Guanosine Attenuates Behavioral Deficits After Traumatic Brain Injury by Modulation of Adenosinergic Receptors

Abstract
Traumatic brain injury (TBI) is a leading cause of disability worldwide, triggering chronic neurodegeneration underlying cognitive and mood disorder still without therapeutic prospects. Based on our previous observations that guanosine (GUO) attenuates short-term neurochemical alterations caused by TBI, this study investigated the effects of chronical GUO treatment in behavioral, molecular, and morphological disturbances 21 days after trauma. Rats subject to TBI displayed mood (anxiety-like) and memory dysfunction. This was accompanied by a decreased expression of both synaptic (synaptophysin) and plasticity proteins (BDNF and CREB), a loss of cresyl violet-stained neurons, and increased astrogliosis and microgliosis in the hippocampus. Notably, chronic GUO treatment (7.5 mg/kg i.p. daily starting 1 h after TBI) prevented all these TBI-induced long-term behavioral, neurochemical, and morphological modifications. This neuroprotective effect of GUO was abrogated in the presence of the adenosine A1 receptor antagonist DPCPX (1 mg/kg) but unaltered by the adenosine A2A receptor antagonist SCH58261 (0.05 mg/kg). These findings show that a chronic GUO treatment prevents the long-term mood and memory dysfunction triggered by TBI, which involves adenosinergic receptors.

Introduction
Traumatic brain injury (TBI) constitutes a major cause of mortality and morbidity; it involves a primary mechanical insult that initiates a cascade of secondary injuries eventually leading to long-term cellular and behavioral deficits [1, 2]. The secondary injury involves the abrupt depolarization of neurons and massive release of glutamate resulting inglutamate excitotoxicity, mitochondrial dysfunction, inflam- matory events, and neuronal cells loss [2, 3], which underlie long-term cognitive and psychiatric complications [1, 2]. Indeed, the vulnerability of the hippocampus, with deficits of synaptic plasticity [4, 5], provides an explanation for the high prevalence of cognitive deficits following TBI [6], which are aggravated over days to weeks by astrocytic scars and microglia-related neuroinflammation [2, 7–9].There are currently few therapeutic options to mitigate the long-term consequences of TBI since most drugs failed trans- lation to clinics, probably as a result of a single mechanism of action [10]. Targeting the purinergic modulation system may offer a unique benefit in view of its multi-target neuroprotec- tive effects in neurons and glia in different models of neuro- degenerative diseases [11]. Among the several purinergic sig- nals, guanosine (GUO) has been less seldom explored, al- though it affords long-term benefits to control brain neurode- generation [12, 13] as well as mood [14–16] and memory impairments [17–19]. Additionally, we have previously shown that GUO prevents several short-term TBI-induced modifications [3, 20], but it is currently unknown if a chronic GUO treatment might be effective to control TBI-induced long-term alterations.A major open question is the mechanism of GUO- mediated neuroprotection, since no membrane receptor for GUO has been identified [12, 13]. The similarity of GUO and adenosine has entrained the hypothesis that GUO could act as a modulator of adenosine receptors, namely A1 and A2A receptors [14, 21, 22].

These two main adenosine receptors in the brain are well established to be involved in the initiation and spreading of degeneration following different insults to the brain [23, 24].We now used a moderate fluid percussion injury (FPI) af- fecting the hippocampus to explore the impact of chronic GUO treatment on the long-term behavioral, neurochemical, and morphological alterations following TBI and further test- ed if the main adenosine receptors in the brain would be in- volved in the effects of GUO.Male adult Wistar rats (n = 166; 120 days old; 280–320 g) were obtained from our animal house colony. The animals were housed in groups of four animals per cage (polypropyl- ene, 41 × 34 × 16 cm L × W × H, 1394 cm2) with the floor covered with autoclaved shavings under controlled conditions (12:12-h light-dark cycle, lights turned on at 07:00 a.m.; 22 ± 2 °C; 45–65% relative humidity) on a ventilated rack with access to water and food (Puro Lab 22 PB) ad libitum. All animals were acclimatized to laboratory conditions for 1 week before the experiments. This study was approved by the Committee on the Ethics of Animal Experiments of the Federal University of Santa Maria, Brazil (Permit Number: 153/2014) and carried out in strict accordance with the recom- mendations in the Guide for the Care and Use of LaboratoryAnimals of the National Institutes of Health. The Ethics Committee imposed the selection of representative drug con- centrations, based on our previous experience in using all tested drugs, to reduce the number of animals used in the experiments, while making all efforts to minimize animal suffering.

The following drugs were used: guanosine (GUO) dissolved in 0.9% saline; 1,3-dipropyl-8-cyclopentylxanthine (DPCPX; adenosine A1 receptor antagonist) and 2-(2-Furanyl)-7-(2- phenylethyl)-7H-pyrazolo[4,3-e][1, 2, 4]triazolo[1,5-c-]pyrimidin-5-amine (SCH58261; adenosine A2A receptor an- tagonist) dissolved in saline with 0.1% DMSO (Sigma, St. Louis, MO, USA). These drugs were administered intraperi- toneally (i.p.) at a constant volume of 1 ml/kg. Trizol® re- agent, iScript™ cDNA synthesis Kit (Biorad®, Hercules, CA, USA), and SYBR® Select Master Mix were from Thermoscientific® (Waltham, MA, USA). All other chemicals of analytical grade were obtained from standard commercial suppliers.The animals were submitted to a surgical procedure to fix the TBI cannula in the right hemisphere and 24 h after the TBI was performed. To test the neuroprotective effect of GUO, rats were randomly divided into four groups: sham-vehicle, sham- GUO, TBI-vehicle, and TBI-GUO. Sham rats underwent all procedures identical to TBI rats, with the exception of the fluid percussion injury (FPI). The dose of GUO was chosen based on previous in vivo studies of excitotoxicity, which demon- strated an optimal effect of guanosine at 7.5 mg/kg i.p. [3, 20]. Rats received a single dose of GUO or vehicle per day and the first administration was performed 1 h after TBI.To assess the involvement of the adenosine modulation system in the neuroprotective effects of GUO, rats were divid- ed in four groups: sham-vehicle, TBI-DPCPX, sham-DPCPX- GUO and TBI-DPCPX-GUO to test A1 receptors and sham- vehicle, TBI-SCH 58261, sham-SCH 58261-GUO and TBI- SCH 58261-GUO to test A2A receptors. Rats received vehicle or doses previously validated as effective [25, 26] of DPCPX (1 mg/kg) or SCH 58261 (0.05 mg/kg) every day (i.p.) 30 min before GUO.

The analysis of all behavioral data showed that the SCH58261 and DPCPX administration did not demon- strate per se effect (not shown; parallel experiments). In order to use a minimum number of animals, the same rats of sham- vehicle group were used for both different adenosine antago- nist’s treatments at the same time. Drug treatments lasted 21 days after TBI (Fig. 1).After behavioral analysis, a part of the animals under deep anesthesia (thiopental sodium, 100 mg/kg i.p.) was perfusedthe administration of DPCPX (1 mg/kg) and SCH 58261 (0.05 mg/kg) 30 min before GUO. The short-term memory (STM) and long-term mem- ory (LTM) was performed by object recognitionfor the histochemistry and immunohistochemistry assays (n = 7 per group). For qPCR analysis (n = 5 per group), the animals were euthanized by decapitation to quickly dissect the ipsilat- eral hippocampus on an ice-cold Petri dish.The model of FPI was carried out as previously described [27] with minor modifications. In brief, rats were anesthetized with a single i.p. injection of Equithesin (4 ml/kg), a mixture con- taining sodium pentobarbital (58 mg/kg), chloral hydrate (60 mg/kg), magnesium sulfate (127.2 mg/kg), propylene gly- col (42.8%), and absolute ethanol (11.6%) and placed in a rodent stereotaxic apparatus. A 3-mm-diameter craniotomy was drilled on the right convexity, 2 mm posterior to the breg- ma and 3 mm lateral to the midline, ensuring that the dura mater remained intact. A plastic injury cannula was fixed with cyanoacrylate, stabilized with dental cement. When dental cement has hardened, the cannula was filled with chloram- phenicol, closed with a proper plastic cap, and the animal removed from the stereotaxic device and returned to its home cage. After 24 h, animals were anesthetized (isoflurane, 1% inhaled) and the injury cannula was attached to the fluid per- cussion device developed in our laboratory. During surgery and FPI, body temperature (37 °C) was monitored rectally and maintained with a heating pad and an overhead incandes- cent bulb. A brief (10–15 ms) transient pressure fluid pulse (1.55 ± 0.09 atm) was applied against the exposed dura in the right hemisphere.

Pressure pulses were measured extracranially by a transducer (fluid control hydraulic automa- tion, Belo Horizonte, MG, Brazil) and recorded on a storage oscilloscope (Tektronix TDS 210, Beaverton, OR, USA). The fluid percussion caused apnea (30–70 s), unconsciousness (7– 10 min) measured through the righting reflex restoration [28]. Overall mortality after TBI was 20.48% (34 out of 167 ani- mals). Twenty-five of these animals died before restoring consciousness, and ten animals (six TBI-saline and three TBI- guanosine animals) died in their home cages before the endof the protocol [7, 28]. Sham animals underwent an identical procedure, with the exception of the FPI.Locomotor Behavior Distance, speed, and rearing were mea- sured 4 min a day for 4 days, to avoid the possible interfer- ences on locomotion caused by the tested drugs [29, 30]. These parameters were analyzed in an activity-monitoring chamber with 50 × 48 × 50 cm (Insight Ltda, Ribeirão Preto, SP, Brazil).Elevated Plus-Maze To evaluate anxiety, rats were exposed to an elevated plus maze, as previously described [31] to measure (a) the time spent in the open arms relative to the total time spent in the plus-maze, expressed as percentage;(b) the number of entries into the open arms; and (c) the number of entries into the closed arms. These parameters were defined as placing all four paws within the boundaries of the arm. The sessions lasted for 5 min, and after each trial, the maze was cleaned with an ethanol solution (20%).Inhibitory Avoidance Task The step-down inhibitory avoid- ance task (IA) has been used to evaluate aversive memory, as we have previously described [32]. Rats were trained using a 50 × 25 × 25-cm plexiglass box with a 5-cm-high, 8-cm- wide, and 25-cm-long platform on the left end of a series of bronze bars which made up the floor of the box.

For the IA training session (acquisition), rats were gently placed on the platform facing the left rear corner of the training box. When they stepped down and placed their four paws on the grid, a foot shock was delivered (2 s, 0.5 mA). The retention trial (test) was performed 24 h after training. Each rat was placed again on the platform, and the transfer latency time (i.e., time took to step down from the platform) was measured in the same way as in the acquisition trial, but foot shock was not delivered and the transfer latency time was recorded to a ceil- ing of 600 s. The criterion for learning was taken as an in- crease in the transfer latency time on retention (second) trial ascompared to the acquisition (first) trial. So, short transfer la- tencies indicate poor retention. To avoid confounds by linger- ing olfactory stimuli, the arena was cleaned with 20% ethanol after each animal was tested.Object Recognition Memory Task Recognition memory was measured using the object recognition task [33]. Rats first underwent a training session in which they were exposed to two identical objects (duple Lego toys) for 5 min. After 90 min (short-term memory; STM) or 24 h (long-term mem- ory; LTM), two dissimilar objects were presented (a familiar and a novel one), and we measured the time spent exploring each object during 5 min, as previously described. A recogni- tion index calculated for each animal in STM task was expressed by the ratio TB/(TA + TB) [TA = time spent explor- ing the familiar object A; TB = time spent exploring the novel object B] and in LTM task was expressed by the ratio TC/ (TA + TC) [TA = time spent exploring the familiar object A; TC = time spent exploring the novel object C].

The experi- ments were performed by a blind observer for the treatment of animals. To avoid confounds by lingering olfactory stimuli and preferences, the objects and the arena were cleaned with 20% ethanol after each animal was tested. Exploration was defined as sniffing or touching the objects with the nose and/ or forepaws. Sitting on or turning around the objects was not considered exploratory behaviors.Total RNA was isolated from hippocampus using Trizol® reagent (Invitrogen®, Carlsbad, CA, USA) immediately after euthanasia. Conversion of total RNA to cDNA was performed with the iScript™ cDNA synthesis Kit (Biorad®), accordingly to the manufacturer’s instructions. Quantitative real-time PCR was performed in 20 μl PCR mixture containing 1 μl of cDNA products as template and a SYBR® Select Master Mix (Applied Biosystems, Forster City, CA, USA). Gene-specific primer sequences were based on published sequences in GenBank Overview (http://www.ncbi.nlm.nih.gov/genbank/) designed with Primer3 program version 0.4.0 (http://frodo.wi.mit.edu/ primer3/) and custom made by Invitrogen® (Table 1). The mixtures were heated at 95 °C for 5 min followed of 40 cycles of 15 s at 95 °C, 15 s at annealing temperature appropriated to each primer sequence, and 25 s at 72 °C for extension in a Thermocycler StepOne Plus (Applied Biosystems, Foster City, CA, USA). All samples were an- alyzed as quadruplicates with a non-template control also included. Samples were quantified using the ΔΔCq meth- od [34], with tubulin and GAPDH serving as reference genes, according the MIQE guidelines [35].For the brain fixation, animals under deep anesthesia (sodium thiopental, 100 mg/kg i.p.) were transcardially perfused. The heart was exposed and, after clamping the descending aorta, a catheter was inserted in the ascending aorta. The animal was then perfused with 600 mL of heparinized saline while open- ing the right atria to allow the outflow of the perfusate. Rats were then perfused with 600 mL of 4% paraformaldehyde in0.1 M phosphate-buffered saline (PBS) at pH 7.4. The brains were removed and immersed in the same fixative overnight and then kept in 30% sucrose in physiological saline (0.9% NaCl) for at least 48 h before sectioning, as previously de- scribed [36].

The dorsal hippocampus of frozen brains was sectioned (30 μm coronal slices) with a Leica CM1850 cryo- stat (Leica Microsystems, Wetzlar, Germany) and hippocam- pal sections were mounted on slides coated with 2% gelatin with 0.08% chromalin and stored at − 20 °C until use. Neuronal morphology in dorsal hippocampal sections was evaluated by cresyl violet staining of Nissl bodies, as previ- ously described [31, 36]. Briefly, sections were incubated for 10 min with cresyl violet (Sigma-Aldrich) solution (0.5% in acetate buffer). Sections were then washed twice with acetate buffer, twice in 100% ethanol, cleared with xylene, and mounted with Vectashield mounting medium (Vector Laboratories, Burlingame, CA, USA). The number of cresyl violet positive neurons in the CA1 region was determined using a light microscope (Leica DM 1000) to collect comput- erized image analyzed with NIH ImageJ software to count the number of cells with round, obvious nuclei and visible nucle- oli in eight sections per rat.Immunohistochemical detection of Iba-1 (a marker of re- active microglia) and of glial fibrillary acidic protein (GFAP) (a marker of astrocytes) was performed to gauge microgliosis and astrogliosis, as previously described [37]. After washing, the sections were permeabilized and blocked before incuba- tion with the anti-Iba-1 rabbit polyclonal antibody (1:1000; 019-19741; RRID: AB_839504; Wako Chemicals, Richmond, VA, USA) and anti-GFAP mouse monoclonal an- tibody (1:1000; ab4648; RRID: AB_449329; Abcam, Inc., Cambridge, MA, USA) for 48 h at 4 °C. Slices were subse- quently incubated with goat anti-mouse or goat anti-rat sec- ondary antibody conjugated with a fluorophore (Alexa Fluor 488; Alexa Fluor 525; Invitrogen) (1:500) for 2 h at room temperature. The mean fluorescence intensities of GFAP and Iba-1 were then semi-quantitatively determined using NIH ImageJ software in images obtained with a Leica (TCS SP5 II) microscope.The normality of the data was analyzed using the D’Agostino and Pearson’s omnibus normality test. Data are mean ± SEMexcept for the inhibitory avoidance memory where data are expressed as median ± interquartile range. Data were analyzed by two-way analysis of variance (ANOVA) followed by post hoc comparisons using Newman-Keuls multiple test, except inhibitory avoidance memory test, which were analyzed by unpaired t test and a Scheirer-Ray-Hare test (an extension of the Kruskal–Wallis test) followed by Mann-Whitney post hoc test. Differences between groups were considered statistically significant when p < 0.05.

Results
Locomotor activity was evaluated during 4 days (6–9th day of protocol; Fig. 1) to discard its interference in the behavioral of animals. We did not observe alterations in the pattern of loco- motion, in particular distance (Fig. 2c) and rearing (Fig. 2d), between sham and TBI animals at any day of analysis.Prevention by Guanosine of TBI-Induced Anxiety Involving Adenosinergic ReceptorsThe percentage of time spent in the open arms (F(7,88) = 10.54; p < 0.05; Fig. 3a) and the number of entries in the closed arms (F(7,88) = 6.45; p < 0.05; Fig. 3c) of the plus maze were signif- icantly altered in TBI as compared to sham animals. The treat- ment with GUO (7.5 mg/kg) inhibited the anxiogenic effect of TBI, as testified by the higher percentage of time spent in the open arms (F(7,88) = 8.02; p < 0.05, Fig. 3a) and lower number of entries in the closed arms (F(7,88) = 8.24; p < 0.05, Fig. 3c). Notably, the administration of the A1 receptor antagonist DPCPX (1 mg/kg) inhibited this anxiolytic-like effect of GUO (F(7,88) = 8.80; p < 0.05, Fig. 3a and F(7,88) = 7.12; p <0.05, Fig. 3c, respectively) and was devoid of effects in the behavior of animals subjected to TBI. Interestingly, it was demonstrated that the administration of A2A receptor antago- nist SCH58261 (0.05 mg/kg) also attenuated the anxiogenic effect of TBI, by the higher percentage of time spent in the open arms (F(7,88) = 4.29; p < 0.05, Fig. 3a) and lower numberof entries in the closed arms (F(7,88) = 7.23; p < 0.05, Fig. 3c). The administration of the A2A receptor antagonist SCH58261 did not affect the anxiolytic-like effect of GUO. Indeed, it showed a similar effect displayed by GUO treatment, with higher percentage of time spent in the open arms (Fig. 3a) and lower number of entries in the closed arms (Fig. 3c). This indicates an ability of GUO to attenuate the impact of TBI on anxiety, which seems to involve modulation of aden- osine receptors. No differences were observed in the number of entries in open arms (Fig. 3b).

The training in the step-down inhibitory avoidance task did not show differences between the groups (Fig. 4a). Analysis of aversive memory revealed a significant decrease on learning of rats submitted to TBI (F(7,88) = 6.45; p < 0.05; Fig. 4b) and GUO treatment avoided this impairment (F(7,88) = 14.41; p < 0.05; Fig. 4b) on retention time. Notably, the administration of DPCPX attenuated this anti-amnesic-like effect of GUO (F(7,88) = 4.96; p < 0.05, Fig. 4b). Interestingly, SCH58261 (0.05 mg/kg) also attenuated the amnesic effect of TBI (F(7,88) = 4,15; p < 0.05, Fig. 4a) on retention time. Theadministration of this antagonist did not affect anti-amnesic- like effect of GUO (Fig. 4b). This indicates an ability of GUO to attenuate the deterioration of aversive memory after TBI.Different parameters of memory can be observed as the functional working memory, in which the animals have a tendency to avoid areas that they have just been and ex- plore novel areas or objects. We first established that none of the treatments affected the exploration of objects dur- ing the training section (Fig. 4c). However, TBI signifi- cantly impaired both the short-term (F(7,88) = 7.66; p < 0.05; Fig. 4d) and long-term object recognition memory tasks (F(7,88) = 7.34; p < 0.05; Fig. 4e), since TBI animals displayed a lower exploration of a novel object in com- parison to sham animals. GUO treatment attenuated this TBI-induced impairment of both short- and longer-term memory tasks (F(7,88) = 4.72; p < 0.05; Fig. 4d and F(7,88) = 4.57; 4E, respectively).

The administration of DPCPX blocked the protective effect of GUO in short- (F(7,88) = 9.86; p < 0.05; Fig. 4d) and longer-term recogni- tion memory (F(7,88) = 6.98; p < 0.05; Fig. 4e) in TBI- induced animals. Newsworthy, SCH58261 (0.05 mg/kg) also reduced the TBI-induced impairment of both short- and longer-term mem- ory tasks (F(7,88) = 4.06; p < 0.05; Fig. 4d and F (7,88) = 4.19;SCH 58261 (0.05 mg/kg) 30 min before GUO. Data were expressed as mean ± S.E.M. (n = 12) and analyzed by two-way ANOVA, followed by Newman–Keuls test when appropriated. Differences were considered significant (*p < 0.05) when compared to sham and TBI groups(d) and long-term memory (e) were performed at 20° and 21° days after TBI. Data were expressed as mean ± S.E.M. (n = 12) and analyzed by two-way ANOVA, followed by Newman–Keuls test when appropriated. In both memory tests, the modulation of adenosinergic system by GUO was evaluated through the administration of DPCPX (1 mg/kg) and SCH 58261 (0.05 mg/kg) 30 min before GUO. Differences were considered significant (*p < 0.05) when compared to sham and TBI groupsPrevention by Guanosine of TBI-Induced Decrease of Plasticity Genes (BDNF and CREB) and Synaptic Markers (Synaptophysin) Involving Adenosinergic ReceptorsBDNF (brain-derived neurotrophic factor) and CREB (cyclic AMP response element-binding protein) are major players in the regulation of functional and structural neuroplasticity un- derlying hippocampal-dependent memory [38]. In accordance with the deterioration of plasticity, TBI decreased the expres- sion in the hippocampus of both BDNF (F(1,32) = 19.23; p < 0.05; Fig. 5a) and CREB (F(1,32) = 6.22; p < 0.05; Fig. 5b).

Treatment with GUO fully prevented this TBI-induced alter- ation of BDNF (F(1,32) = 8.47; p < 0.05; Fig. 5a) and CREB expression (F(1,32) = 8.43; p < 0.05; Fig. 5b). As was observedfor the alterations of memory, DPCPX administration prevented the effects of GUO on TBI-induced decrease of BDNF and CREB gene expression (F(1,32) = 12.42; p < 0.05, Fig. 5a and F(1,32) = 4.50 p < 0.05, Fig. 5b, respectively), confirming the involvement of A1 receptors in the neuropro- tective effects of GUO. The SCH58261 administration also prevented this TBI-induced alteration of BDNF (F(1,32) = 10.1; p < 0.05; Fig. 5a) and CREB expression (F(1,32) = 5.77; p < 0.05; Fig. 5b) and did not alter effects of GUO treatment BDNF (Fig. 5a) and CREB (Fig. 5b) gene expression.Memory impairment is also associated with a synaptotoxicity typified by altered synaptic organization and expression of synaptic proteins [39]. Thus, we investigated the impact of TBI on the expression of two synaptic proteins—a constitutive protein, synaptophysin, and a synaptic protein associated with synaptic regeneration, growth associated pro- tein 43 (GAP-43). Compared to sham rats, TBI led to a de- creased expression of synaptophysin in the hippocampus (F(1,32) = 6.92 p < 0.05, Fig. 5c), whereas no difference in the expression of GAP-43 was found 21 days after traumatic in- jury (Fig. 5d). Treatment with GUO prevented the TBI- induced decrease of synaptophysin gene expression (F(1,32) = 6.29 p < 0.05, Fig. 5c) and this synapto-protectiveevaluated through the administration of DPCPX (1 mg/kg) and SCH 58261 (0.05 mg/kg) 30 min before GUO. Data are expressed as mean ±S.E.M. (n = 5) and were analyzed by two-way ANOVA, followed by Newman–Keuls test when appropriated. Differences were considered significant (*p < 0.05) when compared to sham and TBI groups30 min before GUO.

Data are expressed as mean ± S.E.M. (n = 7) and were analyzed by two-way ANOVA, followed by Newman–Keuls test when appropriated. Differences were considered significant (*p < 0.05) when compared to sham and TBI groups. Scales 50× and 200×effect of GUO was blocked in presence of DPCPX (F(1,32) =7.53; p < 0.05, Fig. 5c) but unaltered in the presence of SCH 58261 (Fig. 5c).We assessed neuronal loss by histological analysis of coronal sections collected at 21 days post-injury and stained with cresyl violet. We observed a significant neuronal loss in the CA1 hippocampal regions 21 days after TBI (F(7,48) = 5.81 p< 0.05, Fig. 6) as compared to sham. Treatment with GUO reduced this TBI-induced hippocampal neurodegeneration (F(7,48) = 5.17 p < 0.05, Fig. 6) and this effect of GUO was prevented in the presence of DPCPX (F(7,48) = 7.03 p < 0.05, Fig. 6). However, SCH 58261 administration also diminished this TBI-induced hippocampal neurodegeneration (F(7,48) =4.27 p < 0.05, Fig. 6) and did not alter the effects of GUO treatment (Fig. 6).The analysis of GFAP immunoreactivity at 21 days post-TBI injury revealed a significant larger astrocytosis in the hippo- campal CA1 region after TBI compared to sham (F(7,48) =11.63 p < 0.05, Fig. 7). Treatment with GUO prevented astrogliosis (F(7,48) = 7.77 p < 0.05, Fig. 7) and the administra- tion of DPCPX blocked this glioprotective effect (F(7,48) =8.38 p < 0.05, Fig. 7).We also assessed microgliosis at 21 days post-TBI inju- ry. IBA-1 immunoreactivity in the hippocampal CA1 re- gion was significantly larger 21 days after TBI compared to sham (F(7,48) = 13.18 p < 0.05, Fig. 8). Treatment with GUO prevented microgliosis (F(7,48) = 7.75 p < 0.05, Fig.8) and DPCPX administration blocked this glioprotective effect (F(7,48) = 10.16 p < 0.05, Fig. 8). In accordance with the results above, SCH 58261 administration prevented both astrogliosis (F(7,48) = 5.11; p < 0.05; Fig. 7) andmicroglial activation (F(7,48) = 4.95; p < 0.05; Fig. 8)(0.05 mg/kg) 30 min before GUO. The mean fluorescence intensity (F.I.) of GFAP was semi-quantitatively determined using NIH ImageJ soft- ware. Data are expressed as mean ± S.E.M. (n = 7) and were analyzed by two-way ANOVA, followed by Newman–Keuls test when appropri- ated. Differences were considered significant (*p < 0.05) when compared to sham and TBI groupsand SCH 58261 (0.05 mg/kg) 30 min before GUO. The mean fluores- cence intensities (F.I.) of IBA-1 were semi-quantitatively determined using NIH ImageJ software. Data are expressed as mean ± S.E.M. (n = 7) and were analyzed by two-way ANOVA, followed by Newman–Keuls test when appropriated. Differences were considered significant (*p < 0.05) when compared to sham and TBI groups21 days post-TBI injury as well as did not modify the effects of GUO treatment (Figs. 7 and 8, respectively).

Discussion
The present study shows that a daily treatment with guanosine (GUO) beginning after moderate percussion brain injury at- tenuates the long-term development of anxiety and memory deficits. This probably involves an ability of GUO to control aberrant plasticity at excitatory synapses since GUO prevented the TBI-induced long-term reduction of the expres- sion of a synaptic marker (synaptophysin) and of genes asso- ciated with synaptic plasticity, namely BDNF and CREB. Additionally, GUO also attenuated TBI-induced long-term astrogliosis and microgliosis that compromise synaptic func- tion and negatively impact on mood and memory [8, 40, 41]. Finally, we concluded that this neuroprotective effect of GUO required functional adenosine receptors. The traumatic event triggering TBI initiates a cascade of excitotoxic events leading to long-term cellular and behavioral deficits, among which stem mood and memory deficits [1, 2]. There are currently no effective strategies to mitigate these long-term sequelae of TBI, probably because most candidate strategies mostly target one of the multiple excitotoxic cas- cades engaged in this pathology [10]. Our previous observa- tions that GUO attenuated at least two of the excitotoxic pro- cesses engaged in TBI, namely the glutamatergic system and mitochondrial function [3, 20], prompted testing if GUO would dampen chronic neurological abnormalities associated with TBI, such as cognitive deficits and emotional distur- bances [30, 42]. The present results confirmed that our model of fluid percussion brain injury caused a long-term develop- ment of anxiety as well as of memory impairment, as reported for different animal models of TBI [8, 43, 44]. Most impor- tantly, we found that the GUO treatment after FPI was effec- tive to reduce these delayed anxiety-like and reference mem- ory deficits. This is in agreement with the previously reported ability of GUO to interfere with memory deficits [16–19] as Guanosine (GUO) was able to block the injuries caused by the trauma through the modulation of purinergic system. The signal (?) indicates a possible mechanism of action involved. (A1, P2 receptors; A1/A2A, A1/ P2Y1 receptors oligomerization) well as with the anxiolytic-like effects of GUO and its deriv- atives in rodents [14, 15], which also produces anti- depressive-like effects in predictive tasks [13–15], the main comorbidity of anxiety.

The mechanism underlying this ability of GUO to attenuate the long-term mood and memory dysfunction after TBI still needs to be defined. The present study unravels the possibility that GUO might control the synaptic dysfunction that has been documented to underlie both mood dysfunction as well as memory impairments [39, 45]. Indeed, we now show that GUO prevented the TBI-induced decrease of the expression in the hippocampus of two genes that are paramount to sustain plastic changes, namely BDNF and CREB. Accordingly, pre- vious studies documented the ability of GUO to control BDNF levels in a different functional context [46]. Furthermore, GUO also prevented the TBI-induced decrease of the expression of a well-established synaptic marker, synaptophysin [47], in agreement with the previously reported ability of GUO to control neurite outgrowth in PC12 cell [48] and to induce synaptogenesis in the adult brain [49]. Notably, GUO failed to affect the expression of GAP-43, which is more associated with the repair of synapses [50]. This suggests that GUO might attenuate hippocampal synaptotoxicity, which un- derlies memory impairment in different neurodegenerative disorders [39, 45]. Synaptic function is also critically dependent on astrocytic support [41, 51] and on microglia reactivity [40, 52] and, accord- ingly, both astrogliosis and microgliosis have been proposed as alternative or concurrent mechanisms to explain the deterioration of memory performance as well as of mood alterations upon TBI [8]. This contention is in agreement with the presently observed long-term modification of the morphology of astrocytes and of microglia in the hippocampus. Furthermore, we observed that GUO attenuated this TBI-induced gliosis, in accordance with previous demonstrations of protective effects of GUO on astrocytes [53, 54] and on microglia reactivity [55].

A major conclusion of the present study is the requirement of adenosine A1 receptors, but it cannot be discarded the in- volvement of adenosine A2A receptors for the long-term neuro- protection afforded by GUO. The SCH 58621 has showed strong neuroprotective effects against TBI and demonstrated the importance of the adenosinergic system in the progression of this pathology. Previous studies had already documented an ability of caffeine and adenosine receptor antagonists to be associated with neuroprotective effects of GUO, in conditions which the effects of GUO appear to be attenuated by adenosine receptor antagonists [13, 14, 21, 56]. In some instance, a pref- erential involvement of A2A receptors was reported [21], whereas in others, A1 receptors were selectively associated with the neuroprotective effects of GUO [14, 56]. Likewise, the mitogenic effects of GUO [57] or the impact on GUO on the synaptosomal K+-stimulated glutamate release [14] was partial- ly dampened upon blockade of A1 receptors and unaffected by manipulation of A2A receptors. This dependency on A1 rather than A2A receptors is somewhat surprising since A2A receptor blockade affords a robust neuroprotection against different nox- ious brain insults [23], whereas the blockade of A1 receptors has paradoxical effects on neuroprotection [24] and a continu- ous activation (or over-expression) of A1 receptors has delete- rious effects on mood and synaptic plasticity [43, 58]. In spite of the presently concluded critical requirement for A1 receptors for the neuroprotective effects of GUO, the exact relation between GUO and A1 receptor function still remains to be elucidated. In fact, it cannot be excluded that GUO might directly interact with A1 receptors, eventually as an allosteric modulator [59] since GUO binds with low affinity to adenosine receptors [60] and GUO binding is unaffected by caffeine [61]; another possibility might be an interaction with A1 receptor-containing heteromers, such as A1-A2A heteromers [21, 62] or A1-P2Y1 heteromers [63, 64]; alternatively, it is also possible that GUO might bolster the extracellular levels of adenosine through an increased outflow or inhibition of reuptake through nucleoside transporters (Fig. 9) [57, 65].

The literature data describes that the glutamatergic system regulation by GUO can involve the modulation of A1 or even a putative oligomeric interaction resulting in A1R activation and A2AR blockage [21, 66] and the consequent activation of the PI3K/Akt and MAPK signaling cascades [21]. In agree- ment, these signaling pathways have been associated with the regulation of astrocytic glutamate transporters [13, 67, 68].
Neuroprotection event caused by GUO is believed to result from the inhibition of the secondary TBI cascade mechanisms (glutamatergic excitotoxicity, mitochondrial dysfunction, calcium homeostasis disruption, oxidative stress, inflammatory process and neuron death) as previ- ously demonstrated by our group [3, 20]. These events im\plicate in the decrease of cellular damage (neuron death, astrogliosys and microglia activation), which is cor- related to synaptic plasticity (expression of BDNF and CREB) and learning/memory (Fig. 9). In summary, we report an ability of GUO to prevent the delayed anxiety and memory deficits following TBI, which might result from the pleiotropic impact of GUO on the ex- pression of genes governing synaptic plasticity, on processes of synaptotoxicity, and on the control of astrogliosis SCH58261 and microgliosis in a manner involving adenosine A1 and A2A receptors (Fig. 9). Altogether, these findings posit GUO as a candidate post-traumatic therapy to alleviate the long-term consequences of TBI.