Paraventricular nucleus: a site for the integration of neuroendocrine and autonomic mechanisms

Vasopressin influences renal function via a spinal action

Intrathecal injections of arginine vasopressin (AVP) (3 pmol) into the lower thoracic region in rats produced consistent, dose-dependent antidiuretic responses accompanied by elevated urine osmolality. That these responses are not due to leakage of the peptide to the periphery is suggested by (a) the lack of response to the analogue, DDAVP, and (b) the lack of antidiuretic response to intrathecal AVP in denervated kidneys.

Comparison of the metabolic and behavioral disturbances following paraventricular- and ventromedial-hypothalamic lesions

Lesions of the ventromedial hypothalamus (VMH) result in an obesity syndrome with several metabolic and behavioral manifestations. It has also been reported that damage to the paraventricular hypothalamus (PVH) leads to changes characteristic of obesity. However, little is known about the consequence of PVH lesions, especially in contrast to the extensive documentation of VMH lesion-induced effects. To assess the basic features of the two hypothalamic obesity syndromes, rats underwent VMH, PVH, or sham lesions and, for 15 weeks, were maintained ad lib on a series of test diets. Both lesions groups were hyperphagic and showed similar weight gains. Although both lesion groups became obese (measured by % carcass fat), VMH rats were fatter than PVH animals. Similarly, only VMH rats were hyperinsulinemic. Further tests were conducted in PVH and VMH rats restricted to control body weights. VMH, but not PVH, rats developed a persisting elevation in basal gastric acid secretion. As well, only VMH, and not PVH, animals developed an obesity when restricted to normal weights. These data indicate similarities in PVH and VMH rats maintained ad lib but experiments on restricted animals reveal fundamental differences in the two obesities and point to different etiologies.

Impaired diet-induced thermogenesis in brown adipose tissue from rats made obese with parasagittal hypothalamic knife-cuts

Two experiments were performed to determine if bilateral parasagittal hypothalamic knife-cuts (KCs), which produce long-term overeating and obesity, after biochemical indices of brown adipose tissue (BAT) reactivity to thermogenic stimuli. In the first study, responses to environmental cold were tested. Four weeks after surgery, KC rats had gained 4-5 times more weight than controls and were obese (increased Lee Obesity Index and weight of gonadal white fat). Before being sacrificed, groups of KC and control rats were exposed to 4 degrees C for 21 hr or remained at 28 degrees C. Interscapular BAT weighed 300% more in KC rats, due largely to increased white fat content. Functional indices of BAT thermogenic capacity (protein content, DNA content, cytochrome oxidase activity and mitochondrial guanosine diphosphate (GDP) binding) were normal at 28 degrees C. Exposure to 4 degrees C produced greatly enhanced responses but these were equivalent for both groups. This suggested an intact capacity for non-shivering thermogenesis in obese KC rats. In the second study, the same BAT responses were examined in other rats fed a palatable "cafeteria" diet (CAFE). One week after surgery, KC and control rats were subdivided into groups that received chow alone or chow plus four different palatable foods daily. Before sacrificing 4-5 weeks later, KC rats had gained 3-4 times more weight than controls and were obese. Interscapular BAT weighed 200-300% more in KC rats. CAFE feeding produced larger increments in all variables for KC vs. control rats. Most importantly, GDP binding was reduced in both KC groups, and significantly more so after CAFE feeding.(ABSTRACT TRUNCATED AT 250 WORDS)

Effects of neural stimuli on paraventricular nucleus neurones

The electrical activity of 122 neurones mostly located within the parvocellular components of the hypothalamic paraventricular nucleus (PVN) was recorded in urethane-anaesthetized male rats. Spontaneous activity and responses to single-shock stimulation of the following sites were recorded: median eminence (ME), neurohypophysis (PIT), preoptic area (POA), nucleus accumbens (ACB), subiculum (SUB), sciatic nerve (ScN), and photic (Ph) stimulation. Stimulation of ME antidromically invaded 14% of the cells recorded, the spontaneous firing rate of these neurones was significantly slower than that of the rest of the population. The responses of the neurones identified as projecting to ME, to the above stimuli, were generally found to be in the same direction as for the whole population but the proportions of the majority responses were greater. Identified cells mainly responded with orthodromic excitation following POA, SUB, ScN, and Ph stimulation whereas ACB stimulation produced a greater proportion of inhibitiory responses. High degrees of convergence were found following afferent stimuli. The possibility that some of the neurones projecting to ME may be candidates for the corticotropin-releasing factor (CRF-) cells, is discussed.

A light microscopic HRP study of limbic projections to the vasopressin-containing nuclear groups of the hypothalamus

The anterograde HRP technique has been used to define the efferent projections from the lateral septum, amygdala and ventral subiculum to the anterior hypothalamus (AH) with particular attention to the paraventricular nucleus (PVN). Each limbic region was found to project to the PVN in a perinuclear fashion leaving the nucleus itself virtually devoid of HRP-labeled terminals. This projection pattern was also characteristic of the limbic innervation of the supraoptic (SON) and suprachiasmatic (SCN) nuclei. HRP injections into limbic sites has also enabled the description of both efferent and afferent projections to the remainder of the diencephalon. These results extend the observations made previously on the projections from limbic structures to the diencephalon in the rat, particularly in regard to the distinctive relationship of efferents to the PVN and SON, the major components of the hypothalamo-hypophyseal neurosecretory system. At the light microscopic level it cannot be stated with certainty whether or not such limbic afferents synapse with the dendrites of cells in the PVN, SON or SCN which extend beyond the cellular boundaries of each nucleus.

The noradrenergic innervation of vasopressin neurons in the paraventricular nucleus of the hypothalamus: an ultrastructural study using radioautography and immunocytochemistry

Immunocytochemical and radioautographic procedures were combined at the ultrastructural level to study the noradrenergic synaptic input to vasopressin neurons in selected portions of the paraventricular nucleus of the hypothalamus (PVN) of the rat. Radioactive norepinephrine (NE) was infused into the lateral ventricle or applied topically to the region of the PVN. After appropriate survival times, brain tissues were processed for ultrastructural immunocytochemical demonstration of vasopressin using a monoclonal antibody. [3H]NE varicosities were detected by electron microscopic radioautography. In the periventricular zone of the PVN, radioactive varicosities were numerous accounting for 20-30% of all nerve terminals in this zones. These NE terminals primarily innervated dendritic processes of non-vasopressinergic neurons. Although an occasional axosomatic synapse was observed, input to vasopressin positive neurons was exclusively to their dendrites. In the lateral magnocellular sub-nucleus of the PVN (designed pvl2), noradrenergic terminals were fewer in number accounting for only 1-2% of the total. These terminals were found predominately but not exclusively making axodendritic synapses onto non-vasopressin processes. In both regions, many of the radiolabeled terminals had well-defined membrane appositions with their post-synaptic partners which included a synaptic cleft and post-synaptic density of varying thickness. In both the periventricular zone and the lateral magnocellular regions, noradrenergic varicosities were seen in close proximity to numerous blood vessels.

Ultrastructural studies on the afferent synaptic input to oxytocin-containing neurons in the paraventricular nucleus of the hypothalamus

A diverse afferent synaptic input to immunostained oxytocin magnocellular neurons of the paraventricular nucleus of the rat hypothalamus is described. By electron microscopy, immunoreactive material is present within cell bodies and neuronal processes and it is associated primarily with neurosecretory granules and granular endoplasmic reticulum. Afferent axon terminals synapse on perikarya, dendritic processes, and possibly axonal processes of oxytocin-containing neurons. The presynaptic elements of the synaptic complexes contain clear spherical vesicles, a mixture of clear spherical and ellipsoidal vesicles, or a mixture of clear and dense-centered vesicles. The postsynaptic membranes of oxytocinergic cells frequently show a prominent coating of dense material on the cytoplasmic face which gives the synaptic complex a marked asymmetry.

Adipose tissue mobilization is unaffected by obesifying hypothalamic knife cuts

Unilateral electrolytic lesions of the medial hypothalamus have previously been reported to spare starvation-induced lipid mobilization from the ipsilateral retroperitoneal fat pad. This suggests that deficient lipid mobilization contributes to the enhanced lipid deposition which results when such lesions are bilateral. In contrast, unilateral parasagittal hypothalamic knife cuts, which also obesify when bilateral, failed to prevent starvation-induced ipsilateral lipid mobilization. This finding indicates that impairment of neurally mediated lipid mobilization is probably neither a necessary nor a sufficient feature of hypothalamic obesity.

Studies of the central cardiovascular effects of vasopressin

Three approaches to the evaluation of the central cardiovascular effects of vasopressin are briefly reviewed. These include assessment of cardiovascular regulation in Brattleboro rats, cardiovascular responses to central nervous system injections of vasopressin, and changes in central nervous system vasopressin content in spontaneous and induced hypertension. The studies indicate that vasopressin could participate in baroregulation by actions within the central nervous system, but they fail to clearly define the importance of vasopressin, if any, in this capacity.

Analysis of the ACTH/beta-End/alpha-MSH-immunoreactive afferent input to the hypothalamic paraventricular nucleus of rat

Immunoreactive fibers and varicosities in the hypothalamic paraventricular nucleus (PVN) were examined by light- and electronmicroscopy, following treatment of brain slices with specific antibodies to adrenocorticotropin (ACTH), beta-endorphin (beta-End) and alpha-melanotropin (alpha-MSH) peptides. In an attempt to provide a more precise, quantitative definition of the densities of immunoreactive elements, sections were analyzed by computer based image-analysis techniques. Fibers and varicosities immunostained with the 3 different antibodies displayed an identical distribution pattern throughout the nucleus suggesting that they are parts of the same, arcuate pro-opiomelanocortin (POMC) neuron system. Although immunoreactive varicosities were found all over the PVN, it was possible to identify a characteristic, density distribution pattern. At the ultrastructural level, immunoreactive presynaptic nerve terminals were observed forming symmetrical synaptic contacts with unlabeled dendrites. The majority of immunoreactive elements were found in the dorsal parvo- and caudal magnocellular subdivisions which give rise to long projections to the lower brainstem. Moderate density of POMC neural elements was observed in the anterior and medial (ventral portion) parvocellular subdivisions which project to the external zone of the median eminence. Only a few, widely scattered immunostained varicosities are found in the medial and lateral magnocellular subdivisions which project to the neurohypophysis. A combined lesion and immunocytochemical approach has shown that the bulk of the afferent neuronal input from arcuate POMC cells enters the PVN from a ventral direction.

Neuronal organization of the avian paraventricular nucleus: intrinsic, afferent, and efferent connections

The heterogeneous paraventricular nucleus (PVN) of birds offers favorable conditions for the analysis of intrinsic, afferent, and efferent connections of neuroendocrine systems. Paraventricular neurons are successfully impregnated with the Golgi-technique. The findings indicate a direct influence of the cerebrospinal fluid (CSF) on the magnocellular neurons that, via their axon terminals in the neural lobe of the pituitary, are also exposed to the hemal milieu. The magnocellular neurons are intermingled with parvocellular elements which may represent local interneurons. A group of parvocellular nerve cells is identified as CSF-contacting neurons. This type of cell forms a basic morphologic component of the avian neuroendocrine apparatus. Immunocytochemical and ultrastructural studies further support the concept of neuronal interactions between parvocellular and magnocellular elements. Moreover, these findings speak in favor of the existence of recurrent collaterals of the magnocellular neurons. Nerve cells giving rise to afferent connections to the PVN are located in the limbic system and autonomic areas of the upper and lower brainstem. Further afferents may originate from the subfornical organ, the organon vasculosum laminae terminalis, the ventral tegmentum, and the area postrema. Via efferent projections, the PVN is connected to the nucleus accumbens, lateral septum, several hypothalamic nuclei, the neural lobe of the pituitary, the organon vasculosum laminae terminalis, the subfornical organ, the pineal organ, the area postrema, the lateral habenular complex, and various autonomic areas of the reticular formation in the upper and lower brainstem and the spinal cord. In conclusion, the PVN may be regarded as an integral component of the neuroendocrine apparatus reciprocally coupled to the limbic system, several circumventricular organs, and various autonomic centers of the brain.

A mechanism by which primary or secondary hypothalamic involvement results in the development of insulin-dependent diabetes mellitus (IDDM)

A literature survey and hypothesis is presented in which it is concluded that an intracellular ventromedial hypothalamic (VMH) glucopenia results in a bibrachial response consisting of adenohypophysial release of growth hormone and ACTH as well as sympathetic discharge, both of which act to elevate plasma glucose and remove the VMH glucopenia. This glucopenia may occur as a result of either a deficiency of circulating insulin or alterations in the kinetic properties of the VMH cellular insulin receptor. Two mechanisms for the development of insulin dependent diabetes mellitus (IDDM) are presented: (1) a defect in VMH glucose transport and/or metabolism such that a VMH glucopenia occurs with a subsequent bibrachial response. The result of this is glucose overproduction and a chronic excess glucose stimulus will eventually cause B-cell exhaustion (primary hypothalamic involvement). (2) reduction of the B-cell population by chemical, genetic and/or viral interactions with a consequential insulopenia results in a VMH glucopenia (secondary to a reduced glucose transport into the VMH cells) and causes a bibrachial response. This VMH response may temporarily restore plasma glucose balance but a chronically enhanced counter-regulatory response to maintain this balance will eventually stress the remaining B-cell population and cause further reductions in B-cell numbers (secondary hypothalmic involvement).

Brain and pituitary receptors for corticotropin releasing factor: localization and differential regulation after adrenalectomy

Specific receptors for corticotropin releasing factor (CRF) were identified in two functionally distinct systems within the brain, the cortex and the limbic system. Autoradiographic mapping of the CRF receptors in the brain revealed high binding density throughout the neocortex and cerebellar cortex, subiculum, lateral septum, olfactory tract, bed nucleus of the stria terminalis, interpeduncular nucleus and superior colliculus. Moderate to low binding was found in the hippocampus, nucleus accumbens, claustrum, nucleus periventricularis thalamus, mammillary bodies, subthalamic nucleus, periaqueductal grey, locus coeruleus and nucleus of the spinal trigeminal tract. As in the anterior pituitary gland, CRF receptors in the brain were shown to be coupled to adenylate cyclase. However, in contrast to the marked decrease in CRF receptors observed after adrenalectomy in the anterior pituitary gland, CRF receptor concentration in the brain and pars intermedia of the pituitary was unchanged. The presence of CRF receptors in areas involved in the control of hypothalamic and autonomic nervous system functions is consistent with the major role of CRF in the integrated response to stress.

Distribution of neurophysin II immunoreactive nerve fibers within the subnuclei of the nucleus of the tractus solitarius of the rat

The location of neurophysin II immunoreactive nerve fibers and preterminal processes has been examined in various functionally distinct subnuclei of the nucleus of the tractus solitarius (nTS) using the indirect immunofluorescence method for immunocytochemistry combined with cytoarchitectonic identification. The nTS is responsible for integrating respiratory and autonomic reflex activity: the vlnTS, vnTS, ni and nI are associated with respiratory activity; the dlnTS and dnTS are important sites for the integration of baroreceptor and chemoreceptor activity; the ncom, dnTS and dlnTS integrate cardiac afferent activity and the mnTS mediates both cardiovascular and gastrointestinal effects. At levels caudal to the obex, the ncom contained the largest number of neurophysin II immunoreactive nerve fibers and the mnTS and dmnX contained moderate neurophysin II immunoreactivity. At levels rostral to the obex the region of the dorsal medulla adjacent to the mnTS and dnTS (PVR and dPSR) showed the densest immunoreactivity and the mnTS, dmnX and vPSR showed moderate immunoreactivity. At the rostral pole of the nTS, neurophysin II immunoreactive nerve terminals were seen in the dendritic regions of cells in dmnX and mnTS. This selective distribution of neurophysin II immunoreactive nerve terminals in the cardiovascular and gastrointestinal subnuclei of the nTS implicates a direct, descending, hypothalamic, oxytocin-neurophysin II containing pathway interacting with these nTS functions. These results confirm the hypothesis (Sawchenko and Swanson) that descending neurophysin II immunoreactive pathways represent an important neuronal system for the hypothalamic regulation of cardiovascular (vasomotor) and gastrointestinal nuclei in the brainstem.

The role of vasopressin and the sympathetic nervous system in the cardiovascular response to vagal cold block in the conscious dog

This study examined the role of arginine vasopressin in the pressor response to vagal cold block and evaluated a possible interaction between vasopressin and the sympathetic nervous system during vagal block in conscious dogs with (carotid sinus intact) and without (sinoaortic denervated) functional arterial baroreflexes. In both carotid sinus intact and sinoaortic denervated dogs, elimination of the arginine vasopressin pressor system by the specific vasopressin antagonist d(CH2)5Tyr(Me)AVP did not alter the response to vagal block, as evaluated by changes in arterial pressure. Subsequent removal of the sympathetic nervous system by ganglionic blockade abolished the response to vagal block. When ganglionic blockade was induced in the absence of the vasopressin antagonist, the pressor response to vagal block was reduced by only 60%. Arginine vasopressin antagonist after ganglionic blockade reduced the response to vagal block by an amount equivalent to 45% of the original increase in pressure. The effects of blockade of either vasopressin or the sympathetic nervous system on the pressor response to vagal block were significantly greater when the other system had previously been eliminated. Data suggest that both arginine vasopressin and the sympathetic nervous system contribute to the pressor response to vagal block. One interpretation of these results is that vasopressin also interacts centrally to inhibit sympathetic outflow and thus modulates the hemodynamic manifestation of interruption of vagal afferents.

Vasopressin and oxytocin in the rat spinal cord: distribution and origins in comparison to [Met]enkephalin, dynorphin and related opioids and their irresponsiveness to stimuli modulating neurohypophyseal secretion

Immunoreactive-vasopressin, -oxytocin, -dynorphin, -dynorphin-(1-8), -alpha-neo-endorphin and -[Met]enkephalin were, in each case, present in greater concentrations in dorsal as compared to ventral, and lumbo-sacral as compared to cervico-thoracic, spinal cord. These differences were significantly more pronounced for vasopressin and oxytocin than for the other peptides. Lesions of the hypothalamic paraventricular nucleus depleted levels of immunoreactive-vasopressin and -oxytocin throughout the cord whereas levels of the opioid peptides therein were unaffected. In contrast, destruction of either the supraoptic or suprachiasmatic nucleus failed to change the content of immunoreactive-vasopressin, -oxytocin or any of the opioid peptides in the cord. Dehydration for 3 days depressed levels of immunoreactive-vasopressin, -oxytocin and -dynorphin in the neurointermediate lobe of the pituitary. In distinction, the levels of these were not modified in the spinal cord. Further, treatment with the synthetic corticosteroid, dexamethasone, elevated levels of immunoreactive-vasopressin, -oxytocin and -dynorphin in the neurointermediate pituitary whereas these were unaffected in the spinal cord. It is concluded that vasopressin and oxytocin in the spinal cord are predominantly derived from the paraventricular nucleus, localized in dorsal lumbo-sacral regions of the cord and insensitive to endocrinological manipulations. These pools may, thus, be modulated differently from their counterparts in the neurohypophysis and have a differing role, possibly in the control of the primary processing, autonomic or motor junctions. Further, there is no evidence from these or our prior studies for a close interrelationship of spinal cord vasopressin with dynorphin-related peptides (or oxytocin with [Met]enkephalin), likewise in contrast to the neurohypophysis.(ABSTRACT TRUNCATED AT 250 WORDS)

Clonidine-induced feeding: analysis of central sites of action and fiber projections mediating this response

Clonidine (CLON), an alpha-adrenergic agonist, was used in conjunction with norepinephrine (NE) to elicit feeding in satiated rats that had sustained hypothalamic electrolytic lesions, or coronal knife cuts at the hypothalamic, midbrain or pontine level of the brainstem. Electrolytic lesions of the paraventricular nucleus (PVN) of the hypothalamus significantly attenuated feeding normally stimulated by intraperitoneal injection of CLON. This contrasts with lesions in the dorsomedial or perifornical hypothalamic regions which had no effect on CLON-elicited eating. Knife cuts placed in the posterior hypothalamus and throughout the midbrain tegmentum also left intact the CLON eating response, in contrast to specific cuts in the dorsal pontine tegmentum which disrupted feeding elicited by PVN injections of NE and CLON, as well as by peripheral administration of CLON. Analyzed together, these results with effective and ineffective cuts relative to CLON and NE feeding provide evidence for an alpha-adrenergic feeding circuit which originates in the PVN and descends from this nucleus, via a dorsal periventricular course, through the diencephalon and midbrain. Further caudally, these fibers mediating NE and CLON feeding then appear to traverse ventrolaterally into the dorsolateral pontine tegmentum on their way to the dorsal medulla.

Distribution and projections of cholecystokinin-immunoreactive neurons in the hypothalamic paraventricular nucleus of rat

Analysis of coronal sections from colchicine-treated rat brains reveals that CCK-immunoreactivity (CCK-ir) is present in two distinguishable neuronal systems in the paraventricular nucleus (PVN). More than 60% of these cells were found to be typical parvicellular neurons; the remainder were magnocellular neurons. The magnocellular CCK-ir neurons were concentrated in the medial magnocellular subdivision, while more caudally they formed a ring around a zone of unstained magnocellular neurons. Immunostained parvicellular neurons predominate in medial and periventricular parvicellular subdivisions. The efferent projections of CCK-ir neurons were investigated by looking for retrograde accumulation of CCK-ir in cell bodies after selective knife cuts. A parasagittal cut of the lateral retrochiasmatic area as well as transection of the rostral median eminence resulted in an accumulation of CCK-ir material in a large number of both parvi- and magnocellular neurons. After pituitary stalk lesions, however, increased staining was only seen in magnocellular neurons. It is inferred that the magnocellular (presumed oxytocin-CCK) cells send their axons to the pituitary, whereas axons of CCK-ir parvicellular neurons appear to terminate in the median eminence. After transection of the medial forebrain bundle (MFB), immunostaining increased in a small number of scattered transected fibers proximal to the knife cut and in a few perikarya in the PVN, indicating that very few CCK cells may send descending fibers to the lower brainstem.

Catecholamine synapses and contents in the paraventricular hypothalamic nucleus and nucleus tractus solitarius of DOCA-salt hypertensive rats

Central catecholamine (CA) neurons in the nucleus tractus solitarius (NTS) and paraventricular hypothalamic nucleus (PVN) were studied in Wistar rats that had been unilaterally nephrectomized. The experimental animals were then treated with deoxycorticosterone acetate (DOCA) and salt water. The control animals were treated with the vehicle and tap water. Blood pressure of animals 4 weeks after DOCA/salt treatment was significantly elevated when compared to control rats. Morphologically, CA terminals showed no noticeable changes in the DOCA/salt hypertensive rats. Furthermore, the density of CA terminals either in the NTS or in the PVN of the DOCA/salt hypertensive rats was not statistically different from that of normotensive controls, suggesting that salt does not cause lesions or destruction of CA terminals. However, an extensive electron-microscopic morphometric analysis indicated that there was an enhancement of CA synaptogenesis (expressed by increased synaptic frequency among all CA boutons labeled with 5-hydroxydopamine) in the PVN, but not in the NTS of DOCA/salt hypertensive rats. In addition, the high-performance liquid chromatography revealed decreased CA contents in the PVN, but not in the NTS, of DOCA/salt hypertensive animals. Since synapses are primary sites for neurotransmitter release, the above results collectively suggest that more CA synapses formed in the PVN may reflect a net CA release from CA terminals resulting in the decreased CA content in the axonal terminals. Such an increased CA release and enhanced CA synaptogenesis may consequently enhance CA function in the PVN of hypertensive rats 4 weeks after DOCA/salt treatment, and relate to the development and/or maintenance of hypertension in the DOCA/salt rats.

Comparison of the distribution of oxytocin and vasopressin in the rat brain

While immunohistochemistry has been used extensively to map both oxytocin (OT) and vasopressin (VP) pathways in the brain, little information is available concerning the quantitative distribution of these hormones--particularly oxytocin. We have isolated oxytocin from extrahypothalamic regions of the rat brain and shown it to behave identically with standard oxytocin in radioimmunoassay (RIA) and on high-performance liquid chromatography. Using sensitive RIA we have measured and compared levels of both oxytocin and vasopressin in the rat brain. Both hormones are widely distributed, with the largest amounts outside the hypothalamus being found in the locus coeruleus. Considerable quantities of both peptides (but particularly oxytocin) are found in mesencephalic, pontine and medullary nuclei. This distribution is similar to that of the catecholamines, and the possible interaction of oxytocin and vasopressin with catecholaminergic pathways in the central control of various functions is discussed.

Localization of forebrain neurons which project directly to the medulla and spinal cord of the rat by retrograde tracing with wheat germ agglutinin

Wheat germ agglutinin (WGA) in a slow-release polyacrylamide gel pellet was implanted in the medulla or spinal cord of the rat. Large numbers of retrogradely labeled cells were visualized by immunocytochemical procedures in specific nuclei of the forebrain mainly ipsilateral to the implant site following implants as far caudal as the sacral segments of the spinal cord. Total average number of labeled forebrain cells (three brains per category; 100 micron per 150 micron of brain tissue were examined microscopically): medulla, 2,115; cervical, 1,878; lumbar, 1,017; sacral, 385. After WGA-gel implants in the medulla or cervical cord the majority of retrogradely labeled neurons were seen in the lateral hypothalamic area, the zona incerta, and in subdivisions of the paraventricular nucleus. A continuum of labeled cells extended from the caudal part of the paraventricular nucleus into the posterior hypothalamus and into the central gray of the midbrain. Labeled cells were also seen in the medial basal hypothalamus and the rostral part of the bed nucleus of the stria terminalis. A few labeled cells were observed in the medial and lateral preoptic areas, the rostral part of the paraventricular nucleus, and in the arcuate nucleus. Following WGA-gel implants in the lumbar or sacral cord many retrogradely labeled cells were observed mainly in the paraventricular nucleus, the lateral hypothalamus, zona incerta, medial basal hypothalamus, and posterior hypothalamic area. The continuum of labeled cells described above was also seen following these implants. Our data indicate that the lateral hypothalamus and zona incerta, as well as specific parts of the paraventricular nucleus, are major loci of neurons which project directly to the medulla and spinal cord of the rat. The consistency with which labeled cells were localized across all brains examined within categories of implant sites and the large numbers of labeled cells counted within these areas appeared to verify the sensitivity of our retrograde tracing method. Therefore, we interpret the paucity or absence of labeled cells in particular brain regions to indicate that cells of these regions do not project to the medulla or spinal cord.

Release of immunoassayable neurohypophyseal peptides from rat spinal cord, in vivo

The subarachnoid space of the spinal cord was perfused in vivo in urethane-anesthetized rats and perfusates were assayed for arginine-vasopressin (AVP) and oxytocin immunoreactivity. In control perfusates, oxytocin concentrations were 3 times those of AVP. Electrical stimulation of the paraventricular nucleus (PVN) of the hypothalamus, but not of other hypothalamic areas, yielded increased amounts of immunoassayable peptides in the spinal cord perfusates. Intravenous infusion of AVP did not elevate AVP concentrations in the cord perfusates. These data suggest that electrical stimulation of PVN neurons caused release of AVP and oxytocin from spinal cord terminals and support the concept that these peptides are neurotransmitters in the cord.

Paraventricular nucleus lesions in weanling female rats result in normophagia, normal body weight and composition, linear growth and normal levels of several plasma substrates

Female weanling rats received small (1 mAmp for 5 sec) electrolytic lesions in the paraventricular nuclei. Sham-operated rats served as controls. The rats were maintained for 42 days and body weight, linear growth, Lee Index, food intake and efficiency of food utilization were determined throughout the study. Plasma glucose, glycerol, free fatty acids, total protein and carcass fat and protein were determined at sacrifice. There was no significant difference between the lesioned and the sham-operated rats in any of the parameters measured. The findings are interpreted to mean that the PVN of the weanling rat is not functionally mature or alternatively, that there exists a sex difference in weanling rats in response to PVN lesions.

Electrophysiological investigations on the central innervation of the rat and guinea-pig pineal gland

The possible influence of central nervous structures on the electrical activity of single pineal cells was investigated in rat and guinea-pig. In the rat electrical stimulation of the hippocampal formation elicited both single cell responses with different latencies and mostly long-term excitations in single pineal cells, while stimulation of the habenular nuclei caused clear orthodromical responses with different latencies, alterations in the rate of spontaneous electrical activity and evoked discharges of "silent" units. In the guinea-pig electrical stimulation of the paraventricular nucleus influenced predominantly cells in the deeper layers of the posterior part of the pineal gland. Electrical stimulation of both the superior and inferior colliculi elicited field potentials with a constant latency, indicating a functional relationship between the corpora quadrigemina and the pineal organ.

Arginine-vasopressin inhibits centrally induced pressor responses by involving hippocampal mechanisms

Administration of arginine-vasopressin (AVP) or prolyl-leucyl-glycinamide (PLG) into a lateral cerebral ventricle reduced the magnitude of systolic blood pressure increase (pressor response) induced by electrical stimulation of the mesencephalic reticular formation (MRF) in urethane-anesthetized rats. Bilateral destruction of the dorsal hippocampus prevented the action of AVP on the pressor response. However, the effect of PLG was only slightly reduced by hippocampal lesion. Microinjection of AVP in the dentate area of the dorsal hippocampus mimicked the action of intracerebroventricularly administered peptides. The effect of a single injection of AVP lasted at least for 60 min. Neither hippocampal damage nor peptide administrations resulted in changes in mean arterial blood pressure (basal BP). Bradycardiac response accompanied the BP increase during MRF stimulation. Hippocampal damage or intracerebroventricular administration of AVP and PLG failed to affect the cardiac response. Injection of AVP into the hippocampus tended to reduce the magnitude of cardiac responses caused by MRF stimulation. It is suggested that the inhibition by AVP of a pressor response produced by MRF stimulation involves the dorsal hippocampus. The action of PLG or related peptides seems to be, at least in part, through mechanisms not involving the hippocampus.

Diurnal variation of uterine contractility

The diurnal uterine activity in four normal women in the secretory phase of their menstrual cycles and one woman suffering from dysmenorrhea were studied in relation to concomitant hormone levels in blood (progesterone, hGH, prolactin, cortisol, vasopressin, and 15-keto-13,14-dihydro-PGF2 alpha). In the four normal women uterine activity decreased after midnight, unrelated to circulating levels of 15-keto-13,14-dihydro-PGF2 alpha. But during a dysmenorrheic episode the uterine hypercontractility pattern correlated well with levels of the PGF2 alpha-metabolite, indicating a role of endogenous-produced PGF2 alpha in this condition. The results demonstrate a diurnal rhythm, possibly related to the wake-sleep cycle. No simple associations were seen between vasopressin, cortisol, prolactin, hGH, the PGF2 alpha-metabolite, and uterine activity.

Metabolic responses to suckling in postpartum lactating rats

The effects of suckling on the metabolic activity both of pituitary tissue and of subnuclei in the supraoptic and paraventricular nuclei of the hypothalamus were assessed by means of the [14C]deoxyglucose (2-DG) method of autoradiography. Three groups of female rats were deprived of their pups and of food for 6 h on the fifth or sixth day after parturition. Metabolic activity was assessed either before, during, or after a bout of suckling. The anterior pituitary is more active in suckled than in either non-suckled or postsuckled females. These differences may be a consequence of depletion/transformation/repletion of prolactin during a bout of suckling. The posterior pituitary is metabolically more active in suckled females and in postsuckled females than in non-suckled rats. These differences may reflect metabolic activity needed for the dynamics of oxytocin release and the restoration of ionic gradients. The dorsal and ventral subdivisions of the supraoptic nucleus and the lateral, dorsal, and posterior subnuclei of the paraventricular nucleus were more active in postsuckled females than in non-suckled females. These metabolic changes in the subnuclei of the magnocellular neurohypophysial system may reflect increased resynthesis of oxytocin (and vasopressin) in the cell bodies following a bout of suckling.

Vasopressin, oxytocin and neurophysins in the human brain and spinal cord

The concentrations of arginine vasopressin, oxytocin, and their related neurophysins were compared in many areas of postmortem human brain and spinal cord using specific radioimmunoassays. In the hypothalamus the ratio of vasopressin to oxytocin was approximately 3:1, and in the extrahypothalamic areas of the brain the greatest amount of both peptides was present in the locus coeruleus, and to a lesser extent the periaqueductal grey. Vasopressin only was found in the substantia nigra, and globus pallidus. In the medulla, including the nucleus of the solitary tract, the dorsal nucleus of the vagus, and the nucleus of the spinal tract of the trigeminal nerve, the amount of oxytocin was greater than that of vasopressin. In the spinal cord oxytocin predominated over vasopressin to an even greater extent, and reached particularly high values in certain segments of the intermediolateral grey column and dorsal horn. Estrogen-stimulated and nicotine-stimulated neurophysins (ESN and NSN) were both found in large amounts in those areas of the brain and spinal cord where the concentrations of the nonapeptides were greatest, but when the molar ratios of ESN to oxytocin and NSN to vasopressin were compared there was an excess of ESN.

Axon collaterals of supraoptic neurones: anatomical and electrophysiological evidence for their existence in the lateral hypothalamus

The magnocellular neurones of the supraoptic nucleus which synthesize and secrete vasopressin and oxytocin have been commonly regarded as simple "output" neurones in that they receive an input, generate an action potential and in turn release hormone from their terminals in the posterior pituitary. Three lines of evidence are presented which suggest that rat supraoptic nucleus neurons also have axon collaterals which terminate in the hypothalamus close to the nucleus. Small injections of horseradish peroxidase were made directly into the nucleus in hypothalamic slices, allowing visualization of the axons of supraoptic neurones. Collaterals of these axons could be observed in regions both dorsal and dorsolateral to the supraoptic nucleus. In a separate series of experiments, sections of perfusion-fixed hypothalamus were stained for vasopressin and oxytocin using specific antisera. Peptide-containing collaterals of both types were observed near the supraoptic nucleus, in a region similar to that seen after horseradish peroxidase injections. Finally, electrophysiological studies were carried out on hypothalamic slices containing the supraoptic nucleus. A small concentric bipolar stimulating electrode was placed directly into the nucleus and activity of lateral hypothalamic neurones within 0.1-1 mm of the nucleus was recorded. Of 68 neurones studied, 52 were excited by supraoptic stimulation via a synaptic pathway that could be blocked by Ca2+ -free solutions containing 18 mM Mg2+. These studies suggest that supraoptic neurones communicate via axon collaterals with other neurones in the lateral hypothalamus, in addition to their previously well characterised functional role in neurosecretion.

Synaptic input to vasopressin neurons of the paraventricular nucleus (PVN)

Following injections of horseradish peroxidase into the PVN, retrogradely filled cells were found in regions of the limbic system known to contain glucocorticoid concentrating neurons [4, 31, 44]. To determine if these regions which include the lateral septum, medial amygdala and ventral subiculum have a monosynaptic input to vasopressin neurons we developed a double label ultrastructural technique [20] to simultaneously visualize immunoreactive neuropeptide and anterogradely transported HRP. Following injections of tracer into all three of these regions, HRP labeled fibers were seen at the light microscopic level to form a halo in the perinuclear, cell poor zone around the PVN. Ultrastructural examination of this area resulted in the discovery of a small number of limbic system synapses on vasopressin dendrites. These synapses were most numerous in the ventral and medial portion of the cell poor zone. A similar pattern of innervation was seen for the supraoptic and suprachiasmatic nucleic which also contain vasopressin cells whose dendrites extend beyond the nuclear boundaries. In a similar fashion we were interested in determining the distribution of noradrenergic terminals on vasopressin neurons in the various subnuclei of the PVN. We have combined immunocytochemistry for vasopressin with radioautography for 3H-norepinephrine (NE) at the ultrastructural level. NE terminals were numerous in the periventricular zone, innervating both vasopressin containing dendrite and non-immunoreactive dendrites and cell bodies. The vasopressin dendrites could originate from cells either resident in the periventricular zone or from cells situated in more lateral subnuclei. In the main, lateral magnocellular region, noradrenergic terminals were very few in number and innervated almost exclusively non-vasopressin containing structures. These studies demonstrate the need for ultrastructural analysis of synaptic input to neurosecretory cells.

Evidence that spinal cord thyrotropin-releasing hormone is independent of the paraventricular nucleus

To determine whether neuronal perikarya containing thyrotropin-releasing hormone (TRH) in the paraventricular nucleus project to the spinal cord, this region of the hypothalamus was ablated by a midline electrolytic lesion and the spinal cord assayed for its content of TRH. In contrast to the marked diminution of TRH in the median eminence, there was no significant change in the TRH content in the spinal cord. These results indicate the absence of significant afferent input of TRH-containing perikarya in the paraventricular nucleus to the spinal cord.

Feeding behavior induced by central norepinephrine injection is attenuated by discrete lesions in the hypothalamic paraventricular nucleus

Extensive brain-cannula mapping studies in the rat have demonstrated that the hypothalamic paraventricular nucleus (PVN) is the most sensitive brain site for eliciting eating behavior with central norepinephrine (NE) injection. The present experiments examined the impact of lesions aimed at the PVN on this NE-elicited eating response. In rats with NE injection cannulas aimed at the lateral ventricle, bilateral lesions of the PVN significantly attenuated, by 60 to 70%, the eating effect induced by NE, at doses ranging from 20 to 160 nmoles. PVN lesions which extended ventrally to damage tissue lying within the periventricular region were more effective in abolishing the NE response than were lesions that remained confined to the dorsal aspects of the PVN. Large lesions located just dorsal to the PVN had no impact on the NE response. This evidence supports the primary role of the PVN in mediating the eating behavior elicited by central noradrenergic activation.

The sensory innervation of the ovary: a horseradish peroxidase study in the rat

Injections of horseradish peroxidase (HRP) into the right or left ovary of the rat produced labeling of perikarya in both nodose ganglia and ipsilateral dorsal root ganglia (DRGs) from T10 to L2. The greatest concentration of labeled cells was in T13 and L1, DRGs. It is suggested that visceral afferent fibers from the ovary may mediate visceral reflexes that modulate ovarian function.

Direct amygdaloid projections to the dorsal motor nucleus of the vagus nerve: a light and electron microscopic study in the rat

The projections from the central nucleus of the amygdala to the dorsal vagal complex were examined in the rat by means of anterograde and retrograde axonal transport of wheat germ agglutinin-horseradish peroxidase and anterograde degeneration. Light microscopic findings confirmed that the amygdala projects to the dorsal motor nucleus (DMV) and the nucleus of the solitary tract. Electron microscopic experiments demonstrated degenerating axosomatic and axodendritic terminals in the DMV following electrolytic lesions in the central nucleus of the amygdala.

ACTH-related peptide containing neurons within the medulla oblongata of the rat

Neurons containing antigenic determinants of 16 Kdalton fragment, ACTH, gamma-LPH, and beta-endorphin have been identified in the nucleus tractus solitarius and lateral reticular formation of the rat brainstem. Immunoreactive fibers extend longitudinally in dorsal and ventral midline tracts throughout the length of the brainstem, and are also concentrated in lateral reticular regions known to contain catecholamine nuclei. ACTH-containing neurons could participate with brainstem catecholamine cell groups in controlling visceral function.

Egg-laying hormone: direct action on the ovotestis of Aplysia

The bag cells are a group of neuroendocrine cells located in the abdominal ganglion of Aplysia californica. These cells induce egg laying in the animal through the release of neurohormone(s). Previous experiments established that an extract of bag cells releases eggs from isolated ovotestis fragments in a dose-dependent manner. (F. E. Dudek and S. S. Tobe, 1978, Gen. Comp. Endocrinol. 36, 618-627.) Experiments presented here were conducted to purify and identify the component(s) of bag cell extract with the egg-releasing activity. Bag cell extracts were fractionated by means of gel-filtration chromatography followed by cation-exchange chromatography. Column eluates were assayed for egg-release activity on isolated ovotestis fragments by the method of Dudek and Tobe (1978). Only one component purified from the crude extract had egg-releasing activity. This component was identified as egg-laying hormone (ELH) based on its purification characteristics, effects on neuronal activity, and migration on thin-layer chromatography. In dose-response studies egg release increased with the concentration of ELH and had a threshold of 8 X 10(-10) M or less. ELH had the same dose-response relationship in egg release assays when present in purified form or as a component of bag cell extract. These data show that ELH acts directly on the ovotestis and that ELH is the only component of bag cell extract with egg-release activity. Taken with the results of other studies (W. D. Branton, S. Arch, T. Smock, and E. Mayeri, 1978. Proc. Nat. Acad. Sci. USA 75, 5732-5736. B. S. Rothman, P. Brownell, and E. Mayeri, 1979. Soc. Neurosci. Abstr. 5, 260. E. Mayeri and B. S. Rothman, 1982.(ABSTRACT TRUNCATED AT 250 WORDS)

Burst discharge in mammalian neuroendocrine cells involves an intrinsic regenerative mechanism

Intracellular recordings from mammalian neuroendocrine cells showed that steady, injected currents can modify and block periodic spike bursts previously associated with increased neurohormone release. Spike afterpotentials could sum to form plateau potentials, which generated bursts and did not depend on axonal conduction or chemical synapses. Therefore, bursting involves a spike-dependent, positive-feedback mechanism endogenous to single neuroendocrine cells.

Selective increase of angiotensin-converting enzyme activity in discrete extrahypothalamic areas of Brattleboro rats

We studied the activity of angiotensin I converting enzyme (ACE, kininase II, E.C. 3.4.15.1) in discrete areas of the brainstem and limbic system, and in circumventricular organs, pineal gland and choroid plexus of homozygous Brattleboro rats (DI) which are characterized by vasopressin deficiency and diabetes insipidus, with or without vasopressin replacement. We also determined ACE activity in heterozygous Brattleboro (HZ) and Long-Evans (LE) control rats. We found changes in ACE activity in several brain areas and the pineal gland of Brattleboro rats. ACE activity was increased in DI rats with respect to HZ and LE controls in the A1 area of the brainstem, locus coeruleus, and triangular nucleus of the septum. ACE activity in the A2 area of the brainstem, the nucleus tractus spinalis nervi trigeminii and the pineal gland was enhanced in both HZ and DI rats with respect to that of LE controls, but was not different between HZ and DI rats. ACE activity did not change in the other extrahypothalamic areas studied. The elevated ACE activity in extrahypothalamic areas of DI rats was not reversed by vasopressin replacement. These results suggest that a relationship between central vasopressin and angiotensin or bradykinin systems may exist in selective extrahypothalamic areas of the rat brain, and that peripherally administered vasopressin cannot influence this relationship.

Rapid and discrete changes in hypothalamic catecholamine nerve terminal systems induced by audiogenic stress, and their modulation by nicotine-relationship to neuroendocrine function

The effects of acute audiogenic stress, with or without simultaneous nicotine treatment (0.3 mg/kg i.v.), on catecholamine levels in discrete dopamine and noradrenaline nerve terminal systems of the hypothalamus, and on the secretion of adenohypophyseal hormones and of corticosterone, were studied using quantitative microfluorometric evaluations of catecholamine stores and radioimmunoassays for the determination of serum hormone levels. Audiogenic stress and nicotine induced very rapid and discrete decreases in noradrenaline levels in the subependymal layer (SEL), in the parvocellular part of nuc. paraventricularis hypothalamic (PA FP) and in the posterior periventricular hypothalamic systems, (PV II); the decreases were apparent 2 min following the onset of treatment. Increases of arterial blood pressure were observed after nicotine treatment but could not have been a major factor in producing the changes in catecholamine levels. These changes in NA levels may be related to the nicotine- and stress-induced increases of ACTH (SEL and PA FP) and prolactin secretion (PV II) found in the present experiments. Stress enhanced the rapid but variable increase in vasopressin secretion induced by nicotine, suggesting one possible mechanism by which stress combined with smoking can contribute to the development of increased arterial blood pressure and finally to sustained hypertension.

Changes in paraventricular vasopressin and oxytocin during the development of spontaneous hypertension

The potential role of central neuroendocrine changes in the development of spontaneous hypertension was evaluated. The developmental changes in blood pressure and hypothalamic and plasma levels of vasopressin (AVP) and oxytocin (OT) were determined in groups of SHR and WKY animals from 3 to 24 weeks of age. Hypothalamic OT content was significantly lower in 3-, 6-, and 12-week-old SHR rats compared to age-matched WKY animals. Hypothalamic AVP content was not different at 3 weeks of age, but was lower in the SHRs at 6 and 12 weeks. To localize strain differences in AVP and OT, specific hypothalamic nuclei were removed from 300 microns frozen brain sections, and hormone content measured. Paraventricular AVP and OT content was lower in the SHRs which had increased blood pressure (6, 12, and 24 weeks of age) but not in the prehypertensive groups (3 weeks of age). Neuropeptide content was unchanged in the supraoptic nucleus or median eminence. Plasma levels of AVP were increased in the SHR, while OT was unchanged. Thus, genetic hypertension is associated with specific and localized changes in hypothalamic AVP and OT. The fact that the peptide deficit occurred in the paraventricular nucleus, a region thought to be involved in the control of autonomic function, may have important implications in terms of the pathogenesis of hypertension.

Corticoliberin-immunoreactive cell bodies localised in two distinct areas of the sheep hypothalamus

The localisation of corticoliberin producing neurones in the sheep hypothalamus was attempted with an antiserum directed against synthetic ovine CRF by the indirect immunofluorescence procedure. Synthetic ovine corticoliberin-immunoreactive fibres were detected, in order of decreasing importance, in the external median eminence, in the caudal neural lobe around capillaries, at the boundary of the neural and intermediate lobe, around the anterior commissure, in the paraventricular nuclei and in the posterior hypothalamus and midbrain, suggesting that synthetic ovine corticoliberin-related substances act not only on anterior pituitary tissue, but also on the intermediate lobe, on central neurones and on peripheral target organs. Two groups of cell bodies reacted to the anti-synthetic ovine corticoliberin antiserum. The first group was located in the paraventricular nuclei and consisted of 15-20 microns diameter cell bodies with a granular cytoplasm. The second group was located mainly in the dorsolateral caudal hypothalamus, and the cell bodies were smaller (10-15 microns) and had a smooth cytoplasm. No cell bodies were detected in the basal hypothalamus). Synthetic ovine corticoliberin-immunoreactive structures did not contain immunoreactive neurophysin. The synthetic ovine corticoliberin-immunoreaction in the paraventricular neurones was abolished by preincubating the antiserum with synthetic ovine corticoliberin but not with sauvagine or several other peptides. The immunoreaction in the posterior hypothalamic groups was abolished by preincubating the synthetic ovine corticoliberin antiserum with both synthetic ovine corticoliberin and sauvagine, but not with other peptides. The results suggest that the immunoreaction was specific for synthetic ovine corticoliberin in the paraventricular but not posterior hypothalamic region. The relative contribution of both areas to synthetic ovine corticoliberin-like peptides containing nerve terminals of the median eminence remains to be established.

Centrally mediated effects of neurohypophyseal hormones

The neurohypophyseal hormones oxytocin and vasopressin cause a variety of biological effects in animals which are mediated by central nervous system mechanisms. Among the best studied of these effects is the modulation of both memory processes and the development of drug tolerance and dependence. Neurohypophyseal hormones have also been shown to alter various physiological parameters such as heart rate and body temperature following central administration. In addition, these peptides can profoundly alter spontaneous, unlearned behavior in several rodent species. Many of the centrally mediated effects of neurohypophyseal hormones have been shown to be elicited at sites within the brain stem and the limbic system where vasopressin and oxytocin occur in cell bodies, axons and nerve terminals, suggesting a physiological role for these peptide effects. The various central effects of neurohypophyseal hormones involve different mechanisms which can be distinguished from one another on the basis of required dose, time-course of action, and structure-activity relationships. Thus, alterations of spontaneous behavior are mediated by putative receptors closely related to vasopressin receptors in blood vessels responsible for the peripheral pressor response while the effects on memory processes are mediated by a mechanism which is not closely related to those involved in the peripheral hormonal effects of the peptides. The influence of neurohypophyseal hormones on memory and attention may be useful clinically. A potential role for these peptides in mental disorders is discussed.

Ultrastructural studies of vasopressin neurons of the paraventricular nucleus of the hypothalamus using a monoclonal antibody to vasopressin: analysis of synaptic input

The ultrastructure of the vasopressin neurons of the paraventricular nucleus of the hypothalamus was studied by immunocytochemical techniques. Tissue antigen was detected in unembedded tissue sections using a monoclonal antibody that recognizes vasopressin but not oxytocin or vasotocin. At the light-microscopic level, reaction product was seen to fill the cytoplasm of the neuron cell body as well as large portions of the dendrite and axon. Immunoreactive spines were seen on both somatic and dendritic surfaces and their presence was confirmed at the ultrastructural level. In the light-microscope, axonal processes do not have spines and are thinner and more varicose than dendritic processes. At the electron-microscopic level, both axons and dendrites of the vasopressin cells are filled with reactive neurosecretory granules. The presence of large numbers of these organelles made it difficult to distinguish proximal dendrites from Herring bodies (axonal swellings). At the ultrastructural level, reaction product was also observed in the cytoplasm of all segments of the vasopressin cells. The presence of reaction product outside of membranous compartments is undoubtably due to disruption of membranes by detergent treatment or exposure to basic pH. However, the staining procedure used did allow us to examine the synaptic input to the vasopressin cells. All portions of the vasopressin neuron receive a diverse innervation. The somata have synapses on their surfaces and on spines. These axo-somatic terminals are primarily, but not exclusively, symmetrical and the presynaptic elements contain spherical or elongate vesicles. On the dendrites, terminals again were observed on the surface or on spines. these axo-dendritic synapses were usually asymmetrical. The presynaptic elements contained clear spherical, elongate or pleomorphic vesicles. Occasional varicosities with dense-core granules were seen to make en passant contacts with dendrites; these contacts did not have obvious membrane specializations. Input to vasopressin axons was studied both along the paraventricular-neurohypophysial tract and in the median eminence. Vasopressin axons receive a synaptic input (axo-axonic), predominately of the asymmetric variety with clear, spherical vesicles in the presynaptic element. These findings demonstrate that the vasopressin neurons of the paraventricular nucleus receive a diverse innervation.