New 3D brain organoids reveal how glioblastoma evades treatment

UCLA scientists have created advanced miniature 3D tumour organoid models that allow glioblastoma to be studied in an environment that closely mirrors the human brain, offering novel insight into how the aggressive cancer interacts with surrounding brain cells and the immune system to evade treatment.

Described in two complementary studies, the models are built from human stem cells and recreate the complex mix of cell types found in the brain. This allows researchers to observe how patient-derived tumours communicate with healthy brain tissue, revealing weaknesses that could be targeted with new personalised therapies.

“Glioblastoma has been incredibly difficult to treat in part because we haven’t had good ways to study how tumours behave in a truly human brain environment,” said senior author Dr Aparna Bhaduri, Assistant Professor of Medicine and Biological Chemistry at the David Geffen School of Medicine at UCLA and Investigator at the UCLA Health Jonsson Comprehensive Cancer Center. “These new models developed by my lab help us understand how interactions with brain tissue and immune cells contribute to therapy failure, which is crucial for developing more effective, patient-specific treatments.”

Revealing a hidden driver of tumour aggression

The first system, known as human organoid tumour transplantation (HOTT), shows how glioblastoma cells interact with neighbouring brain cells to change their identity, invade tissue and resist therapy.

Using HOTT, researchers identified a key communication pathway between tumour cells and surrounding brain cells involving a protein called PTPRZ1, which is expressed by both cell types in glioblastoma tissue. The protein emerged as a major regulator of tumour behaviour that helps determine how aggressive the cancer becomes.

When PTPRZ1 levels were reduced specifically in the brain cells of the organoids, tumour cells shifted into a more invasive state despite not being directly altered. They activated genes linked to movement and tissue invasion and formed longer tumour microtubes, thin extensions that help tumours spread through the brain and resist treatment. Unexpectedly, these effects did not rely on PTPRZ1’s usual enzyme activity, pointing to a previously unrecognised signalling role within the tumour environment.

“This study shows that glioblastoma is strongly influenced by the surrounding brain cells, not just the cancer itself,” said Dr Bhaduri. “By identifying PTPRZ1 as a new regulator of tumour behaviour, we’re revealing hidden communication pathways and demonstrating how these organoid models can help uncover more effective therapeutic targets.”

Mimicking immune responses to therapy

The second model, called immune-human organoid tumour transplantation (iHOTT), adds immune system components to the organoids, allowing scientists to study how immune cells influence tumour growth and resistance to treatment.

The model preserves key features of both tumour and immune cells, including CD4 and CD8 T cells, B cells, NK cells and myeloid cells. This makes it possible to track how immune cells behave, communicate and change in response to the tumour.

When the organoids were treated with pembrolizumab, a widely used PD-1 checkpoint inhibitor, the immune system became more active, with increases in CD4 T cells, B cells and immune signalling. However, tumour cells continued to survive and grow.

“These immune changes observed in the lab closely mirrored what happens in real patients treated with pembrolizumab,” said Dr Bhaduri. “The same shifts in immune cell populations occurred, the same communication pathways between immune cells were activated, and even rare or unconventional immune cell types expanded in similar ways. This demonstrates that iHOTT faithfully reproduces patient-like immune responses in a human-relevant system.”

Further analysis showed that pembrolizumab increased T-cell diversity, but the expanded T cells were unique to each patient rather than shared. The strongest growth was seen in CD4 T cells with stem-like properties, helping explain why immunotherapy has limited success in glioblastoma.

Next steps

These human brain organoid models give scientists a powerful new look into how glioblastoma behaves within the brain and responds to the immune system. By discovering previously hidden interactions that drive tumour aggression and limit the effectiveness of current therapies, the research lays important groundwork for more precise, patient-specific approaches to treating one of the most challenging and deadly brain cancers.

Next, the researchers plan to use these organoid platforms to test new drug combinations, explore ways to overcome immune resistance and refine personalised treatment strategies tailored to the unique biology of each patient’s tumour.

“Together, these studies show that these patient-specific organoid models offer a powerful tool to uncover hidden tumour interactions and test new therapies, bringing personalised treatment for this deadly cancer a step closer,” said Dr Bhaduri.