Fight Aging!

Epigenetic patterns of gene expression regulation change with age in distinct ways, and so it is possible to use machine learning to produce algorithmic combinations of this data that reflect chronological age and biological age. There are any number of such algorithms, and so there are many epigenetic clocks now, with more are being produced every year. These are now accompanied by clocks derived from other omics data, alongside clocks built from combinations of simple measures such as grip strength, gait speed, and so forth. It is commonly accepted, and evidence supports this hypothesis, that when a clock predicts an age higher than an individual's chronological age, this age acceleration is an indication of a greater burden of age-related damage and dysfunction.

None of this tells us how underlying damage and dysfunction produces the observed changes in clock data, for any of the clocks built from omics data. This is a problem, because absent this understanding one can't rely upon a clock to produce the right answer when it is used to assess the results of a potential rejuvenation therapy. For example, senolytic treatments clear senescent cells and reduce their impact on metabolism, but it is by no means a given that any given epigenetic clock is actually measuring that effect. Or perhaps the clock is overly reliant on that effect, and will therefore overestimate the benefit of removing senescent cells. Without, at minimum, calibrating a clock to the specific treatment in mouse life span studies one can't know.

Further, different tissues exhibit different patterns of omics data change with aging. Yet most clocks use blood samples, assessing biological age from immune cell populations. So how should researchers think about epigenetic and other omics clocks in the context of their work? Today's open access commentary is, I think, a useful summary of some of the present wisdom on this topic. Acceptance of the inherent limitations (for now, at least) and avoidance of generalizing results from blood markers given the large variance between tissues sounds like good advice.

Recalibrate concepts of epigenetic aging clocks in human health

First, epigenetic aging in human health should be viewed holistically. The relationship between epigenetic age and health outcomes is often viewed through a narrow lens. Blood epigenetic clocks are predominantly used due to their availability. However, as measured by these clocks, studies and trials frequently focus solely on blood, neglecting broader aging effects on the entire organism. This oversight can lead to exaggerated claims and misconceptions among the public. A younger biological age based on a buccal swab does not necessarily equate to a younger immune system or skeletal muscle composition.

Second, accept the limitations of epigenetic aging clocks. While efforts have been made to develop clocks independent of tissue or cell type, our study demonstrates that entirely eliminating the confounding effects of cell type in epigenetic age acceleration calculation is nearly impossible. Senescence at the cellular level manifests differently across cell types and states, making it challenging to devise a single clock applicable to all cell types. Although universal epigenetic aging marks may exist, modeling them accurately across cell types at the bulk tissue level is complicated by age-related changes in cell composition.

Third, contextualize when interpreting epigenetic aging results. The perception that an accelerated epigenetic age is always detrimental stems from a limited understanding of the clock mechanism. Contrary to this notion, multiple studies have documented beneficial outcomes in specific groups of cancer patients with accelerated epigenetic age. Immune responses play a pivotal role in cancer patient outcomes, and the connection between immune cell composition and epigenetic age established in our study sheds light on how epigenetic age acceleration affects these outcomes, especially in treatments like chemotherapy, radiation therapy, and immunotherapy that influence or rely on the immune system. Moreover, research has revealed significant daily oscillations in epigenetic age, mirroring changes in immune cell composition over the circadian rhythm.

Fourth, model epigenetic biomarkers on biological pathways to avoid black boxes. While direct modeling of DNA methylation changes with age enhances the predictability of biological age, it often obscures specific pathways captured by the algorithm, resulting in black boxes. To mitigate this, further efforts should focus on directly modeling aging and senescence-related pathways for novel biomarkers tracing age. Several markers exemplify the value of this approach. EpiTOC, for instance, functions as a mitotic clock, estimating stem cell divisions. It tracks age and correlates with increased cancer risk, offering insights into the association between cellular aging processes and cancer.