Two Promising Senolytic Therapies Fail to Replicate in Rigorous Cross-Lab Study

Cellular senescence—a state of stable cell-cycle arrest accompanied by the pro-inflammatory senescence-associated secretory phenotype (SASP)—has emerged as a key driver of age-related disease and tissue dysfunction. The idea that eliminating senescent cells pharmacologically (senolytics) could extend healthspan has generated enormous scientific and commercial excitement, with over 20 candidate senolytic drugs reported to date.

Two interventions attracted particular attention: BPTES, a small-molecule inhibitor of the glutaminase enzyme GLS1, and an anti-PD-1 immune checkpoint antibody. Original publications claimed that both agents dramatically reduced p16INK4a-expressing senescent cell burden across multiple organs in aged mice and produced meaningful improvements in physical function and other health parameters. These reports became foundational references in the senolytic literature.

To rigorously evaluate these claims, the current authors conducted blinded, multi-laboratory replication studies spanning three independent research groups in Osaka, Tokyo, and Kyoto, Japan. In vitro experiments across multiple human fibroblast senescence models showed that BPTES eliminated senescent cells only at concentrations (≥10 µM) that were equally toxic to actively proliferating non-senescent cells—demonstrating a lack of senolytic specificity. In mouse experiments, aged animals treated with BPTES under protocols closely mirroring the original study showed no significant reduction in p16INK4a mRNA or protein in liver, lung, or kidney. Grip strength and external aging phenotype scores were also unchanged. Critically, the blinded cross-laboratory design ruled out observer bias as a confounding factor.

For anti-PD-1 intervention, neither antibody-mediated PD-1 blockade nor genetic deletion of the PD-1 gene in aged mice reduced p16INK4a-positive cell burden or improved aging-related functional measures. These null findings held across multiple independent cohorts and readout methods.

The authors explicitly state they do not seek to discredit prior researchers, but rather highlight how subtle differences in experimental conditions, technique proficiency, animal housing, and outcome assessment can yield divergent results—a known challenge in preclinical biomedical research. They call for standardized senescence detection protocols, pre-registered replication studies, and independent validation as prerequisites before any senolytic therapy advances toward clinical trials. The study is a sobering reminder that promising preclinical results require robust, transparent verification before informing human health decisions.