What Can Be Learned About Aging Clocks from Existing Clinical Trial Data for Longevity Interventions?
Aging clocks can be readily produced by using what are by now well established machine learning techniques on any body of biological data that changes with age. Omics data is favored, as there are a great many databases of such data for large study populations, but clocks have been made using clinical chemistry results from simple blood tests, imaging data, and many other items. A clock is calibrated to predict age in the reference population, and when used on people outside that reference populations, most clocks tend to predict a higher age for people more greatly impacted by aging. Thus a higher clock age than chronological age tends to correlate with increased mortality risk, presence of age-related conditions, and so forth, at least over populations.
The most important goal for the development of aging clocks is to prove that one or more of them can be used to assess the quality of novel interventions in aging. Having a way to quickly assess whether or not a potential rejuvenation therapy is any good would transform the field, allowing researchers to rapidly focus on the best ways forward and optimize them. Unfortunately, there is no way to trust that a clock will be reliable in such circumstances other than to calibrate it against the intervention by running slow and expensive life span studies - which defeats the purpose. Unfortunately, there remains no well mapped direct connection between the data used to derive clocks and the underlying mechanisms of aging. Any given clock might underestimate or overestimate the effects of any given approach to treating aging, and whether that is the case or not is presently unknowable in advance.
The approach to this challenge taken by the research community is to gather as much data as possible on the behavior of as many different clocks as possible, both in animal studies and in human clinical trials. The hope is that out of this growing body of data, patterns and understanding will emerge as to which forms of clock, and which measures making up clocks, can be trusted. Today's open access paper is illustrative of progress on this front, an assessment of much of the human clinical trial data gathered to date for interventions that are known or hoped to affect late life health and life expectancy. Obviously, the catalog of interventions rigorously assessed in humans since the advent of aging clocks is at the present time largely modest in effect: exercise, diet (including calorie restriction), metformin, therapeutic plasma exchange, and so forth. Nonetheless, there is still something to be learned, even at the present state of progress.
Responsiveness of epigenetic aging biomarkers to longevity interventions in humans
Aging biomarkers can potentially allow researchers to rapidly monitor the impact of an aging intervention without the need for decade-spanning trials. However, before the use of aging biomarkers, such as epigenetic clocks, as surrogate endpoints, their responsiveness to interventions that target aging must be tested. Here we curate TranslAGE, a harmonized database of 51 public and private longitudinal interventional studies, and calculate a consistent set of 16 prominent epigenetic clocks for each study, along with 94 other DNA methylation (DNAm) biomarkers that can help explain the changes observed for each clock.
Using this database, we discover patterns of responsiveness across a variety of interventions and DNAm biomarkers. For example, clocks trained to predict mortality or pace of aging show the strongest responses across all interventions and show consistent results with one another; pharmacological and lifestyle interventions drive the strongest responses from DNAm biomarkers; and the characteristics of the study population and study duration are key factors in determining the responsiveness of DNAm biomarkers to an intervention. Moreover, clocks with multiple subscores (that is 'explainable clocks') provide specificity and greater mechanistic insight into the responsiveness of interventions than single-score clocks.
These findings can help to design future clinical trials by guiding the choice of interventions and of specific subsets of DNAm biomarkers to minimize multiple testing, study duration, study population and sample size, with the eventual aim of uncovering DNAm biomarkers that can be used as surrogate aging endpoints.