Thalamic input to parvalbumin-immunoreactive GABAergic interneurons: organization in normal striatum and effect of neonatal decortication.

The neocortex and thalamus send dense glutaminergic projections to the neostriatum. The neocortex makes synaptic contact with spines of striatal projection neurons, and also targets a distinct class of GABAergic interneurons immunoreactive for the calcium-binding protein parvalbumin. We determined whether the parafascicular thalamic nucleus also targets striatal parvalbumin-immunoreactive interneurons. The anterograde tracer biotinylated dextranamine was injected into the parafascicular nucleus of adult rats. Double-labeled histochemistry/immunohistochemistry revealed overlapping thalamic fibers and parvalbumin-immunoreactive neurons in the neostriatum. Areas of overlap within the sensorimotor striatum were analysed by electron microscopy. Of 311 synaptic boutons originating from the parafascicular nucleus, 75.9% synapsed with unlabeled dendrites, 22.5% with unlabeled spines, and 1.3% had parvalbumin-immunoreactive dendrites as a postsynaptic target. Only 4% of all asymmetric synapses on parvalbumin-immunoreactive dendrites were derived from the parafascicular nucleus. A separate group of animals underwent bilateral neocortical deafferentation on the third postnatal day, prior to injection of anterograde tracer into the parafascicular nucleus of adult animals. These experiments were performed with the dual purpose of (i) reducing the possibility that thalamic inputs to parvalbumin-immunoreactive neurons are the result of transsynaptic uptake of tracer by a thalamo-cortico-striatal route, and (ii) determining whether competitive interactions between developing corticostriatal and thalamostriatal fibers may account for the relatively sparse thalamic input onto parvalbumin-immunoreactive interneurons. In decorticates, 219 striatal synaptic contacts derived from the parafascicular nucleus, out of which 77.2% were on unlabeled dendrites, 20.9% were upon unlabeled spines, and 0.9% targeted parvalbumin-immunoreactive dendrites. We conclude that the thalamic parafascicular nucleus indeed sends synaptic input to parvalbumin-immunoreactive striatal neurons. Parafascicular nucleus inputs to striatal parvalbumin-immunoreactive interneurons are sparse in comparison to other asymmetric inputs, most of which are likely to be of cortical origin. The synaptic profile of thalamostriatal inputs to parvalbumin-immunoreactive neurons and unlabeled elements is unchanged following neonatal decortication. This suggests that competitive interaction between developing thalamostriatal and corticostriatal projections is not a major mechanism determining synaptic input to striatal subpopulations.