Search

Synaptic plasticity is thought to be integral to learning and memory. The two most common forms of plasticity are long-term potentiation (LTP) and long-term depression (LTD), both of which can be supported either by presynaptic changes in transmitter release probability (Pr), or by postsynaptic changes in AMPA receptor number. It is generally thought that the induction of LTP and LTD at Schaffer-collateral synapses in the hippocampus depends on the activation of NMDA receptors (GluN). Recent studies, however, have demonstrated that both increases and decreases in Pr can be induced under blockade of postsynaptic GluN receptors, suggesting that the activation of postsynaptic GluN receptors by glutamate is only a strict requirement for postsynaptic plasticity. In this thesis, I therefore re-examined the role of glutamate in presynaptic plasticity. I used single synapse imaging along with electrophysiological and pharmacological techniques to independently manipulate and monitor the levels of glutamatergic signalling during synaptic activity. I discovered that glutamate is inhibitory and unnecessary for the induction of LTP at the presynaptic locus. My findings support a novel model of presynaptic plasticity in which the net activity-dependent changes in Pr at an active presynaptic terminal is jointly determined by two opposing processes that can be simultaneously active: 1) postsynaptic depolarization, which, via the activation of L-type voltage-gated Ca2+ channels, increases Pr by driving the synthesis and release of nitric oxide from neuronal dendrites and 2) glutamate release, which through the activation of presynaptic GluN receptors, decreases Pr. Computationally, this model suggests that plasticity functions to reduce prediction-errors that arise during synaptic activity, and, thereby offers a biologically plausible mechanism by which neuronal networks may optimize learning at the level of single synapses.

Learning and memory within the brain are thought to be based on long lasting changes in synaptic efficacy. Purkinje neurones, which focus their output on descending projection pathways and constitute the sole inhibitory output from the cerebellum, display two forms of synaptic plasticity termed 'depolarisation-induced suppression of inhibition' (DSI) and 'rebound potentiation' (RP). Purkinje neurone depolarisation induces a rapid rise in [Ca2+ ]I triggering both the release of a retrograde transmitter and activation of a variety of protein kinases. The phenomena of DSI underlies a transient (lasting < 60s) decrease in the mean frequency of mIPSCs, occurring immediately after stimulus cessation, while RP manifests itself as a robust increase in the mean amplitude of spontaneous and miniature inhibitory postsynaptic currents (IPSCs). The present study examined the relationship between pre- and postsynaptic plasticity during the induction phase of DSI and rebound potentiation in cultured cerebellar Purkinje neurones. Depolarisation caused an immediate decrease in the frequency of mIPSCs (lasting ~40s), followed by a transient increase in mIPSC frequency lasting approximately 5 minutes before recovering. A robust potentiation of the mean mIPSC amplitude was observed throughout all experiments and persisted for the duration of recording. The initial frequency decrease (DSI), was abolished by a specific group II mGluR antagonist, LY 341495, while the subsequent transient frequency potentiation was abolished by the specific N-methyl-D-aspartate receptor (NMDAR) antagonist, d-APV. Removal of extracellular sodium, the main current carrying ion through NMDARs, mimicked the application of d-APV by abolishing the frequency potentiation while having no effect on the induction of DSI. Immunocytochemical staining of mixed cerebellar preparations identified cerebellar basket/stellate cells as displaying immunoreactivity for NMDAR NR1 subunits but not mGluR2/3 at putative presynaptic release sites. These results provide the first evidence for, 1) the involvement of presynaptic NMDARs in the transient enhancement of GABA release during rebound potentiation and 2) the possibility that a novel group II mGluR splice variant/subtype underlies the induction of cerebellar DSI. A model is proposed to explain the relationship between DSI and rebound potentiation.