Can you briefly explain what interindividual epigenetic variation is?
Epigenetics is the study of stable changes to the genome that can influence gene expression, but do not involve changes to the underlying DNA sequence. In our study, we focus on DNA methylation, a major epigenetic mark that can affect gene expression potential and is implicated in various diseases and ageing.
In particular, we looked for genomic regions of interindividual epigenetic variation in which DNA methylation is consistent across various tissues within an individual but shows variability among different individuals. To quantify this we developed a new algorithm called the systemic interindividual variation index (SIVI), and calculated it at more than 6 million regions spanning the human genome.
A high SIVI indicates that one individual has similar DNA methylation status in different tissues in their body, and another individual also has consistent DNA methylation across their tissues, but the level of DNA methylation between the two individuals is different.
What were the key findings from your research?
Using this SIVI screen we identified a gene, VTRNA2-1, as a human ‘metastable epiallele’, and using another independent genome-wide screen we further demonstrated that DNA methylation in infants at this gene is associated with maternal environment around the time of conception.
There is currently a lot of interest in the phenomenon of epigenetic metastability. To be considered a metastable epiallele, a stochastic process must occur prior to gastrulation in the early embryo, leading to consistent or ‘systemic’ methylation across tissues.
Because of the stochastic nature of this process, interindividual variation in methylation at metastable epialleles is determined probabilistically, rather than by genetic differences. To exclude possible genetic influences on methylation at VTRNA2-1 we looked for SNP effects.
SNPs are single nucleotide polymorphisms, or mutations at a single base pair of DNA. The lack of SNP-associated methylation effects around the VTRNA2-1 gene suggests that the systemic interindividual variability in methylation is not mediated by genetics, making VTRNA2-1 a true metastable epiallele. Other genomic regions showing the greatest systemic interindividual variability without evidence of genetic effects were included in a shortlist of candidate metastable epialleles, which should help guide future studies by us and others.
Using biomarkers measured in the blood of mothers of the Gambian infants in the study, we found that low levels of maternal riboflavin (vitamin B2) and methionine around the time of conception corresponded to a higher chance of hypomethylated VTRNA2-1 in her infant, while low levels of dimethylglycine corresponded to a decreased chance of hypomethylation. These biomarkers were generally present at different levels between the two seasons, suggesting nutrition as a likely driver of the season of conception effect.
Why was it important to apply two independent genome wide approaches?
Our genomewide screen in blood and hair follicle DNA from two Caucasians allowed us to search for metastable regions in an unbiased fashion, in other words, without making assumptions about where such regions might be located in the genome.
Our second screen in rural Gambian infants used an entirely different measurement technology and analytical approach to identify genomic regions where DNA methylation is associated with maternal environment around the time of conception.
The fact that these effects are evident in several unrelated ethnic groups indicates that epigenetic metastability at VTRNA2-1 is an ancient feature of the human genome.
For us, the fact that these two independent approaches converged on the same gene, VTRNA2-1, is remarkable, suggesting that DNA methylation status at this gene is established in the very early embryo, and further that this state can be influenced by environmental factors around this time.
Using three other tissues from a different set of samples, we were able to confirm the cross-tissue concordance of DNA methylation at VTRNA2-1. Notably, the two genomewide screens were performed in Caucasians and Gambians, while the validation in the other tissues was performed on samples from an Asian cohort. The fact that these effects are evident in several unrelated ethnic groups indicates that epigenetic metastability at VTRNA2-1 is an ancient feature of the human genome.
Why did you decide to look at variants in rural Gambia?
The climate of The Gambia gives us a unique ‘natural experiment’ for this type of study. There is a wet and a dry season, with markedly different availability of foods, which in turn affects nutrition. The UK Medical Research Council’s International Nutrition Group has a field station in rural Gambia which allows us to study how this seasonality affects establishment of DNA methylation during human development.
Additionally, the longstanding relationship between the Medical Research Council and the local population meant we were able to collect samples from dozens of the same individuals over a ten year period. This allowed us to demonstrate that DNA methylation at VTRNA2-1 is stable from childhood into adulthood, providing further evidence of the potential for the changes we observed to have lifelong effects on health and disease.
It is important to note, though, that since our studies showed that VTRNA2-1 is a metastable epiallele across Africans, Caucasians, and Asians, periconceptional environment likely affects establishment of DNA methylation in various contexts and cultures worldwide.
Were you surprised by any of your findings?
The biggest surprise came from the fact that VTRNA2-1 was identified as the gene exhibiting the strongest systemic interindividual variation in DNA methylation and – independently – had the strongest season of conception effects on DNA methylation.
Since both of these analyses were conducted on a genomewide scale, in different populations, and in search of potentially unrelated effects, we had no idea if any of the top genes from the two studies would overlap.
We were also excited to find that this new metastable epiallele has been identified by others as being associated with tumor suppression, and as having a role in regulating the immune system. Previous metastable epialleles identified by our groups have not previously been associated with phenotypic effects.
VTRNA2-1 is a ‘non-coding’ RNA gene that has been determined to be genomically imprinted. Since we have one copy from each of our parents, it is possible to have one copy turned off through epigenetic mechanisms – a phenomenon known as genomic imprinting.
The results of one study suggest that this gene is polymorphically imprinted – having some of the effects of genomic imprinting but showing variation among individuals. Our season of conception data provides insight into the role of early environment in this polymorphic imprinting.
What are the possible implications?
In this study we were able to show that DNA methylation at VTRNA2-1 is established during a critical window very early in development, is influenced by periconceptional environment, including maternal nutrition status; and that once established, these methylation levels are persistent, lasting over the course of at least ten years.
Accordingly, establishment of DNA methylation at VTRNA2-1 appears to be the first example of ‘metabolic imprinting’, a theoretical construct first proposed in 1999 relating early nutritional exposure to disease later in life. It is exciting to speculate that there may be other genes exhibiting these effects, perhaps caused by other nutritional indicators in early pregnancy or mediated by other epigenetic mechanisms.
This would have important implications for our efforts to understand maternal nutritional factors that might be influencing offspring health through the life course.
- What can we learn from single cells? - 1st June 2016
- Scaling down to single cells - 14th December 2015
- Guidance for the open sharing of epigenome data: an author Q+A - 22nd July 2015