Published online 24 June 2008 | Nature | doi:10.1038/news.2008.913


Lifetime lessons of DNA change

Study of how genome is chemically altered as we age could help us understand disease.

familyDifferent families have different 'styles' of lifelong DNA change.Getty

The chemical markers on people's DNA change throughout their life, an international team of researchers has found.

And the way they change – that is, whether markers are gained or lost — runs in the family, showing that these ‘epigenetic’ features, although not inherited in the conventional, strictly genetic way, are influenced by our genetic make-up.

Researchers led by Andrew Feinberg of Johns Hopkins University in Baltimore looked at changes in the 'methylation' of people's DNA over time. Attaching methyl groups – which contain one carbon atom – to particular DNA bases can influence a gene’s activity.

Epigenetic changes — potentially reversible changes that do not alter the actual nucleotide sequence of the DNA but do affect its functioning — are thought to be involved in many cancers and other diseases of old age, such as diabetes.

This study is a step towards linking such changes with diseases, says Feinberg. “The question of whether DNA methylation changes over a lifetime hadn’t been asked before,” he says. “A fairly large number of people thought it didn’t change.”

All change

To address the question, the researchers analysed samples from 111 Icelandic men and women collected first in the early 1990s, with a second sample taken, 11 years later on average. By the time of the second sample, the subjects were aged 69 and over. The results are reported in the medical journal JAMA1.

The team measured the level of methylation over the whole genome, using two enzymes that cut a specific sequence of four DNA letters – CCGG. One enzyme cuts this sequence only when the second ‘C’ has a methyl group attached, the other cuts it regardless. So the difference between the number of cuts made by each enzyme is a measure of methylation.

Nearly two-thirds of the Icelanders showed at least a 5% change in the amount of methylation, and in nearly one-third the change was more than 10%. People were equally likely to show more or less methylation: the range of the change was from -30% to +26%.

Changes in methylation could influence the development of diseases that are influenced by the complex action of many genes. “This may explain why some people don’t get disease, but we don’t know,” says team member Vilmundur Gudnason, of the Icelandic Heart Association in Kopavogur.

Nor do the researchers know why the amount of methylation goes down or goes up, or how this relates to disease. They plan to address this last question by comparing the changes in methylation with the individual's health.

Family clusters

The team repeated the analysis using samples collected from 126 individuals from 21 families living in Utah in the United States. The average interval between samples was 16 years. As well as looking at methylation across the genome, they used a different test to analyse changes in 807 specific genes.

The results were similar: two-fifths of the individuals showed a change in methylation of at least 5%, with a range of change from -49% to +39%.

In the Utah sample, members of the same family tended to show the same direction of change. “We were surprised to see such clusters,” says Feinberg.

The four individuals who showed the biggest loss of methylation all belonged to one family. “It would be worth studying the genetics of this family on its own,” says James Flanagan, who studies epigenetic influences on cancer at University College London, UK.


But Flanagan does not think that this study resolves the question of how methylation changes over time. “It’s a difficult question to answer,” says Flanagan. “This adds fuel to the fire, but I’m not sure it answers it.”

The fact that change went in all directions suggests that a person’s biological make-up is a bigger influence on epigenetic change than their age, he says. And comparing individuals is not the whole story, as there is as much variation within individuals as there is between them. 

  • References

    1. Bjornsson, H. T. et al. J. Am. Med. Assoc. 299, 2877–2883 (2008).
Commenting is now closed.