Introduction

Early anatomical and functional studies have proposed a subdivision of the hippocampus in dorsal (DH) and ventral (VH) hippocampus [5, 79]. The hippocampus is a structure of the limbic system [60]. The limbic system and visceral brain were terms coined by MacLean in 1949 to refer a group of cortical and subcortical structures forming an integrated neural system related to the integration of autonomic and behavioral responses [49, 50].

Hippocampus stimulation may cause behavioral changes such as escape, defensive, attack, and attention responses [43, 51, 74] as well as signs of autonomic nervous system activation such as salivation, pupillary changes, inhibition of pyloric antral peristalsis, and respiratory inhibition [3, 9, 42]. Likewise, electrical stimulation of the hippocampus can evoke decreases in heart rate (HR) and increases in pulse pressure [76], decreases in blood pressure (BP) with either increases or decreases in HR [3], and sympathetic activation as indicated by pupillary dilatation along with the increase in the BP [4].

Moreover, previous work from our group has demonstrated the involvement of the hippocampus in the modulation of cardiovascular reflex responses. Administration of cobalt chloride (CoCl2), a calcium-dependent synaptic neurotransmission blocker, into the DH increased the heart rate reflex responses promoted by vasoactive drugs (phenylephrine or sodium nitroprusside) intravenously injected. Thus, it suggests that the DH has an inhibitory influence on cardiac baroreflex function [28].

Glutamate (glut) is the primary excitatory neurotransmitter in the central nervous system (CNS). Glut administration into several areas of the CNS promotes changes in both BP and HR [7, 8, 77, 84]. N-methyl-d-aspartate (NMDA) and α-amino-3-hydroxy-5-methyl-4-isoxazole propionate (AMPA)/Kainate receptors are extensively expressed in the hippocampus [10, 34, 38, 83] and unilateral administration of glut into the DH of awake rats elicit marked decrease in BP, HR, and respiratory rate [34, 68].

In contrast, gamma-aminobutyric acid (GABA) is the main inhibitory neurotransmitter in the CNS. GABA is highly expressed within the DH [20, 39], but GABA content is reduced in the hippocampus of spontaneous hypertensive rats (SHR) [36, 41]. Moreover, intracerebroventricular injection of GABA reduces the sympathetic nerve activity, BP, and HR in a dose-dependent manner in both normotensive rats and SHR, but the magnitude of these effects is larger in SHR than normotensive animals [71].

Furthermore, the endocannabinoid system is implicated in several hippocampal functions [34, 35, 44, 48, 85]. The enzymes, fatty acid amide hydrolase (FAAH) and monoacylglycerol lipase (MAGL), responsible for degradation of endocannabinoids anandamide and 2-arachidonoylglycerol, respectively, are found in the hippocampus [18, 64] as well as the cannabinoid receptors (subtype 1 and 2) are widely expressed in the hippocampus [45, 64]. The cannabinoid receptor subtype 1 (CB1) may be found in presynaptic glutamatergic axon terminals [45, 64], and the binding of endocannabinoids in the CB1 receptors initiates different signaling pathways that hyperpolarize the cell membrane triggering the decrease in the exocytose of glut from presynaptic neurons in the hippocampus [35, 44, 48, 85].

Therefore, considering that the hippocampus can modulate the cardiac baroreflex function, administration of glut in the DH promotes cardiovascular responses, GABAergic neurotransmission is impaired in hippocampus of SHR, and activation of CB1 receptors in the hippocampus may reduce glutamate exocytose; we tested the hypothesis that glutamatergic, GABAergic, and endocannabinoid neurotransmissions interact with each other within the DH to modulate the cardiac baroreflex function in rats.

Materials and methods

Ethical approval

Experimental procedures were carried out according to the protocols approved by the Ethical Review Committee of the School of Medicine of Ribeirao Preto (Protocol number 128/2010) that follows the rules of the European Communities Council Directive of 24 November 1986 (86/609/EEC) and the Guiding Principles for Research Involving Animals and Human Beings of the American Physiological Society.

Drugs

The following drugs were used: DL-AP7 (Tocris Bioscience, Bristol, UK; NMDA receptor antagonist), NBQX (Tocris Bioscience, Bristol, UK; AMPA/Kainate receptor antagonist), AM251 (Tocris Bioscience, Bristol, UK; CB1 receptor antagonist), Bicuculline (Tocris Bioscience, Bristol, UK; GABAA receptor antagonist), lidocaine 2% (S.C., S.S. White, Rio de Janeiro, RJ, Brazil), ketamine (I.P., 100 mg/kg, Sespo, Ind e Comércio, Paulínia, SP, Brazil), xylazine (I.P., 20 mg/kg, Ind e Comércio, Paulínia, SP, Brazil), urethane (I.P., 1.25 g/kg, Sigma-Aldrich, St. Louis, MO, USA), flunixin meglumine (S.C., 5 mg/kg, Banamine®, Schering-Plough, Cotia, São Paulo, Brazil), and polyantibiotic preparation of streptomycin and penicillins (I.M., Pentabiotico®, Fort Dodge, Campinas, São Paulo, Brazil).

Animal surgeries

We utilized in this work male Wistar rats weighing 230–270 g. Animals were obtained from the animal breeding facility of the University of Sao Paulo (Ribeirao Preto, SP). They were housed individually in plastic cages in a temperature-controlled room at 25 °C and kept under a 12- and 12-h light–dark cycle (lights on between 06:00 and 18:00 h) with food and water ad libitum in the Animal Facility of the Department of Pharmacology, School of Medicine of Ribeirao Preto.

Before starting the surgical procedures, rats were anesthetized with ketamine (100 mg/kg, I.P., Sespo, Ind e Comércio, Paulínia, SP, Brazil) and xylazine (20 mg/kg, I.P., Sespo, Ind e Comércio, Paulínia, SP, Brazil). To assess the level of consciousness by the degree of antinociception (lack of response to noxious stimuli) after the anesthesia procedure, we gently pinched the back paw of the animals and observed the musculoskeletal response and cardiovascular and respiratory function.

Six days before cardiovascular recordings, rats were anesthetized with ketamine and xylazine. After local anesthesia with 2% lidocaine, the skull was surgically exposed and stainless-steel guide cannulae (26 G) were bilaterally implanted into the DH using a stereotaxic apparatus (Stoelting, Wood Dale, IL, USA). Stereotaxic coordinates for cannula implantation (AP = − 3.8 mm, L = +2.8 mm from the medial suture and V = − 2.2 mm from the skull) were selected from the Rat Brain Atlas of Paxinos and Watson (2007). Cannulae were fixed to the skull with dental cement and one metal screw. After the surgery, animals were treated with the polyantibiotic preparation (to prevent infection) and nonsteroidal anti-inflammatory (for postoperative analgesia).

One day before cardiovascular recordings, rats were again anesthetized with ketamine and xylazine and a catheter (4-cm segment of PE-10 heat-bound to a 13-cm segment of PE-50, Clay Adams, Parsippany, NJ, USA) was inserted into the abdominal aorta through the femoral artery for blood pressure recording. A second catheter was implanted into the femoral vein, and it was used to infuse vasoactive drugs to evoke arterial BP alterations. Both catheters were tunneled under the skin and exteriorized on the animal’s dorsum. The treatment with nonsteroidal anti-inflammatory was repeated after the surgery.

Drug injection

We have diluted the drugs in appropriate vehicles according to the datasheet of the manufacturer. DL-AP7 has been diluted in NaOH 100 nM and Bicuculline in DMSO 2%, while the NBQX and the AM251 were diluted in DMSO 10%. The vehicles and concentrations used in this work did not produce any effects on the baroreflex function as well as they did not generate toxic effects when they were used in previously published works [13, 26, 47, 59, 69]. Needles (33 G, Small Parts, Miami Lakes, FL, USA) used for microinjection into the DH were 1 mm longer than guide cannulas. After the syringe was filled with the drug, the needle was connected to a 2-μL syringe (7002-H, Hamilton Co., Reno, NV, USA) through the PE-10 tubing. The needle was carefully inserted into the guide cannula, and drugs injected in a final volume of 500 nL over a 60-s period.

Baroreflex assessment

Twenty-four hours after femoral artery surgery, we connected the animals to the acquisition system to collect the pulsatile arterial pressure using an AECAD 04P preamplifier (AVS projects, Sao Paulo, Sao Paulo, Brazil) and an acquisition board (PowerLab 8/30, AD Instruments, New South Wales, Australia) connected to a computer. Mean arterial pressure (MAP) and HR values were derived from these pulsatile recordings and were online processed. We activated the baroreflex through the intravenous infusion of phenylephrine (70 μg/ml at 0.4 ml/min/kg), an alpha 1-adrenoreceptor agonist, or sodium nitroprusside (SNP, 100 μg/ml at 0.8 ml/min/kg), a nitric oxide donor, with the aid of an infusion pump (KD Scientific, Holliston, MA, USA). Phenylephrine and SNP were injected, in a random way, in each animal over 30 s and caused, respectively, an increase or a decrease in the BP [26]. All animals received phenylephrine and SNP at three different times: before, 10 min after, and 60 min after administration of drugs into DH.

Baroreflex function analysis

We constructed the baroreflex sigmoidal curves and the linear regressions by matching the BP-evoked increases and the reflex HR decreases as well as by matching the BP-evoked decrease and the reflex HR increases. These correlations were generated at three different times (before, 10 min after, and 60 min after the administration of drugs into the DH) and used to build the sigmoidal curves and the linear regressions for each rat [25, 26]. We used Boltzmann sigmoidal function [Y = Bottom + (Top − Bottom) / 1 + exp (BP50 – X / Slope)] to build the sigmoidal curves and determine sigmoidal baroreflex parameters: (i) lower heart rate plateau (P1, bpm), (ii) upper heart rate plateau (P2, bpm), (iii) difference between upper and lower plateau levels (range, bpm), and (iv) the average slope of the sigmoidal curve (G, bpm/mmHg) [37, 46]. According to Head and Mccarty, while P1 plateau is influenced by parasympathetic activity, the P2 plateau may be used as a measure of the baroreflex sympathetic tonus [37]. On the other hand, linear regressions were utilized to analyze the slopes of tachycardic and bradycardic baroreflex responses separately [25, 29]. The results of the sigmoid curves and linear regression curves were separately analyzed using one-way ANOVA followed by Dunnett’s post hoc test. Moreover, we collected the average 5 min of stable cardiovascular recordings to assess the baseline BP and HR. These baseline values before and after pharmacological treatment in the DH were compared using the Student’s t test. Results of statistical tests with P < 0.05 were considered significant.

Histological procedure

We euthanized the animals at the end of the experiments using an overdose of urethane. After, 500 nL of 1% Evan’s blue dye was bilaterally injected into the DH as a marker of injection sites. The chest was surgically opened, the descending aorta occluded, the right atrium cut off, and the brain perfused with saline and 10% formalin through the left ventricle. Brains were postfixed for 24 h at 4 °C, and 40-μm sections were cut with a cryostat (CM-1900, Leica, Wetzlar, Germany). We used the Rat Brain Atlas of Paxinos and Watson to localize the placement of the injection needles [62].

Experimental protocols

Seventeen groups of animals were used in this study (Fig. 1): three sets of animals that received bilateral microinjections of 500 nL of NMDA receptor antagonist (DL-AP7): 1 nmol (n = 8), 10 nmol (n = 8), or 50 nmol (n = 14) into the DH [27]; three sets of animals that received bilateral microinjections of 500 nL of AMPA/Kainate receptor antagonist (NBQX): 10 nmol (n = 7), 30 nmol (n = 5), or 100 nmol (n = 7) into the DH [66]; three sets of animals that received bilateral microinjections of 500 nL of GABAA receptor antagonist (Bicuculline): 0.01 nmol (n = 7), 0.1 nmol (n = 9), or 1 nmol (n = 10) into the DH [78]; three sets of animals that received bilateral microinjections of 500 nL of CB1 receptor antagonist (AM251): 1 pmol (n = 10), 10 pmol (n = 11), or 100 pmol (n = 9) into the DH [24]; one set of animals that received bilateral microinjections of 500 nL of NMDA receptor antagonist (1 nmol) plus CB1 receptor antagonist (100 pmol) (n = 12) into the DH; another two sets that received microinjections of 500 nL of DMSO 10% (n = 7) or NaOH 100 nM (n = 7) into the DH; and finally those ones that received bilateral microinjections of 500 nL of NMDA receptor antagonist (10 nmol) (n = 5) or CB1 receptor antagonist 100 pmol (n = 7) in the surrounding structures of the DH.

Fig. 1
figure 1

Schematic drawing of the general experimental procedure. Vasoactive drugs (phenylephrine and sodium nitroprusside) were intravenously administered before, 10 min after, and 60 min after the pharmacological manipulation of the DH

Full size image

Results

Figure 2 shows a photomicrograph of a coronal section of rat brain implanted with bilateral guide cannulae into the DH and diagrammatic representations of injection sites of drugs and vehicles administered into the DH or in the surrounding structures of the DH.

Fig. 2
figure 2

Photomicrograph of a coronal section of the hippocampus and the injection sites within the DH. Diagrammatic representations with the injection sites of the vehicle (gray circles), drugs into the DH (black circles), and in surrounding structures (white circles) of the DH. Coordinates according to the Rat Brain Atlas of Paxinos and Watson (1997)

Full size image

Effects of NMDA receptor antagonist (DL-AP7 1, 10, or 50 nmol) injected into the DH

NMDA receptor antagonist, DL-AP7, injected into the DH did not alter the basal levels of MAP or HR (Table 1). Non-linear regression analysis showed that NMDA receptor antagonist 1 nmol (n = 8) did not change any parameters analyzed in sigmoid curve (Table 2 and Fig. 3). However, both doses 10 nmol (n = 8) and 50 nmol (n = 14) reduced the cardiac baroreflex activity (Table 2 and Fig. 3). Moreover, linear regression analysis shows that there were no differences in linear regression slopes of bradycardic (before = − 1.78 ± 0.15, 10 min after = − 1.69 ± 0.29, and 60 min after = − 1.68 ± 0.13 bpm/mmHg; F (2, 23) = 0.10; P > 0.05) and tachycardic (before = − 2.49 ± 0.33, 10 min after = − 2.41 ± 0.15, and 60 min after = − 2.37 ± 0.28 bpm/mmHg; F (2, 23) = 0.10; P > 0.05) responses after administration of DL-AP7 1 nmol (Fig. 3). However, DL-AP7 10 nmol reduced the bradycardic (before = − 1.95 ± 0.17 and 10 min after = − 1.33 ± 0.13 bpm/mmHg; F (2, 23) = 3.55; P < 0.05) and tachycardic (before = − 2.33 ± 0.18 and 10 min after = − 1.43 ± 0.15 bpm/mmHg; F (2, 23) = 5.40; P < 0.05) responses (Fig. 2). Likewise, DL-AP7 50 nmol reduced the bradycardic (before = − 2.23 ± 0.25 and 10 min after = − 1.47 ± 0.18 bpm/mmHg; F (2, 41) = 3.61; P < 0.05) and tachycardic (before = − 2.07 ± 0.22 and 10 min after = − 1.32 ± 0.19 bpm/mmHg; F (2, 41) = 4.15; P < 0.05) responses (Fig. 3). The effect of DL-AP7 on the baroreflex activity was reverted after 60 min (Fig. 3).

Table 1 Injection effects of DL-AP7 (NMDA antagonist), NBQX (AMPA/Kainate antagonist), Bicuculline (GABAA antagonist), or AM251 (CB1 antagonist) into the DH on basal levels of MAP and HR
Full size table
Table 2 Parameters of sigmoidal curves before, 10 and 60 min after injections of DL-AP7 1 nmol (n = 8), DL-AP7 10 nmol (n = 8), DL-AP7 100 nmol (n = 14), NBQX 10 nmol (n = 7), NBQX 30 nmol (n = 5), or NBQX 100 nmol (n = 7) into the DH
Full size table
Fig. 3
figure 3

Left side: sigmoid curves generated before (r 2 = 0.86; 0.93; 0.84) and 10 min (r 2 = 0.86; 0.94; 0.78) and 60 min (r 2 = 0.88; 0.91; 0.84) after the administration into the DH of the NMDA receptor antagonist (DL-AP7) in doses of 1 nmol (n = 8), 10 nmol (n = 8), or 50 nmol (n = 14). Right side: bradycardia linear regression before (r 2 = 0.83; 0.78; 0.54) and 10 min (r 2 = 0.53; 0.75; 0.50) and 60 min (r 2 = 0.85; 0.73; 0.56) after the DL-AP7 injection into the DH. Tachycardia linear regression before (r 2 = 0.65; 0.85; 0.57) and 10 min (r 2 = 0.90; 0.75; 0.59) and 60 min (r 2 = 0.70; 0.69; 0.59) after the DL-AP7 injection into the DH

Full size image

Effects of AMPA/Kainate receptor antagonist (NBQX 10, 30, or 100 nmol) injected into the DH

Bilateral administration of AMPA/Kainate receptor antagonist NBQX into the DH did not alter the basal levels of MAP or HR (Table 1). Non-linear regression analysis showed that AMPA/Kainate receptor antagonist NBQX 10 nmol (n = 7) and 30 nmol (n = 5) into the DH did not change any parameters analyzed in sigmoid curve (Table 2). However, the dose of 100 nmol (n = 7) decreased the cardiac baroreflex activity (Table 2 and Fig. 4). Linear regression analysis showed that there were no differences in linear regression slopes after administration of NBQX 10, 30, or 100 nmol. NBQX 10 nmol did not change bradycardic (before = − 1.80 ± 0.36, 10 min after = − 2.07 ± 0.23, and 60 min after = − 2.02 ± 0.27 bpm/mmHg; F (2, 20) = 0.22; P > 0.05) and tachycardic (before = − 1.83 ± 0.19, 10 min after = − 1.92 ± 0.36, and 60 min after = − 1.98 ± 0.26 bpm/mmHg; F (2, 20) = 0.10; P > 0.05) responses. NBQX 30 nmol did not change bradycardic (before = − 2.00 ± 0.31, 10 min after = − 1.77 ± 0.38, and 60 min after = − 1.80 ± 0.35 bpm/mmHg; F (2, 14) = 0.13; P > 0.05) and tachycardic (before = − 2.13 ± 0.69, 10 min after = − 1.54 ± 0.37, and 60 min after = − 2.04 ± 0.61 bpm/mmHg; F (2, 14) = 0.31; P > 0.05) responses. NBQX 100 nmol did not change bradycardic (before = − 1.52 ± 0.24, 10 min after = − 1.91 ± 0.28, and 60 min after = − 1.62 ± 0.15 bpm/mmHg; F (2, 20) = 0.73; P > 0.05) and tachycardic (before = − 2.72 ± 0.56, 10 min after = − 2.03 ± 0.35, and 60 min after = − 2.83 ± 0.31 bpm/mmHg; F (2, 20) = 1.06; P > 0.05) responses (Fig. 4).

Fig. 4
figure 4

Left side: sigmoid curves generated before (r 2 = 0.86; 0.75; 0.78) and 10 min (r 2 = 0.85; 0.80; 0.81) and 60 min (r 2 = 0.88; 0.77; 0.89) after the administration into the DH of the AMPA/Kainate receptor antagonist (NBQX) in doses of 10 nmol (n = 7), 30 nmol (n = 5), or 100 nmol (n = 7). Right side: bradycardia linear regression before (r 2 = 0.50) and 10 min (r 2 = 0.67) and 60 min (r 2 = 0.68) after the NBQX injection into the DH. Tachycardia linear regression before (r 2 = 0,78) and 10 min (r 2 = 0.53) and 60 min (r 2 = 0.69) after the NBQX injection into the DH

Full size image

Effects of GABAA receptor antagonist (Bicuculline 0.01, 0.1, or 1 nmol) injected into the DH

GABAA receptor antagonist Bicuculline injected into the DH did not alter the basal levels of MAP or HR (Table 1). Non-linear regression analysis showed that GABAA receptor antagonist 0.01 nmol (n = 7) and 0.1 nmol (n = 9) did not change any parameters analyzed in sigmoid curve (Table 2 and Fig. 5). However, the dose of 1 nmol (n = 10) increased the cardiac baroreflex activity (Table 3 and Fig. 5). Moreover, linear regression analysis shows that there were no differences in linear regression slopes of bradycardic (before = − 2.26 ± 0.22, 10 min after = − 2.59 ± 0.22, and 60 min after = − 2.01 ± 0.21 bpm/mmHg; F (2, 20) = 1.35; P > 0.05) and tachycardic (before = − 2.36 ± 0.21, 10 min after = − 2.41 ± 0.27, and 60 min after = − 2.19 ± 0.28 bpm/mmHg; F (2, 20) = 0.20; P > 0.05) responses after administration of GABAA 0.01 nmol (Fig. 5) as well as there were no differences in linear regression slopes of bradycardic (before = − 2.25 ± 0.27, 10 min after = − 2.54 ± 0.39, and 60 min after = − 2.31 ± 0.23 bpm/mmHg; F (2, 26) = 0.26; P > 0.05) and tachycardic (before = − 2.03 ± 0.25, 10 min after = − 2.28 ± 0.34, and 60 min after = − 2.20 ± 0.36 bpm/mmHg; F (2, 26) = 0.15; P > 0.05) responses after administration of GABAA 0.1 nmol. However, GABAA 1 nmol increased the bradycardic (before = − 1.59 ± 0.19 and 10 min after = − 2.19 ± 0.13 bpm/mmHg; F (2, 29) = 4.59; P < 0.05) response, but this dose did not change the tachycardic (before = − 2.16 ± 0.27, 10 min after = − 2.75 ± 0.40, and 60 min after = − 2.17 ± 0.30 bpm/mmHg; F (2, 29) = 1.06; P > 0.05) response (Fig. 5). The effect of GABAA 1 nmol on the baroreflex activity was reverted after 60 min (Fig. 5).

Fig. 5
figure 5

Left side: sigmoid curves generated before (r 2 = 0.90; 0.86; 0.85) and 10 min (r 2 = 0.91; 0.82; 0.89) and 60 min (r 2 = 0.88; 0.84; 0.81) after the administration into the DH of the GABAA receptor antagonist (Bicuculline) in doses of 0.01 nmol (n = 7), 0.1 nmol (n = 9) or 1 nmol (n = 10). Right side: bradycardia linear regression before (r 2 = 0.60; 0.72; 0.71) and 10 min (r 2 = 0.77; 0.78; 0.82) and 60 min (r 2 = 0.72; 0.76; 0.70) after the AM251 injection into the DH. Tachycardia linear regression before (r 2 = 0,73; 0,64; 0,55) and 10 min (r 2 = 0.70; 0.87; 0.66) and 60 min (r 2 = 0.59; 0.63; 0.57) after the AM251 injection into the DH

Full size image
Table 3 Parameters of sigmoidal curves before, 10 and 60 min after injections of Bicuculline 0.01 nmol (n = 7), Bicuculline 0.1 nmol (n = 9), Bicuculline 1 nmol (n = 10), AM251 1 pmol (n = 10), AM251 10 pmol (n = 11), AM251 100 pmol (n = 9), or AM251 100 pmol + DL-AP7 1 nmol (n = 12) into the DH
Full size table

Effects of CB1 receptor antagonist (AM251 1, 10, or 100 pmol) injected into the DH

CB1 receptor antagonist injection into the DH did not alter the basal levels of MAP or HR (Table 1). Non-linear regression analysis showed that CB1 receptor antagonist AM251 1 pmol (n = 10) did not change any parameters analyzed in sigmoid curve (Table 3 and Fig. 6). However, both doses of 10 pmol (n = 11) and 100 pmol (n = 9) increased the cardiac baroreflex activity (Table 3 and Fig. 6) analyzed in the sigmoid curve. Moreover, linear regression analysis shows no differences in linear regression slopes of bradycardic (before = − 1.51 ± 0.28, 10 min after = − 1.75 ± 0.16, and 60 min after = − 1.57 ± 0.16 bpm/mmHg; F (2, 29) = 0.37; P > 0.05) and tachycardic (before = − 1.73 ± 0.17, 10 min after = − 2.23 ± 0.24, and 60 min after = − 1.80 ± 0.25 bpm/mmHg; F (2, 29) = 1.55; P > 0.05) responses after the administration of AM251 1 pmol (Fig. 6). However, AM251 10 pmol increased the bradycardic (before = − 1.45 ± 0.14 and 10 min after = − 1.94 ± 0.13 bpm/mmHg; F (2, 32) = 3.84; P < 0.05) and tachycardic (before = − 1.92 ± 0.12 and 10 min after = − 2.59 ± 0.15 bpm/mmHg; F (2, 32) = 3.71; P < 0.05) responses (Fig. 6). Likewise, AM251 100 pmol increased bradycardic (before = − 1.55 ± 0.17 and 10 min after = − 2.28 ± 0.18 bpm/mmHg; F (2, 26) = 4.28; P < 0.05) and tachycardic (before = − 1.62 ± 0.20 and 10 min after = − 2.39 ± 0.20 bpm/mmHg; F (2, 26) = 3.46; P < 0.05) responses (Fig. 6). The effect of AM251 on the baroreflex activity was reverted after 60 min (Fig. 6).

Fig. 6
figure 6

Left side: sigmoid curves generated before (r 2 = 0.85; 0.81; 0.87) and 10 min (r 2 = 0.89; 0.91; 0.89) and 60 min (r 2 = 0.85; 0.84; 0.86) after the administration into the DH of the CB1 receptor antagonist (AM251) in doses of 1 pmol (n = 10), 10 pmol (n = 11), or 100 pmol (n = 9). Right side: bradycardia linear regression before (r 2 = 0.60; 0.72; 0.71) and 10 min (r 2 = 0.77; 0.78; 0.82) and 60 min (r 2 = 0.72; 0.76; 0.70) after the AM251 injection into the DH. Tachycardia linear regression before (r 2 = 0.73; 0.64; 0.55) and 10 min (r 2 = 0.70; 0.87; 0.66) and 60 min (r 2 = 0.59; 0.63; 0.57) after the AM251 injection into the DH

Full size image

Effects of CB1 receptor antagonist (AM251 100 pmol) + NMDA receptor antagonist (DL-AP7 1 nmol) injected into the DH

Bilateral administration of AM251 100 pmol and DL-AP7 1 nmol into the DH (n = 12) did not alter the basal levels of MAP (before = 101 ± 3 and after = 104 ± 4 mmHg; t = 1.2; P > 0.05) or HR (before = 358 ± 9 and after = 353 ± 8 bpm; t = 1.3; P > 0.05). Non-linear regression analysis showed that DL-AP7 1 nmol into DH was able to inhibit the effect of AM251 100 pmol on baroreflex activity (Table 3 and Fig. 7). Corroborating these results, linear regression analysis showed that DL-AP7 1 nmol into DH was able to inhibit the effect of AM251 100 pmol on linear regression slopes of bradycardic (before = − 1.72 ± 0.26, 10 min after = − 2.11 ± 0.27, and 60 min after = − 1.91 ± 0.24 bpm/mmHg; F (2, 35) = 0.58; P > 0.05) and tachycardic (before = − 2.39 ± 0.26, 10 min after = − 2.68 ± 0.32, and 60 min after = − 2.52 ± 0.26 bpm/mmHg; F (2, 35) = 0.26; P > 0.05) responses (Fig. 7).

Fig. 7
figure 7

Left side: sigmoid curves before (r 2 = 0.86) and 10 min (r 2 = 0.85) and 60 min (r 2 = 0.87) after the co-administration of the drugs AM251 100 pmol (CB1 receptor antagonist) and DL-AP7 1 nmol (NMDA receptor antagonist) into the DH (n = 12). Right side: bradycardia linear regression before (r 2 = 0.51) and 10 min (r 2 = 0.59) and 60 min (r 2 = 0.58) after the co-administration of the drugs AM251 100 pmol and DL-AP7 1 nmol into the DH. Tachycardia linear regression before (r 2 = 0.70) and 10 min (r 2 = 0.61) and 60 min (r 2 = 0.67) after the co-administration of the drugs AM251 100 pmol and DL-AP7 1 nmol into the DH

Full size image

Effect on baroreflex activity of vehicles (DMSO 10% or NaOH 100 nM) administered into the DH and drugs (DL-AP7 10 nmol or AM251 100 pmol) injected in the surrounding structures of the DH

DMSO 10% bilaterally administered into the DH (n = 7) did not alter the basal levels of MAP (before = 92 ± 2 and after = 88 ± 1 mmHg; t = 1.1; P > 0.05) or HR (before = 362 ± 11 and after = 358 ± 9 bpm; t = 0.9; P > 0.05). Moreover, linear regression analysis showed that there were no differences in linear regression slopes of bradycardic (before = − 1.90 ± 0.16 and after = − 2.11 ± 0.28 bpm/mmHg; t = 0.6; P > 0.05) and tachycardic (before = − 2.42 ± 0.23 and after = − 2.41 ± 0.19 bpm/mmHg; t = 0.1; P > 0.05) responses (Fig. 8). Likewise, bilateral administration of NaOH 100 nM into DH (n = 7) did not alter the basal levels of MAP (before = 99 ± 3 and after = 96 ± 3 mmHg; t = 1.3; P > 0.05) or HR (before = 361 ± 9 and after = 357 ± 8 bpm; t = 1.1; P > 0.05). Linear regression analysis showed that there were no differences in linear regression slopes of bradycardic (before = − 1.71 ± 0.17 and after = − 1.62 ± 0.18 bpm/mmHg; t = 0.3; P > 0.05) and tachycardic (before = − 1.73 ± 0.14 and after = − 1.91 ± 0.15 bpm/mmHg; t = 0.74; P > 0.05) responses (Fig. 8).

Fig. 8
figure 8

NaOH 100 nM: linear regression generated before (r 2 = 0.78) and 10 min (r 2 = 0.85) after the NaOH 100 nM injection into the DH (n = 7). DMSO 10%: linear regression generated before (r 2 = 0.85) and 10 min (r 2 = 0.82) after the DMSO 10% injection into the DH (n = 7). DL-AP7 10 nmol: linear regression generated before (r 2 = 0.86) and 10 min (r 2 = 0.85) after the DL-AP7 10 nmol injection in the surrounding structures of the DH (n = 5). AM251 100 pmol: linear regression generated before (r 2 = 0.86) and 10 min (r 2 = 0.82) after the AM251 100 pmol injection in the surrounding structures of the DH (n = 7)

Full size image

In the same way, bilateral administration of DL-AP7 10 nmol in the surrounding structures of the DH (n = 5) did not alter the basal levels of MAP (before = 103 ± 5 and after = 104 ± 4 mmHg; t = 1.1; P > 0.05) or HR (before = 352 ± 7 and after = 355 ± 9 bpm; t = 1.0; P > 0.05). Linear regression analysis showed that there were no differences in linear regression slopes of bradycardic (before = − 2.21 ± 0.15 and after = − 2.45 ± 0.16 bpm/mmHg; t = 1.1; P > 0.05) and tachycardic (before = − 2.27 ± 0.28 and after = − 2.42 ± 0.15 bpm/mmHg; t = 0.4; P > 0.05) responses (Fig. 8). Moreover, bilateral administration of AM251 100 pmol in the surrounding structures of the DH (n = 7) did not alter the basal levels of MAP (before = 99 ± 7 and after = 103 ± 6 mmHg; t = 0.8; P > 0.05) or HR (before = 348 ± 9 and after = 352 ± 9 bpm; t = 0.9; P > 0.05). Linear regression analysis showed that there were no differences in linear regression slopes of bradycardic (before = − 1.79 ± 0.22 and after = − 1.68 ± 0.18 bpm/mmHg; t = 0.5; P > 0.05) and tachycardic (before = − 1.80 ± 0.12 and after = − 1.82 ± 0.18 bpm/mmHg; t = 0.6; P > 0.05) responses (Fig. 8).

Discussion

The present work is the first to demonstrate that the glutamatergic, GABAergic, and endocannabinoid neurotransmissions within the DH are involved in the modulation of the cardiac baroreflex function. Our results show that glutamate within the hippocampus facilitates the cardiac baroreflex function through acting at NMDA and AMPA/Kainate receptors while both neurotransmitters GABA, acting at GABAA receptors, and endocannabinoids, acting at CB1 receptors, reduce cardiac baroreflex responses. Moreover, we found an interaction between glutamatergic and endocannabinoid neurotransmission in the DH in modulating the cardiac baroreflex function once NMDA receptor antagonist in the dose of 1 nmol inhibited the enhancement of the cardiac baroreflex response promoted by the dose of 100 pmol of CB1 receptor antagonist. Furthermore, the vehicles of the drugs administered into the DH did not alter cardiac baroreflex responses, as well as doses of the antagonists administered in the DH surrounding areas, did not change the cardiac baroreflex function. Thus, these results reinforce a site-dependent effect of the drugs injected into the DH.

The baroreflex provides an essential feedback to the CNS for moment-to-moment control of the cardiovascular function to maintain BP within a narrow functional range [33, 63]. Baroreceptors are peripheral afferent neurons located in the aortic arch and carotid sinus [22]. They convert mechanical stimuli derived from oscillations in BP into action potentials driven to the nucleus of the solitary tract (NTS) [22, 33]. The NTS is a medullary area that integrates the BP information from the baroreceptors and sends projections to other brainstem areas influencing the sympathetic and parasympathetic drive to heart and vessels [33]. Several studies have shown that forebrain areas can also modulate the baroreflex function: medial prefrontal cortex (MPFC) [26], bed nucleus of stria terminalis (BNST) [15], hypothalamus [14], amygdala [29], as well as the hippocampus. We have shown that the DH blockade increases the heart rate reflex responses, suggesting that the DH has an inhibitory influence on cardiac baroreflex function [28].

In spite of DH does not project directly to brainstem areas involved in cardiovascular control [11, 33, 57], it has extensive connections with the BNST [17, 19, 61]. BNST plays a role in the baroreflex modulation [1, 2, 15] and sends projections to brainstem areas related to the neurocircuitry of the baroreflex [31, 32, 40]. Thus, BNST may be a “relay area” to the modulatory action of the DH in the cardiac baroreflex function. Further studies are required to verify the pathways involved in the baroreflex modulation promoted by the DH.

Although the DH has been classically implicated in memory and spatial navigation [21, 65], studies show the involvement of the DH in autonomic responses during aversive conditions. The rats’ DH blockade before the re-exposure to an aversive context did not change the behavioral consequences (“freezing”) of animals, but it reduced the increase in both BP and HR [67]. Indeed, DH inhibition may also attenuate both BP and HR increase elicited by acute restraint stress [73]. Thus, these results suggest that DH plays a major role in cardiovascular modulation in threat situations.

In the present study, glut facilitates baroreflex function via NMDA and non-NMDA receptor activation. In the same way, a previous work showed that the administration of a NMDA receptor antagonist in the ventral hippocampus (VH) was able to inhibit the increase in the HR and BP promoted by the local infusion of glut in the VH [70]. Moreover, the intravenous pretreatment with either a ganglionic blocker or a β1-adrenergic receptor antagonist abolished those cardiovascular responses glut-evoked in the VH [70]. Thus, these results suggest that glut elicits cardiovascular responses through NMDA receptor in the VH and these cardiovascular responses glut-evoked are mediated by the sympathetic nervous system.

The injection of a NMDA receptor antagonist into the DH attenuates the increase in the BP and HR as well as the reduction in tail skin temperature of rats submitted to the restraint stress model [56]. Moreover, administration of a NMDA receptor antagonist into the DH reduces the “freezing behavior” and cardiovascular responses of animals submitted to the contextual fear conditioning model [23]. Taken together, these results suggest that NMDA receptors within the DH have a facilitative influence in stress-evoked behavior and autonomic responses.

On the other hand, administration of a CB1 receptor antagonist into the DH reduces the “freezing behavior” and cardiovascular responses of animals submitted to the contextual fear conditioning model [78]. Furthermore, the NMDA receptor antagonist was able to inhibit the increase in both “freezing behavior” and cardiovascular responses promoted by the CB1 receptor antagonist [78], suggesting that the changes evoked by CB1 receptors involve the glutamate and its action on NMDA receptors. These results are in accordance with the present study showing that endocannabinoid system modulates hippocampal glutamate-related responses.

In the hippocampus, CB1 receptors are found in glutamatergic neurons [45]. Excessive activation of glutamatergic transmission in the hippocampus is considered as a key pathogenic event leading to epileptiform seizures [6]. Endocannabinoid system is important in the control of neuronal activity through CB1 receptors [30, 52]. Balanced control of neuronal activity is central in maintaining function and viability of neuronal circuits [6]. The endocannabinoid system tightly controls neuronal excitability, providing substantial endogenous protection against kainic acid (KA)-induced seizures [55]. Cardiovascular autonomic dysfunction in seizures is a major cause of sudden unexpected death in epilepsy [53]. Seizures are associated with altered autonomic activity [54] and decrease heart rate variability [80]. Moreover, a study in rats showed that sustained hippocampal seizure activity was accompanied by progressive baroreflex impairment [81]. Since our results showed that endocannabinoids and glutamate in the hippocampus are important to control of baroreflex function, we suggest that the imbalance in endocannabinoid and glutamatergic activity found in seizure patients may contribute to autonomic and baroreflex failure in the epileptic crisis.

Glutamatergic pyramidal neurons are the main postsynaptic target of GABAergic interneurons [12]. These interneurons control different domains of glutamatergic principal neurons at precise moments during hippocampal activity [12, 75], and the activation of these interneurons results in inhibitory postsynaptic currents (IPSCs) in pyramidal cells [82]. Therefore, since the GABA interneurons in the hippocampus are arranged in a configuration that permits to control the activity of glutamatergic pyramidal neurons, we speculate that GABAergic and endocannabinoid neurotransmissions modulate the glutamatergic neurotransmission activity within the DH controlling the baroreflex function (Fig. 9).

Fig. 9
figure 9

A putative mechanism involving glutamate, GABA, and endocannabinoids within the dorsal hippocampus in the control of baroreflex function. We have suggested that presynaptic glutamate release facilitates the baroreflex function through NMDA and AMPA/Kainate receptors activation. However, the NMDA receptor activation may elicit the endocannabinoid synthesis in the postsynaptic neuron that in turn activate cannabinoid receptor subtype 1 (CB1) in the presynaptic neuron decreasing the baroreflex function through the inhibition of glutamate exocytose. On the other hand, GABA interneurons also modulate the activity of glutamatergic principal neurons. The inhibitory actions of GABA on GABAA receptors hyperpolarize the glutamatergic neurons, decreasing the baroreflex function

Full size image

Conclusion

During defensive reactions, there is a “delay” in the baroreflex response that allows an increase in HR along with a rise in BP [16, 58, 72]. In the present study, we have shown that endocannabinoid neurotransmission through the CB1 receptor or GABAergic neurotransmission through the GABAA receptor in the DH may blunt heart rate baroreflex responses. Therefore, we speculate that the activation of these pathways within the DH may modulate the baroreflex function in threat situations to adjust the blood perfusion to organs and tissues needed for flight or fight since the baroreflex function is reduced during both an aversive condition and activating CB1 or GABAA receptors.

However, our results are limited in demonstrating the involvement of the glut, GABA, and endocannabinoids within the DH in the cardiac baroreflex modulation. The neural pathways upon which the DH modulates the baroreflex function as well as the conditions in which the DH is necessary for the baroreflex function will be the target of future studies. Therefore, our findings support the possibility that the endocannabinoid, glutamatergic, and GABAergic neurotransmissions interact with each other within the DH to modulate the cardiac baroreflex function (Fig. 9).