3D Brain Vasculature Microenvironment Model Captures Invasive Behavior and Gene Signatures of Glioblastoma

Introduction: Glioblastoma (GBM) is the most aggressive and lethal form of brain cancer, with a median survival of only 15-18 months despite multimodal treatment strategies. Traditional two-dimensional (2D) cell culture systems fail to replicate the complex tumor microenvironment (TME) and GBM heterogeneity observed in vivo, which are crucial for understanding tumor biology and drug response. Conventional in vivo models like xenografts and genetically engineered mouse models offer valuable insights but present the complexity of the TME all at once, making it challenging to dissect the intricate interactions between tumor cells and the surrounding stromal, neuronal, and glial networks. The TME of GBM is characterized by a unique extracellular matrix (ECM), dysfunctional dense vascular networks, various non-neoplastic glial cell types, and immune cells, all contributing to tumor progression, invasion, immune suppression, and therapy resistance.
Advanced biomimetic three-dimensional (3D) culture systems that mimic the brain microenvironment with tissue-level complexity are needed to address these limitations. Brain organoids and organ-on-a-chip platforms have shown promise in GBM research by providing a more realistic representation of tumor-host interactions. However, these systems lack functional human vasculature, limiting their ability to fully mimic the complex vascular network of the brain. We developed an improved 3D brain vascular model incorporating brain-specific endothelial cells, mural cells, and patient-derived GBM cells. This study aimed to compare tumor cell behavior across traditional 2D cultures, various 3D models, and an orthotopic in vivo intracranial transplant model, focusing on two lines of patient-derived GBM stem cells (GSC).

Materials and

Methods: To build a biomimetic brain vascular niche model, we seeded brain endothelial cells (bECs) in a 3D fibrin matrix along with astrocytes and brain pericytes. The vascular network formation was monitored and optimized by adjusting the seeding densities of astrocytes and pericytes. We tested different combinations of cell densities to achieve optimal vascular network connectivity, vessel width, hollow tube formation, and overall vascular area. For modeling GBM invasion and interaction with brain vasculature, we used patient-derived GBM stem cell lines (SD2 and SD3) and compared their behavior in 2D cultures, 3D gel (3D G; Glioblastoma) without vasculature, and the optimized 3D human brain vascular niche model (3D GAPE; Glioblastoma with Astrocytes, Pericytes, and Endothelial cells) (Figure 1a). We further compared the effects of dispersed cell culture versus spheroid aggregates in these models. RNA sequencing (RNA-seq) was conducted on GBM cells isolated from 2D, 3D G, 3D GAPE models, and in vivo intracranial transplants to analyze differential gene expression.

Results, Conclusions, and Discussions: Optimization of the 3D human brain vascular niche model demonstrated that co-culturing bECs with both astrocytes and pericytes significantly enhanced vascular network formation compared to using either cell type alone (Figure 1b). The optimized 3D GAPE model enabled long-term culture of patient-derived GBM cells, maintaining structural integrity without significant matrix degradation. GBM cells in the 3D GAPE model exhibited distinct invasion patterns and a stronger vascular association compared to the 3D G model (Figure 1c).
RNA-seq analysis revealed that GBM cells in the 3D GAPE model had 1438 differentially expressed genes (DEGs) compared to 2D cultures, with a significant overlap of 533 DEGs with in vivo samples (Figure 1d). These DEGs were mainly associated with CNS functions, including neurogenesis, synaptic signaling, and gliogenesis (Figure 1e-f). The shared gene signature between the 3D GAPE model and in vivo conditions predicted worse prognosis for GBM patients (Figure 1g).
Our 3D brain vascular niche model effectively recapitulates key features of GBM heterogeneity, including invasion behavior, vascular association, and gene reprogramming. This model bridges the gap between traditional 2D cultures and complex in vivo models, offering a stable and biomimetic environment for long-term study of tumor cell behavior. The 3D GAPE model revealed phenotypic differences between different patient-derived GBM lines and provided insights into the transcriptional changes associated with tumor microenvironment interactions.
The study's findings underscore the critical role of the brain glia-vascular unit in GBM gene reprogramming, particularly in synaptic signaling and immune suppression. The ability of the 3D GAPE model to maintain a quiescent niche and provide protection against chemotherapy further demonstrates its potential as a tool for GBM research and therapeutic development. Future directions include incorporating neurons and immune cells to better understand the malignant network between GBM cells and the neuronal/glial environment.
In summary, our advanced 3D brain vascular niche model offers a versatile and physiologically relevant platform for studying GBM biology, advancing drug discovery, and improving therapeutic strategies for managing this aggressive brain cancer.

Acknowledgements (Optional):