Epigenome-Wide Association Studies and Disease

Epigenetic Mechanisms of Disease: Viable Yellow Mouse Model

The back-translated meaning of epigenetics ( epi, above, upon; genetic, DNA sequence) implies that information encoded in the DNA sequence may be modifiable in some way by higher-order information that regulates the levels of activity of specific genes. Such a concept is attractive when trying to understand why monozygotic twins, who have identical DNA sequences, are sometimes discordant for certain heritable diseases, such as Alzheimer disease and type 1 diabetes mellitus. Genetic predisposition that fails to account fully for the development of a disease (or other) phenotype has been called “missing heritability,” a gap that epigenetic regulatory processes have been proposed to fill. Furthermore, because the environment influences the risk of certain disorders by modifying an underlying genetic predisposition, environmental stimuli may act through epigenetic regulatory processes of gene expression.

The most compelling evidence for the epigenetic, higher-level regulation of genes and predisposition to disease was the viable yellow mouse model ( Fig. 100.1 ). This mouse was found to have a mutation involving an endogenous retrovirus, a component of the genome that can replicate itself and move to a new location. In the case of the viable yellow mouse, the endogenous retrovirus was the type called an intracisternal A particle (IAP) , which inserted upstream from a gene called a ( nonagouti ). The nonagouti gene encodes agouti-signaling protein precursor, which binds to and has a negative effect on melanocortin receptors. When it stimulates melanocytes in hair follicles, it causes the production of the yellow pheomelanin pigment rather than black eumelanin. Without the upstream IAP element, nonagouti would normally switch on for a very short burst of activity and stimulate a limited amount of yellow pigment production. The presence of the active IAP element upstream was found to create a new, constitutively active start site for the nonagouti gene, leading to the hair being produced with pheomelanin throughout its length, and a distinctive yellow fur phenotype. Because the agouti-signaling protein precursor is also expressed in other cell types, the extra activity of the nonagouti gene driven by the IAP element caused the yellow mice to become obese (due to actions on adipocytes), creating a syndrome comparable to human type 2 diabetes mellitus in these animals.

These mice became an intriguing model of a potential epigenetic role in disease risk because of the unexpected observation that pups from the same litter, all containing the same IAP insertion mutation, differed strikingly in their amount of yellow fur and associated adult obesity. Some of the mice had so little yellow fur that they had no visible evidence of having a mutation at all. The IAP element in these littermates was active in the cells of the yellow mice, as expected, but had undergone silencing in the mice with the brown fur. The inactive IAP element was distinctive for having acquired DNA methylation , the modification of cytosines located immediately before guanines (CG or CpG dinucleotides) to 5-methylcytosine. Methylation of cytosines at CG dinucleotides is the default state throughout the genome, but it is usually absent at the sites regulating expression of nearby genes, so its acquisition at these sites indicates that the gene has undergone silencing. This suggested that an influence on how genes are expressed overrode innate genetic susceptibility, modifying the risk of acquiring a disease. Further, researchers modified the diets of mothers pregnant with a litter of pups with the IAP insertion by supplementing folic acid , a single-carbon donor that increased the availability of a cofactor needed for DNA methylation. The outcome was a higher proportion of pups born with DNA methylation and inactivation of the IAP mutation ( Fig. 100.2 ).

Therefore, a reason for the variability in whether the mice developed the yellow fur and obesity could be influences during pregnancy, such as maternal diet. This supported suggestions that intrauterine stresses were associated with increased risks of certain adult conditions, such as cardiovascular, renal, and metabolic diseases. This field of study is often known as the Developmental Origins of Health and Disease (DOHaD) , which asks how someone's cells remember an intrauterine stress years or decades later. The viable yellow mouse model suggested that such memory could be mediated by regulators of gene expression and influenced by environmental factors such as maternal diet during pregnancy.

Epigenetics and Regulation of Gene Expression

Two examples of gene regulation provide a model for locking in a regulatory pattern early in development and maintaining it indefinitely thereafter. The first is X chromosome inactivation . Because males have only 1 X chromosome, it does not undergo inactivation. However, a person with 2 X chromosomes will inactivate 1, a person with trisomy for the X chromosome will inactivate 2, and so on. The result is that males and females have 1 active X chromosome per cell, despite starting with different numbers of X chromosomes, a process referred to as dosage compensation .

X chromosome inactivation is generally a random event, choosing the maternal X for inactivation in half the cells of the body and the paternal X in the other half. The inactivation occurs very early during development, when the blastocyst is implanting itself into the uterine wall. However, once established in this small number of pluripotent cells, the inactivation persists in all the cells of the individual throughout life.

The other relevant model of gene regulation is genomic imprinting (see Chapter 97 ). Gene activation in a specific cell type usually switches on the copies present on both the paternal and the maternal chromosomes. However, an imprinted locus is distinctive because only the copy on the paternal chromosome is switched on for some imprinted genes, while other imprinted genes are distinctive for only switching on the maternal copy. The timing of this inactivation event is even earlier than X chromosome inactivation, occurring during the formation of the male or female gametes. Again, these patterns of inactivation persist throughout life into old age.

Pediatric Diseases Involving Epigenetic Processes

Prime examples of epigenetics are those involving imprinted loci, exemplified by the Prader-Willi and Angelman syndromes (see Chapter 97 ). Each of these syndromes may be caused by the same deletion on chromosome 15, distinguished by the deletion occurring on the paternal chromosome 15 causing Prader-Willi syndrome and the maternal chromosome 15 causing Angelman syndrome . There are imprinted genes located within the 15q11-q13 region, some of which are expressed only on the paternal chromosome, some only on the maternal chromosome. When an individual is missing the region on the paternally inherited chromosome, the person still has a copy of the gene on the remaining maternal chromosome, but if it is silenced by imprinting, the individual effectively has no functional copy of the gene, leading to the Prader-Willi phenotype. The converse happens for Angelman syndrome; a deletion of the maternal chromosome leaves a silenced copy of the gene on the paternal chromosome.

Although deletions cause these syndromes in the majority of affected individuals, a subset results from uniparental disomy (UPD) , in which there are 2 intact chromosomes 15, but both are inherited from 1 parent . Maternal UPD has the same effect as a paternal deletion in that there is no contribution of a paternally inherited chromosome, causing Prader-Willi syndrome, with paternal UPD causing Angelman syndrome. UPD is thought to start with trisomy for that chromosome, with a 2nd event occurring early in development in which 1 of the 3 chromosomes is lost, occasionally leaving 2 chromosomes derived from the same parent. In a further, very small proportion of individuals, mutations within the 15q11-q13 region seem to affect the imprinted domain as a whole.

Prader-Willi and Angelman syndromes occur because of genetic mutations : large deletions, nondisjunction events leading to whole chromosomal gains or losses, or smaller DNA mutations. These mutations reveal the underlying pattern of genomic imprinting, a distinctive organization of gene regulation at chromosome 15q11-q13 that reflects a memory of the gamete of origin of each chromosome, described as epigenetic. What is not occurring in these individuals is an alteration of the normal epigenetic regulation of the locus, as exemplified by the yellow agouti mice. To find examples of altered epigenetic regulation associated with disease, researchers take advantage of assays that studied DNA methylation patterns throughout the genome. If we had never known about the IAP element insertion in the yellow agouti mice, for example, the locus would have revealed itself by having distinctive DNA methylation in the yellow, obese animals compared with the brown, lean, genetically identical littermates. This approach, referred to as an epigenome-wide association study (EWAS) , was initially applied to study individuals who had intrauterine perturbations, environmental exposures, or various types of cancer, to look for cellular reprogramming events.