New Cellular "Blueprint" of Glioblastoma Explains Why It's So Hard to Treat
Key findings
- To investigate the transcriptional heterogeneity of glioblastoma, researchers at the Massachusetts General Hospital Cancer Center used single-cell RNA sequencing to analyze gene expression in 24,131 glioblastoma cells from 20 adult and eight pediatric patients
- Malignant cells in glioblastoma existed in a limited set of cellular states that recapitulated programs of neural development, which the researchers named neural progenitor–like, oligodendrocyte progenitor-like, astrocyte-like and mesenchymal-like
- Each patient had a unique combination of states, with differences in the frequencies of each state being associated with genetic alterations in the CDK4, EGFR, NF1 and PDGFRA loci, that each favored a particular state
- In vivo lineage-tracing experiments demonstrated plasticity between the four states
- The convergence of cellular signatures into just four patterns explains how glioblastoma can adapt to current therapeutic efforts and might help identify new treatment strategies
Subscribe to the latest updates from Cancer Center Advances in Motion
A critical requirement for better treatment of glioblastoma is to understand its heterogeneity. Bulk transcriptional subtypes that differ between patients and can also differ within tumors—and can change over time and during therapy—have been used to study tumor heterogeneity, but their underlying biology significance remained unclear. Single-cell analyses are offering a renewed understanding of glioblastoma heterogeneity.
Mario L. Suvà, MD, PhD, of the Center for Cancer Research at the Massachusetts General Hospital Cancer Center, and colleagues have reported in Cell on the most comprehensive analysis of glioblastoma heterogeneity to date. Using single-cell RNA sequencing, they identified four cellular states, each favored by defined genetic alterations—and demonstrated that cells can switch between these states.
The Four States
The research team analyzed the RNA content of 24,131 glioblastoma cells from 20 adult and eight pediatric patients. Almost all malignant cells existed in one of four gene expression states: mesenchymal-like, astrocyte-like, oligodendrocyte precursor-like or neural progenitor–like. As their names suggest, these states mimicked normal brain cell development to some extent, but they had distortions that made them cancerous.
Genetic Drivers
The states were largely consistent between adult and pediatric tumors. Each glioblastoma sample contained cells in multiple states, but the relative frequency of each state varied between tumors. Differences in frequency were influenced by alterations in CDK4, EGFR, NF1 and PDGFRA that each favored a particular state.
Plasticity of Glioblastoma Cells
Intriguingly, 15% of glioblastoma cells highly expressed two distinct states. Using several in vivo models and single-cell barcoding technologies, the researchers observed that glioblastoma cells can transition from one state to another; in fact, a single cell can give rise to all four states.
Implications for Drug Development
These results provide clues to why previous approaches to glioblastoma have failed. Targeting a single genetic pathway, such as EGFR, might simply cause a different state to emerge and stabilize the tumor ecosystem. By exploring the newly discovered mechanisms of adaptation, it may be possible to develop novel drugs that overcome resistance to therapy.
view original journal article Subscription may be required
Learn more about the Suvà Lab
Refer a patient to the Mass General Cancer Center