Chick Chorioallantoic Membrane as a Model for Glioblastoma Growth, Invasion and Treatment

Abstract

Glioblastoma is the most aggressive primary brain tumour, with a median patient survival of only 15 months. Its rapid proliferation, high invasion, poor therapeutic response, and recurrence, combined with substantial inter- and intra-tumoural heterogeneity, complicate the development of efficient therapies and highlight the need for personalised treatment strategies. Personalised treatment strategies require rapid and efficient experimental models capable of capturing diverse tumour behaviours. The chick chorioallantoic membrane (CAM) model represents a promising intermediate platform between in vitro assays and rodent xenografts. This thesis evaluated the suitability of the CAM model for studying glioblastoma tumour growth, invasion, and treatment response using 17 patient-derived Human Glioblastoma Cell Culture (HGCC) cell lines. All 17 cell lines formed tumours with high xenografting efficiency (> 90%), demonstrating that the CAM environment efficiently supports tumour development. Histological analyses revealed morphological heterogeneity and variation in invasion levels, reflecting the heterogeneity of glioblastoma phenotypes. The CAM vasculature consistently formed around the tumours, further supporting glioblastoma growth. Ki67 immunostaining confirmed active proliferation across all xenografts, with proliferation rates varying between cell lines. High Nestin expression demonstrated preservation of glioblastoma stem/progenitor like characteristics in the CAM model. Tumour weight analysis revealed variable tumour growth progression, underlining limitations of weight-based growth assessment. To evaluate the CAM model’s suitability for drug screening, nine GFP-Luc labelled HGCC cell lines were treated with the proteasome inhibitor bortezomib, previously identified as a potent agent against glioblastoma in large-scale drug screening. Due to the limitations in weight-based growth assessment, bioluminescence imaging (BLI) was used before and after the treatment to accurately monitor treatment responses. The change in cellular viability in response to bortezomib treatment was analysed by calculating relative change in BLI signal intensity for each group. Five cell lines exhibited significant decrease in cellular viability, whereas four cell lines showed resistance against the treatment. These sensitivity patterns aligned with in vitro and mouse xenograft results, supporting the translational relevance of CAM based drug testing. Overall, this study demonstrates that the CAM model enables efficient glioblastoma tumour formation and analysis of tumour growth and invasion. The CAM model holds promise as an effective model for glioblastoma treatment screening and has potential as a bridge between large-scale in vitro drug screening and mammalian in vivo studies. The short experimental timeline of the CAM model is both an advantage and a limitation. It enables rapid results, which is an important characteristic of a model for patient-specific treatment testing, but the short duration does not enable long-term studies of tumour progression and treatment effects. Additional research is needed to further optimise the CAM model for glioblastoma studies and drug screening.