利用单细胞多组学解析胶质母细胞瘤中细胞状态的基因调控

Glioblastoma (GBM) is an incurable and aggressive brain cancer characterized by profound intra- and intertumoral heterogeneity and remarkable cellular plasticity. Single-cell transcriptomic analyses have revealed several major cell states, including NPC-like, OPC-like, GPC-like, AC-like and MES/Hypoxia-like. However, the cis-regulatory networks that govern GBM cell state transitions remain poorly understood. In this study, we performed single-cell chromatin accessibility profiling and multi-omics analysis on 35 GBM IDHwt samples. Firstly, we developed a new scATAC-seq data analysis framework and reconstructed six malignant consensus cis-regulatory element (CRE) modules. Four of these modules were specifically associated with malignant cell states corresponding to the MES/Hypoxia-like, AC-like, OPC-like, and NPC-like identities. Interestingly, cycling cells exhibited broadly open chromatin across all four CRE modules, while GPC-like cells showed accessibility in both the AC-like and OPC-like states, suggesting a role as an intermediate or hybrid regulatory state. Further epigenetic information quantification revealed that NPC-like malignant cells harbor higher regulatory information content compared with other cell states. Master regulator enrichment analysis identified AP-1 transcription factors as key regulators of differentiated (MES/AC-like) malignant state-associated CRE modules, whereas neuronal-development transcription factors were enriched in stem-like (NPC/OPC-like) state-associated modules. Through in vitro gain- and loss-of-function experiments, we screened and validated several transcription factors that modulate malignant cell-state transitions. Additionally, our scATAC-seq-based copy number alteration (CNA) analysis captured hallmark GBM genomic events at high resolution, including EGFR focal amplification, CDKN2A/B deletion, and CDK4 and MDM2 amplifications. Leveraging these CNA profiles, we successfully constructed a high-resolution phylogenetic tree, capturing the clonal architecture and evolutionary trajectory of GBM. By integrating transcriptomic, chromatin accessibility, and genetic CNA data, we elucidated the evolutionary landscape of GBM progression and cellular plasticity. Our findings provide significant insights into the regulatory architecture of GBM and establish a foundational framework for precision therapies targeting distinct cell states.