An Investigation into the Vascular Phenotypes of Breast Patient- Derived Xenografts

Abstract

Blood vessel networks in tumours are often chaotic, dense and immature resulting in reduced blood perfusion and oxygen delivery, leading to hypoxia (low oxygen levels). Hypoxic tumours are more aggressive, therapy resistant and likely to metastasise, particularly in breast cancer. Equally, hypoxic tumours encourage the growth of new blood vessels from existing vasculature, termed angiogenesis, and high rates of angiogenesis result in immature and chaotic vessels forming, creating a feed-forward loop of poor perfusion and oxygenation. Photoacoustic imaging (PAI) can visualise vascular features in the tumour microenvironment at multiple scales, building a complete picture of the vascular phenotype in a single tumour, which can be monitored longitudinally over time and noninvasively in vivo. It’s high spatiotemporal resolution, low cost, use of nonionizing radiation and noninvasive properties make PAI an attractive option for monitoring dynamic vascular features not only in a preclinical setting but also throughout a patient’s treatment regime, particularly in breast cancer management. Thus far there has been a reliance on cell-line mouse models to provide insight into tumour vascular phenotypes captured with PAI. As a result, several questions remain regarding the sensitivity of PAI to discriminate between patient vascular phenotypes, and which vascular features are important to monitor in patients. In order to translate PAI into the clinic, the field must begin to use more clinically-relevant preclinical models and assess their ability to recapitulate the phenotypes seen in patients. This thesis proposes the use of breast cancer patient-derived xenografts (PDXs) in PAI, to begin to answer the aforementioned questions in clinically-relevant models. However, whether PDXs are good vascular models themselves remains unknown. This thesis conducts a careful evaluation of whether vascular phenotypes differ between 4 breast PDX models and how they evolve over time using PAI with corresponding ex vivo immunohistochemistry, used to biologically validate the phenotypes seen in vivo and provide additional molecular information. The work assessed how vascular phenotypes change across PDX passages and briefly compared them to originating patient tissue sections. Finally, the origin of these vascular phenotypes is investigated by measuring the underlying hypoxic gene expression of the cancer cells, assuming that the cancer cells shape the mouse host vasculature. The 4 breast PDXs studied displayed different vascular features on ex vivo IHC, which PAI was sensitive to in vivo. Overall, the PDXs were robust and reliable vascular models, with little inter-passage variability and similarity to patient vascular phenotypes shown on IHC. Demand and supply of oxygen through the blood vessel network appears to influence the extent of hypoxia in the tumour tissue. Inherent hypoxic phenotypes were measured using PAI, IHC and RNA-seq, which could drive formation of immature vascular networks in some PDX models. This thesis is the first multiparametric investigation into breast PDX vasculature across scales using PAI, IHC and RNA-seq.