Despite the challenges inherent in the practical use of epigenetic clocks based on age-related changes in DNA methylation, clinical trials are forging ahead in the employment of these clocks. The assays are cheap enough that there is a sense of “why not?” and, considered over patient populations rather than in individuals, an epigenetic age higher than chronological age correlates well with risk and progression of age-related disease. It remains problematic for any given individual to extract meaning from an epigenetic clock assay, however. It is unclear as to what exactly the measured age-related changes in DNA methylation reflect, in terms of the underlying damage and dysfunction of aging, and thus results are not yet actionable for the individual.


Geroscience is a developing discipline based on the premise that health can be improved by targeting aging. Clinical trials are underway to test the geroscience hypothesis in humans. Definitive tests of the hypothesis must demonstrate reduced rates of age-related diseases and death, but the length of time and size of trial needed to test the hypothesis are both substantial. Therefore, objective, quantifiable characteristics of the aging process – known as biomarkers – that can be tracked in clinical trials are needed for the field to progress.

Useful biomarkers should meet several criteria: i) their measurement should be reliable and feasible; ii) they should be relevant to aging; iii) they should robustly and consistently predict trial endpoints, such as functional ability, disease, or death; and iv) they should be responsive to interventions such as treatments targeting aging biology. Practically speaking, this means that a change in the level of a biomarker should parallel changes in the susceptibility to disease, age of death, or loss of function. Interventions that target aging and support the geroscience hypothesis should therefore also lead to changes in these biomarkers, which will be reflected in the incidence or severity of age-related diseases and functional decline.

Biomarkers based on DNA methylation levels look promising. Briefly, these biomarkers quantify the proportion of cells in which a gene locus is methylated. Small but consistent changes in the methylation of some loci occur in organisms with older ages, and early methods for estimating age using epigenetics took advantage of these chronologic changes. However, critics argue that while these ‘clocks’ may be associated with chronological age, it is uncertain whether they reflect meaningful change in the context of interventions affecting the underlying biology.

Estimators based on the levels of DNA methylation are now being developed to detect a myriad of disease states and predict mortality and adverse health events, and each is unique to its calibration method. Now researchers report the development of a new epigenetic biomarker called Dunedin Pace of Aging methylation (DunedinPoAm) that is able to detect how aging phenotypes change over time. The new biomarker relies on a composite measure called the Pace of Aging. Pace of Aging is calculated based on a number of age-related phenotypic changes that occur over time. In the new work this measure was used to calibrate and validate a DNA-wide methylation clock in four independent cohorts.

Is DunedinPoAm developed to the point where it could be relied upon as a biomarker for clinical trials targeting biological aging? DunedinPoAm appears to satisfy the criteria aside from being responsive to interventions. One of the cohorts used to validate the new approach consisted of middle-aged, non-obese adults enrolled in the CALERIE trial. This trial tested the effects of caloric restriction – an intervention that has been successful in animal models – over a period of two years. DunedinPoAm was able to predict changes in the Pace of Aging measure in the control group, but not in the group that had been calorie restricted. However, it remains to be seen whether interventions which affect aging biology change DunedinPoAm in a way that is consistent with the phenotypic changes observed in the trial. Testing the geroscience hypothesis in clinical trials is still in its early days, so it is not surprising that DunedinPoAm does not yet meet the primary criterion for an aging biomarker.

Link: https://doi.org/10.7554/eLife.58592