COMMISSURAL CONNECTIONS OF THE CAT PERIAQUEDUCTAL GRAY-MATTER STUDIEDWITH ANTEROGRADE AND RETROGRADE TRACT-TRACING TECHNIQUES

The commissural connections of the periaqueductal gray matter were inv estigated by light and electron microscopy by using the anterograde tr acer Phaseolus vulgaris leucoagglutinin and the retrograde tracer hors eradish peroxidase. In the first group of seven animals (1-7), single injections of Phaseolus vulgaris leucoagglutinin were performed iontop horetically (4.5 mu A for 30 min) into various subdivisions of the per iaqueductal gray matter. On light microscopic examination, injection s ites were characterized by several immunolabeled neurons of different sizes and morphology, with the cytoplasm, nucleus and neuronal process es intensely stained. Many labeled fibers turned from injection sites toward all contralateral periaqueductal gray matter subdivisions, but anterograde labeling was densest in the regions homotopic to those inj ected. Commissural fibers bore along their course many en passant bout ons of different sizes and morphology, and gave off spine-like process es, at the end of which one terminal bouton was observed. Labeled fibe rs branched into numerous collaterals which ended in a terminal array of 10-20 en passant and en grappe boutons. At the electron microscopic level, commissural axons were observed in close proximity to the cyto plasmic membranes of cells. Axon terminals formed symmetric or asymmet ric synapses mainly on dendritic shafts of neurons and rarely on vesic le-containing profiles. Horseradish peroxidase experiments were carrie d out in four cats (1-4). The tracer was injected iontophoretically in to different regions of the periaqueductal gray matter of three cats ( 1-3). Retrogradely labeled neurons giving rise to commissural connecti ons had a morphology similar to that of polygonal, triangular and fusi form cells described in previous Golgi studies. The perikaryal cross-s ectional area of commissural neurons was smaller than that of neurons projecting outside the periaqueductal gray matter (mean value of commi ssural neurons 149.77 mu m(2) vs 261.19 mu m(2) for projecting neurons ), which were retrogradely labeled by pressure-injecting horseradish p eroxidase into several targets of periaqueductal gray matter (4). More over, since the distribution of sizes of the two populations of the pe riaqueductal gray matter overlapped in the range of 90-300 mu m(2), a considerable number of projecting neurons were as small as commissural neurons. The present results suggest that commissural fibers could re ciprocally connect zones of the periaqueductal gray matter with simila r functions, and originate from small and medium-sized neurons, some o f which are also projecting neurons.