INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS ELECTROPHYSIOLOGY EXPERIMENTS RESULTS DISCUSSION REFERENCES

ARGININE VASOPRESSIN is a peptide hormone produced by magnocellular neurosecretory cells in the supraoptic (SON) and paraventricular nuclei and secreted into the blood stream from the posterior pituitary (7). On release vasopressin acts as a potent antidiuretic and vasoconstrictor (7) while augmenting baroreflex inhibition of sympathetic nerve activity (15). The ability of this hormone to alter blood pressure and blood volume is indicative of the important role of vasopressin in body fluid homeostasis. Vasopressin secretion is regulated by changes in blood pressure, volume, and osmolality (37). Decreases in blood pressure, decreases in blood volume, and increases in osmolality increase the activity of vasopressin-secreting neurons. Conversely, increases in blood pressure, increases in blood volume, and decreases in osmolality decrease the activity of vasopressin-secreting neurons (37). It has been demonstrated that changes in the firing frequency and firing patterns of magnocellular neurons regulate the amount of vasopressin that they secrete (37). The central nervous system pathways responsible for regulating the activity of vasopressin-secreting neurons in response to various stimuli are still being investigated.

Changes in blood volume and blood pressure that influence the activity of vasopressin neurons are detected by two different populations of peripheral baroreceptors. Arterial baroreceptors or high pressure baroreceptors are located in the aortic arch and carotid sinus. An increase in blood pressure will stretch these receptors, increasing the activity of the cells, and ultimately decrease vasopressin release (9-11, 34, 37). Low-pressure baroreceptors, or cardiopulmonary receptors, are sensitive to changes in blood volume and are found in the atria, lungs, great veins, and ventricles. An increase in volume will increase the activity of cardiopulmonary receptors, decreasing vasopressin secretion (11, 14, 29, 32). To date, few experiments have examined the mechanism by which a select volume stimulus inhibits vasopressin neurons.

Recent experiments from this laboratory have investigated volume regulation of supraoptic neurons. To selectively activate cardiopulmonary receptors without stimulating arterial baroreceptors, a small latex balloon was inflated at the junction of the superior vena cava and right atrium. This technique increases release of atrial natriuretic peptide (27), elicits diuresis and natriuresis (27, 39), decreases plasma renin activity (39), and inhibits sympathetic nerve activity (28) without altering blood pressure (14, 26). In the rat, caval-atrial stretch decreases the activity of most (86%) but not all vasopressinergic neurons (14). This is in contrast to the sensitivity of vasopressinergic neurons to an increase in blood pressure; almost all of the vasopressin-secreting cells in the SON are inhibited by an increase in blood pressure (9, 10, 13, 21, 24, 34, 37). Whether there is overlap between the neural pathways that are activated by arterial baroreceptor stimulation and caval-atrial stretch to ultimately decrease the activity of supraoptic vasopressin neurons remains to be determined.

In the rat a series of experiments have been performed to identify the central networks that inhibit vasopressin neurons following an increase in blood pressure (9, 10, 13, 21, 23, 24, 34, 37). A region of the lateral hypothalamus that surrounds the SON and is known to participate in this inhibitory pathway is the perinuclear zone (PNZ) of the SON. The PNZ is a GABAergic region that projects to the SON (2, 23, 38) as well as the paraventricular nucleus (38). Previous studies have shown that a GABA_A antagonist applied to the SON reversibly blocks arterial baroreceptor-mediated inhibition of vasopressin neurons (24). Excitotoxic lesions of the PNZ significantly decrease the number of SON vasopressin neurons that are inhibited by arterial baroreceptor activation (10, 34). An isotonic volume expansion that significantly increases central venous pressure without increasing mean arterial pressure increases Fos expression in the PNZ (8, 36). Fos is the protein product of the immediate early gene c-fos and is commonly used as an indication of synaptic activation (12). Together, these data suggest that the PNZ may be a necessary component of the neural pathways responsible for both arterial and cardiopulmonary baroreceptor regulation of vasopressin release.

The purpose of the present study was to determine whether lesions of the PNZ interrupt the pathway activated by caval-atrial stretch that ultimately decreases the activity of vasopressinergic neurons in the SON. Rats received unilateral injections of ibotenic acid into the PNZ. Ibotenic acid is an excitotoxic glutamate analog that acts at both N-methyl-D-aspartate and metabotropic receptors (18). Because magnocellular neurons of the SON are resistant to ibotenic acid (16, 18), the vasopressin-secreting neurons are not damaged despite their proximity to the PNZ. The present study used in vivo electrophysiology techniques to obtain extracellular recordings from putative vasopressinergic neurons of the SON. The effect of an increase in blood pressure and stretch of the caval-atrial junction on the firing activity of those neurons was measured and the results were compared with cells recorded from vehicle-injected rats. We hypothesized that lesion of the PNZ would attenuate both the arterial baroreceptor and the cardiopulmonary baroreceptor sensitivity of supraoptic vasopressinergic neurons.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS ELECTROPHYSIOLOGY EXPERIMENTS RESULTS DISCUSSION REFERENCES

Nineteen adult male Sprague-Dawley rats (Harlan, Indianapolis, IN) weighing 225-300 g were housed in a temperature- and humidity-controlled room that was maintained on a 12:12-h light-dark cycle with food and water available ad libitum. Experiments were conducted in accordance with the guidelines of the American Physiological Society and the Society for Neuroscience. The University of Missouri Institutional Animal Care and Use Committee approved all protocols.

Neurotoxic lesions. Rats were anesthetized with pentobarbital sodium (50 mg/kg ip). Animals were placed in a stereotaxic frame and their skulls leveled between bregma and lambda. Rats were divided into two groups: lesion and vehicle control. Lesion animals received three unilateral injections of ibotenic acid (250 nl, 5 µg/µl; Sigma, St. Louis, MO) into the right PNZ, whereas control rats were injected with the same volume of PBS vehicle. Previous studies indicate that ibotenic acid produces extensive neuronal loss in the lateral hypothalamus (10, 16, 34). Because most afferent projections from the PNZ to the SON are ipsilateral (2, 23), all lesions were unilateral. Injections were delivered via a 30-gauge injector connected to a 5-µl syringe. The injector was oriented at a 10° angle to prevent leakage of ibotenic acid into the ventricular system. The coordinates used for the three injections were 1.1 mm caudal to bregma, 3.4 mm lateral to bregma, and 8.5 mm ventral to bregma; 0.9 mm caudal to bregma, 3.2 mm lateral to bregma, and 8.5 mm below bregma; and 0.7 mm caudal to bregma, 3.0 mm lateral to bregma, and 8.5 mm below bregma (35). After each injection the injector was left in the tissue for 5 min and then slowly removed from the tissue over a second 5-min time period to minimize leakage up the cannula tract.

	ELECTROPHYSIOLOGY EXPERIMENTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS ELECTROPHYSIOLOGY EXPERIMENTS RESULTS DISCUSSION REFERENCES

Male rats that were previously injected with ibotenic acid or PBS vehicle were anesthetized with pentobarbital sodium (50 mg/kg ip). Supplemental anesthetic sufficient to suppress withdrawal reflexes (2-5 mg pentobarbital sodium as needed) was delivered via a catheter (PE-50, A-M Systems) placed in the right femoral vein. Bolus injections of phenylephrine were delivered via a catheter placed in the left femoral vein. Blood pressure was recorded via a catheter in the left femoral artery using a pressure transducer connected to preamplifiers (Dalton Cardiovascular Research Center Electronics Core) and a Pentium computer running Spike2 data acquisition software (Cambridge Electronic Design). Balloons were made from fine latex spread over the tip of PE-10 tubing. The latex was sealed around the PE-10 tubing with a short piece of Silastic tubing. Balloons were inflated using a saline-filled syringe attached to a three-way stopcock (14). Before a balloon was inserted into the jugular vein, the volume of saline necessary to inflate the balloon to a diameter of 2.5-3.0 mm was determined and this volume subsequently was used to inflate the balloon during each test. Balloons were inserted in the right jugular vein and advanced to the junction of the superior vena cava and right atrium. Inflation of a small latex balloon at the caval-atrial junction activates cardiac receptors without altering a rat's venous return to the heart (26). This is because the left jugular vein, which is unobstructed, is connected to the right jugular vein by collateral vessels and drains into the inferior vena cava rather than the superior vena cava (26).

After catheterization, each rat was tracheotomized and the ventral surface of the brain was exposed using a transpharyngeal surgical approach (9, 10, 13, 14, 34). Once the pituitary and SON were exposed, a pressure foot was placed on the ventral surface of the hypothalamus to stabilize the nucleus. A bipolar-stimulating electrode placed in the posterior pituitary was used to antidromically activate neurosecretory neurons in the SON. Neurons were identified as supraoptic magnocellular neurosecretory cells if they were antidromically activated from posterior pituitary stimulation, showing a constant latency at threshold of activation and collision cancellation with spontaneously occurring action potentials (13, 14). Single 1-Hz suprathreshold current pulses of 1-ms duration with an intensity 10 mA were applied via the bipolar stimulating electrode discharged to identify single units. Once a unit was identified the current intensity was reduced to determine the threshold for antidromic activation. Extracellular action potentials from SON neurons were recorded using glass micropipettes (WPI) filled with 3 M sodium chloride (resistance 15-45 M). Signals were amplified (DAGAN 2400 extracellular Preamplifier, Minneapolis, MN), low-pass- and high-pass-filtered, and relayed through a window discriminator to an analog-to-digital converter (CED 1401, Cambridge Electronic Design). The signal for the window discriminator and the raw signal from the preamplifier were both sent to a Pentium computer running Spike2 data acquisition software.

Antidromically identified SON neurons ipsilateral to the stereotaxic injections with phasic firing patterns were included in this study. Supraoptic neurons with phasic patterns of activity are vasopressinergic, whereas neurons with continuous firing patterns may be oxytocinergic or vasopressinergic and must therefore be classified based on their sensitivity to an increase in blood pressure (37). Because lesions of the PNZ have been shown to alter the blood pressure sensitivity of vasopressin-secreting neurons (10, 34), supraoptic neurons with continuous firing patterns could not be identified as oxytocinergic or vasopressinergic. Oxytocin-secreting cells in the SON are insensitive to increases in blood pressure (37) and caval-atrial stretch (14). Antidromically activated continuous neurons were recorded in this study but only analyzed by their sensitivity to increases in blood pressure and caval-atrial stretch.

Supraoptic neurons were considered sensitive to an increase in blood pressure if they stopped firing within 20 s of an increase in blood pressure of at least 40 mmHg (phenylephrine, 5 µg/µl). In phasic neurons, bolus injections of phenylephrine were delivered 5-15 s following the initiation of a spontaneous phase of activity. Previous studies demonstrate that this increase in blood pressure inhibits 95-100% of phasic neurons in the SON (9, 10, 13, 14, 34). Cells were tested for baroreceptor sensitivity no more than five times. Each supraoptic neuron was also tested for sensitivity to caval-atrial stretch. Ratemeter records were used to compare the spontaneous firing rate (in spikes/s) before, during, and after the stimulus. For phasic neurons, baseline readings were taken from a comparable point in the phase previous to the phase that was stimulated, and recovery readings were taken from a comparable point in the phase following the stimulated phase. A neuron was classified as sensitive to balloon inflation if there was a 30% change in firing activity from baseline during balloon inflation. Cells were tested for balloon sensitivity no more than four times. Because neurons that were tested multiple times exhibited similar changes in firing activity to each stretch of the caval-atrial junction, responses were averaged for a single neuron. The order of blood pressure and balloon tests was varied between cells.

On completion of the experiment, rats were given an overdose of pentobarbital and their chests were opened for visual verification of balloon placement (14). Each balloon was inflated at least three times with the same volume of saline used during the experiment to determine placement of the balloon. The inflation had to produce an obvious and unmistakable distension of the caval-atrial junction for the animal to be included in the study. Animals whose balloons were outside of the caval-atrial junction were excluded from further analysis.

Histology. After determination of balloon placement, animals were perfused transcardially with PBS followed by 4% paraformaldehyde. Their brains were subsequently removed and placed in a 30% sucrose PBS solution to cryoprotect the tissue. After sectioning of the brains in a cryostat at 40 µm, sections were mounted on gelatin-coated glass slides and stained for nissl with Giemsa (Merck, Rathaway, NJ) to determine the extent of the lesions. The borders of the lesion were easily visible under high magnification. Lesioned areas were differentiated by neuronal loss and glial infiltration (11, 18, 34). The extent of each lesion was assessed using light microscopy and recorded (Olympus, Melville, NY). Only animals with lesions of the region of the hypothalamus previously described as the PNZ (10, 23, 34) were included in the study as PNZ-lesioned rats. All of these lesions were functional PNZ lesions; phasic neurons recorded from each rat were insensitive to increases in blood pressure and caval-atrial stretch. Brains from the vehicle injected control rats were also analyzed to determine whether those injections were made in the region of the PNZ.

Statistics. All statistical tests were performed using commercially available software (SigmaStat 2.03, Jandel Scientific). The effect of PNZ lesion on the sensitivity of phasic vasopressin neurons to increases in blood pressure and caval-atrial stretch was evaluated with paired 2 × 2 Fisher Exact tests. The P value for statistical significance for these tests was adjusted for the number of tests using a Bonferroni correction (30). Significance was defined when P < 0.05. All other statistical analysis was performed using one-way ANOVA with Student-Newman-Keuls follow-up tests.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS ELECTROPHYSIOLOGY EXPERIMENTS RESULTS DISCUSSION REFERENCES

Electrophysiology. Stable extracellular recordings were obtained from a total of 107 antidromically identified supraoptic neurons, 55 of which displayed the phasic firing pattern characteristic of vasopressin neurons (37). As many as eight phasic cells were recorded from an individual animal. Latencies of antidromic activation ranged from 5 to 22 ms with a mean latency of 16.1 ±0.4 ms, and thresholds for antidromic activation ranged from 0.6 to 7.5 mA with a mean antidromic activation of 3.4 ±0.5 mA. Phasic vasopressin cells were recorded from the SON of nine rats with PNZ lesions and balloon placement at the caval-atrial junction and 10 rats with PBS vehicle injections in the PNZ and balloon placement at the caval-atrial junction. The baseline blood pressures of the different groups were not statistically different from one another (vehicle control animals 92.4 ± 2.2 mmHg, lesion animals 90.9 ± 3.0 mmHg, P > 0.05).

Phasic neurons. Lesion of the PNZ significantly attenuated the arterial baroreceptor and caval-atrial stretch sensitivities of phasic vasopressinergic cells of the SON (P < 0.05). Phasic vasopressin neurons insensitive to both caval-atrial stretch and an increase in blood pressure were recorded from all PNZ-lesion rats. Of the 26 phasic neurons recorded from vehicle animals, all 26 (100%) were sensitive to an increase in blood pressure (Table 1). In contrast, only 12 of the 29 phasically active cells (41%) were inhibited by equivalent blood pressure increases in experiments with PNZ-lesion rats (Fig. 1, Table 1). There was no difference in the mean increase in blood pressure following phenylephrine infusion between groups (vehicle animals 80.6 ± 3.9 mmHg, lesion animals 72.8 ± 3.0 mmHg, P > 0.05).

View larger version (28K): [in this window] [in a new window]   Fig. 1.   Ratemeter records (top) and corresponding blood pressure (BP) tracing (bottom) for 2 vasopressin neurons with phasic-firing patterns. A: neuron was recorded from a vehicle-control animal. Bolus intravenous injection of phenylephrine (Phe, 5 µg, solid box) increased blood pressure and inhibited the firing of the neuron. Distension of a balloon at the caval-atrial junction (open box) also caused a transient inhibition of firing activity. B: neuron was recorded from an animal with a lesion of the perinuclear zone (PNZ). Neither bolus Phe (5 µg, solid box) to increase blood pressure nor balloon inflation at the caval-atrial junction (open box) affected the firing rate of this neuron.

Continuous neurons. PNZ lesions had a significant effect on both the arterial and cardiopulmonary baroreceptor sensitivity of continuously firing supraoptic neurons (P < 0.05, Table 4). Most likely, both oxytocinergic and vasopressinergic neurons are included in this category. Eighteen of the 26 continuous supraoptic neurons (69%) recorded from control rats were inhibited by an increase in blood pressure, whereas only 9 of 26 continuous neurons (35%) recorded from PNZ-lesioned rats were inhibited by an increase in blood pressure. Eighteen of the 26 continuous neurons (69%) recorded from control rats were sensitive to caval-atrial stretch, whereas just 9 of the 26 continuous cells (35%) from PNZ-lesioned rats were sensitive to caval-atrial stretch.

Histology. Histological examination of animals with ibotenic acid injections into the region of the PNZ showed extensive cell body loss and glial infiltration in the regions previously identified as the PNZ (Fig. 2; 10, 23, 34). As previously reported, the cell bodies of the magnocellular neurons in the SON remained intact (10, 23, 34). All of the effective lesions included in the study extended medially and dorsally from the SON and surrounded the anterior portion of the nucleus (Fig. 2). The lesion usually spread through at least three-fourths of the rostral-caudal length of the SON. Three of the PNZ lesions spanned the entire rostral-caudal extent of the SON. Figure 3 is a representative tracing illustrating the PNZ lesion of an animal included in this experiment. This correlates well with the area of the PNZ where Fos expression was increased following isotonic volume expansion (8, 36). If cell loss was not apparent in this region of the PNZ, the animals were not included in the study. Regions distal from the injection sites such as the hippocampus, diagonal band of Broca, and paraventricular nucleus demonstrated normal cell histology. Vehicle-injected rats had injection sites confined to the same regions as the PNZ-lesioned rats.

View larger version (73K): [in this window] [in a new window]   Fig. 2.   Photomicrographs of Giemsa-stained coronal sections including the PNZ from a rat injected with ibotenic acid. A: section taken from the nonlesioned, intact side of the animal. B: section from the side of the brain injected with ibotenic acid illustrates the histology of a lesion following ibotenic acid injection into the hypothalamus. Note the glial infiltration and lack of cell bodies within the dashed line in the region of the PNZ on the side injected with ibotenic acid. Scale bar represents 100 µm. SON, supraoptic nucleus; OT, optic tract.

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Schematic representation of rat brain illustrating the rostral-caudal progression of a lesion in the hypothalamus that includes the PNZ. Shaded areas indicate the extent of glial infiltration and neuronal loss following three serial injections of ibotenic acid. The distance (in mm) from bregma is indicated by the numbers on top.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS ELECTROPHYSIOLOGY EXPERIMENTS RESULTS DISCUSSION REFERENCES

The ability of atrial stretch to decrease the firing activity of vasopressin neurons in the SON (14, 29, 32) and paraventricular nucleus (29) and to decrease the circulating levels of vasopressin is well established. Previous studies have shown that this effect is vagally mediated (14, 29). As of yet, there is little in the way of experimental data to indicate the neural pathway responsible for cardiopulmonary regulation of vasopressin release, although the literature suggests that the PNZ is a potential part of the pathway (8, 36). The PNZ is part of the neural network responsible for arterial baroreceptor mediated inhibition of vasopressin release (10, 34). Because cardiopulmonary and arterial baroreceptor activation elicit the same effect on neurohypophysial hormones, vasopressin secretion decreases, and the literature suggests that the PNZ may be involved in both pathways (8, 10, 34, 36), we hypothesized that the PNZ lesion would interrupt arterial baroreceptor and caval-atrial stretch-mediated inhibition of supraoptic vasopressin neurons. These data show that lesions of the PNZ significantly attenuate the ability of caval-atrial stretch and increased blood pressure to decrease the activity of supraoptic vasopressinergic neurons. Thus the PNZ is part of the neural pathway by which arterial baroreceptors inhibit vasopressin release and the neural pathway by which cardiac baroreceptors inhibit vasopressin release. This is the first study to describe part of the central network that relays information from caval-atrial stretch receptors to vasopressin-secreting neurons in the SON.

The current study coupled with previous work (10, 34) indicates that the PNZ of the SON is an important region in neural regulation of vasopressin neurons. Lesions of the PNZ significantly attenuate the ability of caval-atrial stretch to decrease the activity of supraoptic vasopressinergic neurons. It is interesting that the PNZ lesion affects the blood pressure and caval-atrial stretch sensitivities of most vasopressin neurons in the same manner. If the sensitivity of a vasopressin neuron to one stimulus is attenuated, so is the neuron's sensitivity to the other stimulus. In other words, removal of the same population of PNZ neurons affects the caval-atrial and arterial baroreceptor sensitivity of an individual vasopressin neuron in the same way. Whether a single PNZ neuron is capable of modulating the response of a supraoptic vasopressin neuron to both cardiac and arterial baroreceptor activation remains to be seen. The PNZ is not a common synapse for all pathways that inhibit supraoptic vasopressin cells because inhibition of neurons following median preoptic nucleus stimulation is preserved in PNZ-lesion animals (34). Whether the PNZ is part of the circuitry that inhibits supraoptic neurons after stimulation of the amygdala and lateral septum (6) is not yet known.

It is established that arterial baroreceptor inhibition of vasopressin release at the level of the SON is mediated by GABA_A receptors (24). The SON is densely innervated with GABAergic terminals (38) from the PNZ and from other regions such as the anterior hypothalamic area (23, 38) and the arcuate nucleus (31). When GABA is iontophoresed into the SON, phasic supraoptic neurons are inhibited (4). Most of the GABA-mediated inhibition of supraoptic neurons is thought to be via GABA_A receptors, although activation of GABA_B receptors will also inhibit the spontaneous activity of magnocellular neurons (19). Although likely, it has yet to be determined whether caval-atrial stretch inhibits vasopressin-secreting supraoptic neurons via GABAergic afferent terminals from the PNZ.

As previously reported (10, 34) chronic excitotoxic lesion of the PNZ does not alter the spontaneous discharge properties of phasic supraoptic neurons. If the PNZ exerts a tonic inhibitory influence on vasopressin cells, then lesion of the PNZ would be expected to increase the firing rate of those neurons. PNZ lesion does not, however, increase the basal firing rate of vasopressin neurons. Yet both arterial baroreceptors and cardiopulmonary baroreceptors have been postulated to tonically inhibit vasopressin release (5, 33), and the PNZ is responsible for mediating the inhibitory effects of cardiac and arterial baroreceptor activation on vasopressin neurons. The recovery period allowed between PNZ lesion and the electrophysiology experiments was 3-6 days. Previous studies indicate that while sinoaortic denervation will acutely increase plasma vasopressin levels, vasopressin concentration has returned to control levels 3 days after denervation (5, 33). Therefore any transient effect of removing PNZ afferent input from supraoptic vasopressin neurons may be compensated for by the time of the electrophysiological experiments.

Because the pathways by which arterial and cardiac receptors inhibit vasopressin neurons both involve the PNZ, it is possible that the two pathways share further common nuclei. Much work has been performed to elucidate the neural pathway responsible for arterial baroreceptor control of vasopressin release (9, 10, 13, 23, 24, 34, 37). Information from the arterial baroreceptors travels via the IX and X cranial nerves to enter the central nervous system at the nucleus of the solitary tract (11). The nucleus of the solitary tract is also the likely entry point of vagal afferents activated by atrial stretch (17). A series of experiments indicate that the diagonal band of Broca, part of the telencephalon, is necessary for arterial baroreceptor-mediated inhibition of vasopressin release (9, 23, 37), although the diagonal band does not appear to be activated by volume expansion (14). The locus coeruleus is a catecholaminergic nucleus in the pons that projects to the diagonal band of Broca (25) and may also be involved in inhibition of vasopressin neurons following an increase in blood pressure (13). The firing frequency of locus coeruleus neurons increases following caval-atrial stretch, suggesting that this region may also be involved in the response to that stimulus (22). The parabrachial nucleus has been shown to project to the PNZ, and electrical stimulation of the parabrachial nucleus affects the activity of neurons in the PNZ region (21). Yet whereas the parabrachial nucleus may be involved in the neural response to arterial baroreceptor activation, it does not seem to be associated with caval-atrial stretch. An increase in blood pressure will increase or decrease the activity of parabrachial nucleus neurons, although caval-atrial stretch does not alter the activity of those cells (20). Therefore, the PNZ, nucleus of the solitary tract, and locus coeruleus may all be involved in vasopressin regulation following an increase in blood pressure and an increase in volume. There is probably some differentiation of the two pathways centrally because there is no evidence that the pathway by which cardiac receptor activation inhibits vasopressin release involves the diagonal band of Broca or the parabrachial nucleus (8, 20, 36).

The excitotoxic lesions produced in the current study are specific for cell bodies of nonmagnocellular neurons because magnocellular neurosecretory cells are resistant to the neurotoxicity of excitatory amino acids (16, 18). Neural pathways that affect vasopressin secretion but do not synapse in the PNZ would be unaffected because axons passing through the lesioned area are not damaged by ibotenic acid. This is supported by both histological (16, 40) and electrophysiological data (10, 34). Indeed, previous studies indicate that pathways that increase and decrease the firing activity of SON vasopressin neurons remain intact following similar lesion of the PNZ (10, 34). These pathways include afferents from the subfornical organ and the median preoptic nucleus (10, 34). Furthermore, if axons of passage were damaged by ibotenic acid, supraoptic neurons could not be antidromically activated because axons from the SON project to the posterior pituitary through the lesioned area (1). The antidromic latencies of supraoptic neurons recorded from vehicle rats are not different from the latencies reported in PNZ-lesioned animals. The resistance of magnocellular neurons to ibotenic acid and the fact that this excitotoxin spares axons of passage make this chemical ideal for use in this particular study since the toxin will be specific for PNZ neurons without damaging SON neurons.

PNZ lesion suppressed the sensitivity of most vasopressin neurons to arterial baroreceptor stimulation and caval-atrial stretch, but a population of supraoptic vasopressin secreting neurons retained their sensitivity. The distribution of PNZ neurons appears to be diffuse (23), so that it may be difficult to lesion the entire nucleus. There is also evidence of PNZ afferent projections to the contralateral SON (2). These projections would remain intact with the unilateral PNZ lesions of the current study and may explain the vasopressin neurons that retain sensitivity to cardiac and arterial baroreceptor stimulation. Another potential explanation for the incomplete effect of a PNZ lesion is the existence of other inputs to the SON responsible for the remaining cardiopulmonary and arterial baroreceptor sensitivity of vasopressin neurons. Indeed, the majority of GABAergic terminals within the SON do not originate in the PNZ (23, 38), so an alternative inhibitory pathway may exist that involves the arcuate nucleus (31) or anteroventral third ventricle region (3). Nonetheless, the effects of PNZ lesion on arterial and cardiac receptor sensitivity are significant, with most of the vasopressin neurons recorded from those rats insensitive to both stimuli.

The present study supports the hypothesis that the PNZ of the SON is a brain region common to the neural pathways controlling cardiopulmonary and arterial baroreflex regulation of supraoptic vasopressin neurons. Lesion of the PNZ attenuates the sensitivity of supraoptic vasopressin neurons to both caval-atrial stretch and arterial baroreceptor activation. Whether the PNZ is the final step in the pathway by which caval-atrial stretch depresses the activity of vasopressin-secreting supraoptic neurons and whether the inhibition is GABAergic remains to be determined.

Perspectives