Mechanism and function of GABA and glutamate co-packaging at pallidohabenular synapses

Abstract

Classic views on neurotransmission suggested for a long time that a single neuron releases a single neurotransmitter at all of its synaptic terminals. However, recent studies reveal that many neurons throughout the brain can synthesize, store, and release two or more neurotransmitters. We examined a genetically- and anatomically- defined population in the mammalian brain that co-releases common fast-acting neurotransmitters, glutamate and GABA, which act antagonistically on a postsynaptic neuron. These co-releasing neurons are located in the internal globus pallidus, a major output nucleus of the basal ganglia, a set of subcortical brain structures implicated in motor control and learning. Curiously, the co-releasing neurons participate in the limbic pathway through their projections to the lateral habenula, a brain center for aversion and reward. Aim of this thesis work was to explore the mechanism and function of glutamate and GABA co-release in the pallidohabenular circuit. We first tested whether these opposing transmitters are co-released from individual synaptic vesicles via co-packaging, using interdisciplinary approaches that combined optical, electrophysiological, and analytical methods to repeatedly examine the physiology of a single synapse. We discover that the opposing transmitters are released from the same vesicles in individual presynaptic terminals in the lateral habenula, which is known to receive multiple glutamate/GABA co-releasing and converging inputs. Furthermore, our findings reveal that neuromodulators targeting sertonin and adenosine receptors downregulate the release of both glutamate and GABA, consistent with the co-packaging of these transmitters. Next, we investigated the function of the glutamate/GABA co-releasing neurons and the potential synaptic function of co-release using animal behavior and bulk calcium photometry recording. We found a complex and dynamic representation of a footshock and its predictive cue by the glutamate/GABA co-releasing population and their putative postsynaptic neurons, consistent with previous results showing the pallidohabenular circuit’s role in aversion and learning. Moreover, we observed differential encoding of the learned cues and outcomes by these two neural populations. Lastly, we discovered that the functional connectivity between the co-releasing neurons and their putative downstream targets varied in different behavioral contexts. Our findings here inform a biologically realistic model of the pallidohabneular synapses and the potential function of co-release and co-packaging during synaptic adaptation and learning. We suggest that this circuit may employ co-packaging of glutamate and GABA to allow the postsynaptic value assignment of a set of inputs that convey multimodal representation of ongoing events. Co-packaging is also beneficial for relaying a reliable and consistent signal with every impulse. We speculate that habenula neurons can extract either positive or negative value from the same set of inputs based on this synaptic structure by modulating the type and number of receptors. Using value assignment for learning association may be useful for the habenula to better tune itself to upcoming events through prediction and enhance its signal-to-noise ratio of aversive information.