Are 'epigenetic clocks' reliable? CNIO researchers refine the accuracy of these tests that measure ageing

• Epigenetic clocks use 'DNA chips' to measure molecular changes in the genome that correlate with ageing in various organs.
• Having an 'epigenetic age' that is higher than one's chronological age (the length of time someone has been alive) is associated with an increased risk of disease and mortality.
• CNIO researchers have detected significant inaccuracies in the most commonly used epigenetic clocks and have developed a more accurate model.
• The study is published in the journal Genome Medicine.

Among the most widely used tools today to measure the deterioration of the organism over time are the so-called 'epigenetic clocks'. Epigenetics refers to the chemical modifications that accumulate in the DNA of cells throughout life; these changes are known to vary with age, and are, in fact, an indicator of ageing. Just over a decade ago, researchers discovered a pattern of epigenetic changes that correlates with ageing in various organs, leading to the use of epigenetic clocks to measure the progression of the ageing process.

Having an epigenetic age that is greater than one's chronological age (how long a person has actually been alive) has been associated with an increased risk of disease and mortality. The biology underlying epigenetic clocks—the relationship between epigenetic changes and disease—is not yet well understood, but it is accepted that epigenetic clocks do measure critical biological processes.

"Epigenetic clocks have not only highlighted the potential of epigenetic changes as biomarkers of ageing, but also represent a valuable tool to learn more about various pathologies," write researchers Miguel Quintela and Leonardo Garma from the Spanish National Cancer Research Centre (CNIO).

Now, in a study published in Genome Medicine, they have identified significant inaccuracies in the most widely used epigenetic clocks and warn that these tools suffer from a problem of "mismatched technologies". To overcome this problem, they have developed a new epigenetic clock model that provides more accurate measurements.

'Epigenetic age' of breast cancer patients

Quintela heads the CNIO's Breast Cancer Clinical Research Unit, which is leading the 'Digital Twins' project to create virtual models of cancer patients (models that, it is hoped, will help in the future to determine the best treatment for each patient). The team thought it would be interesting to include the epigenetic age of the patients in the model, and they used the usual epigenetic clocks.

Several studies have shown that the epigenetic age of tissues may be relevant to cancer. "In cancer, accelerated epigenetic age [higher than chronological age] has been linked to an increased risk of developing breast and colon cancer," Quintela and Garma write in Genome Medicine. Breast tissue from breast cancer patients has also been shown to exhibit epigenetic ageing, with the acceleration rate dependent on the intensity of the treatment.

However, when using epigenetic clocks, CNIO researchers found that the results varied by an average of three years, and up to 25 years for some of the epigenetic clocks used.

Mathematical models

In practice, epigenetic clocks are mathematical models that interpret data provided by DNA analysis devices, or DNA chips. What explains the inaccuracies in the results, say the CNIO researchers, is that existing epigenetic clocks consider points in the genome that are not represented on the DNA chips currently in use. "It's a question of technological evolution," says Garma, lead author of the study.

The new epigenetic clock model developed by the CNIO researchers adapts to the latest version of DNA chips available. They have found that the variation they obtain between measurements from a single subject or from different subjects is less than a year. "This means that it's a robust and accurate model," says Garma.

Health indicator at a population level?

This is important because "there has been talk that these models may reflect pathological states, but here we see that this is rather questionable. If we want to use them to assess the differences between, for example, an ill person and one with cancer, we have to take into account the technical noise that can cause age disparities in the results," explains the lead author of the paper.

Garma adds: "For an individual it may not matter that the biological age is three years older, but for populations it does. For example, if the epigenetic age of a group of smokers is three years higher than that of the non-smoking population, this data could be relevant for drawing conclusions about the influence of tobacco on health."