Tumor microenvironment-preserving gliosarcoma organoids as an in vitro preclinical platform: a comparative analysis with glioblastoma models

Abstract

Background: Gliosarcoma (GS) is a rare variant of glioblastoma (GBM), characterized by biphasic glial and sarcomatous histology and poor prognosis. Despite its distinct clinical features, GS remains underrepresented in glioma research due to the lack of biologically faithful and experimentally tractable models. Methods: We established patient-derived gliosarcoma organoids (GSOs) from freshly resected tumors using a suspension culture system without enzymatic dissociation. Histological, molecular, and functional properties were evaluated using H&E staining, immunohistochemistry, whole-exome sequencing, 3D invasion assays, and single-cell RNA sequencing (scRNA-seq). Drug response assays were performed using temozolomide and ANA-12 (NTRK2 inhibitor). Results: GSOs preserved key histological features, genetic alterations, and tumor microenvironmental cell populations from the original tumors. They stable growth and viability over extended culture periods and maintained integrity following cryopreservation and recovery, supporting their long-term utility. 3D invasion assays further revealed infiltrative behavior, consistent with the aggressive nature of GS. Histological and molecular analyses revealed that GSOs retained glial and mesenchymal differentiation, diverse non-malignant stromal cells, and case-specific somatic alterations. Comparative scRNA-seq revealed distinct transcriptional programs: GSOs were enriched for fibroblast-like and oligodendrocyte progenitor-like states, while glioblastoma organoids (GBOs) displayed astrocyte-like differentiation and high connectivity signatures. Functional assays confirmed consistent sensitivity of GSOs to temozolomide, and selective therapeutic response to NTRK2 inhibition was observed in the GSO harboring an NTRK2 alteration, supporting the utility for genomic-context-guided therapy. Conclusion: Together, GSOs constitute a tractable and translationally relevant model that faithfully reflects the cellular complexity, genetic landscape, and tumor-specific heterogeneity. This model addresses a critical gap in GS biology and supports its integration into precision oncology. © The Author(s) 2025.