Exploring vasculogenic mimicry and anti-angiogenic therapy resistance within breast cancer patient-derived tissue xenograft models

Therapeutically targeting tumour neovascularisation with anti-angiogenic therapy (AAT) holds much promise to halt tumour growth and inhibit metastasis, however AAT trials have not delivered the predicted clinical success in a large amount of the cancer types tested, including breast cancer. It has become increasingly evident that tumours have complex and alternative neovascularisation pathways and mechanisms which may be involved in causing AAT resistance. Vasculogenic mimicry (VM) is a unique non- angiogenic vascularisation method whereby tumour cells acquire endothelial cell-like characteristics and create tumour-lined pseudo-blood vessels. VM has been associated with poor prognosis, metastasis and AAT resistance, however the exact mechanisms and processes that drive VM are still relatively poorly defined. Furthermore, the functionality and relevance of VM within tumours has been debated due to the drawbacks of the current methods used to study and visualise it. Hence, my project aims to further understand VM and its relationship with AAT resistance in patient-derived tissue xenograft (PDTX) models of breast cancer by utilising bulk and single-cell transcriptomic data and developing optimised VM vessel visualisation modalities. A 3D imaging modality which is capable of visualising both host vessels and functional vessels lacking endothelial cells within mouse models was developed and optimised. This method provided a more robust method of VM visualisation than alternative 2D techniques, which in general were found to overestimate the abundance of functional VM vessels. This 3D vessel imaging modality was applied to breast cancer PDTX models to identify VM-capable models which were further characterised to understand VM. Six PDTX models with varying VM-capabilities were subjected to AAT to investigate changes to the vasculature and gene expression upon treatment and identify possible AAT resistance mechanisms. The different models presented an array of responses including a significant reduction in host vessel number resulting in delayed tumour growth; no response to AAT; and vasculature normalisation and improved blood flow into the tumour. Whilst inherent VM-capability was linked with AAT resistance in a subset of models, it was proposed that the structure of the host vasculature and the heterogeneity between the host vessel types of models played a larger role in intrinsic or acquired AAT resistance and the type of response to AAT. This study highlighted the challenges and complex nature of using AAT to treat cancer within patients and our need to develop better biomarkers to predict response. Using single-cell transcriptomic data from the PDTX models, a subset of cells were identified within the models which displayed VM-capability and AAT-resistance that expressed high levels of endothelial, hypoxia and VM-related genes and were proposed to be the VM-capable tumour cells within these models. Biomarkers were identified from these cells in the form of “VM” gene sets that could be used to predict VM-capability from bulk RNA sequencing data; or proteins which were expressed within the tumour cells participating in VM vessels to be used to pinpoint these vessels within tissue. The genes identified were enriched in extracellular matrix proteins, tubulogenesis, cell to cell adhesion and anticoagulation, all of which are processes which are known to be important within VM. Furthermore, genes associated with caveolae membranes were identified and in particular, caveolin-1 was found to be associated with endothelial- negative, perfused vessels, implicating caveolae proteins in VM, and warranting further investigation. Altogether, although there is still much to be learnt about these fascinating tumour-lined vessels, the VM-capable PDTX models identified, the VM gene sets formulated, and the improved VM visualisation modalities developed within this thesis may help answer some of the questions surrounding VM. It is increasingly clear that VM plays a role within AAT resistance in some tumours, and hence, our improved knowledge about the mechanisms that drive VM will help us develop therapeutics to target these deadly vessels. The VM gene sets developed within this thesis could be used to enable the stratification of patients who are VM-capable and AAT-resistant and may benefit from a combinational therapy which inhibits both VM and AAT to successfully cut off the blood supply to the tumour.

Name dc:type.qualificationname

Doctor of Philosophy (PhD)

Level dc:type.qualificationlevel

Doctoral

Grantor dc:publisher.institution

University of Cambridge

Year dc:date.issued

2024

Author dc:creator

• Deighton, Lauren

Advisors dc:contributor.advisor

• Sawicka, Kirsty
• Hannon, Gregory

• Breast Cancer
• Angiogenesis
• Vasculogenic Mimicry
• Patient Derived Xenograft Models
• 3D imaging
• Single-cell sequencing
• anti-angiogenic therapy

Licence

• https://apollo8-f-pro.lib.cam.ac.uk/bitstreams/0a23a7d4-9760-4e13-8381-0e7b7d42adfc/download
• http://purl.org/NET/rdflicense/allrightsreserved

Language dc:language

eng

DOI dc:identifier.doi

https://doi.org/10.17863/CAM.116515

OAI identifier oai:identifier

oai:www.repository.cam.ac.uk:1810/381232