// NATURE NEWS — SPAZIO & SCIENZA
Decoding the regulatory programs governing glioblastoma plasticity
Nature
(2026) Cite this article
Glioblastoma is incurable and resistant to therapy despite genomic stability under treatment1,2,3,4,5,6,7,8, suggesting that non-genetic cellular plasticity is a central driver of progression. Whether such plasticity reflects stochastic state switching or is governed by predictable gene regulation has remained unclear9,10,11,12,13. Here we introduce scDORI (single-cell deep multi-omic regulatory inference)—a deep-learning framework that infers enhancer-driven gene-regulatory networks at single-cell resolution—and apply it to single-nucleus multiome profiles of more than one million cells from primary glioblastomas. Integrating computational inference with systematic perturbation screens across patient-derived models, we show that a structured regulatory program with an asymmetric interplay of activators and repressors governs cellular plasticity in glioblastoma. The resulting transition landscape is hierarchical and predictable, but permits cross-lineage trajectories that are absent from normal neurodevelopment. Within this landscape, low-plasticity states such as the neuronal-like state are not stable differentiation end points but actively maintained configurations, stabilized by a network of safeguard repressors that silence alternative fates. We identify MYT1L as the dominant member of this network, and show that it binds to and represses master regulators of every alternative state to consolidate neuronal-like identity. MYT1L restricted plasticity even under standard-of-care therapy and suppressed tumour growth and invasion in vivo. Conversely, loss of MYT1L reactivated plasticity and accelerated malignant features, indicating that low-plasticity identity requires continuous repression. These findings reframe malignant cell state stability as a balance between activation and continuous repression, and nominate safeguard repressors as candidate therapeutic targets that are distinct from differentiation-inducing activators, with potential relevance across cancers in which phenotypic switching drives therapy resistance.
The capacity of cancer cells to switch phenotypic states, also called cellular plasticity, is a hallmark of cancer and is involved in tumour initiation and progression, including therapeutic resistance1. Glioblastoma (GB), the most common primary brain tumour in adults, is characterized by profound cellular heterogeneity and a dismal prognosis2,3. Although the genomic landscape of GB has been catalogued extensively, actionable genetic drivers remain scarce, and clinical responses to targeted therapies have been minimal4,5,6. Longitudinal studies have underscored the limited contribution of therapy-induced genetic evolution in GB and revealed remarkable clonal stability despite treatment7,8, suggesting that non-genetic mechanisms, such as epigenetic plasticity and transcriptional reprogramming, are central to adaptation and treatment resistance in GB. Seminal single-cell transcriptomic studies have shown that malignant GB cells occupy a continuum of cellular states resembling developmental progenitors (for example, oligodendrocyte precursor cell (OPC)-like and neural progenitor cell (NPC)-like), as well as astrocytic and mesenchymal-like phenotypes9,10. Patient-derived xenograft models have demonstrated that individual GB cells can give rise to multiple states, implying substantial plasticity10,11. However, whether these state transitions follow stereotyped routes or variable trajectories12,13, and how gene-regulatory mechanisms promote or constrain them, remain poorly understood.
Here, and in a companion study14, we address these questions by integrating single-nucleus RNA sequencing (snRNA-seq) and chromatin accessibility (single-nucleus assay for transposase-accessible chromatin using sequencing; snATAC-seq) profiles from more than one million cells from isocitrate dehydrogenase (IDH)-wild-type primary GBs. Using a computational approach (singl