Organization principles of thalamocortical input onto deep-layer inhibitory neurons of the barrel cortex

Abstract

Thalamus functions as the main route by which incoming sensory information reaches the cerebral cortex. Thalamic nuclei will receive input from the sensory receptors in the periphery and relay it to different areas in the cortex. (Boivie & Perl, 1975; Mountcastle, 1980; Brodal, 1981; Heimer, 1983). For this thesis, I use the vibrissal system as a model of thalamocortical input onto the cortex. In this system, the cortex will receive most of the thalamocortical (TC) projections from two thalamic nuclei, the ventral posteromedial nucleus (VPm) and the posteromedial complex (POm) (Landisman et al, 2007). A canonical pathway of information flow has been described for this system, in which thalamocortical projections will target neurons in layer 4, which in turn will synapse with neurons in layer 2/3, to finally end in pyramidal neurons in layer 5 (Gilbert et al., 1979). Although the canonical pathway is able to explain the activation of layer 4 neurons after the thalamic activation, it has become increasingly clear that neuronal responses from layer 5 do not fit with the signal flow of the canonical pathway (Egger et al., 2020). Considering the new evidence of thalamic activation of excitatory neurons in the deep layers, an investigation of the role of thalamocortical input onto the inhibitory population in the deep layers is required, as also is the evaluation of the role of the deep layer inhibitory neurons in modulating the responses of L5PTs. <br /> To this end, I systematically investigated how sensory input from primary thalamus is relayed to inhibitory neurons in the deep layers of the barrel cortex. I first recorded and labeled inhibitory neurons in the deep layers of the barrel cortex. Before the recording session, animals were injected with a modified rAAV conjugated with channelrhodopsin for optogenetic manipulation of thalamocortical synapses, and with m-Cherry as a fluorescent reporter. I recorded the spontaneous activity of these neurons, and their response to first, a multi-whisker deflection provided by a 700 ms airpuff, and second to a direct optogenetic activation of the thalamocortical synapses with an LED light. Once all recordings were finalized, I labeled these neurons with biocytin and identified them as inhibitory neurons based on their morphological characteristics (Yen et al., 1985). These neurons were then reconstructed to evaluate their axo-dendritic arborization and registered to a standard model of the barrel cortex (Egger et al., 2012). The soma sections of these experiments were then stained against the non-overlapping molecular identities for this population, parvalbumin, somatostatin, and VIP (Meyer et al., 2013). Afterwards, these same sections were used to quantify the number and density of putative synapses between dendrites of the inhibitory neurons, and boutons from the axons of VPm neurons (reported with m-Cherry). <br /> Finally, I quantified the distribution and density of inhibitory neurons in the barrel cortex that receive monosynaptic input from the primary thalamus. I achieve this by the simultaneous injection of a trans-synaptic viral injection of pEEN-AAV1 in the VPm nucleus and AAV2-DIO-eGFP in the barrel cortex (Zingg et al., 2014, 2017). On top of that, I stained these sections against parvalbumin and somatostatin to evaluate the possible differences between thalamocortical input in these populations. <br /> I find that inhibitory neurons in the deep layers exhibit either a reliable fast evoked response that precedes the surrounding excitatory neurons or a delayed response that succeeds the excitatory activation by a time consistent with a monosynaptic jump (Miles & Wong, 1986; Doyle & Andresen, 2001). This fast-evoked response was accompanied by a significantly higher density of putative synaptic contacts between these neurons and the axons from the primary thalamus. Both these results indicate that this subpopulation of inhibitory neurons receives a direct, strong input from the thalamus and these neurons are likely involved in feedforward inhibitory circuits. I report that the composition of these circuit configurations seems to be more heterogeneous than those described for layer 4. In support of that statement, I find that those neurons preceding the excitatory population can have distinct morphologies, molecular identities, response patterns, and cortical depth. The same heterogeneity can be found in those neurons that exhibit a delayed response to the stimulus presentation. I conclude that in the deep layers of the barrel cortex, the thalamus recruits a highly heterogeneous population for feedforward inhibitory circuits, but also spares an equally heterogeneous population.