The technique of intracellular recording and staining of the same neuron with horseradish peroxidase (HRP) was used to study the soma-dendritic and axonal morphology of nigrothalamic and nigrotectal cells in the rats. The nigrothalamic and nigrotectal cells were spread throughout the dorsoventral extent of the pars reticulata (SNR) and exhibited the same soma-dendritic and axonal features. Both populations consisted of medium-sized and large cells with extensive dendritic fields overlapping in all three directions. Their axons collateralized within the substantia nigra (SN) and in the mesencephalic tegmentum. The intrinsic collaterals were thin and branched partly within the dendritic field of a parent cell partly in remote regions of the SNR, and even in the pars compacta (SNC). The extrinsic branches involved thin arborizations in the rostroventral mesencephalic reticular substance and thicker descending and ascending collaterals. This material was supplemented by physiologically nonidentified HRP stained medium-sized and large neurons located in the SNR. The two kinds displayed the same extent and orientation of their dendrites but the branching patterns differed slightly. Proximal dendrites of all cells were coarse and smooth; thinner distal dendrites had varicosities and spinelike appendages. Some dendrites, specially those near the crus cerebri, terminated in dendritic thickets bearing many pleomorphic appendages. The orientation of dendritic fields varied with dorsoventral position of cells within the SNR. The most ventral region of the SNR contained neurons with dendrites oriented parallel to the crus cerebri and thus remained confined to the deepest stratum. The dendrites of cells in the central region of SNR were oriented mainly anteroposteriorly and ventrally, the ventral dendrites terminating in the ventralmost layer. Cells in the dorsolateral part of the SNR were characterized by the large dorsoventral extent of their dendrites which penetrated the entire thickness of SN. This variation in the arrangement of dendritic fields indicates that the SN is organized in three dorsoventral layers.

Choline acetyltransferase immunhistochemistry was employed at light and electron microscopic levels in order to determine the distribution of cholinergic neurons in two subdivisions of the rat pedunculopontine tegmental nucleus that were previously defined on cytoarchitectonic grounds, and to compare the synaptic inputs to cholinergic and non-cholinergic somata in the subnucleus dissipatus, which receives major input from the substantia nigra. Large cholinergic neurons were found in both the pars compacta and the pars dissipata of the pedunculopontine nucleus. However, they were intermingled with non-cholinergic neurons and did not respect the cytoarchitectural boundaries of the nucleus. Ultrastructural study showed that all cholinergic neurons in the subnucleus dissipatus exhibited similar features. The majority had large somata (largest diameter greater than or equal to 20 microns) containing abundant cytoplasmic organelles and nuclei displaying a few shallow invaginations. Synaptic terminals on the cholinergic cell bodies were scarce and unlabeled boutons containing spherical synaptic vesicles and establishing asymmetric synaptic junctions were the dominant type. In contrast, the non-cholinergic neurons presented prominent differences in the size of their somata as well as in the distribution of axosomatic synapses. Two almost equally represented classes of non-cholinergic neurons which are referred to as large (largest diameter greater than or equal to 20 microns) and small (largest diameter less than 20 microns) were recognized. Large non-cholinergic cell bodies were ultrastructurally similar to the cholinergic ones, but they received rich synaptic input by unlabeled nerve terminals which contained pleomorphic vesicles and were engaged in symmetric synaptic junctions. Small non-cholinergic cell bodies were characterized by deeply invaginated nuclei surrounded by a narrow rim of cytoplasm, and were often found near or in direct apposition to the cholinergic somata. Their major input consisted of axosomatic boutons containing round synaptic vesicles. These results demonstrate that cells in the pedunculopontine tegmental nucleus are differentiated with regard to their axosomatic synaptic inputs which may influence their firing properties. Furthermore, they support previous suggestions that nigral afferents may be preferentially distributed to a subpopulation of the pedunculopontine neurons.

The termination of the substantia nigra pars reticulata efferents in the nucleus tegmenti pedunculopontinus was studied in the rat by using the anterograde tracer Phaseolus vulgaris-leucoagglutinin (PHA-L). Both large and small injections of PHA-L in various portions of the substantia nigra pars reticulata labeled varicose fibers in the ipsilateral and contralateral nucleus tegmenti pedunculopontinus, subnucleus dissipatus as well as in the ipsilateral nucleus tegmenti pedunculopontinus, subnucleus compactus. However, the bulk of the nigral fibers appeared to terminate in the medial two-thirds of the ipsilateral subnucleus dissipatus of the pedunculopontine nucleus and exhibited a discrete dorsoventral topographical pattern. The terminal plexus displayed patches of uneven density, which was partly due to the numerous fiber bundles passing through the pedunculopontine nucleus, but also to an obvious preference of nigral fibers for some cells. Electron microscopic examination confirmed that nearly all of the varicosities observed in the light microscope contained synaptic vesicles and represented either terminal boutons or boutons en passant. The labeled boutons were elongated (average length: 1.5 microns) and consistently contained a prominent group of mitochondria. The results suggest that the nigral input to the nucleus tegmenti pedunculopontinus may be directed toward specific subpopulation(s) of pedunculopontine neurons and may influence not only cells in the subnucleus dissipatus, but also in the subnucleus compactus.

The distribution and collateralization of ascending and descending projections from neurons in the nucleus tegmenti pedunculopontinus (PPN) were studied in the rat by using retrograde transport of HRP, HRP/WGA, and fluorescent dyes. The PPN and its two subdivisions, the subnucleus compactus (PPNc) and subnucleus dissipatus (PPNd), were delineated on sagittal Nissl-stained sections by using cytoarchitectural features as guidelines. Large bilateral pressure injections of HRP and/or fluorescent dyes into the cervical cord retrogradely labeled moderate numbers of fusiform and polygonal PPN cells which ranged in size between 65 and 390 microns2. The labeled cells were scattered throughout the PPNd and were somewhat more numerous in the medial half of the subnucleus. The PPNc contained only occasional labeled cells in its ventralmost portion. Following single unilateral HRP/WGA injections in the striatum, globus pallidus, entopeduncular nucleus, subthalamus, or the substantia nigra, the distribution of the labeled cells was similar to that of the spinal cord-projecting PPN neurons. Multiple HRP injections were then made bilaterally in the substantia nigra and the entopeduncular nucleus and/or subthalamus in order to label the entire population of PPN neurons projecting to the basal ganglia. In this case, not only the PPNd but also the PPNc contained a substantial number of retrogradely labeled cells. The rostrally projecting PPN cells outnumbered 5.4 times those projecting to the spinal cord, and their somata were somewhat larger, ranging between 114 and 472 microns2. While both fusiform and polygonal shapes were encountered, the polygonal cell somata were more numerous. In the double-labeling experiments, Granular Blue and Diamidino Yellow Dihydrochloride were injected into the cervical cord and the entopeduncular nucleus or subthalamus. In general, these experiments confirmed the extensive overlap of forebrain- and spinal cord-projecting neurons within the PPNd and the quantitative preponderance of ascending neurons. They also demonstrated that these two projection systems originate largely from separate cell populations since the double-labeled cells always composed less than 5% of the labeled neurons. The results confirm the existence of a direct PPN projection to the spinal cord. This pathway originates mainly in the PPNd and appears to be quantitatively weaker than the PPN projections to the forebrain. The spinal cord-projecting cells are not spatially segregated from the cells projecting to the basal ganglia, but they represent a separate population of the PPN projection neurons.

Substantia nigra pars reticulata (SNr) neurons, antidromically activated following stimulation of the dorsal thalamus and/or superior colliculus were intracellularly stained with HRP. Light microscopic analysis revealed that the labeled SNr neurons have axon collaterals arborizing within SNr. Axon collaterals of SNr neurons partially overlapped with the dendritic fields of their parent cells and also extended beyond the parent dendritic fields. The labeled axon terminals did not closely appose the parent cell processes, suggesting that the collaterals most likely terminate on neurons other than the parent cell. Electrical stimulation of either the thalamus or the superior colliculus induced monosynaptic and polysynaptic IPSPs in SNr cells. The polysynaptic IPSPs evoked from thalamic stimulation disappeared following hemitransection of the brain just rostral to the thalamus while the monosynaptic IPSPs remained the same. Since there are no known afferents from either thalamus or superior colliculus to SNr, we consider that these monosynaptic IPSPs are due to activation of the recurrent collaterals of SNr projection cells. The results of this study indicate that projection neurons of SNr also have an inhibitory role within the SNr.