Introduction

Anxiety is a common mental disorder worldwide and is strongly associated with poor quality of life [1]. Despite advances in the elucidation of the pathophysiology of anxiety disorders, their causal factors, etiology, and mechanisms remain relatively poorly understood [2]. However, there is a consensus that the mechanisms involved in anxiety include multiple neurotransmitter systems, resulting in a wide range of pharmacological targets for treating anxiety disorders [3].

In addition to the classical pharmacological targets, the GABAergic [4] and serotoninergic [5] systems, particular emphasis has been placed on the development of novel drugs that focus on the glutamatergic and purinergic systems [68]. Glutamate (GLU) is the main excitatory neurotransmitter in the central nervous system (CNS) and acts through ionotropic and metabotropic receptors [9]. Changes in anxiety phenotypes have been observed after the administration of GLU receptor modulators [2, 3, 8, 10]. Moreover, glutamatergic neurotransmission is directly influenced by extracellular GLU levels, which are regulated by its release from neuronal pre-synaptic terminals and/or by its uptake mainly by astrocytic GLU transporters [8, 11]. Recently, the control of extracellular GLU levels has been proposed as a target for the development of novel anxiolytic drugs [2, 3, 8, 10].

Adenosine (ADO) neurotransmission has also been associated with anxiety-related responses [12] based on the modulatory effects of ADO on vesicular pre-synaptic GLU release [13]. ADO receptors have dual effects on GLU release: the activation of ADO A1 receptors is associated with inhibition [13], whereas the activation of ADO A2a receptors is more likely to facilitate GLU release [7, 14]. Therefore, anxiolytic/anxiety-related behaviors have been demonstrated by pharmacological and/or genetic manipulation of both receptors [7, 12, 15]. Moreover, although less abundant than A1R, the ADO A2aR as well as A1R receptors play important roles in glutamatergic neurons of the hippocampus and other limbic structures, brain regions closely related to anxiety behavior, hence strengthening their potential as targets for treating anxiety [15].

Several findings support the hypothesis that guanine-based purines (GBPs) are able to modulate both adenosinergic and glutamatergic neurotransmission in vivo [1618]. Indeed, the administration of GBPs, especially the nucleoside guanosine (GUO), counteracts the harmful effects of glutamatergic excitotoxicity in several experimental models [1619]. Experimental evidence indicates that the neuroprotective effects exerted by GUO appear to be mediated by the stimulation of astrocytic GLU uptake, which clears excessive GLU from the synaptic cleft [20, 21] and keeps the extracellular GLU levels at physiological concentrations. Interestingly, at least some of the observed neuroprotective effects of GUO appear to be blocked by ADO receptor antagonists [17, 22]. However, the stimulation of astrocytic GLU uptake by GBPs is not influenced by caffeine (CAF), a nonspecific ADO receptor antagonist [20], suggesting additional mechanisms of action for GUO that involve both the adenosinergic and glutamatergic systems. Thus, to the best of our knowledge, the interplay between GUO and the adenosinergic and glutamatergic systems remains elusive.

Additionally, our group has already shown that systemic administration of guanosine-5′-monophosphate (GMP) induced anxiolytic-like behaviors in rats [6], and the induction of a similar effect by chronic GUO administration was also previously described in mice [23]. However, the mechanisms involved in these effects are still not clear. Thus, the present study aimed to investigate the anxiolytic potential of acute in vivo GUO administration in rats, with particular emphasis on mechanisms involving the adenosinergic and glutamatergic neurotransmitter systems. Here, the potential anxiolytic-like effect of GUO was evaluated in three anxiety-related paradigms: the elevated plus maze (EPM), and light/dark and round open field tasks. Additionally, ADO and GLU levels in cerebrospinal fluid (CSF), GLU uptake (ex vivo and in vitro protocols), and in vitro synaptosomal GLU release were also evaluated to investigate the putative molecular targets and signaling pathways recruited by GUO. The results presented here illustrate the involvement of the adenosinergic and glutamatergic systems on in vivo GUO anxiolytic-like behavior.

Materials and Methods

Animals

Male Wistar rats, 60–90 days old (250–300 g), were kept under a 12-h light/dark cycle (light on at 7:00 AM) at 22 ± 1 °C in plastic cages (5 per cage) with water and food available ad libitum. On the day of the behavioral tasks, they were acclimated to the behavioral room with appropriate lighting (25 lx) for 1 h before the behavioral procedures. The animals were maintained according to the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All protocols followed the Ethical Committee of the Federal University of Rio Grande do Sul (Project number 18236).

Chemicals

GMP, GUO, CAF, 2-chloro-N 6-cyclopentyladenosine (CCPA, an ADO A1 receptor agonist), 2-p-(2-carboxyethyl)phenethylamino-5′-N-ethylcarboxamidoadenosine hydrochloride hydrate (CGS21680), 8-cyclopentyl-1,3-dipropylxanthine (DPCPX, an ADO A1 receptor antagonist), and ZM241385 (an ADO A2a receptor antagonist) were from Sigma Chemicals (St. Louis, MO, USA). l-[3H] glutamate (specific activity of 50 Ci/mmol) was from Amersham International, UK. The anesthetic sodium thiopental was from Cristália (Itapira, SP, Brazil).

Drugs were dissolved in saline (0.9 % NaCl) and administered i.p. for in vivo evaluation. GUO and CAF were dissolved in Hank’s buffered salt solution (HBSS) containing (in mM) the following: 137 NaCl, 0.63 Na2HPO4, 4.17 NaHCO3, 5.36 KCl, 0.44 KH2PO4, 1.26 CaCl2, 0.41 MgSO4, 0.49 MgCl2, and 5.5 glucose (pH = 7.4) for in vitro GLU uptake and release assays. CCPA, CGS21680, DPCPX, and ZM241385 were dissolved with DMSO (0.1 %) in HBSS.

For the Western blotting assay, bovine serum albumin, a protease and phosphatase inhibitor cocktail, and antibodies against glial fibrillary acid protein (GFAP) and synaptosomal-associated protein 25 (SNAP 25) were from Sigma Chemicals. The antibody against the N-methyl-d-aspartic receptor subunit (NR1) was from Chemicon; the antibody against postsynaptic density protein 95 (PSD 95) was from Affinity BioReagents; the antibodies against excitatory amino acid carrier 1 (EAAC1) and glutamate transporter-1 (GLT-1) were from Alpha Diagnostic; the antibody against 2-amino-3-(5-methyl-3-oxo-1,2-oxazol-4-yl) propanoic acid (AMPA) receptor subunit 1 (GluA1; AMPA receptor subunit) was from UpState; and the antibodies against monoamine oxidase A (MAO A), β-tubulin III, and synaptobrevin (VAMP) were from Santa Cruz Biotechnology.

Drug Administration

The experimental design and drug administration schedules are shown in Fig. 1.

Fig. 1
figure 1

Experimental design and drug administration. Schematic representation of the experimental protocols, including the drug administrations, behavioral tasks, and neurochemical analysis performed, for protocol I (A and B) and protocol II

Protocol I

As depicted in Fig. 1—protocol IA, the rats were divided into five groups: saline, 50 mg/kg GMP (positive control), and 7.5, 30, or 60 mg/kg GUO. GMP 50 mg/kg was used as a positive control for anxiolytic-like effects [6]. The GUO doses were carefully chosen based on previously studies from our group, which demonstrated important neuromodulatory effects of GUO in different in vivo protocols [18, 19]. Sixty minutes after intraperitoneal (i.p.) administration of these compounds, the animals were subjected to an EPM task. After the behavioral procedure, CSF was immediately collected for purine content evaluation.

After assessing the potential anxiolytic GUO dose in the EPM, a well-established and widely used task to detect anxiolytic/anxiety-like phenotypes in rodents, the anxiolytic-like effect of 7.5 mg/kg GUO was also evaluated in two other anxiety-related paradigms, the light/dark and the round open field tasks (Fig. 1—protocol 1B). The rats were divided into two groups: saline or 7.5 mg/kg GUO. The same schedule of administration was maintained as previously mentioned.

Protocol II

As depicted in Fig. 1, the rats were divided into four groups (saline/saline, 10 mg/kg CAF/saline, saline/7.5 mg/kg GUO, and 10 mg/kg CAF/7.5 mg/kg GUO). The CAF dose was chosen based on the literature [24] and by preliminary pilot experiments (Material Supplementary 1). Pre-administration of saline or CAF i.p. was performed 15 min before i.p. administration of saline or GUO. Sixty minutes after the last administration, the EPM task was performed. It is important to highlight that 10 mg/kg CAF per se did not affect the anxiety-related behavior assessed by the EPM task. On the other hand, this CAF dose modulated rats’ performance on the light/dark and round open field tasks (Material Supplementary 1), demonstrating that under different stimulus, CAF 10 mg/kg can act as a psychostimulant.

Immediately after the behavioral task, CSF was collected for the measurement of purines and GLU content and the brain was immediately processed for GLU uptake.

EPM Task

The EPM task was performed as previously described [25]. The apparatus consisted of two open (50 × 10 cm, length × width) and two enclosed (50 × 10 × 40 cm, length × width × height) arms that were separated by a central platform (5 × 5 cm); the height of the maze was 70 cm. The animal’s behavior was recorded for 5 min using ANY-Maze software. The percentage of time spent in the open arms, the total distance travelled in open arms (cm), and the total distance travelled (cm) were determined. When administration of a drug significantly increased the first two behavioral parameters without changing the total number of transitions, the drug was considered anxiolytic.

Light/Dark Task

The light/dark task was performed as previously described [26]. The apparatus consisted of a rectangular acrylic box with two separated compartments. One compartment had black walls and a black floor (without illumination) with a size of 21 × 35 × 41 cm (height × length × width). The other had white walls and a white floor, with a size of 21 × 45 × 41 cm (height × length × width, illuminated by a 100-W lamp placed 45 cm above the center of the box). An 8 × 5 cm (height × length) opening joined both compartments. Each rat was placed in the light compartment facing away from the opening and allowed to explore the box for 5 min. The following behavioral parameters were analyzed by ANY-Maze software: the number of transitions between compartments and the time spent in the light compartment. When the administration of a drug increased these parameters, the drug was considered anxiolytic.

Round Open Field Task

The rats were placed in the center of a round open field apparatus (60 cm in diameter) and let to freely explore the arena for 15 min, as previously described with minor modifications [27]. Total locomotor activity, distance travelled in the center zone, and total time spent in the center zone were recorded for each animal using ANY-Maze software. The treatment was considered anxiolytic when it increases the time spent or the distance travelled in the center zone, in accordance with previous studies [27].

CSF Analysis

Preliminarily, two CSF sampling protocols were tested in order to avoid purine degradation (Supplementary Material 2). The following methodology (method II) showed to be more reliable and it was used in the experiments. The CSF was collected immediately after the EPM task to evaluate purines and GLU levels. The rats were anesthetized with sodium thiopental (40 mg/kg, i.p.) and placed in a stereotaxic apparatus. The CSF (100/150 μL) was collected by direct puncture of the cisterna magna with an insulin syringe (27 gauge × 0.5 in. in length). Samples with visible blood contamination were discarded. All samples were centrifuged at 10,000g at 4 °C for 10 min in an Eppendorf® centrifuge to obtain cell-free supernatants. After the centrifugation, the samples were immediately frozen in dry ice and stored at −80 °C until analysis.

High-Performance Liquid Chromatography

The high-performance liquid chromatography (HPLC) procedures were performed with cell-free CSF aliquots in a Shimadzu Class-VP chromatography system. The system consisted of a quaternary gradient pump with vacuum degassing and piston desalting modules, a Shimadzu SIL-10AF auto injector valve with 50 μL loop and UV and fluorescence detectors, which were used to detect purines and GLU, respectively (Shimadzu, Kyoto, Japan).

The HPLC for purine analysis was performed as previously described [16, 28]. Briefly, the levels of the following purines were determined: ADO, GMP, GUO, inosine (INO), hypoxanthine (HYPOX), xanthine (XANT), and uric acid (UA). The mobile phase flow rate was 1.2 mL/min, and the column temperature was 24 °C. The buffer compositions were unchanged (A: 150 mmol/L phosphate buffer, pH 6.0, containing 150 mmol/L KCl; B: the same buffer with 15 % acetonitrile). The gradient profile was modified according to the content of buffer B in the mobile phase: 0 % at 0.00 min, 2 % at 0.05 min, 7 % at 2.45 min, 50 % at 10.00 min, 100 % at 11.00 min, and 0 % at 12.40 min. Samples of 25 μL were injected. Absorbance was read at 254 nm in a UV detector.

The HPLC for GLU analysis was performed as previously described [29]. The samples were derivatized with ο-phthalaldehyde, the mobile phase flow rate was 1.4 mL/min, and the column temperature was 24 °C. The buffer compositions were as follows: A: 0.04 mol/L sodium phosphate buffer, pH 5.5, containing 20 % methanol; and B: 0.01 mol/L sodium dihydrogen phosphate monohydrate buffer, pH 5.5, containing 80 % methanol. The gradient profile was modified according to the content of buffer B in the mobile phase: 0 % at 0.00 min, 25 % at 13.75 min, 100 % at 15.00–20.00 min, and 0 % at 20.01–25.00 min. Absorbance values were measured at 360 nm excitation and 455 nm emission in a fluorescence detector. Samples of 50 μL were injected.

The CSF levels of the purines and GLU were expressed in micromolar (as the mean ± S.E.M.).

Ex Vivo Na+-Dependent Hippocampal GLU Uptake

Immediately after the EPM task (protocol II), the brain was processed for an ex vivo GLU uptake assay as previously described [30]. From each brain, hippocampal slices of 400 μm were obtained using a McIlwain chopper and individually placed into 24-well plates containing HBSS at 37 °C. The slices were washed once with 1 mL of 37 °C HBSS and then pre-incubated at 37 °C for 15 min. The incubation was started by the addition of 0.33 Ci/mL l-[3H]GLU plus 100 μM (final concentration) GLU and stopped after 5 min with two ice-cold washes with 1 mL of HBSS. The washes were immediately followed by the addition of 0.5 N NaOH. Na+-independent uptake was measured using the same protocol, with modifications in the temperature (4 °C) and medium composition (choline chloride instead of sodium chloride). Na+-dependent uptake was defined as the difference between both uptakes. The incorporated radioactivity was measured in a Hidex 300 SL scintillation counter.

Hippocampal Synaptosomal Preparation

Synaptosomes were prepared from the hippocampus of animals that were not subjected to any treatment or behavioral task according previously published procedures [31]. The hippocampus of each rat was manually homogenized (small capacity Teflon/glass homogenizer in 10 × mL/g) in 10 mM Tris buffer (pH 7.4) with 0.32 M sucrose, 1 mM EDTA, and 0.25 mM dithiothreitol (DTT). The homogenate (H) was centrifuged in microfuge tubes (1.5 mL per tube) at 1000×g for 10 min at 4 °C, using a fixed-angle rotor. The resulting supernatant (S1) was centrifuged at 11,000×g for 20 min at 4 °C to obtain a synaptosomal-enriched pellet (SP), which was washed twice with HBSS (pH 7.4) by centrifugation at 16,000×g for 10 min at 4 °C to remove excess sucrose. The final pellet was resuspended in HBSS buffer (900 μL).

To characterize the quality of the three different stages of the preparation of hippocampal synaptosomes (H, S1, and SP), we performed Western blot analysis to evaluate the immunocontent of neuronal and glial proteins, evaluating the enrichment of each protein in the SP. The following proteins were investigated: β Tub III for the neuronal cytoskeleton; SNAP 25 and VAMP for synaptic vesicles; PSD 95 for postsynaptic densities; EAAC1, NR1, and GLUA1 for neuronal plasma membranes; MAO A for mitochondrial membranes; and GFAP and GLT-1 for astrocytes (Supplementary Material 3).

In Vitro l-[3H] GLU Uptake by Hippocampal Synaptosomal Preparations

Five minutes prior to the measurement of GLU uptake in the SP, 100 μM GUO was added to the incubation medium. Then, 1 μCi/mL l-[3H]GLU plus 50 μM (final concentration) GLU was added, and the mixture was incubated for 15 min at 37 °C. The reaction was terminated by four washes of ice-cold HBSS through centrifugation at 14,000×g for 1 min at 4 °C. The final pellet was resuspended in HBSS, and an aliquot was separated. The radioactivity of this aliquot was determined using a Hidex 300 SL scintillation radioactivity counter.

In Vitro l-[3H] GLU Release from Hippocampal Synaptosomal Preparations

l-[3H] GLU release was measured according to [31], with minor modifications. First, the SP was loaded with l-[3H] GLU using the in vitro l-[3H] GLU uptake assay protocol (described above), without GUO in the incubation medium. Basal l-[3H] GLU release was initiated by the addition of aliquots of loaded synaptosomes in HBSS buffer at 37 °C for 1 min and terminated by immediate centrifugation (14,000×g for 1 min at 4 °C). The percentage of previously loaded radioactivity present in the supernatant was considered the amount of GLU released. K+-stimulated l-[3H]GLU release was assayed as described for basal release, except that the incubation medium contained 40 mM KCl (NaCl decreased accordingly) to induce synaptosomal depolarization. The K+-stimulated GLU release was calculated as the delta (Δ) between both GLU release activities. GLU release was increased by approximately 70 % by high K+, indicating the viability of our preparations, in comparison between the nonstimulated l-[3H]GLU release and the K+-stimulated l-[3H]GLU release (Fig. 6c, d).

To determine the effects of different drugs on GLU release, the final incubation medium (for basal or stimulated GLU release) contained GUO (100 μM), CCPA (100 nM), CGS21680 (30 nM), CAF (1 μM), DPCPX (100 nM), ZM241385 (50 nM), or one of the following combinations: CCPA/GUO, CGS21680/GUO, CAF/GUO, DPCPX/GUO, or ZM241385/GUO. The concentrations of CAF, CCPA, CGS21680, and ZM241385 were chosen based on literature [17, 32, 33], and the concentration of DPCPX was chosen based on our previous data (Supplementary Material 4). Radioactivity was separately determined for supernatants and pellets using a Hidex 300 SL scintillation counter.

Protein Determination

Protein content was measured using the BCA® protein assay kit with bovine serum albumin as a standard.

Statistical Analysis

One-way analysis of variance (ANOVA) followed by Tukey’s post hoc test was used to analyze the effects of GUO on the EPM task. Student’s t test was used to evaluate the effects of GUO on the light/dark and round open field tasks, on CSF purine concentrations, and on both ex vivo and in vitro GLU uptake and to compare basal versus K+-stimulated l-[3H]GLU release from the SP. Correlations between EPM behavioral parameters and ADO or GLU CSF levels were analyzed by Pearson’s correlation. Two-way ANOVA followed by Bonferroni’s post hoc test was used to analyze the influence of CAF pre-administration on the GUO effect, on the EPM task, and on ADO and GLU CSF levels [factors: (1) pre-administration with saline or CAF and (2) administration with saline or GUO] and to analyze the effects of GUO, CCPA, CGS21680, CAF, DPCPX, and ZM241385 on the effect of GUO on synaptosomal GLU release [factors: (1) incubation with CCPA, CGS21680, CAF, DPCPX, or ZM241385 and (2) incubation with GUO]. Differences were considered statistically significant at p < 0.05.

Results

Systemic GUO Administration Induced Anxiolytic-Like Effects and Increased CSF ADO Levels

GMP (50 mg/kg) and GUO (7.5 mg/kg) presented anxiolytic-like effects in the EPM task by significantly enhancing the percentage of the time spent in the open arms (Fig. 2a, **p = 0.0029 and **p = 0.0025, respectively) and the open arms distance travelled (Fig. 2b, *p = 0.0130 and *p = 0.0160, respectively), without significantly affecting total distance travelled (Fig. 2c); higher doses of GUO had no anxiolytic-like effect. Based on these results, only 7.5 mg/kg GUO was used in subsequent experiments. In the light/dark task, 7.5 mg/kg GUO presented anxiolytic-like effects by significantly increasing the number of transitions between the compartments (Fig. 3a, *p = 0.0395) and the time spent in the light compartment (Fig. 3b, *p = 0.0086). In the round open field task, 7.5 mg/kg GUO presented an anxiolytic-like effect by significantly enhancing the time spent in the center zone (Fig. 3d, *p = 0.0278), without affecting the total distance travelled or the distance travelled in the center zone (Figs. 3c and 4e, p = 0.5342 and p = 0.0698, respectively).

Fig. 2
figure 2

Systemic administration of GUO-induced anxiolytic-like behavior in the EPM task. The percentage of time spent in the open arms (a), the total distance travelled in open arms (cm) (b), and the total distance travelled (cm) (c) were evaluated in the EPM task 60 min after i.p. administration of saline, 50 mg/kg GMP or 7.5, 30, or 60 mg/kg GUO. Data are reported as the mean ± S.E.M. and were analyzed by one-way ANOVA followed by Tukey’s multiple comparisons test (n = 10 animals/group). *p < 0.05 and **p < 0.01 compared to the saline group

Fig. 3
figure 3

Systemic administration of GUO-induced anxiolytic-like behavior in the light/dark task and in the round open field task. The number of transitions (a) and the time spent in the light compartment (b) were evaluated in the light/dark task 60 min after i.p. saline or 7.5 mg/kg GUO administration. The total distance travelled (c), the time spent in the center zone (d), and the distance travelled in center zone (e) were evaluated in the round open field task 60 min after i.p. saline or 7.5 mg/kg GUO administration. Data are reported as the mean ± S.E.M. and were analyzed by unpaired Student’s t test. (n = 12 animals/group). *p < 0.05 compared to the saline group

Fig. 4
figure 4

Systemic CAF pre-administration inhibited the anxiolytic-like effects of GUO. The percentage of time spent in the open arms (a), the total distance travelled in open arms (cm) (b), and the total distance travelled (cm) (c) were evaluated in the EPM task 60 min after i.p. saline or 7.5 mg/kg GUO administration, which was preceded by 15 min of i.p. saline or 10 mg/kg CAF pre-administration. Data are reported as the mean ± S.E.M., and differences among groups were determined by two-way ANOVA followed by Bonferroni’s post hoc test when applicable (n = 12 animals/group). *p < 0.05 compared to the saline/saline group and ## p < 0.01 compared to the saline/GUO group

The administration of 7.5 mg/kg GUO simultaneously exerted an anxiolytic effect (Fig. 2) in the EPM task and increased the CSF levels of ADO (Fig. 5a, **p = 0.0041) without affecting the levels of other purines (INO, GUO, HYPOX, XANT, and UA—Supplementary Material 5 A and B).

Fig. 5
figure 5

CAF prevented the increase of ADO levels and the decrease of GLU levels on CSF of rats following systemic GUO administration induced anxiolytic-like effects. ADO levels (a) were measured on CSF collected immediately after EPM task performance from rats that received i.p. saline or 7.5 mg/kg GUO administration. Data are reported as the mean ± S.E.M. and were analyzed by unpaired Student’s t test. ADO levels (b) and GLU levels (e) were measured on CSF collected immediately after EPM task performance from rats that received i.p. saline or 7.5 mg/kg GUO administration preceded by i.p. saline or 10 mg/kg CAF pre-administration. Data are reported as the mean ± S.E.M., and differences among groups were determined by two-way ANOVA followed by Bonferroni’s post hoc test when applicable. Linear correlations between the percentage of time spent in the open arms or the total distance travelled in open arms (cm) in the EPM task and ADO CSF levels, and the percentage of time spent in the open arms in the EPM task and GLU CSF levels in the animals that received saline or GUO are presented in (c, d, and f, respectively). Each point represents one animal. *p < 0.05, **p < 0.01, and *** p < 0.001, compared to the saline group (n = 10–12 animals/group)

The Anxiolytic-Like Effects Promoted by Systemic GUO Administration Were Completely Prevented by Pre-administration of CAF, a Nonspecific ADO Receptor Antagonist

CAF (10 mg/kg) administration per se did not modulate any anxiety-like behavioral parameter in the EPM task (Fig. 4 and Supplementary Material 1). On the other hand, GUO 7.5 mg/kg, again, increases the percentage of time spent in the open arms (Fig. 4a, ** p = 0.0010) and in open arms distance travelled (Fig. 4b, *p = 0.0425) in the EPM task. Additionally, the effect of GUO on the percentage of time spent in the open arms was completely abolished by i.p. pre-administration of 10 mg/kg CAF (Fig. 4a, ## p = 0.0078). The total distance travelled was not affected by any drug administration (Fig. 4c).

Fig. 6
figure 6

In vitro GUO incubation decreased the release of GLU from hippocampal synaptosomal preparations, without any change in GLU uptake. Ex vivo hippocampal GLU uptake was evaluated 60 min after saline or GUO administration (a) (n = 12 animals/group), and in vitro hippocampal synaptosomal GLU uptake was evaluated 5 min after the synaptosomal incubation with 100 μM GUO (b) (n = 5 animals/group). Both methods were described in the “Materials and Methods” section. Data were compared by unpaired Student’s t test. In vitro synaptosomal differences between the treatments on non-stimulated L-[3H]GLU release (basal) were measured in c), and in vitro K+-stimulated GLU release after 1 min of synaptosomal depolarization were evaluated in d, as described in the “Materials and Methods” section. The different treatments tested were GUO 100 μM, the ADO receptor agonists (CCPA 100 nM and CGS21680 30 nM), the ADO antagonists (CAF 1 μM, DPCPX 100 nM and ZM241385 50 nM) with or without GUO 100 μM. Data are reported as the mean ± S.E.M., and differences among groups were determined by two-way ANOVA followed by Bonferroni’s post hoc test. (n = 25 animals). *p<0.05, **p<0.01 and ***p<0.001 compared to the saline group

The Increase in CSF ADO Levels Promoted by GUO Administration Was Partially Prevented by CAF Pre-administration

CAF (10 mg/kg) did not present any effect per se on CSF ADO levels. However, GUO 7.5 mg/kg effect on increase CSF ADO levels (Fig. 5b, ***p = 0.0001) was partially prevented by pre-administration of CAF. A positive correlation between ADO levels and the percentage of time spent in the open arms in the EPM task can be observed regardless of CAF pre-treatment (Fig. 5c, r 2 = 0.3228, **p = 0.0016), as well as between ADO levels and the open arm distance travelled (Fig. 5d, r 2 = 0.2021, *p < 0.0377), without any correlation with total distance travelled in the EPM task (r 2 = 0.1903, p = 0.5016).

Interestingly, the animals that received CAF showed an increase in CSF levels of XAN and UA regardless of posterior administration of Sal or GUO (Supplementary Material 5 C and D). Accordingly, it is noteworthy that XAN and UA are products of CAF metabolism [34], which corroborates to the accuracy of our analysis.

Systemic GUO Administration Decreased CSF GLU Levels, an Effect that Was Partially Prevented by CAF Pre-administration

CSF GLU levels significantly decreased following systemic i.p. administration of 7.5 mg/kg Sal/GUO (Fig. 5e, ***p = 0.0354). CAF per se did not change significantly but partially blocked GUO effect on CSF GLU levels (Fig. 5e, p = 0.9891 and p = 0.4345, respectively). As opposed to CSF ADO levels, a negative correlation between CSF GLU levels and the percentage of time spent in the open arms in the EPM task (Fig. 5f, r 2 = 0.2977, *p = 0.0235) could be observed. No significant correlations were observed with the open arm distance travelled (r 2 = 0.1637, p = 0.2170) or with the total distance travelled (r 2 = 0.01949, p = 0.6652) in the EPM task.

Ex Vivo and In Vitro Approaches to Investigate Putative GUO Mechanisms of Action

Considering that extracellular GLU levels are modulated by cellular uptake/release, we performed ex vivo GLU uptake assays in hippocampal slices, and no significant effect of in vivo GUO administration was observed (Fig. 6a, p = 0.9933). Thus, we further investigated the putative mechanisms of GUO action in vitro, using hippocampal SP preparations from naïve rats (not subjected to any behavioral task). The SP preparation was greatly enriched in synaptic components (Supplementary Material 3) and was sensitive to high K+. GLU uptake by the hippocampal synaptosomal preparations (Fig. 6b, p = 0.7212) was not influenced by GUO. Interestingly, although basal GLU release was not affected by GUO (Fig. 6c, p = 0.999), K+-stimulated GLU release was significantly decreased by 100 μM GUO in vitro (Fig. 6d, ***p < 0.001).

To examine the putative involvement of ADO receptors in the GUO effect on K+-stimulated GLU release, ADO receptor modulators were used, to identify putative adenosinergic involvement in the GUO effects. Basal GLU release was not affected by GUO, the ADO A1 receptor agonist CCPA (100 nM), the ADO A2a receptor agonist CGS21680 (30 nM), the nonspecific ADO receptor antagonist CAF (1 μM), the ADO A1 receptor antagonist DPCPX (100 nM), the ADO A2a receptor antagonist ZM241385 (50 nM), or their combinations with GUO (Fig. 6c). However, although none of the ADO receptor antagonists affected K+-stimulated GLU release, the effect of GUO on reducing K+-stimulated GLU release was not prevented by the A2a antagonist ZM241385 (Fig. 6d, **p = 0.0076), while CAF partially prevented (Fig. 6d, p = 0.089) and DPCPX totally prevented (Fig. 6d, # p = 0.013) the GUO-mediated reduction in K+-stimulated GLU release. Accordingly, the ADO A1 receptor agonist CCPA also significantly decreased K+-stimulated GLU release (Fig. 6d, **p = 0.008) and the ADO A2a receptor agonist CGS21680 did not affect the GUO effect (Fig. 6d, p = 0.460). Finally, the combination of GUO plus CCPA or CGS21680 continued to hold the decrease in glutamate release (Fig. 6d, *p = 0.0101, **p = 0.0018, respectively).

Discussion

Acute/systemic GUO administration in rats induced anxiolytic-like effects in three different anxiety paradigms; these effects were correlated with the increase of ADO nucleoside levels and with the decrease of GLU levels in CSF. Additionally, these GUO effects were prevented in vivo by CAF pre-administration. Interestingly, in hippocampal synaptosomal preparations, GUO and the ADO A1 receptor agonist CCPA decreased K+-stimulated GLU release, while the ADO A1 receptor antagonist DPCPX and the nonspecific ADO receptor antagonist CAF reversed the GUO effect. Here, we provided experimental evidence in support to a new GUO mechanism of action involving the adenosinergic and glutamatergic systems, which seems to be closely related to the anxiolytic-like effects observed in vivo.

Anxiolytic Potential of GUO Administration

The EPM is currently the first-choice task for screening anxiolytic/anxiety-like behavior. Here, in the EPM task, acute administration of only the lower dose of GUO produced an anxiolytic-like behavior. The U-shape dose curve response for GUO was similar to that previously reported for the anxiolytic-like behavior promoted by GMP [6]. However, comparing the effective anxiolytic dosages, GMP showed anxiolytic-like effects in a higher dose than GUO (GMP at 50 mg/kg and GUO at 7.5 mg/kg), which suggests that GUO is more potent than GMP. In addition, the anxiolytic potential of GUO was further demonstrated in the light/dark and round open field tasks. Thus, pharmacologically, all three paradigms presented reasonable sensitivity to 7.5 mg/kg GUO.

To the best of our knowledge, this is the first study to evaluate the potential anxiolytic effects of acute/systemic GUO administration. Importantly, it was previously demonstrated that chronic oral administration of GUO exerted anxiolytic-like effects in mice [23]. Regarding depression, for which the main comorbidity is anxiety, the administration of GMP or GUO also produces antidepressive-like effects in predictive tasks [35, 36]. It has been shown that some of the behavioral effects of GMP, such as its anticonvulsant and antinociceptive effects, depend on the enzymatic conversion of GMP to GUO [37]. Although similarities exist, it remains to be investigated whether the mechanisms underlying the anxiolytic-like effects induced by GMP depend on its conversion to GUO. In fact, this topic is currently under investigation in our laboratory.

Interplay Between the Anxiolytic Effects of GUO and the ADO System

It has been previously demonstrated that CSF GUO levels increased 30 min after acute/systemic administration of 7.5 mg/kg GUO in rats [37]. By contrast, here, we did not observe any modulation in CSF GUO levels 60 min after systemic GUO administration. However, it has also been shown that GUO levels are increased in brain tissue homogenates 60 min after i.p. administration of GUO [38], suggesting that GUO uptake by neural cells could be at least partially involved in the decrease in GUO CSF levels, which could explain the unaltered CSF GUO levels observed in our study.

Additionally, we showed an increase in ADO CSF levels after GUO administration, and this increase was correlated with the anxiolytic parameters evaluated in the EPM task. The modulation of CSF ADO by GUO seems to occur later than 30 min after the GUO administration, as it cannot be observed earlier (30 min after systemic GUO 7.5 mg/kg administration) [37]. Intriguingly, our in vivo results are consistent with those of in vitro studies, in which GUO stimulates extracellular ADO enhancement in astrocytes and in other types of cells in culture [39, 40]. Furthermore, recently reported evidence indicates that the effect of GUO on regulating extracellular ADO levels may be a result of reducing its uptake by uncharacterized specific transporters [40, 41]. Together, all of these data suggest a close association between extracellular levels of GUO and ADO.

Interestingly, the increase in CSF ADO levels and the correlated anxiolytic-like effect promoted by GUO were sensitive to CAF. The literature supports the hypothesis that GUO presents effects that are either dependent on or independent of ADO receptors. In fact, using distinct CAF pre-administration doses, different results were obtained; in vivo, CAF pre-administration did not prevent the anticonvulsant (at a dose of 30 mg/kg) [42] and amnesic (at a dose of 5 mg/kg) effects of 7.5 mg/kg GUO [23] but prevented the antinociceptive effect (at a dose of 10 mg/kg) induced by 7.5 mg/kg GUO [16]. In vitro, some of the neural effects produced by exogenous GUO are prevented by ADO receptor antagonists, such as mitogenic activity, which could be partially inhibited by ADO A1 and A2b receptor antagonists [39], whereas other effects do not involve adenosinergic neurotransmission, such as the stimulatory effect on in vitro astrocytic GLU uptake activity. What seems clear is that GUO does not bind to ADO receptors, as shown in a preliminary study in which the binding of GUO to its putative G-protein receptor did not appear to be significantly sensitive to adenosinergic modulator compounds, including CAF [43].

Interactions Between the Anxiolytic Effects of GUO and Glutamatergic Neurotransmission

Acute systemic administration of GUO significantly decreased CSF GLU levels, an effect that was correlated with its anxiolytic-related behavior, as demonstrated by the increase in the percentage of time spent in the open arms of the EPM task. Previously, our group demonstrated that chronic systemic GUO administration prevented an increase in CSF GLU concentration in rats subjected to a chronic hepatic encephalopathy model [44]. Moreover, our group also showed that acute systemic administration of GUO prevented an increase in CSF levels of excitatory amino acids (including GLU) in the CSF of mice subjected to hyperalgesia induced by MK-801 [45]. Here, we provided the first report of a direct effect of systemic administration of GUO on CSF GLU levels in rats subjected to an anxiety-related paradigm.

No changes were observed in the GLU uptake activity induced by GUO in either the ex vivo (brain slices) or in vitro (synaptosomal preparations) protocols. These data indicated that there is at least one additional mechanism by which extracellular GLU levels are regulated by GUO in basal conditions. In this context, CAF pre-administration partially blocked the decrease in CSF GLU levels promoted by GUO, suggesting that the adenosinergic system is involved in the effects of GUO under physiological conditions.

In Vitro Effects of GUO on Hippocampal GLU Release

In vitro GUO significantly decreased K+-stimulated GLU release in the hippocampal synaptosomal preparation, without affecting basal GLU release. Our data is in accordance with GUO effect on preventing extracellular GLU accumulation induced by excitotoxic conditions in a more complex system, i.e., hippocampal slices [46]. In fact, decreasing hippocampal GLU neurotransmission is a feasible mechanism of inducing anxiolytic-like effects [3] because the hippocampus is part of the limbic system, which is linked to emotional behaviors, and GLU is one of the neurotransmitters that is intimately responsible for exciting the limbic pathway during anxiety-related behaviors [47].

Notably, the GUO effect on the K+-stimulated GLU release was similar to that of the ADO A1 receptor agonist CCPA. Moreover, this GUO effect was completely blocked by an ADO A1 receptor antagonist but not by an ADO A2a antagonist. These data suggest that by inhibiting pre-synaptic GLU release through ADO A1 receptor activation, GUO could decrease extracellular GLU levels. Interesting, ADO A2a receptors seem to play a role in GUO stimulatory effect on glutamate uptake in excitotoxic conditions [17], corroborating with GUO effect on the interplay between adenosinergic and glutamatergic system.

Recently, it was also showed that GUO recovered the impairment caused by oxygen glucose deprivation in hippocampal slices through stimulation of glutamate uptake by a pathway that, among others, involves antagonism of ADO A2a receptors [17]. Here we did not observe any counteractive effect of CGS21680, a specific ADO A2a receptor agonist, on GUO effect. Collectively, these data suggest that the modulatory effects of GUO on different parameters of the glutamatergic neurotransmission through the adenosinergic system are regulated by different mechanisms, especially in comparison between physiologic and excitotoxic conditions.

In agreement with this hypothesis, the following should be highlighted: (i) an increase in endogenous ADO is capable of increasing the activation of ADO A1 receptors [48]; (ii) ADO A1 receptors are the most abundant of the four known ADO receptors in the brain [49], including in the hippocampus; and (iii) in some regions of the brain, one of the physiological roles of ADO is to inhibit GLU release via pre-synaptic ADO A1 receptors, which consequently decreases neuronal excitability [50, 51]. Interestingly, there are other reports specifically related to the involvement of ADO A1 receptors that are also consistent with the GUO effects described here: (i) the blockade of ADO A1 receptors induces an anxiety phenotype in zebrafish [52], an effect opposite to the effect of GUO administration specified here; and (ii) the activation of ADO A1 receptors in rats leads to memory impairment [53], an effect similar to that promoted by GUO [18, 23].

Considering these findings, our results indicate that GUO-induced enhancement in extracellular (CSF) ADO concentration might be able to increase the activation of ADO A1 receptors, which could contribute to the inhibition of GLU excitability in the hippocampus, promoting attenuation of the anxiety-like phenotype.

Conclusions

In summary, the present work demonstrated the potential anxiolytic effects of acute/systemic GUO administration, which could be exerted by orchestration of the activities of the adenosinergic and glutamatergic systems. More specifically, a novel mechanism of action could be proposed for the anxiolytic effects of GUO, which include a modulatory effect on extracellular ADO and GLU levels. In addition, because high levels of extracellular GLU are involved in several neuropsychiatric conditions, these results reinforce the previously observed neuroprotective effects of GUO. Thus, we contributed to the goal of identifying the molecular targets and signaling pathways that are recruited by GUO and that promote anxiety-like behavior. Finally, GUO is an endogenous, orally active compound that is apparently well tolerated and that deserves attention as a potential novel therapeutic drug in anxiety disorders.