Human Brain Tumor Organoid Platform Reveals Immune Responses to Glioblastoma

Glioblastoma (GBM) is the deadliest common brain cancer in adults, and despite the revolution of checkpoint immunotherapy in other cancers, GBM remains largely refractory. A core obstacle is the absence of human, immunocompetent models that faithfully replicate the tumor microenvironment and allow mechanistic testing of immune therapies. Mouse and engineered rodent models fail to capture human cytokine signaling and immune diversity, making translational insights unreliable.

To fill this gap, researchers at UCLA built iHOTT by extending a previously established human organoid tumor transplantation (HOTT) platform. Human cortical organoids grown over 8–12 weeks serve as the host tissue. On the day of surgery, freshly dissociated patient tumor cells are GFP-labeled with lentivirus, then co-transplanted with matched peripheral blood mononuclear cells (PBMCs) at a 1:1 ratio onto those organoids. The system is cultured for 7 days before analysis. PBMCs were chosen over tumor-infiltrating lymphocytes because they have not undergone chronic exhaustion, providing a more responsive immune compartment to model early tumor recognition and clonal dynamics.

Key results spanned multiple analytic modalities. Flow cytometry and immunostaining confirmed engraftment of tumor cells and retention of CD3+ T cells and CD19+ B cells within organoids, mirroring distributions in patient peripheral blood. A 37-plex cytokine panel showed significant upregulation of GBM-associated cytokines—IL-6, IL-8, IL-10, and G-CSF—in co-cultures versus tumor-only or PBMC-only controls, with control experiments confirming the cytokine signatures were driven by tumor-immune interaction rather than lentiviral transduction. Single-cell RNA sequencing of three patient co-cultures confirmed preservation of all canonical tumor subtypes (including oligodendrocyte-like, progenitor-like, and mesenchymal-like populations) and major immune populations. Immune cells showed low exhaustion markers (HAVCR2, LAG3) and high activation markers (GZMB, CD69). Cell-cell interaction analysis revealed tumor oligodendrocyte-like cells as dominant signaling senders, with CD8+ T cells as primary recipients via MHC class I and MIF pathways—consistent with cytotoxic immune surveillance observed in patient tumors.

The platform was then benchmarked against clinical pembrolizumab treatment. iHOTT samples treated with pembrolizumab versus IgG control reproduced cell-type composition shifts—modest expansion of CD4+ T cells and B cells—and immune pathway enrichment (B-cell-mediated immunity, humoral responses, MHC class II signaling) matching patterns observed in pembrolizumab-treated patient tumors. TCR sequencing identified pembrolizumab-driven clonal expansion of stem-like CD4 T cell clonotypes with patient-specific repertoires, underscoring the importance of individualized antigen-targeted strategies.

Caveats include limited myeloid cell recovery ex vivo despite cytokine supplementation (IL-2, IL-15, GM-CSF), meaning the model is better suited for studying T cell dynamics than myeloid biology. The 7-day culture window captures early immune events but may miss longer-term adaptive responses. The organoids themselves are hypoimmunogenic (lacking MHC I/II), which minimizes allogeneic confounding but also means the neural microenvironment contribution is incomplete. Despite these limitations, iHOTT represents a meaningful advance: a fully human, autologous, three-dimensional system for dissecting GBM tumor-immune dynamics and evaluating personalized immunotherapy.