Direct and indirect roles of Fgf3 and Fgf10 in innervation and vascularisation of the vertebrate hypothalamic neurohypophysis.

The neurohypophysis is a crucial component of the hypothalamo-pituitary axis, serving as the site of release of hypothalamic neurohormones into a plexus of hypophyseal capillaries. The growth of hypothalamic axons and capillaries to the forming neurohypophysis in embryogenesis is therefore crucial to future adult homeostasis. Using ex vivo analyses in chick and in vivo analyses in mutant and transgenic zebrafish, we show that Fgf10 and Fgf3 secreted from the forming neurohypophysis exert direct guidance effects on hypothalamic neurosecretory axons. Simultaneously, they promote hypophyseal vascularisation, exerting early direct effects on endothelial cells that are subsequently complemented by indirect effects. Together, our studies suggest a model for the integrated neurohemal wiring of the hypothalamo-neurohypophyseal axis.

Secretory cells of the supraoptic nucleus have central as well as neurohypophysial projections.

Conventional neuroanatomical methods may fail to demonstrate the presence of axons that are finer than 1 microm in diameter because such processes are near or below the limit of resolution of the light microscope. The presence of such axons can, however, be readily demonstrated by recording. The most easily interpreted type of recording for this purpose is the demonstration of antidromic activation of the cell body following stimulation of the region through which the axon passes. We have exploited this technique in the hypothalamus and have demonstrated the presence of double axonal projections or axons branching very near the cell bodies of the secretory cells of the neurohypophysial system in the rat supraoptic nucleus. We found that a small proportion of supraoptic magnocellular cells could be antidromically activated both from the neural stalk and from elsewhere in the hypothalamus, including the suprachiasmatic nucleus (8 cells of a total of 182) and the antero-ventral third ventricular region (AV3V; 4 of 182 cells) near the organum vasculosum of the lamina terminalis (OVLT). Collision of antidromic and orthodromic spikes showed that the cells were clearly antidromically (rather than synaptically, or orthodromically) activated from both sites. A stimulus applied to one of the axons prevented propagation of a spike evoked by a pulse delivered to the other axon until sufficient time had elapsed after the first stimulus for the resultant spike to have propagated from the first stimulus site along one cell process (towards the cell body or branch point), and from this point along the other axonal branch to the second stimulus site (there was also a short additional delay period during which the axon at the site of the second stimulus recovered from its absolute refractory period). If the interval between the stimuli was progressively reduced, there came a point where the second spike failed. Such a clear demonstration of dual projections in a system where the cells were previously thought to have only a single axon raises the possibility that many nerve cells in the CNS have previously unsuspected projections.

Secretion-related uptake of horseradish peroxidase in neurohypophysial axons.

During secretion of the neurohypophysial hormones, oxytocin and vasopressin, secretory granule membrane is added to the plasma membrane of the axon terminals. It is generally assumed that subsequent internalization of this additional membrane occurs by endocytosis. In order to study this process, we have traced the uptake of intravenously injected horseradish peroxidase by neurohypophysial axons in rats and golden hamsters. Peroxidase reaction product within the secretory axons was found mainly in vacuolar and C-shaped structures of a size comparable with or larger than the neurosecretory granules. Our observations suggest that these large horseradish peroxidase (HRP)-impregnated vacuoles arise directly by a form of macropinocytosis. Morphometric analysis indicated that this form of membrane retrieval increased significantly after the two types of stimuli used, reversible hemorrhage and electrical stimulation of the pituitary stalk. Microvesicular uptake of HRP was found to be comparatively less.

Isolated nerve endings (neurosecretosomes) from the posterior pituitary. Partial separation of vasopressin and oxytocin and the isolation of microvesicles.

Subcellular fractions of the bovine posterior pituitary, including one composed almost exclusively of pinched-off nerve endings (neurosecretosomes), were characterized electron microscopically, hormonally, and enzymically. 15% of the nerve terminals in the gland were isolated as neurosecretosomes, as estimated from determinations of lactic dehydrogenase, a soluble, cytoplasmic enzyme. Neurosecretosomes were subdivided into three fractions by density-gradient centrifugation. The three subfractions, each shown to be nearly homogeneous populations of neurosecretosomes by means of electron microscopic and enzymic criteria, differed from each other in their vasopressin/oxytocin (VP/OT) ratios. The VP/OT ratio increased from the lightest to the densest fraction, indicating that VP is localized to denser and OT to lighter neurosecretosomes; similar results have been obtained previously for subfractions of neurosecretory granules (NSG). No morphological differences were apparent in neurosecretosomes among the three subfractions. Although complete separation of VP and OT was not achieved, the findings suggest that VP and OT are each stored in a different species of nerve ending and support the hypothesis that a given neurosecretory cell synthesizes, stores, and secretes only one of the peptide hormones. Microvesicles, 40-80 mmicro diameter and contained in typical neurosecretory cell terminals, are believed to be degradation products of membrane ghosts of depleted NSG; electron micrographs indicative of this transformation are presented. A fraction rich in microvesicles, but containing some NSG membranes, was prepared by density-gradient centrifugation of an osmolysate of neurosecretosomes. Smaller, apparently nonneurosecretory nerve endings, lacking NSG but filled with small vesicles, are occasionally seen in sections from whole gland. The vesicles in these atypical posterior pituitary nerve endings may be true neurohumor-containing, "synaptic" vesicles.

Microvesicles of the neurohypophysis are biochemically related to small synaptic vesicles of presynaptic nerve terminals.

Nerve endings of the posterior pituitary are densely populated by dense-core neurosecretory granules which are the storage sites for peptide neurohormones. In addition, they contain numerous clear microvesicles which are the same size as small synaptic vesicles of typical presynaptic nerve terminals. Several of the major proteins of small synaptic vesicles of presynaptic nerve terminals are present at high concentration in the posterior pituitary. We have now investigated the subcellular localization of such proteins. By immunogold electron microscopy carried out on bovine neurohypophysis we have found that three of these proteins, synapsin I, Protein III, and synaptophysin (protein p38) were concentrated on microvesicles but were not detectable in the membranes of neurosecretory granules. In addition, we have studied the distribution of the same proteins and of the synaptic vesicle protein p65 in subcellular fractions of bovine posterior pituitaries obtained by sucrose density centrifugation. We have found that the intrinsic membrane proteins synaptophysin and p65 had an identical distribution and were restricted to low density fractions of the gradient which contained numerous clear microvesicles with a size range the same as that of small synaptic vesicles. The peripheral membrane proteins synapsin I and Protein III exhibited a broader distribution extending into the denser part of the gradient. However, the amount of these proteins clearly declined in the fractions preceding the peak of neurosecretory granules. Our results suggest that microvesicles of the neurohypophysis are biochemically related to small synaptic vesicles of all other nerve terminals and argue against the hypothesis that such vesicles represent an endocytic byproduct of exocytosis of neurosecretory granules.

Glutamatergic innervation of the hypothalamic median eminence and posterior pituitary of the rat.

Recent studies have localized the glutamatergic cell marker type-2 vesicular glutamate transporter (VGLUT2) to distinct peptidergic neurosecretory systems that regulate hypophysial functions in rats. The present studies were aimed to map the neuronal sources of VGLUT2 in the median eminence and the posterior pituitary, the main terminal fields of hypothalamic neurosecretory neurons. Neurons innervating these regions were identified by the uptake of the retrograde tract-tracer Fluoro-Gold (FG) from the systemic circulation, whereas glutamatergic perikarya of the hypothalamus were visualized via the radioisotopic in situ hybridization detection of VGLUT2 mRNA. The results of dual-labeling studies established that the majority of neurons accumulating FG and also expressing VGLUT2 mRNA were located within the paraventricular, periventricular and supraoptic nuclei and around the organum vasculosum of the lamina terminalis and the preoptic area. In contrast, only few FG-accumulating cells exhibited VGLUT2 mRNA signal in the arcuate nucleus. Dual-label immunofluorescent studies of the median eminence and posterior pituitary to determine the subcellular location of VGLUT2, revealed the association of VGLUT2 immunoreactivity with SV2 protein, a marker for small clear vesicles in neurosecretory endings. Electron microscopic studies using pre-embedding colloidal gold labeling confirmed the localization of VGLUT2 in small clear synaptic vesicles. These data suggest that neurosecretory neurons located mainly within the paraventricular, anterior periventricular and supraoptic nuclei and around the organum vasculosum of the lamina terminalis and the preoptic area secrete glutamate into the fenestrated vessels of the median eminence and posterior pituitary. The functional aspects of the putative neuropeptide/glutamate co-release from neuroendocrine terminals remain to be elucidated.

Possible involvement of brain-derived neurotrophic factor (BDNF) in the innervation of dopaminergic neurons from the rat periventricular nucleus to the pars intermedia.

Developing neurons are guided to their appropriate targets by specific guidance substances that have neurotrophic actions. The aim of the present study was to elucidate the mechanism by which hypothalamic neurons reach the pars intermedia (PI) by correlating the development of dopaminergic (DA) neurons arising in the periventricular nucleus (PeV) of fetal rats with the expression of brain-derived neurotrophic factor (BDNF) in the rat pituitary. The differentiation of DA neurons was observed by immunohistochemistry using an antibody against tyrosine hydroxylase (TH), whereas the ontogenesis of BDNF mRNA in the PI was examined by in situ hybridization and RT-PCR. Immunoreactive TH-neurons were first observed in the PeV at embryonic day (E) 16.5, following which time their axons elongated toward the pituitary. TH-positive reactions were observed in the connective tissue between the PI and the pars nervosa at E20.5. Innervation of the PI by TH-positive neurons was determined at postnatal day (P) 1.5; however, BDNF mRNA was first detected in the PI cells at E17.5, with an increase in its expression clearly visible at E21.5 and continuing high expression levels in the PI thereafter. These results suggest that BDNF is a specific guidance cue for DA neurons elongating from the PeV to the PI.

Dopamine D2 receptor expression in the corticotroph cells of the human normal pituitary gland.

The dopamine D2 receptor is the main dopamine receptor expressed in the human normal pituitary gland. The aim of the current study was to evaluate dopamine D2 receptor expression in the corticotroph cell populations of the anterior lobe and pars intermedia, as well as posterior lobe of the human normal pituitary gland by immunohistochemistry. Human normal pituitary gland samples obtained from routine autopsies were used for the study. In all cases, histology together with immunostaining for adrenocorticotropic hormone, melanocyte-stimulating hormone, prolactin, and neurofilaments were performed and compared to the immunostaining for D2 receptor. D2 receptor was heterogeneously expressed in the majority of the cell populations of the anterior and posterior lobe as well as in the area localized between the anterior and posterior lobe, and arbitrary defined as "intermediate zone". This zone, characterized by the presence of nerve fibers included the residual pars intermedia represented by the colloid-filled cysts lined by the remnant melanotroph cells strongly expressing D2 receptors, and clusters of corticotroph cells, belonging to the anterior lobe but localized within the cysts and adjacent to the posterior lobe, variably expressing D2 receptors. D2 dopamine receptor is expressed in the majority of the cell populations of the human normal pituitary gland, and particularly, in the different corticotroph cell populations localized in the anterior lobe and the intermediate zone of the pituitary gland.

Innervation of the rat anterior and neurointermediate pituitary visualized by immunocytochemistry for the growth-associated protein GAP-43.

Immunocytochemical localization of the neuronal growth associated protein GAP-43 revealed a dense axonal plexus throughout the neurointermediate lobe of the rat pituitary. These axons were fine, presumably monoaminergic fibers, whereas magnocellular neurosecretory axons did not appear to contain detectable GAP-43. These experiments also revealed the presence of an extensive nerve plexus within the anterior lobe. Fine beaded fibers were present throughout the parenchyma of the anterior lobe, and punctate staining suggestive of nerve terminals was seen surrounding numerous endocrine cells. Nerve fibers did not appear to cross directly between the intermediate and anterior lobes, but rather entered the anterior lobe directly from its margins or in association with blood vessels. Preabsorption of antisera with GAP-43 purified from neonatal rat brain completely eliminated immunoreactivity. These findings confirm the existence of a direct innervation of the anterior pituitary of the rat; moreover, the presence of GAP-43 in these fibers suggests that they may be capable of growth and terminal reorganization in the adult animal.

Marked changes of arginine vasopressin, oxytocin, and corticotropin-releasing hormone in hypophysial portal plasma after pituitary stalk damage in the rat.

Mechanical compression of the pituitary stalk with the help of a blunt stereotaxic knife results in posterior pituitary denervation (PPD) and sprouting proximal to the injury, leading to formation of an ectopic neurohypophysis in the stalk. This provides an experimental model for those cases in which traumatic damage severs the nerve fibers to the neural lobe but does not obliterate the hypophysial-portal circulation. The effect of PPD on the hypophysial-portal concentration profile of putative ACTH secretagogues as well as basal and stimulated ACTH secretion in vitro were investigated at varying times after PPD. The contents of arginine vasopressin (AVP) and oxytocin (OT) in extracts of the stalk median eminence 1 week after PPD were markedly elevated, whereas corticotropin-releasing hormone (CRH) content was unaffected. Levels of these three neuropeptides in hypophysial-portal blood collected under anesthesia from the proximal stump of the transected stalk (or the ectopic neural lobe) were measured at weekly intervals in groups of rats after sham or PPD surgery. Hypophysial-portal AVP levels showed a monotonic increase with time after PPD from a 1.8-fold elevation at 1 week post-PPD to a maximum concentration 6-fold greater than that in sham groups at 4 weeks post-PPD. Portal plasma OT levels also exhibited extreme elevation. In contrast, portal plasma CRH levels showed an initial 72% decline 1 week post-PPD. We suggest that mechanical damage to the pituitary stalk and the subsequent sprouting redirected secretion of AVP and OT from the neural lobe to the pituitary stalk. This caused sustained elevations of portal plasma concentrations of AVP and OT. The resulting tonic exposure to AVP and/or OT may down-regulate anterior pituitary receptors to these neurohypophyseal peptides and indirectly decrease CRH release into the portal circulation.

Cells of origin of the afferent fibers to the median eminence in the rat.

The cell bodies of origin of axons terminating in the median eminence have been identified by retrograde axonal transport of horseradish peroxidase (HRP) or 125I-wheat germ agglutinin (WGA). The tracers were injected into the median eminence by pressure and under direct visual control, using a ventral surgical approach. The retrogradely labeled cells are exclusively located within the hypothalamus. The most heavily labeled cells are parvocellular neurons in the arcuate nucleus, the periventricular area, the medial part of the paraventricular nucleus, and the rostral paraventricular nucleus; a few cells are also located in the rostral part of the preoptic area immediately lateral and dorsal to the organum vasculosum of the lamina terminalis (OVLT). Less heavily located cells are found in the magnocellular neurosecretory nuclei, including the lateral parts of the paraventricular and rostral paraventricular nuclei, the supraoptic nucleus, and the accessory magnocellular nuclei. Retrogradely labeled cells are not found in the ventromedial hypothalamic nucleus, except for a few lightly labeled cells in the posterior division of the nucleus. However, if the injected tracer spreads into the arcuate nucleus, labeled cells are present throughout the ventromedial nucleus. Labeled cells are not found in other parts of the hypothalamus, including lateral and posterior portions of the preoptic area, the anterior hypothalamic area, and the suprachiasmatic nucleus, or in catecholaminergic cell groups of the midbrain, pons, and medulla. Control injections of HRP into the posterior pituitary and the ventromedial nucleus produce patterns of cell labeling which are very distinct from that seen with injections into the median eminence. Following injections into the posterior pituitary, the cells of the magnocellular neurosecretory nuclei are all heavily labeled, but small cells in the parvocellular neuronal groups are not labeled. Direct injections into the ventromedial nucleus resulted in labeled cells in widespread parts of the hypothalamus, as well as in the bed nucleus of the stria terminalis and the lateral septum, in parts of the amygdaloid complex and the subiculum, and in several cell groups in the midbrain, pons, and medulla.

Immunocytochemistry and in situ hybridization of neuropeptide Y in the hypothalamus of Xenopus laevis in relation to background adaptation.

The amphibian Xenopus laevis is able to adapt to a dark background by releasing melanophore-stimulating hormone from the pars intermedia of the pituitary gland. The inhibition of melanophore-stimulating hormone release is accomplished by neuropeptide Y-containing axons innervating the pars intermedia. To determine the production site of neuropeptide Y involved in this inhibitory control, the distribution of neuropeptide Y in the brain has been investigated by immunocytochemistry and in situ hybridization. Immunoreactive cell bodies were visualized in, among others, the ventromedial and posterior thalamic nuclei, and the suprachiasmatic and ventral infundibular hypothalamic nuclei. A positive hybridization signal with a Xenopus-specific probe for preproneuropeptide Y-RNA was found in the diencephalic ventromedial thalamic nucleus and in the suprachiasmatic nucleus. With both immunocytochemistry and in situ hybridization, suprachiasmatic neurons appeared to be stained only in animals adapted to a white background; animals adapted to a black background showed no staining. Quantitative image analysis revealed that this effect of background adaptation is specific for suprachiasmatic neurons because no effect could be demonstrated of the background light condition on the ventral infundibular nucleus (immunocytochemistry) or the ventromedial thalamic nucleus (in situ hybridization). These results indicate that neurons in the suprachiasmatic nucleus enable the adaptation of X. laevis to a white background, by producing and releasing neuropeptide Y that inhibits the release of melanophore-stimulating hormone from the melanotrope cells in the pars intermedia of the pituitary gland.

The relationship between the membrane potential of neurosecretory nerve endings, as measured by a voltage-sensitive dye, and the release of neurohypophysial hormones.

The membrane potential of isolated rat neurohypophyses and isolated neurosecretosomes (neurosecretory nerve endings) was monitored with the voltage sensitive fluorescent probe diS-C3-(5). K ions, in contrast to Na or Cl ions, give rise to large changes of the fluorescent signal. The fluorescent response is linearly related to log[K+]0 at values higher than 10 mM, whereas at lower [K+]0 the permeability of the membrane for Na ions has to be taken into account. Veratridine increases the fluorescent signal only in the presence of external sodium; this effect is blocked by tetrodotoxin. After prolonged K-induced depolarisation, addition of veratridine to the medium gives a further change in fluorescence of diS-C3-(5) associated with release of vasopressin. Vasopressin release from isolated neurohypophyses started to increase significantly only above 25 mk [K+]0, while the depolarization of the membrane was linearly related to log[K+]0. The results are consistent with the view that neurosecretory nerve endings have voltage-dependent calcium channels that regulate the amount of hormone released during depolarisation.

Intragranular colocalization of immunoreactive methionine-enkephalin and oxytocin within the nerve terminals in the posterior pituitary.

To determine differential tissue antigens in the same section immunocytochemically using the electron microscope, the neurohypophysis was examined following the application of a freeze-drying tissue preparation and staining with the protein A-colloidal gold-antibody complex method (Hisano S, Adachi T, Daikoku S: J Histochem Cytochem 32:705, 1984). At the light microscopic level, colocalized immunostaining for methionine-enkephalin (ENK) and oxytocin (OXT) was found in the rat neurohypophysis under different physiological states. Small pieces of the neurohypophysial tissue were frozen and dried. The dried tissue was fixed with paraformaldehyde vapor and embedded. The ultrathin sections were stained with the antibody for ENK coupled with protein A-small colloidal gold, and antibody for OXT or vasopressin (VP) conjugated with protein A-large colloidal gold. The ultrastructures of the nerve terminals were well preserved and showed many membrane-limited secretory granules. It was possible to identify both OXT- and VP-containing nerve terminals as their secretory granules were differentially labeled with protein A-colloidal gold anti-OXT or anti-VP complex, respectively. The secretory granules, which were labeled with large gold particles for OXT, also carry small gold particles. It is evident that ENK coexists with OXT in the same granules.

Denervation of the rat posterior pituitary gland: validation of a stereotaxic method.

A stereotaxic surgical method was developed for interrupting the nerve fibres running through the rat pituitary stalk to the posterior pituitary gland without obliterating the hypothalamo-pituitary portal circulation. The pituitary stalk was compressed by the blunt tip of an L-shaped rotating knife. Successful operations produced mild diabetes insipidus, disappearance of arginine vasopressin from the neural lobe, accumulation of arginine vasopressin and neurosecretory material in the pituitary stalk and no infarction in the anterior lobe of the pituitary gland. In female rats, the oestrous cycle was only temporarily disturbed. Plasma prolactin and corticosterone levels were high during the first 24 h after the stalk compression but returned to normal baseline levels from the second day after the operation. One week after the operation plasma adrenocorticotropin and prolactin levels were in the control range while plasma alpha-melanocyte-stimulating hormone was elevated. Denervation of the posterior pituitary gland may help in studying the neural control of intermediate lobe function and the role of the neural lobe in various endocrine conditions, and may serve as a model for lesions of the pituitary stalk and formation of ectopic neurohypophysis in the human.

Galanin stimulates the release of cholecystokinin from nerve fibres in the pituitary neurointermediate lobe.

Galanin (GAL), a 29 amino acid peptide, is present in magnocellular neurones in the paraventricular and supraoptic nuclei with projections to the neurohypophysis. The effect of GAL on the release of vasopressin, oxytocin and cholecystokinin (CCK) from rat neural lobes was investigated using an in vitro method. GAL in a concentration of 10(-6) M did not affect basal or K(+)-induced release of vasopressin or oxytocin. In contrast, GAL (10(-6) M) significantly stimulated basal and K(+)-stimulated release of CCK. Double-labelling immunofluorescence histochemistry of the neurohypophysis showed that GAL-immunoreactive (-IR) fibres co-contained vasopressin-like immunoreactivity (-LI), whereas the majority of oxytocin-IR fibres were CCK-IR. There was no evidence for colocalization of GAL with CCK or oxytocin. The data suggest a stimulatory role of GAL on CCK release via a paracrine effect on neighbouring oxytocin-CCK-containing nerve fibres.

Oxytocin stimulates oxytocin release from isolated nerve terminals of rat neural lobes.

Nerve endings from rat neural lobes isolated by homogenization were placed on a filter and constantly superfused. The effects of exogenous oxytocin and vasopressin (both added at 1 nM concentration) on basal and stimulated release of oxytocin and vasopressin were investigated. Stimulated release was evoked by 30 mM potassium and a simultaneous increase in osmolarity. A stimulatory effect of oxytocin on basal and evoked release of oxytocin was found while there was no effect on vasopressin release. The addition of vasopressin did not induce any change in the release of either hormone. The positive feedback mechanism observed with oxytocin might be active during the discharge of oxytocin which is known to occur in bursts.

Modulation of oxytocin and vasopressin release from rat neurosecretosomes: the roles of VIP oxytocin and GABA.

The effects of vasoactive intestinal polypeptide (VIP), of a selective oxytocin antagonist and of GABA on basal and stimulated oxytocin and vasopressin release from isolated neurosecretory endings were investigated. Superfusion of the secretosomes with VIP (10(-7) M) induced an increased basal and stimulated release of both oxytocin and vasopressin. Addition of the oxytocin antagonist induced a decrease of the stimulated oxytocin release as compared to the control which indicated a positive feedback mechanism of oxytocin on oxytocin release. In presence of GABA (1 or 50 microM) no change in basal or stimulated oxytocin and vasopressin release was observed.

Morphological evidence for the involvement of calcium in neurohypophysial hormone release.

An elevated intracellular Ca2+ concentration in the secretory nerve endings of the rat neurohypophysis was detected histochemically by means of light microscopy concomitant with the vasopressin secretion evoked by hypertonic saline. The electron microscopic and X-ray microanalytical results furnish morphological evidence for the function-dependent Ca2+ storage capacity of the mitochondria, and suggest their role in the regulation of the free Ca2+ level in the neurosecretory axon terminal.

Electrophysiological study of paraventricular nucleus neurons projecting to the dorsomedial medulla and their response to baroreceptor stimulation in rats.

In male rats anesthetized with urethane, extracellular recordings were made from 415 neurons in the paraventricular nucleus (PVN) and adjacent areas. Of these neurons 64 were excited antidromically by stimulation of the dorsomedial medulla but not by stimulation of the pituitary stalk (first group). Seventy-three neurons were antidromically excited by stimulation of the pituitary stalk but not of the dorsomedial medulla (second group, neurosecretory cells). The other 2 neurons were antidromically excited by stimulation of both the dorsomedial medulla and the pituitary stalk (third group). Latencies of antidromically evoked action potentials by stimulation of the dorsomedial medulla and of the pituitary stalk ranged between 8 and 60 ms (mean +/- S.D., 38.5 +/- 9.8, n = 66) and from 7 to 24 ms (mean +/- S.D., 13.0 +/- 3.6, n = 75), respectively, suggesting unmyelinated fiber projections in both instances. PVN neurons of these 3 groups were found to be dispersed throughout the PVN and no difference in specific locations between the neuron groups existed. Their characteristics, however, were different. The first group of neurons discharged at a slower rate and showed no phasic pattern of firing, while 28% of the second group of neurons ('identified' neurosecretory cells) showed phasic patterns of firing and their rates of discharge were higher than those of the first group of neurons. The two neurons belonging to the third group showed irregular spontaneous discharges. The areas within the dorsomedial medulla stimulation of which evoked antidromic excitation of PVN neurons were located within and adjacent to the nucleus of the tractus solitarius (NTS) and the dorsal motor nucleus of the vagus (DMV). Among PVN neurons which were antidromically excited by stimulation of dorsomedial medulla, 51 cells were examined for their responses to excitation of baroreceptors. An increase in pressure of the 'isolated' carotid sinus excited 2 neurons, and inhibited 7 (14%). On the other hand, 27% (11 out of 41) of neurosecretory cells (second group) were inhibited by baroreceptor stimulation. From these results, it was concluded that essentially separate populations of PVN neurons project to the neurohypophysis and to the NTS, DMV and their vicinities, and that some of the caudally-projecting PVN neurons receive synaptic input from carotid baroreceptor reflex pathway, suggesting the possible involvement of these PVN neurons in central cardiovascular regulation.