Molecular Neuro-Oncology Research

Medulloblastoma (MB) arises in the cerebellum, from where it can invade locally and spread to the meninges of the brain and spinal cord. The neural tumor microenvironment (nTME) of the cerebellum, with its unique cellular, structural, and chemical composition, shapes how MB cells behave. In turn, the growing tumor continuously remodels the nTME through direct contact and by releasing proteins, lipids, nucleic acids, and metabolites. This reciprocal interplay, and its consequences for tumor growth and progression, remains incompletely understood.

Using organotypic cerebellar slice cultures and advanced imaging technologies, we investigate how the growth and invasion of Sonic Hedgehog (SHH) MB cells modify the nTME, and how these changes in turn control tumor cell behavior.

Our primary goal is to identify the molecular mediators of this dynamic reciprocity and define their role in tumor growth and tissue invasion. Our secondary goal is to explore strategies that target these mediators to suppress MB growth and invasion.

Cells communicate in part by releasing extracellular vesicles (EVs), which carry proteins, RNA, and metabolites from the cell of origin. How medulloblastoma (MB)-derived EVs affect cells in the surrounding cerebellar tissue remains poorly understood.

Using biochemical, proteomic, and sensor-based methods, we determine the composition, regulation, and release–uptake dynamics of MB-derived EVs, identify their target cells, and characterize the transcriptomic changes they induce in those cells.

Our primary goal is to identify druggable modulators of EV biogenesis and release in tumor cells. Our secondary goal is to establish a rational basis for diagnostic or predictive analysis of EV content in blood or cerebrospinal fluid.

The tumor-associated surface molecule B7-H3 (CD276) is consistently expressed across many primary pediatric brain tumors, including medulloblastoma (MB), and is associated with poor outcomes. In recent years, B7-H3 has emerged as one of the most compelling cell-surface targets in pediatric CNS malignancies. Yet little is known about its biology in MB: how B7-H3 shapes tumor cell behavior and tissue responses, and how B7-H3-targeted therapies affect signaling between the tumor and its microenvironment.

Using live-cell imaging, gain- and loss-of-function studies, and B7-H3-directed nanobodies, we characterize B7-H3 function in the cerebellar tissue context and explore nanobodies as a strategy to deliver diagnostic and therapeutic payloads to tumor cells.

We use live- and fixed-cell imaging to analyze single cells, both in isolation and in their tissue context, qualitatively and quantitatively. The lab has developed pipelines for robust quantification of cell dynamics and functions in two- and three-dimensional environments, and uses fluorescent probes and sensors to interrogate specific proteins and measure cellular responses to treatment.

Conventional image analysis cannot capture the changes in F-actin architecture that define motile, invasive cancer cell behavior. To quantify these phenotype-defining features, we are integrating deep learning into our phenotype analyses. We are developing image classifiers that predict treatment efficacy, and generative models that build an interpretable representation of the phenotypic information in the F-actin cytoskeleton of control and treated cells. We use these tools to assess both new and established compounds in an unbiased way.