THE JOURNAL OF COMPARATIVE NEUROLOGY 195~253-263 11981) Thalamocortical Synapses Involving Identified Neurons in Mouse Primary Somatosensory Cortex: A Terminal Degeneration and GolgilEM Study

Synapses involving thalamocortical afferents and hitherto unexamined neuron types of the posteromedial barrel subfield (PMBSF) of the mouse have been identified using a combined degeneration and GolgiIEM technique (Peters e t al., '77). Degeneration of thalamocortical axon terminals was produced with electrolytic lesions of the nucleus ventralis posterior, pars lateralis thalami, and the nucleus posterior thalami. Four days after receiving lesions, the animals were perfused, and blocks of cortex containing the PMBSF were processed by the Golgi method. The blocks were tissue-chopped a t 125 Fm and examined with the light microscope. Sections containing neurons of interest were gold-toned and deimpregnated in preparation for electron microscopy (Fairen et al., '77). Portions of selected neurons contained in layers 111-V were serially thin-sectioned and examined with an electron microscope to determine if they were involved in synapses with degenerating thalamocortical axon terminals. Results showed thalamocortical synapses on the apical dendrites of five different sized pyramidal cells whose somata occurred in layers V and VI, and on dendrites of one spiny bitufted neuron and one non-spiny multipolar neuron with somata in layer V. A non-spiny bitufted neuron of layer IV which was not impregnated also received thalamocortical synapses. Although every neuron examined formed a t least one thalamocortical synapse, some formed very few, whereas others formed many. Of the pyramidal cells, small layer V and VI pyramidal cells and a large deep layer V pyramidal cell were involved in small numbers of thalamocortical synapses, while a medium superficial layer V pyramidal cell and a large layer VI pyramidal cell each formed many. The spiny bitufted neuron formed a small number of thalamocortical synapses, while the non-spiny bitufted neuron formed very many. The non-spiny multipolar neuron was involved in a moderate number of thalamocortical synapses. The findings suggest that, whereas any type of neuron with a dendrite in layer IV likely receives some synaptic input from the thalamus, individual neurons were involved in very different quantities of thalamocortical synapses. The bulk of the thalamic projections to the primary areas of sensory cortex terminates in layer IV (see review by White, '79). In the mouse, layer IV of primary somatosensory cortex contains about 200 multicellular units known as barrels (Woolsey and Van der Loos, '70), and the barrels which contain the cortical representation of the mystacial vibrissae comprise the posteromedial barrel subfield (PMBSF). By combining the GolgiiEM method of Fairen e t al., ('77) with the technique of anterograde degeneration, White ('78) showed tha t thalamocortical synapses in mouse PMBSF involve the basal dendrites of layer 111 pryamidal cells, spiny and non-spiny layer IV nonpyramidal cells, and the apical dendrites of small superficial layer V pyramidal cells. In fact, this study and a similar one of rat visual 0021-9967/81/1952-0253$02.30 LJ 1981 ALAN R. LISS. INC. 254 S.M. HERSCH AND E L . WHITE cortex (Peters et al., '79) suggested that every postsynaptic surface present in layer IV capable of forming asymmetrical synapses is postsynaptic to thalamic axon terminals (see Peters, '79; White, '79). However, many types of pyramidal and nonpyramidal cells with dendrites contained in layer IV have yet to be examined. Particularly interesting in this regard are the pyramidal cells with perikarya in the deeper layers of sensory cortex. These pyramidal cells have been shown by light microscopy of Golgi-impregnated tissue to possess spine populations that are very sensitive to both perinatal deafferentation and sensory manipulation (for example, Globus and Scheibel, '67; Valverde, '68), and the findings of such studies have been interpreted as evidence that nearly one third of the spines of pyramidal cell apical dendrites passing through layer IV receive thalamocortical input. The present study is a survey of the thalamocortical connectivity of hitherto unexamined neuron types in mouse PMBSF cortex, with an emphasis on the thalamic synapses involving pyramidal cells of layers V and VI. Our intention is to further test the hypothesis that all neurons with dendrites contained in layer IV form thalamocortical synapses, and to learn if individual cells are involved in very different numbers of thalamocortical synapses. An indication that this might be true is provided by a light microscopic Golgi study that showed superficial and deep layer V pyramidal cells to be differentially sensitive to postnatal deafferentation (Ryugo et al., '75a,b). MATERIALS AND METHODS Electrolytic lesions were made in the nucleus ventralis posterior pars lateralis thalami of male CD/1 mice ranging in age from 5 5 8 0 days. A transcollicular approach was used to avoid passing the lesion electrode through neocortex. Four days after the lesion, the mice were perfused intracardially with an initial weak fixative containing 1.072 paraformaldehyde and 1.2Wo glutaraldehyde, followed by a fixative containing 2.0% of each of these aldehydes. Both fixatives were buffered with 0.1 M sodium cacodylate and adjusted to pH 7.2. The thalami were removed, frozen-sectioned at 40 pm, mounted on glass slides, and then stained with cresyl violet to assess the placement and extent of the lesions. Blocks containing all of PMBSF cortex ipsilateral to the lesion were then processed by the rapid Golgi method (Valverde, '70). After impregnation, the tissue was dehydrated through graded glycerin solutions to 100% glycerin and tissue-chopped in the coronal plane a t 125 pm with a Sorvall tissue chopper. Sections were wet mounted in 100% glycerin and, using a microscope equipped with a drawing tube, they were compared to drawings of similar Nissl-stained sections to identify the posteromedial barrel subfield. Sections containing Golgi-impregnated PMBSF neurons were then prepared for electron microscopy using a procedure based on the Golgi/EM method of Fairen et al. ('77): Tissue-chopped sections in 10Wt glycerin were passed into 40% and 20% glycerin solutions for 1 minute each, then into .05% gold chloride for 15-20 minutes. The gold was then reduced with .05% oxalic acid and the precipitate removed with 1.0%~ sodium thiosulfate. The result of this procedure is to replace the dense Golgi precipitate (silver chromate) with a deposit of fine gold particles. Gold-toned neurons are still visible in the light microscope, and the gold deposit does not obscure their fine structure (Fairen et al., '77). Sections containing selected gold-toned neurons were subsequently postfixed for 30 minutes with 2.0% osmium tetroxide in 0.1 M cacodylate buffer, block-stained for 2 hours in 1 .Wr uranyl acetate in 7Wc methanol, and embedded in a thin layer of Epon-Araldite. The selected neurons were then photographed or drawn using a light microscope equipped with a drawing tube, and mounted on Epon-Araldite cylinders for thin sectioning. Using a MTB-2B ultramicrotome, unbroken series of thin sections were taken through the portions of the neurons contained in layers 111-V, subsequently mounted on formvarcoated slot grids, and stained with lead citrate. All sections through each neuron were then examined with an AEI Corinth electron microscope, and all synapses occurring between gold-toned processes and degenerating thalamocortical axon terminals were photographed a t x 15,000. In addition to the gold-toned neurons, an unimpregnated neuron was also examined. On the basis of the form of its cell body and dendrites in a three-dimensional reconstruction, this neuron was identified as a layer IV nonspiny bitufted cell (White and Rock, '80). Its thalamocortical connections were assessed by tracing its dendrites through the series and photographing its synapses with degenerating thalamocortical axon terminals a t x 15,000. To determine if the dendrites of each neuron examined passed through fields containing comparable numbers of degenerating thalamoTHALAMOCORTICAL SYNAPSES ON IDENTIFIED NEURONS 255 cortical axon terminals, counts were made in layer IV of a l l degenerating and normal presynaptic elements occurring in small volumes of neuropil surrounding each dendrite.