Abstract Glioblastoma (GBM) remains the most aggressive and lethal primary brain tumor, with limited therapeutic options. A major contributor to treatment failure is intratumoral heterogeneity driven by dynamic and reversible cellular state transitions allowing GBM to adopt distinct transcriptional states in response to environmental and therapeutic stress. How individual anti-tumor drugs reshape this cellular state landscape, and whether drug-induced state transitions can be therapeutically exploited, however, remains to be elucidated. In this study, we systematically characterized drug-mediated cellular state dynamics in GBM using two patient-derived glioblastoma stem cell (GSC) lines. We performed single-cell RNA sequencing before, during, and after treatment with a panel of 13 anti-tumor drugs spanning diverse mechanisms of action. Therapies induced pronounced, drug-specific shifts in cellular state distributions, most of which were partially reversible following drug withdrawal, highlighting the intrinsic plasticity of GSCs. Among the compounds tested, the histone deacetylase inhibitor panobinostat produced the most robust shift, toward the Mesenchymal-like (MES-L) state, whereas the PI3K inhibitor paxalisib preferentially enriched Astrocytic-like (AC-L) states. To resolve whether these changes reflected phenotypic plasticity or clonal selection, we performed Whole Genome Sequencing and lineage tracing using transcriptomics barcoding in an ex vivo glioblastoma cerebral organoid (GLICO) model. Clonal analysis demonstrated that drug-induced state transitions occurred largely within existing clones, supporting phenotype plasticity as the dominant mechanism. Panobinostat treatment preferentially depleted AC-L populations while enriching MES-L states, revealing distinct cell-state–specific sensitivities. Genetic manipulation of the MES-L associated transcription factor FOSL1 further demonstrated that enforced MES-L identity conferred resistance to panobinostat, whereas knock-down of FOSL1 sensitized cells to treatment. Building on these observations, we tested a therapeutic strategy called state-selective lethality (SSL), in which drugs that drive opposing cellular state transitions are combined to restrict adaptive escape. Combination treatment with panobinostat and paxalisib resulted in synergistic anti-glioma activity in both 2D cultures and GLICOs. Single-cell analysis of residual cells following combination therapy revealed a small, predominantly MES-L, non-cycling population, suggesting that SSL may function by depleting proliferative states and limiting transitions into therapy-tolerant compartments. Collectively, our findings establish cell state plasticity as a key mechanism of therapeutic response in GBM and provide proof-of-concept that rationally designed, state-targeted drug combinations can enhance treatment efficacy. Citation Format: Stefano M. Cirigliano, Richa Singhania, James Nicholson, Isha Monga, Caroline Haywood, Ashlesha Muley, Skylar Giacobetti, Howard A. Fine. Targeting Glioblastoma Cell State Plasticity for Enhanced Therapeutic Efficacy abstract. In: Proceedings of the AACR Special Conference in Cancer Research: Brain Cancer; 2026 Mar 23-25; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2026;86 (6Suppl): Abstract nr B057.
Cirigliano et al. (2026) studied this question.