← Back to Home

Million-Cell Atlas Reveals Shared Paths of State Change in Glioblastoma

From the edges invading brain tissue to the oxygen-deprived core, seemingly disordered tumor cells show recurring state transitions. Tissue from 12 patients and organoid experiments offer clues for identifying points for therapeutic intervention, while also defining the limits of the findings’ clinical applicability.

By SURL BioNews

Glioblastoma is difficult to treat partly because cancer cells can change their state, allowing multiple survival strategies to coexist within a single tumor. To find more effective treatment strategies, researchers need to understand not only what cells look like at a given moment, but also how they might change next. A study published in *Nature* on November 4 identified shared pathways that recur across different patients’ tumors, using data from more than a million cells.

The research team analyzed tumor tissue from 12 patients who underwent surgery at Cambridge University Hospitals, combining single-cell and spatial molecular analyses, imaging, and whole-genome sequencing. The publicly available GBM-Space atlas contains 1,025,329 cell nuclei, with 4 to 15 sites sampled in each tumor, and identifies 28 malignant cell states. This multiregion sampling allowed researchers to place gene activity within its tissue context and understand how different cells together make up a tumor.

The patterns that emerged corresponded to the environments within the tumors. At the edges invading surrounding brain tissue, cancer cells more closely resembled developing brain stem cells and could give rise to other cancer cell states. Farther inward, cells showed responses resembling those seen when brain tissue is injured and repair begins. In the oxygen-deprived core, cells shifted toward states associated with hypoxia and stress responses. These findings support a trajectory from developmental-like states through injury responses toward hypoxic states, but should not be interpreted to mean that every cell necessarily follows the same path in sequence.

Artificial intelligence and machine learning were used to integrate large amounts of molecular and spatial data and infer possible transitions between cell states. The team also tested the trajectory using three-dimensional brain organoids grown in the laboratory, obtaining evidence that supported it. The research therefore goes beyond computational inference alone. However, organoids still cannot fully reproduce the environment inside a patient’s brain or replace clinical validation.

Another study published in *Nature* on the same day used multiomics data from the same patients to further investigate which mechanisms permit or constrain cell transitions. According to the German Cancer Research Center, the team identified the transcription factor MYT1L, which can suppress regulatory programs associated with other cell states and limit cancer cell plasticity. The researchers tested predictions in patient-derived cell lines and found in mice that tumors with increased MYT1L activity grew more slowly and invaded less surrounding brain tissue, while the animals also survived longer. This offers a mechanistic point for intervention, but has not yet established a treatment that can be used in patients.

The two studies map where cells might go and which regulatory factors influence those transitions, respectively. Because they share data, they cannot be regarded as independent replication in different patient groups. GBM-Space is now available for interactive exploration and data downloads, allowing researchers to further compare gene expression, cell states, and tissue locations to test whether early states can be targeted or the microenvironment supporting tumor progression can be altered.

The atlas’s value lies in giving treatment research more specific, testable questions. More than a million cell nuclei provide fine resolution, yet the number of patients remains just 12. Whether the shared pathways apply to a broader range of tumors, whether blocking state transitions is safe, and whether it can improve patient survival all require further research. The findings provide a new basis for understanding the disease and designing experiments, but do not yet demonstrate that any new therapy is effective in humans.

References

  1. Cancer Research UK
  2. Wellcome Sanger Institute
  3. German Cancer Research Center (DKFZ)
  4. GBM-Space research consortium