Epigenomics – Understanding the Epigenetic Landscape of Cells

Epigenomics – Understanding the Epigenetic Landscape of Cells

Epigenomics – Understanding the Epigenetic Landscape of Cells F Eckhardt, Inventages, London, UK J Walter, University of Saarland, Saarbrücken, German...

439KB Sizes 1 Downloads 65 Views

Epigenomics – Understanding the Epigenetic Landscape of Cells F Eckhardt, Inventages, London, UK J Walter, University of Saarland, Saarbrücken, Germany Ó 2014 Elsevier Inc. All rights reserved.

Glossary DNA methylation The covalent modification of DNA bases by the addition of a methyl group. Epigenetics The science of gene expression and cellular phenotype that are caused by mechanisms other than changes in the underlying DNA sequence.

Introduction The systematic sequencing of the human and other genomes laid the ground to further our understanding how the genome is organized and what mechanisms are at play that regulate human genes. It has now become clear that there exist multiple layers of information that are encoded beyond the pure sequence of the DNA. Epigenetics is an umbrella term used to describe such mechanisms which introduce heritable changes of gene expression on a cellular or organism level. In contrast to genetic variation, epigenetic variants are not necessarily caused by changes in the respective DNA sequence of a cell or organism. It is a network of histone modifications and various types of DNA methylation that form a cell-specific chromatin structure of the genome and consequently, during development and differentiation, the genome of a cell is converted into a cell-type-specific epigenome. While epigenetics aims to understand heritable, epigenetic changes at single genes, the term epigenomics refers to the comprehensive genome-wide analysis of epigenetic modifications in isolated tissues or cells. Epigenomic data can be read and interpreted by bioinformatic methods as a functional map along the genome which helps to understand the local and global regulation of genes within the genomic context. In addition, epigenomic data can provide information about the developmental and disease state of a cell, the cell-type-specific interpretation of genetic variation and the influence of environment on the cellular physiology. Since the completion of the human genome sequence (Venter et al., 2001; Lander et al., 2001), research on epigenetics and epigenomics has become center stage. This was furthermore boosted with the advent of array and next generation sequencing (NGS) technologies over the last two decades. While in the year 2002, the term ‘epigenomics’ appeared only in 4 abstracts/titles of Pubmed cited articles, this number increased to 77 in year 2012. Several large epigenomic programs have been launched and completed over the years in various plant and animal model organisms. Particularly dedicated to human epigenomes are several large national programs: the NIH Roadmap Epigenome initiative, the EU high impact program BLUEPRINT, the German program DEEP, the Canadian program CEEHRC, the Japanese program CREST, the Korean epigenome program on Metabolic Epigenomes,

Reference Module in Biomedical Research, 3rd edition

Genome The complete set of DNA within a single cell of an organism. Histones A group of proteins found in the nucleus of eukaryotic cells that organize and aggregate DNA into higher-order structures, called nucleosomes.

the Italian program EPIGEN were formed and joined in 2011 to form the International Human Epigenome Consortium (IHEC) (see below). In the present article, we review the current knowledge of the key features of cellular epigenomes and strategies to produce and interpret such data. In our examples, we will particularly focus on DNA methylation mapping.

From Epigenetics to Epigenomics The field of modern epigenetics started in the mid-1980s with the discovery of mechanisms of epigenetic inheritance such as transcriptional gene silencing in plants, genomic imprinting in mammals and plants, dosage compensation in drosophila and mammals, mating-type control in yeast, and position effect variegation in drosophila. The molecular analyses of these effects provided a deep insight in the various molecular layers of chromatin and DNA associated epigenetic modifications and the enzymes controlling these processes. Starting from such mainly gene focused analyses, the field quickly developed tools for genome-wide approaches. Consequently, first epigenomic maps were generated in model organisms such as budding and fission Yeast, Drosophila, and Arabidopsis using tiling array technologies. After the year 2000, first approaches for the analysis of larger genomes such as mouse and human were performed. Numerous technological advances both on the side of enrichment of specific epigenetic modifications and their respective detection were made accompanied by the development of new bioinformatics tools (Bock, 2012). Particularly, the development of chromatin-immunoprecipitation (ChIP)based technologies (see below) currently allows locating up to 40–60 different histone modifications in chromatin. The development of sodium-bisulfite-modified DNA sequencing technologies generated the possibility to directly visualize cytosine modifications at single base resolution. Combined with NGS technologies these molecular approaches allow to obtain high-resolution epigenomic maps. For medical research such maps provide a unique insight into functional changes associated with human diseases. The prerequisite for the human epigenome project was the completion of the human genome as a reference for epigenomic mapping. Recent comparative sequencing of genomes

http://dx.doi.org/10.1016/B978-0-12-801238-3.00009-X

1

2

Epigenomics – Understanding the Epigenetic Landscape of Cells

(e.g., in the 1000 genomes project (Hayden, 2008)) shows many small structural variations in the genome of the human population. The impact of such variations on epigenomic plasticity in specific cells still needs to be clarified. Epigenetic information can be used for epigenetic association studies (EWAS) of common human diseases (Rakyan et al., 2011) and first comparative analyses reveal that epigenomics greatly helps to interpret the functional consequence of some of such genetic variants identified in genetic association studies (GWAS, Maurano et al., 2012).

Basic Principles of Epigenomics Each cell of a multicellular organism has a characteristic cell-type-specific epigenetic profile along its chromosomes reflecting the differential use and expression of genes in a particular cell type and developmental stage. Early epigenetic analyses suggested that the chromatin is packaged in distinct euchromatic and heterochromatic regions defined by a general ‘epigenetic code’ i.e., combinations of histone modifications and DNA methylation that are strictly associated with distinct chromatin domain structures. While this concept is still a good approximation, the view on the combination and the use of histone and DNA modifications in a specific chromatin context has become more complex over time. A number of key modifications (see below) can be associated with both ‘active ¼ euchromatic’ and ‘repressed ¼ heterochromatic’ chromatin regions (ENCODE Project Consortium, 2007). Moreover, the functional assignment of chromatin modifications can even be different in different organisms or are even absent. DNA methylation, for example, plays a key epigenetic role in all higher plants and all vertebrates but is absent in some important model organisms such as Caenorhabditis elegans or Saccharomyces cerevisiae. In S. cerevisiae the same holds for some heterochromatic histone modifications such as H3K9me2/3 modifications. The combination of modifications at particular chromatin regions is often development and cell-type-specific and may change upon cellular aging, differentiation, and tissue position. Particularly, gene-specific epigenetic modifications at promoters and regulatory regions are responding to extracellular signals such as hormones, nutrients, stress, and damage (Jirtle and Tyson, 2013). Hence in addition to providing information about the cell-specific ‘interpretation’ of the unique genome, epigenome information allows to identify responsive regions and even discriminate how (many), when (in time), and where (in the tissue) cells are responding to extrinsic signals and environmental changes. A key for such functional interpretation of comprehensive epigenomic mapping data is the precise knowledge of the transcriptional status across the genome ideally using deep sequencing approaches as these provide the highest resolution.

Key Modifications of the Epigenome and Their Mapping Resolution Histone and DNA methylation constitute the main layers of epigenomic information. The currently identified 140

modifications at various histones form the most complex class of epigenetic modifications. A number of these modifications have been investigated on a genome-wide scale. About 7–10 modifications comprehensively describe most of the functional states along the genome (ENCODE Project Consortium, 2007, 2012). Active genes display a reduced nucleosome occupancy (Lee et al., 2004) and increased histone acetylations and ‘key’ modifications are monomethylations of lysine residue 4 of histone 3 (H3K4me1), trimethylations of the same lysine residue (H3K4me3), and acetylation of lysine residue 4 of histone 3 (H3K4Ac) as marks of open accessible chromatin mostly concentrated in small localized domains around promoters (H3K4me3) and regulatory regions (H3K4me1 and H3K9Ac) (Pokholok et al., 2005; Barksi et al., 2007; Hon et al., 2009). H3K36me3 and the histone variant H2AZ are associated with the transcribed region of the genome and sites of DNA repair, while heterochromatic modifications such as H3K27me3 are located at silent promoters and H3K9me3 covers extended constitutive heterochromatic regions (Barksi et al., 2007). Histone modifications are extremely informative for functional annotating of the genome since all functional changes in gene regulation are associated with histone modification changes. Recent findings by the ENCODE consortium (ENCODE Project Consortium, 2012) even suggest that based on the comparative mapping of 7–10 modifications, a large part of our genome can be assigned as biochemically functional and potentially transcribed. Further comparative analyses and functional studies using knock-down (KD) and knock-out (KO) models will tell us which combinations of chromatin modifications are indeed necessary to maintain a functional chromatin state in a cell type. A number of findings including mutational studies in various cancer genome projects (such as The Cancer Genome Atlas (TCGA) and the International Cancer Genome Consortium (ICGC)) already show that some histone modifying enzymes and chromatin remodeling enzymes are key players in cancer development. These findings are in line with the frequently occurring epigenetic changes in cancer indicating that the deregulation of DNA and histone modifications are crucial in oncogenesis (You and Jones, 2012; Weisenberger, 2014). All current epigenomic technologies for the mapping of chromatin modifications rely on ChIP methods. Combined with high-resolution sequencing, ChIP allows to locate particular histone modifications to rather small regions (ultimately defined by the nucleosome size and position) in the genome. The exact concentration and molecular combination of such modifications on individual nucleosomes still remain beyond resolution. Epigenomic data generated by ChIP therefore represent an average sum of antibody enriched histone modifications across multiple (ideally similarly positioned) nucleosomes. Moreover, the specificity and sensitivity to detect individual histone modifications depends on certain technical procedures such as chromatin fixation and chromatin shearing protocols as well as the use of specific primary and secondary antibodies. Particularly, the quality of the primary antibody defines the relative accuracy of the mapping. Unless strictly standardized (as, e.g., propagated in IHEC recommendations), histone maps produced under different conditions (and antibodies) are only qualitatively comparable to each other. High quality histone maps as, e.g., provided by the ENCODE or the

Epigenomics – Understanding the Epigenetic Landscape of Cells

IHEC consortium are of tremendous value to annotate genome function in a given cell type. Histone modifications are catalyzed by four groups of enzymes; histone acetyltransferases (HATs) and histone deacetylases (HDACs) regulate the acetylation and deacetylation of histones, respectively, while histone methyltransferases (HMTs) and histone demethylases regulate the methylation status of histones. In various cancers such as colon, uterine, lung, and leukemias, numerous mutations affecting both HATs and HDACs have been found (Esteller 2007). HDACs that remove acetyl residues from histones have been implicated in cancer due to their aberrant binding and gene-silencing activity (Ocker and Schneider-Stock, 2007). Additional evidence for their tumorigenic potential comes from the observation that germline mutations of HDACs increase the risk of breast and lung cancers and several HDACs are overexpressed in various cancers (Miremadi et al., 2007). Associated with histones and nucleosome, RNAs are likely to play a major role in shaping the local distribution of modifications in chromatin. Various classes of small and long noncoding RNAs have been found to be necessary for the establishment and/or maintenance of local/regional chromatin structures, e.g., by attracting histone and DNA-modifying (see below) enzymes to certain regions. Long noncoding RNAs (lncRNAs) such as XIST, HOTAIR, and AIR are found to be enriched in silenced heterochromatic regions. It seems likely that more such examples of chromatin structuring RNA will be found in the near future. However, the mapping of chromatin associated RNAs remains a technically challenging effort and only a few examples are well characterized. Hence while the cell- and tissue-specific expression of more and more small and long noncoding RNAs are being discovered, their local assignment as epigenetic components is not yet a reality in epigenomic mapping. In addition, small RNAs in particular are discussed as molecules for transmitting short- and long-term memory effects of epigenetic information and even in a systemic way as shown for plants in posttranscriptional gene silencing (reviewed in Rogers and Chen, 2013). In contrast to the extreme variation of histone modifications, DNA methylation forms a rather simple type of modification covalently linked to the DNA. In mammals/human, DNA methylation almost exclusively occurs at the 50 position of cytosines forming 5-methylcytosine (5mC). In a mammalian somatic cell, about 3–6% of all cytosines are methylated (Esteller, 2005) and 5mC is almost exclusively located within the nucleotide context of a downstream guanosine (abbreviated CpG, Bird, 2002). Recently, work of several groups has shown that 5mC can be further oxidized by Tet (ten-eleven-translocation) enzymes in various steps to 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC) and 5-carboxycytosine (5caC). Among these oxidative modifications, 5hmC shows the highest abundance with 0.1–0.3%, the other two modifications are much less abundant. As 5mC and 5hmC are indistinguishable in conventional bisulfite sequencing, 5hmC has been elusive for quite some time. Novel protocols using additional chemical modification or enzymatic oxidation steps allow to map 5hmC at base resolution (Song et al., 2012). These methods are just beginning to find their way in systematic epigenome analyses. As an alternative, various antibody-based enrichment strategies such as hyMeDIP (Stroud et al., 2011) or GLIB (which

3

combines chemical modifications with antibody capturing, Pastor et al., 2012) have been developed and have already been used. The functional interpretation of such data remain however inconclusive. First reports of genome-wide analysis of 5mC distribution reported that in embryonic stem (ES) cells, 5mC may also occur in the context of mCHG and mCHH (with H being an adenosine, cytosine, or thymine). More detailed studies suggest that such non-CG methylation appears to be strictly linked to methylation at neighboring CpGs on the same DNA strand mediated by the DNA methyltransferases Dnmt3A and Dnmt3B (Arand et al., 2012) suggesting that such non-CpG methylation may be a ‘by-product’ in DNMT3A/B expressing cells such as pluripotent stem cells (Ramsahoye et al., 2000; Lister et al., 2009) and its functional relevance remains unclear. Several findings suggested the presence of other base modifications as well, such as adenine methylation, in mammals. However, the evidence for this is very limited and such modifications are – if existing – mostly beyond detection limits and many data have not been reproduced in independent experiments. The same holds for recent reports on the presence of cytosine DNA methylation in mitochondria. Here the literature is still controversial but the evidence for DNA methylation in mitochondrial DNA is very scarce. Methylated CpG dinucleotides are nonrandomly distributed across the mammalian nuclear genome. Large regions have a rather lower frequency of CpG dinucleotides as statistically predicted. This relative depletion of CpGs in large regions of the genome has been contributed to the fact that methylated cytosines have a higher likeliness to mutate as 5mCs undergo spontaneous deamination which converts methylated cytosines into thymines that are difficult to be recognized by the DNA repair machinery. This discrepancy has lead to an approximately fivefold underrepresentation of CpG dinucleotides within the genome sequence over evolutionary time (Gardiner-Garden and Frommer, 1987) and is also reflected in the higher number of SNPs in CpG positions across the genome. In contrast to these CpG poor regions there are distinctive CpG-rich regions termed CpG islands (CGIs) that display a high content of CpG dinucleotides. Although different definitions for CGIs exist, the original and most frequently used definition by Gardiner-Garden and Frommer (1987) coins a region that is at least 200 bases long, has a (G þ C) content of greater than 50% and a ratio of observed CpG to expected CpG dinucleotides of at least 0.6. Such CGIs span the promoter regions of approximately 60% of all human genes and are located throughout the genome. Depending on the exact definition used there are about 29 000 CGIs in the repeat-masked and about 50 000 in the nonmasked human genome sequence located (International Human Genome Sequencing Consortium, 2001). CGIs are mostly unmethylated although approximately 9% of all CGIs in a given cell type are methylated (Eckhardt et al., 2006). Currently, it is not known how such methylated CGIs escape the mutational pressure arising from spontaneous deamination and maintain their methylation. Thymine DNA gylcosylase along with de novo DNA methyltransferases are currently discussed as potential players (Jacobs and Schär, 2012).

4

Epigenomics – Understanding the Epigenetic Landscape of Cells

Computational analysis has identified a common architectural feature that about 10% of all human genes are located in a head-to-head arrangement with their promoters being less than 1000 bp apart and many of them having a CGI covering their promoters (Trinklein et al., 2004). Expression studies of some of these head-to-head genes and a correlative methylation analysis revealed that the head-to-head genes are coexpressed and their expression level inversely correlates with the methylation status of their promoters (Cortese et al., 2008). In contrast to most CGI containing promoters that are mostly unmethylated, several large-scale studies have found that the majority (>50%) of non-CGI containing promoters are hypermethylated and only a minority (<20%) are nonmethylated with the rest having an intermediary methylation status (Weber et al., 2005; Eckhardt et al., 2006). Comparative analysis of the methylation profiles of different cell types and development stages revealed that the methylation status of many DNA regions is specific for the respective stage and cell type. For example, one study compared the methylation status of different promoter regions on chromosomes 6, 20, and 22 in 12 different tissues and cell types and found about 17% of all promoter regions are being differentially methylated in at least one tissue (Figure 1; Eckhardt et al., 2006). Such regions have been coined as tissue-specific differentially methylated regions (T-DMRs). A recent survey of a data set of 42 whole genome methylomes supports this interpretation on a more comprehensive scale (Ziller et al., 2013). Of interest to note is as well that most of the T-DMRs are located outside of CGIs and T-DMRs within CGIs are strongly underrepresented. DMRS frequently contain single nucleotide polymorphisms associated with cell-type-related diseases (Ziller et al., 2013). Some of these T-DMRs have been shown to correlate inversely with the expression of the cognate gene (Futscher et al., 2002; Song et al., 2005) while for other T-DMRs, no direct correlation to the expression of the cognate gene has been observed. In light of such findings, DNA methylation may only play a permissive role and additional modifications such as specific histone modifications as mentioned above are necessary to control gene expression. Other experiments, primarily in T-cell lymphocytes have shown as well that the role of T-DMRs goes beyond a direct promoter-mediated transcriptional effect as some enhancers and silencers mediate their effect by differential methylation as well (Ansel et al., 2006). A hallmark of many different cancerous cell types are many CGIs that are unmethylated in their healthy cells but show abnormal gains of DNA methylation (Bird, 2002). This aberrant methylation is affecting up to 10% of all CGI containing promoters and is often accompanied by transcriptional repression and loss of gene function and promoter hypermethylation of tumor suppressor genes is often observed in cancers and is likely to drive tumorigenesis (Baylin and Jones, 2011). Consequently, at least one allele of DNA repair genes and genes controlling the cell cycle such as PTEN, BRCA1/2, and CDKNB is frequently hypermethylated (Hatziapostolou and Iliopoulos, 2011). Similar to DNA mutations that frequently affect specific genes responsible for growth initiation and progression such as TP53 and KRAS, some genes are hotspots for epigenetic modifications such as the DNA repair gene O6-methylguanineDNA methyltransferase (MGMT) or the cell cycle regulator gene

CDKN2B. In fact, epigenetic mutations are much more frequent than genetic lesions of the same gene (Chan et al., 2008; Schuebel et al., 2007; Yi et al., 2011) and one study found a four times higher frequency of epigenetic mutations compared to genetic alterations in colon cancer (Baylin and Jones, 2011). MGMT acts as a cellular defense mechanism by removing carcinogen-induced O6-methylguanine adducts that lead, if unrepaired, to G to A transitions. Cells that harbor a hypermethylated MGMT gene and lack therefore a functioning repair mechanism are susceptible to genetic mutation affecting many genes including critical genes such as KRAS or p53. Alkylating agents such as temozolomide induce DNA crosslinks and trigger thereby apoptosis, a feature that makes such agents increasingly used as cancer therapy. Consequently, assessing the methylation status of MGMT helps to identify patients that will benefit from treatment with alkylating agents (Hegi et al., 2005) for some cancer types. Despite such distinct causative example, the role of the several hundred hypermethylated genes within a given cancer type remains elusive. One hypothesis is that the aggregation of different epigenetic mutations in the same signaling pathway may contribute toward the progression of an oncogenic phenotype (Wood et al., 2007). In addition to aberrant promoter hypermethylation of tumor cells, the DNA of malignant cells is hypomethylated on a genome-wide level compared to normal cells. In particular, repetitive elements tend to be heavily hypermethylated in benign cells which is generally accompanied by heterochromatin formation and this condensation contributes to the overall chromosomal integrity by preventing translocation and possibly gene disruption by transposons and other parasitic DNA elements (Walsh et al., 1998; Gaudet et al., 2003; Esteller and Almouzni, 2005; Weisenberger, 2014). Thus, it is likely that the observed genome-wide hypomethylation contributes as well to tumorigenesis by contributing to large-scale chromosomal changes.

Enzymes Regulating DNA Methylation and Removal The presence and genome-wide distribution of DNAmethylation is controlled by three DNA methyltransferases Dnmt1, Dnmt3a, and Dnmt3b. Dnmt1 is the main enzyme responsible for the maintenance (i.e., copying) of DNA methylation in dividing somatic cells. A loss of the maintenance DNA methyltransferase Dnmt1 leads to uncontrolled development and death during early embryogenesis (gastrulation). Dnmt3a and 3b predominantly function as de novo DNA methyltransferases. As for Dnmt1 both Dnmt3a and Dnmt3b are essential for development and play a central role in the generation of novel methylation patterns at early steps of stem cell differentiation. Several recent reports highlight as well the importance of Dnmt3a and Dnmt3b for maintaining DNA methylation in ES-cells (Ficz et al., 2013; Habibi et al., 2013). Moreover, they also show that Dnmt3a and Dnmt3b are responsible for sporadic nonsymmetrical methylation. This nonsymmetrical activity of both enzymes requires Dnmt3L as a cofactor (Arand et al., 2012). In cultivated ES-cells the loss in Dnmt3b expression leads to a dramatic reduction of 5mC in most parts of the genome despite of the presence of an intact

Epigenomics – Understanding the Epigenetic Landscape of Cells

5

Figure 1 An example of the methylome of different primary cells and tissues. The CpG methylation of different CpG positions of different distinct regions are shown and color coded. Several T-DMRs for genes with diverse functions like oncostatin, SMTN, and RNF 185 are shown. Rows present different samples, and are grouped according to tissue or cell type. Columns represent similar CpGs positions in different sample types, the methylation status is color coded (bottom right side) with yellow indicating an unmethylated CpG, blue a methylated CpG position, and shades of green an intermediate methylation status. Printed with permission from Eckhardt, F., Lewin, J., Cortese, R., et al., 2006. DNA methylation profiling of human chromosomes 6, 20 and 22. Nat. Genet. 38, 1378–1385.

and well expressed Dnmt1 highlighting that the level and distribution of DNA methylation is controlled by a complex interplay of all three DNA methyltransferases. Before establishing cell-type-specific de novo DNA methylation the level of DNA methylation is reduced to a low level ‘ground state’ in pluripotent cells of the early embryo. This

suggests the presence of mechanism controlling the reprogramming/demethylation of 5mC in the genome. Recent data suggest that the oxidation of 5mC to 5hmC, 5fC, and 5caC (Kriaucionis and Heintz, 2009; Tahiliani et al., 2009), catalyzed by the family of Tet oxigenases, may play a crucial role in this demethylation and reprogramming process. At very early

6

Epigenomics – Understanding the Epigenetic Landscape of Cells

embryonic stages, 5hmC, 5fC, and 5caC appear to be highly abundant and undergo fast and dynamic changes (Pastor et al., 2013; Ficz et al., 2011). Similar genome-wide processes are observed during early germ cell development but may also occur at specific loci in differentiated cells such as neurons. Several enzymatic scenarios are currently discussed to contribute to these dynamic methylation changes. The oxidative forms may influence the copying of DNA methylation promoting a ‘passive,’ replication-dependent loss of methylation on a genome-wide scale. Other scenarios involve DNA repair coupled mechanisms to directly and locally remove modified bases followed by replacement by unmodified nucleotides (Lepikhov et al., 2010; Seisenberger et al., 2013). So far several methods to identify and localize the various modified nucleotides have been proposed and will provide a new and deeper insight in the DNA methylome.

Technologies The advances in epigenetics and epigenomics have been driven to a large extent by advances in the associated discovery technologies. The discovery of the first acetylation modification of histones in yeast by Nelson in 1982 (Nelson, 1982) has led to a plethora of research of different histone modifications as described previously. While in its earlier days, histone modifications were detected by radioactive labeling, research was greatly propelled by the establishment of chromatin immunoprecipitation and subsequently analysis of the immunoprecipitated DNA. Here, DNA binding proteins are covalently attached to the DNA at their binding site by chemically cross linking them to the DNA and are than immunoprecipitated by specific antibodies. Finally, the crosslinking is reversed and the isolated DNA is detected by either specific polymerase chain reaction (PCR), hybridization on a microarray, or sequencing approaches that allow to map the histone modification to the respective genomic locus (ChIP on chip). Now, there is a plethora of specific antibodies for various modifications such as H3K27 acetlyation or H3K27 methylation available that readily allow the analysis of such highly specific modifications. Technologies for the detection of DNA methylation have seen similar improvements. Although historically, such methods have concentrated on the analysis of the most frequent methylation form 50 methylcytosine, newer technologies allow now as well to detect other modifications such as 50 hydroxylmethylcytosines. These different technologies differ in their precision of assessing DNA methylation, their specificity for 50 -methylcytosines and their throughput. All of these technologies use one of the three principles to detect DNA methylation which makes use of either digestion by methylation-sensitive restriction enzymes, or affinity enrichment by methylation-specific antibodies or proteins, or chemical (bisulfit) conversion protocols and detection of nucleotide modifications by their physical properties (e.g., by mass spectrometry). For an extensive discussion on these technologies the reader is referred to recent excellent reviews (Laird, 2010; Esteller, 2007). Here, we review only the most widely used technologies. Restriction endonucleases are bacterial enzymes that can cut DNA at specific sequences, their recognition sequences. In

bacteria, restriction endonucleases act as a defense mechanism against invading bacteriophages by distinguishing the bacterial DNA by virtue of its methylation profile at the endonuclease recognition site which is different compared to the methylation profile of the parasitic DNA. The first studies used methylationspecific restriction enzymes such as HpaII or MspI to differentiate between methylated and unmethylated DNA followed by gel electrophoresis and subsequent Southern blotting (Kaput and Sneider, 1979) to identify the specific DNA position. In the last decade, such approaches have been scaled-up and applied on a genome-wide scale as well. The first of such approaches was restriction landmark genome scanning that visualized methylation differences by a methylation-specific restriction endonuclease digestion followed by a radioactive labeling of the DNA fragments and a separation by two-dimensional gel electrophoresis (Costello et al., 2000). A variation on the theme of using restriction enzymes to decipher DNA methylation is the more recent differential methylation hybridization approach. Here, a sample DNA is split into two pools with one pool of DNA being digested with a methylation-sensitive restriction endonuclease while the other pool is not digested. Similar to other array hybridization approaches these two samples are subsequently labeled with two different fluorescent dyes and hybridized onto an array (Huang et al., 1999). The relative intensities of the fluorescent hybridization signals are proportional of the methylation level at the interrogated site. Advantages such as a reduced required input DNA, greater flexibility, and a reduced risk for artifacts such as crosshybridization have spurred the current trend away from microarrays toward high-throughput or NGS approaches. Consequently, restriction-enzyme-based methods such as the HELP assay (Oda et al., 2006) have been adapted to sequencing technologies as well. The second approach seeks to enrich methylated DNA regions by antibodies specific for methylated cytosines (called methyl-DNA immunoprecipitation (MeDIP), Weber et al., 2005) or by the use of proteins having a high affinity for methylated DNA such as the methyl-binding protein MECP2 (Cross et al., 1994). After such enrichment, the DNA is analyzed either by sequencing or array-based DNA detection methods. The first application on a genome-wide methylation analysis using methylation-cytosine-specific antibodies used either BAC arrays (Weber et al., 2005) or tilling arrays (Zhang et al., 2006) but more and more, the array-based detection methods are replaced by sequencing approaches. However, array-based approaches have proven their merits when profiling multiple samples for distinct loci (Bibikova et al., 2009). 5mC undergoes spontaneous hydrolytic deamination of 5mC to form thymine resulting in a mismatch of T:G which is difficult to repair. It is estimated that about one-third of all disease causing mutations including SNPs occur at methylated CpG sites. Wang et al. (1980) and Hayatsu (2008) discovered that this deamination can be chemically catalyzed by sodium bisulfite that converts unmethylated cytosines into uracils while leaving methylated cytosines unaffected. Subsequent cloning or amplification techniques convert then the uracil by its base pairing properties to thymine which is now a surrogate for the methylation and is readily analyzed by Sanger

Epigenomics – Understanding the Epigenetic Landscape of Cells

sequencing or other technologies. Frommer and colleagues used this bisulfite conversion approach to first analyze DNA methylation (Frommer et al., 1992) in the early 1990s and since then it has become the workhorse of epigenetic research with further improvements such as the direct quantitative Sanger sequencing (Clark et al., 1994). The introduction of NGS allows to perform genome-wide bisulfite sequencing (whole genome bisulfite sequencing, WGBS). This is the current gold standard technology to monitor and quantify CpG methylation at a single-base resolution on a genome-wide scale. This method is becoming increasingly used but remains rather expensive since genome-wide coverage of >20x is needed for a decent quantification. Meissner et al. introduced reduced representation bisulfite sequencing (RRBS) to reduce sequence complexity and redundancy by size-fractionating of MspI restriction fragments and subsequent bisulfite conversion and sequencing (Meissner et al., 2005). RRBS concentrates on sequencing of CpG rich regions. This targeted approach allows a more inexpensive analysis of potentially epigenetically modified regions by NGS technologies (Boyle et al., 2012). WGBS and RBBS cause technical (PCR-amplification) and bioinformatics (read mapping) challenges as the basecomposition of the bisulfite-treated DNA is generally reduced from four bases (A,G,T,C) to only three (A,G,T) with only methylcytosines being conserved. Several bisulfite-based array-detection methods have been described. The reduced complexity of bisulfite-converted DNA can cause crosshybridizations and decreased hybridization specificity. The combination of methylation detection with primer extension technologies using multiplexed bead arrays provides a fast and reliable semiquantitative method for a comprehensive screen of CpG methylation across the human genome.

The Epigenome and Disease With the advancements of epigenetic research, aberrant epigenetic modifications have been implicated in the etiology of autism, Parkinson’s disease, diabetes, and different autoimmune diseases such as lupus erythematosus, but the show case example remains cancer. Our view of the epigenome of a malignant cell has grown from the evidence that genespecific methylation may silence e.g., tumor suppressor genes and genome-wide hypomethylation may cause genomic instability to a more comprehensive analysis provided by cancer sequencing projects. The comparison between healthy and malignant cells shows that in almost all cancer types, between 5 and 10% of all human genes have an aberrantly upmethylated promoter (Bird, 2002). In some cancers such as breast and colon, there are four times more aberrantly hypermethylated genes than mutated genes (Zeng et al., 2012). Most of these upmethylated genes are scattered throughout the genome but entire subchromosomal bands with multiple, coordinately suppressed genes exist as well (Frigola et al., 2006). Such multiple lesions of the epigenome are likely to be the result of a misregulated epigenetic machinery. Evidence for this view comes from a number of large-scale cancer sequencing projects that cataloged mutations in different cancer types. Many identified few, but high-frequency epigenetic mutations affecting specific genes that are involved in the

7

regulation of the epigenome (Ley et al., 2010; Ernst et al., 2010; Nikoloski et al., 2010; Uno et al., 2002; Jiao et al., 2011; Jones et al., 2010). For example, analysis of the coding genome of diffuse large B-cell lymphoma revealed high-frequency mutations of mixed-lineage leukemia gene 2 (MLL2) that encodes a HMTs (Pasqualucci et al., 2011; Morin et al., 2011). MLL2 promotes gene transcription by methylating the lysine-4 position of histone 3 (H3K4). In the same experiment, other genes directly involved in epigenomic regulation were frequently mutated as well, such as the HATs genes CREBP and EP300 and the HMT gene Enhance of Zeste Homolog 2 (EZH2, Pasqualucci et al., 2011). Further work is required to understand if the majority of the many hypermethylated genes observed in cancer are just ‘bystanders’ affected by dysregulated epigenetic modifiers or if their aggregation in the same signaling pathway helps to drive cells to a malignant phenotype (Wood et al., 2007; Baylin and Jones, 2011). In addition, hypermethylation of genes is only one side of the coin. Recent evidence shows that DNA demethylation enzymes such as the TET family proteins that are crucial for developmental reprogramming are also linked to tumorigenesis (Moran-Crusio et al., 2011). The potential to target and reverse these observed epigenetic alterations in cancer has attracted widespread interest from researchers and drug companies. Two inhibitors (5-azacytidine and 50 -aza-20 -deoxycytidine) of DNA methyltransferases were the first to gain approval by the FDA for myelodysplastic syndromes and hematological malignancies. Both being nucleoside analogs, these drugs inhibit DNA methyltransferases by binding covalently to them. As these drugs are not very specific, side-effects are a major issue of concern. Besides DNA-methyltransferase inhibitors, HDAC inhibitors have made their way into the clinic as well with Vorinostat, valproic acid, and pyroxamide being approved for the treatment of various leukemias and solid tumors (reviewed in Costa, 2010). Targeting other components of chromatin and nucleosome remodeling proteins such as histone methylases is also a focus of intense clinical research but none has yet been approved for routine use (Fiskus et al., 2009; Brueckner et al., 2005; Lara et al., 2009).

Epigenome Projects Publication of the Human Genome in 2000 (Venter et al., 2001; Lander et al., 2001) exemplified the success and the importance of large-scale, multinational, concerted efforts to provide a blue print and a biological resource for research for decades to come. Soon after the genomic layer was decoded, interest turned toward epigenetic layers of genomes and several EU-funded projects were initiated such as the Human Epigenome Project (HEP), the EPIgenetic Treatment of Neoplastic disease project (EPITRON), and the Highthroughput Epigenetic Regulatory Organisation In Chromatin (HEROIC) project. In 2008, the American Association for Cancer Research Human Epigenome Task Force and the European Union, Network of Excellence announced the Alliance for the Human Epigenome and Disease (AHEAD) to provide high-resolution reference epigenome maps (The American Association for

8

Epigenomics – Understanding the Epigenetic Landscape of Cells

Cancer Research Human Epigenome Task Force and the European Union, Network of Excellence, 2008). The goal of this project as it was originally stipulated was to focus on human epigenomes but using as well model organisms to further our understanding of epigenetic regulation. Fostered by AHEAD, the NIH Roadmap Epigenomics Program was initiated in 2008. This NIH program was funded with $200 million for 5 years with the goal to provide about 150 reference epigenomes of key cellular states including stem cells, and proliferating, differentiating, and aging phenotypes (Bae, 2013). As NIH and EU funded programs were initially not coordinated, efforts were started to pursue the analysis of multiple epigenomes as an international collaboration. As a result, the IHEC was launched by seven member organizations/countries (NIH Roadmap Epigenomics, EU BLUEPRINT, Germany, Deutsches Epigenom Program (DEEP), Canadian Institutes for Health Research (CIHR), Japan Science and Technology Agency (JST), Korea KNIH, and Italy). In addition, research organizations from France, UK, and Australia are IHEC member observers supporting the IHEC consortium and may join-in as full members at a later stage. Goal of IHEC is to provide about 1000 epigenomes of approximately 250 human tissues and primary cells using comparable standards and methods over the coming 5 years. IHEC will produce human reference epigenome maps of healthy and diseased states and will address as well how the human epigenome is modified by the environment and how epigenomes of the human population have been modified over past generations. Furthermore, IHEC will establish standardized protocols and provide publically available high quality reference data sets for biomedical research. IHEC complements other large-scale projects such as ENCODE (ENCyclopedia Of DNA Elements) and modENCODE (Model Organism ENCyclopedia Of DNA Elements) that have a focus on identifying the functional sequences in the genome rather than defining epigenetic patterns of such sequences as IHEC aims.

Outlook Besides providing a huge amount of information for basic cell and developmental biology epigenomics will largely contribute to enhance molecular diagnostics in medicine. The ever increasing knowledge on genomics and epigenomics and how they interact and are modified by the environment will prove essential for the development of future therapies of cancer and other complex diseases. The fast development and scaling-up of new sequencing technologies will soon bring this knowledge into daily clinical decision making. In the not too distant future, physician will use a patient’s genome and epigenome information for routine diagnostic workup. The miniaturization and automatization of epigenomic technologies will allow to obtain unprecedented insights into ‘cell-to-cell’ variation in tissues allowing to better monitor progression from healthy to diseased states. In the near future, epigenomics will ‘go’ into three dimensions by integrating the linear epigenomic layers obtained by sequential sequencing with topological data of the intact nucleus. Conformation capture sequencing technologies (‘from 3c’ to ‘5c’) or the dam-ID technologies are emerging as

approaches to provide such comprehensive insights in the nuclear architecture of cells. Complemented with transcriptionfactor-ChIP-Seq data, the combination of epigenomic and topological data will provide deep new insights in the organization of chromatin and the dynamic control of transcription. The complexity of such huge epigenetic and transcriptional data sets will require the development of novel and integrative tools for data handling and data visualization. These challenges are met by bioinformaticians who started to develop fast and comprehensive tools for statistical data analysis, interactive visualization, and systemic modeling. The field of computational epigenomics is one of the fastest growing and most important areas in systems biology.

See also: Aberrant DNA Methylation; Cancer Epigenetics and Identity, Richard Meehan; DNA Methylation Mediating Long-term Genome Responses to the Environment, Moshe Szyf; Dynamics of DNA Methylation Patterns in Human Cells; Epigenetic Impact of Nutrition; Epigenetics: Introduction, Caveats, and Historic Overview; Human Embryonic Stem Cells; Structure and Function of Human DNA Methyltransferases; The Human Epigenome Project: Past, Present, and Future.

References Ansel, K.M., Djuretic, I., Tanasa, B., Rao, A., 2006. Regulation of Th2 differentiation and IL4 locus accessibility. Annu. Rev. Immunol. 24, 607–656. Arand, J., Spieler, D., Karius, T., et al., 2012. In vivo control of CpG and non-CpG DNA methylation by DNA methyltransferases. PLOS Genet. 8, e1002750. Bae, J.B., 2013. Perspectives of international human epigenome consortium. Genomics Inform. 11, 7–14. Barksi, A., Cuddapah, S., Cui, K., et al., 2007. High-Resolution profiling of histone methylations in the human genome. Cell 129, 823–837. Baylin, S.B., Jones, P.A., 2011. A decade of exploring the cancer epigenomebiological and translational implications. Nat. Rev. Cancer 11, 726–734. Bibikova, M., Le, J., Barnes, B., et al., 2009. Genome-wide DNA methylation profiling using Infinium assay. Epigenomics 1, 177–200. Bird, A., 2002. DNA methylation patterns and epigenetic memory. Genes Dev. 16, 6–21. Bock, C., 2012. Analysing and interpreting DNA methylation data. Nat. Rev. Genet. 13, 705–719. Boyle, P., Clement, K., Gu, H., Smith, Z.D., Ziller, M., Fostel, J.L., Holmes, L., Meldrim, J., Kelley, F., Gnirke, A., Meissner, A., 2012. Gel-free multiplexed reduced representation bisulfite sequencing for large-scale DNA methylation profiling. Genome Biol. 13, R92. Brueckner, B., Garcia Boy, R., Siedlecki, P., 2005. Epigenetic reactivation of tumor supressor genes by a novel small-molecule inhibitor of human DNA methyltransferases. Cancer Res. 65, 6305–6311. Chan, T.A., Glockner, S., Yi, J.M., et al., 2008. Convergence of mutation and epigenetic alterations identifies common genes in cancer that predict for poor prognosis. PLOS Med. 5, e114. Clark, S.J., Harrison, J., Paul, C.L., Frommer, M., 1994. High sensitivity mapping of methylated cytosines. Nucleic Acids Res. 22, 2990–2997. Cortese, R., Hartmann, O., Berlin, K., Eckhardt, F., 2008. Correlative gene expression and DNA methylation profiling in lung development nominate new biomarkers in lung cancer. Int. J. Biochem Cell Biol. 40, 1494–1508. Costa, F.F., 2010. Epigenomics in cancer management. Cancer Manage. Res. 2, 255–265. Costello, J.F., Frühwald, M.C., Smiraglia, D.J., et al., 2000. Aberrant CpG-island methylation has non-random and tumor-type-specific patterns. Nat. Genet. 24, 132–138. Cross, S.H., Charlton, J.A., Nan, X., Bird, A.P., 1994. Purification of CpG islands using a methylated DNA binding column. Nat. Genet. 6, 236–244. Eckhardt, F., Lewin, J., Cortese, R., et al., 2006. DNA methylation profiling of human chromosomes 6, 20 and 22. Nat. Genet. 38, 1378–1385.

Epigenomics – Understanding the Epigenetic Landscape of Cells

ENCODE Project Consortium, 2007. Identification and analysis of functional elements in 1% if the human genome by the ENCODE pilot project. Nature 447, 799–816. ENCODE Project Consortium, 2012. An integrated encyclopedia of DNA elements in the human genome. Nature 489, 57–74. Ernst, T., Chase, A.J., Score, J., et al., 2010. Inactivating mutations of the histone methyltransferase gene EZH2 in myeloid disorders. Nat. Genet. 42, 722–726. Esteller, M., 2005. Aberrant DNA methylation as a cancer-including mechanism. Annu. Rev. Pharmacol. Toxicol. 45, 629–656. Esteller, M., Almouzni, G., 2005. How epigenetics integrates nuclear functions. Workshop on epigenetics and chromatin. Transcriptional regulation and beyond. EMBO Rep. 6, 624–628. Esteller, M., 2007. Cancer epigenomics: DNA methylomes and histone-modification maps. Nat. Rev. Genet. 8, 286–298. Ficz, G., Branco, M.R., Seisenberger, S., et al., 2011. Dynamic regulation of 5-hydroxymethylcytosine in mouse ES cells and during differentiation. Nature 473, 398–402. Ficz, G., Hore, T.A., Santos, F., et al., 2013. FGF signaling inhibition in ESCs drives rapid genome-wide demethylation to the epigenetic ground of state of pluripotency. Cell Stem Cell 13, 351–359. Fiskus, W., Wang, Y., Sreekumar, A., et al., 2009. Combined epigenetic therapy with the histone methyltransferases EZH2 inhibitor 3-deazaneplanocin A and the histone deacetylase inhibitor panobinostat against human AML cells. Blood 114, 2733–2743. Frigola, J., Song, J., Stirzaker, C., et al., 2006. Epigenetic remodeling in colorectal cancer results in coordinate gene suppression across an entire chromosome band. Nat. Genet. 38, 540–549. Frommer, M., McDonald, L.E., Millar, D.S., et al., 1992. A genomic sequencing protocol that yields a positive display of 5-methylcytosine residues in individual DNA strands. Proc. Natl. Acad. Sci. U.S.A. 89, 1827–1831. Futscher, B.W., Oshiro, M.M., Wozniak, R.J., et al., 2002. Role for DNA methylation in the control of cell type specific maspin expression. Nat. Genet. 31, 175–179. Gardiner-Garden, M., Frommer, M., 1987. CpG islands in vertebrate genomes. J. Mol. Biol. 196, 261–282. Gaudet, F., Hodgson, J.G., Eden, A., et al., 2003. Induction of tumors in mice by genomic hypomethylation. Science 300, 489–492. Habibi, E., Brinkman, A.B., Arand, J., et al., 2013. Whole-Genome bisulfite sequencing of two distinct interconvertible DNA methylomes of mouse embryonic stem cell. Cell Stem Cell 13, 360–369. Hatziapostolou, M., Iliopoulos, D., 2011. Epigenetic aberrations during oncogenesis. Cell. Mol. Life Sci. 68, 1681–1702. Hayatsu, H., 2008. Discovery of bisulfate-mediated cytosine conversion to uracil, the key reaction for DNA methylation analysis – a personal account. Proc. Jpn. Acad. Ser. B Phys. Biol. Sci. 84, 321–330. Hayden, E.C., 2008. International genome project launched. Nature 451, 378–379. Hegi, M.E., Diserens, A.C., Gorlia, T., et al., 2005. MGMT gene silencing and benefit from temozolomide in glioblastoma. N. Engl. J. Med. 352, 997–1003. Hon, G.C., Hawkins, R.D., Ren, B., 2009. Predictive chromatin signatures in the mammalian genome. Hum. Mol. Genet. 18, 195–201. Huang, T.H., Perry, M.R., Laux, D.E., 1999. Methylation profiling of CpG islands in human breast cancer cells. Hum. Mol. Genet. 8, 459–470. International Human Genome Sequencing Consortium, 2001. Initial sequencing and analysis of the human genome. Nature 409, 860–921. Jacobs, A.L., Schär, P., 2012. DNA glycosylases: in DNA repair and beyond. Chromosoma 121, 1–20. Jiao, Y., Shi, C., Edil, B.H., et al., 2011. DAXX/ATRX, MEN1, and mTOR pathway genes are frequently altered in pancreatic neuroendocrine tumors. Science 331, 1199–1203. Jirtle, R., Tyson F.L. (Eds.), 2013. Environmental Epigenomics in Health and Disease, vol. 1. “Epigenetics and Complex Diseases” in “Epigenetics and Human Health” Feil, R., Noyer-Weidner, M., Walter, J. (Eds), Springer, Heidelberg, New York. Jones, S., Wang, T.L., Shih l.M., et al., 2010. Frequent mutations of chromatin remodeling gene ARID1A in ovarian clear cell carcinoma. Science 330, 228–231. Kaput, J., Sneider, T.W., 1979. Methylation of somatic vs germ cell DNAs analyzed by restriction endonuclease digestions. Nucleic Acid Res. 7, 2303–2322. Kriaucionis, S., Heintz, N., 2009. The nuclear DNA base 5-hydroxymethylcytosine is present in Purkinje neurons and the brain. Science 324, 929–930. Laird, P.W., 2010. Principles and challenges of genome-wide DNA methylation analysis. Nat. Rev. Genet. 11, 191–203. Lander, E.S., Linton, L.M., Birren, B., et al., 2001. Initial sequencing and analysis of the human genome. Nature 409, 860–921. Lara, E., Mai, A., Calvanese, V., et al., 2009. Salermide, a Sirtuin inhibitor with a strong cancer specific proapoptotic effect. Oncogene 28, 781–791.

9

Lee, C.K., Shibata, Y., Rao, B., Strahl, B.D., Lieb, J.D., 2004. Evidence for nucleosome depletion at active regulatory regions genome-wide. Nat. Genet. 36, 900–905. Lepikhov, K., Wossidlo, M., Arand, J., Walter, J., 2010. DNA methylation reprogramming and DNA repair in the mouse zygote. Int. J. Dev. Biol. 54, 1565–1574. Ley, T.J., Ding, L., Walter, M.J., et al., 2010. DNMT3A mutations in acute myeloid leukemia. N. Engl. J. Med. 363, 2424–2433. Lister, R., Pelizzola, M., Dowen, R.H., et al., 2009. Human DNA methylomes at base resolution show widespread epigenomic differences. Nature 462, 315–322. Maurano, M.T., Humbert, R., Rynes, E., et al., 2012. Systematic localization of common disease-associated variation in regulatory DNA. Science 337, 1190–1195. Meissner, A., Gnirke, A., Bell, G.W., et al., 2005. Reduced representation bisufite sequencing for comparative high-resolution DNA methylation analysis. Nucleic Acid Res. 33, 5868–6877. Miremadi, A., Oestergaard, M.Z., Pharoah, P.D., Caldas, C., 2007. Cancer genetics of epigenetic genes. Hum. Mol. Genet. 16, R28–R49. Moran-Crusio, K., Reavie, L., Shih, A., et al., 2011. Tet2 loss leads to increased hematopoietic stem cell self-renewal and myeloid transformation. Cancer Cell 20, 11–24. Morin, R.D., Mendez-Lago, M., Mungall, A.J., et al., 2011. Frequent mutation of histone-modifying genes in non-Hodkin lymphoma. Nature 476, 298–303. Nelson, D.A., 1982. Histone acetylation in baker’s yeast. Maintenance of the hyperacetylated configuration in log phase protoplasts. J. Biol. Chem. 257, 1565–1568. Nikoloski, G., Langemeiher, S.M., Kuiper, R.P., et al., 2010. Somatic mutations of the histone methyltransferases gene EZH2 in myelodysplastic syndromes. Nat. Genet. 42, 665–667. Ocker, M., Schneider-Stock, R., 2007. Histone deacetylase inhibitors: signaling towards p21cip1/waf1. Int. J. Biochem. Cell Biol. 39, 1367–1374. Oda, M., Glass, J.L., Thompson, R.F., et al., 2006. High-resolution genome-wide cytosine methylation profiling: the HELP assay. Genome Res. 16, 1046–1055. Pasqualucci, L., Terifonov, V., Fabbri, G., et al., 2011. Analysis of the coding genome of diffuse large B-cell lymphoma. Nat. Genet. 43, 830–837. Pastor, W.A., Huang, Y., Henderson, H.R., Agarwal, S., Rao, A., 2012. The GLIB technique for genome-wide mapping of 5-hydroxymethylcytosine. Nat. Protoc. 7, 1909–1917. Pastor, W.A., Pape, U.J., Huang, Y., et al., 2013. Genome-wide mapping of 5-hydroxymethylcytosine in embryonic stem cells. Nature 473, 394–397. Pokholok, D.K., Harbison, C.T., Levine, S., et al., 2005. Genome-wide map of nucleosome acetylation and methylation in yeast. Cell 122, 517–527. Rakyan, V.K., Down, T.A., Balding, D.J., Beck, S., 2011. Epigenome-wide association studies for common human diseases. Nat. Rev. Genet. 12, 529–541. Ramsahoye, B.H., Biniszkiewics, D., Lyko, F., et al., 2000. Non-CpG methylation is prevalent in embryonic stem cells and may be mediated by DNA methyltansferase 3a. Proc. Natl. Acad. Sci. U.S.A. 97, 5237–5242. Rogers, K., Chen, X., 2013. Biogenesis, Turnover, and Mode of Action of plant MicroRNAs. Plant Cell 25, 2383–2399. Schuebel, K.E., Chen, W., Cope, L., et al., 2007. Comparing the DNA hypermethylome with gene mutations in human colorectal cancer. Plos Genet. 3, 1709–1723. Seisenberger, S., Peat, J.R., Reik, W., 2013. Conceptual links between DNA methylation reprogramming in the early embryo and primordial germ cells. Curr. Opin. Cell Biol. 25, 281–288. Song, C.H., Yi, C., He, C., 2012. Mapping recently identified nucleotide variants in the genome and transcriptome. Nat. Biotechnol. 30, 1107–1116. Song, F., Smith, J.F., Kimura, M.T., et al., 2005. Association of tissue-specific differentially methylated regions (TDMs) with differential gene expression. Proc. Natl. Acad. Sci. U.S.A. 102, 3336–3341. Stroud, H., Feng, S., Morey Kinney, S., Pradhan, S., Jacobsen, S.E., 2011. 5-Hydroxymethylcytosine is associated with enhancers and gene bodies in human embryonic stem cells. Genome Biol. 12, R54. Tahiliani, M., Koh, K.P., Shen, Z., et al., 2009. Conversion of 5-methylcytosine to 5-hydroxymethylcytosine in mammalian DNA by MLL partner TET1. Science 324, 930–935. The American Association for Cancer Research Human Epigenome Task Force and the European Union, Network of Excellence, 2008. Moving AHEAD with an international human epigenome project. Nature 454, 711–715. Trinklein, N.D., Aldred, S.F., Hartman, S.J., et al., 2004. An abund bidirectional promoters in the human genome. Genome Res. 14, 62–66. Uno, K., Takita, J., Yokomori, K., et al., 2002. Aberrations of the hSNF5/INI1 gene are restricted to malignant rhabdoid tumor or atypical teratoid/rhabdoid tumors in pediatric solid tumors. Genes Chromosomes Cancer 34, 33–41. Venter, J.C., Adams, M.D., Myers, E.W., et al., 2001. The sequence of the human genome. Science 291, 1304–1351.

10

Epigenomics – Understanding the Epigenetic Landscape of Cells

Walsh, C.P., Chailet, J.R., Bestor, T.H., 1998. Transcription of IAP endogenous retroviruses is constrained by cytosine methylation. Nat. Genet. 20, 116–117. Wang, R.Y., Gehrke, C.W., Ehrlich, M., 1980. Comparison of bisulfite modification of 5-methyldeoxycytidine and deoxycytidine residues. Nucleic Acid. Res. 8, 4777–4790. Weber, M., Daviess, J.J., Wittig, D., et al., 2005. Chromosome-wide and promoterspecific analyses identify sites of differential DNA methylation in normal and transformed human cells. Nat. Genet. 37, 853–862. Weisenberger, D.J., 2014. Characterizing DNA methylation alterations from the cancer genome atlas. J. Clin. Invest. 124, 17–23. Wood, L.D., Parsons, D.W., Jones, S., et al., 2007. The genomic landscapes of human breast and colorectal cancers. Science 318, 1108–1113. Yi, J.M., Dhir, M., Van Neste, L., et al., 2011. Genomic and epigenomic integration identifies a prognostic signature in colon cancer. Clin. Cancer Res. 17, 1535–1545. You, J.S., Jones, P., 2012. Cancer genetics and epigenetics: two sides of the same coin? Cancer Cell 10, 9–20. Zeng, J., Konopka, G., Hunt, B.G., et al., 2012. Divergent whole-genome methylation maps of human and chimpanzee brains reveal epigenetic basis of human regulatory evolution. Am. J. Hum. Genet. 91, 455–465. Zhang, X., Yazaki, J., Sundaresan, A., et al., 2006. Genome-wide high resolution mapping and functional analysis of DNA methylation in Arabidopsis. Cell 126, 1189–1201. Ziller, M.J., Gu, H., Müller, F., et al., 2013. Charting a dynamic DNA methylation landscape of the human genome. Nature 500, 477–481.

Relevant Websites www.1000genomes.org – A Deep Catalogue of Human Genetic Variation. http://cancergenome.nih.gov – The Cancer Genome Atlas. www.epitron.eu – EU Funded Project EPITRON (EPIgenetic Treatment of Neoplastic disease). http://www.genome.gov – The ENCODE (The ENCyclopedia of DNA Elements) Project. http://genome.ucsc.edu/ENCODE/ – Genomebrowser for the ENCODE Project. http://icgc.org – International Cancer Genome Consortium. http://ihec-epigenomes.org/ – The International Human Epigenome Consortium (IHEC). http://www.modencode.org – The modENCODE (Model Organism ENCyclopedia of DNA Elements) Project. http://projects.ensembl.org/heroic/ – EU Funded Project HEROIC (High-throuphput Epigenetic Regulatory Organisation in Chromatin).