Epigenomics in Health and Disease

Mario Fraga, Agustin Fernandez Fernandez

• Publisher: Academic Press
• Year: 2015
• Print ISBN: 9780128001400
• eBook ISBN: 9780128004968
• Topic: Biological Sciences
• Subtopic: Genetics & Genomics
• Index: Biological Sciences

Epigenomics in Health and Disease discusses the next generation sequencing technologies shaping our current knowledge with regards to the role of epigenetics in normal development, aging, and disease. It includes the consequences for diagnostics, prognostics, and disease-based therapies made possible by the study of the complete set of epigenetic modifications to the genetic material of human cells. With coverage pertinent to both basic biology and translational research, the book will be of particular interest for medical and bioscience researchers and students seeking current translational knowledge in epigenesis and epigenomics. Coverage includes the latest findings on epigenome-wide research in disease-based profiling, epidemiological implications, epigenome-wide epigenetic studies, the cancer epigenome, and other pervasive disease categories. - Presents critical reviews that provide the means for reviewing and analyzing the epigenome as a whole, also discussing its translational potential - Combines basic epigenomic knowledge with methodological and biostatistical topics related to technology and data analysis - Includes coverage of relatively new topics, including DNA methylation dynamics during development and differentiation, genome-wide histone post-translational modifications during development and differentiation, and genome-wide DNA methylation changes during aging

Chapter 1

The Role of the Genetic Code in the DNA Methylation Landscape Formation

Holger Heyn, Cancer Epigenetics and Biology Program, Bellvitge Biomedical Research Institute, Barcelona, Catalonia, Spain

Abstract

The genetic and epigenetic codes are highly interconnected, and the understanding of causalities within this cross talk guides our understanding of the biology of healthy life and diseases. Particularly, epigenetic regulation helps interpret genetic variants, and in return functional variance in the epigenetic landscape can be identified through its relation to the genetic background. Epigenetic mechanisms, particularly DNA methylation, shaped the genetic sequence during evolution, and the intrinsic properties of the DNA have impact on the formation of the DNA methylation landscape. Here it is the complex process of gene regulation that is actively controlled by differential DNA methylation patterns. In contrast, transcriptional regulation, such as transcription factors binding to genetic consensus sequences, leaves its marks in the DNA methylome. Elucidating consequences of genetic variance and particularly its impact on epigenetic regulation provides important insights into phenotype formation in natural variation and pathologic contexts, including disease risk.

Keywords

Epigenetics; genetics; DNA methylation; transcription factor; genome-wide association studies; methylation quantitative trait loci; cancer risk; meQTL; GWAS

Chapter Outline

1.1 Bringing the Genetic Code to Life 1

1.2 Intrinsic Properties of DNA 3

1.3 Sequence-Pattern-Dependent DNA Methylation Profiles 5

1.4 DNA-Binding Factors Shaping the Epigenetic Landscape 6

1.5 The Dynamics of Transcription-Factor-Mediated DNA Hypomethylation 8

1.6 Translational Potential of Demethylation Dynamics 9

1.7 DNA Methylation Quantitative Trait Loci 9

1.8 Genotype-Driven Variance in Human DNA Methylation Profiles 10

1.9 Epigenetic Mediator Function for Human Risk Phenotype Formation 11

1.10 Epitype Association Guiding the Interpretation of Cancer Risk Polymorphisms 12

1.11 Closing Remarks 13

References 14

1.1 Bringing the Genetic Code to Life

At the beginning of this millennium, the human genome was decoded [1,2]. The sequencing of the major fraction of the genetic sequence and the resulting first version of an assembled human reference genome displayed an important step toward the understanding of cellular biology and human life. Using the then-current knowledge of genomic organization and sequence characteristics, the genetic sequence was estimated to be roughly 1% coding and the majority of the genome to be noncoding, from which a major part was considered “junk” DNA [3]. However, realizing that the human genome differs only marginally between individuals and is even conserved to around 99% between humans and chimpanzees [4], the noncoding part of the genome was quickly recognized to contribute to phenotype formation and interindividual and interspecies differences.

To illustrate the impact of gene regulation on phenotype formation, the human body presents the best paradigm. Considering the fact that every single cell in the human body carries exactly the same genetic information, it appears remarkable that cells can differ to such a large extent in appearance and function. The difference between a pancreatic cell that produces insulin and a hematopoietic stem cell, giving rise to numerous specialized blood cell types, does not lie within their genetic sequence but is grounded in the activity of the genome. A distinct set of active and repressed genomic regions defines intraindividual differences between cell phenotypes but is also responsible for variation between human individuals (interindividual) and across species (interspecies). Following the sequencing of more than 1000 human genomes [5] and the analyses of variable genetic loci in thousands of individuals in genome-wide association studies (GWAS) [6], it also became apparent that the majority of human phenotypes (traits), ranging from the color of the hair to the susceptibility to develop breast cancer, is encoded in the noncoding fraction of the genome.

Although the protein-coding fraction of the genome, including transcription and translation start sites, was well annotated, the knowledge of the noncoding genome was grounded in annotations based on distance to known genes (promoter regions) or single examples of cis-acting regulatory regions (enhancers). As the function of the genome outside the coding context could not be inferred by the sequence itself, efforts were focused on the comprehensive annotation of the noncoding part of the genome that was no longer considered silent or “junk,” but rather as important driver for phenotype formation. Here seminal contribution was made by the work performed within the ENCyclopedia Of DNA Elements (ENCODE) project, which, after annotating selected parts of the genome in a pilot phase (1% of the genome) [7], presented a comprehensive catalog of regulatory elements within the human genome and led to different estimates as to the extent of active regions in the human genome [8]. However, although the features tested to determine functionality (ranging from transcription activity to DNA factor binding occupancy and chromosomal conformation) covered a wide range of mechanisms involved in known gene regulatory processes, the tissue specificity of the “regulome” remained a major challenge to drawing a general conclusion. Also, the studies used cell lines from single individuals and hence could not address interindividual differences, and results could have been influenced by biases introduced by cell culture. However, due to its extensive number of performed experiments and width of analyzed features, the resulting annotation of the human genome presented a highly valuable insight into the actual activity of the genetic sequence. Importantly, integrated analysis of the wealth of data sets produced by ENCODE enabled the segmentation of the genome into distinct regulatory elements [9–11]. Specifically, active regulatory regions were annotated as promoter, enhancer, or insulator regions, wherein transcriptional activity was determined as a measure for actively transcribed regions that mark putative genic loci or those for which transcription contributes to the regulatory process (enhancers). In addition, transcriptional silent regions or in those carrying actively repressing marks were annotated in repressed or silent segments.

Segmentation algorithm are extremely powerful to summarize the data produced by multidimensional experiments related to gene regulation and enable an estimate of the activity and function of a given region of interest. However, the diversity of regulatory processes and their individual particularities can only be assessed using in-depth analysis and, importantly, large sample cohorts and particular tailored study designs. This chapter focuses on a regulatory mechanism directly affecting the genetic sequence: DNA methylation. DNA methylation involves the covalent modification of DNA, converting cytosine in 5-methyl-cytosine (5-mC). As part of epigenetic regulatory processes, DNA methylation was shown to be stably inherited throughout cell division and to be crucial in developmental and differentiation processes [12]. However, DNA methylation represents the oldest described epigenetic mechanism and its role in gene regulation is still subject of discussion. In general, high levels of 5-mC at proximal promoters are associated to transcriptional silencing; however, outside these regions, its role is highly dependent on the genomic context [13]. For example, increased 5-mC levels within the gene body were related to an elevated transcriptional activity [14].

Similar to the controversial function of DNA methylation, the question of which elements determine the temporal and spatial distribution of 5-mC within the genome remains unresolved. Although the mechanisms covalently modifying the genetic code are clearly defined, with DNA methyltransferases catalyzing the addition of the methyl-group to cytosine, the driving forces guiding the activity of such enzymes is under intense discussion, and putative causalities are summarized in this chapter.

1.2 Intrinsic Properties of DNA

Taking into account that DNA methylation varies among different cell types within the same individual but is relatively stable in the same tissue of different individuals, the genetic code appears to play an inferior role in the variable nature of DNA methylation profiles. However, there are several examples clearly displaying how the genetic sequence is necessary and sufficient for the formation of the DNA methylation landscape [15]. Here, the density of CpG dinucleotides presents the most prominent example of a genetic feature directly associated to the DNA methylation status of the respective regions [16]. In particular, an elevated CpG density was related to an unmethylated (hypomethylated) status of DNA, whereas CpG-poor regions are mainly highly methylated (hypermethylated). Intriguingly, it was the DNA methylation itself that was shaping the distribution of CpG dinucleotides in the genome throughout evolution. Methylated cytosines tend to be mutated by deamination, which results in the formation of uracil and the incorporation of thymine after cell division (without DNA repair). As these transition mutations mainly affect methylated cytosine and almost exclusively occur in a CpG context, this specific dinucleotide was depleted in the genome during evolution. Consequently, the proportion of CpG dinucleotide within the human genome is far less than would be expected if all dinucleotides would underlie equal evolutionary pressures. Moreover, because hypomethylated CpG sites, a frequent feature of promoter regions, were protected from deamination processes, it led to their enrichment in promoters of constitutively transcribed genes and the formation of CpG clusters or CpG islands. However, CpG islands being located outside proximal promoter regions and a specific set of islands being mainly hypermethylated suggest that additional features shape the human DNA methylome [17].

One of these features was discovered as an intrinsic property of the DNA molecule itself, and such parameters as DNA stiffness or flexibility enlarged our understanding of the code encoded by the one-dimensional succession of nucleotides. Sophisticated analysis and computational modeling of DNA molecule properties resulted in the redefinition of the promoter signature and supported a hypothesis of an ancient regulatory mechanism encoded by the intrinsic physical properties of DNA [18]. Computational strategies, linking the DNA sequence to epigenetic mechanisms, consistently support an intrinsic association between the genome and epigenome. Here the cumulative effects of many weak interactions are thought to guide the epigenetic code and to present the blueprint for subsequent formation processes [19]. In this regard, the epigenetic landscapes could be partially predicted computational models and remarkably could be assembled in vitro, suggesting DNA methylation profiles to be partially encoded in the genetic code itself [20].

Interestingly, DNA methylation was described to be capable of altering physical DNA properties and hence modify downstream regulatory processes by altering its flexibility [21]. Specifically, molecular dynamics simulations observed that the high flexibility of the CpG sites is not extrapolated in CpG-dense regions. In contrast, poly-CpG fragments are hardly distinguishable from equally sized dinucleotides in terms of global unwinding and isotropic bending. Importantly, it is the covalent modification of cytosine in CpG-dense regions that dramatically alters the winding properties of DNA, which are particularly important, when DNA is wrapped around nucleosomes, an additional layer in the epigenetic code. Here, DNA methylation changes the winding properties of the DNA by making it stiffer. This was shown to be particularly important in CpG islands, wherein the altered flexibility of the DNA molecule changed the affinity of DNA to assemble into nucleosomes. As nucleosome positioning presents an important feature of transcriptional regulation at transcription start sites, differential DNA methylation can indirectly influence epigenetic regulation through its impact of the physical properties of the genetic sequence.

In conclusion, there is evidence of a direct influence of the DNA sequence on the spatial formation of DNA methylation profiles. These intrinsic properties of the genome on the epige...

Table of contents

• Cover image
• Title page
• Table of Contents
• Copyright
• List of Contributors
• Preface
• Chapter 1. The Role of the Genetic Code in the DNA Methylation Landscape Formation
• Chapter 2. DNA Methylation Microarrays
• Chapter 3. Ultra-Deep Sequencing of Bisulfite-Modified DNA
• Chapter 4. Bioinformatics Tools in Epigenomics Studies
• Chapter 5. Noncoding RNA Regulation of Health and Disease
• Chapter 6. Genome-Wide DNA Methylation Changes During Aging
• Chapter 7. The Dynamics of Histone Modifications During Aging
• Chapter 8. Epigenomic Studies in Epidemiology
• Chapter 9. The DNA Methylomes of Cancer
• Chapter 10. Genome-Wide Epigenetic Studies in Neurologic Diseases
• Chapter 11. Epigenetic Deregulation in Autoimmune Disease
• Chapter 12. Genome-Wide DNA and Histone Modification Studies in Metabolic Disease
• Chapter 13. Clinical Applications of Epigenomics
• Index