Efforts to extend longevity and healthspan have benefited from the use of in vitro systems, short-lived model organisms and computational approaches, but translation to humans depends on demonstrating efficacy and safety in mammalian in vivo models. However, rigorously controlled rodent lifespan studies are slow, costly and low-throughput. Hormesis, the induction of adaptive stress response pathways by low-level toxic or metabolic challenges, provides a mechanistic means for improved resilience, but testing hormetic dose-response relationships in mammals is logistically challenging. Regulatory long-term toxicity studies, mandated for chemical authorisation, routinely generate lifespan and chronic toxicity data from large cohorts across multiple dose groups and therefore represent an underused resource for detecting unexpected survival benefits. We mined a random subset of the German Federal Institute for Risk Assessment (BfR) electronic archive for long-term rodent pesticide studies in ≥ 50 animals per dose and sex. From qualifying studies with 154 substances, we identified 27 substances (18% of entries) that significantly increased survival at study termination (≥ 10%, p < 0.05). Fifteen publicly disclosable substances were examined further; roughly half perturbed mitochondrial oxygen consumption rates in mammalian cells, and bromoxynil acted as a mitochondrial uncoupler. Back-translation to Caenorhabditis elegans showed bromoxynil improved multiple healthspan indicators without compromising organismal health. Regulatory toxicity archives can thus be repurposed to identify candidate compounds, targets and mechanisms for mammalian healthy ageing research, complement conventional screening pipelines, align with 3Rs by maximising existing data, and supply high-quality datasets for AI-driven prioritisation of interventions for rigorous preclinical follow-up.
