Glioblastoma cells do not switch between different states at random, according to two coordinated studies published in Nature. Researchers mapped predictable pathways between cell states and identified molecular regulators that may help restrict the cancer’s ability to adapt.
The teams from EMBL, the German Cancer Research Center and the Wellcome Sanger Institute built a multi-omic atlas using data from more than one million patient-derived cells. The work combined single-cell molecular profiles with spatial information to examine how different cell states are organized within tumors.
The analysis found that cell states gathered in distinct tumor microenvironments and followed hierarchical transition paths. Researchers also observed similar patterns in patient-derived cells grown in three-dimensional brain organoid models.
To investigate what controls these transitions, the researchers developed scDORI, a deep-learning framework that infers gene-regulatory networks from single-cell data. The method used variation among tumor cells to identify regulators involved in cell-fate decisions.
One regulator highlighted by the analysis was MYT1L. In experiments with patient-derived glioblastoma cells, increasing MYT1L pushed cells toward a neuronal-like state with reduced plasticity. In preclinical mouse models, tumors with increased MYT1L showed less growth and invasion, and animals lived longer than untreated controls.
The findings suggest that some less-plastic cell states are actively maintained by regulatory networks rather than being fixed endpoints. The researchers say this approach could inform efforts to manipulate cancer cell identity, although clinical translation remains in development.
Sources: Neuroscience News.