Different immune cells are expected to have unique, obligatory, and stable epigenomes for cell-specific functions. Histone methylation is recognized as a major layer of the cellular epigenome. However, the discovery of histone demethylases raises questions about the stability of histone methylation and its role in the epigenome. In this study, we used chromatin-immunoprecipitation combined with microarrays to map histone H3K9 dimethylation (H3K9Me2) patterns in gene coding and CpG island regions in human primary monocytes and lymphocytes. This chromosomal mark showed consistent distribution patterns in either monocytes or lymphocytes from multiple volunteers despite age or gender, but the pattern in monocytes was clearly distinct from lymphocytes of the same population. Gene Set Enrichment analysis, a bioinformatics tool, revealed that H3K9Me2 candidate genes are enriched in many tightly controlled signaling and cell-type specific pathways. These results demonstrate that monocytes and lymphocytes have distinct epigenomes and H3K9Me2 may play regulatory roles in the transcription of genes indispensable for maintaining immune responses and cell-type specificity.

Various cells in humans have uniform genomes but diverse phenotypes. The development from a single cell to an embryo is largely an epigenetic process, in which some 23,000 genes are expressed in specific cells in a time-dependent manner by epigenetic modulation of the chromatin without changing genome sequences (1). Gene expression patterns in specific cell types are established during this process of development. It is therefore thought that the epigenome determines the differential expression of genes in specific cell types. The key layers of epigenetic control of gene expression include histone post translational modifications (PTMs),3 DNA methylation, and protein occupancies. Ample evidence now demonstrates that histone PTMs in the chromatin, such as methylation, acetylation, and phosphorylation, play vital roles in determining whether a gene is activated, repressed, or silenced and that specific patterns of histone PTMs are required for cellular development (2, 3, 4, 5). DNA methylation and histone acetylation patterns have been shown to change over time. Recently, it was reported that homozygous twins are epigenetically indistinguishable in their early years but exhibit marked variations in DNA methylation and histone acetylation patterns later in their lifetime (6). In addition, the more recent discoveries of histone demethylases provide convincing evidence of the dynamic nature of histone methylation (7, 8). These studies raise questions about the stability of histone marks, especially if histone methylation, the most stable histone PTM, is dynamically altered (9). However, specific human cell types have to maintain distinct, stable core epigenomes if they are involved in controlling cell-specific gene expression patterns and function. We sought to obtain evidence for this by mapping a key chromatin mark in a genome-wide scale in human blood cells.

Peripheral blood T cells, B cells, neutrophils, and monocytes are involved in innate and adaptive immunity and are derived by hemopoietic stem cell differentiation. Beyond transcription factors, it is now obvious that chromatin modifications and structure play key roles in cell differentiation and development (10, 11). Examining and mapping the landscape of histone modifications in T cells could provide a wealth of resources to understand T cell-specific functions (12). However, it is not clear whether the global histone lysine methylation patterns of genes in blood cells are cell-type specific and whether they differ between individuals. In the current study, we have examined these aspects because such data will be an invaluable epigenomic resource, particularly due to the recent discoveries of several histone methylases and demethylases.

Chromatin immunoprecipitation coupled to DNA microarray analysis, or ChIP-chip, is currently a widely used approach for acquiring genome-wide information on histone modifications (12, 13, 14, 15, 16, 17). In this study, by evaluating histone H3K9Me2 for variability, as a measure of epigenetic stability within cell types, we compared the profile of histone H3K9Me2 in primary lymphocytes vs monocytes isolated from peripheral blood of normal volunteers. H3K9Me2 was chosen because it is widespread in chromatin, is frequently associated with DNA methylation, and is both a repressive mark in euchromatin and a hallmark feature of heterochromatin (3, 4, 5). Increases or decreases in H3K9 methylation can change chromatin structure and affect gene expression (18). Most importantly, we observed that the chromatin mark, histone H3K9Me2, is maintained in a relatively stable and inheritable state within the coding and promoter regions of core genes in human primary lymphocytes or monocytes despite differences in age and gender, but the pattern of histone H3K9Me2 was highly specific to cell type.

Abs specific to H3K9Me2 (07–441) and H3K4Me2 (07–030) were purchased from Upstate Biotechnology. Human 12K cDNA arrays were from the University of Pennsylvania Functional Genomics Core. Human 12K CpG island arrays were from the Universal Health Network Microarray Center. The sequences of CpG islands on the array and alignment data are available through http://data.microarrays.ca/. Human 6K promoter arrays were from Aviva Systems Biology. Human promoter tiling arrays were produced by NimbleGen Systems. The design is a two-array set, containing 5.0 kb of each promoter region (from build HG17) that extends 4.2 kb upstream and 800 bp downstream of the transcription start site. Where individual 5.0-kb regions overlap, they are merged into a single larger region, preventing redundancy of coverage. The promoter regions thus range in size from 5.0 to 50 kb. These regions are tiled at a 110-bp interval using variable length probes with a target melting temperature of 76°C.

Informed consents were obtained from all volunteers before blood sampling. A total of 50 ml of blood from unrelated adult normal, healthy volunteers (n = 8) was collected in the presence of anticoagulant in accordance with an approved Institutional Review Board protocol. PBMCs were isolated by Ficoll-Paque density gradient centrifugation. Blood was diluted with equal volumes of PBS. An equal volume of diluted blood was overlaid on Ficoll-Paque-plus in 1:1 ratio and centrifuged at 1200 × g for 20 min at 18–20°C. The leukocyte population was collected from the interface and washed with PBS several times to remove plasma and Ficoll. Approximately 50 million washed cells in 10 ml of RPMI 1640 medium containing 10% FCS were plated in 100-mm culture dishes to allow monocytes to adhere on the surface of the dish for 2–3 h. The nonadherent cells (lymphocyte population) were removed, washed with fresh medium, and cultured in RPMI 1640 medium. Attached monocytes were washed twice with warm RPMI 1640 medium containing 10% FCS and allowed to remain in the dish overnight at 37°C in 5% CO2. During this period, the monocytes detached from the dish. They were collected in fresh RPMI 1640 medium. Monocyte purity is ∼85%, and this is similar (86%) to what we get with the monocyte isolation kit II from Miltenyi Biotec. Lymphocyte preparations contain <5% monocytes based on CD14 expression. Viability was 98% and 99% in monocyte and lymphocyte fractions, respectively (trypan blue staining). Both lymphocytes and monocytes were used for ChIP experiments.

Purified blood monocytes and lymphocytes (∼107) from normal and T1D subjects were crosslinked in 1% formaldehyde, washed, and then sonicated to shear DNA. IP was then performed with anti-H3K9Me2 or -H3K4Me2, as described earlier (15). One-tenth of the total lysate was used for “no antibody” control. IP was then performed with the methylated histone Abs. Precipitates are washed, eluted, and crosslinks reversed. Part of the DNA was retained for conventional ChIP PCR. The remaining ChIP-enriched DNA was amplified by ligation-mediated PCR. DNA was blunted with T4 DNA polymerase, purified, and ligated with linker (5′-GCGGTGACCCGGGAGATCTGAATTC-3′ and 5′-GAATTCAGATC-3′). DNA was purified on Qiagen spin columns, and used for PCR amplification (20 cycles) with the primer 5′-GCGGTGACCCGGGAGATCTGAATTC-3′. The PCR products were purified using Qiagen spin columns used in ChIP-chip experiments. The ChIP-chips were performed with human 12K cDNA, 12K CpG, and 6K promoter arrays using protocols described by us (15). All NimbleGen Systems arrays were hybridized, and the data were extracted according to standard operating procedures by NimbleGen Systems. Signal map software provided by NimbleGen Systems was used to visualize the array peaks.

After washing, hybridized microarray slides were scanned using GenePix 4000B scanner (Axon Instruments). Acquired microarray images were analyzed with Genepix v.6 software. Preprocessing of raw data and statistical analyses were performed (15). To examine the methylation pattern in monocytes and lymphocytes, an unsupervised hierarchical clustering method was used to group the samples. We focused on probes whose ChIP-enriched signal was 1.5-fold higher than the no-antibody control signal in at least four of the 16 samples. A total of 2682 probes on cDNA array and 2878 probes on CpG array satisfied the above criteria and were applied to Cluster v2.11 to generate a hierarchical clustering diagram. Pearson correlation was used as distance measurement and average linkage method was used to generate the dendrogram, which was visualized using Java Treeview V1.0.12. The unbiased probe selection criterion assures that these observations are not due to probes that either already shows consistent patterns within the cell type groups or show difference between the groups. To generate the venn diagrams, dimethylation candidates in lymphocyte and monocyte samples were selected separately using Bioconductor package LIMMA. The criteria used in candidate selection were >1.5-fold enrichment in ChIP-enriched DNA compared with input DNA and with a value of p < 0.05.

Gene set enrichment analysis (GSEA) was performed as described (6). This method analyzes expression data at the level of predefined gene sets instead of individual genes to detect significant concordant differences in biological processes between two phenotypes. GSEA 2.0 software and generic pathways in the Molecular Signature Database of genesets C2 version 2 were used for the analysis. All genes with known symbols in the data set were ranked based on their correlation to the lymphocyte phenotype, and the rank positions of all members of a given gene set were used to calculate an enrichment score (ES). Subsequently, 1000 permutations were used to determine which gene sets were significantly enriched in monocytes or lymphocytes.

Genome-wide histone methylation data sets for histone H3K9Me2 were obtained by the ChIP-chip approach (13, 19). Peripheral blood lymphocyte and monocyte fractions were prepared from eight normal healthy volunteers, ages ranging from 36 to 71 (three males and five females) according to standard procedures as described earlier (19). All blood samples were obtained under protocols approved by the City of Hope IRB. Blood cells were then processed for conventional ChIP assays and ChIP-chip profiling. In brief, isolated cells (lymphocytes or monocytes) from peripheral blood were crosslinked with 1% formaldehyde. Anti-dimethyl-histone H3K9 (Upstate 07-441), specific to H3K9Me2 but not to H3K9Me or H3K9Me3, was used in our ChIP-chip experiments. The Ab-enriched DNA samples and no-antibody controls were prepared from lymphocytes and monocytes of each volunteer separately, amplified by ligation-mediated PCR, and labeled with Cy5 and Cy3 dyes. They were then analyzed by ChIP-chips using human 12K CpG island arrays and 12K cDNA arrays as described under Materials and Methods.

To analyze histone H3K9Me2 profiling data sets, we used an unsupervised hierarchical clustering method to group the samples. Fig. 1,A depicts the 12K cDNA array data. Two striking features are observed. First, the H3K9Me2 distribution patterns in the coding regions are remarkably similar in all the individuals within their monocyte or lymphocyte groups irrespective of age or gender. Second, the distribution patterns are clearly distinct between lymphocyte and monocyte groups. To elaborate the cell-type distinct patterns, we selected probes consistently methylated in all samples using LIMMA (detailed candidate list in Tables S1–S4, published as online supplement information).4 The venn diagram (Fig. 1,C) shows a striking difference with very little overlap of methylated probes between monocytes and lymphocytes. Similar results were also found using CpG island arrays that are mostly representative of promoter regions (20) (Fig. 1, B and D). This genome-wide profiling demonstrates for the first time, with resolution at the gene level, that two closely related human blood cells with different functions have distinct H3K9Me2 patterns, whereas patterns within a specific cell type are remarkably similar despite age or gender, suggesting that stability of core epigenome integrity is indispensable for cell-specific functions. However, in analyzing lymphocytes and monocytes separately, we did find that a portion of the epigenome does indeed show key differences among individuals, with person-to-person variations as previously indicated (6). This emphasizes a critical need to analyze blood cells separately, rather than whole blood when performing these kinds of profiling studies.

FIGURE 1.

Comparison of histone H3K9Me2 profiles in monocytes and lymphocytes within coding regions and CpG islands. Lymphocytes and monocytes were isolated from normal volunteers. Each sample was processed for ChIP-chips with anti-H3K9Me2 Ab. A, Hierarchical clustering with Spearman Correlation and average linkage applied to selected probes from cDNA arrays. Each column represents methylation profile in one individual, and the columns represent H3K9Me2 profiles in eight monocyte and eight lymphocyte samples. Color bar shows H3K9Me2 enrichment level compared with the average of all samples, where red indicates increased methylation and green indicates decreased methylation. Color intensity correlates to the magnitude of change. B, Similar hierarchical clustering patterns from 12K CpG arrays. C, Venn diagram of selected histone H3K9 methylated genes from 12K cDNA array. These candidates in lymphocyte and monocytes were selected separately and only probes with p < 0.05- and >1.5-fold were selected. F = Female; M = male. D, Venn diagram of selected histone H3K9 candidates from CpG arrays.

FIGURE 1.

Comparison of histone H3K9Me2 profiles in monocytes and lymphocytes within coding regions and CpG islands. Lymphocytes and monocytes were isolated from normal volunteers. Each sample was processed for ChIP-chips with anti-H3K9Me2 Ab. A, Hierarchical clustering with Spearman Correlation and average linkage applied to selected probes from cDNA arrays. Each column represents methylation profile in one individual, and the columns represent H3K9Me2 profiles in eight monocyte and eight lymphocyte samples. Color bar shows H3K9Me2 enrichment level compared with the average of all samples, where red indicates increased methylation and green indicates decreased methylation. Color intensity correlates to the magnitude of change. B, Similar hierarchical clustering patterns from 12K CpG arrays. C, Venn diagram of selected histone H3K9 methylated genes from 12K cDNA array. These candidates in lymphocyte and monocytes were selected separately and only probes with p < 0.05- and >1.5-fold were selected. F = Female; M = male. D, Venn diagram of selected histone H3K9 candidates from CpG arrays.

Close modal

The size of the human genome and complexity of histone PTMs are major impediments to mapping the entire human epigenome (21, 22). However, the majority of the human genome is composed of repetitive elements with only 4% coding for proteins that include the coding and promoter regions of genes. In this study, we used two very affordable DNA microarrays, human 12K cDNA arrays and 12K CpG island arrays. The human 12K CpG island array contains a significant percentage of the CpG islands found in the human genome and ∼68% of them were located near a transcription start site (20). Together, these cDNA and CpG arrays cover about a third of all human genes and a significant fraction of CpG islands and promoters. It should be noted that although these arrays have lower resolution and accuracy than high density tiling arrays, they are highly sensitive, requiring as little as 0.1 μg ChIP DNA compared with 5 μg for tiling arrays. This is an important factor for ChIP-chip analysis due to limited numbers of primary human cells, like blood lymphocytes and monocytes obtained in this study. Tiling arrays that cover the whole human gene coding regions for ChIP-chip are still beyond the reach of most laboratories. Most importantly, the 12K cDNA and CpG arrays used in this study do provide valid information at gene resolution.

Based on a previous study (6), it can be predicted that individual epigenomes have significant variations. However, it is also reasonable that different cell types would have specific and obligatory epigenomes for cell-specific functions. Indeed, results from this study (Fig. 1) demonstrate that, despite noticeable variations among individuals, the histone H3K9Me2 distribution pattern among core genes is cell-type specific and stable.

An unexpected advantage of using cDNA arrays in ChIP-chip analysis is that data mining and pathway analyses packages used in gene expression analyses can be directly applied without any modification. Therefore, we used GSEA software (23) to examine whether genes within specific signaling pathways are H3K9Me2 targets in monocytes or lymphocytes. GSEA, which is still not applicable to tiling array data, is a computational method that determines whether an a priori defined set of genes in biological pathways shows statistically significant and concordant differences between two biological states. It is a powerful data mining bioinformatics tool that can complement single-gene studies to examine changes occurring in several gene members of biological networks. For each gene set, an ES is calculated to measure the degree of enrichment in lymphocytes or monocytes. Nominal p values derived from permutation, as well as multiple comparison corrected p values (false discovery rate (FDR) q value and family-wise error rate (FWER) p values), are provided to assess the significance of the ES.

As illustrated in Table I, it is evident that histone H3K9Me2 is substantially enriched in many tightly regulated pathways in monocytes and lymphocytes, and especially in some cell-type specific pathways when lymphocytes and monocytes are compared. Among the top twenty-four H3K9Me2 enriched pathways in lymphocytes, several are clearly associated with lymphocytes or lymphocyte-specific functions. Notably, T cell, IL4 signaling, and GATA3 transcription, which are known to be T cell response pathways, are among the top pathways. Similar results are seen in monocytes (Table I, right panel) where TNFR1, inflammation, and matrix metalloproteinases/cytokines are present and known to have monocyte-specific functions. Although the mechanism for this is not completely understood, they do confirm that chromosomal H3K9Me2 is an euchromatin mark because these identified genes are expressed or are inducible in the respective cells. They suggest that this chromatin mark has evolved as an important modulator of gene regulation, especially for tightly controlled and cell-type specific genes. As indicated by previous studies (18, 24, 25, 26), H3K9Me2 most likely plays a role in transient transcriptional repression of these genes. This may be mediated by G9a, the major mammalian enzyme responsible for H3K9Me2 in cells that is also essential for early embryogenesis (18, 27).

Table I.

GSEA of histone H3K9Me2-methylated candidates in lymphocytes and monocytes (cDNA array)

LymphocytesSizeESNESaNominal p ValFDR q ValMonocytesSizeESNESaNominal p ValFDR q Val
Thrombin signaling 11 −0.62 −1.87 0.00 0.40 Fibrinolysis 0.80 1.79 0.00 1.00 
Wnt/β catenin signaling 18 −0.50 −1.81 0.00 0.36 Cell cycle G1 to S 34 0.42 1.76 0.00 0.72 
Rhodopsin-like G protein receptor 27 −0.48 −1.77 0.01 0.36 IL17 signaling 0.81 1.76 0.00 0.49 
T cell signaling 23 −0.46 −1.76 0.01 0.30 TNFR1 signaling 19 0.56 1.75 0.02 0.40 
Myocyte adrenergic Receptor 12 −0.53 −1.74 0.01 0.29 Cytokine/matrix metalloproteinases Connection 0.79 1.74 0.00 0.35 
B cell Ag receptor 25 −0.43 −1.72 0.01 0.29 ETS-mediated macrophage differentiation 13 0.73 1.73 0.00 0.31 
Oxidative phosphorylation 45 −0.42 −1.69 0.01 0.32 Cell cycle KEGGb 37 0.39 1.70 0.00 0.33 
VIPb inhibit the apoptosis in T cells 15 −0.56 −1.66 0.00 0.36 TNF silencer of death domain signaling 0.74 1.65 0.01 0.43 
G α 13 pathway 23 −0.41 −1.66 0.01 0.34 Cascade of cyclin gene expression 0.81 1.65 0.01 0.39 
Adrenergic receptor signaling 20 −0.45 −1.64 0.01 0.33 Cytokines and inflammatory response 12 0.67 1.64 0.01 0.36 
fMLP-induced chemokine activation 27 −0.44 −1.63 0.02 0.33 Secretin-like G protein-coupled receptors 0.83 1.63 0.01 0.37 
B cell immune responses (TACI)b 10 −0.56 −1.62 0.03 0.33 IL10 signaling 0.54 1.63 0.01 0.34 
Steroids biosynthesis 11 −0.52 −1.62 0.02 0.31 NK T cell pathway 0.67 1.60 0.04 0.39 
IL 4 signaling −0.60 −1.61 0.03 0.31 MAPK signaling 64 0.39 1.59 0.04 0.37 
GATA3 transcription 10 −0.62 −1.60 0.03 0.31 Cysteine metabolism 0.68 1.58 0.01 0.36 
Methane metabolism −0.60 −1.59 0.03 0.31 Wnt signaling 45 0.45 1.56 0.08 0.41 
Programmed cell death −0.52 −1.57 0.03 0.33 Cyclin E destruction 0.66 1.56 0.04 0.38 
Myosin phosphorylation −0.56 −1.52 0.04 0.41 Tyrosine metabolism 10 0.67 1.55 0.05 0.38 
Regulation of ck1/cdk5 10 −0.61 −1.51 0.03 0.42 Cytokines-mediated hematopoiesis 0.78 1.55 0.05 0.36 
Erk1/Erk2 MAPK signaling 24 −0.42 −1.50 0.05 0.41 Cell cycle morphological progression 37 0.34 1.54 0.03 0.34 
N-Glycan degradation 10 −0.53 −1.50 0.05 0.39 Ubiquitin-mediated proteolysis 21 0.46 1.55 0.01 0.35 
Heme biosynthesis −0.58 −1.49 0.08 0.39 FASb signaling 17 0.48 1.54 0.03 0.34 
Protein kinase C signaling −0.58 −1.47 0.02 0.44 Methionine metabolism 0.56 1.53 0.05 0.36 
CD40 signaling 11 −0.45 −1.46 0.04 0.44 Keratinocyte differentiation 35 0.37 1.51 0.09 0.39 
LymphocytesSizeESNESaNominal p ValFDR q ValMonocytesSizeESNESaNominal p ValFDR q Val
Thrombin signaling 11 −0.62 −1.87 0.00 0.40 Fibrinolysis 0.80 1.79 0.00 1.00 
Wnt/β catenin signaling 18 −0.50 −1.81 0.00 0.36 Cell cycle G1 to S 34 0.42 1.76 0.00 0.72 
Rhodopsin-like G protein receptor 27 −0.48 −1.77 0.01 0.36 IL17 signaling 0.81 1.76 0.00 0.49 
T cell signaling 23 −0.46 −1.76 0.01 0.30 TNFR1 signaling 19 0.56 1.75 0.02 0.40 
Myocyte adrenergic Receptor 12 −0.53 −1.74 0.01 0.29 Cytokine/matrix metalloproteinases Connection 0.79 1.74 0.00 0.35 
B cell Ag receptor 25 −0.43 −1.72 0.01 0.29 ETS-mediated macrophage differentiation 13 0.73 1.73 0.00 0.31 
Oxidative phosphorylation 45 −0.42 −1.69 0.01 0.32 Cell cycle KEGGb 37 0.39 1.70 0.00 0.33 
VIPb inhibit the apoptosis in T cells 15 −0.56 −1.66 0.00 0.36 TNF silencer of death domain signaling 0.74 1.65 0.01 0.43 
G α 13 pathway 23 −0.41 −1.66 0.01 0.34 Cascade of cyclin gene expression 0.81 1.65 0.01 0.39 
Adrenergic receptor signaling 20 −0.45 −1.64 0.01 0.33 Cytokines and inflammatory response 12 0.67 1.64 0.01 0.36 
fMLP-induced chemokine activation 27 −0.44 −1.63 0.02 0.33 Secretin-like G protein-coupled receptors 0.83 1.63 0.01 0.37 
B cell immune responses (TACI)b 10 −0.56 −1.62 0.03 0.33 IL10 signaling 0.54 1.63 0.01 0.34 
Steroids biosynthesis 11 −0.52 −1.62 0.02 0.31 NK T cell pathway 0.67 1.60 0.04 0.39 
IL 4 signaling −0.60 −1.61 0.03 0.31 MAPK signaling 64 0.39 1.59 0.04 0.37 
GATA3 transcription 10 −0.62 −1.60 0.03 0.31 Cysteine metabolism 0.68 1.58 0.01 0.36 
Methane metabolism −0.60 −1.59 0.03 0.31 Wnt signaling 45 0.45 1.56 0.08 0.41 
Programmed cell death −0.52 −1.57 0.03 0.33 Cyclin E destruction 0.66 1.56 0.04 0.38 
Myosin phosphorylation −0.56 −1.52 0.04 0.41 Tyrosine metabolism 10 0.67 1.55 0.05 0.38 
Regulation of ck1/cdk5 10 −0.61 −1.51 0.03 0.42 Cytokines-mediated hematopoiesis 0.78 1.55 0.05 0.36 
Erk1/Erk2 MAPK signaling 24 −0.42 −1.50 0.05 0.41 Cell cycle morphological progression 37 0.34 1.54 0.03 0.34 
N-Glycan degradation 10 −0.53 −1.50 0.05 0.39 Ubiquitin-mediated proteolysis 21 0.46 1.55 0.01 0.35 
Heme biosynthesis −0.58 −1.49 0.08 0.39 FASb signaling 17 0.48 1.54 0.03 0.34 
Protein kinase C signaling −0.58 −1.47 0.02 0.44 Methionine metabolism 0.56 1.53 0.05 0.36 
CD40 signaling 11 −0.45 −1.46 0.04 0.44 Keratinocyte differentiation 35 0.37 1.51 0.09 0.39 
a

NES: Normalized enrichment score.

b

KEGG: Kyoto Encyclopedia of Genes and Genomes; VIP: Vasoactive intestinal peptide; TACI: TNF receptor superfamily member 13B; FAS: TNF receptor superfamily member 6.

A recent study of G9a that greatly enhances our understanding of the role of histone H3K9Me2 in cell development (24) showed that, surprisingly, only eight genes were up-regulated by 2-fold in G9a knock out cells despite a significant reduction of global H3K9Me2. This suggests that, although H3K9Me2 is an epigenetic mark of heterochromatin formation and transcription silencing in euchromatin, this mark alone probably cannot lead to or maintain these events unless other repression marks, such as H3K9Me3, H3K27Me3, H4K20Me3, or DNA methylation, are also involved. Notably, evidence shows that histone H3K9Me2 and heterochromatin protein 1γ, a protein containing a chromodomain that recognizes H3K9 methylation, are present in transcribed genes and both were associated with elongation of RNA polymerase II (28). The wide distribution of H3K9Me2 in gene promoter and coding regions in lymphocytes and monocytes noted in our current study indicates that this mark has key roles in transcription at the level of chromatin genome. It has been demonstrated that the distribution of H3K9Me2 and H3K4Me2/3 in gene promoter and coding regions are exclusive of each other (12, 15), indicating that these marks define chromatin states of transcribed genes in euchromatin. Thus, chromatin with H3K4Me2/3 (along with H3K9 acetylation) would represent proactive transcription state, whereas that with H3K9Me2 would represent repressed transcription but not gene silencing. Such repression can be reversed to active transcription in response to proper signals such as the pathways shown in this study (Table I). This suggests a “fine-tuning” role for H3K9Me2 in regulating the transcription of these genes through unknown mechanisms occurring in the coding regions of these genes. Moreover, such repression is most likely critical for a cell to maintain its unique identity and explain, at least in part, the stable histone methylation distribution patterns we observed (Fig. 1).

Evidence shows that H3K4Me2 and H3K9Me2 are present in partitioned chromosome regions that are structurally and functionally distinct in eukaryotes (29). H3K4Me2 is closely associated with H3K4Me3, which mainly occurs concomitantly on active loci (14, 15, 30), especially in the first exon. H3K4Me2 can be present on active as well as inactive genes. Its distribution can extend to promoter regions and gene coding regions (29, 30). We therefore tested whether patterns of K4Me2, a chromosomal mark associated with gene activation, are also conserved among the different individuals. We performed ChIP-chip experiments in primary lymphocytes from the volunteers using anti-histone H3K4Me2 and human 6K promoter arrays (Aviva Systems Biology). Similar to chromatin mark histone H3K9Me2, H3K4Me2 distribution patterns at promoter region are comparable in lymphocytes among individuals (Fig. 2). Analysis results in 1240 candidate genes were identified as enriched H3K4Me2 in promoter region. Detailed genes list in Tables S5, published as online supplement.

FIGURE 2.

Hierarchical clustering of histone H3K4Me2 profiling in lymphocytes among individuals. Lymphocytes were isolated from normal volunteers. Each sample was processed for ChIP-chip analysis with human 6K promoter arrays separately with anti-H3K4Me2 Ab. Hierarchically clustered histone H3K9Me2 profiles of lymphocyte sample from six normal subjects (columns) and 5026 probes (rows). Only well-measured probes were included for analysis. The filtering yielded 5026 probes, corresponding to 6000 probes. Hierarchical clustering with Spearman Correlation and average linkage applied to selected probes.

FIGURE 2.

Hierarchical clustering of histone H3K4Me2 profiling in lymphocytes among individuals. Lymphocytes were isolated from normal volunteers. Each sample was processed for ChIP-chip analysis with human 6K promoter arrays separately with anti-H3K4Me2 Ab. Hierarchically clustered histone H3K9Me2 profiles of lymphocyte sample from six normal subjects (columns) and 5026 probes (rows). Only well-measured probes were included for analysis. The filtering yielded 5026 probes, corresponding to 6000 probes. Hierarchical clustering with Spearman Correlation and average linkage applied to selected probes.

Close modal

The accuracy of the analysis of histone H3K4Me2-methylated candidates (Table S5, online supplement) were further validated by an independent ChIP-chip experiment with the same Ab but this time with a high resolution human promoter tiling array (NimbleGen Systems). Fig. 3 shows the promoter regions of selected candidate methylated genes (RPL30, RPL26, NOTCH4, SLCA1, DUSP2, CDC25A, CDCL5, AND OSR1) from Table S5 (online supplement) and demonstrates that promoter region of these gene are, indeed, nicely enriched in H3K4Me2 peaks. In contrast, no peaks are seen in promoters of none candidate genes (LU17A4, PLC, MOG, and ADH6) that are not determined to be methylated candidates from the analysis (Fig. 3) and provide further evidence for the accuracy of the microarray data (Fig. 2). Together, these experiments are consistence with histone H3K9Me2 results (Fig. 1) and prove that H3K4Me2 distribution patterns in lymphocytes are also cell-specific among individuals regardless of age or gender.

FIGURE 3.

Representative tiling array views of histone H3K4Me2 patterns across promoter regions of selected genes from histone H3K4Me2 candidates. Eight H3K4Me2 candidate genes (RPL30, RPL26, NOTCH4, SLCA1, DUSP2, CDC25A, CDCL5, and OSR1) (Table S5 in online supplement) and four non-candidates genes (LU17A4, PLC, MOG, and ADH6) were selected to validate their promoter H3K4Me2 patterns in the promoter tiling arrays. Dots represent the transcription start sites and arrows indicate the transcription direction. Histone H3K4Me2 data within promoter regions is visualized by signal map software provided by NimbleGen Systems.

FIGURE 3.

Representative tiling array views of histone H3K4Me2 patterns across promoter regions of selected genes from histone H3K4Me2 candidates. Eight H3K4Me2 candidate genes (RPL30, RPL26, NOTCH4, SLCA1, DUSP2, CDC25A, CDCL5, and OSR1) (Table S5 in online supplement) and four non-candidates genes (LU17A4, PLC, MOG, and ADH6) were selected to validate their promoter H3K4Me2 patterns in the promoter tiling arrays. Dots represent the transcription start sites and arrows indicate the transcription direction. Histone H3K4Me2 data within promoter regions is visualized by signal map software provided by NimbleGen Systems.

Close modal

This study shows, for the first time, that monocytes and lymphocytes of different individuals, regardless of age or gender, have stable, consistent, and cell-type specific histone methylation patterns in gene coding, promoter, and CpG islands regions. It is highly likely that this conclusion can be extended to the whole human genome. Although our result could be anticipated, this genome-wide study provides direct experimental proof that histone methylation distributions are not random but exhibit specific and stable patterns among different individuals. Notably, this is analogous to gene expression patterns, but there are key differences between them. Whereas gene expression profiling measures thousands of RNA transcripts simultaneously, histone methylation profiling provides genome-wide information of chromatin status depending on the type of microarray used.

The implications of our results are the following. First, the specific blood-cell type maintains its histone methylation patterns and perhaps the epigenome too, although factors such as age, gender, and cellular states may contribute to significant variations. Second, the methylation patterns of core genes in differentiated cells are relatively stable and very likely heritable. Thus, a cell might have system(s) to repair imbalances or “damages” involving nucleosome deletion or alterations in PTMs induced by events such as DNA repair. Third, both histone methylation and gene expression have cell-type specific patterns and provide strong evidence that histone PTMs control gene expression. Fourth, histone H3K9Me2 is enriched in specific cell-type genes and pathways, suggesting that the H3K9Me2 methyltransferase, G9a, plays a regulatory role in the transcription of these genes, and also explaining the requirement for this chromosomal mark and G9a during differentiation. Last, but not least, human peripheral blood remains the most easily accessible and noninvasive source of human tissue, which can provide valuable information such as genomic mutations and gene expression, particularly with respect to immune and inflammation responses. Approaches such as DNA sequence analyses, mRNA profiling, and proteomics have yielded immense amounts of valuable data for medical research. However, there is little data in the context of histone methylation in specific human blood cells. Our results demonstrate that histone PTMs have the potential to become chromatin information databases for individuals that can be enriched by profiling these chromatin marks in human primary tissues and cells using the ChIP-chip approach.

We are deeply grateful to all the volunteers who donated blood.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by grants from the National Institutes of Health (R01 DK065073) and the Juvenile Diabetes Research Foundation International (to R.N.) and in part by a General Clinical Research Center grant from the National Center for Research Resources (M01RR00043 to City of Hope).

3

Abbreviations used in this paper: PTM, post translational modification; H3K9Me2, H3 lysine-9 dimethylation; ChIP-chip, chromatin immunoprecipitation coupled to DNA microarray analysis; GSEA, gene set enrichment analysis; ES, enrichment score; FDR, false discovery rate.

4

The online version of this article contains supplemental material.

1
Rodenhiser, D., M. Mann.
2006
. Epigenetics and human disease: translating basic biology into clinical applications.
CMAJ
174
:
341
-348.
2
Strahl, B. D., C. D. Allis.
2000
. The language of covalent histone modifications.
Nature
403
:
41
-45.
3
Sims, R. J., III, K. Nishioka, D. Reinberg.
2003
. Histone lysine methylation: a signature for chromatin function.
Trends Genet.
19
:
629
-639.
4
Shilatifard, A..
2006
. Chromatin modifications by methylation and ubiquitination: implications in the regulation of gene expression.
Annu. Rev. Biochem.
75
:
243
-269.
5
Li, B., M. Carey, J. L. Workman.
2007
. The role of chromatin during transcription.
Cell
128
:
707
-719.
6
Fraga, M. F., E. Ballestar, M. F. Paz, S. Ropero, F. Setien, M. L. Ballestar, D. Heine-Suner, J. C. Cigudosa, M. Urioste, J. Benitez, et al
2005
. Epigenetic differences arise during the lifetime of monozygotic twins.
Proc. Natl. Acad. Sci. USA
102
:
10604
-10609.
7
Tsukada, Y., J. Fang, H. Erdjument-Bromage, M. E. Warren, C. H. Borchers, P. Tempst, Y. Zhang.
2006
. Histone demethylation by a family of JmjC domain-containing proteins.
Nature
439
:
811
-816.
8
Shi, Y. J., C. Matson, F. Lan, S. Iwase, T. Baba, Y. Shi.
2005
. Regulation of LSD1 histone demethylase activity by its associated factors.
Mol. Cell.
19
:
857
-864.
9
Trojer, P., D. Reinberg.
2006
. Histone lysine demethylases and their impact on epigenetics.
Cell
125
:
213
-217.
10
Reiner, S. L..
2005
. Epigenetic control in the immune response.
Hum. Mol. Gen.
14 Spec No. 1
:
R41
-R46.
11
Kioussis, D., K. Georgopoulos.
2007
. Epigenetic flexibility underlying lineage choices in the adaptive immune system.
Science
317
:
620
-622.
12
Barski, A., S. Cuddapah, K. Cui, T. Y. Roh, D. E. Schones, Z. Wang, G. Wei, I. Chepelev, K. Zhao.
2007
. High-resolution profiling of histone methylations in the human genome.
Cell
129
:
823
-837.
13
Ren, B., F. Robert, J. J. Wyrick, O. Aparicio, E. G. Jennings, I. Simon, J. Zeitlinger, J. Schreiber, N. Hannett, E. Kanin, et al
2000
. Genome-wide location and function of DNA binding proteins.
Science
290
:
2306
-2309.
14
Bernstein, B. E., M. Kamal, K. Lindblad-Toh, S. Bekiranov, D. K. Bailey, D. J. Huebert, S. McMahon, E. K. Karlsson, E. J. Kulbokas, III, T. R. Gingeras, et al
2005
. Genomic maps and comparative analysis of histone modifications in human and mouse.
Cell
120
:
169
-181.
15
Miao, F., R. Natarajan.
2005
. Mapping global histone methylation patterns in the coding regions of human genes.
Mol. Cell. Biol.
25
:
4650
-4661.
16
Mito, Y., J. G. Henikoff, S. Henikoff.
2005
. Genome-scale profiling of histone H3.3 replacement patterns.
Nat. Genet.
37
:
1090
-1097.
17
Schubeler, D., D. M. MacAlpine, D. Scalzo, C. Wirbelauer, C. Kooperberg, F. van Leeuwen, D. E. Gottschling, L. P. O’Neill, B. M. Turner, J. Delrow, et al
2004
. The histone modification pattern of active genes revealed through genome-wide chromatin analysis of a higher eukaryote.
Genes Dev.
18
:
1263
-1271.
18
Tachibana, M., K. Sugimoto, M. Nozaki, J. Ueda, T. Ohta, M. Ohki, M. Fukuda, N. Takeda, H. Niida, H. Kato, Y. Shinkai.
2002
. G9a histone methyltransferase plays a dominant role in euchromatic histone H3 lysine 9 methylation and is essential for early embryogenesis.
Genes Dev.
16
:
1779
-1791.
19
Miao, F., X. Wu, L. Zhang, Y. C. Yuan, A. D. Riggs, R. Natarajan.
2007
. Genome-wide analysis of histone lysine methylation variations caused by diabetic conditions in human monocytes.
J. Biol. Chem.
282
:
13854
-13863.
20
Heisler, L. E., D. Torti, P. C. Boutros, J. Watson, C. Chan, N. Winegarden, M. Takahashi, P. Yau, T. H. Huang, P. J. Farnham, et al
2005
. CpG island microarray probe sequences derived from a physical library are representative of CpG islands annotated on the human genome.
Nucleic Acids Res.
33
:
2952
-2961.
21
Jones, P. A., R. Martienssen.
2005
. A blueprint for a Human Epigenome Project: the AACR Human Epigenome Workshop.
Cancer Res.
65
:
11241
-11246.
22
Bernstein, B. E., A. Meissner, E. S. Lander.
2007
. The mammalian epigenome.
Cell
128
:
669
-681.
23
Subramanian, A., P. Tamayo, V. K. Mootha, S. Mukherjee, B. L. Ebert, M. A. Gillette, A. Paulovich, S. L. Pomeroy, T. R. Golub, E. S. Lander, J. P. Mesirov.
2005
. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles.
Proc. Natl. Acad. Sci. USA
102
:
15545
-15550.
24
Tachibana, M., M. Nozaki, N. Takeda, Y. Shinkai.
2007
. Functional dynamics of H3K9 methylation during meiotic prophase progression.
EMBO J.
26
:
3346
-3359.
25
Ueda, J., M. Tachibana, T. Ikura, Y. Shinkai.
2006
. Zinc finger protein Wiz links G9a/GLP histone methyltransferases to the co-repressor molecule CtBP.
J. Biol. Chem.
281
:
20120
-20128.
26
Tachibana, M., K. Sugimoto, T. Fukushima, Y. Shinkai.
2001
. Set domain-containing protein, G9a, is a novel lysine-preferring mammalian histone methyltransferase with hyperactivity and specific selectivity to lysines 9 and 27 of histone H3.
J. Biol. Chem.
276
:
25309
-25317.
27
Tachibana, M., J. Ueda, M. Fukuda, N. Takeda, T. Ohta, H. Iwanari, T. Sakihama, T. Kodama, T. Hamakubo, Y. Shinkai.
2005
. Histone methyltransferases G9a and GLP form heteromeric complexes and are both crucial for methylation of euchromatin at H3-K9.
Genes Dev.
19
:
815
-826.
28
Vakoc, C. R., S. A. Mandat, B. A. Olenchock, G. A. Blobel.
2005
. Histone H3 lysine 9 methylation and HP1γ are associated with transcription elongation through mammalian chromatin.
Mol. Cell.
19
:
381
-391.
29
Noma, K., C. D. Allis, S. I. Grewal.
2001
. Transitions in distinct histone H3 methylation patterns at the heterochromatin domain boundaries.
Science
293
:
1150
-1155.
30
Schneider, R., A. J. Bannister, F. A. Myers, A. W. Thorne, C. Crane-Robinson, T. Kouzarides.
2004
. Histone H3 lysine 4 methylation patterns in higher eukaryotic genes.
Nat. Cell. Biol.
6
:
73
-77.