Final Report Summary - INTERNEURON NETWORK (cerebellar molecular layer Interneuron network imaging in awake animal) The objective of this project was to monitor the activity of cerebellar molecular layer interneurons (MLIs) during behavior and to understand how they are involved in motor coordination and learning. The main results of this project consist two parts: one is a technical development for fast calcium imaging of a population of neurons in vivo; the other is the measure of spatial-temporal activation property of metabotropic glutamate receptor (mGluR) in vivo. Classical raster scan of laser scanning microscope has low temporal resolution. For a 200 um * 100 um area of scanning in cerebellar molecular layer, it takes half a second. The modern genetically encoded calcium indicators (GECIs) have rising time in the order of tens of milliseconds. The scanning speed became a limiting factor for imaging a population of neurons. Resonant scanning mirror and acousto-optic method can improve the imaging speed, but they require additional high cost investment. We have developed a multi-patch scan (MPS) method based on a traditional galvos-mirror scan system to decrease the scan time required for imaging multiple neurons or structures. This method can achieve simultaneous imaging of 20 MLIs at a rate of 50 Hz and it requires only additional software control for the scan path, which can be implemented in most of the current laser scan microscope. It has been applied to image MLIs when the animal perform a behavior task and will soon reveal the functional relevance of MLI network in cerebellar control of motor behavior (an ongoing project in the host laboratory). In cerebellar slices, activation of mGluRs has been shown important for several neuronal functions, including changes in synaptic efficacy (long-term synaptic plasticity), control of AMPAR Ca2+ permeability and modulation of Ca2+ channels. In parallel fiber (PF) synapses and in the glutamatergic synapses of the CA1 region of the hippocampus, type 1α mGluRs are excluded from the postsynaptic density (PSD) and are instead localized toperisynaptic regions. Compared to AMPARs, the kinetics of glutamate binding and unbinding to mGluRs are much slower. In accord with these spatial and kinetic constraints, mGluR-evoked intracellular Ca2+ (Cai) rises in vitro require repetitive synaptic stimulation at high frequency. mGluR1s have been implicated in cerebellar control of motor behavior, but little is known on the conditions governing their activation in vivo. Significant differences between in vitro and in vivo may arise from perturbations in the extracellular space due to tissue slicing as well as from modifications in either glutamate homeostasis or in the composition and/or functional state of molecules involved in mGluR signaling. In this project we explored these issues using a combination of experimental and modelling approaches, focusing on the synapses between PFs and MLIs. Whereas PF-PC synapses are encapsulated by glial membranes, a geometry that could retard the diffusion of synaptically released glutamate and favor activation of slow-activating perisynaptic receptors, PF-MLI synapses lack ensheathing elements and thus transmitter availability for receptors distant from the PSD may be less favored. While repetitive PF stimulation activates mGluR1s on MLIs in vitro, whether the same occurs in vivo is unknown. We monitored Cai changes in MLIs in vivo using PV-cre dependent expression of genetically encoded Ca2+ indicators (GECIs). We used beam-activation of PFs, a pattern of activity recently shown to occur during forelimb stimulation. We found that trains of PF stimulations resulted in significant activation of postsynaptic mGluR1s in MLIs. We further determined the temporal pattern of PF stimulation required for mGluR1 activation in these neurons. We performed quantitative electron microscopy (EM) analysis of mGluR1α localization at PF-MLI synapses and numerical simulations. The results argue that mGluR1 activation occures at sites in close proximity to the glutamate release sites. Our modeling also showed that Cai rises linked to mGluR necessitated repetitive stimulation of the synapse, but did not rely on glutamate spillover from neighboring synapses. This project also yielded another surprising result that calcium permeable AMPA receptors seem to be absent from adult MLIs.
