GABA and glycine receptors in the nucleus ambiguus mediate tachycardia elicited by chemical stimulation of the hypothalamic arcuate nucleus

Abstract

We have previously reported that stimulation of the hypothalamic arcuate nucleus (ARCN) by microinjections of N-methyl-d-aspartic acid (NMDA) elicits tachycardia, which is partially mediated via inhibition of vagal inputs to the heart. The neuronal pools and neurotransmitters in them mediating tachycardia elicited from the ARCN have not been identified. We tested the hypothesis that the tachycardia elicited from the ARCN may be mediated by inhibitory neurotransmitters in the nucleus ambiguus (nAmb). Experiments were done in urethane-anesthetized, artificially ventilated, male Wistar rats. In separate groups of rats, unilateral and bilateral microinjections of muscimol (1 mM), gabazine (0.01 mM), and strychnine (0.5 mM) into the nAmb significantly attenuated tachycardia elicited by unilateral microinjections of NMDA (10 mM) into the ARCN. Histological examination of the brains showed that the microinjections sites were within the targeted nuclei. Retrograde anatomic tracing from the nAmb revealed direct bilateral projections from the ARCN and hypothalamic paraventricular nucleus to the nAmb. The results of the present study suggest that tachycardia elicited by stimulation of the ARCN by microinjections of NMDA is mediated via GABA_A and glycine receptors located in the nAmb.

the hypothalamic arcuate nucleus (ARCN), located bilaterally at the base of the caudal two-thirds of the third ventricle, plays an important role in central cardiovascular regulation in normal and pathological states (41). We have recently shown that microinjections of N-methyl-d-aspartic acid (NMDA) into the ARCN elicited an increase in heart rate (HR) in rats regardless of whether the baroreceptors were intact or denervated (26, 27, 36). Tachycardic responses elicited from the ARCN are mediated via the release of glutamate and α-melanocyte-stimulating hormone (α-MSH) in the hypothalamic paraventricular nucleus (PVN) because blockade of ionotropic glutamate receptors (iGLURs) and melanocortin 3/4 receptors in the ipsilateral PVN attenuated these responses (27). In this context, it may be noted that proopiomelanocortin-immunoreactive cells are present in the ARCN (27). Proopiomelanocortin is the precursor of α-MSH, and a subpopulation of proopiomelanocortin neurons in the ARCN is glutamatergic (29, 33).

The increases in HR elicited by NMDA-induced stimulation of the ARCN are partially mediated via activation of sympathetic input to the heart (36). The activation of sympathetic input to the heart, after the stimulation of the ARCN, is mediated via projections of the PVN to the intermediolateral cell column of the spinal cord (IML) or projections of the PVN to the rostral ventrolateral medullary pressor area (RVLM), which, in turn, projects to the IML (14, 19, 36, 39, 40). Glutamate has been implicated as the neurotransmitter in the IML because blockade of iGLURs in the IML attenuated tachycardic responses elicited from the ARCN (36).

Tachycardic responses elicited NMDA-induced stimulation of the ARCN are also partially mediated via inhibition of the vagal input to the heart (36). The neuronal pools mediating the inhibition of vagal input to the heart in response to ARCN stimulation and the neurotransmitters and receptors involved in these neuronal pools have not been identified. Although cardiac vagal neurons (CVNs) in the rat are located in the dorsal motor nucleus of the vagus as well as the nucleus ambiguus (nAmb), the predominant vagal input to the heart is provided by the CVNs located in the nAmb (13, 43, 44). The axons of CVNs located in the nAmb pass through the vagus nerves and synapse on cardiac ganglia located in the atrioventricular node and fat tissue surrounding the sinoatrial node (2). Typically, CVNs located in the nAmb are silent and need excitatory and inhibitory synaptic inputs for their activation and inhibition, respectively (34, 35). CVNs in the nAmb are known to receive GABA (46, 48) and glycinergic (48) inhibitory inputs. Based on this information, we hypothesized that GABA and glycine may be released in the nAmb in response to ARCN stimulation, causing an inhibition of CVNs via GABA and glycine receptors, respectively. Consequently, the parasympathetic input to the heart is decreased, resulting in tachycardia. This hypothesis was tested in the present study.

METHODS

General procedures.

Adult male Wistar rats (Charles River Laboratories, Wilmington, MA) weighing 300–360 g were used. A total number of 69 rats was used. The distribution of rats in different groups was as follows: microinjections of NMDA into the ARCN (n = 9), inhibition of nAmb neurons using muscimol (n = 15), blockade of GABA_A receptors in the nAmb using gabazine (n = 17), blockade of glycine receptors using strychnine in the nAmb (n = 10), combined blockade of GABA_A and glycine receptors in the nAmb (n = 5), blockade of iGLURs in the nAmb (n = 7), and retrograde tracing experiments (n = 6).

Animals were housed under controlled conditions with a 12:12-h light-dark cycle. Food and water were available to animals ad libitum. The experimental protocols, designed according to the National Institutes of Health Guide for the Care and Use of Laboratory Animals, were reviewed and approved by the Institutional Animal Care and Use Committee of Rutgers New Jersey Medical School.

Details of all procedures used in the present study have been reported in our previous publications (1, 6, 8–10, 24–27, 36). Rats were initially anaesthetized with inhalation of isoflurane (2–3% in 100% O₂) using a vaporizer (Fluotec-3, Cyprane) to cannulate the trachea and the femoral vein and artery on one side. Isoflurane administration was then continued via the trachea using a rodent ventilator (model 683, Harvard Apparatus, Holliston, MA). A solution of urethane (800 mg/ml) was then injected intravenously in six to nine aliquots (each 0.05–0.1 ml containing 40–80 mg urethane) at 2-min intervals. The inhalation of isoflurane was discontinued after the administration of four to five aliquots of urethane. The injection of urethane was completed within 12–18 min. The total dose of urethane injected was 1.2–1.4 gm/kg iv. Adequate depth of anesthesia, indicated by the absence of an increase in blood pressure (BP) and/or withdrawal of the limb in response to pinching of a hindpaw, was tested periodically throughout the experiment. End-tidal CO₂ was measured continuously from expired gases using an infrared CO₂ analyzer modified for use in small animals (Micro-Capnometer, Columbus Instruments, Columbus, OH). The frequency and tidal volume were adjusted on the ventilator to maintain end-tidal CO₂ at 3.5–4.5%. Rectal temperature was monitored using a rectal probe (RET-1) connected to a temperature controller (model TCAT-2A, Physitemp Instruments, Clifton, NJ) and was maintained at 37 ± 0.5°C. BP and HR were recorded on a computer hard drive using a 1401 plus analog-to-digital converter and Spike 2 software (Cambridge Electronic Design, Cambridge, UK).

Microinjections into the ARCN.

Rats were placed in a prone position in a stereotaxic instrument (David Kopf Instruments, Tajunga, CA) with a bite bar 18 mm below the interaural line. A hole (8–10 mm in diameter) was drilled in the midline at the junction of the two parietal bones caudal to the bregma. Our experiments required microinjections into the ARCN and nAmb in the same animal. Therefore, a double-barrel glass micropipette (tip size: 20–40 μm), used for microinjections into the ARCN, was mounted on a micromanipulator (David Kopf Instruments) and fixed at an acute angle (22°) pointing caudally. The barrels in the micropipette were connected to a picospritzer (General Valve, Fairfield, NJ). The glass micropipettes were inserted into the brain tissue through the previously made hole in the parietal bones on either side of the midline. To approach the ARCN, the following coordinates were used: 1.72–4.36 mm caudal to the bregma, 0.2–1 mm lateral to the midline, and 8.8–10.1 mm ventral to the dura. Using this approach, the tip of the micropipette reached the ARCN at the following coordinates: 1.92–3.72 caudal to the bregma, 0.2–0.6 lateral to the midline, and 9.6–10.1 ventral to the dura. Using a modified binocular horizontal microscope (model PZMH, WPI) with a graduated reticule in one eyepiece, the volume of microinjection was visually confirmed by the displacement of fluid meniscus in the micropipette barrel. Microinjections of NMDA (10 mM) were used to identify the ARCN (1, 11, 26–28, 36). In this and other series of experiments, unless indicated otherwise, the duration of microinjections was 5–10 s, and microinjections of artificial cerebrospinal fluid (aCSF; pH 7.4) were used as controls. As described above, our experiments required microinjections into the ARCN and nAmb in the same animal. Therefore, the interval between microinjections into the ARCN and nAmb was at least 20 min to allow the HR response elicited from the ARCN to return to the basal level before the pharmacological manipulations in the nAmb were attempted.

Microinjections into the nAmb.

Microinjections into the nAmb were made using a dorsal approach (6, 8–10). The dorsal neck muscles and part of the occipital bone were removed, and the medulla was exposed by incisions in the atlantooccipital membrane and dura. Four-barrel micropipettes (tip size: 20–40 μm) were lowered into the nAmb perpendicularly. The coordinates for the nAmb were as follows: 0.12 caudal to 0.64 mm rostral with reference to the calamus scriptorius (CS), 1.8–2 mm lateral to the midline, and 2–2.4 mm deep from the dorsal surface of the medulla. The remaining procedure for microinjections into the nAmb was similar to that described for microinjections into the ARCN except that l-glutamate (l-Glu; 5 mM) instead of NMDA was used to identify nAmb sites eliciting bradycardia.

Retrograde tracing of ARCN projections.

The work station and stereotaxic assembly were sanitized by Clorox spray. Rats were anesthetized with intraperitoneal injections of pentobarbital sodium (50 mg/kg) and fixed in a prone position in the stereotaxic instrument. All surgical instruments were sterilized using an autoclave. The surgical procedure for approaching the nAmb was identical to that described in Microinjections into the nAmb. A microinjection of either green retrobead solution [20 nl, 10% (wt/vol), n = 4] or fluorogold (2%, 5 nl, n = 2) was made into the nAmb. After the microinjection was completed, the exposed brain surface was covered with a small piece of absorbable gelatin sponge (Surgifoam, Ethicon, Somerville, NJ), and the skin over the wound was sutured. After recovery from pentobarbital anesthesia, rats survived for 7–12 days. Rats were administered an antibiotic [cefazolin (30 mg/kg)] subcutaneously twice a day for 3 days and one dose of a slow-release dosage form of an analgesic [buprenorphine SR (1 mg/kg)]. On the day of euthanization, animals were deeply anesthetized with pentobarbital and perfused with heparinized normal saline followed by 2% paraformaldehyde solution. Brains were removed and stored for 48 h in 2% paraformaldehyde. After the completion of the fixation procedure, the left or right side of the brain surface was marked by a superficial cut, and serial sections of the brain containing the hypothalamus and nAmb were cut (40 μm) in a vibratome (1000 Plus Sectioning System, The Vibratome Company, St. Louis, MO). Sections were mounted on subbed slides, covered with Citifluor mountant medium (Ted Pella, Redding, CA), and coverslipped. The microinjection sites of green retrobead solution [wavelength of maximal absorbance (A_max): 460 nm, and wavelength of maximal emission (E_max): 505 nm] and of fluorogold (A_max: 360 nm, and E_max: 515 nm) in the nAmb and retrogradely labeled cells in the brain sites of our interest were visualized under a microscope (model AX70, Olympus Provis, Middlebush, NJ). In retrograde tracing experiments, the brain sites of our interest included the ARCN, caudal ventrolateral medullary depressor area (CVLM), RVLM, nucleus tractus solitarius (NTS), and PVN. Brain sections were compared with a standard atlas (38) and photographed using Neurolucida software (version 7.5, MicroBrightField, Williston, VT).

Histological verification of microinjection sites.

At the end of the experiment, microinjections of green retrobeads [0.2% (wt/vol)] were made into the nAmb (20 nl) and ARCN (30 nl) to mark the microinjection sites. Animals were then perfused with heparinized normal saline and paraformaldehyde, and brains were removed and prepared as sections as described in Retrograde tracing of ARCN projections. The microinjection sites of green retrobeads in the nAmb and ARCN were visualized under a microscope, photographed using Neurolucida software, and compared with a standard atlas (38).

Drugs and chemicals.

The following drugs and chemicals were used in this study: d(−)-2-amino-7-phosphono-heptanoic acid (d-AP7; NMDA receptor antagonist), gabazine bromide (GABA_A receptor antagonist), l-Glu monosodium, isoflurane, 2,3-dioxo-6-nitro-1,2,3,4-tetrahydrobenzo-[f]quinoxaline-7-sulfonamide (NBQX) disodium salt (non-NMDA receptor antagonist), muscimol hydrobromide, strychnine hydrochloride, and urethane. Green retrobead solution and fluorogold were procured from Lumafluor (Durham, NC) and Fluorochrome (Denver, CO), respectively. Isoflurane was purchased from Piramal Critical Care (Bethlehem, PA). All other drugs were purchased from Sigma-Aldrich. All solutions for the microinjections were freshly prepared in aCSF (pH 7.4).

Statistical analyses.

Maximum changes in mean arterial pressure (MAP) and HR in response to microinjections of different drugs are expressed as means ± SE. One-way ANOVA followed by Tukey-Kramer's multiple-comparison test was used to evaluate the significance of differences in maximum changes in MAP and HR elicited by microinjections of different drugs in different groups of rats. Student's paired t-test was used for comparison of the following responses: decreases in MAP and increases in HR induced by microinjections of NMDA into the ARCN and decreases in HR elicited by microinjections of d-AP7, gabazine, l-Glu, NBQX, and strychnine and increases in HR by microinjections of muscimol into the nAmb. Differences were considered significant at P < 0.05 in all cases.

RESULTS

Baseline values for MAP and HR in urethane-anesthetized rats were 101.9 ± 3.4 mmHg and 399.5 ± 4.5 beats/min, respectively (n = 69).

Effect of NMDA microinjection into the ARCN.

In this and other series of experiments, the ARCN was always identified by microinjections of NMDA (10 mM), which elicited decreases in MAP (19 ± 1 mmHg) and increases in HR (46 ± 3 beats/min, n = 9) in nonbarodenervated rats (26). Time intervals between the microinjection and beginning of the response (onset) and maximum response (peak) and the end of the response (duration) were noted. The onset, peak, and duration of depressor responses to microinjections of NMDA (1 mM) were 14.1 ± 2.4 s, 2.8 ± 0.5 min, and 10.1 ± 1.6 min, respectively. The onset, peak, and duration of tachycardic responses (16 ± 3 s, 3 ± 1 min, and 13 ± 2 min, respectively) were not statistically different from those of depressor responses. These time intervals are similar to those previously reported by us and others after microinjections of NMDA into the ARCN or PVN (24, 26, 30). In the ARCN, all microinjections were unilateral, and the volume of microinjections was 30 nl. Microinjections of aCSF (30 nl) into ARCN elicited no cardiovascular responses.

Effect of muscimol microinjections into the nAmb on NMDA-induced HR responses from the ARCN.

In one group of rats (n = 9), the ARCN was identified by microinjections (30 nl) of NMDA (10 mM); as described above, increases in HR were elicited. Twenty minutes later, when the effects of NMDA in the ARCN had abated, the ipsilateral nAmb was identified by microinjections of l-Glu (5 mM, 20 nl) in the same rats; decreases in HR (76.4 ± 9.3 beats/min) were elicited within 2 ± 0.1 s, and this effect lasted for 74.8 ± 10.7 s. In this and other experiments, microinjections of aCSF (20 nl) into the nAmb elicited no cardiovascular responses. Next, muscimol (1 mM, 20 nl) was microinjected into the nAmb. Muscimol was used as a pharmacological tool to inhibit nAmb neurons in this experiment. The concentration of muscimol used in the present study was selected from one of our previous reports (9, 10). Unilateral microinjections of muscimol into the nAmb elicited an increase in HR (31.4 ± 4.1 beats/min); the onset and duration of this effect were 4.9 ± 0.9 s and 6.1 ± 1 min, respectively. At this time, microinjection of NMDA into the ARCN was repeated. The increases in HR induced by NMDA in the ARCN were significantly (P < 0.0001) reduced; the increases before and after unilateral microinjections of muscimol into the ipsilateral nAmb were 46.1 ± 3.2 and 25.1 ± 2.2 beats/min, respectively (Fig. 1, A and B).

Fig. 1.

Attenuation of tachycardia elicited from the arcuate nucleus (ARCN) by inhibition of neurons in the nucleus ambiguus (nAmb). A: increase in heart rate (HR) elicited by unilateral microinjections of N-methyl-d-aspartic acid (NMDA; 10 mM) into the ARCN [n = 9; baseline value of HR: 387.3 ± 14.4 beats/min (bpm)]. B: in the same group of rats, the increase in HR elicited by unilateral microinjections of NMDA (10 mM) into the ARCN was significantly attenuated by prior unilateral microinjections of muscimol (Mus; 1 mM) into the ipsilateral nAmb (baseline value: 388.9 ± 12.2 beats/min). C: in another group of rats (n = 6), the increase in HR elicited by unilateral microinjections of NMDA (10 mM) into the ARCN (baseline value: 363.2 ± 11.3 beats/min) is shwon. D: in the same rats, the increase in HR elicited by unilateral microinjections of NMDA (10 mM) into the ARCN was significantly attenuated by prior bilateral microinjections of Mus (1 mM) into the nAmb (baseline value: 368.2 ± 11.6 beats/min). E–J: tracings from one experiment [top trace: mean arterial pressure (MAP); in mmHg: bottom trace: HR, in beats/min]. E: microinjection of NMDA (10 mM) into the ARCN elicited a decrease in MAP and an increase in HR. F and G: 20 min later, the nAmb was identified bilaterally with microinjections of l-glutamate (l-Glu; 5 mM); a decrease in HR was elicited, with no changes in MAP. H and I: after 20 min, Mus (1 mM) was bilaterally microinjected into the nAmb; an increase in HR was elicited. J: 5 min later, tachycardia elicited from the ARCN was significantly attenuated (compare with E). Lt, left; Rt, right. ****P < 0.0001.

In another group of rats (n = 6), the ARCN was identified by microinjections of NMDA, as described above. Twenty minutes later, nAmb was identified bilaterally by microinjections of l-Glu (5 mM, 20 nl each side) in the same rats. Next, muscimol was microinjected into the nAmb bilaterally (1 mM, 20 nl each side). Bilateral microinjections of muscimol into the nAmb elicited an increase in HR (46.6 ± 1.8 beats/min) within 5.9 ± 0.4 s, and this effect lasted for 12.4 ± 0.9 min. At this time, the increases in HR induced by NMDA in the ARCN were significantly reduced; the increases before and after bilateral microinjections of muscimol into the nAmb were 55.8 ± 3.4 and 10.1 ± 1.3 beats/min, respectively (P < 0.0001; Fig. 1, C and D). The attenuation of tachycardic responses elicited by microinjections of NMDA into the ARCN by bilateral microinjections of muscimol into the nAmb was significantly greater (P < 0.001) compared with the attenuation caused by unilateral microinjection of muscimol into the nAmb. Typical tracings of the effect of bilateral microinjections of muscimol into the nAmb on tachycardic responses elicited from the ARCN are shown in Fig. 1, E–J.

Effect of microinjections of gabazine into the nAmb on NMDA-induced HR responses from ARCN.

In one group of rats (n = 9), the ARCN was identified by microinjections of NMDA. Twenty minutes later, the ipsilateral nAmb was identified by microinjections of l-Glu (5 mM, 20 nl) in the same rats. Next, gabazine (0.01 mM, 20 nl) was microinjected into the ipsilateral nAmb. The concentration of gabazine used in the present study was selected from our previous report (28). Unilateral microinjections of gabazine into the nAmb elicited a decrease in HR (17.7 ± 1.1 beats/min); the onset and duration of this effect were 6.6 ± 0.8 s and 5.7 ± 0.7 min, respectively. The increases in HR induced by NMDA in the ARCN were significantly (P < 0.001) attenuated by unilateral microinjections of gabazine into the nAmb; increases in HR elicited by ARCN stimulation before and after microinjections of gabazine into the ipsilateral nAmb were 56.7 ± 9.6 and 33 ± 5.4 beats/min, respectively (Fig. 2, A and B).

Fig. 2.

Attenuation of tachycardia elicited from the ARCN by GABA_A receptor blockade in the nAmb. A: increase in HR elicited by unilateral microinjections of NMDA (10 mM) into the ARCN (n = 9; baseline value of HR: 370.4 ± 15.5 beats/min). B: in the same group of rats, the increase in HR elicited by unilateral microinjections of NMDA (10 mM) into the ARCN was significantly attenuated by prior unilateral microinjections of gabazine (Gabaz; 0.01 mM) into the ipsilateral nAmb (baseline value: 346.1 ± 17.2 beats/min). C: in another group of rats (n = 8), the increase in HR elicited by unilateral microinjections of NMDA (10 mM) into the ARCN (baseline value: 355.3 ± 15.5 beats/min) is shown. D: in the same rats, the increase in HR elicited by unilateral microinjections of NMDA (10 mM) into the ARCN was significantly attenuated by prior bilateral microinjections of Gabaz (0.01 mM) into the nAmb (baseline value: 330.5 ± 14.7 beats/min). E–J: tracings from one experiment. E: microinjection of NMDA (10 mM) into the ARCN elicited a decrease in MAP and an increase in HR. F and G: 20 min later, the nAmb was identified bilaterally with microinjections of l-Glu (5 mM); a decrease in HR was elicited, with no changes in MAP. H and I: after 20 min, Gabaz (0.01 mM) was bilaterally microinjected into the nAmb; no significant change in HR was elicited by this dose of Gabaz. J: 5 min later, tachycardia elicited from the ARCN was significantly attenuated (compare with E). ***P < 0.001; ****P < 0.0001.

In another group of rats (n = 8), the ARCN was identified by microinjections of NMDA. Twenty minutes later, nAmb was identified bilaterally by microinjections of l-Glu (5 mM, 20 nl) in the same rats. Next, gabazine was microinjected into the nAmb bilaterally (0.01 mM, 20 nl each side). Bilateral microinjections of gabazine into the nAmb elicited a decrease in HR (21.7 ± 1.8 beats/min) within 7.4 ± 0.5 s, and this effect lasted for 10.6 ± 1.6 min. At this time, the increases in HR induced by NMDA in the ARCN were significantly reduced; the increases in HR before and after bilateral microinjections of gabazine into the nAmb were 50.5 ± 4.3 and 10.6 ± 1 beats/min, respectively (P < 0.0001; Fig. 2, C and D). The attenuation of tachycardic responses elicited from the ARCN by NMDA was significantly (P < 0.01) greater with bilateral microinjections of gabazine into the nAmb compared with unilateral microinjection of gabazine into the nAmb. Typical tracings of the effect of bilateral microinjections of gabazine into the nAmb on tachycardic responses elicited from the ARCN are shown in Fig. 2, E–J.

Effect of strychnine microinjections into the nAmb on NMDA-induced HR responses from the ARCN.

In one group of rats (n = 5), the ARCN was identified by microinjections of NMDA. Twenty minutes later, the ipsilateral nAmb was identified by microinjections of l-Glu (5 mM, 20 nl) in the same rats. Next, strychnine (0.5 mM, 20 nl) was microinjected into the ipsilateral nAmb. The concentration of strychnine used in the present study was selected from our previous report (7). Unilateral microinjections of strychnine into the nAmb elicited a decrease in HR (14.3 ± 1.8 beats/min); the onset and duration of this effect were 5.9 ± 0.8 s and 5.8 ± 0.7 min, respectively. At this time, increases in HR elicited by microinjections of NMDA into the ARCN were significantly (P < 0.05) reduced; the increases in HR before and after microinjections of strychnine into the ipsilateral nAmb were 43.8 ± 7 and 23.4 ± 2.8 beats/min, respectively (Fig. 3, A and B).

Fig. 3.

Attenuation of tachycardia elicited from the ARCN by glycine receptor blockade in the nAmb. A: increase in HR elicited by unilateral microinjections of NMDA (10 mM) into the ARCN (n = 5; baseline value of HR: 341.4 ± 5.9 beats/min). B: in the same group of rats, the increase in HR elicited by unilateral microinjections of NMDA (10 mM) into the ARCN was significantly attenuated by prior unilateral microinjections of strychnine (Strych; 0.5 mM) into the ipsilateral nAmb (baseline value: 345 ± 12.2 beats/min). C: in another group of rats (n = 5), the increase in HR elicited by unilateral microinjections of NMDA (10 mM) into the ARCN (baseline value: 349.2 ± 13 beats/min) is shown. D: in the same rats, the increase in HR elicited by unilateral microinjections of NMDA (10 mM) into the ARCN was significantly attenuated by prior bilateral microinjections of Strych (0.5 mM) into the nAmb (baseline value: 329.4 ± 16.1 beats/min). E–J: tracings from one experiment. E: microinjection of NMDA (10 mM) into the ARCN elicited a decrease in MAP and an increase in HR. F and G: 20 min later, the nAmb was identified bilaterally with microinjections of l-Glu (5 mM); a decrease in HR was elicited, with no changes in MAP. H and I: after 20 min, Strych (0.5 mM) was bilaterally microinjected into the nAmb; no significant change in HR was elicited by this dose of Strych. J: 5 min later, tachycardia elicited from the ARCN was significantly attenuated (compare with E). *P < 0.05; **P < 0.01.

In another group of rats (n = 5), the ARCN was identified by microinjections of NMDA. Twenty minutes later, the nAmb was identified bilaterally by microinjections of l-Glu (5 mM, 20 nl each side) in the same rats. Next, strychnine (0.5 mM, 20 nl each side) was microinjected into the nAmb bilaterally. Bilateral microinjections of strychnine into the nAmb elicited a decrease in HR (22. 9 ± 2.3 beats/min) within 7.1 ± 0.6 s, and the duration of this effect was 12.3 ± 1.2 min. At this time, the increases in HR induced by NMDA in the ARCN were significantly (P < 0.01) reduced; the increases before and after bilateral microinjections of strychnine into the nAmb were 49.2 ± 6.3 and 12.2 ± 0.5 beats/min, respectively (Fig. 3, C and D). The attenuation of NMDA-induced tachycardia in the ARCN elicited by bilateral microinjections of strychnine into the nAmb was significantly (P < 0.05) greater compared with the attenuation caused by unilateral microinjections of strychnine into the nAmb. Typical tracings of the effect of bilateral microinjections of strychnine into the nAmb on tachycardic responses elicited from the ARCN are shown in Fig. 3, E–J.

Effect of combined blockade of GABA_A and glycine receptors in the nAmb on NMDA-induced HR responses from the ARCN.

In one group of rats (n = 5), the ARCN was identified by microinjections of NMDA. Twenty minutes later, the ipsilateral nAmb was identified by microinjections of l-Glu (5 mM, 20 nl) in the same rats. Next, gabazine (0.01 mM) and strychnine (0.5 mM, 20 nl each) were microinjected sequentially within 2–3 min into the ipsilateral nAmb. Combined microinjections of gabazine and strychnine into the nAmb elicited a decrease in HR (20.2 ± 2.2 beats/min); the onset and duration of this effect were 5.8 ± 0.4 s and 6.9 ± 0.7 min, respectively. At this time, increases in HR elicited by microinjections of NMDA into the ARCN were significantly (P < 0.001) reduced; the increases in HR before and after combined microinjections of gabazine and strychnine into the ipsilateral nAmb were 43.2 ± 4.8 and 16.8 ± 6.1 beats/min, respectively (Fig. 4, A and B).

Fig. 4.

A and B: attenuation of tachycardia elicited from the ARCN by combined microinjections of GABA_A and glycine receptor blockade in the nAmb. A: increase in HR elicited by unilateral microinjections of NMDA (10 mM) into the ARCN (n = 5; baseline value: 397.2 ± 4.2 beats/min). B: in the same rats, the increase in HR elicited by unilateral microinjections of NMDA (10 mM) into the ARCN was significantly attenuated by prior unilateral combined microinjections of Gabaz (0.01 mM) and Strychn (0.5 mM) into the ipsilateral nAmb (baseline value: 358.8 ± 5.7 beats/min). C and D: attenuation of tachycardia elicited from the ARCN by blockade of ionotropic glutamate receptors in the nAmb. C: the increase in HR elicited by unilateral microinjections of NMDA (10 mM) into the ARCN (n = 7; baseline value: 361.7 ± 7.4 beats/min) is shown. D: in the same rats, the increase in HR elicited by unilateral microinjections of NMDA (10 mM) into the ARCN was significantly attenuated by prior unilateral microinjections of d(−)-2-amino-7-phosphono-heptanoic acid (d-AP7; 5 mM) and 2,3-dioxo-6-nitro-1,2,3,4-tetrahydrobenzo-[f]quinoxaline-7-sulfonamide (NBQX; 2 mM) into the nAmb (baseline value: 367.9 ± 10 beats/min). *P < 0.05; ***P < 0.001.

Effect of blockade of iGLURs in the nAmb on NMDA-induced HR responses from the ARCN.

In a group of rats (n = 7), the ARCN was identified by microinjections of NMDA. Twenty minutes later, the ipsilateral nAmb was identified by microinjections of l-Glu (5 mM, 20 nl) in the same rats. Next, d-AP7 (5 mM) and NBQX (2 mM) were microinjected (20 nl each) into the ipsilateral nAmb. The concentrations of d-AP7 and NBQX used in the present study were selected from one report (9). Unilateral microinjections of iGLUR antagonists into the nAmb elicited a decrease in HR (16.4 ± 2.1 beats/min); the onset and duration of this effect were 2.8 ± 1 s and 5.9 ± 0.5 min, respectively. Blockade of iGLURs resulted in significantly (P < 0.05) reduced HR increases to NMDA microinjections into the ARCN; the increases in HR before and after the blockade of iGLURs in the ipsilateral nAmb were 54.6 ± 6.2 and 37.9 ± 3.7 beats/min, respectively (Fig. 4, C and D).

Histological verification of microinjection sites in the ARCN.

A typical microinjection site in the ARCN marked by green retrobeads is shown in Fig. 5A. The spot was in the ARCN at a level 2.28 caudal to the bregma, and the center of the spot was 0.35 mm lateral to the midline and 9.7 mm deep from the dura. Microinjection sites in the ARCN at different rostrocaudal levels (3.72-1.92 caudal to the bregma) are shown in Fig. 5, B–G, in which each dark spot represents a site in one animal (all microinjection sites are not shown because of overlapping). Rats in which the diffusion sphere of the marker was not within the boundaries of the ARCN, as shown in a standard atlas (38), were not included in the study.

Fig. 5.

Histological identification of microinjection sites in the ARCN. A: a coronal section at a level 2.28 mm caudal to the bregma showing a microinjection site in the ARCN marked with green retrobead solution (arrow); the center of the spot was 0.35 mm lateral to the midline and 9.7 mm deep from the dura. Scale bar = 500 μm. B–G: drawings of coronal sections 3.72 (B), 3.48 (C), 3.12 (D), 2.76 (E), 2.28 (F), and 1.92 (G) mm caudal to the bregma showing ARCN microinjection sites as dark spots; each spot represents a site in one animal. Microinjection sites were located in the ARCN, 0.1–0.6 mm lateral to the midline and 9.6–10.1 mm deep from the dura. 3V, third ventricle; f, fornix. Scale bar = 1 mm.

Histological verification of microinjection sites in the nAmb.

A typical microinjection site in the nAmb marked with green retrobeads is shown in Fig. 6A; the center of the spot shown in Fig. 6A was located 0.24 mm rostral to the CS, 1.9 mm lateral to the midline, and 2.1 mm deep from the dura. Diagrams showing other microinjection sites in the nAmb at different rostrocaudal levels (0.64 mm rostral to 0.12 mm caudal to the CS) are shown in Fig. 6, B–H. In these diagrams, each dark spot represents a microinjection site in one animal (all microinjection sites are not shown because of overlapping). Rats in which the green retrobeads or fluorogold microinjections were not within the boundaries of the nAmb, as shown in a standard atlas (38), were not included in the study.

Fig. 6.

Histological identification of microinjection sites in the nAmb. A: a coronal section at a level 0.24 mm rostral with reference to the calamus scriptorius (CS) showing a microinjection site in the nAmb on the right side marked with green retrobead solution (arrow); the center of the spot was 1.9 mm lateral to the midline and 2.1 mm deep from the dura. B–H: drawings of coronal sections at different levels of the medulla showing nAmb microinjection sites as dark spots; each spot represents one animal. B–F: 0.64 (B), 0.48 (C), 0.36 (D), 0.24 (E), and 0.12 (F) mm rostral to the CS. G: at the level of the CS. H: 0.12 mm caudal to the CS. AP, area postrema; CC, central canal; py, pyramid; 10N, dorsal motor nucleus of the vagus; 12N, hypoglossal nucleus. Scale bar = 1 mm.

Retrograde labeling in the medulla.

Unilateral microinjections of retrobead solution (20 nl, n = 4) or fluorogold (5 nl, n = 2) into the nAmb (Fig. 7A) resulted in bilateral retrograde labeling of cells in the medial NTS with ipsilateral preponderance (Fig. 7B). Retrogradely labeled cells were also observed in the ipsilateral CVLM (Fig. 7, C and D) and RVLM (Fig. 7, E and F).

Fig. 7.

Retrograde tracing of medullary projections to the nAmb. A: a coronal section at a level 0.20 mm rostral to the CS showing a microinjection site (arrow) in the nAmb marked with fluorogold and retrograde labeling in the nucleus tractus solitarius (NTS; box). Scale bar = 1 mm. B: higher magnification of the boxed area in A showing retrograde labeling of NTS neurons. Scale bar = 200 μm. C: a coronal section at a level 1.32 mm rostral to the CS showing retrograde labeling of fluorogold in the caudal ventrolateral medullary depressor area (CVLM; box). Scale bar = 1 mm. D: magnified boxed area in C showing retrograde labeling of CVLM neurons. Scale bar = 200 μm. E: a coronal section at a level 1.92 mm rostral to the CS showing retrograde labeling of fluorogold in the rostral caudal ventrolateral medulla (RVLM; box). Scale bar = 1 mm. F: higher magnification of the boxed area in E showing retrogradely labeled RVLM neurons. Scale bar = 200 μm.

Retrograde labeling in the hypothalamus.

The ARCN (∼2.64 mm long in the rostrocaudal direction) was arbitrarily divided into the following three equal segments: rostral (1.72–2.6 mm caudal to the bregma), middle (2.6–3.48 mm caudal to the bregma), and caudal (3.48–4.36 mm caudal to the bregma). Unilateral microinjections of retrobead solution (20 nl) or fluorogold (5 nl) into the nAmb resulted in retrograde labeling of cells in the ipsilateral as well as contralateral rostral (Fig. 8, A and B) and middle (Fig. 8, C and D) segments of the ARCN. The PVN (∼1.68 mm long in the rostrocaudal direction) was also arbitrarily divided into the following three equal segments: rostral (0.6–1.16 mm caudal to the bregma), middle (1.16–1.72 mm caudal to the bregma), and caudal (1.72–2.28 mm caudal to the bregma). Unilateral microinjections of retrobead solution (20 nl) or fluorogold (5 nl) into the nAmb resulted in retrograde bilateral labeling of cells in the rostral (Fig. 9, A and B) and middle (Fig. 9, C and D) segments of PVN with ipsilateral preponderance. No retrogradely labeled cells were observed in the caudal segments of either the ARCN or PVN.

Fig. 8.

Retrograde tracing of ARCN projections to the nAmb. A: a coronal section at a level 1.80 mm caudal to the bregma showing bilateral retrograde labeling of the rostral ARCN (box) after a microinjection of fluorogold into the nAmb. Scale bar = 200 μm. B: higher magnification of the boxed area in A showing retrogradely labeled ARCN cells. Scale bar = 100 μm. C: a coronal section at a level 3.36 mm caudal to bregma showing bilateral labeling of cells in the middle ARCN. Scale bar = 200 μm. D: magnified boxed area in C showing retrograde labeling of ARCN neurons. Scale bar = 100 μm.

Fig. 9.

Retrograde tracing of paraventricular nucleus (PVN) projections to the nAmb. A: a coronal section at a level 1.08 mm caudal to bregma showing retrogradely labeled cells in the ipsilateral rostral PVN after a microinjection of fluorogold into the nAmb. Scale bar = 200 μm. B: magnification of the boxed area in A showing retrograde labeled cells in the ipsilateral rostral PVN. Scale bar = 100 μm. C: a coronal section at a level 1.32 mm caudal to the bregma showing retrogradely labeled cells in the middle PVN. Scale bar = 200 μm. D: magnified boxed area in C showing retrogradely labeled cells in the ipsilateral middle PVN. Scale bar = 100 μm.

DISCUSSION

The major findings of the present study are as follows: 1) microinjections of NMDA into the ARCN elicit increases in HR, 2) tachycardic responses elicited from the ARCN are blocked by prior unilateral and bilateral inhibition of nAmb neurons using microinjections of muscimol, and 3) increases in HR are attenuated by prior unilateral and bilateral blockade of GABA_A and glycine receptors in the nAmb.

We have previously reported that tachycardic responses elicited by chemical stimulation of the ARCN are mediated via stimulation of the sympathetic inputs to the heart as well as inhibition of vagal inputs to the heart (36). The pathways mediating sympathetic stimulation of the heart after ARCN stimulation have been described in our previous publications (26, 27, 36, 41). Tachycardic responses elicited by microinjections of NMDA into the ARCN are not reflex responses to the concomitant falls in BP because similar tachycardic responses are elicited in barodenervated rats (27). As described above in the Introduction, the brain areas and neurotransmitters/receptors in them that mediate the inhibition of the vagal input to the heart after ARCN stimulation are not known.

It has been well established that the vagal input to the heart originates predominantly from the nAmb (13, 43, 44). In the present study, unilateral and bilateral microinjections of muscimol into the nAmb attenuated tachycardic responses elicited by microinjections of NMDA into the ARCN, indicating that the nAmb is important in mediating these responses. Furthermore, blockade of either GABA_A receptors by gabazine or glycine receptors by strychnine in the nAmb significantly attenuated the tachycardia elicited from the ARCN, suggesting that both GABA and glycine may be released in the nAmb after chemical stimulation of the ARCN. The attenuation of tachycardic responses elicited by ARCN stimulation after simultaneous blockade of GABA_A and glycine receptors in the nAmb was not significantly greater compared with the attenuation of HR responses after blockade of either GABA_A or glycine receptors alone. It is possible that either GABA or glycine is released in the nAmb after ARCN stimulation. The situations under which either one of these pathways is activated cannot be identified based on our results. Thus, inhibition of CVNs by release of either GABA or glycine in the nAmb may be the mechanism by which ARCN stimulation decreases the vagal input to the heart, resulting in tachycardia.

The nAmb in the rat consists of a rostral compact portion (nAmbC), a semicompact portion (nAmbS; located ventral and posterior to the nAmbC), and a loose formation of neurons (nAmbL; located posterior to nAmbS). The external formation of the nAmb (nAmbE) is located ventral to the nAmbC and nAmbS at rostral levels and surrounds the nAmbL at caudal levels (3, 18). The precise location and distribution of CVNs in different regions of the nAmb have not been firmly established. Previous reports have suggested that CVNs are located primarily in the nAmbE at the level of the CS (3, 18). In recent studies (5, 45), the distribution of CVNs in the different regions of the nAmb in the rat has been reported as follows: nAmbC (26.2%), nAmbE (19.2%), nAmbL (10.2%), and nAmbS (8%). The possibility of spread of retrograde tracers into the thoracic esophagus or other tissues was excluded in these studies (5, 45). Our microinjections of muscimol, gabazine, and strychnine into the nAmb extended from 0.12 mm caudal to 0.64 mm rostral with reference to the CS. The nAmbC and nAmbS are not present at this rostrocaudal level of the medulla (38). Therefore, CVNs located in the nAmbE and nAmbL at levels 0.12 mm caudal to 0.64 mm rostral with reference to the CS were involved in our pharmacological manipulations. Significant bradycardic responses were elicited by microinjections of l-Glu into the nAmb at these rostrocaudal medullary levels (Ref. 32 and the present study). We placed microinjections into the center of the nAmb to avoid diffusion of the injectate into the adjacent areas (e.g., CVLM) (40, 49). However, the short onset of bradycardia (2 ± 0.1 s) induced by microinjections of l-Glu into the nAmb indicated that the injected material reached CVNs located in the nAmbE and nAmbL within seconds.

The pathways from the ARCN to the nAmb that are involved in mediating increases in HR elicited by ARCN stimulation cannot be identified with certainty based on our results. One possibility is that ARCN neurons containing α-MSH and/or glutamate (26) may be activated by NMDA, one or both of these excitatory neurotransmitters may be released in the nAmb at the terminals of direct projections to the nAmb, interneurons containing GABA and/or glycine in the nAmb may be activated, and the release of GABA and/or glycine in the nAmb may cause inhibition of CVNs with consequent tachycardia. Alternatively, GABA neurons present in the ARCN (20, 26, 41) may be stimulated by NMDA microinjections into the ARCN; GABA may be released at terminals of direct projections of these neurons in the nAmb, causing inhibition of CVNs with subsequent tachycardia. The source of glycinergic input to CVNs in the nAmb is not known.

Another possibility is that microinjections of NMDA into the ARCN may activate glutamate- and/or α-MSH-containing cells projecting from the ARCN to the PVN (27), releasing one or both of these neurotransmitters in the PVN, which results in the activation of PVN neurons projecting to the nAmb; an excitatory neurotransmitter is then released in the nAmb, interneurons containing GABA and/or glycine in this nucleus are stimulated, GABA and/or glycine is released in the nAmb, CVNs are inhibited, and HR is increased. Supporting the possible role of this pathway in mediating tachycardia elicited by ARCN stimulation is our previously reported observation that tachycardia elicited by microinjections of NMDA into the ARCN is attenuated by blockade of iGLURs and melanocortin 3/4 receptors in the PVN (27). Furthermore, blockade of iGLURs in the nAmb significantly attenuated tachycardic responses elicited by microinjections of NMDA into the ARCN (present study).

The implication of the pathways described above in mediating tachycardic responses elicited from the ARCN is supported by our anatomic tracing experiments. Direct bilateral projections from the ARCN to the nAmb were identified after microinjection of tracers (green retrobeads and fluorogold) into the nAmb. In agreement with Ciriello et al. (12), direct bilateral projections from the PVN to the nAmb were also identified after microinjection of these tracers into the nAmb. Direct projections to the nAmb from the medial NTS, CVLM, and RVLM were also identified in our study; these observations are consistent with a report (16) showing that CVNs in the nAmb receive inputs from the medullary areas located directly ventral, medial, and dorsomedial to the nAmb. GABAergic neurons located in areas ventral and medial to the nAmb may be involved in respiratory modulation of CVNs and cardiorespiratory functions (16). In the anatomic retrograde tracing techniques used in the present study, there is a possibility that axons passing through the injection site in the nAmb, but not terminating there, could have taken up the tracer, resulting in the labeling of neurons in the medulla and hypothalamus. This problem was minimized in some experiments by using green retrobeads, which are not taken up by fibers of passage (23, 42). However, uptake of retrobeads by fibers of passage damaged during microinjections cannot be definitely excluded.

CVNs located in the nAmb do not possess inherent pacemaker properties and depend on synaptic neurotransmission for their activation or inhibition (34, 35). GABA and glycine have been implicated as major inhibitory neurotransmitters in the nAmb (6, 46, 47). Endogenous ACh, norepinephrine, ATP, and nitric oxide facilitate GABAergic and/or glycinergic neurotransmission to CVNs (4, 21, 22, 46–48). Facilitation of GABAergic and/or glycinergic neurotransmission in the nAmb have been implicated in some physiological and pathophysiological conditions. For example, ACh-induced facilitation of GABAergic and/or glycinergic neurotransmission to CVNs in the nAmb may be involved in the generation of respiratory sinus arrhythmias (48). Activation of α₁-adrenergic receptors by norepinephrine facilitates GABAergic and glycinergic neurotransmission in the nAmb, reducing the activity of CVNs, which may explain tachycardia observed in norepinephrine-dependent behavioral arousal (4). One of the mechanisms by which CVNs are excited during hypoxia may be disinhibition of these neurons via withdrawal of GABAergic and glycinergic neurotransmission (15). The inhibitory pathway to the nAmb identified in our study may play a role in stress-induced tachycardia. In this context, it may be noted that the ARCN is one of the brain sites for the location of neurons that are activated during stress (31, 37).

There are some technical issues that are relevant to this study and need some discussion. The onset, peak, and duration of NMDA effects in the ARCN are relatively long. These longer time intervals cannot be ascribed to diffusion of the injected material to areas located adjacent to the ARCN because the volume of microinjection in our study was small. We have previously shown that microinjections of NMDA (30–50 nl) into the ARCN do not diffuse to the dorsal median nucleus of the hypothalamus, which is located immediately dorsal to the ARCN (36). Longer onset and peak times of the responses to NMDA microinjections into the ARCN may be due to the fact that the ARCN consists of heterogeneous populations of neurons, and longer times may be needed for NMDA to reach the target neurons. The possible involvement of relay neurons (e.g., PVN neurons) may also contribute to these longer time intervals. NMDA is a nonmetabolizable analog of d-aspartic acid, and it may remain at the injection site for a longer time, thus increasing the duration of NMDA effects. At the beginning of each experiment, the ARCN was functionally identified by microinjections of NMDA instead of l-Glu because very high doses of l-Glu are needed for stimulation of this nucleus (36). On the other hand, the nAmb was functionally identified by microinjections of l-Glu, consistent with our previous reports (6, 8–10). Cardiovascular responses elicited from the ARCN or nAmb were not due to the distortion of the brain tissue because microinjections of aCSF into these nuclei elicited no cardiovascular responses.

Perspectives

In the present study, we have demonstrated that ARCN stimulation elicits an increase in HR that is mediated via GABA_A and glycine receptors in the nAmb. The ARCN is located close to the median eminence, which has a high capillary density and high capillary blood flow and lacks a blood-brain barrier (17). Because of these anatomic features, the ARCN may be accessible to physiologically active substances circulating in the blood. The results of the present study provide the groundwork for testing the role of ARCN in mediating HR and other cardiovascular responses to circulating substances that are released during stress.

GRANTS

This work was supported in part by National Heart, Lung, and Blood Institute Grants HL-024347 and HL-076248 (to H. N. Sapru).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: V.C.C., K.K., and H.N.S. conception and design of research; V.C.C. and K.K. performed experiments; V.C.C., K.K., and H.N.S. analyzed data; V.C.C., K.K., and H.N.S. interpreted results of experiments; V.C.C., K.K., and H.N.S. prepared figures; V.C.C. and H.N.S. drafted manuscript; V.C.C. and H.N.S. edited and revised manuscript; V.C.C., K.K., and H.N.S. approved final version of manuscript.