Search

SummaryAging is a major risk factor for human morbidity and mortality. Thus, the identification of compounds that defer aging, also known as ‘geroprotectors’, could greatly improve our health and promote a longer life. Here we screened for geroprotectors, employing the power of human transcriptomics to predict biological age. We used age-stratified human tissue transcriptomes to generate machine-learning-based classifiers capable of distinguishing transcriptomes from young versus old individuals. Then we applied these classifiers to transcriptomes induced by 1300 different compounds in human cell lines and ranked these compounds by their ability to induce a ‘youthful’ transcriptional state. Besides known geroprotectors, several new candidate compounds emerged from this ranking. Testing these in the model organismC. elegans, we identified two Hsp90 inhibitors, Monorden and Tanespimycin, which substantially extended the animals’ lifespan and improved their health. Hsp90 inhibition specifically induces the expression of heat shock proteins, known to improve protein homeostasis. Consistently, Monorden treatment improved the survival ofC. elegansunder proteotoxic stress, and its lifespan benefits were fully dependent on the master regulator of the cytosolic unfolded protein response, the transcription factor HSF-1. Taken together, we present an innovative transcriptomics-based screening approach to discover aging-preventive compounds and highlight Hsp90 inhibitors as powerful geroprotectors that could be of great value, to target the aging process in humans.