Epigenetics & Longevity: Unraveling Aging Mysteries with Dr. David Meyer and Prof. Sarah Voisin at ARDD — The VitaDAO Aging Science Podcast
This special the episode of the Aging Science podcast is brought to you from the ARDD conference that took place at the end of August in Kopenhagen. I sat together with Dr. David Meyer from CECAD and Prof. Sarah Voisin from the University of Copenhagen to talk about epigenetics, clocks, our favorite conference presentations, stochastic and adaptive changes during aging. Finally, we also talked about the all-important questions of sustainability and overpopulation, and whether longevity researchers should have a say in this matter.
My interview partners
David is a bioinformatician with a strong interest in genome instability and aging research. His research utilises data analysis, machine learning, and the development of computational tools to help advance our understanding of aging and age-related diseases.
Sarah is an Assistant Professor at the University of Copenhagen and a former Australian NHMRC Early Career Fellow (2019–2022). She works at the intersection of epigenetics, genetics, statistics and bioinformatics, with particular focus on exercise, ageing and human health. She completed her PhD at Uppsala University and Pierre & Marie Curie University, and published 58 peer-reviewed papers. Her current research work and interest is epigenetic ageing and exercise, sex differences in biology, and statistical methods to develop personalised health interventions.
What is epigenetics?
The term epigenetics describes a form of gene regulation that relies on modifications to the (chemical) structure and organization of DNA and chromatin. This is in contrast to genetics which deals with changes to the DNA sequence itself.
Epigenetic marks are a type of modification that is attached to DNA that help genes to be expressed at the proper time. There are two major types of epigenetic marks, those found on histones and those found directly on the DNA. Histone modifications are highly diverse, whereas regulatory DNA modification is largely limited to the methylation of cytosines.
The different epigenetic marks are recognized by reader proteins that affect gene expression. Although these marks can have complex effects, some broad patterns are evident, like for example repression of transcription through CpG methylation.
Although methylation serves an important regulatory role, it turns out not all species have methylation at CpG sites. Bacteria have a different type of DNA methylation whereas C. elegans seems to have lost this feature altogether during evolution.
The rise of epigenetic clocks
When we age, so does our epigenome and one way to measure this epigenetic aging is through so-called DNA methylation clocks. Many studies, initially pioneered by Steve Horvath, have shown that methylation at certain sites in the genome can be used to construct a clock that closely tracks the passage of time.
These clocks are exciting because they are much better at predicting chronologic and biological age than the biomarkers we used to have. In fact, I still remember going to conferences and listening to talks that painted a rather disappointing picture of blood based aging biomarkers.
These clocks are also exciting because they promise to become a surrogate endpoint for medium sized phase II-like studies. A surrogate endpoint is a cheap, easy-to-measure marker that substitutes for something else that is more expensive. The hope is that a slowing of epigenetic aging in small studies would be predictive of success in larger trials that focus on harder outcomes like health and survival. If true, this would push down the cost of trials by orders of magnitude because we could pre-select promising compounds before going into large studies. However, even the best surrogate endpoints like LDL or HDL cholesterol cannot substitute for large studies, as the countless failures and disappointments in the clinic show (e.g. CETP-inhibitors for atherosclerosis and the clinical failure of most non-statin drugs).
Another reason why clocks are important is their link with the biology of aging. Understanding why methylation at certain sites changes and what harmful effects these changes have, could improve our understanding of aging and allow us to find new therapies. To give an example of such advances, as mentioned in a prior podcast, the study of epigenetics has uncovered not only the importance of entropy in the process of aging, but also ways to reset some of these age-related changes (partial reprogramming).
Stochastic Changes and Aging
These changes refer to random alterations in the DNA, epigenome or even proteome that accumulate over time. David Mayer’s work suggests that these stochastic changes, or noise, may play a significant role in aging and may be the age component that epigenetic clocks actually measure. Understanding these changes could provide insights into why people age at different rates, if we find the drivers of increasing noise or even means to reverse it.
In a way the idea that stochastic changes drive aging is a very old one and dates to aging theories like the error catastrophe theory. Initially, it was believed that DNA will accumulate damage, which is then transcribed to produce proteins of lower quality which will further increase damage accumulation. Although there may be no error catastrophe per se, the concept has been rediscovered recently. It may very well be that such stochastic errors accumulate, both on the gene and epigenetic level, which decrease the functioning of gene expression networks, finally resulting in reduced resilience and aging.
There are several interesting aspects of these stochastic theories. One is the potential irreversibility. Aging may be hard to reverse if stochastic, very random, scattered, changes are key drivers rather than a few well-defined damage types or “hallmarks”. This worry has been recently echoed by voices like Peter Fedichev. Because stochastic changes are hard to reverse, and hard to “titrate” in experiments, these theories are also very difficult to prove. For example, there is still no smoking gun experiment showing that reduced DNA damage accumulation slows aging or that increased accumulation accelerates aging. The former has proven devilishly hard since DNA damage repair is so efficient. Whereas the latter is always controversial since models of accelerated damage accumulation often show damage levels that are either too high, too low or too specific (hence it is hard to “titrate” to the correct level).
In this context, we also discussed attempts to disentangle the stochastic component of clocks from an adaptive one, as recently proposed by Vadim Gladyshev. It remains to be seen if this will allow us to pinpoint genes that have important functions with aging.
