Immunohistochemical results on the distribution of somatostatin in the hypothalamus and in limbic structures of the rat.

For several years it has been known that somatostatin is synthesized by paraventricular cells in the wall of the third ventricle (Elde and Parsons, 1975; H#{246}kfelt et al., 1975; Alpert et al., 1976). Moreover, these cells are not aggregated into a nucleus, but are always located separately in maximally three to four short lines that run parallel to the ventricular wall within an area extending from the preoptic nucleus to the rostral pole of the median eminence. In a parasagittal section, the region of somatostatin-producing cells occupies an approximatively kidney-shaped area which is sloping upward in caudal direction. In frontal and in parasagittal sections the cells appear to be predominantly bipolar. On the ultrastructural level, the product of the immunocytochemical reaction is found exclusively above small elementary granules in the paraventricular cells. Granules with the same diameters are also found within numerous axons in the surrounding neuropil. Cells, positively reacting to antisomatostatin can be observed in the newborn rat, but the band of cytoplasm is still quite narrow at this time. However, the large number of positively reacting points and dots, representing fiber sections in the paraventricular neuropil, is striking. This last observation can no longer be made after the first postnatal week; at that time the stainability of fibers adjacent to the cells has decreased. It increases again slowly but steadily during the next weeks, just like the amount of stainable cytoplasm. From the fourth postnatal week on there is no observable difference between the immunohistochemical features of these perikarya and those of adult animals. Beyond the paraventricular neuropil no cells are stained immunohistochemically in such a way that somatostatin-producing properties could be attributed to them. However, “labeled” cells can be found in other regions too, the labeling of which is not quite unambiguous; they will be referred to later. Since the number of processes leaving directly from the cell soma is small, as observed in frontal and sagittal sections, the number of axon collaterals must be very large. Short distance connections of positive reacting fibers are demonstrated within the hypothalamus not only to the neurohemal regions (the distribution of somatostatin in the neurohemal regions is the subject of Dr. Pelletier’s report, this volume), but also to several hypothalamic nuclei. Moreover, longdistance projections of positive-reacting fibers are observed in structures of the limbic system (Fig. 1). Short-distance projections: The following hypothalamic nuclei exhibit a labeling by positive-reacting fibers: preoptic nucleus, suprachiasmatic nucleus, ventromedial and arcuate nucleus, ventral premammillary nucleus. The caliber of the labeled fibers vanes to some degree in these nuclei, but their perikarya are always unstained. In the arcuate nucleus, synapse-like contacts are observed between