Unmasking phenotypic heterogeneity and extracellular matrix architecture in neuroblastoma through orthotopic xenograft modeling

Neuroblastoma (NB) is a highly heterogeneous pediatric tumor in which the extracellular matrix (ECM) critically influences tumor growth, cell lineage plasticity, and therapeutic resistance. However, the interplay of ECM remodeling with the traceable origin and fate of tumor cell phenotypes remains poorly understood. Paired-like homeobox 2b (PHOX2B), a pivotal diagnostic marker reflecting neuronal differentiation states of NB, was examined across multiple NB models. Our analysis revealed heterogeneous PHOX2B expression, including PHOX2B-negative foci, indicating pronounced intratumoral heterogeneity. To unravel the mechanisms underlying this behavior, we developed an integrated platform combining orthotopic xenografts-derived from 2D cultures and 3D tyramine-modified gelatin/silk fibroin (GTA-sf) hydrogel systems, and digital analysis. Original and NB models diverged in both cellular and microenvironmental features, despite shared PHOX2B positivity and reticulin fibers. Original tumors were identified as poorly differentiated NB, showing high PHOX2B expression, whereas SH-SY5Y xenografts displayed mesenchymal-like PHOX2B-negative foci, in which SOX9 and vimentin marked plastic niches within a dense, aligned reticulin fiber network arranged in trabeculae. In contrast, SK-N-BE(2) and patient-derived xenograft (PDX1) tumors showed no upregulation of mesenchymal-type markers within PHOX2B-negative niches and exhibited shorter, curved, and disorganized fibers. Unlike other models, PDX2 exhibited uniform PHOX2B expression without negative foci, characterized by densely packed tumor cells and thick fibrillar structures consistent with high matrix stiffness. Collectively, these findings demonstrate that PHOX2B heterogeneity and ECM stiffness, including fibrillar organization, interact in a model-dependent manner to shape tumor architecture and microenvironmental remodeling. Unmasking these complex interactions is essential for identifying novel therapeutic vulnerabilities and improving clinical outcomes.