Aging is a complex stochastic process of damage accumulation, shaped at the detail level by choices made, environments encountered, and the responses and interactions of countless cells. As the paper here notes, this gives rise to a sizable variance between individuals, even those with the same genetics or similar lifestyles. The research and medical community largely react to this variance with a drive towards developing personalized medicine, but the underlying mechanisms of aging are the same from individual to individual, even if outcomes vary. A therapy that addresses one underlying mechanism in one tissue will tend to produce differing levels of benefit from person to person, but in principle it should always produce benefit. Undergoing some degree of repair should always result in better outcomes than not undergoing repair.
Most human omic studies remain cross-sectional, despite growing evidence that transcriptomic and metabolomic states are temporally dynamic. Cross-sectional designs are limited in their capacity to separate stable interindividual differences from within-participant change and may obscure the degree to which molecular aging varies between individuals. Longitudinal multiomic studies address this gap but are challenging to implement. To address these challenges, we established the MultiMuTHER study within the deeply phenotyped TwinsUK cohort. MultiMuTHER includes repeated whole-blood RNA sequencing and serum metabolomic profiling at three or more time points over up to 8 years in 335 females. We used this resource to ask four questions: which genes and metabolites change over time; how much individual trajectories diverge from population-level trends; to what extent multiomic trajectories are shaped by intrinsic factors including host genetics and cell type composition as well as extrinsic elements such as circadian timing, seasonality, and environmental exposures; and whether transcriptomic and metabolomic changes are coordinated across omic layers.
We identified 5,061 genes and 181 metabolites whose levels changed over time, with participants displaying distinct longitudinal trajectories, sometimes opposite to population-level trends. Longitudinal changes in gene expression exhibited marked cell type specificity, with different patterns in the adaptive versus innate immune compartments. Longitudinally variable genes were enriched for functional categories relevant to aging, including cardiometabolic and neurodegenerative disorders. Gene expression and metabolite levels showed strong context-specificity, with 25% of genes and 24% of metabolites associated with seasonality, and 26 and 39%, respectively, associated with circadian variation.
Our findings show molecular aging to be a dynamic, context-dependent, and highly individualized process. Longitudinal multiomic profiling revealed molecular changes shared across the population and trajectories that diverged between individuals, while highlighting the influence of cell type composition, host genetics, biological rhythms, environmental exposures, and cross-omic connectivity. This study, along with the MultiMuTHER resource, provides a framework for understanding biological aging as a set of interacting molecular trajectories, with utility for future precision medicine efforts.
Link: https://doi.org/10.1126/science.aed6452
View the full article at FightAging
