INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE RESPONSE TO HEMORRHAGE includes autonomic changes, which serve to maintain blood pressure during decreased blood volume. A reflex increase in sympathetic nerve activity and heart rate maintains blood pressure initially (stage I) and diverts blood flow to vital organs. When blood loss becomes severe with a 20-30% loss of blood volume, arterial blood pressure falls dramatically due to sympathoinhibitory and bradycardic responses (stage II). Initiation of hemorrhagic hypotension of stage II has been attributed to activation of cardiac vagal afferents (30, 35) resulting from forceful contraction of the heart around empty cardiac chambers. However, the central mechanisms generating the sympathoinhibition during stage II remain undetermined. The sympathoinhibition can be attenuated (24, 29) and the depressor effect delayed (11, 29) after blockade of serotonin synthesis by administration of p-chlorophenylalanine (24) or blockade of serotonin [5-hydroxytryptamine (5-HT)] receptors (11, 24, 29), indicating that serotonin is involved in the mediation of the renal sympathoinhibitory and hypotensive responses to hemorrhage. The lack of effect of the broad-spectrum 5-HT receptor antagonist methysergide administered intravenously vs. intracerebroventricularly on the sympathoinhibition during severe hemorrhage indicated that the serotonergic component is centrally mediated (24, 29). However, the 5-HT receptor subtype and site of action have not been determined.

Sympathoinhibition and hypotension can be evoked by activation of 5-HT_1A receptors in the rostroventrolateral medulla (RVLM) (2). Furthermore, sympathoinhibition and hypotension can also be elicited by activation of neurons in the ventrolateral periaqueductal gray matter (vlPAG) in the midbrain, a response that is mediated via 5-HT_1A receptors located in the sympathoexcitatory region of the RVLM (1). Although direct projections from the vlPAG to the RVLM have been identified (4), neurons in the vlPAG have also been shown to project to the caudal raphe region (15), including the midline raphe obscurus and pallidus. Sympathoinhibition and hypotension can be elicited from the caudal raphe (5, 34), mediated by inhibition of sympathoexcitatory neurons in the RVLM (34), and the hypotensive response to hemorrhage is attenuated after blockade of neurotransmission through this midline region (16). Therefore, the present study was performed to determine whether blockade of 5-HT_1A receptors in the sympathoexcitatory region of the RVLM influences the sympathoinhibitory response to severe hemorrhage, suggesting that descending information from sites such as the periaqueductal gray matter or caudal raphe may be involved in evoking this response.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The protocol for this study was approved by the Animal Care and Use Committees at the Medical College of Wisconsin and the Zablocki Department of Veterans Affairs Medical Center. The experiments were performed in Sprague-Dawley rats (250-380 g) anesthetized with pentobarbital sodium (50 mg/kg ip) with a catheter inserted into a femoral vein for supplemental administration of anesthetic. Arterial blood pressure was monitored continuously from a femoral arterial cannula connected via a pressure transducer (Statham) to a polygraph (model 7, Grass) and recorded on tape (model 3000A PCM recording adapter, Vetter). A catheter was placed in a common carotid artery for withdrawal of arterial blood, and the trachea was cannulated through a midline cervical incision. A heating pad was used to maintain body temperature at 37°C. Blood samples were taken at intervals, and arterial blood gases were maintained within physiological limits by infusion of bicarbonate.

The head of the animal was fixed in a stereotaxic frame (Kopf), and sympathetic nerve activity was recorded using flexible silver wire electrodes positioned on a renal nerve via a retroperitoneal approach. The electrodes were fixed in position with Silastic gel, allowing adjustment of the body of the animal without disturbing neural recordings. The electrophysiological signal was directed first to a high-impedance differential preamplifier (gain = 1,000, pass band = 0.1-10-kHz) and then to a filter/amplifier (gain  400, high and low pass filtering = 10 Hz-3 kHz) and recorded on tape. The amplifier output was directed to a precision full-wave rectifier and averaged using a Bessel linear averaging filter (averaging interval = 100 ms) to obtain an on-line moving time average (17). The averaged activity was displayed on a recorder (model 7, Grass) to observe trends during the experiment.

In all rats, hemorrhage was performed by withdrawal of blood from the arterial line. Unless indicated, blood was withdrawn continuously at a rate of 1 ml/min until mean arterial pressure decreased to ~50 mmHg. The amount of blood withdrawn varied among groups.

Three experimental groups of animals were utilized to examine whether 5-HT_1A receptors in the RVLM were involved in the sympathoinhibition accompanying severe hemorrhage. Because of nonrecovery from the hemorrhage, one group of rats was used as controls for two other experimental groups, which included one group in which 5-HT_1A receptors were blocked in the sympathoexcitatory region of the RVLM and an additional group in which 5-HT_1A receptors were blocked in RVLM regions adjacent to the sympathoexcitatory region.

For control animals (n = 7), hemorrhage was performed until arterial pressure reached 50 mmHg. Blood pressure and sympathetic nerve activity were monitored continuously for 5 min before hemorrhage, during hemorrhage, and for 15 min after hemorrhage.

For animals in which rostral medullary 5-HT_1A receptors were to be blocked, a dorsal craniotomy was performed and the dura was reflected to allow the insertion of a multibarreled glass micropipette. A triple-barreled glass micropipette (10-20 µm total tip diameter), attached to a pressure ejection system developed and built in the laboratory, was used to pressure eject agents into the brain stem. The micropipette was advanced into the RVLM slowly using a microdrive, and initial coordinates (in mm) with respect to bregma, midline, and dorsal surface were 11.8, 1.2, and 8.5, respectively, targeting the sympathoexcitatory region with a site identified functionally as described below. The volume of ejectate was measured by observing the level of the fluid meniscus through a graduated monocular microscope eyepiece (7 nl/division), and the amount of drug administered could be controlled by altering the ejected volume by changing driving pressure, duration, or frequency of the driving pressure. An ejection marker was recorded on tape for later analysis. Artificial cerebrospinal fluid vehicle (20 nl) was injected as control for volume and pressure effects of ejected solution. The synaptic excitant DL-homocysteic acid (DLH, 0.1 M; Sigma-Aldrich) and the specific 5-HT_1A receptor antagonist WAY-100635 [WAY, N-(2-(4,2-methoxyphenyl)-1-piperazinyl)ethyl)-N-2-pyridinyl-cyclohexane-carboxamide maleate, 1 mM; Research Biochemicals International] were used. WAY was diluted in artificial cerebrospinal fluid containing 1% pontamine sky blue dye adjusted to pH 7.2-7.4 to mark ejection sites.

A micropipette was inserted at the coordinates targeting the sympathoexcitatory region of the RVLM, and vehicle was ejected (20 nl) from one of the three barrels as a control. In the absence of vehicle effects, DLH (7 nl) was microinjected from a second barrel, with fine adjustment of the micropipette as necessary to locate a site at which a pressor response with an increase >20 mmHg and an accompanying increase in sympathetic nerve activity were evoked. This protocol was used to indicate tip placement in the RVLM region containing sympathoexcitatory neurons (1). A recovery period of 30 min was allowed. The 5-HT_1A receptor antagonist WAY (14 nl) was then ejected from the third barrel at the RVLM pressor site. Two minutes after the microinjection of WAY, hemorrhage was performed as described above for control animals.

For animals (n = 6) in which 5-HT_1A receptors in regions adjacent to the sympathoexcitatory region of the RVLM were blocked, WAY was microinjected at sites within the RVLM from which no changes in blood pressure or sympathetic nerve activity were seen in response to DLH microinjections. Two minutes after the microinjection of WAY, hemorrhage was performed as described above for control animals.

Two other groups were used as additional controls. To determine the response to withdrawal of a standard volume of blood after blockade of 5-HT_1A receptors in the RVLM vs. removal of variable amounts of blood to reach the predetermined pressure of 50 mmHg, WAY was microinjected into the sympathoexcitatory RVLM in three animals as described above, and 2 min later hemorrhage was performed at a rate of 1 ml/min for 4 min. Renal sympathetic nerve activity and blood pressure were monitored before, during, and after hemorrhage. To confirm that WAY did not have a nonspecific effect on neurons in the sympathoexcitatory region of the RVLM, the viability of the baroreflex was tested in four animals before and after blockade of 5-HT_1A receptors in the sympathoexcitatory region of the RVLM. The reflex response of renal sympathetic nerve activity to bolus administration of phenylephrine (6 µg/kg iv) to raise arterial pressure by ~50 mmHg was observed before and after microinjection of WAY into the sympathoexcitatory region of the RVLM. A recovery period of 30 min was provided between the two phenylephrine injections, and a recovery period of 2 min was provided after microinjection of WAY. Hemorrhage was performed in these rats after recovery of baseline levels of sympathetic nerve activity.

Finally, to assess the role of baroreceptors/chemoreceptors in the renal sympathetic response to hemorrhage after blockade of 5-HT_1A receptors in the sympathoexcitatory region of the RVLM, sinoaortic denervation (SAD) was performed in three additional animals. Via a midline cervical incision, the carotid sinus nerves were sectioned bilaterally, and loose ties were placed around the aortic depressor nerves. The animal was then placed in the stereotaxic frame, and the protocol was followed to locate and microinject WAY into the pressor region of the RVLM. The aortic depressor nerves were then bisected bilaterally. Effectiveness of the SAD was confirmed by the absence of a reflex decrease in renal sympathetic nerve activity in response to phenylephrine administration (6 µg/kg iv) before hemorrhage as outlined above. To confirm that sympathetic nerve activity was not maximized and thus prevented from further increases after SAD, DLH microinjection into the pressor site was repeated after hemorrhage. Sympathetic nerve activity increased in these animals, indicating the lack of a "ceiling" effect of SAD on sympathetic nerve activity.

Data analysis. Analysis of taped data of blood pressure and renal nerve activity was performed by sampling blood pressure and averaged nerve activity at a rate of 20 Hz before, during, and after hemorrhage using a Hewlett-Packard 310 computer equipped with an Infotek 16-channel, 12-bit analog-to-digital converter. An analysis program was used to display blood pressure and nerve activity on the computer monitor, along with a movable cursor. The cursor was set at the onset of hemorrhage and acted as a zero time marker for the analysis. Nerve activity and blood pressure were averaged over sequential 30-s periods before, during, and after hemorrhage and subsequently expressed as percent changes from four averaged 30-s prestimulus baseline periods. To eliminate noise, zero nerve activity was obtained at the end of the experiment by crushing the nerve proximal to the recording electrodes, averaging the remaining noise level, and subtracting it from the averaged activity (17). The percent change in activity from baseline levels was compared between groups before and after antagonist administration for each protocol using one-way analysis of variance. Duncan's post hoc analysis was employed with the level of significance set at P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animals were hemorrhaged to the same level of blood pressure in the control group and in those in which WAY was microinjected into the sympathoexcitatory region of the RVLM (49.9 ± 1.8 and 52.8 ± 5.5 mmHg, respectively). However, to reduce blood pressure to this level, a significantly greater amount of blood was withdrawn from the animals with blockade of RVLM 5-HT_1A receptors (5.8 ± 0.5 vs. 3.5 ± 0.3 ml in control animals). In control animals, this was equivalent to withdrawal of 15% of total blood volume vs. ~23% in animals with RVLM 5-HT_1A receptors blocked.

An example of a hemorrhage response in a control rat is shown in Fig. 1. Hemorrhage first produced a decrease in arterial blood pressure that was initially accompanied by an increase in renal sympathetic nerve activity. After 2 min and loss of ~7.5% of blood volume (about half of the total hemorrhage amount), a further decrease in arterial blood pressure was accompanied by a decrease in renal sympathetic nerve activity (Fig. 1A). These data are typical of responses reported during the initial sympathoexcitatory (stage I) and decompensatory (stage II) phases of severe hemorrhage (17, 24, 29, 30, 35). With continued hemorrhage, blood pressure fell to 50 mmHg after 4 min, at which point hemorrhage was stopped. Blood pressure then increased slowly but remained below baseline levels (Fig. 1, A and B). Renal sympathoinhibition also reached its nadir at 4 min (end of hemorrhage) and then gradually increased toward baseline levels over the following 10 min (Fig. 1, A and B). Summed data for all control rats are shown in Fig. 2. These data show that, at the end of hemorrhage, blood pressure was significantly decreased by 78.8 ± 5.8 mmHg from a baseline level of 133.7 ± 4.7 mmHg (n = 7), and renal sympathetic nerve activity was significantly decreased to 33.3 ± 9.0% below baseline levels (n = 7).

View larger version (35K): [in this window] [in a new window]   Fig. 1.   A: control response to hemorrhage on arterial blood pressure (BP), renal sympathetic nerve activity (RSNA), and averaged RSNA (avg RSNA). C: response to hemorrhage after microinjection of the 5-hydroxytryptamine type 1A (5-HT1A) receptor antagonist WAY-100635 (WAY, 1 mM, 14 nl) into the sympathoexcitatory region of the rostroventrolateral medulla (RVLM). au, Arbitrary units. B and D: responses at 15 min after onset of hemorrhage before and after WAY, respectively.

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Change in RSNA and BP in response to hemorrhage in control animals (n = 7) and animals subjected to blockade of 5-HT1A receptors in the sympathoexcitatory region of the RVLM (n = 6) by microinjection of 1 mM WAY (14 nl). *Significantly different from baseline and control; significantly different from baseline; significantly different from control (P < 0.05).

In animals in which WAY was microinjected into the sympathoexcitatory region of the RVLM, blockade of 5-HT_1A receptors did not alter baseline levels of blood pressure or nerve activity (Fig. 3). For the group data before WAY administration, mean arterial blood pressure was 142.1 ± 7.8 mmHg (n = 6), while after WAY administration mean pressure was 141.4 ± 7.3 mmHg (n = 6), and baseline nerve activity was 100.2% of control (n = 6). Baseline blood pressure after WAY administration was not significantly different from baseline blood pressure of the control group of rats. Respiratory rate, as measured by the oscillations on the blood pressure traces, was 57 ± 7.0 breaths/min (n = 6) before WAY and remained constant at 58 ± 6.7 breaths/min after microinjection of WAY. An example of the response of a rat to hemorrhage after RVLM microinjection of WAY is shown in Fig. 1C. This example shows that, after blockade of 5-HT_1A receptors in the sympathoexcitatory region of the RVLM, the hypotensive response to hemorrhage was delayed. Approximately 7 min of hemorrhage were required to reach the predetermined pressure of 50 mmHg. In addition, no sympathoinhibition was seen in response to hemorrhage. Renal nerve activity increased and remained above baseline throughout the experimental period (Fig. 1D). Summed data for these animals are shown in Fig. 2. At 2-3 min after the onset of hemorrhage, the hypotensive response in the animals with blockade of 5-HT_1A receptors in the sympathoexcitatory region of the RVLM was significantly less than in control animals. In addition, renal nerve activity in these animals actually increased over time, reaching a peak at 3 min, after which activity decreased slowly but remained above baseline. The sympathoinhibition that was seen in control animals was eliminated. After hemorrhage, blood pressure remained lower in the WAY-treated rats than in the control animals, although not significantly, perhaps because of the greater decrease in blood volume needed to drop arterial pressure during the hemorrhage.

View larger version (30K): [in this window] [in a new window]   Fig. 3.   Microinjection of 1 mM WAY (14 nl) into the sympathoexcitatory region of the RVLM does not alter BP, RSNA, and avg RSNA. Animal is the same as that used in Fig. 1, C and D.

As a control for the response to hemorrhage of a comparable volume of blood, 4 ml were withdrawn in a group of rats in which 5-HT_1A receptors in the sympathoexcitatory region of the RVLM were blocked by microinjection of WAY. This volume was less than that withdrawn from the group described above, in which on average 5.8 ml of blood were withdrawn to reduce pressure to 50 mmHg. In animals in which 4 ml of blood were withdrawn, the increase in sympathetic nerve activity was the same as that in rats that were hemorrhaged to 50 mmHg (Fig. 4). However, the accompanying decrease in blood pressure was significantly less than in the other WAY-treated group by ~40 mmHg, and recovery improved to control levels after the hemorrhage (Fig. 4).

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Changes in RSNA and BP in response to hemorrhage in animals subjected to blockade of 5-HT1A receptors in the sympathoexcitatory region of the RVLM with 1 mM WAY (14 nl). Blood was withdrawn until pressure fell to 50 mmHg (from Fig. 2), and 4 ml of blood were withdrawn (n = 3). *Significantly different from baseline and BP = 50 mmHg; significantly different from baseline; significantly different from BP = 50 mmHg (P < 0.05).

Microinjection sites of WAY in the pressor region of the RVLM were verified histologically to be located in the lateral region of the nucleus paragigantocellularis lateralis at the rostral pole of the inferior olives (Fig. 5). As a control, hemorrhage was performed in a group of rats (n = 6) in which WAY was microinjected into adjacent sites in the RVLM from which no pressor or sympathoexcitatory effects could be elicited by microinjection of DLH. Histologically, most sites at which no pressor response was elicited were located more rostrally or medially in the RVLM (Fig. 5). In these animals, the renal sympathetic nerve and blood pressure responses to hemorrhage were the same as those elicited in control animals (Fig. 6).

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Diagrammatic, transverse sections through the rostral medulla indicating sites of microinjection of DL-homocysteic acid (DLH) and WAY into the RVLM (10.0 and 10.3-10.5 mm caudal to bregma). , Sites from which DLH could elicit a pressor and sympathoexcitatory response; shaded squares, control microinjection sites outside the pressor region of the RVLM. NA, nucleus ambiguus; TB, trapezoid body; VII, facial nucleus.

View larger version (25K): [in this window] [in a new window]   Fig. 6.   Change in RSNA and BP in response to hemorrhage in control animals (from Fig. 2) and animals in which 1 mM WAY (14 nl) was microinjected outside the sympathoexcitatory region of the RVLM (n =6). Significantly different from baseline (P < 0.05).

The baroreceptor reflex, induced by an intravenous injection of a bolus of phenylephrine (Fig. 7A), remained viable after blockade of 5-HT_1A receptors in the sympathoexcitatory region of the RVLM with WAY. Reflex renal nerve sympathoinhibition was still observed in response to the acute elevation in pressure (Fig. 7B). As a further test of the involvement of the baroreceptors/chemoreceptors in the sympathetic response to hemorrhage after 5-HT_1A receptor blockade, hemorrhage was performed in a group of SAD rats (n = 3) after RVLM microinjection of WAY. The hypotensive response to hemorrhage in the 5-HT_1A receptor-blocked SAD rats was compared with the response previously obtained in barointact, but 5-HT_1A blocked, animals (Fig. 2). The hypotensive response between groups was similar, but the increase in renal sympathetic activity in barointact animals was eliminated by SAD (Fig. 8).

View larger version (47K): [in this window] [in a new window]   Fig. 7.   Baroreflex inhibition of avg RSNA in 1 animal in response to a phenylephrine-induced (6 µg/kg iv, at arrow) increase in BP before (A) and after (B) microinjection of 1 mM WAY (14 nl) into the pressor region of the RVLM.

View larger version (26K): [in this window] [in a new window]   Fig. 8.   Change in RSNA and BP in response to hemorrhage in animals with blockade of 5-HT1A receptors in the sympathoexcitatory region of the RVLM by microinjection of 1 mM WAY (14 nl) in barointact (from Fig. 2) and sinoaortic-denervated (SAD) animals (n = 3). *Significantly different from barointact; significantly different from baseline (P < 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The initial response to hemorrhage included a decrease in arterial blood pressure accompanied by a baroreflex increase in renal sympathetic nerve activity. With increasing blood loss, a decompensatory phase included a rapid decrease in blood pressure accompanied by a renal sympathoinhibition. The neural and cardiovascular responses evoked during hemorrhage are similar in anesthetized and unanesthetized animals (11, 24, 28-30, 38). However, the initial compensatory stage is shortened in the presence of anesthetics, including pentobarbital sodium (38), presumably by a central action of the anesthetic drug. In general, in unanesthetized animals, the decompensatory stage of hemorrhage begins after 30% of total blood volume is lost (28, 38). Data from the present study are in accordance with other studies performed in anesthetized animals (16, 28), in that the onset of the decompensatory stage of hemorrhage and the sympathoinhibition occur after loss of only 7.5% of total blood volume.

Data from the present study indicate that the renal sympathoinhibition that accompanies the hypotensive response to severe hemorrhage is mediated by 5-HT_1A receptors in the sympathoexcitatory region of the RVLM. When hemorrhaged by a blood volume similar to that of control animals, the sympathoinhibitory response to hemorrhage was eliminated and hypotension was attenuated after blockade of 5-HT_1A receptors in the sympathoexcitatory region of the RVLM. Several other studies have implicated a serotonergically mediated mechanism in the elaboration of the hypotensive and sympathoinhibitory responses to severe hemorrhage (10, 24, 29). The broad-spectrum 5-HT receptor antagonist methysergide has been reported to improve recovery of arterial pressure after hemorrhage in cats (9). Blockade of serotonin synthesis by administration of p-chlorophenylalanine (24) or blockade of 5-HT receptors with methysergide delayed the hypotensive (11, 24, 29) and abolished the sympathoinhibitory (24, 29) responses to severe hemorrhage. Methysergide was ineffective or significantly less effective when administered intravenously than when administered intracerebroventricularly (11, 24, 29), suggesting the involvement of a central serotonergic mechanism, although the site of action was unknown. The actions of methysergide as a partial agonist at 5-HT receptors (26) and an affinity for 5-HT₁ and 5-HT₂ receptor subtypes have prevented the identification of the 5-HT receptor subtype involved in the mediation of the sympathoinhibition during severe hemorrhage. With acknowledgment of different species and varying hemorrhagic conditions, 5-HT_1A and 5-HT_2C receptors remained the likely subtypes responsible for the mediation of the sympathoinhibitory response to severe hemorrhage, primarily by a process of elimination by use of pharmacological receptor blockade (10, 11, 29). In the present study, we microinjected WAY, which is considered to be a selective and silent 5-HT_1A receptor antagonist (14), into the sympathoexcitatory region of the RVLM for localized receptor blockade. Local administration of WAY can block the inhibitory effects of 5-HT_1A receptor agonists on the discharge of putative sympathoexcitatory neurons (unpublished observations). However, in the dose used in vivo, there is no evidence that WAY blocks 5-HT_1A receptors selectively. The lack of effect of WAY on baseline nerve activity and blood pressure indicates no agonist activity or tonic serotonergic control at 5-HT_1A receptors at this level of the nervous system. Microinjection of WAY did not abolish the baroreflex inhibition of renal sympathetic nerve activity, mediated through GABA receptors on sympathoexcitatory neurons in the RVLM (32), ruling out a nonspecific effect of WAY on sympathoinhibitory responses mediated through RVLM sympathoexcitatory neurons. The ventral respiratory group, including the respiratory rhythm generator in the pre-Botzinger complex, is also located in the RVLM. It is unlikely that the 5-HT_1A receptor antagonist influenced sympathetic nerve responses indirectly by an action at neurons in the rostral ventral respiratory group, because respiratory rate was unchanged after microinjection of WAY into the RVLM, although we did not measure tidal volume.

Administration of WAY into the sympathoexcitatory region of the RVLM attenuated the sympathoinhibitory response to severe hemorrhage, delayed the hypotension, and improved blood pressure recovery. In contrast, microinjection of WAY at adjacent, nonpressor sites had no effect on the response to hemorrhage. These data indicate that the central site of action of 5-HT in the generation of sympathoinhibition during severe hemorrhage is at 5-HT_1A receptors in the sympathoexcitatory region of the RVLM. Electrophysiological studies demonstrate that the discharge rate of putative sympathoexcitatory neurons in the RVLM can be reduced by local application of 5-HT_1A receptor agonists (19, 35). In addition, activation of 5-HT_1A receptors in the sympathoexcitatory region of the RVLM by microinjection of the 5-HT_1A receptor agonist 8-hydroxy-2-(di-n-propylamino)tetralin results in differential patterns of sympathoinhibition that correlate to the topographical organization of sympathoexcitatory neurons in the cat (2, 8). Taken together, these data suggest that the serotonergic sympathoinhibition is likely to be mediated at sympathoexcitatory neurons in the RVLM.

The central pathway mediating the sympathetic response to severe hemorrhage has not been conclusively determined, but a recent study has shown that hypotension and sympathoinhibition elicited by activation of neurons in the vlPAG can be attenuated by blockade of 5-HT_1A receptors in the sympathoexcitatory region of the RVLM (1). The vlPAG has been shown to have inputs to the RVLM (4), but there are few serotonergic neurons in the vlPAG (31), and it has been suggested that inhibition of RVLM neurons by neurons in the vlPAG may be mediated at least in part via the caudal midline raphe (15). Serotonin-containing neurons have been located in the caudal raphe obscurus and raphe pallidus (7), and projections from the caudal raphe to RVLM have been identified (25, 34). A role for the caudal raphe in the cardiovascular response to severe hemorrhage is suggested by data that showed that inactivation of neurons in a vasodepressor area of the caudal midline medulla resulted in a delayed onset and an attenuation of the hypotension produced by hemorrhage (16). Furthermore, depressor and sympathoinhibitory responses elicited from caudal midline medullary neurons have been shown to be mediated by inhibition of the discharges of sympathoexcitatory neurons in the RVLM (5, 34). It has been suggested that a polysynaptic pathway mediates the inhibition via GABA receptors in the RVLM (6), but a monosynaptic serotonergic pathway could mediate the inhibition. Alternatively, a polysynaptic pathway from the raphe to the RVLM with a local serotonergic projection from the adjacent B3 serotonergic cells or within the RVLM could mediate the observed sympathoinhibitory response.

Baroreceptor afferents mediate the initial sympathoexcitation evident during stage I, but the baroreflex is attenuated or overridden during the generation of the sympathoinhibition of stage II (3), and the mechanism is not known. The present data demonstrate that a sympathoexcitatory influence overridden during the response to hemorrhage is abolished by SAD, suggesting the involvement of the baroreceptors and/or chemoreceptors. Hypotension to 50 mmHg has been reported to increase chemoreceptor discharge (20, 21), probably through the decrease in flow, which may elicit the increase in renal sympathetic nerve activity. As a whole, the response to hemorrhage is complex and likely involves inputs integrated at several levels, but the present data suggest that RVLM 5-HT_1A receptors are involved in the mediation of the sympathoinhibitory component and the suppression of a sympathoexcitatory influence during severe hemorrhage.

Activation of the decompensatory cardiovascular response to hemorrhage is considered to be via cardiac afferent stimulation, probably triggered by the fall in blood volume. Hemorrhage after section of vagal afferents produced only an increase in renal sympathetic nerve activity in the rat model (24, 29, 35), and blockade of cardiac nerves abolished the sympathoinhibition and hypotension in hemorrhage in rabbits (3, 12). Bradykinins, prostaglandins (30), and 5-HT₃ receptors (29) are not thought to be involved in the activation of cardiac afferents, but the initiating mechanism remains unresolved. It is possible that there is a paradoxical activation of cardiac mechanoreceptors due to contraction around emptying chambers. Regardless of the mechanism in the integrated sympathetic response to severe hemorrhage, the emphasis shifts from the baroreceptor/chemoreceptor afferents to the cardiac afferents with increasing blood loss, but the site of their interaction is not known.

Perspectives