Scientists have long suspected that aging, at its molecular core, follows a conserved script — that a mouse dying of old age and a human doing the same are, in some fundamental sense, running the same program. A landmark study published in Nature now provides the most comprehensive evidence yet that this is true, and offers a toolkit to read and score that program in real time from a blood sample or tissue biopsy.
Researchers at Harvard Medical School’s Brigham and Women’s Hospital, led by Vadim Gladyshev and Alexander Tyshkovskiy, assembled more than 11,000 gene expression profiles spanning 25-plus tissues from four mammals — mouse, rat, crab-eating macaque, and human — and used them to train a suite of transcriptomic clocks. Unlike DNA methylation clocks (which read chemical marks on DNA), these new “tAge” clocks read which genes are being switched on or off, and at what level, to predict not just how old an organism is, but how close to death it is — right now.
The mortality clock is the centerpiece. Trained on expected all-cause hazard rates derived from Gompertz survival models, it distinguishes animals running on borrowed time from those aging slowly, predicts time-to-death in the Framingham Heart Study (n = 3,698 people) with accuracy comparable to second-generation DNA methylation clocks like DunedinPACE, and does so with fully interpretable gene-level readouts.
The team then dissected the architecture of biological aging into 28 co-regulated gene modules — distinct cellular subsystems each with its own aging “clock” — revealing that different interventions target different parts of the machinery. Caloric restriction primarily reverses metabolic ageing (mitochondrial, lipid, and haem metabolism modules), while inflammatory stress (LPS, chronic disease) predominantly drives immune-module aging. Crucially, heterochronic parabiosis — sharing young blood with old mice — produced a systemic tAge reduction spanning most modules, suggesting its rejuvenating effects are genuinely multi-pathway rather than pathway-specific.
Three genes emerged as the most consistent molecular markers of mortality across every species, tissue, and disease model tested: CDKN1A (encoding p21, a cell-cycle brake and senescence enforcer), LGALS3 (galectin-3, a pro-inflammatory lectin), and GPNMB (an inflammation-associated glycoprotein). All three predict all-cause mortality and a wide spectrum of chronic diseases in over 50,000 UK Biobank participants at the protein level — bridging the gene-expression findings to clinical reality.
Perhaps most striking is the embryogenesis finding: a genuine molecular “ground zero” occurs around embryonic day 10 in mice, during which the transcriptomic mortality signature resets to its lowest recorded level, with CDKN1A and LGALS3 among the key genes suppressed. Early embryogenesis, caloric restriction, and heterochronic parabiosis all share this suppressive signature — suggesting the biology of rejuvenation may converge on the same molecular levers, regardless of how it is triggered.
Actionable Insights
For individuals and clinicians:
The most immediately actionable finding is the identification of three plasma proteins — GPNMB, CDKN1A/p21, and LGALS3 — as validated mortality and multimorbidity predictors in over 50,000 people. Effect sizes from UK Biobank Cox models (adjusted for age and sex) show standardized hazard ratios of approximately 1.1 to 1.6 per standard deviation of circulating protein level for all-cause mortality. This means that for each meaningful step up in the blood level of these proteins, your risk of dying from any cause increases by 10% to 60%. For comparison, many commonly used clinical risk markers (like LDL cholesterol for heart disease) have hazard ratios in a similar range. The fact that these proteins predict all-cause mortality — not just one disease — is what makes them particularly notable. Of the three proteins, galectin-3 (LGALS3) is the broadest predictor — it’s not just tied to one organ or one disease, but to six major disease categories spanning the heart, liver, kidneys, and metabolic system. This makes biological sense: galectin-3 drives fibrosis (progressive scarring) and chronic inflammation, which are underlying mechanisms in all of those conditions. A protein that contributes to organ damage across multiple systems naturally shows up as a predictor of multiple diseases.
Pathway-level intervention targeting:
Caloric restriction’s mortality benefit operates primarily through metabolic modules (mitochondrial, lipid, haem/ROS pathways), not inflammatory ones. Conversely, chronic inflammatory load — from any source — drives the immune/interferon mortality modules hardest. This provides a rational basis for combining a metabolic intervention (caloric restriction, rapamycin, canagliflozin) with an anti-inflammatory one (e.g. PectaSol-C) to address distinct aging subsystems simultaneously rather than overlapping mechanisms.
Senescence as the common currency: CDKN1A/p21 upregulation is the single most consistent pro-mortality signal across disease, damage, and ageing. Senolytic or CDKN1A-targeting strategies therefore address a genuine pan-tissue ageing driver. [Confidence: Medium — protein correlation, not yet causal proof in humans]
Source:
- Open Access Paper: Universal transcriptomic hallmarks of mammalian ageing and mortality
- Institution: Division of Genetics, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA (primary); with collaborators from Moscow State University, ETH Zurich, McGill University, Tohoku University, The Jackson Laboratory, and UT Health Science Center San Antonio.
- Country: USA (primary lead institution), with international collaboration (Russia, Switzerland, Canada, Japan)
- Journal: Nature (Springer Nature)
- Impact Evaluation: The impact score of this journal is approximately 50.5 (2023 Journal Impact Factor), evaluated against a typical high-end range of 0 to 60+ for top general science multidisciplinary journals. Therefore this is an Elite impact journal.
