建立具有临床相关性的 ETMR 体外模型及其对药物疗效研究的意义

Background: Patient-derived 3D models of pediatric brain tumors are increasingly used in preclinical drug screening due to their greater clinical relevance. However, no guidelines or widely adopted methods exist, and most drug screening studies still rely on small tumorspheres. The primary ETMR line, BT183, retains key molecular features, including C19MC amplification and LIN28A overexpression, yet its conventional tumorsphere format fails to recapitulate the characteristic multilayered rosettes observed in patients, lacks cellular complexity, and remains size-restricted for functional studies and drug screening. To overcome these limitations, we established a more clinically relevant 3D pseudo-organoid model by optimizing the culture medium, promoting growth from 700-3000 µm in diameter, and co-culturing with non-tumor cell populations, including blood brain barrier (BBB) and immune cells. Methods: BT183 cells were cultured as tumorspheres in NeuroCult (NC) or as pseudo-organoids in tumor stem cell medium (TSM). Rosette abundance was quantified in H&E sections, while rosette architecture and size were assessed by 3D multiplex immunofluorescence (mIF) of optically cleared whole pseudo-organoids followed by confocal imaging and morphometric analysis. Drug screening was performed across multiple sizes and drug penetration was evaluated by mass spectrometry imaging (MSI). Multicellular BBB pseudo-organoids were generated by staged seeding of tumor, microglia, neuronal, endothelial, and pericyte populations in ratios reflecting ETMR patient samples; Integration and layering were evaluated by mIF. Large (> 2 mm) organoids were produced with or without Matrigel under constant agitation. Results: Pseudo-organoids grown in TSM contained significantly more rosettes compared to tumorspheres grown in NC. There was a divergence in the size of the spheroids once they reached 500 µm, with the rosette forming model growing to larger sizes faster than the conventional model. Notably, the rosettes forming pseudo-organoid model demonstrated drug-dependent enhanced chemoresistance to clinically relevant agents compared to the standard model. MSI provided quantification of drug penetration and spatial distribution across models to correlate with efficacy. BBB pseudo-organoids showed successful incorporation of cell populations, including endothelial cells and pericytes. Prolonged culture under agitation, coupled with Matrigel dome embedding, further supported the formation of large, structurally complex organoids. Conclusions: Our ETMR pseudo-organoid model recapitulates the multilayered rosette architecture characteristic of patient tumors. Ongoing work is being carried out to further refine this system into multicellular, large-scale 3D organoids that further advance preclinical modeling and therapeutic testing in embryonal tumors with multilayered rosettes.