Repository logo
 

In vitro modelling of human glioblastoma migration and angiogenesis


Type

Thesis

Change log

Authors

Stavropoulou-Tatla, Agavi 

Abstract

Glioblastoma (GBM) invasion and angiogenesis are critical for the tumour's fast growth and recurrence. A significant limitation in the development of innovative anti-invasive and anti-angiogenic treatments is the lack of personalised in vitro experimental models to accurately recreate the physiological and pathophysiological properties of the brain.

This project focuses on the development of an injury-free ‘wound healing’ and a vascularised tumoroid in vitro model using patient-derived GBM cells. These models are used to recapitulate how individual components of the GBM's complex brain microenvironment, such as low oxygen tension (e.g. 1% O2), vasculature related stromal cells (e.g. human umbilical vein endothelial cells (HUVECs), human dermal fibroblasts (HDF)) and growth factors (e.g. VEGF, bFGF), support the GBM migration and angiogenesis.

In the first half of the thesis, an injury-free ‘wound healing’ model was developed that aimed at introducing an accurately defined cell-free area on a confluent GBM monolayer to investigate GBM migration in 2D. The cells became progressively motile perpendicularly to the border. Quantitative results were reported by mapping the velocity field of collectively migrating cells using particle image velocimetry. Hypoxia (1% O2) showed a higher average cell velocity and polar order parameter (meaning that the cells move more in the outward direction normal to the initial border) compared to normoxia (20% O2). Replacing one side of the ‘wound’ with a HUVECs monolayer was also found to increase the GBM average cell velocity and polar order parameter. These results are consistent with in vivo observations of hypoxia promoting GBM invasion and blood vessels chemotactically attracting GBM cells.

In the second half of the thesis, a GBM vascularised tumouroid model was developed to investigate GBM angiogenesis in 3D by embeddeding GBM/HUVECs spheroids in HDF containing fibrin gel. Co-culturing HUVECs with HDF lead to HUVECs undergoing an angiogenic-like response, sprouting radially outwards forming capillaries. Primary, stem-like cell rich, GBM cells were found to integrate with the capillaries. Their presence made HUVECs produce vasculature of higher explant area but lower interconnectivity. They were able to trigger some HUVECs sprouting even in the absence of HDF. The addition of exogenous growth factors (200 ng/mL VEGF and bFGF) enhanced this effect for cells cultured at normoxia (20% O2) but not at hypoxia (1% O2) suggesting that the GBM cells at hypoxia are already producing a significant amount of growth factors. Finally, the primary GBM cells and HUVECs were found to be strongly co-localised as indicated by the Pearson's correlation coefficient. CD31 staining hinted that the primary GBM cells have acquired an endothelial-like behaviour. Therefore, in all cases, primary GBM cells were found to promote capillary-like networks and even to acquire endothelial-like characteristics, which is reported to occur in vivo.

The findings presented throughout this thesis demonstrate that the two tissue-engineered GBM models analysed can be used as platforms to study GBM invasion and angiogenesis mechanisms in vitro. We observe some key behaviours which are reported to occur in vivo. Still, we need to advance our knowledge a lot more before those type of personalised models can serve as clinical prediction models.

Description

Date

2020

Advisors

Markaki, Athina

Keywords

glioblastoma, migration, angiogenesis

Qualification

Awarding Institution

University of Cambridge