Immune Cells Cluster, Not Crowd, in Melanoma That Spreads to the Brain

When melanoma spreads to the brain, the outlook for patients has historically been grim, and one of the reasons lies in how the brain’s unique environment reshapes the immune cells that try to fight the tumor. A new study from researchers at Radboud University Medical Center in Nijmegen, the Netherlands, published in Cancer Immunology, Immunotherapy, reveals that the architecture of tumor-infiltrating lymphocytes in melanoma brain metastases is fundamentally different from that in metastases elsewhere in the body. Remarkably, the study suggests that how immune cells are arranged in space, rather than how many of them are present, may be the decisive factor that predicts whether patients survive longer and respond to modern immunotherapy.

The research team, led by Mark W. D. Sweep and senior author Kalijn F. Bol, analyzed tissue from 55 patients with melanoma, examining 60 brain metastases and 53 paired extracerebral metastases, meaning tumors from the same patients that had spread to other organs. Using multiplex immunohistochemistry, a staining technique that allows several immune cell markers to be visualized simultaneously on a single tissue section, the researchers could identify and localize individual lymphocytes within each tumor. Crucially, they went beyond simply counting cells: they recorded the two-dimensional coordinates of every lymphocyte and applied a mathematical tool called the local correlation function to quantify whether immune cells were scattered randomly across the tumor or gathered into tight clusters.

The first major finding was stark. Brain metastases exhibited a sixfold reduction in lymphocyte density compared with paired metastases elsewhere in the body from the same patients. In other words, the tumors that had colonized the brain were far more sparsely patrolled by immune cells than their counterparts in the skin, lung, or other organs. Gene expression analysis performed on a subset of these samples, 13 brain metastases and 21 paired extracerebral tumors, reinforced this picture, showing markedly reduced expression of immune signatures, sets of genes whose collective activity indicates an active immune response, in the brain lesions.

This paucity of immune infiltration is consistent with the brain’s reputation as an immunologically privileged site. The central nervous system is shielded by the blood-brain barrier and maintained in a deliberately suppressive state to protect delicate neural tissue from inflammation. Tumors that seed the brain appear to exploit this neuroprotective environment, creating conditions in which fewer lymphocytes gain entry or survive once inside. For patients whose brain metastases cause neurological symptoms, the situation is often compounded by dexamethasone, a glucocorticoid steroid routinely prescribed to reduce swelling around brain tumors, which itself is a powerful immunosuppressant.

Yet the study’s most striking insight came from the spatial analysis of the lymphocytes that were present. Despite their lower abundance, lymphocytes within brain metastases displayed a higher degree of spatial clustering than lymphocytes in extracerebral metastases. The immune cells in brain tumors were not merely fewer; they were organized differently, gathered into localized groups rather than dispersed evenly throughout the tumor tissue. This finding implies that the few immune cells that do penetrate brain metastases congregate, perhaps around particular tumor regions, vascular structures, or chemokine signals, and that this congregated pattern carries biological meaning.

That meaning became clear when the researchers linked spatial organization to clinical outcomes. Low spatial clustering, particularly of CD8 T cells, the cytotoxic lymphocytes responsible for directly killing tumor cells, correlated with favorable survival after neurosurgery and with better responses to immune checkpoint inhibitors, the antibody therapies that unleash T cells against cancer. This result inverts the conventional emphasis in tumor immunology, where the density of tumor-infiltrating lymphocytes has long served as the standard biomarker of an inflamed, immunotherapy-responsive tumor. In melanoma brain metastases, density alone was not the predictive variable; the spatial pattern of infiltration was. A tumor containing relatively few but broadly distributed CD8 T cells fared better than one in which lymphocytes were packed into isolated islands.

The mechanistic interpretation of this finding remains an open question, but the authors suggest that the spatial pattern of infiltration may be functionally more important than infiltration numbers. Scattered lymphocytes may be better positioned to survey tumor cells across the entire lesion, recognize antigens, and sustain tumor control, whereas tightly clustered cells may be sequestered in niches where they exert little influence on the bulk of the malignancy. If validated, this principle could reshape how immunotherapy strategies are tailored for patients with brain metastases, focusing attention not just on attracting immune cells into the brain but on how they distribute themselves once there.

The study also delivered a sobering finding about a routine clinical practice. Prolonged exposure to dexamethasone, defined in the study as more than 30 days, correlated with reduced immune signatures and increased tumor-promoting signatures in brain metastases relative to shorter exposures. Because dexamethasone is frequently given to manage symptoms in patients with symptomatic brain metastases, sometimes for extended periods, this observation raises the possibility that the very medication used to keep patients comfortable may be actively degrading the immune landscape of their tumors. The authors emphasize that the negative effects of prolonged dexamethasone treatment on brain metastases should be taken into account and clarified in follow-up studies, a point with direct implications for how clinicians weigh steroid duration against the goal of preserving immunotherapy sensitivity.

Half of all patients with advanced cutaneous melanoma eventually develop brain metastases, making this one of the most feared complications of the disease and a major driver of mortality. Immune checkpoint inhibitors have transformed outcomes for many melanoma patients, but responses in the brain remain less predictable than elsewhere, and this study offers a potential explanation rooted in tumor geography. The brain’s suppressive milieu, potentially intensified by corticosteroid exposure, produces lesions that are immunologically sparse yet spatially clustered, a configuration unlike anything seen in extracerebral disease and one that standard density-based measures fail to capture.

Methodologically, the work demonstrates the power of combining multiplex immunohistochemistry with spatial statistics in oncology. The local correlation function, adapted from spatial ecology and physics, converts raw cell coordinates into a quantitative description of tissue organization, allowing researchers to test whether immune architecture, not merely immune abundance, shapes disease course. As single-cell and spatial technologies spread through cancer research, this study stands as an early signal that the geography of immune cells inside tumors will join density, phenotype, and gene expression as a critical dimension of tumor immunology, particularly in immunologically specialized sites like the brain where every infiltrating lymphocyte may count for more.

Subject of Research: Spatial organization of tumor-infiltrating lymphocytes in melanoma brain metastases and its link to immunotherapy response

Article Title: Spatial clustering of tumor-infiltrating lymphocytes in symptomatic melanoma brain metastases

Article References: Sweep, M. W. D., Gorris, M. A. J., van Mierlo, G., Verrijp, K., Küsters, B., ter Laan, M., Textor, J., van Herpen, C. M. L., de Vries, I. J. M., & Bol, K. F. (2026). Spatial clustering of tumor-infiltrating lymphocytes in symptomatic melanoma brain metastases. Cancer Immunology, Immunotherapy. https://doi.org/10.1007/s00262-026-04550-5

Image Credits: AI Generated

DOI: 10.1007/s00262-026-04550-5

Keywords: melanoma, brain metastases, tumor-infiltrating lymphocytes, spatial clustering, immunotherapy, immune checkpoint inhibitors, CD8 T cells, dexamethasone, multiplex immunohistochemistry, transcriptomics, tumor microenvironment, Radboud University Medical Center