In Search of a Better DNA Methylation Biomarker of Aging

In the open access paper I'll point out today, the authors dig into some of the details of DNA methylation changes that occur with aging, seeking to build a better biomarker of aging. This methylation is one of the epigenetic decorations to DNA that act to alter the expression of particular genes, determining whether or not the encoded proteins are produced. The methylation status of genes changes constantly in response to circumstances, differently in every tissue, one small portion of the countless interacting feedback loops that drive the behavior of cells. The cell and tissue damage of aging is the same for everyone, however, and so are the reactions to that damage, even though happenstance, lifestyle choices, and genes conspire to create some variation in the pace at which aging progresses. Thus there are patterns of DNA methylation that are distinctive for people at a given point in the progression of degenerative aging, and those patterns can be picked out of the constant changes that occur due to other environmental factors. Some of these patterns better reflect chronological age, others better reflect biological age, but there is much left to be done to expand and improve upon the existing discoveries in this field.

Why is this important? Primarily for economic reasons. At present it costs a great deal of time and money to assess whether or not a potential therapy that might produce a slowing or reversal of aging in fact works. Researchers have to run life span studies, and as the focus moves from lower animals into mice the cost per study rises to millions of dollars and the time taken rises to years. Yet without those studies, obtaining the proof and support to justify further development is impossible. You might look at the progress towards senescent cell clearance, one of the SENS approaches to rejuvenation biotechnology, over the past decade as an example. Despite the established body of evidence for the role of senescent cells in aging, that line of research didn't start to pull in meaningful support until researchers managed, against all the odds, to raise enough funding to run a study in mice and demonstrate extended life span through removal of senescent cells. Now, five years after those results were published, there are funded startups and numerous research groups working on building a variety of senescent cell clearance therapies. Looking ahead for the field of aging research as a whole, imagine that these lengthy and expensive mouse life span studies could be replaced with very short studies that assess a biomarker of aging, apply the therapy, wait a few weeks, then assess the biomarker again. That would dramatically reduce the cost, get many more research groups into the field, and allow many more approaches to be proved or disproved fairly rapidly. The iterative process of research and development would speed up considerably.

So, to the degree that DNA methylation is a path to a good biomarker of biological age, one that will change quickly and predictably when a real, actual, working rejuvenation therapy is applied, we should all be cheering progress in DNA methylation research. The present DNA methylation clocks are not as accurate as researchers would like them to be, however. There is definitely room for improvement, and all such improvement will - in the end - be reflected in the bottom line: the cost of running studies to assess potential treatments for aging, and thus the cost of progress in the treatment of aging as a whole. Greater reductions in costs will bring larger increases in the output of the research and development communities. The more progress here the better, as no-one is getting any younger yet. All of that said, I should note that the publicity materials here make what is to my eyes a complete hash of the meaning and significance of this research, so you might want to just skip straight to the paper.

Youthful DNA in old age

The DNA of young people is regulated to express the right genes at the right time. With the passing of years, the regulation of the DNA gradually gets disrupted, which is an important cause of ageing. A study of over 3,000 people shows that this is not true for everyone: there are people whose DNA appears youthful despite their advanced years. The researchers charted the regulation of the DNA of over 3,000 people by measuring the level of methylation at close to half a million sites across the human DNA. They were looking for sites where the difference in regulation increased between people as life progressed. Unexpectedly, these sites were closely linked to the activity of genes that were known from studies in worms and mice to play a central role in the ageing process. Not everyone in the study showed equal evidence of an age-related dysregulation of the DNA. Some elderly people had DNA that was regulated as if they were still 25 years old. In these individuals, genes characteristic of the ageing process were much less active. The next step will be to find out whether such people stay healthier for longer. "Obviously, health depends on more than just the regulation of our DNA. But we do think that the dysregulation of the DNA is a fundamental process that could push the risk of different diseases in the wrong direction. In cancer cells, we found changes in the regulation of the DNA at the same sites as if the differences occurring with ageing were a precursor of the disease. We therefore want to study whether a dysregulated DNA increases the risk of different forms of cancer and, conversely, a "youthful" DNA is protective."

Age-related accrual of methylomic variability is linked to fundamental ageing mechanisms

Studies of model organisms such as yeast, nematodes, and mice have shown that the accumulation of cellular damage is a fundamental cause of ageing across species. Epigenetic dysregulation is thought to play a key role in this process. Numerous human population studies have now shown that changes in DNA methylation of CpG dinucleotides, a key epigenetic mechanism, are strongly associated with chronological age. Although these epigenetic changes are in part a by-product of age-related changes in the cellular composition of the studied tissue, many age-related differentially methylated positions (aDMPs) observed in blood samples are independent of cell composition, and aDMPs have proven to be a useful tool to predict chronological age. However, aDMPs may not be the most informative marker of the ageing process since they were discovered as close correlates of chronological age instead of biological age. Moreover, only a small proportion of aDMPs are associated with expression changes, suggesting that their functional implication may be limited. In contrast, DNA methylation changes that increasingly diverge from chronological age may reflect the increasing inter-individual variation in health that occurs with increasing age. Initial studies, although small or lacking a genome-wide view, indicated that an increasing variability of DNA methylation with age indeed exists.

In the current study, we charted the occurrence of age-related variably methylated positions (aVMPs) across the genome. We evaluated the methylation at 429,296 CpG sites for increased variability with age in whole blood samples from 3295 individuals aged 18 to 88 years. We discovered and validated 6366 age-related variably methylated positions (aVMPs). While aVMPs were commonly associated with the expression of (neuro)developmental genes in cis, they were linked to transcriptional activity of genes in trans that have a key role in well-established ageing pathways such as intracellular metabolism, apoptosis, and DNA damage response. Of interest, tumors were found to accumulate DNA methylation changes at CpG sites of aVMPs, thus supporting the long-standing notion that ageing and cancer are in part driven by common mechanisms.

Our data show that the genomic regions accumulating variability in ageing populations are highly specific and reproducible. Hence, although the increase in variability may have a stochastic component, the regions affected by this phenomenon are well-defined and not stochastic. Intriguingly, associations of aVMP methylation with gene expression in trans extended to genes known to play a role in ageing. In older individuals who had an aged DNA methylation profile as compared with young individuals, we observed a downregulation of genes involved in metabolism. The upregulation of ageing pathways, as observed in old individuals with an aged methylome, has been reported previously in hematopoietic stem cells in mice and humans, for which macromolecular or DNA damage may be the driving force. Of note, many of the trans-genes we identified are involved in the DNA damage response and are frequently mutated in various cancers. Hence, genomic stress, due either to hyperproliferation or DNA damage, may drive upregulation of well-established ageing pathways, downregulation of intra-cellular metabolism, and altered regulation by proteins associated with increased variability of DNA methylation. In contrast to aDMPs, aVMPs show a striking variability in DNA methylation at higher ages. Two individuals of the same age may display highly distinct methylation patterns across aVMPs, where one of them may have a DNA methylation profile at aVMPs that is similar to that of young individuals. Therefore, aVMPs fulfill a primary prerequisite for a biomarker of biological age.