leukotriene C4 in different human tissues is attributable to
distinct membrane bound biosynthetic enzymes. J. Biol. Chem. 272, 10182–
Shen, Y., Yue, F., McCleary, D.F., Ye, Z., Edsall, L., Kuan, S., Wagner, U.,
Dixon, J., Lee, L., Lobanenkov, V.V., and Ren, B. (2012). A map of the cis-reg-
ulatory sequences in the mouse genome. Nature 488, 116–120.
Sjo ¨stro ¨m, M., Jakobsson, P.J., Heimburger,
Haeggstro ¨m, J.Z. (2001). Human umbilical vein endothelial cells generate
leukotriene C4 via microsomal glutathione S-transferase type 2 and express
the CysLT(1) receptor. Eur. J. Biochem. 268, 2578–2586.
M., Palmblad, J., and
Stahl, E.A., Raychaudhuri, S.,Remmers, E.F., Xie, G.,Eyre, S.,Thomson, B.P.,
Li, Y., Kurreeman, F.A., Zhernakova, A., Hinks, A., et al.; BIRAC Consortium;
YEAR Consortium. (2010). Genome-wide association study meta-analysis
identifies seven new rheumatoid arthritis risk loci. Nat. Genet. 42, 508–514.
Thieu, V.T., Yu, Q., Chang, H.C., Yeh, N., Nguyen, E.T., Sehra, S., and Kaplan,
M.H. (2008). Signal transducer and activator of transcription 4 is required for
the transcription factor T-bet to promote T helper 1 cell-fate determination.
Immunity 29, 679–690.
Visel, A., Blow, M.J., Li, Z., Zhang, T., Akiyama, J.A., Holt, A., Plajzer-Frick, I.,
Shoukry, M., Wright, C., Chen, F., et al. (2009). ChIP-seq accurately predicts
tissue-specific activity of enhancers. Nature 457, 854–858.
Weaver, C.T., Hatton, R.D., Mangan, P.R., and Harrington, L.E. (2007). IL-17
family cytokines and the expanding diversity of effector T cell lineages.
Annu. Rev. Immunol. 25, 821–852.
Wei, G., Wei, L., Zhu, J., Zang, C., Hu-Li, J., Yao, Z., Cui, K., Kanno, Y., Roh,
T.Y., Watford, W.T., et al. (2009). Global mapping of H3K4me3 and H3K27me3
reveals specificity and plasticity in lineage fate determination of differentiating
CD4+ T cells. Immunity 30, 155–167.
factor CREB in immune function. J. Immunol. 185, 6413–6419.
hibitors. Pharmacol. Ther. 112, 701–718.
Zhang, Y., Liu, T., Meyer, C.A., Eeckhoute, J., Johnson, D.S., Bernstein, B.E.,
Nusbaum, C., Myers, R.M., Brown, M., Li, W., and Liu, X.S. (2008). Model-
based analysis of ChIP-Seq (MACS). Genome Biol. 9, R137.
Zhang, J.A., Mortazavi, A., Williams, B.A., Wold, B.J., and Rothenberg, E.V.
(2012). Dynamic transformations of genome-wide epigenetic marking and
transcriptional control establish T cell identity. Cell 149, 467–482.
Zhou, L., Chong, M.M., and Littman, D.R. (2009). Plasticity of CD4+ T cell line-
age differentiation. Immunity 30, 646–655.
Enhancers for Early Th1 and Th2 Polarization
1284 Immunity 38, 1271–1284, June 27, 2013 ª2013 Elsevier Inc.