Brain metastasis is a significant clinical problem, and 20-40% of cancer patients will develop metastatic spread to the brain. There remains, however, an urgent need to better understand the pathogenesis of brain metastasis and, in particular, the contribution of the tumour microenvironment. Astrocytes are activated in response to metastatic growth and in vitro studies suggest that this activation promotes tumour growth and invasion. However, the result of astrocyte activation has yet to be probed in relevant in vivo models. The primary aim of this thesis, therefore, was to determine the role of astrocyte activation in metastatic progression. To address this question, two routes of metastatic induction were used; intracardiac and intracerebral. Astrocyte activation was demonstrated quantitatively in both models across a protracted time course, increasing both spatially and temporally. Subsequently, a CNTF-lentivirus model of chronic astrocyte activation was used to probe the effect of astrocyte activation on metastasis pathogenesis. No significant changes in either tumour seeding or growth were observed. However, based on subsequent assessment of the molecular phenotype of astrocytes incubated with either CNTF or tumour-conditioned media, I conclude that this lack of effect in vivo is likely to reflect the phenotypic state of the astrocytes in the CNTF-mediated model. For the converse experiment, I aimed to inhibit astrocyte activation with a targeted steroid previously shown to have anti-astrocytic effects. No significant effects on tumour burden were observed overall, although the data suggest a reduction in prefrontal areas in steroid-treated animals. In vitro and ex vivo PCR studies, together, suggested that astrocytic IL-6 may be important in metastasis progression. Finally, in light of the robust activation of astrocytes in response to metastasis, I hypothesised that astrocyte activation could be used as a diagnostic biomarker for tumour growth, potentially facilitating diagnosis at earlier time points than currently possible. To test this hypothesis, radiolabelled small molecule ligands, targeting the translocator protein (TSPO) on activated glia, were used to image astrocyte activation with either SPECT or PET. Micrometastases were detected with both agents used, and at time points prior to that possible with current clinical methods of passive contrast enhancement across a compromised blood-brain barrier. In conclusion, a dynamic interaction exists between astrocytes and tumour cells in brain metastases. Although the approaches used here have not yet revealed the contribution of astrocytes to metastatic growth, a framework has been established to further probe the role of astrocytes at the molecular level; in vivo modulation of astrocyte-derived factors will potentially yield not just mechanistic insights, but therapeutic gains. Importantly, regardless of the role of astrocytes, their robust association with tumour growth renders their activation an ideal surrogate for tumour growth.