Paradoxical sleep (PS) is a state characterized by cortical activation, rapid eye movements and muscle atonia. Fifty years after its discovery, the neuronal network responsible for the genesis of PS has been only partially identified. We recently proposed that GABAergic neurons would have a pivotal role in that network. To localize these GABAergic neurons, we combined immunohistochemical detection of Fos with non-radioactive in situ hybridization of GAD67 mRNA (GABA synthesis enzyme) in control rats, rats deprived of PS for 72 h and rats allowed to recover after such deprivation. Here we show that GABAergic neurons gating PS (PS-off neurons) are principally located in the ventrolateral periaqueductal gray (vlPAG) and the dorsal part of the deep mesencephalic reticular nucleus immediately ventral to it (dDpMe). Furthermore, iontophoretic application of muscimol for 20 min in this area in head-restrained rats induced a strong and significant increase in PS quantities compared to saline. In addition, we found a large number of GABAergic PS-on neurons in the vlPAG/dDPMe region and the medullary reticular nuclei known to generate muscle atonia during PS. Finally, we showed that PS-on neurons triggering PS localized in the SLD are not GABAergic. Altogether, our results indicate that multiple populations of PS-on GABAergic neurons are distributed in the brainstem while only one population of PS-off GABAergic neurons localized in the vlPAG/dDpMe region exist. From these results, we propose a revised model for PS control in which GABAergic PS-on and PS-off neurons localized in the vlPAG/dDPMe region play leading roles.

Adequate fixative in immunohistochemistry requires not only a rapid and total immobilization of the antigen, but also a sufficient preservation of its immunoreactivity and maintenance of its accessibility to the immunochemical reagents for localization. Thus, the optimal fixation condition for a specific antigen necessitates a compromise between these opposing variables and can be determined by the preparation of a series of tissues with a progressively increasing degree of fixation. Unless the results of localization using such a series is available, one must be satisfied with adequate but less than optimal results. In the present study, this principle is demonstrated using the localization of tyrosine hydroxylase in the dopaminergic system with formaldehyde as the fixative. The rate and degree of fixation with formaldehyde was shown to be highly pH dependent. By perfusing the tissue with formaldehyde at pH 6.5 (where the rate of fixation is extremely slow) it is possible to rapidly distribute the fixative homogeneously into the tissue. By suddenly changing to a formaldehyde perfusate of higher pH, the cross-linking reaction is rapidly increased. This two-step fixation procedure provides a means of obtaining a rapid and uniform immobilization of the antigen, so that its translocation can be avoided. The final degree of fixation is controlled by the duration and pH of the second fixative solution. The results obtained by increasing the pH of the second solution demonstrated that complete fixation of tyrosine hydroxylase in the dopaminergic system with formaldehyde maybe obtained using a very basic formaldehyde solution (pH 11) while still retaining immunoreactivity of the enzyme. The localization that was achieved at lower pH appeared adequate until it was compared to the results obtained by perfusion at pH 11 in the second step.

The distribution of catecholaminergic fibers and cell bodies in the paraventricular and supraoptic nuclei of the hypothalamus was investigated with immunohistochemical methods in the adult albino rat. Sections through the nuclei were stained with antisera to the catecholamine synthesizing enzymes tyrosine hydroxylase (TH), dopamine-beta-hydroxylase (DBH), and phenylethanolamine-N-methyltransferase (PNMT). The results suggest that adrenergic (PNMT-stained) fibers innervate the entire parvocellular division of the paraventricular nucleus, although the highest density of fibers was found in the medial part of the division. Only widely scattered adrenergic fibers are found in the magnocellular division of the nucleus and in the supraoptic nucleus. Noradrenergic fibers appear to innervate the periventricular zone of the paraventricular nucleus and those parts of the paraventricular and supraoptic nuclei that contain predominantly vasopressinergic neurons in both the normal and in the homozygous Brattleboro rat. Significant numbers--somewhat more than 500--of dopaminergic (TH-stained) neurons are found in the paraventricular nucleus; the cells are distributed throughout the nucleus but are concentrated in the medial and periventricular parts of the parvocellular division. Double-labeling experiments with the retrogradely transported tracer true blue indicate that between 4% and 8% of the dopaminergic neurons in the paraventricular nucleus project to the region of the dorsal vagal complex and/or thoracic levels of the spinal cord. It is concluded that adrenergic inputs to the paraventricular nucleus may influence cells that project to the median eminence and to preganglionic autonomic cell groups in the medulla and spinal cord. Noradrenergic inputs to the supraoptic and paraventricular nuclei may influence primarily vasopressinergic cells that project to the posterior lobe of the pituitary, as well as cells in the periventricular part of the paraventricular nucleus that project to the median eminence.