A Cross-Species Transcriptomic Aging Clock
If one can develop a single aging clock that works in much the same way in both mice and humans, could it be used to determine which of the interventions to treat aging that have been tested in mice are more likely to work well in humans? It is clearly the case that most of the established approaches to slowing the progression of aging, largely derived from manipulation of stress response mechanisms that clean up damage and improve cell function, produce much larger increases in life span in short-lived species than in long-lived species. How will that difference manifest in an aging clock designed to work similarly in both short-lived and long-lived mammals? That is an interesting question, still awaiting an answer.
Ageing and interventions modulate health and mortality, yet the underlying molecular mechanisms of this modulation remain unclear. Here we integrate more than 11,000 transcriptomes from more than 25 tissues across 4 mammals (mouse, rat, macaque, and human) to develop accurate, interpretable rodent and multi-species biomarkers of chronological age and expected mortality, predicting lifespan-modulating interventions, time to death, chronic diseases, and rejuvenation. Ageing-related changes were conserved across species and cell types, revealing universal transcriptomic signatures of mammalian ageing and mortality, including CDKN1A and LGALS3, whose protein levels were also associated with mortality and multimorbidity in UK Biobank.
Mortality-associated features were recapitulated across in vivo and in vitro damage-accumulation models, including inflammation, replicative senescence, metabolic inhibition, and γ-irradiation, and were attenuated or reversed by cell immortalization, reprogramming, heterochronic parabiosis, and early embryogenesis. Network analysis uncovered a modular architecture of ageing- and mortality-associated hallmarks, encompassing inflammation, interferon signalling, mitochondrial function, chromatin modification, and extracellular matrix organization.
To quantify ageing of individual cellular components, we developed module-specific clocks, which revealed pathway-specific effects of interventions: chronic diseases primarily accelerated inflammatory-module ageing, whereas caloric restriction and Klotho deficiency targeted mitochondrial and metabolic modules. Transcriptomic and DNA methylation clocks showed correlated age acceleration in human blood, which was strongest for the chromatin-associated module clock, highlighting mechanistic links between molecular ageing modalities. This study reveals conserved signatures and a modular architecture of mortality regulation, providing a framework for quantifying and targeting ageing of cellular subsystems across species and tissues.