The molecular components of the epigenome have been extensively studied: they include cytosine methylation, the proteins that attach and remove methylation, histones and their modifications, the proteins that modify histones, the proteins that are recruited by histone modifications, and a variety of “chromatin” proteins. The association of DNA with these modifications forms chromosomes, and the modifications present at a given locus control the state of gene expression of the locus. The complexity of epigenetic modifications explains why epigenetic inheritance is so different from Mendelian inheritance. Genetics is based on the faithful replication of a stable molecule, DNA. Epigenetics on the other hand is based on structures containing an assortment of molecules that associate with varying stability. Because the epigenome has different compositions at different sites, each example of epigenetic inheritance displays a different pattern. A locus in a single individual can be described by one or two genetic alleles, but by an undetermined number of “epialleles” created by combinations of molecular accretions to the genome.
From a molecular perspective, an epigenetic variant can be defined as a change in the molecular composition of the epigenome at some site (for example, methylation of a CpG dinucleotide, or gain or loss of a histone modification). Many such molecular variants may be functionally trivial. From a functional perspective an epigenetic variant is a stable change in a transcriptional regulatory element, which changes the expression of a gene without any change in DNA sequence. In our work we use gene expression as the measure of the epigenetic state of a gene: differences in gene expression between isogenic individuals (such as monozygotic twins or isogenic mice) reflect differences in epigenetic states that are independent of genetic variation.
We have developed a mouse model of multigenerational epigenetic inheritance induced by paternal obesity and prediabetes. In our model, a single generation of paternal obesity and prediabetes programs male offspring with a latent metabolic defect that is exposed by overnutrition. The latent phenotype is transmitted through the paternal line for two generations without further exposure to obesity. Paternal transmission is associated with changes in the sperm small RNAs that are predicted to regulate transcriptional processes. This system permits a direct test of the epigenetic inheritance model: genetic variants can be ruled out as a factor in transmission because we study isogenic mice, transmission through the paternal line rules out in utero metabolic exposure as a cause, and the high penetrance of the metabolic phenotype makes it amenable to mechanistic studies. We are particularly interested in determining the full set of epigenetic and transcriptomic changes that occur in the sperm of mice with the latent phenotypes, and how such changes affect early embryonic development. We are also interested in the possibility that somatic changes in the presence of the latent metabolic phenotype are the responsible for changes in the sperm.
The unexpected multigenerational effect of paternal obesity/prediabetes on the metabolism of genetically identical offspring challenges established views on the causes of obesity. Understanding the scope and the mechanism of this heritable epigenetic programming phenomenon will be critical in developing new strategies to manage or prevent the effects of paternal obesity.
The epigenome is interposed between the genome and the environment, and allows a genome to respond to environmental cues. This suggests that the epigenome can be a mechanism of phenotypic response to environmental factors, one that can readily be reversed when such factors change. Under stable environmental conditions, the reference epigenetic state at a locus is likely to be more stable than a variant state; however a change in environmental conditions could perturb the stability profile of the epigenome at a locus, changing the probability of variant states arising, and even making a variant state more stable than the previous reference state. In practice it can be difficult to establish that a specific epigenetic variant is spontaneous, or instead induced by an environmental agent.
We use exposure to ethanol as a model of environmental change. Evidence that exposure to ethanol in parents can have effects on their offspring has raised the possibility that ethanol has effects on the epigenome. We are studying the extent to which epigenetic variants can be induced by ethanol, and transmitted from parents to offspring through multiple generations. We hypothesize that paternal exposure to ethanol before conception induces epigenetic variants that are heritable in the absence of any genetic variation. We ask if paternal preconception ethanol exposure can induce epigenetic variants that persist for a generation past the cessation of ethanol exposure. This work will assess the frequency with which paternal preconception ethanol exposure induces gene expression variants in the offspring of exposed males, and the frequency with which such variants are transmitted to the next generation.
We explored the possibility that the epigenome carries heritable information in one of the first comparative epigenomic studies, where we compared genome-wide cytosine methylation patterns in neutrophils of multiple humans, chimpanzees, and orangutan. Cytosine methylation is an important component of the epigenome; in vertebrates, it is associated with stable silencing of gene expression. Comparative analysis of rates of CG decay in the methylome of human, chimp, and orangutan showed that somatic methylation states reflect methylation states in the germline, and that human-specific methylation changes are also present in the germline. This finding is consistent with the possibility that germline epigenetic states carry heritable phenotypic information; however analysis of outbred populations, such as human and chimp, does not allow to tell if these changes are pure epigenetic variants or if they are encoded by underlying genetic differences. The study of pure epigenetic variation and inheritance requires working with inbred populations such as the mouse.
We are using an isogenic mouse model to determine the number of epigenetic variants that can be transmitted from parents to offspring and the persistence of transmission through multiple generations. We are interested in epigenetic variants that have functional consequences: for this reason we use gene expression variants as the functional readout of epigenetic variation. We study isogenic mice to minimize genetic and environmental confounding factors, and use a single, highly homogeneous cell type purified from mice to minimize artifacts due to cell heterogeneity and culture. We have found that hundreds of gene expression variants are present even when confounding factors are controlled. The long-term goal is to define how many epigenetic variants are transmitted from parents to offspring, how stable is their heritability through multiple generations, and how heritability is affected by selection. The ultimate goal of this work is to produce evidence for or against the view that inherited epigenetic variation can be an important component of disease risk.
Drosophila melanogaster is a species with very low levels of cytosine methylation. We have been able to construct the first map of cytosine methylation in Drosophila using a method that we developed to detection of very low levels of methylation. The function of DNA methylation in Drosophila is still unknown but likely to be very different from its function in other animals: it is very rare, apparently present only during early embryonic development, and associated with sequence motifs that are very different from those found in other animal species. We find that methylation is particularly enriched on chromosome X, suggesting that it might participate in sex chromosome dosage compensation during early development before the Male-Specific Lethal complex becomes active at blastoderm. Several Drosophila species have neo-sex chromosomes that have evolved either partial or full dosage compensation; these species provide an ideal experimental system to further investigate the functional association between methylation and dosage compensation.
Monday, September 19, 2016 11:30 AM