Abstract Objective Pediatric brain tumors, including relpased medulloblastoma and diffuse midline gliomas, are the amongst most lethal childhood cancers and remain largely unresponsive to current immunotherapies. Given their considerable difference from adult brain tumor counterparts and the general lack of immune cell infiltration, pediatric–specific brain tumor models are crucial for immunotherapy testing. Here, we developed a method to generate and maintain patient-derived tumor organoids (PDTOs) for characterization of the immune landscape and comparison to the primary tumor. Further, we identified mechanisms of immune cell activation and evasion for the pre-clinical testing of targeted immunomodulatory strategies. Materials and Methods Fresh tumor specimens and matched peripheral blood mononuclear cells (PBMCs) were obtained from pediatric patients undergoing neurosurgery at Cohen Children’s Medical Center. We generated PDTOs and performed baseline immunophenotyping of the organoids in monoculture by multiplex flow cytometry. Following this initial characterization, PDTOs were co-cultured with PBMCs obtained either from the patient or from an age and sex-matched healthy donor. The co-culture system was subsequently immunophenotyped using flow cytometry at Days 1 and 7 of co-culture. Results from Day 1 were compared to Day 0 immunophenotyping of the ‘untouched’ primary tumor. Lastly, matched FFPE specimens from day of surgery were analyzed by immunofluorescence microscopy to validate immune cell states and distinguish tumor heterogeneity. Results We successfully generated 10 PDTOs from pediatric brain tumor specimens. Baseline (Day 0) PDTOs exhibited heterogeneous immune infiltration, including CD8+, CD4+, CD11b+, and CD56+ populations, which was preserved through day 7 in monoculture. Flow cytometric analysis demonstrated that PDTOs co-cultured with healthy donor PBMCs recapitulated tumor–immune interactions observed in Day 0 primary PDTOs. Consistently, multiplex imaging of archival FFPE tumor specimens revealed mixed immune populations concordant with baseline PDTO immunophenotyping. PBMC co-culture induced IFN-γ release, accompanied by PD-L1 upregulation and subsequent T cell exhaustion in a subset of samples. Distinct tumor subtypes demonstrated differential inhibitory receptor expression and signaling profiles. Conclusion This study establishes a robust platform to model immune features of pediatric brain tumors and highlights the utility of tumor-immune co-cultures for identifying and testing immunomodulatory therapies that selectively kill tumor cells.
Rachelkar et al. (Tue,) studied this question.