Localisation of arginine vasopressin V1a receptors on sympatho-adrenal preganglionic neurones

Vasopressin-containing nerve terminals are present in the spinal cord of several species. This study was designed to determine whether sympatho-adrenal preganglionic neurones (SPN) express vasopressin receptors (VPRs). SPN in the spinal cord were revealed by retrograde labelling of Fluorogold following its unilateral injection into the adrenal medulla of 12–20 day postnatal rats. VPRs were simultaneously visualised in the Fluorogold-labelled slices of spinal cord using a recently developed biotinylated vasopressin receptor antagonist [1-phenylacetyl,2-O-methyl-D-tyrosine,6-arginine,8-arginine,9-lysinamide(N?-biotinamidocaproamide)]vasopressin, PhAcAL(Btn)VP. The VPR:PhAcAL(Btn)VP complexes were visualised either with Texas Red-conjugated avidin or with a Vectastain avidin:alkaline phosphatase detection kit. These dual-labelling experiments revealed VPRs to be present in the spinal grey matter and to be particularly dense in the intermediate grey matter and adjacent regions of the ventral horn. Many SPN were associated with receptor-specific labelling of PhAcAL(Btn)VP, thereby demonstrating that VPRs are expressed by these neurones. These VPRs were pharmacologically defined as the V_1a subtype. It is concluded that sympatho-adrenal preganglionic neurones express VPRs and that these are of the V_1a subtype. The distribution of VPRs is not, however, restricted to these SPN in the spinal cord.     

Central distributions of the efferent and afferent components of the pharyngeal branches of the vagus and glossopharyngeal nerves: an HRP study in the cat

Summary The central distributions of efferent and afferent components of the pharyngeal branches of the vagus (PH-X) and glossopharyngeal (PH-IX) nerves in the cat were studied by soaking their central cut ends in a horseradish peroxidase (HRP) solution. HRP-labelled PH-X neurones were distributed ipsilaterally in the rostral part of the nucleus ambiguus (NA) and the retrofacial nucleus (RFN); HRP-labelled PH-IX neurones were found in the ipsilateral RFN and the bulbopontine lateral reticular formation (RF). Vagal pharyngeal neurones constituted a large population of brainstem motoneurones. The population of HRP-labelled glossopharyngeal neurones was divided into two components. Indeed, on the basis of their location and somal morphology, the most ventral cells were identified as cranial motoneurones and those scattered in the lateral RF as parasympathetic preganglionic neurones. Application of HRP to the PH-IX nerve resulted also in the labelling of fibres and terminals in the alaminar spinal trigeminal nucleus and the nucleus of the solitary tract (NTS). The afferent fibres entered the lateral medulla with the glossopharyngeal roots, ran dorsomedially, then turned caudally toward the NTS and the caudal part of the alaminar spinal trigeminal motor (V) nucleus. In the NTS, labelled fibres ran mainly along the solitary tract, projecting to terminals in the dorsal and dorsolateral nuclei of the NTS.     

The effect of previous transection on quantitative estimates of the preganglionic neurones projecting in the cervical sympathetic trunk of the guinea-pig and the cat made by retrograde labelling of damaged axons by horseradish peroxidase

The numbers of sympathetic preganglionic neurones in the upper thoracic spinal cord retrogradely labelled after application of horseradish peroxidase to the severed axons of the cervical sympathetic trunk have been determined in guinea-pigs and in cats. Neuronal counts were made on both sides of spinal segments C8 to T10 in control animals and in others in which the enzyme was applied 5–8 days after the left cervical sympathetic trunk had been transected. Labelling was not significantly different between sides of unoperated animals, but after previous cervical trunk section labelling was always less, by from 5 to 100% of that on the control side. The results were not significantly different if the application of horseradish peroxidase was made close to the original lesion or 1 cm more proximal along the nerve trunk. Analysis of the dimensions of labelled cells suggested that the deficit in labelling was greatest for small preganglionic neurones.These findings do not support an earlier report¹⁴ that retrograde labelling from regenerating axons is enhanced at this stage after axotomy, and the possible reasons for the discrepancy in results are discussed. It is suggested that labelling of cell bodies may not be restricted if intact axon sprouts are exposed to horseradish peroxidase. Reduced labelling from axons at the time of a second transection might, at least in part, be due to axon atrophy, emphasizing limitations to labelling when horseradish peroxidase is applied to severed axons of fine diameter.The high numbers and reproducibility of our control data enabled estimates to be made of the segmental distribution of the cell bodies of origin of the axons of the cervical sympathetic trunk in both species.     

Presence of the N-methyl-D-aspartic acid R1 glutamatergic receptor subunit in the lumbosacral spinal cord of male rats

The lumbosacral spinal cord contains neurones that control the lower urogenital and digestive tracts. Spinal neurones respond to activation from the periphery and supraspinal nuclei. Glutamate, acting through a variety of receptors, is an established transmitter of excitatory pathways to the spinal cord. Using immunohistochemical methods, we reveal the presence of the N-methyl-D-aspartic acid R1 (NMDAR1) glutamatergic receptor subunit in the lumbosacral spinal network that controls urogenital and digestive functions: the dorsal horn; the area around the central canal including the dorsal grey commissure; the sacral parasympathetic nucleus; and pudendal motoneurones. A complete thoracic spinal section did not alter labelling. Using retrograde labelling techniques, we identify sacral preganglionic neurones and pudendal neurones that are NMDAR1 immunoreactive. Glutamate, acting at NMDA receptors, can therefore co-ordinate the activity of the autonomic and somatic outflows to the pelvic organs.     

Projections from the diencephalon and mesencephalon to nucleus paragigantocellularis lateralis in the cat

The distribution of labelled cells in the diencephalon and mesencephalon has been mapped following injections of horseradish peroxidase into nucleus paragigantocellularis lateralis in the cat. Most of the labelled cells were found ipsilateral to the injection site. A group of small and medium-sized labelled perikarya (11-40 microns in diameter) was present in the caudal part of the periaqueductal grey matter (A3-P2) and in the adjacent tegmentum. Small, round or fusiform cells (8-25 microns were labelled in the tuberal region of the hypothalamus in the dorsomedial hypothalamus and in the lateral hypothalamic area. It is suggested that the cardiovascular responses which can be elicited by stimulation in these regions of the periaqueductal grey and hypothalamus are mediated via a relay on to spinally projecting neurones in nucleus paragigantocellularis lateralis which synapse on sympathetic preganglionic neurones in the intermediolateral cell column.     

Properties of interneurones in the intermediolateral cell column of the rat spinal cord: role of the potassium channel subunit Kv3.1

Sympathetic preganglionic neurones located in the intermediolateral cell column (IML) are subject to inputs descending from higher brain regions, as well as strong influences from local interneurones. Since interneurones in the IML have been rarely studied directly we examined their electrophysiological and anatomical properties. Whole cell patch clamp recordings were made from neurones in the IML of 250 microM slices of the thoracic spinal cord of the rat at room temperature. Action potential durations of interneurones (4.2+/-0.1 ms) were strikingly shorter than those of sympathetic preganglionic neurones (9.4+/-0.2 ms) due to a more rapid repolarisation phase. Low concentrations of tetraethylammonium chloride (TEA) (0.5 mM) or 4-aminopyridine (4-AP) (30 microM) affected interneurones but not sympathetic preganglionic neurones by prolonging the action potential repolarisation as well as decreasing both the afterhypolarisation amplitude and firing frequency. Following recordings, neurones sensitive to TEA and 4-AP were confirmed histologically as interneurones with axons that ramified extensively in the spinal cord, including the IML and other autonomic regions. In contrast, all cells that were insensitive to TEA and 4-AP were confirmed as sympathetic preganglionic neurones. Both electrophysiological and morphological data are therefore consistent with the presence of the voltage-gated potassium channel subunit Kv3.1 in interneurones, but not sympathetic preganglionic neurones. Testing this proposal immunohistochemically revealed that Kv3.1b was localised in low numbers of neurones within the IML but in higher numbers of neurones on the periphery of the IML. Kv3.1b-expressing neurones were not sympathetic preganglionic neurones since they were not retrogradely labelled following intraperitoneal injections of Fluorogold. Since Kv3.2 plays a similar role to Kv3.1 we also tested for the presence of Kv3.2 using immunohistochemistry, but failed to detect it in neuronal somata in the spinal cord. These studies provide electrophysiological and morphological data on interneurones in the IML and indicate that the channels containing the Kv3.1 subunit are important in setting the firing pattern of these neurones.     

D1-like dopamine receptors on retrogradely labelled sympathoadrenal neurones in the thoracic spinal cord of the rat

Cellular localization of dopamine D₁-like receptors was accomplished on target-specified sympathoadrenal preganglionic neurones using the radioligand [³H]SCH23390. Sympathoadrenal neurones were retrogradely labelled with cholera B subunit conjugated to horseradish peroxidase and were detected in segments T₁ to T₁₃ with a predominance at T₈/T₉. Binding of the selective D₁-like radioligand [³H]SCH23390 was associated with the retrogradely labelled sympathoadrenal neurones in longitudinal/horizontal sections of thoracic spinal cord. D₁-like receptor localization on target-specific neurones was determined in more than half of the spinal cord sections and was associated predominantly with the cell soma and principal proximal dendrites in the intermediolateral cell column of the spinal grey matter. D₂-like receptor localization was not associated with retrogradely labelled sympathoadrenal neurones but a higher degree of specific binding was noted in more medial aspects of the spinal grey matter. This is the first successful demonstration of receptor localization combining two quite different techniques and provides conclusive anatomical evidence for D₁-like receptor localization on sympathetic preganglionic neurones that project to the adrenal medulla. Received: 27 November 1998 / Accepted: 6 May 1999     

Age-related loss of cardiac vagal preganglionic neurones in spontaneously hypertensive rats

Despite the findings that impaired vagal control of the heart rate occurs in human hypertension, leading to greater cardiovascular risk, the mechanism of this impairment is as yet unknown. Observations in humans and experiments in the spontaneously hypertensive rat (SHR) suggested that such impairment may be related to an anomaly in central vagal neurones. We therefore set out to determine whether the numbers and distribution of cardiac-projecting vagal preganglionic neurones in the medulla of adult (12 week) hypertensive SHR are different from those in young (4 week) prehypertensive SHR and in age-matched Wistar-Kyoto (WKY) rats of two age groups. The number of vagal neurones, identified by labelling with the fluorescent tracer DiI applied to the heart, was essentially similar in the three areas of the medulla analysed (dorsal vagal nucleus, nucleus ambiguus and intermediate reticular zone) in young SHR and young or adult WKY rats. In contrast, fewer vagal neurones were labelled in adult SHR compared with young SHR or WKY rats. This difference was due to highly significant reductions in vagal neurones in the dorsal vagal nucleus and nucleus ambiguus on the right side of the medulla. These observations suggest that a loss of parasympathetic preganglionic neurones supplying the heart with axons in the right vagus nerve, or a remodelling of their cardiac projections, may explain the known impairment of the baroreceptor reflex gain controlling heart rate in hypertension.     

The relationship between serotonin, dopamine beta hydroxylase and GABA immunoreactive inputs and spinal preganglionic neurones projecting to the major pelvic ganglion of Wistar rats

Preganglionic neurones in the lumbosacral spinal cord give rise to nerves providing the parasympathetic and sympathetic innervation of pelvic organs. These neurones are modulated by neurotransmitters released both from descending supra-spinal pathways and spinal interneurones. Though serotonin has been identified as exerting a significant influence on these neurones, few studies have investigated the circuitry through which it achieves this particularly in relation to sympathetic preganglionic neurones. Using a combination of neuronal tracing and multiple immunolabeling procedures, the current study has shown that pelvic preganglionic neurones receive a sparse, and probably non-synaptic, axosomatic/proximal dendritic input from serotonin-immunoreactive terminals. This was in marked contrast to dopamine beta hydroxylase-immunoreactive terminals, which made multiple contacts. However, the demonstration of both serotonin, and dopamine beta hydroxylase immunoreactive terminals on both parasympathetic and sympathetic preganglionic neurones provides evidence for direct modulation of these cells by both serotonin and norepinephrine. Serotonin-containing terminals displaying conventional synaptic morphology were often seen to contact unlabeled somata and dendritic processes in regions surrounding the labeled preganglionic cells. It is possible that these unlabeled structures represent interneurones that might allow the serotonin containing axons to exert an indirect influence on pelvic preganglionic neurones. Since many spinal interneurones employ GABA as a primary fast acting neurotransmitter we examined the relationship between terminals that were immunoreactive for serotonin or GABA and labeled pelvic preganglionic neurones. These studies were unable to demonstrate any direct connections between serotonin and GABA terminals within the intermediolateral or sacral parasympathetic nuclei. Colocalization of serotonin and GABA was very rare but terminals immunoreactive for each were occasionally seen to contact the same unlabeled processes in close proximity. These results suggest that in the rat, the serotonin modulation of pelvic preganglionic neurones may primarily involve indirect connections via local interneurones.     

Long-lasting discharge of postganglionic neurones to skin and muscle of the cat's hindlimb after repetitive activation of preganglionic axons in the lumbar sympathetic trunk

1. Postganglionic neurones to hairy skin and to skeletal muscle of the cat's hindlimb, probably vasoconstrictor in function, have been investigated for their response characteristics on repetitive stimulation of the preganglionic axons in the lumbar sympathetic trunk in chloralose anaesthetized cats.
2. Repetitive stimulation of the preganglionic axons with 50 stimuli at 25 to 50 Hz and stimulus strengths which excited all classes of preganglionic axons generated eartly short latency responses and late low frequency responses in postganglionic neurones. These late responses occurred predominantly in postganglionic neurones supplying skeletal muscle and in only 30% of postganglionic neurones supplying hairy skin. In most neurones the late discharges could only be elicited by trains of more than 5–10 preganglionic stimuli at frequencies of more than 5–10 Hz.
3. The threshold stimulus strength applied to the preganglionic axons to generate the late discharge was on average 4 times higher than for the early discharge. This finding implies that the late discharge requires the activation of thin preganglionic axons.
4. In most postganglionic neurones, the late discharges could be blocked by atropine, indicating that they were generated by muscarinic action of acetylcholine. The late discharges in a few cutaneous postganglionic neurones were unaffected by atropine. The early discharges were blocked by hexamethonium.
5. The results argue that the transmission of impulses from pre- to postganglionic neurones may be accomplished by two separate pathways acting via nicotinic and muscarinic receptors respectively.

     

Reflex activation of postganglionic vasoconstrictor neurones supplying skeletal muscle by stimulation of arterial chemoreceptors via non-nicotinic synaptic mechanisms in sympathetic ganglia

Postganglionic sympathetic neurones supplying skeletal muscle and skin can be activated from the preganglionic site via cholinergic nicotinic, muscarinic and noncholinergic synaptic mechanisms. The experiments described in this paper were designed in order to show that postganglionic vasoconstrictor neurones supplying skeletal muscle can be activated by the naturally occurring discharge pattern in preganglionic axons when the nicotinic transmission is blocked. For this purpose, the activity was recorded simultaneously from postganglionic vasoconstrictor axons supplying skeletal muscle and vasoconstrictor axons supplying hairy skin. The preganglionic neurones were driven reflexly by stimulation of the arterial chemoreceptors. 1) During blockade of nicotinic transmission muscle vasoconstrictor neurones were activated via the CNS during stimulation of arterial chemoreceptors. This activation is either generated by muscarinic action of released acetylcholine or by a noncholinergic synaptic mechanism. 2) Postganglionic cutaneous vasoconstrictor neurones were inhibited during stimulation of arterial chemoreceptors. During blockade of cholinergic nicotinic transmission these neurones were not activated reflexly by stimulation of arterial chemoreceptors although they received inputs via cholinergic muscarinic and noncholinergic synaptic mechanisms. 3) The results illustrate that postganglionic vasoconstrictor neurones supplying skeletal muscle can not only be activated via non-nicotinic synaptic mechanisms through synchronous repetitive electrical stimulation of preganglionic axons but also by the discharge pattern produced in preganglionic neurones during stimulation of arterial chemoreceptors.     

Projections of the mediolateral part of the lateral septum to the hypothalamus, revealed by Fos expression and axonal tracing in rats

 The lateral septum participates in a variety of functions involving the hypothalamus. The present study investigated the effect of an electrical stimulation of the mediolateral part of the lateral septum on the expression of Fos in the hypothalamic nuclei by using immunohistochemical methods in anaesthetised and free-moving rats. We analysed in another series of rats the direct projections of the lateral septum by axonal anterograde tracing with biotinylated dextran-amine. Tracing was used in combination with Fos labelling in a third series of animals. Stimulation induced an expression of Fos in neurones located in anteroventral and anterodorsal preoptic nuclei, medial preoptic area, anterior hypothalamic nucleus, subparaventricular zone, dorsomedial nucleus, lateral hypothalamic area and mammillary nucleus. The distribution of Fos-immunoreactive neurones conforms to the topographic organisation of direct projections from the lateral septum, as revealed by axonal tracing. These results suggest that the lateral septum activates definite hypothalamic structures by a direct link. Some structures displayed substantial Fos labelling whereas they received a slight, or no projection, from the lateral septum. This was particularly evident in the core of the ventromedial nucleus and in areas known to contain tubero-infundibular neurones. This observation suggests that the lateral septum may also exert an indirect control, via polysynaptic links, on hypothalamic structures including nuclei involved in neuroendocrine mechanisms. Accepted: 24 August 1998     

Mechanisms underlying recurrent inhibition in the sacral parasympathetic outflow to the urinary bladder.

1. In cats with the sacral dorsal roots cut on one side electrical stimulation (15-40 c/s) of the central end of the transected ipsilateral pelvic nerve depressed spontaneous bladder contractions. The depression was abolished by transecting the ipsilateral sacral ventral roots. 2. Electrical stimulation of acutely or chronically transected ('deafferented') sacral ventral roots depressed spontaneous bladder contractions and the firing of sacral parasympathetic preganglionic neurones innervating the bladder. The depression of neuronal firing occurred ipsilateral and contralateral to the point of stimulation, but only occurred with stimulation of sacral roots containing preganglionic axons and only with stimulation of sacral roots containing preganglionic axons and only at intensities of stimulation (0-7-4V) above the threshold for activation of these axons. 3. The inhibitory responses were not abolished by strychnine administered by micro-electrophoresis to preganglionic neurones, but were blocked by the intravenous administration of strychnine. 4. The firing of preganglionic neurones elicited by micro-electrophoretic administration of an excitant amino acid (DL-homocysteic acid) was not depressed by stimulation of the ventral roots. 5. It is concluded that the inhibition of the sacral outflow to the bladder by stimulation of sacral ventral roots is related to antidromic activation of vesical preganglionic axons. Collaterals of these axons must excite inhibitory interneurones which in turn depress transmission at a site on the micturition reflex pathway prior to the preganglionic neurones.     

Location of bladder preganglionic neurons within the sacral parasympathetic nucleus of the cat.

Sacral parasympathetic preganglionic neurons innervating the urinary bladder were labelled by applying horseradish peroxidase (HRP) to branches of the pelvic nerve at the neck of the bladder. Bladder preganglionic cells were located primarily in the intermediolateral grey matter of sacral cord segments 1, 2 and 3 and extended from the ventral end of the central canal to the inferior margin of the dorsal horn. Dendrites extended into the dorsolateral funiculus, toward the lateral edge of the dorsal horn, medially beneath the dorsal horn and ventrally within the nucleus. Although a small number of cells were found medially beneath the dorsal horn, it is concluded that the majority of bladder preganglionic neurons occupy a lateral position within the sacral parasympathetic nucleus.     

Preganglionic sympathetic neurones, innervating the guinea pig adrenal medulla, immunohistochemically contain choline acetyltransferase and also leu-enkephalin

Applying retrograde neuronal tracing combined with double labelling immunofluorescence, preganglionic nerve cell bodies in the intermediate grey matter of the guinea pig thoracic spinal cord, projecting to the adrenal gland, co-exhibited immunolabelling for choline-acetyltransferase (ChAT) and sometimes, also for leu-enkephalin. Likewise, ChAT-immunoreactive nerve fibers, forming a dense meshwork in the adrenal medulla, partly contained immunostaining also for leu-enkephalin. Some of the intramedullary nerve cell bodies were ChAT-positive but were non-reactive for leu-enkephalin. The findings provide evidence for an extrinsic (preganglionic) and an intrinsic (postganglionic) cholinergic nerve system in the guinea pig adrenal medulla, the preganglionic syste, utilising leu-enkephalin as co-mediator.     

Reticulo-collicular projections: a neuronal tracing study in the rat

Neuroanatomical tract-tracing methods were used to study the topography of the reticulocollicular projections. Injections of gold-HRP or BDA tracers into the medial and/or central portions of the superior colliculus resulted in labelled neurones mainly in the medial reticular formation, whereas injections into the lateral portion of the superior colliculus showed labelling in the medial and lateral reticular formation. When tracer was injected into the lateral portion of the caudal superior colliculus, extensive lateral labelling was observed in the contralateral parvocellular reticular nucleus and the contralateral dorsal medullary reticular nucleus, two areas involved in reflex blinking. The present study shows that these reticular areas project to the lateral superior colliculus, which is known to be involved in the coordination of eye and eyelid movements.     

Vasoconstrictor and pilomotor fibres in skin nerves to the cat's tail

Summary Postganglionic neurones to the tail's skin of the cat were investigated with regard to their spontaneous activity, response characteristics to somatic stimuli and asphyxia, the conduction velocity of their axons, and the conduction velocity of the preganglionic axons converging on them. The cats were anaesthetized with chloralose, immobilized, and arteficially ventilated. With this regimen the postganglionic neurones were divided into two types: 1. Type 1 neurones are spontaneously active and exhibit reflexes upon somatic stimulation. During asphyxia they are mostly first depressed and then excited for about 2–3 min. Their axons conduct with 0.57±0.13 m/s (mean ± SD). The preganglionic axons converging on them conduct with 5.4±1.6 m/s. 2. Type 2 neurones are not spontaneously active and exhibit with few exceptions no reflexes on somatic stimuli. During asphyxia they are activated after 3–4 min, concomitantly with piloerection, when the activity in type 1 neurones is already decreasing. Their axons conduct with 0.84±0.14 m/s, the preganglionic axons converging on them conduct with 9.9±2.9 m/s. 3. From these characteristics it is concluded that type 1 neurones have vasomotor function and most type 2 neurones pilomotor function.     

Specific innervation of guinea-pig superior cervical ganglion cells by preganglionic fibres arising from different levels of the spinal cord.

1. The synaptic contribution of preganglionic nerve fibres arising from the last cervical (C8) and the first seven thoracic spinal cord segments (T1-T7) to neurones of the guinea-pig superior cervical ganglion has been studied by means of intracellular recording during ventral root stimulation in vitro. 2. The majority of neurones received innervation from the middle segments (T2 and T3) of the length of spinal cord from which preganglionic fibres derive; an intermediate number of ganglion cells were innervated by fibres from the segments adjacent to these (T1, T4, and T5), and relatively few neurones by fibres from the most rostral and caudal segments supplying innervation to the ganglion (C8, T6 and T7). 3. Each neurone received preganglionic terminals from multiple thoracic segments (range 1-7, mean = 4-0). The estimated minimum number of preganglionic fibres contacting each neurone was 10, on average. 4. As a rule, the spinal segments innervating a neurone were contiguous. Thus we rarely encountered neurones innervated by segments located both rostrally and caudally to a segment which failed to provide innervation. 5. Neurones tended to be innervated predominantly by axons arising from a single spinal segment, with adjacent segments contributing a synaptic influence that diminished as a function of their distance from the dominant segment. All segments provided dominant innervation to at least some neurones. 6. Stimulating the ventral roots of C8-T7 in vivo showed that the axons arising from each segment produced a characteristic pattern of peripheral effects. Thus different populations of neurones in the superior cervical ganglion of the guinea-pig are innervated by preganglionic axons from different levels of the spinal cord, as originally suggested by Langley (1892) for the cat, dog, and rabbit. 7. On the basis of our in vitro studies we conclude that underlying the specificity of innervation of neurones of the superior cervical ganglion that can be inferred from in vivo experiments is a tendency for individual neurones to be innervated in a systematically graded fashion by a contiguous subset of the eight spinal segments which provide innervation to the ganglion.     

Effects of substance P on sympathetic preganglionic neurones

Antidromically identified sympathetic preganglionic neurones, located in the second thoracic segment of the rat spinal cord, were tested for their response to iontophoretically applied substance P. Substance P increased the firing rate of both 'spontaneously' active and glutamate-activated neurones. As substance P-like immunoreactivity is present in the intermediolateral cell column of the thoracic spinal cord, it is concluded that substance P may be an excitatory transmitter or modulator involved in mediating excitatory drive to sympathetic preganglionic neurones.     

Localization of neurones projecting to the hypothalamic paraventricular nucleus area of the rat: a horseradish peroxidase study

To detect the cell bodies of neurones which project to the area of the hypothalamic para-ventricular nucleus, 10–40 nl of a solution containing horseradish peroxidase and poly-l-ornithine were pressure-injected into one paraventricular nucleus of the rat. After 24 or 48 h, the enzyme remaining at the site of injection was detected by the diaminobenzidine procedure. Retrogradely transported horse-radish peroxidase was visualized by using o-dianisidine as the chromogen substrate.The extent and the intensity of labelling correlated with the apparent volume of the injection site. Labelled cell bodies were observed, ipsilateral to the injection, in the mediobasal hypothalamus, in the limbic system (lateral septum, posteromedial amygdala, ventral subiculum) and in several cell clusters in the brain stem (dorsal raphe nucleus, locus coeruleus, parabrachial nucleus, nucleus of the solitary tract and lateral reticular nucleus). In some animals, light labelling in the organum vasculosum laminae terminalis and in the subfornical organ was observed. No labelled neurones could be detected in the spinal cord.