Cellular memory after cancer therapy: persistence after drug withdrawal, maintenance across cell divisions, and responses to rechallenge

A fraction of tumor cells can survive therapy without acquiring stable resistance mutations and enter a drug-tolerant persister (DTP) state. After drug withdrawal, these cells may resume proliferation and regain drug sensitivity, retain phenotypic features that alter survival or death upon retreatment, or continue to evolve toward stable resistance. The DTP designation describes survival under therapeutic pressure but does not establish whether the associated state persists after treatment ends. The term post-treatment cellular memory is used here operationally for a treatment-associated state that remains detectable after adequate drug washout and can influence a later response. Evidence can be evaluated across three experimental dimensions: persistence after withdrawal, maintenance across subsequent cell divisions or clonal descendants, and a functional consequence upon re-exposure. These dimensions indicate different strengths of support rather than mandatory diagnostic criteria. The review therefore organizes representative studies by what was measured after withdrawal rather than by DTP markers or mechanisms alone. Recent work indicates that AP-1 can contribute to encoding treatment-associated transcriptional states and that CBP/p300 activity can help maintain an established reporter state, including after drug withdrawal; however, molecular persistence does not substitute for demonstrating altered function upon rechallenge with the original agent. Chromatin remodeling, transcriptional state transitions, antioxidant metabolism, autophagy, organelle homeostasis, and lineage plasticity may prolong residual-cell survival. Their biological consequences differ: some states decay after withdrawal, some create time-limited vulnerabilities to ferroptosis or to therapeutic targeting of anti-apoptotic dependencies, and others extend the interval during which genetic resistance can emerge. Lineage tracing, single-cell multi-omics, withdrawal-rechallenge experiments, and longitudinal sampling of minimal residual disease can help distinguish transient stress responses from persistent post-treatment states. These observations provide a rationale for testing whether treatment sequence and interventions during residual disease can be matched to the reversibility and vulnerabilities of residual states.