Neurodegenerative diseases are characterised by the accumulation of misfolded proteins and the progressive loss of synapses, leading eventually to neuronal demise. Synapse loss occurs early in disease and is a reversible process: the pruning and regeneration of synapses, known as structural plasticity, is a continuous process in the brain and represents an important point for therapeutic intervention. Recent evidence has shown that synapse loss in mice models of neurodegeneration is due to a failure of synapse regeneration, driven by failed induction of the cold shock protein, RNA-binding motif protein 3 (RBM3). Overexpression of neuronal RBM3, either by cooling or lentiviral delivery, restores synapse regeneration, preventing subsequent neuronal loss and significantly prolonging survival. How RBM3 regulates structural synaptic plasticity, however, is not well understood.

Alongside neuronal pathology, astrocytic activation is a hallmark of many neurodegenerative diseases. Astrocytes play a key role in support of synapses and in regulating structural plasticity. In this thesis, I use in vitro and in vivo approaches to examine the hypothesis that astrocytic RBM3 expression plays a role in synapse formation and structural plasticity, in health and during neurodegeneration.

In vitro data indicate that whilst neuronal RBM3 expression is necessary for synapse formation both in development and in cooling-induced structural plasticity, the role of astrocytic RBM3 expression is restricted to non-developmental structural plasticity. Mice with a complete knockout of RBM3 display accelerated disease progression compared to wild type mice with prion disease, supporting a role of both neuronal and astrocytic RBM3 in maintaining structural synaptic plasticity in disease.

Finally, expression of RBM3 protein in human blood and brain was quantified. RBM3 was shown to form part of the human cold-shock response in individuals undergoing hypothermia in a number of situations. Given the established link between RBM3 and neuroprotection in mouse models, the data presented support the possibility of therapeutic manipulation of RBM3 for neuroprotection in humans.