Persistence and/or senescence: not so lasting at last?

Therapy-exposed surviving cancer cells may have encountered profound epigenetic remodeling that renders these drug-tolerant persisters candidate drivers of particularly aggressive relapses. Typically presenting as slow-to-nongrowing cells, persisters are senescent or senescence-like cells. In this issue of Cancer Research, Ramponi and colleagues study mTOR/PI3K inhibitor–induced embryonic diapause–like arrest (DLA) as a model of persistence in lung cancer and melanoma cells and compare this persister condition with therapy-induced senescence in the same cells. The DLA phenotype recapitulated some but not all features attributed to senescent cells, lacking, for instance, an inflammatory secretome otherwise known as the senescence-associated secretory phenotype. A CRISPR dropout screen pointed to methyl group–providing one-carbon metabolism and further to H4K20me3-mediated repression of senescence-associated secretory phenotype–related IFN response genes selectively in DLA-like persister cells. Conversely, inhibition of H4K20-active KMT5B/C methyltransferases derepressed inflammatory programs and was toxic in DLA cells. These findings not only suggest exploitable vulnerabilities of DLA-like persister cells but also unveil general technical and conceptual challenges of cultured multipassage cell line–based persister studies. Collectively, the approach chosen and insights obtained will stimulate a productive scientific debate on senescence-like features and their reversibility across drug-tolerant persister cells.