Gliomas are among the most difficult cancers to treat, not only because of their location and invasive growth, but also because they develop inside an organ whose relationship with the immune system is fundamentally different from that of most tissues. A new perspective published in Nature Reviews Cancer argues that the limited success of immunotherapy in brain tumours cannot be explained solely by the biology of glioma cells. Instead, the architecture of the central nervous system (CNS), and the specialized barriers that control movement between the brain and the circulation, may determine whether immune cells and therapeutic molecules can even reach the tumour. These barriers preserve the delicate environment required for neuronal function, but they can also leave malignant cells hidden from immune surveillance and restrict the activity of modern cancer treatments.
The CNS is protected by a network of interfaces rather than by a single wall. The best-known structure is the blood–brain barrier, formed primarily by tightly connected endothelial cells lining the brain’s blood vessels. These cells work with pericytes, astrocytes and components of the extracellular matrix to regulate the passage of substances from the bloodstream into the neural tissue. Tight junctions between endothelial cells limit the movement of water-soluble molecules, while transport proteins selectively control the entry and removal of nutrients, metabolites and drugs. Additional interfaces, including the blood–cerebrospinal fluid barrier and barriers associated with the meninges, create separate compartments around the brain. Together, these systems help prevent toxins, pathogens and uncontrolled immune activity from disturbing the CNS.
This organization has important consequences for immune surveillance. In many organs, immune cells continuously patrol tissue, enter sites of inflammation and rapidly encounter abnormal cells. In the healthy brain parenchyma, however, immune access is more tightly regulated. Immune monitoring is concentrated at CNS borders, particularly in the subarachnoid space surrounding the brain and spinal cord, the cerebrospinal fluid and perivascular spaces that accompany blood vessels. These regions function as strategic observation points, allowing immune mediators and cells to monitor the nervous system without exposing neurons to the potentially damaging effects of unrestricted inflammation. The result is a carefully controlled form of surveillance that protects brain function but may be poorly suited to detecting tumours developing deep within the neural tissue.
Gliomas arise from, or resemble, cells within the CNS parenchyma and can grow through regions that are naturally shielded from circulating immune components. As tumour cells expand, they may therefore remain outside the main routes used by immune cells to enter or inspect the brain. This spatial separation creates a major obstacle for immunotherapies. Immune checkpoint inhibitors, for example, are designed to release molecular brakes that prevent T cells from attacking cancer. Yet removing those brakes has limited value if activated T cells cannot efficiently enter the tumour or remain unable to penetrate its surrounding tissue. The same principle applies to cancer vaccines, which stimulate tumour-specific immune responses, and to cellular therapies that depend on the physical delivery of immune cells to malignant sites.
Adoptive T cell therapies illustrate the problem particularly clearly. Chimeric antigen receptor, or CAR, T cells are engineered to recognize specific molecules on cancer cells, while T cell receptor, or TCR, transgenic cells are modified to detect tumour-derived peptides presented by major histocompatibility complex molecules. Once activated, these cells must circulate, cross relevant CNS interfaces, move through the tumour microenvironment and maintain their function in a setting that can suppress immune activity. Brain barriers can restrict each stage. They may limit the number of therapeutic cells entering the CNS, control the molecules that reach the tumour and create compartmentalized environments in which immune cells receive incomplete or altered signals. A powerful T cell product may therefore show striking activity in laboratory systems while producing weaker responses in patients with parenchymal gliomas.
The review also highlights that gliomas are not passive occupants of the brain’s protected environment. Emerging evidence suggests that these tumours can actively remodel barrier function, reinforcing the separation between the tumour and the immune system. Cancer cells and associated stromal or vascular cells may alter endothelial properties, modify extracellular matrix structures and influence the activity of astrocytes and pericytes. Such changes can affect vascular permeability, transport systems and the routes through which immune cells communicate with or enter the tumour. Importantly, barrier remodeling is not necessarily equivalent to simply making the blood–brain barrier “leaky.” A tumour may disrupt some barrier functions while preserving or intensifying others, producing a complex interface that permits selected molecules or cells to pass but continues to block effective immune access.
This complexity helps explain why the presence of blood vessels inside a glioma does not guarantee that immune therapies can reach their targets. Tumour-associated vessels are often structurally abnormal, yet their permeability and transport properties can vary across the same lesion. Some regions may display barrier breakdown, while neighbouring areas retain restrictive endothelial characteristics. The tumour microenvironment can also contain suppressive myeloid cells, altered glial cells and molecular signals that diminish T cell activation. In this setting, improved access alone may not be sufficient; therapy must account for the identity, activation state and distribution of the immune cells that arrive. Brain barriers should therefore be viewed as dynamic neuroimmunological interfaces rather than static obstacles.
These observations carry consequences for the design of clinical trials. Immunotherapy studies in glioma commonly focus on tumour genetics, antigen expression, T cell activity and radiographic changes. The authors argue that barrier status and CNS compartmentalization should also become central considerations. Researchers may need to determine whether a treatment reaches the relevant tumour regions, which barrier components regulate that access and how those properties change during disease progression or therapy. Biomarkers reflecting vascular transport, cerebrospinal fluid immune activity, perivascular inflammation or regional barrier function could eventually help identify patients most likely to respond. Treatment schedules might also need to be coordinated with approaches that influence barrier permeability or immune trafficking, although such interventions would have to preserve the protective functions of the CNS and avoid harmful neuroinflammation.
The perspective does not suggest that disrupting brain barriers indiscriminately would solve the problem. These structures are essential for neuronal survival, and uncontrolled opening could allow toxic substances, infectious agents or excessive immune activity into sensitive tissue. Any strategy aimed at improving access must therefore be precise, temporary and compatible with the brain’s physiological requirements. Future therapies may combine tumour-directed immune activation with methods that guide T cells toward CNS borders, enhance their passage through selected interfaces or alter the tumour-associated vasculature without damaging healthy neural tissue. Understanding how CAR T cells, TCR therapies, vaccines and checkpoint inhibitors interact with these barriers could also reveal why responses differ between patients whose tumours appear similar by conventional molecular classifications.
The central message is that successful glioma immunotherapy will require more than identifying the right tumour antigen or engineering a stronger immune cell. It will require understanding the routes that connect the bloodstream, the meninges, the cerebrospinal fluid, the perivascular spaces and the brain parenchyma, as well as the ways gliomas reshape those routes. By placing brain barriers at the centre of glioma immune surveillance, the new analysis frames treatment resistance as a problem of access, communication and compartmentalized biology. For patients with brain tumours, that shift could guide the development of immunotherapies designed not only to recognize cancer, but also to reach it.
Subject of Research: The role of CNS brain barriers in glioma immune surveillance and immunotherapy response.
Article Title: Brain barriers at the crossroads of glioma immune surveillance and immunotherapy response
Article References: Engelhardt, B., Platten, M. Brain barriers at the crossroads of glioma immune surveillance and immunotherapy response. Nature Reviews Cancer (2026). https://doi.org/10.1038/s41568-026-00960-w
Image Credits: AI Generated
DOI: 10.1038/s41568-026-00960-w
Keywords: Glioma, brain barriers, blood–brain barrier, CNS immune privilege, neuroimmunology, immune surveillance, immunotherapy, CAR T cells, TCR therapy, immune checkpoint inhibitors
