Vasopressin is a major vasoconstrictor involved in hindlimb vascular responses to stimulation of adenosine A₁ receptors in the nucleus of the solitary tract

Abstract

Our previous study showed that stimulation of adenosine A₁ receptors located in the nucleus of the solitary tract (NTS) exerts counteracting effects on the iliac vascular bed: activation of the adrenal medulla and β-adrenergic vasodilation versus vasoconstriction mediated by neural and unknown humoral factors. In the present study we investigated the relative contribution of three major potential humoral vasoconstrictors: vasopressin, angiotensin II, and norepinephrine in this response. In urethane-chloralose anesthetized rats we compared the integral changes in iliac vascular conductance evoked by microinjections into the NTS of the selective A₁ receptor agonist N⁶-cyclopentyladenosine (CPA; 330 pmol in 50 nl) in intact (Int) animals and following: V₁ vasopressin receptor blockade (VX), angiotensin II AT₁ receptor blockade (ATX), bilateral adrenalectomy + ganglionic blockade (ADX + GX; which eliminated the potential increases in circulating norepinephrine and epinephrine), ADX + GX + VX and ADX + GX + VX + ATX. In Int animals, stimulation of NTS A₁ adenosine receptors evoked typical variable responses with prevailing pressor and vasoconstrictor effects. VX reversed the responses to depressor ones. ATX did not significantly alter the responses. ADX + GX accentuated pressor and vasoconstrictor responses, whereas ADX + GX + VX and ADX + GX + VX + ATX virtually abolished the responses. Stimulation of NTS A₁ adenosine receptors increased circulating vasopressin over fourfold (26.4 ± 10.4 vs. 117.0 ± 19 pg/ml). These data strongly suggest that vasopressin is a major vasoconstrictor factor opposing β-adrenergic vasodilation in iliac vascular responses triggered by stimulation of NTS A₁ adenosine receptors, whereas angiotensin II and norepinephrine do not contribute significantly to the vasoconstrictor responses.

it is now widely accepted that adenosine operating via A₁ and A_2a receptors modulates neural cardiovascular control at the level of the nucleus tractus solitarii (NTS) and other brainstem cardiovascular centers (2, 4, 5, 20, 31, 36, 38, 41). Under normal, physiological conditions a natural source of adenosine is ATP released synaptically from neurons as well as from glial cells activated by neighboring neurons (6, 8, 9, 14, 38). This extracellular ATP is catabolized via ectonucleotidases to adenosine, which further acts more broadly as a neuromodulator operating via pre- or postsynaptic A₁ and A_2a adenosine receptor subtypes (8, 26, 38, 46). Under pathological conditions such as ischemia, hypoxia, and severe hemorrhage, a global release of adenosine from many cell types occurs via the breakdown of intracellular ATP (25, 43, 45). Thus, adenosine, which is generated in or released into the extracellular space under physiological or pathological conditions, acts not via specific synapses in specific neuronal pathways but rather spatially, reaching all adenosine receptor subtypes in the vicinity; this may resemble signaling via diffusive neurotransmitters/neuromodulators like nitric oxide or carbon monoxide. This spatial aspect of the action of naturally released adenosine seems especially important for the NTS, where groups of functionally different neurons usually overlap allowing for adenosine crosstalk. Specific physiological effects exerted via nonselective, spatial spread of adenosine in the NTS may depend on differential expression of adenosine receptor subtypes on functionally distinct NTS neurons/terminals, as our laboratory suggested previously (31, 36).

In the central nervous system adenosine may inhibit or facilitate release of neurotransmitters from synaptic terminals as well as directly inhibit or activate neurons via pre- and postsynaptic A₁ or A_2a receptors, respectively (8, 26, 31, 36). Since A₁ versus A_2a adenosine receptors exert contrasting effects on central neurons/terminals, reciprocal effects are usually, although not always, observed in response to selective stimulation of the two receptor subtypes. NTS A₁ adenosine receptor stimulation predominately yields differential regional sympathoactivation (adrenal > renal ≥ lumbar) and pressor responses (2, 5, 30, 35). In contrast, NTS A_2a receptor stimulation typically evokes depressor responses accompanied by contrasting regional sympathetic responses: decreases in renal sympathetic nerve activity (RSNA), no changes in lumbar sympathetic nerve activity (LSNA), and increases in preganglionic adrenal sympathetic nerve activity (pre-ASNA) (2, 5, 16, 32–34). Note that whereas stimulation of NTS A₁ and A_2a receptors evoke contrasting changes in RSNA and LSNA, both adenosine receptor subtypes activate the sympathetic output to the adrenal medulla (33, 35).

Recent studies from our laboratory showed that the pressor and sympathoexcitatory responses evoked by stimulation of NTS A₁ adenosine receptors are mediated mostly via inhibition of baroreflex mechanisms at the level of the NTS, whereas hemodynamic and differential sympathetic responses evoked by stimulation of NTS A_2a adenosine receptors are mediated mostly via activation of nonbaroreflex mechanisms (16, 30, 33–35). It should be stressed that A₁ adenosine receptors may also modulate nonglutamatergic, nonbaroreflex mechanisms operating in the NTS. For example, sinoaortic denervation or ionotropic glutamatergic blockade abolished A₁-adenosine receptor-mediated increases in RSNA and LSNA, whereas this attenuated, but did not abolish, the increases in pre-ASNA (35). The activation of pre-ASNA, which persisted after sinoaortic and the glutamatergic blockade, was most likely mediated via A₁ adenosine receptor modulation of nonglutamatergic pathways descending into the NTS from higher structures, such as from hypothalamic paraventricular and/or dorsomedial nuclei (12, 28, 42). A₁ adenosine receptors located on these descending nonglutamatergic pathways and/or NTS interneurons could selectively activate the sympathetic output to the adrenal medulla (but not RSNA and LSNA) via disinhibition of direct NTS-rostral ventrolateral medulla pathways (10). NTS A₁ adenosine receptors may also modulate the control of heart rate (HR) via both baroreflex and nonbaroreflex mechanisms (30, 35). Taken together, these observations strongly suggest that A₁ adenosine receptors are differentially located on functionally different NTS neurons/terminals, which are mostly, but not exclusively, glutamatergic and involved in the baroreflex arch.

Although selective stimulation of NTS A₁ adenosine receptors evokes predominantly pressor responses (2, 5), occasionally biphasic or even depressor responses are observed (35). This variability of the responses is a natural consequence of simultaneous activation of at least two counteracting mechanisms: sympathetic vasoconstriction and β-adrenergic vasodilation mediated via epinephrine released from the activated adrenal medulla. Since β-adrenergic receptors are preferentially located in the muscle vascular bed (44) and both pre-ASNA and LSNA increase upon stimulation of NTS A₁ adenosine receptors, it was likely that these two counteracting factors may significantly contribute to the variability of the iliac vascular responses. We confirmed this hypothesis in a recent study from our laboratory by showing that removal of the vasodilatory mechanism (via bilateral adrenalectomy as well as peripheral blockade of β-adrenergic receptors) abolished the variability of the responses normally observed in intact animals and markedly increased the pressor and hindlimb vasoconstrictor responses (21). In contrast, bilateral lumbar sympathectomy tended to increase the vasodilatory component of the responses although the variability still persisted. To our surprise, following combined adrenalectomy plus lumbar sympathectomy, a marked, consistent vasoconstrictor component still persisted, suggesting that some unknown humoral vasoconstrictor factor(s) are involved (21).

The most likely humoral candidates contributing to the iliac vasoconstriction are vasopressin, angiotensin II, and norepinephrine. Since activation of A₁ adenosine receptors may inhibit glutamate release in baroreflex pathway at the level of the NTS (30, 35), and given that the NTS is a crucial, primary relay station for tonic baroreflex inhibition of vasopressin release (10, 29, 39), it is likely that the activation of A₁ adenosine receptors may inhibit the baroreflex restraint of vasopressin release. When released, vasopressin could evoke powerful peripheral vasoconstriction. However, whether A₁ adenosine receptors are present on those NTS baroreflex neurons/terminals, which inhibit the release of vasopressin, is unknown. The present study was designed to test this hypothesis.

Since RSNA directed to the kidney has been shown to increase following stimulation of NTS A₁ adenosine receptors (35), this may facilitate the renin/angiotensin mechanism leading to humoral vasoconstriction of the hindlimb vasculature mediated via AT₁ angiotensin II receptors (11). In addition, since stimulation of NTS A₁ adenosine receptors increases the activity of various sympathetic outputs (35), it is also possible that circulating norepinephrine released from other synaptic terminals may reach the iliac vasculature and cause vasoconstriction. Therefore, in the present study we investigated the extent to which these three potential humoral vasoconstrictors (vasopressin, angiotensin II, and/or norepinephrine) contribute to NTS-A₁ adenosine receptor-elicited iliac vasoconstriction (21).

MATERIALS AND METHODS

All protocols and surgical procedures employed in this study were reviewed and approved by the Institutional Animal Care and Use Committee and were performed in accordance with the Guiding Principles in the Care and Use of Animals endorsed by the American Physiological Society and published by the National Institutes of Health.

Design

This study investigates further the mechanisms responsible for the consistent variability of hemodynamic responses elicited by activation of adenosine A₁ receptors in the NTS (21, 35). Previously, our laboratory showed that the variability of the pressor/depressor and iliac vasoconstrictor/vasodilator responses is not a simple effect of competitive interactions between sympathetic vasoconstriction versus β-adrenergic vasodilation, but some unknown, powerful, humoral vasoconstrictor factor(s) are also involved (21). Therefore, the present study assessed the relative contribution of potential humoral vasoconstricting factors to the iliac vascular responses evoked by selective stimulation of NTS A₁ adenosine receptors. Experiments were performed on a total of 102 male Sprague-Dawley rats. In 63 rats, we compared the relative vasoconstrictor effects potentially mediated via vasopressin, angiotensin II, norepinephrine, and sympathetic innervation of the hindquarters, the effects normally opposed by simultaneous β-adrenergic vasodilation mediated via activation of the adrenal medulla. In an additional 20 rats, respective time controls were performed, and in six rats the effectiveness of vasopressin and angiotensin II receptor blockades was assessed. These functional experiments strongly suggested that the major vasoconstrictor factor triggered by activation of adenosine A₁ receptors in the NTS may be vasopressin. Therefore, in an additional group of 13 animals, the levels of circulating vasopressin were evaluated before and following microinjections into the NTS of the selective A₁ adenosine receptor agonist N⁶-cyclopentyladenosine (CPA) or vehicle control.

Instrumentation and measurements.

All the procedures were described in detail previously (4, 19, 21, 32–35). Briefly, male Sprague-Dawley rats (350–400 g; Charles River) were anesthetized with a mixture of α-chloralose (80 mg/kg) and urethane (500 mg/kg ip), tracheotomized, connected to a small animal respirator (SAR-830; CWE, Ardmore, PA), and artificially ventilated with 40% oxygen-60% nitrogen mixture. Catheterization of the right femoral artery and vein were performed to monitor arterial blood pressure and infuse drugs, respectively. Arterial blood gases were tested occasionally for appropriate experimental values (Radiometer, ABL500, OSM3). Averaged values measured at the end of each experiment were the following: pH = 7.38 ± 0.01, Po₂ = 140.1 ± 3.8 mmHg, and Pco₂ = 36.2 ± 0.7 mmHg.

From a midabdominal incision, the left iliac artery was exposed. A pulse Doppler blood flow velocity transducer was placed around the artery and connected to the flowmeter (Baylor Electronics). From the same midabdominal incision in some animals, bilateral adrenalectomy or lumbar sympathectomy (L1-L6) was performed. The intermesenteric nerves were also severed in sympathectomized animals.

Arterial blood pressure and iliac flow signals were digitized and recorded with an analog-digital converter (Modular Instruments) interfaced to a laboratory computer. The signals were recorded continuously using Biowindows software (Modular Instruments), averaged over 5-s intervals and stored on hard disk for subsequent analysis.

Microinjections into the NTS.

Animals were placed in stereotaxic frame with head tilted down at 45°. After the exposure of the brainstem via dissected atlantoccipital membrane, the animals were allowed to stabilize for at least 30 min before microinjections. Unilateral microinjections of the selective A₁ adenosine receptor agonist CPA (Tocris, 330 pmol) dissolved in 50 nl of artificial cerebrospinal fluid (ACF) were made through multibarrel, glass micropipettes into the medial region of the caudal subpostremal NTS as described previously (4, 19, 32–35). This dose of CPA produced the most consistent, predominantly pressor responses in a previous study from our laboratory (35). The CPA was dissolved in ACF, and the pH was adjusted to 7.2. In several previous studies our laboratory has shown that microinjections of the same amount of ACF into the same site of the NTS did not markedly affect mean arterial pressure (MAP), HR, RSNA, LSNA, and pre-ASNA and vascular flows in iliac, renal, and mesenteric arteries (4, 32–35). The changes in all these variables were either not different from zero or smaller than natural, random fluctuations of these variables over the time of measurements. To avoid the effect of desensitization of A₁ adenosine receptors, in all experiments only one dose of the agonist was microinjected into left or right side of the NTS. All microinjection sites were marked with fluorescent dye (DiI; Molecular Probes) and verified histologically (Fig. 1) as described previously (4, 19, 32–35).

Fig. 1.

Microinjection sites in the caudal subpostremal nucleus tractus solitarii (NTS) for all experiments. Schematic diagrams of transverse sections of the medulla oblongata from a rat brain are shown. AP, area postrema; c, central canal; 10, dorsal motor nucleus of the vagus nerve; 12, nucleus of the hypoglossal nerve; ts, tractus solitarius; Gr, gracile nucleus; Cu, cuneate nucleus. Scale is shown at the bottom; the number on the left side of the schematic diagram denotes the rostro-caudal position in millimeters of the section relative to the obex according to the atlas of the rat subpostremal NTS by Barraco et al. (Ref. 1). Microinjection sites were marked with fluorescent dye and are denoted with filled symbols for the pressor responses to N⁶-cyclopentyladenosine (CPA) and corresponding open symbols for the depressor responses. A: microinjections of CPA in intact animals (●, ○), after vasopressin V₁ receptor blockade (VX; ◊), lumbar sympathectomy plus VX (LX + VX_i; ▴, ▵), and after angiotensin II AT₁ receptor blockade (ATX; ■, □). B: microinjections of CPA after bilateral adrenalectomy (ADX) plus ganglionic blockade (GX) (▴), following ADX + GX + VX (dotted closed circle, dotted open circle), after ADX + GX + VX + ATX (dotted closed square, dotted open square). In experiments where vasopressin assay was performed, the microinjection sites were denoted: + and x for CPA and artificial cerebrospinal fluid (ACF) microinjections, respectively.

We believe that the microinjection technique mimics natural, spatial (not strictly synaptic) action of adenosine in the central nervous system since adenosine is naturally produced in the intracellular space by ectonucleotidases from extracellular ATP (released from neurons and glial cells under physiological conditions) (6, 8, 14, 38, 46) or is directly released into the intracellular space from ischemic/hypoxic neurons and glial cells under pathological conditions (25, 43, 45).

Experimental Protocols

In a previous study from our laboratory we showed that in addition to sympathetic iliac vasoconstriction and β-adrenergic vasodilation, some unknown humoral vasoconstrictor(s) contribute to the consistent variability of hemodynamic responses evoked by selective stimulation of adenosine A₁ receptors located in the NTS (21). This conclusion was based on comparing of the responses observed in intact animals and following four experimental protocols: 1) β-adrenergic blockade, 2) adrenalectomy, 3) lumbar sympathectomy, and 4) combined adrenalectomy plus lumbar sympathectomy. In the last experimental condition a powerful iliac vasoconstriction was observed, indicating that nonsympathetic, humoral vasoconstrictors are involved (21). The present study is a direct extension of our previous findings and focuses on the relative contribution to the responses of three potential humoral vasoconstrictors: vasopressin, angiotensin II, and norepinephrine. Six experimental protocols were designed, according to the diagrams presented in Fig. 2. Data collected in each protocol were compared with responses observed in the intact group. In protocols 1 and 2, the contribution of vasopressin and angiotensin II was assessed by comparing hemodynamic responses elicited by stimulation of NTS A₁ adenosine receptors in intact animals with those obtained following selective blockade of vasopressin V₁ receptors and angiotensin II AT₁ receptors via intravenous injections of selective antagonists: [β-mercapto-β,β-cyclopentylmethylenepropionyl,¹-O-Me-Tyr²,Arg⁸ ]-vasopressin (20 μg/kg; Sigma) and losartan (5 mg/kg; Merck), respectively. To evaluate the potential contribution of circulating norepinephrine to the responses, ganglionic blockade (hexamethonium bromide, 25 mg/kg iv; Sigma) was combined with adrenalectomy (protocol 3), and these data and the responses observed previously following adrenalectomy alone (21) are discussed together. This indirect evaluation of norepinephrine contribution to the responses was necessary because total sympathetic denervation is impossible and ganglionic blockade, which prevents secretion of norepinephrine from sympathetic terminals, also impairs/abolishes the effects of activation of the adrenal medulla; thus ganglionic blockade removes β-adrenergic vasodilation simultaneously. Therefore, the appropriate reference point for the responses obtained following the ganglionic blockade (protocol 3) were the responses obtained following adrenalectomy alone, which has been already performed in a previous study from our laboratory (21). Protocol 4 removed the combined contribution of norepinephrine and vasopressin to the responses (ganglionic blockade + vasopressin V₁ receptor blockade), whereas protocol 5 removed the combined effect off all three potential vasoconstrictors considered via ganglionic blockade + vasopressin V₁ receptor blockade + angiotensin AT₁ receptor blockade. Protocols 3–5 were performed in adrenalectomized animals to clarify the experimental conditions by removing any residual adrenal responses, which may potentially persist following the ganglionic blockade. Since preliminary results of the above five experimental protocols strongly suggested that only vasopressin has a marked contribution to the responses, in protocol 6 the magnitude of β-adrenergic vasodilation alone, not opposed by major vasoconstrictor factors (sympathetic vasoconstriction and vasopressin), was assessed. In this protocol bilateral lumbar sympathectomy was combined with blockade of V₁ vasopressin receptors, and these data were compared with data following V₁ vasopressin receptor blockade alone (protocol 1) and discussed together with previous data obtained following bilateral lumbar sympathectomy alone (21).

Fig. 2.

Time line of how experimental protocols are executed. Abbreviations are as in Fig. 1. Time control experiments were performed for protocols including ATX and/or GX (protocols 2–5) according to the respective diagrams; however, microinjections of CPA were omitted. PE, phenylephrine.

The effectiveness of vasopressin V₁ and angiotensin AT₁ receptor blockades were tested in separate groups of animals (n = 3 for each blockade) with intravenous injections of arginine-vasopressin (50 mU/kg; Sigma) and angiotensin II (300 ng/kg; Sigma), respectively, before and after the blockade. Both blockades remained effective for over 1 h. Blockade of V₁ vasopressin receptors caused relatively small decreases in MAP and increases in iliac vascular conductance (IVC; Table 1), which spontaneously returned toward resting values in ∼10 min; therefore, ∼10 min after V₁ vasopressinergic blockade the microinjection of CPA was performed in protocols 1 and 6 (Fig. 2). However, following blockade of angiotensin AT₁ receptors and ganglionic blockade, marked and sustained decreases in MAP and increases in IVC were observed (Table 1). Therefore, in protocols 2–5, where these blockades were performed, intravenous infusions of phenylephrine (PE; 200 μg/ml; Sigma) were used to return the hemodynamic parameters toward baseline, preblockade values. Table 2 presents the rates of PE infusion needed for the compensation. No differences were observed in PE infusion rates between protocols 2 and 4. However, significantly greater PE infusion rates were required when angiotensin AT₁ receptor antagonist, losartan, was combined with ganglionic and V₁ vasopressin receptor blockades in protocol 5 (Table 2). During the responses to stimulation of NTS A₁ adenosine receptors, PE infusion was continued at the same rates as needed to compensate for the altered hemodynamic values in protocols 2–5. The effect of PE infusion on baseline hemodynamic values was estimated in respective time controls for protocols 2–5 (Table 2). In the time-control experiments, all procedures except microinjections of CPA were performed in the same time pattern as in experimental protocols 2-5. Figure 3 shows an example of a time control for the most complex experimental protocol 5. PE infusion rates were similar in the experimental protocols and respective time controls (Table 2).

Table 1.

Maximal hemodynamic responses evoked by blockade of V₁ vasopressin receptors, AT₁ angiotensin II receptors, and ganglionic blockade

Experimental Groups	n	Δ% Mean Arterial Pressure, mmHg	Δ% Heart Rate, beats/min	Δ% Iliac Blood Flow	Δ% Iliac Vascular Conductance
VX	16	−4.7±1.8*	1.2±0.9	10.2±1.6*	15.3±2.7*
GX	39	−39.7±1.9*	3.1±1.9	16.6±3.3*	99.9±9.9*
ATX	15	−49.2±2.8*	−13.9±3.5*	8.6±6.1	125.1±18.2*

Values are means ± SE. Abbreviations are as in Fig. 1. Numbers of responses to VX, GX, and ATX were combined from those protocols where these blockades were applied as a first pharmacological manipulation: n_VX = n_VX + n_LX+VX; n_GX = n_ADX+GX + n_ADX+GX+VX + n_{ADX+GX+VX+ATX} + n of respective controls; n_ATX = n_ATX + n of respective control (see Table 2). The small changes in mean arterial pressure and iliac vascular conductance caused by VX were allowed to return spontaneously toward resting levels, whereas the large, sustained changes in these variables caused by ATX and GX were compensated via intravenous infusion of phenylephrine.

^*P < 0.05 vs. zero.

Table 2.

Infusion rates of phenylephrine used to maintain mean arterial pressure and iliac vascular conductance at preblockade levels in experimental and respective time control groups in which ATX and/or GX were performed

Protocol Number	Experimental Procedure	n	Infusion Rate, ml·h−1·kg−1	Time Controls	n	Infusion Rate, ml·h−1·kg−1
2	ATX	10	3.18±0.12*	ATX	5	2.39±0.25*
3	ADX + GX	8	3.02±0.24*	ADX + GX	5	3.20±0.33*
4	ADX + GX + VX	8	3.45±0.4*	ADX + GX + VX	5	3.94±0.62
5	ADX + GX + VX + ATX	8	6.33±0.24	ADX + GX + VX + ATX	5	5.29±0.62

Values are means ± SE; n = number of rats. Abbreviations are as in Fig. 1. There were no differences in infusion rates between experimental groups versus respective time control groups (P > 0.05).

^*P < 0.05 vs. ADX + GX + VX + ATX.

Fig. 3.

An example of time control experiment following ADX + GX + VX + ATX; no CPA was microinjected in this experiment. The dashed arrow denotes a potential microinjection of CPA and the subsequent 20-min integration of the response. Note that although the infusion of PE did not fully compensate for the decrease in mean arterial pressure (MAP), the iliac vascular conductance (IVC) gradually declined constricting iliac vasculature slightly below the baseline level.

Vasopressin Assay

Since the hemodynamic experiments (protocols 1–6) suggested that vasopressin plays a dominant role in the iliac vascular responses to stimulation of NTS A₁ adenosine receptors, in an additional group of animals the effect of microinjections into the NTS of CPA (n = 8) or respective volume control (50 nl of ACF; n = 5) on circulating vasopressin was evaluated. We compared the levels of plasma vasopressin measured 30 min before and ∼8 min after the microinjections (the average time when maximal hemodynamic responses to stimulation of NTS A₁ receptors occur). Arterial blood samples (∼1 ml) were slowly withdrawn from the femoral artery into prechilled, heparinized tubes. Blood volume was kept unchanged via simultaneous infusion of the same volume of donor blood into the femoral vein. The samples were immediately placed on ice and centrifuged at 5,000 g for 10 min at 4°C. Plasma was collected and stored at −70°C. Plasma vasopressin concentration was assessed via standard radioimmunoassay procedures in our laboratory as described previously (13, 24, 27). The sensitivity of the vasopressin assay was 0.1 pg/ml and 50% displacement was 4.1 pg/tube. Intra- and interassay variability was 7.0% and 13.4%, respectively.

Data Analysis

Hemodynamic responses were analyzed over a 20-min period following the microinjections, similar to a previous study from our laboratory (21). The responses were quantified as an integration of the differences between the baseline and response values averaged in 1-min periods and summed for 20 min of the response, i.e., when the majority of the responses occur. The integral reflects the predominant trend of the responses despite transient, sometimes large, bidirectional fluctuations in each variable. Because hemodynamic effects evoked by stimulation of NTS A₁ adenosine receptors were variable, often biphasic, or even polyphasic, as we previously reported (21, 35), we used the integral values for the comparison between the experimental groups. The absolute values of blood flow depend to some extent on positioning of the probe around the iliac artery; therefore, the comparison between the relative changes in MAP, iliac blood flow (IBF), and IVC was more reliable. The HR responses, calculated from pulse intervals through the flow probe, were expressed in absolute values (beats/min). IVC was calculated by dividing IBF, expressed as a Doppler shift (in Hz), by MAP (in mmHg). In experimental protocols 2-5, where PE was infused to compensate for the decreased MAP and increased IVC, the direct effect of PE on baseline hemodynamic variables was evaluated in respective time control experiments and subtracted from experimental data. Specifically, changes occurring in each variable during time-control experiments were integrated for 20 min and subtracted from the respective 20-min integral values obtained in each animal of the experimental groups (protocols 2–5).

One-way ANOVA for independent measures was used to compare hemodynamic responses versus experimental conditions. Differences observed were further evaluated by t-test with Bonferroni adjustment for independent measures. Differences between circulating vasopressin levels measured before and after microinjections of ACF or CPA were evaluated using paired t-test; the differences in vasopressin levels between the groups (ACF vs. CPA) were evaluated using unpaired t-test. The changes in all recorded variables were also compared with zero by means of SYSTAT univariate F test. An α-level of P < 0.05 was used to determine statistical significance.

RESULTS

Resting values of MAP, HR, IBF, and IVC for each experimental group, measured just before stimulation of NTS A₁ adenosine receptors, are presented in Table 3. The resting MAP and IVC for all the groups where blockades were performed were not different from those for intact animals; this provided reliable comparison between the experimental protocols. The direct effects of ganglionic blockade, and blockade of V₁ vasopressin and AT₁ angiotensin II receptors on all hemodynamic variables, are presented in Table 1. The large decreases in MAP and increases in IVC evoked by blockade of AT₁ angiotensin II receptors and/or ganglionic blockade required additional compensation with PE, whereas the small decreases in MAP and increases in IVC evoked by V₁ vasopressin receptor blockade were allowed to partially recover without compensation.

Table 3.

Resting values of hemodynamic parameters in each experimental group

Protocol Number	Experimental Procedure	n	Mean Arterial Pressure, mmHg	Heart Rate, beats/min	Iliac Blood Flow, Hz	Iliac Vascular Conductance, Hz/mmHg
	Intact	13	98.1±.1	353.9±4.7	1030.2±3.6	10.9±3.3
1	VX	8	105.3±4.9	359.7±11.2	1091.0±204.2	10.6±2.0
2	ATX	10	95.3±4.9	342.2±8.2	992.8±122.5	10.8±1.5
3	ADX + GX	8	93.1±5.1	383.6±11.4	1190.4±167.2	13.2±2.1
4	ADX + GX + VX	8	88.7±1.3	383.3±9.0	1394.0±222.7	15.8±2.7
5	ADX + GX + VX + ATX	8	88.1±2.1	405.0±10.9*	1271.0±102.9	14.4±1.1
6	LX + VX	8	95.8±4.1	340.6±7.0	1336.4±132.8*	14.3±1.7
(2)	Control (ATX)	5	99.3±2.8	330.0±11.8	973.7±173.2	10.0±2.0
(3)	Control (ADX + GX)	5	84.2±2.8	365.8±17.2	1073.1±144.7	12.8±1.8
(4)	Control (ADX + GX + VX)	5	88.1±3.5	381.6±7.8	1121.6±141.1	12.7±1.4
(5)	Control (ADX + GX + VX + ATX)	5	90.2±4.7	398.9±13.3	1144.6±302.0	12.2±2.4

Values are means ± SE; n = number of rats. Numbers in parentheses show time controls for respective protocols. Abbreviations are as in Fig. 1.

^*P < 0.05 vs. Intact.

Effects of V₁, AT₁, and Ganglionic Blockades on Responses to Stimulation of NTS A₁ Adenosine Receptors

Figure 4 presents examples of the responses evoked by selective stimulation of NTS A₁ adenosine receptors, which were observed most often under each experimental condition. The average integral responses for each experimental group are presented in Fig. 5. In intact animals the typical variability in the responses to stimulation of NTS A₁ adenosine receptors was observed: the pressor and vasoconstrictor responses prevailed (Table 4 and Figs. 4 and 5), although biphasic, polyphasic, or, more rarely, depressor and vasodilatory responses were also observed. As we previously demonstrated, these patterns and variability of the responses reflected counteracting effects of β-adrenergic vasodilation versus sympathetic and humoral vasoconstriction (21, 35). V₁ vasopressin receptor blockade reversed iliac vasoconstrictor responses observed in the intact group into slight iliac vasodilation (P = 0.0001 vs. intact). Blockade of angiotensin II AT₁ receptors alone did not significantly alter the responses compared with the intact group (P > 0.05 for all variables). Elimination of adrenal and sympathetic neural effects on the iliac vasculature (protocol 3) increased the iliac vasoconstrictor responses almost fourfold compared with the intact group, indicating that other humoral factor(s) different than circulating norepinephrine play a crucial role in the iliac vasoconstrictor responses. Subsequent blockade of vasopressin V₁ receptors (protocol 4) virtually abolished the exaggerated iliac vasoconstriction observed following adrenalectomy plus ganglionic blockade alone (protocol 3; Figs. 4 and 5), indicating that the humoral iliac vasoconstriction evoked by stimulation of NTS A₁ adenosine receptors is mediated mostly via the release of vasopressin. Combined blockade of neural and all considered humoral factors (adrenalectomy + ganglionic + V₁ vasopressinergic + AT₁ angiotensinergic blockades; protocol 5) had very similar effects on the responses to that observed in protocol 4 (adrenalectomy + ganglionic + V₁ vasopressinergic blockades); there were no significant differences between these two groups with respect to MAP, HR, and IVC responses (P > 0.05 for all comparisons). The lack of differences between responses observed in protocols 4 and 5 additionally confirmed that vasopressin is the dominant humoral vasoconstrictor factor triggered by stimulation of NTS A₁ adenosine receptors. The analysis of absolute values of the integral responses presented in Table 5 leads to the same conclusions as those based on the relative responses (Fig. 5). The absolute values of the responses, presented in Table 5, show residual iliac vasoconstrictor effects in protocols 4 and 5. However, this vasoconstriction was abolished when the respective time control values were subtracted from the direct experimental values presented in Table 5.

Fig. 4.

MAP, iliac blood flow (IBF), and IVC responses to microinjection of adenosine A₁ receptor agonist (CPA; 330 pmol/50 nl) into the subpostremal NTS in intact rats (INT) and each experimental group. Abbreviations are as in Fig. 1. Microinjections of CPA, marked by vertical arrows, were applied ∼10–20 min after the blockades, when baseline levels of all variables stabilized (see Fig. 2). Note that VX reversed pressor and vasoconstrictor responses most often observed in Int group into depressor and vasodilatory responses. ADX + GX exaggerated the pressor and vasoconstrictor response. These exaggerated vasoconstrictor responses were virtually abolished following ADX + GX + VX and ADX + GX + VX + ATX.

Fig. 5.

Integral responses of MAP, IBF, and IVC evoked by microinjections of CPA (330 pmol/50 nl) into the caudal subpostremal NTS. Abbreviations are as in Fig. 1. Data are means ± SE. In groups ATX, ADX + GX, ADX + GX + VX, and ADX + GX + VX + ATX, respective time control values were subtracted. *Different vs. intact group (P < 0.05); #different vs. zero (P < 0.05). VX reversed iliac vasoconstrictor responses observed in intact group into vasodilation and virtually abolished exaggerated vasoconstrictor responses observed following ADX + GX.

Table 4.

Number of individual experiments where overall increments or decrements were observed for each recorded hemodynamic parameter based on its integral values

Protocol Number	Experimental Procedure	n	Mean Arterial Pressure, mmHg		Heart Rate, beats/min		Iliac Blood Flow, Hz		Iliac Vascular Conductance, Hz/mmHg
			Incr	Decr	Incr	Decr	Incr	Decr	Incr	Decr
	Intact	13	10	3	1	12	1	12	1	12
1	VX	8	0	8	0	8	3	5	7	1
2	ATX	10	9	1	4	6	2	8	0	10
3	ADX + GX	8	8	0	1	7	0	8	0	8
4	ADX + GX + VX	8	5	3	2	6	0	8	1	7
5	ADX + GX + VX + ATX	8	7	1	2	6	1	7	0	8
6	LX + VX	8	1	7	1	7	5	3	8	0

Number of increments (Incr) and decrements (Decr) observed in each experimental group. Abbreviations are as in Fig. 1.

Table 5.

Absolute values of integral changes in mean arterial pressure, heart rate, illiac blood flow, and illiac vascular conductance evoked by microinjections of 330 pmol in 50 nl N⁶-cyclopentyladenosine into the nucleus tractus solitarii in each experimental group

Protocol Number	Experimental Procedure	n	Mean Arterial Pressure, mmHg	Heart Rate, beats/min	Iliac Blood Flow, Hz	Iliac Vascular Conductance, Hz/mmHg
	Intact	13	86.0±37.7†	−379.3±96.4†	−1480.3±588.7†	−23.1±7.1†
1	VX	8	−195.2±63.9*†	−411.9±99.3†	59.6±784.5	20.8±9.3*†
2	ATX	10	198.4±47.3†	−70.6±126.0	−1585.8±445.4†	−34.8±6.4†
3	ADX + GX	8	312.3±67.1*†	−139.0±48.6†	−5758.2±1284.4*†	−91.6±20.2*†
4	ADX + GX + VX	8	7.1±46.3	−237.0±137.5	−2752.3±715.7†	−31.3±11.5†
5	ADX + GX + VX + ATX	8	74.9±19.0†	−249.8±134.8	−1784.9±447.1†	−30.9±5.5†
6	LX + VX	8	−162.0±71.5*†	−502.0±177.3†	2127.6±1489.5*	57.5±11.4*†
(2)	Control (ATX)	5	80.0±37.6	504.1±91.8*†	−303.2±573.6	−9.8±3.6†
(3)	Control (ADX + GX)	5	22.3±36.2	54.5±104.8*	−166.1±364.4	−5.9±3.8
(4)	Control (ADX + GX + VX)	5	15.6±39.7	163.3±54.5*†	−1224.1±380.7†	−13.8±5.1†
(5)	Control (ADX + GX + VX + ATX)	5	−11.4±23.7	206.5±61.6*†	−2694.9±1428.2	−24.7±13.0

Values are means ± SE. Numbers in parentheses show time controls for respective protocols. Abbreviations are as in Fig. 1.

^*P < 0.05 vs. intact;

^†P < 0.05 vs. zero.

Data collected in protocols 1–5 (Figs. 4 and 5) strongly suggested that the only significant humoral vasoconstrictor contributing to iliac vascular responses evoked by selective stimulation of NTS A₁ adenosine receptors is vasopressin. Therefore, in protocol 6 vasopressin and lumbar sympathetic vasoconstrictor components of the responses were eliminated to unmask β-adrenergic vasodilation alone, not opposed by neural and humoral vasoconstriction. Following V₁ vasopressin receptor blockade combined with bilateral lumbar sympathectomy (protocol 6), the iliac vasodilation response doubled compared with that observed following V₁ vasopressin receptor blockade alone (405.1 ± 58.4 ∫Δ% vs. 195.8 ± 51.2 ∫Δ%; P = 0.0175).

The above observations are supported by comparison of the frequency of which we observed increases versus decreases in hemodynamic variables in each experiment for all experimental groups (Table 4). In intact animals, decreases in IVC prevailed, whereas following V₁ vasopressinergic blockade the increases in IVC prevailed. Combined V₁ vasopressinergic blockade and lumbar sympathectomy completely eliminated iliac vasoconstrictor responses (protocol 6). AT₁ angiotensinergic blockade increased the number of pressor and vasoconstrictor events compared with the intact group (Table 4), although there were no significant differences between the averaged responses observed in this versus intact groups (Fig. 5). Adrenalectomy plus ganglionic blockade completely eliminated the depressor and iliac vasodilatory responses as expected (protocol 3). When V₁ vasopressinergic blockade was added to adrenalectomy and ganglionic blockade (protocol 4) or all the blockades were performed in adrenalectomized animals (protocol 5), again variable vasoconstrictor/vasodilatory responses were observed; however, this variability reflected rather random variations of IVC since vascular responses were virtually abolished in these groups (changes in IVC were not different from zero; Fig. 5). Prevailing pressor and depressor responses were consistent with prevailing vasoconstrictor and vasodilatory responses in each experimental group. However, the decreases in HR dominated in all experimental groups despite the different experimental conditions.

Release of Vasopressin in Response to Stimulation of NTS A₁ Adenosine Receptors

The above results, obtained via pharmacological blockade approaches, strongly suggested that vasopressin is a dominant vasoconstrictor factor released into the circulation following the selective stimulation of NTS A₁ adenosine receptors. Therefore, in an additional two groups of animals, we tested this hypothesis more directly by measuring whether the plasma levels of vasopressin indeed increase as a result of stimulation of NTS A₁ adenosine receptors (Fig. 6). The resting levels of vasopressin were similar in both groups (P = 0.931), and they were moderately elevated compared with vasopressin levels normally observed in intact conscious animals (7, 13, 15, 23, 24), consistent with levels associated with anesthesia and surgical stress (7, 15). Stimulation of NTS A₁ adenosine receptors increased circulating vasopressin levels over fourfold (P = 0.0006). Microinjection of ACF into the NTS also tended to increase the level of circulating vasopressin; however, the increase did not reach statistical significance (P = 0.083). Notably, the difference between vasopressin levels measured across the groups, i.e., following microinjections of CPA versus ACF, was also significant (P = 0.041).

Fig. 6.

Comparison of plasma vasopressin levels measured before (pre) and after (post) microinjections into the NTS of vehicle (ACF; n = 5) or selective A₁ adenosine receptor agonist (CPA; n = 8). *Differences between pre- vs. postmicroinjections (P < 0.05); #differences between the groups (P < 0.05). CPA increased circulating vasopressin levels over 4-fold, whereas ACF did not significantly increase the level of circulating vasopressin.

DISCUSSION

The present study assessed the relative roles of three major humoral vasoconstrictor factors in mediating the responses to stimulation of A₁ adenosine receptors located in the NTS. The release of humoral vasoconstrictor factor(s) had been implied by a previous study from our laboratory, which showed that the large iliac vasoconstriction, triggered by stimulation of NTS A₁ adenosine receptors, persisted after bilateral lumbar sympathectomy and adrenalectomy (21). The major finding of the present study is that vasopressin is released into the circulation upon stimulation of NTS A₁ adenosine receptors and that it is the primary humoral factor contributing to the iliac vasoconstriction. The potential release of norepinephrine and angiotensin II did not contribute significantly to the iliac vascular responses.

NTS A₁ Adenosine Receptors Are Involved in Control of Vasopressin Release

Our previous studies strongly suggested that A₁ adenosine receptors may modulate vasopressin release at the level of the NTS similarly as they modulate baroreflex control of regional sympathetic outputs and HR (30, 35). However, with the consideration of differential localization/expression of adenosine receptor subtypes on functionally different NTS neurons (31, 36), it remained unknown whether A₁ adenosine receptors are indeed located on these baroreflex neurons controlling vasopressin release. The present study showed that activation of A₁ adenosine receptors in the NTS disinhibits vasopressin release. Therefore, pre- and/or postsynaptic A₁ adenosine receptors are likely located on those NTS baroreflex neurons/terminals, which control vasopressin release. This hypothesis based on integrative, systemic data may be confirmed in future studies at the cellular level.

The most likely mechanism responsible for the release of vasopressin into the circulation in response to stimulation of NTS A₁ adenosine receptors is the inhibition of baroreflex transmission in the NTS resulting in the disinhibition of vasopressin release from hypothalamic nuclei (paraventricular and supraoptic) (10). The role of NTS in baroreflex control of vasopressin release has been well established (29, 39). Bilateral inhibition, anesthetization, or lesioning of the NTS, which removes baroreflex mechanisms, results in 10-fold increases of circulating vasopressin levels and vasopressin-mediated hypertension (39). Sinoaortic baroreceptor denervation prevents this response (29). In the present study the unilateral inhibition of NTS baroreflex mechanisms via unilateral activation of A₁ adenosine receptors resulted in over fourfold increases in circulating vasopressin.

Relative Roles of Vasoactive Factors Triggered by Activation of A₁ Adenosine Receptors in the NTS

A previous study from our laboratory showed that activation of A₁ adenosine receptors in the NTS evokes neural and humoral iliac vasoconstriction opposed by β-adrenergic vasodilation (21). The counteracting vascular effects contributed to the variability of the overall pressor/depressor responses with prevailing vasoconstriction and increases in MAP. These conclusions were based on comparison of iliac vascular responses observed in intact animals and following β-adrenergic blockade, bilateral adrenalectomy, lumbar sympathectomy, and combined adrenalectomy plus lumbar sympathectomy. The present study investigated potential humoral vasoconstrictors (vasopressin, angiotensin II, and norepinephrine) by assessing the effects of vasopressin V₁ receptor and angiotensin AT₁ receptor blockade as well as ganglionic blockade performed in adrenalectomized animals. Perhaps the most compelling results were that the blockade of peripheral V₁ vasopressin receptors reversed the normal iliac vasoconstriction into marked vasodilation. This vasodilation was further enhanced by adding lumbar sympathectomy to the V₁ receptor blockade. It is likely that sympathetic nerves have a smaller role than vasopressin in mediating this iliac vasoconstriction since lumbar sympathectomy alone, performed in a previous study from our laboratory (21), had a relatively smaller effect on the IVC responses then V₁ vasopressinergic blockade alone, performed in the present study. Collectively, these data indicate that vasopressin alone can override epinephrine-induced iliac vasodilation and induce vasoconstriction, but that sympathetic nerves alone do not. The roles of vasopressin and sympathetic nerves appear additive since the sum of the individual effects approximate that of the combined effects. Vasopressin together with increases in LSNA can induce significant vasoconstriction without which a powerful β-adrenergic vasodilation is revealed. We conclude that the variability often seen in the arterial pressure response to NTS A₁ adenosine receptor stimulation is the direct result of these push-pull counteracting mechanisms.

Circulating norepinephrine potentially released upon stimulation of NTS A₁ adenosine receptors likely has little role since ganglionic blockade did not lessen the intense iliac vasoconstriction seen after bilateral adrenalectomy (21). Probably the amount of norepinephrine released from sympathetic terminals was too small to exert measureable effects beyond that of circulating norepinephrine already likely to be elevated as a result of anesthesia and surgical stress. A previous study from our laboratory showed that pre-ASNA increased much more than RSNA and LSNA in response to stimulation of NTS A₁ adenosine receptors (35). It is possible that other sympathetic outputs responded even less or might even be inhibited with the stimulation. This could cause the increase of circulating norepinephrine in the present study to be functionally irrelevant.

Angiotensin II blockade alone markedly decreased baseline MAP and increased IVC. However, the normal pressor and vasoconstrictor responses to microinjections of CPA did not decrease but rather tended to increase. This indicates that angiotensin II, similarly as circulating norepinephrine, does not contribute to the iliac vasoconstriction elicited by stimulation of NTS A₁ adenosine receptors. Although increases of RSNA may cause an increase in the release of renin from the kidney (11), the total time from activation of the renal nerve to the subsequent release of renin and conversion of angiotensinogen to angiotensin I and then to angiotensin II may have exceeded the time of the analyzed responses. In addition, losartan used to block AT₁ receptors in this study most likely crossed the blood-brain barrier and might evoke central effects counteracting the peripheral actions. Nevertheless, our data show that circulating angiotensin II had no functional effect as a potential humoral iliac vasoconstrictor triggered by stimulation of NTS A₁ adenosine receptors.

Vasoactive Effects of NTS A₁ versus A_2a Adenosine Receptor Subtypes

The predominately pressor but often variable responses to selective stimulation of NTS A₁ adenosine receptors are mediated most likely via inhibition of baroreflex glutamatergic transmission in the NTS and resetting of baroreflex functions toward higher MAP (30, 35). In contrast, selective stimulation of NTS A_2a adenosine receptors evokes decreases in MAP and preferential iliac vasodilation mediated via nonbaroreflex, nonglutamatergic mechanism(s) (16, 33, 34). These responses were mediated mostly via activation of the adrenal medulla and β-adrenergic mechanism and to a lesser extent via lumbar sympathetic nerves (19). However, no other humoral mechanisms were involved since bilateral adrenalectomy and lumbar sympathectomy abolished the responses (19). How do these two adenosine receptor subtypes contribute to the responses mediated by adenosine operating in the NTS under physiological (9, 38) or pathological (25, 37, 43, 45) conditions? It is well known that microinjections of adenosine into the NTS result in depressor responses similar to those observed following selective activation of A_2a adenosine receptors (3, 22, 40). However, both adenosine receptor subtypes may contribute to the depressor responses since they both activate the adrenal medulla and facilitate β-adrenergic vasodilation via baroreflex and unknown, nonbaroreflex mechanisms (33, 35). Other components of the responses elicited by NTS A_2a receptors (decreases in RSNA and HR) may further contribute to the depressor responses, whereas A₁ receptor-mediated sympathoactivation and vasopressin release would oppose these responses. Under physiological conditions adenosine may contribute to the pressor effects evoked by stimulation of the hypothalamic defense area (9, 38) via A₁ receptor-mediated inhibition of baroreflex mechanisms and resulting sympathoactivation and vasopressin release (30, 35). In contrast, under pathological conditions naturally released adenosine, acting via both A₁ and A_2a adenosine receptor subtypes, may contribute to the paradoxical sympathoinhibition and cardiac slowing observed during hypovolemic phase of hemorrhagic shock (37).

Limitations of the Method

In the central nervous system adenosine is not released synaptically but rather produced in the extracellular space via the degradation of ATP, which is released at synaptic terminals from neurons as well as released from glial cells activated by extracellular glutamate diffused from nearby active nerve terminals (6, 8, 14, 46). In addition, under pathological conditions adenosine is released into the extracellular space in a global, nonsynaptic manner from hypoxic/hypoperfused neurons and glial cells (25, 43, 45). Therefore, the natural, spatial (not strictly synaptic) action of adenosine in the NTS may be simulated well via microinjections of selective agonists of adenosine receptor subtypes. Importantly, although adenosine is released nonselectively in the NTS, it does produce specific differential responses in regional sympathetic outputs and vascular beds as shown by numerous studies from our laboratory (4, 30, 32–35). We believe that these specific, differential effects evoked by nonspecific, spatial activation of adenosine receptors result from differential location/expression of adenosine receptor subtypes on NTS neurons and synaptic terminals controlling different sympathetic outputs and involved in different mechanisms integrated in the NTS (31, 36). To confirm this hypothesis, based on systemic, integrative approaches, further studies at the cellular level are required.

The experiments were performed in anesthetized animals, and these conditions most likely attenuated baroreflex mechanisms, which may have contributed to the increased vasopressin levels. In the conscious rat circulating the vasopressin level is ∼2 pg/ml (7, 13, 15, 23, 24, 27). With even limited surgery under anesthesia, baseline vasopressin rises to ∼6–10 pg/ml (7, 15, 29, 39). In our study after extensive surgery under anesthesia, baseline vasopressin levels were ∼25 pg/ml. However, despite the increased baseline vasopressin levels, activation of NTS A₁ adenosine receptors triggered marked release of vasopressin into the circulation (Fig. 6), and a powerful V₁ receptor-mediated vasoconstrictor effect in the hindlimb vasculature was apparent (Fig. 5). In contrast, no functional effects of potential, A₁ adenosine receptor-mediated increases in circulating norepinephrine or angiotensin II were observed.

Glucocorticoid deficiency, which may be observed following chronic adrenalectomy, increases circulating levels of vasopressin (18). However, this effect seems to be mediated at the level of transcription of the vasopressin gene and not an immediate release (18). Therefore, in the animals in which the adrenal glands were removed acutely, as in the present study, increased resting levels of vasopressin were most likely due to anesthesia and surgical stress.

The large and sustained changes in resting hemodynamic variables following AT₁ angiotensinergic and ganglionic blockades required compensation with intravenous infusions of phenylephrine to make the relative responses comparable across the experimental groups. It was extremely difficult to return both MAP and IVC to the preblockade levels. Since we tried to compensate most accurately for IVC, the resting MAP, preceding the microinjections into the NTS in the groups where ganglionic blockade was combined with V₁ vasopressinergic and with AT₁ angiotensinergic blockades (protocols 4 and 5), was slightly lower compared with other experimental groups. The long lasting compensatory infusions of phenylephrine in protocols 2–5 could facilitate central baroreflex mechanisms (17). Therefore, the inhibition of enhanced baroreflex activity via stimulation of NTS A₁ adenosine receptors could result in relatively greater pressor and vasoconstrictor responses compared with those observed in intact animals and following V₁ vasopressinergic blockade alone (protocol 1) and V₁ vasopressinergic blockade combined with lumbar sympathectomy (protocol 6). In fact, pressor and vasoconstrictor responses tended to increase following AT₁ receptor blockade although this tendency did not reach statistical significance. The pressor and vasoconstrictor responses observed after adrenalectomy combined with ganglionic blockade (protocol 3) were not different from those observed following adrenalectomy alone in a previous study from our laboratory (21). Finally, these responses (even if slightly exaggerated) were virtually abolished by the vasopressinergic blockade (protocols 4 and 5). Therefore, the potential central effects of infusions of phenylephrine on responses to stimulation of NTS A₁ adenosine receptors were likely negligible and did not affect the primary conclusions of the present study. The infused phenylephrine activated α₁-adrenergic receptors located on vascular smooth muscles, including those in the iliac vascular bed. The occupation of α₁-adrenergic receptors by an exogenous agonist could mask the effect of circulating endogenous norepinephrine, which may be potentially increased following stimulation of NTS A₁ adenosine receptors. The peripheral interactions between simultaneously activated V₁, α₁-_, and AT₁ receptors could further complicate the comparison of relative roles of humoral vasoconstrictor factors, obscuring the smaller ones. Nevertheless, the marked effects of vasopressin V₁ receptor blockade were evident in experimental groups with and without phenylephrine compensation (protocols 1 and 6 vs. protocols 4 and 5). This indicates that vasopressin contribution to the iliac vasoconstriction was much larger than the contribution (if any) of other potential vasoconstrictors (norepinephrine and/or angiotensin II).

Conclusion

Selective stimulation of NTS A₁ adenosine receptors triggers the release of vasopressin into the circulation, strongly suggesting that A₁ adenosine receptors are likely located on afferent terminals and/or NTS interneurons mediating baroreflex control of vasopressin release. Vasopressin is a major humoral vasoconstrictor factor contributing to the iliac vascular responses. The natural variability of MAP and IVC responses to stimulation of NTS A₁ adenosine receptors observed in intact animals is a result of the simultaneous triggering of sympathetic and vasopressinergic vasoconstriction counteracted by β-adrenergic vasodilation resulting mainly from epinephrine release in response to the large increase in adrenal preganglionic sympathetic activity.

GRANTS

This study was supported by National Heart, Lung, and Blood Institute Grant HL-67814 (to D. S. O'Leary and T. J. Scislo) and a Merit Review Award by Department of Veterans Affairs (to N. F. Rossi and T. J. Scislo).