Neurochemical Changes in the Mouse Hippocampus Underlying the Antidepressant Effect of Genetic Deletion of P2X7 Receptors

Recent investigations have revealed that the genetic deletion of P2X7 receptors (P2rx7) results in an antidepressant phenotype in mice. However, the link between the deficiency of P2rx7 and changes in behavior has not yet been explored. In the present study, we studied the effect of genetic deletion of P2rx7 on neurochemical changes in the hippocampus that might underlie the antidepressant phenotype. P2X7 receptor deficient mice (P2rx7−/−) displayed decreased immobility in the tail suspension test (TST) and an attenuated anhedonia response in the sucrose preference test (SPT) following bacterial endotoxin (LPS) challenge. The attenuated anhedonia was reproduced through systemic treatments with P2rx7 antagonists. The activation of P2rx7 resulted in the concentration-dependent release of [3H]glutamate in P2rx7+/+ but not P2rx7−/− mice, and the NR2B subunit mRNA and protein was upregulated in the hippocampus of P2rx7−/− mice. The brain-derived neurotrophic factor (BDNF) expression was higher in saline but not LPS-treated P2rx7−/− mice; the P2rx7 antagonist Brilliant blue G elevated and the P2rx7 agonist benzoylbenzoyl ATP (BzATP) reduced BDNF level. This effect was dependent on the activation of NMDA and non-NMDA receptors but not on Group I metabotropic glutamate receptors (mGluR1,5). An increased 5-bromo-2-deoxyuridine (BrdU) incorporation was also observed in the dentate gyrus derived from P2rx7−/− mice. Basal level of 5-HT was increased, whereas the 5HIAA/5-HT ratio was lower in the hippocampus of P2rx7−/− mice, which accompanied the increased uptake of [3H]5-HT and an elevated number of [3H]citalopram binding sites. The LPS-induced elevation of 5-HT level was absent in P2rx7−/− mice. In conclusion there are several potential mechanisms for the antidepressant phenotype of P2rx7−/− mice, such as the absence of P2rx7-mediated glutamate release, elevated basal BDNF production, enhanced neurogenesis and increased 5-HT bioavailability in the hippocampus.


Introduction
Mood disorders, including major depressive disorder (MDD) and bipolar disorder (BPD), are the most common psychiatric illnesses [1,2], which are caused by complex interactions between genes, developmental and environmental factors. Genetic research has identified several chromosomal regions and genes involved in the susceptibility to mood disorders, but no clear etiological mechanism has been determined. Previous studies revealed mutations in the gene encoding the P2X7 purinergic receptor (P2rx7), which are associated with the development of MDD and BPD [3,4,5,6,7,8]. Although subsequent studies have failed to confirm this association [9,10,11], more recent data have again reaffirmed the potential of these polymorphisms to affect the pathology of mood disorders [12,13]. In addition, a clinical study showed the reduced mRNA expression of the P2X7 receptor in a set of patients suffering from depression and posttraumatic stress disorder characterized by the increased risk of suicide [14]. P2rx7 belongs to the family of ionotropic P2X receptors that are sensitive to ATP and other related nucleotides. These receptors are distributed in hematopoietic cells, epithelial cells, neurons, astrocytes, oligodendrocytes and microglia. P2rx7 plays an important role in the processing and secretion of mature proinflammatory cytokines, such as interleukin (IL)-1b, IL-18, and tumor necrosis factor (TNF)-a, and in ATP-mediated apoptosis. The primary role of P2rx7s in the brain is regulation of neurotransmitter release [15,16,17]. The activation of P2rx7 results in Ca 2+ influx [18], increased glutamate and GABA release from brain slices [19,20] and nerve terminals [21,22].
Previous studies have demonstrated that the genetic deletion and pharmacological antagonism of P2rx7 leads to an antidepressant phenotype in several behavioral models [23,24,25]. We found that P2rx72/2 mice displayed a decreased immobility response in the forced swim (FST) and tail suspension (TST) tests, which can be reproduced by subacute administration of the selective P2rx7 antagonist Brilliant Blue G (BBG). We also presented evidence that the deletion of P2rx7 in non-hematopoietic cells leads to the observed antidepressant phenotype. When bone marrow chimeras were generated that lacked the P2rx7 only in their hematopoietic compartment, no difference was found in behavioral tests, indicating that the antidepressant phenotype found in P2rx72/2 mice was not transferred to wild-type recipients with the engraftment of the P2rx72/2 bone marrow cells [25]. Consequently, the expression of P2rx7 in other cell types, particularly neurons or astrocytes, might be responsible for the associated changes in mood. In addition, we found that the deficiency of P2X7 receptors leads to a widespread alteration of the gene expression in the limbic system, including the up and downregulation of genes crucial for synaptic transmission and plasticity, such as glutamatergic and GABAergic receptor subunits [25]. Consistent with these results, other studies have shown enhanced c-Fos expression in the hippocampus of P2rx72/2 mice after repeated forced swim tests, which indicates that the hippocampus is an important target area that mediates the effect of P2rx7 activation on emotional behavior [24].
Despite these data, the key question how the activity of P2X7 receptor leads to alterations in animal behavior remains unknown. It has been repeatedly shown that major depression is characterized by a reduction of neuronal plasticity and a 2-week administration of antidepressant treatment is sufficient to normalize this deficit and generate a neuroadaptive mechanism that might enhance neuronal plasticity [26,27]. Neurotrophins, such as brain-derived neurotrophic factor (BDNF), promote neuronal survival and enhance synaptic plasticity [28]. BDNF is synthesized as a 30-35-kDa-precursor protein, which is proteolytically cleaved to produce a mature and functional enzyme that plays a crucial role in the development, differentiation and survival of neuronal populations within the central and peripheral nervous system. Several studies have demonstrated a strong association between the expression of BDNF and the onset of depression. Reduced BDNF mRNA and protein expression have been detected in the hippocampus of postmortem brains from suicide victims [29] and serum BDNF levels are reduced in depressed patients [30]. Consistent with these findings, it has been shown that the infusion of BDNF into the brain results in an antidepressant-like phenotype [31]. However, the lack of BDNF is not sufficient to produce a depressive phenotype, although it is required for the behavioral response to antidepressants [32].
As a consequence of an alteration in the level of neurotrophic factors, or independently from it, an aberrant regulation of adult neurogenesis in the hippocampus has also been implicated in the pathophysiology of MDD. Human in vivo imaging studies have demonstrated a reduction in the volume of the hippocampus in depressed patients [33], and this effect can be ameliorated through antidepressant treatment [34]. Post-mortem studies also verified a decrease in the size and/or number of neuronal cell bodies in this brain area [35]. Studies in rodents have consistently shown that stress is one of the most powerful stimuli that leads to a decrease in adult neurogenesis in the hippocampus [36]. Moreover, chronic, but not acute antidepressant treatment normalizes neurogenesis in the adult hippocampus, and the inhibition of neurogenesis prevents the behavioral response to antidepressants, providing a direct link between neurogenesis and the antidepressant response [37].
Nevertheless, other mechanisms, such as the alteration in extracellular levels of monoamines in the brain might also be responsible for the characteristic action of antidepressant drugs observed in rodent models of depression [38].
In the present study, to gain insight into the cellular basis of differential behavioral responses, we used a P2rx72/2 mouse model to demonstrate that (1) the genetic deletion of P2rx7 induced neurochemical changes characteristic of antidepressant treatment, such as the dysregulation of monoaminergic transmission and elevated BDNF levels in the hippocampus, and (2) the lack of P2rx7-mediated glutamate release and inhibition of BDNF production, resulting in an enhanced hippocampal neurogenesis is a potential mechanism for the antidepressant phenotype.

Animals
This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and the local Animal Care Committee of the IEM HAS approved all experimental procedures (Permission No: 22.1/3671/003/2008). This study used 2-to 3-months old (approx. 30 g) male wild type (P2rx7+/+), and P2rx7 knockout (P2rx72/2) mouse littermates, which were housed under a 12 h on/12 h off light cycle with 60610% humidity in a temperature-controlled room (2362uC). Food and water were supplied ad libitum. Homozygous P2rx72/2 mice were bred to C57BL/6 mice. Christopher Gabel (Pfizer Inc., Groton, CT, USA) kindly supplied the original breeding pairs of P2rx72/2 mice (C57BL/6J based). An overall six backcrosses on C57BL/6 were performed for the P2rx7 knockout mouse colony used in our experiments. Offspring of this mouse line were cross-bred with P2rx7+/+ mice and the resulting heterozygotes were used as breeding stock for the F1 generation offspring employed in the behavior studies. The animals contained the DNA construct (P2X7-F1 (59-CGGCGTGCGTTTTGACATCCT-39) and P2X7-R2 (59-AGGGCCCTGCGGTTCTC-39)), which have been previously shown to generate the genetic deletion of P2rx7 [39]. The animals were genotyped using PCR analysis as described earlier [39].

Behavior Experiments
Automated Tail Suspension Test (TST). All experiments and treatments were performed during the light phase (7.00 AM-7.00 PM). The TST was conducted using an automated tail suspension device (BIO-TST2, Bioseb, France) in accordance with a previously described method [40]. The device consisted of a single tail suspension module with three chambers, connected to a personal computer that recorded the activity of the animals during the test sessions. Three mice were individually suspended by the tail onto the hooks of the device using adhesive tape (distance from tip of tail was 1-2 cm). The measurements were automatically recorded within 5-10 seconds after placing the last animal into the chamber, and each measurement lasted 6 min. During the test, the animals showed several escaping behaviors with temporary periods of immobility. The threshold level was set at 6. To examine endotoxin-induced depressive behavior, the animals were subjected to bacterial lipopolysaccharide injection (LPS, serotype 055:B5, Sigma, St. Louis, MA, USA, 1 mg/kg i.p.) at 24 h before a single 6-min test period. Brilliant blue G (BBG, 50 mg/kg i.p.) or an equal volume of saline was injected at 30 min before the test period. The time of immobility was expressed in sec. The P2rx7+/ + and P2rx72/2 mice were subjected to alternation testing. In accordance with the observations of Mayorga and Lucki (2001), some of the animals (0-16%, depending on the experiment) displayed tail-climbing behavior. The data of these animals were excluded from the calculations during the post hoc analysis.
Sucrose Preference Test (SPT). The experiments were performed using the two-bottle sucrose preference paradigm [41]. and SPT (C-H) tests. A/Genetic disruption of P2rx7 expression decreases basal immobility in the TST (n = 6,10, ***P,0.001 vs. P2rx7+/+, Student t test). The immobility time is expressed in sec. The total test period was 360 sec. B/Effect of acute BBG treatment (50 mg/kg i.p.) on LPS-induced depressive behavior in TST. LPS was added at a dose of 1 mg/kg at 24 h before testing (n = 7-9, **P,0.01 vs. SAL, Student t test). BBG was administered intraperitoneally at 30 min before testing; saline-treated mice (SAL) were injected with equal volume of saline. C/LPS-induced decline in sucrose preference is attenuated in P2rx72/2 mice. Baseline sucrose intake was measured under a habituation period of 3-4 days before treatment, and the mean volume ingested over the last 3 training sessions was used as a baseline value. The mice were treated with LPS (0.2 mg/kg i.p.) or an The decreased preference for sucrose vs. water in this test has been proposed to reflect impaired sensitivity towards reward and to model anhedonia [42]. Mice were offered a choice of 2% sucrose solution or water during 12-hour nocturnal periods (starting at 19:00). The bottles were weighed before and after each test session to monitor sucrose and water consumption. The relative position of the sucrose and water bottles were swapped every night to avoid place preferences. The baseline sucrose intake was measured under a habituation period of 3-4 days before administering the treatments, and the mean volume ingested over the last 3 training sessions was used as a baseline value. The sucrose consumption was measured for 4 consecutive nights after LPS (0.2 mg/kg i.p.) or saline administration and expressed as g/100 g body weight or as a percentage of the baseline (%). The LPS dosage was chosen based on preliminary experiments, which showed that a 0.2 mg/ kg dose elicits a relatively selective anhedonia response.

[ 3 H]Glutamate ([ 3 H]Glu) Release Experiments
The [ 3 H]Glu release experiments were conducted using the method described in our previous papers [20]. Briefly, the mice were anaesthetized under light CO 2 inhalation, and subsequently decapitated. The hippocampus was dissected from each mouse in ice-cold Krebs solution saturated with 95% O 2 and 5% CO 2  Subsequently, the perfused samples were collected over a 3-min period and measured for tritium content. At 6 min after the start of the collection, the slices were subjected to a 6-min perfusion of the agonist (ATP) at various concentrations (1, 3, 6 and 10 mM) and then changed to normal Krebs solution until the end of the collection period. In some experiments, the P2X1 receptor antagonist, NF449 (100 nM), or the selective P2rx7 antagonist Brilliant blue G (100 nM) was applied to the perfusion solution at 15 min before the beginning of the ATP perfusion.
The radioactivity released from the preparations was measured using a Packard 1900 Tricarb liquid scintillation spectrometer, using Ultima Gold Scintillation cocktail. The release of tritium was expressed as a percentage of the amount of radioactivity in the tissue at the sample collection time (fractional release). The tritium uptake in the tissue was determined as the sum of release+the tissue content after the experiment and expressed in Bq/g. For the evaluation of the basal tritium outflow the fractional release measured in four consecutive 3 min samples under drug free conditions were taken into account. The ATP-induced [ 3 H]Glu efflux calculated as the net release in response to the respective stimulus by subtracting the release before the stimulation from the values measured after stimulation. Selected samples collected under resting conditions and during the peak of the ATP-evoked response were subjected to further HPLC analyses.

Analysis of NR2B mRNA Expression Using Quantitative Real-time PCR
P2rx7 wild type and knockout mice were anesthetized under light CO 2 inhalation, and subsequently decapitated. The hippocampus from each mouse was dissected in ice-cold Krebs solution saturated with 95% O 2 and 5% CO 2 , sectioned (400 mm thick slices) using a McIlwain tissue chopper and incubated in 2 ml of modified Krebs solution in the presence of 1 ml Hibernate medium (Invitrogen Life Technology, Grand Island, NY, USA), ascorbic acid (300 mM) and Na 2 EDTA (30 mM) for 60 min. The medium was bubbled continuously with 95% O 2 and 5% CO 2 and maintained at 37uC. The hippocampal slices were collected, frozen on dry ice and stored at 270uC until further investigation. Each experimental group contained 3-4 mice. Total RNA samples were isolated and purified from the cell lysates using the RNeasy Lipid Tissue Mini Kit (Qiagen) according to the manufacturer's instructions. The RNA (2 ml) was reverse transcribed using the RevertAid First Strand cDNA Synthesis Kit (Fermentas, Vilnius, Lithuania) as described in our previous study [20,43]. Briefly, 1 mg of total RNA reverse transcribed using 1 ml of RevertAid H Minus M-MuLV reverse transcriptase in a mixture containing 5 ml of 5X reaction buffer, 1 ml of random hexamer primer (10 pmol/ml), 1 ml of RiboLock TM RNase Inhibitor (20 u/ml), and 2 ml of 10 mM dNTP mix in a final volume of 20 ml with 0.1% diethylpyrocarbonate-treated distilled water. The reverse transcription reaction was performed at 70uC for 5 min, followed by incubation at 25uC for 5 min, synthesis at 25uC for 10 min, and a final incubation at 42uC for 60 min. The resulting cDNA samples were stored at 220uC. The expression level of the target gene was determined using the cDNA samples with quantitative real-time PCR (Rotor-Gene 3000; Corbett Research, Sydney, Australia). The Real-time PCR analysis was performed according to standard protocols using a LightCycler DNA Master SYBR Green I Kit (Roche, Indianapolis, IN, USA). The PCR conditions were optimized for primers, templates and MgCl 2 . The PCR cycling protocols was set to the following conditions: initial denaturation at 95uC for 10 min followed by 40 cycles at 94uC for 15 sec, 64uC for 30 sec, and 72uC for 10 sec. The PCR primers were based on previous study of Xiaoping Du et al. [44]. The following primers were used for mRNA detection: NR2B forward primer, 59 GTG equal volume of saline (SAL), indicated by an arrow, and the sucrose intake was monitored over subsequent days. LPS substantially decreased sucrose consumption as a sign of anhedonia (F LPS (1, 63) = 1193; P,0.0001). The sucrose consumption was expressed as a percentage of the baseline (%). N = 9-14, *P,0.05 vs. P2rx7+/+. Two-way ANOVA, followed by Fischer LSD test. D-H/The P2X7 receptor antagonists Brilliant blue G (BBG), AZ-10606120 (AZ), and the selective 5-HT re-uptake inhibitor citalopram (CIT) inhibit the development of LPS-induced anhedonia in P2rx7+/+ mice. D/ The antidepressant effect of a subacute, 4-day treatment with BBG (50 mg/kg i.p./day, n = 8-20, *P,0.05, vs. control). E/The effect of subacute (AZ 2 subacute) AZ-10606120 treatment on LPS-induced anhedonia. Notably, in this protocol, AZ-10606120 significantly attenuated the decline in sucrose consumption at 24 h after LPS treatment. (F AZ (1,39) = 1179.4, P,0.00001; n = 5-9, **P,0.01, ***P,0.001 vs. control). F/The effect of acute AZ-10606120 (AZ 2, acute) treatment on LPS-induced anhedonia. When mice were pretreated with AZ-10606120 for 30 min before LPS injection at the doses indicated in the legend (mg/kg), sucrose consumption on the 2 nd and 3 rd day after LPS injection was significantly higher (n = 7-8, **P,0.01, ***P,0.001 vs. control). G and H/. Effect of 4-day subacute (CIT 15 subacute, G), and single acute treatment with citalopram (CIT 15, acute, H) on LPSinduced anhedonia. Citalopram elicited a comparable effect to AZ-10606120 using both application protocols (n = 5-10, **P,0.01, ***P,0.001 vs. control). The sucrose consumption in all experiments was evaluated at 0, 24, 48 and 72 h after LPS injection and expressed as a percentage of the baseline. Two-way ANOVA, followed by Fischer LSD test was performed as statistical analyses in SPT datasets. 2-3 months old drug-and test-naive male homozygous mice (P2rx7+/+ and P2rx72/2), weighing approximately 30 g were used in the experiments. doi:10.1371/journal.pone.0066547.g001 AGA GCT CCT TTG CCA AC; NR2B reverse primer, 59 GTC AGG GTA GAG CGA CTT GC; 18S forward primer, 59-GTAACCCGTTGAACCCCATT, and 18S reverse primer, 39-CCATCCAATCGGTAGTAGCG.
Analysis of real-time PCR measurements. To ensure reaction specificity and accurate quantification, a melting curve analysis was performed after each reaction, which confirmed the lack of primer-dimer artifacts or contamination in all cases. All DCt values were calculated using Rotor Gene 5 software (Corbett Research, Sydney, Australia). The expression level of the target genes was normalized to the expression level of 18S rRNA as a reference or housekeeping gene. The target gene and the reference gene were measured together within the same experiment. The efficiency calibrated model of Pfaffl was applied to compare the expression level of target genes between the different experimental groups [45]. Differences in the gene expression levels between the experimental groups were considered significant when the P level was ,0.05. Data are presented as the mean normalized expression ratio 6 SEM.

NR2B Immunostaining
Male P2rx7+/+ or P2rx72/2 mice (62-65 days old, appr. 30 g) from our in-house colony were used. Animals were deeply anesthetized and perfused transcardially with 4% paraformaldehyde in 0.1 M phosphate buffer (PB, pH 7.4). Brains were removed and post-fixed overnight at 4uC. The block containing the whole hippocampus was dissected out and 100 mm coronal sections were serially sectioned with a Leica vibratome and immersed free-floating in 0.1 M PB. After extensive washing, sections were treated with 0.2M HCl containing 0.2 mg/ml pepsin at 37uC for 15 min, then rinsed in PB three times and washed for 3X10 min in 0.1M Tris buffer (TBS). Nonspecific binding sites were blocked by 10% normal horse serum (Vector Laboratories, Burlingame, CA) for 2 hours and incubated with a mouse anti-NR2B (GluN2B/NR2B antibody (UC Davis/NIH NeuroMab Facility) diluted 1:1000 in TBS containing 2% normal horse serum for 24 h at 4uC. After washing in TBS three times (10 min each), anti-mouse Cy3 (Jackson ImmunoResearch Europe Ltd. Suffolk CB8 1JX, UK) secondary antibody was applied in 1:1000 for 2 hours at room temperature. Sections were washed in TB, mounted on polylysine-coated slides and cover slipped with Vectashield (Vector Laboratories). Specificity of NeuroMab GluN2B/NR2B antibody was validated on KO tissues and it did not cross react with glutamate receptors NR2A, NR2C or NR2D (see information http://catalog.antibodiesinc.com/item/neuromabs/ receptors/75-097?&plpver = 10&origin = advsrch&by = prod&filter = 0&categid = 1025&prodid = 1028).
Likely due to the HCl-pepsin treatment, careful blocking with 10% NHS and 2% NHS content in the first and second antibody Although both citalopram and AZ-10606120 alleviated anhedonia over subsequent days in these mice, the effect of AZ-10606120, but not citalopram, was significantly attenuated compared with the effect measured in P2rx7+/+ mice (n = 5-11, **P,0.01, ***P,0.001 vs. control, ## P,0.01 vs. subacute AZ in P2rx7+/+ mice, twoway ANOVA, followed by Fischer LSD test). doi:10.1371/journal.pone.0066547.g002 Confocal images were acquired at the same depth of the sections at same acquisition parameters with a Nikon A1R confocal system on an inverted Nikon Ti-E microscope (objective 20X DIC N1, numerical aperture 0.45 for large images and Plan Apo VC 606 Oil DIC N260X, numerical aperture 1.4 for details) equipped with NIS-Elements C software. Images were edited, brightness as well as contrast were adjusted using Adobe Photoshop CS3 (San Jose, CA, USA). The average intensity of NR2B immunostaining was quantified with NIH ImageJ program (U.S. National Institutes of Health, Bethesda, MD).

BDNF Protein Assay
Ex vivo study. The P2rx7 wild type and knockout mice received an intraperitoneal injection of sterile saline (0.9% NaCl) or LPS from E. Coli (Sigma, 055:B5; 1 mg/kg; 0.1 ml/mouse) and decapitated at 24 h later. The hippocampus was collected, frozen on dry ice and stored at 270uC until further investigation. Each experimental group contained 4-6 mice.
BDNF protein measurement. At the time of analysis, the samples were removed from the freezer and weighed. The BDNF extraction procedure was performed in accordance with the methods of a previous study [49]. For the BDNF assays, 2 ml lysis buffer (100 mM PIPES, pH 7, 500 mM NaCl, 0.2% Triton X-100, 0.1% NaN 3 , 2% BSA, 2 mM EDTA, 200 mM PMSF, 10 mg/ mL aprotinin, 1 mg/mL leupeptin, and 0.5 mM sodium vanadate) was added to each sample, and the hippocampus was sonicated at power level 2 using pulses at 1 sec intervals for 10-15 sec. Subsequently, the samples were centrifuged at 16,000 6 g for 30 min at 4uC. 100 ml aliquots of the resulting supernatants were removed and diluted with 400 ml of DPBS buffer (137 mM NaCl, 2.68 mM KCl, 1.47 mM KH 2 PO 4 , 8.1 mM Na 2 HPO 4 (pH 7.35), 0.9 mM CaCl 2 ?H 2 O, and 0.5 mM MgCl 2 ?H 2 O). The supernatants were collected and stored at 270?C until further analysis. For the total free BDNF measurement, the samples were acid treated with 1 ml of 1 N HCl for each 50 ml of diluted samples to decrease the pH to 2.5, followed by incubation at room temperature for 15 min. The samples were neutralized with 1 ml of 1 N NaOH for each 50 ml of diluted samples. The levels of BDNF (both precursor and mature forms) expression were evaluated using an enzyme-linked immunosorbent assay (ELISA) kit and a BDNF Emax ImmunoAssay System (Promega Corporation, Madison, WI, USA), which is specific for mouse BDNF protein, according to the manufacturer's instructions. The BDNF levels were calculated by plotting the optical density (OD) of each sample against the standard curve. A seven-point standard curve using two-fold serial dilutions in Reagent Diluent (according the manufacturer's instructions) and a high standard of 500 pg/ml were used for the determination of BDNF levels. The assay detection limits were ,15 pg/ml. Absorbance was measured at 450 nm using a Perkin-Elmer Victor 3V 1420 Multilabel Counter.

5-Bromo-2-deoxyuridine (BrdU) Incorporation Experiment
BrdU staining procedure. Male P2rx7+/+ and P2rx72/2 mice were used. The animals received intraperitoneal injections of 150 ml (3650 mg/kg body weight) BrdU Labeling Reagent (Invitrogen) for 3-5 consecutive days. The animals were anesthetized and sacrificed on the day after the last injection of BrdU. The animals were perfused transcardially with 4% paraformaldehyde in 0.1 M phosphate buffer (PB, pH 7.4). The brains were removed and post-fixed overnight at 4uC. The entire hippocampus was dissected, and 40 mm coronal sections were serially sectioned using a Leica vibratome and immersed free-floating in 0.1 M PB. A onein-six series of sections from every animal was used for the cell counts. The sections were washed (365 min) in 0.1 M PBS (pH 7.4). Peroxidase labeling was used to assess BrdU incorpora-(A) as S1 and S2 were analyzed. ATP perfusion (10 mM) significantly increases the endogenous Glu level in the effluent. Results are expressed as pmol/3 min. N = 8, ***P,0.001. C/Concentration-response relationship of ATP-evoked [ 3 H]Glu efflux in P2rx7+/+ and P2rx72/2 mice. Experiments were performed according to the protocol shown in (A), using different concentrations of ATP, as indicated in the abscissa. The net ATP-induced release was calculated and expressed as the fractional release (%). The curves represent the mean 6 S.E.M. of 4-12 identical experiments. D/Changes in the mRNA expression levels of the NMDA-NR2B receptor in hippocampus obtained from P2rx7+/+ and P2rx72/2 mice. Immediately after the 60min incubation, the brain slices were removed and total RNA was extracted from the hippocampus and reverse transcribed to cDNA. Quantitative SYBR Green real-time PCR was performed using specific primers, as described in Methods, and cDNA as a template. The experiments were repeated two times with similar results. The expression level of the NR2B receptor was normalized to that of the distinct housekeeping gene, 18S rRNA. The data are displayed as the mean 6 SEM. Asterisk indicates significant difference from the P2rx7+/+ mice (*P,0.05, Student's t-test). E-I/ Immunofluorescence staining for NR2B on hippocampal sections of P2rx7+/+ and P2rx72/2 mice. Besides the typical dotted staining most likely presenting NR2B -immunolabelled terminals (arrows, F, H) some staining is observable around cell bodies especially in the pyramidal cell layer (three arrowheads G). The whole staining is more intense in the hippocampal section of P2rx72/2 mouse (G, I). The most intense staining is observed in the CA3 regions (CA3 on E, G and F, H), while stratum oriens (or) shows the least immunoreactivity either in P2rx7+/+ and P2rx72/2 sample. Contrary to P2rx7+/+ staining (E), intense immunofluorescence illustrates hilus region on P2rx72/2 (G) sample. Images acquired at higher magnification (F,H) also show blood vessels (arrowheads), stained at the same level in both sections (background staining). bars: 50 mm in F, H, 500 mm in E, G. I. Immunofluorescence staining intensity for NR2B in P2rx7+/+ and P2rx72/2 mice. Average intensity was quantified with NIH ImageJ program (U.S. National Institutes of Health, Bethesda, MD) and is expressed in arbitrary units. Asterisk indicates significant difference from the wild type mice (*P,0.05, Student's t-test, n = 3). doi:10.1371/journal.pone.0066547.g003 The animals were sacrificed by decapitation, and the hippocampus was dissected and frozen in liquid nitrogen. The frozen tissue was weighed and homogenized in ice-cold 0.1 M perchloric acid containing 10 mM theophylline as internal standard and 0.5 mM sodium metabisulfite. The suspension was centrifuged at 3006g for 10 min at 4uC. The perchloric anion was precipitated using 1 M KOH and separated through centrifugation. The protein content of the pellet fraction was determined according to the method of Lowry [50]. The supernatant was stored at 220uC until further analysis. The biogenic amines were measured using two-dimensional reversed-phase and ion-pair reversed-phase chromatography as previously described [51] using a Gilson The separation of pre-column dansylated amino acids was performed using a gradient elution-working mode at ambient temperature. The mobile phase A consisted of 5/95 (v/v) 78/22 acetonitrile/methanol in 15 mM ammonium formate buffer, and the mobile phase B was composed of 90/10 (v/v 78/22) BzATP (C), Mg 2+ was omitted from the incubation solution (see Methods). The data are given as the mean level of neurotrophin 6 SEM. Asterisks indicate significant differences calculated by one-way ANOVA followed by Dunnett's test (multiple comparisons) and Student's t-test (pairwise comparisons) (*P,0.05, **P,0.01,***P,0.001). D/Effect of glutamate receptor antagonists and agonist on the basal and BzATP-induced decrease in BDNF expression in hippocampal slices of P2rx7+/+ mice. The inhibitory effect of BzATP (100 mM) could be counteracted using CNQX (10 mM), the non-NMDA-type glutamate receptor antagonist, TCN-201 (10 mM), the NMDA-NR1/NR2A glutamate receptor antagonist and RO-256981 (3 mM) the NMDA-NR2B glutamate receptor antagonist but not using MCPG (200 mM), the group I mGluR antagonist. CNQX, TCN-201 and RO-256981 was added for 15 min, MCPG was applied for 30 min, and the group I mGluR agonist, DHPG (100 mM), was administered for 10 min. Data are given as the mean level of neurotrophin 6 SEM with and without (first columns) the indicated antagonists in the presence (BzATP) or absence (CTRL) of BzATP. Asterisks indicate significant differences calculated using one-way ANOVA followed by Dunnett's test and Student's t-test (ns P.0.05, **P,0.01), as appropriate, indicated by the horizontal bars. doi:10.1371/journal.pone.0066547.g004 acetonitrile/methanol in ammonium formate buffer, pH 3.7. The mobile phase B increased linearly (at 0.11 min to 50% 17 min to 72% and 26 min to 100%, followed by a final run to 54 min), the flow rate was 0.7 ml/min. The analytical and the trap column were equilibrated for 10 min, and enrichment and clean-up procedures were performed. The dansylated derivatives were detected using an absorbance detector (Agilent 1100) at a 319 nm wavelength. The retention time of the dansylated Glu was 11.4 min. Pre-column derivatization was performed by mixing 50 ml of dansyl chloride prepared by dissolving 5 mg 5-(dimethylamino)naphthalene-1-sulfonyl chloride in acetonitrile daily and adding 50 ml of 2 M sodium carbonate, which contained norvaline (20 mM) as an internal standard in 25 ml of sample. After a 10-min reaction time at 60uC, the mixture was acidified using 25 ml of 6 M formic acid and injected onto the ''trap-column''.
The concentrations of the separated compounds were calculated using a two-point calibration curve internal standard method:    Asterisks denote significant differences related to genotype or LPS treatment (*P,0.05, **P,0.01, ***P,0.001). C/The number (B max ) of [ 3 H]Citalopram recognition sites was increased in the membrane preparation, and the synaptosomal 5-HT uptake was enhanced in the hippocampus of P2rx72/2 mice (black bars) compared with their P2rx7+/+ littermates (clear bars). Citalopram binding was expressed as fmol/mg protein, and the 5-HT uptake was expressed as fmol/mg protein/10 minutes. Experiments in P2rx7+/+ and P2rx72/2 mice were performed in the same assay. Binding parameters were calculated using Prism3 software. For experimental procedures, see Methods section. Data are given as the mean 6 SEM of four independent experiments, * P,0.05. doi:10.1371/journal.pone.0066547.g006 ascorbic acid (300 mM) and Na 2 EDTA (30 mM). Thereafter, slices were transferred to tissue chambers, and preperfused for 1 hour (flow rate: 0.65 ml/min) with modified Krebs' solution. In [ 3 H]5-HT release experiments, the medium also contained the serotonin reuptake inhibitor citalopram (300 nM). After washing, 3 min perfusate samples were collected and assayed for [ 3 H]5-HT/ [ 3 H]NA. At 6 and 36 min, two identical periods of electrical field stimulations (S1, S2; 25 V, 1 msec, 2 Hz, 240 shocks) were delivered by a Grass S88 stimulator (Grass Instruments, Quincy, MA, USA). The temperature was maintained at 37uC throughout the experiment. The radioactivity released from the preparations was measured as described above ([ 3

H-Citalopram, H-Nisoxetine and H-Dihydroalprenolol Binding Experiments
Male mice (80-85 days old) were decapitated, and the hippocampus was dissected, weighted, and frozen overnight at 270uC. The next day, the tissue from four animals was pooled, homogenized in incubation buffer using an Ultra-Turrax and centrifuged at 40 0006g for 15 minutes. used in three parallels. The incubation time was 1 hour at 25uC for citalopram, 2.5 hrs at 4uC for nisoxetine and 30 minutes at 25uC for dihydroalprenolol. The non-specific binding was determined in the presence of 1 mM paroxetine (citalopram), 10 mM desmethylimipramine (nisoxetine) and 1 mM (-) alprenolol (DHA). The membranes of P2rx7+/+ and P2rx72/2 mice were assayed in parallel. The binding was terminated using vacuum filtration on GF/B filters soaked in 0.05% polyethylenimine. The K d and the number of binding sites (B max ) were calculated using Prism3 software. The protein concentration (approx. 0.5 mg/ml) was measured according to Lowry method [50] using CuEDTA.

Measurement of 3 H-5HT Uptake in Synaptosomal (P2) Preparation of Mouse Hippocampus
The hippocampus of each mouse (80-85 days old) was homogenized in 0.32 M sucrose solution (1:10 g/ml) using a Teflon pestle homogenizer and centrifuged at 1,0006g for 10 minutes. The supernatant was centrifuged at 12,0006g for 20 minutes. The pellet containing synaptosomes was suspended in 0.32 M sucrose (1:30 g/ml). Aliquots of the synaptosomal suspension were preincubated for 5 min at 376C, in a final volume of 1 ml oxygenated (95% O 2 +5% CO 2 ) Krebs solution containing 118 mM NaCl, 4.7 mM KCl, 2.5 mM CaCl 2, 1.2 mM KH 2 PO 4 , 1.2 mM MgSO 4 , 10 mM D-glucose, 25 mM NaHCO 3 , 0.3 mM ascorbic acid and 0.01 mM pargyline (pH 7.4). After preincubation, 3 H-5HT (10 nM) was added, and the incubation was continued for 10 min. The uptake was terminated by addition of cold physiological saline, and by vacuum filtration of the samples through GF/B filters. The non-specific uptake was determined by incubation of the samples at 0uC. The protein content was measured using Lowry method [50]. The uptake in synaptosomal preparations from P2rx7+/+ and P2rx72/2 mice was measured in the same assay.

Statistics
All data were presented as the mean 6 SEM of n determinations. The statistical analyses were performed using one-way ANOVA followed by Dunnett's post-hoc test (multiple comparisons) and the unpaired Student's t-test (pair-wise comparisons), as appropriate. The SPT datasets and resting [ 3 H]Glu outflow were analyzed by two-way ANOVA followed by Fischer LSD test. The level of significance was set at P,0.05.

Behavior Experiments
Using automated recording, the basal time of immobility was 189.769.7 sec (n = 6) during the 6-min tail suspension test (TST) in P2rx7+/+ mice. This value significantly decreased in the P2rx72/2 mice (Fig. 1A). The time of immobility was also examined 24 h after LPS (1 mg/kg i.p.) injections in P2rx7+/+ mice, i.e., when acute symptoms of the LPS induced sickness behavior is restored [41]. The basal time of immobility was significantly increased in response to systemic LPS challenge, i.e., the mice displayed depressive-like behavior (Fig. 1B). The LPSinduced depressive behavior was significantly attenuated in mice pretreated with the selective P2rx7 antagonist, Brilliant blue G (50 mg/kg i.p.) (Fig. 1B).
In addition to TST, which examines active coping with behavioral stress, we also employed the sucrose preference test (SPT) to assess anhedonia, which is another core symptom of depression [53]. The basal sucrose consumption in P2rx7+/+ mice was 25.0460.76 g/100 g (n = 74), and their water and food consumption was 7.160.32 g/100 g and 10.5760.27 g/100 g (n = 74), respectively. In these experiments, a lower (200 mg/kg i.p.) LPS dose was applied, which elicited a relatively selective anhedonia response in our preliminary experiments. Indeed, the LPS (200 mg/kg i.p.) injection elicited a significant decrease in sucrose, but not water consumption at 24 h later in P2rx7+/+ mice (Fig. 1C), indicating the development of anhedonia; the sucrose consumption was restored in the subsequent days and approached the baseline value at 3 days after the treatment (Fig. 1C). The basal sucrose consumption in P2rx72/2 mice was 25.8960.94 g/100 g (n = 60), which was not significantly different from that observed in P2rx7+/+ mice. The water consumption was 6.5860.29 g/100 g (n = 60, P.0.05), and the food consumption was 9.8260.25 g/100 g (n = 60, P.0.05), which was also similar to those observed in P2rx7+/+ mice. The LPS treatment also decreased sucrose consumption in P2rx72/2 mice; however, the decrease was less pronounced (Fig. 1C).
Next, we examined the effect of two P2rx7 antagonists, BBG and AZ-10606120, in comparison with the selective 5-HT reuptake inhibitor antidepressant drug, citalopram, which was used as a reference compound. A subacute, 4-day application of BBG (50 mg/kg i.p.) did not change the sucrose consumption in P2rx7+/+ mice at one day after LPS injection, but this treatment significantly increased sucrose consumption over subsequent days (Fig. 1D), i.e., the mice exhibited an antidepressant phenotype.
The administration of AZ-10606120 (2 mg/kg i.p.), another P2rx7 antagonist, elicited a more robust antidepressant response, which not only resulted in the restoration of the decline in sucrose consumption, but the amount of sucrose consumed on the day after LPS injection was significantly higher after the subacute treatment (Fig. 1E). Treatment with AZ-10606120 (2 mg/kg i.p.) also promoted the restoration of the anhedonia upon acute injection, and on the 3 rd day after LPS injection, the mice consumed significantly higher amount of sucrose (Fig. 1F). Similar responses were observed after 4-day subacute (Fig. 1G) and acute (Fig. 1H) treatments with citalopram. None of the treatments significantly altered water consumption under basal conditions or after LPS treatments (data not shown).

[ 3 H]Glutamate Release in Acute Hippocampal Slices
After loading the hippocampal slices with [ 3 H]Glu, the uptake of radioactivity was 380668 kBq/g (n = 12) in the hippocampal slices of P2rx7+/+ mice and 390646 kBq/g (n = 8, P.0.05) in the P2rx72/2 mice; these results were not significantly different. The basal efflux of [ 3 H]Glu, when measured in a single sample was 3.3360.08% (n = 12) and 3.1460.05% (n = 8, P.0.05) in P2rx7+/+ and P2rx72/2 mice, respectively. However, when four consecutive samples under drug free conditions were taken into account, two-way ANOVA indicated a significant genotype effect (F(1,27), P,0.001, Fig. 3A), which is indicative for a lowered extracellular Glu level in the hippocampus of P2rx7 deficient mice. A 6-min perfusion with the P2rx7 agonist, ATP (1-10 mM), elicited a rapid and concentration-dependent increase in the efflux of [ 3 H]Glu, which was reversible upon washout (Fig. 3A, C). The net release of tritium evoked in response to 10 mM ATP was 3.3460.57% (n = 12, Fig. 3C). The endogenous glutamate content of the samples collected during resting condition and at the peak of ATP-evoked response was also determined using HPLC analysis (Fig. 3B), and a remarkable elevation in the level of glutamate was observed in response to ATP application.

Analysis of NR2B mRNA Expression Using Quantitative Real-time PCR and NR2B Protein Immunofluorescence
Hippocampal slices from P2rx7+/+ and P2rx72/2 mice were incubated in vitro with Hibernate medium containing Krebs solution for 60 min. Changes in the level of mRNA transcripts of the NR2B glutamate receptor subunits were measured using real-time RT-PCR. The gene expression level was normalized to the expression of the 18S rRNA reference gene. The results revealed that the NR2B receptor mRNA level was upregulated to 1.8960.0015 of the corresponding wild type values, (established as 1) normalized to 18S rRNA (Fig. 3D, n = 4, P,0.05).
The NR2B protein immunostaining was not evenly distributed on the hippocampal sections of P2rx7+/+ and P2rx72/2 mice (Fig. 3E, F, G, H). In general, the staining was more intense on sections of P2rx72/2 origin (Fig. 3I). Whereas high immunoreactivity for N2RB was found in hilus region in hippocampal sections of P2rx72/2 mice and weak in the P2rx7+/+ mice, the staining pattern was otherwise rather similar in both cases (Fig. 3  E, G). The most intensive punctate immunostainng was observed in the CA3 region, while stratum oriens was weakly stained.
To explore the regulatory role of P2X7 receptors in the local modulation of BDNF production, the neurotrophin levels in hippocampal slices in P2rx7+/+ and P2rx72/2 mice were subsequently analyzed in vitro. In these experiments, hippocampal slices were incubated in the presence of agonist and antagonist of P2rx7 and their effects on BDNF production were examined. Consistent with the ex vivo results, the basal BDNF level in the hippocampal slices of P2rx72/2 mice in these experiments was significantly higher than detected in P2rx7+/+ mice (Fig. 4B, C, 23.9460.49 pg/ml and 18.3260.39 pg/ml in P2rx72/2 and P2rx7+/+ mice, n = 8/group, P,0.001).

BrdU Incorporation Experiments
In these experiments, we used the proliferation marker BrdU to evaluate adult neurogenesis and examined the average number of BrdU-positive cells in the dentate gyrus (DG), which contains the granular cell layer and its subgranular zone at no more than 50 mm apart. There was a significantly higher average number of BrdU-positive cells in the DG of the rostral hippocampal sections of P2rx72/2 mice compared with P2rx7+/+ mice (Fig. 5A, B, C). Although the difference between the two genotypes was not robust, we consistently observed, except in one case, more BrdUpositive cells in the DG of P2rx72/2 mice than in their P2rx7+/+ littermates.
LPS treatment significantly increased 5-HT levels in the hippocampus (Fig. 6A) of P2rx7+/+ mice. This elevation was absent in P2rx72/2 mice; moreover, the 5-HT level was decreased in response to endotoxin treatment (Fig. 6A). By contrast, LPS treatment did not affect the NA content in the hippocampus in either genotype (Fig. 6B).  (Table 1). As a potential correlate of decreased NA levels in the deficiency of functional P2rx7, the electrical stimulation evoked [ 3 H]NA efflux in the hippocampus was slightly, but significantly decreased in P2rx72/2 mice (Table 1).

[ 3 H]5-HT/[ 3 H]NA Release Experiments
To examine, whether alterations in tissue monoamine levels results from a change in the uptake, we examined the 5-HT and NA transporters using the specific ligands, [ 3 H]Citalopram and [ 3 H]Nisoxetine, and measured the 5-HT uptake in synaptosomal preparations. As shown in Fig. 6C, the uptake in the synaptosomal preparation of the P2rx7+/+ mouse hippocampus was 1627647 fmol/mg protein (n = 4) using 5610 28 M 5-HT, whereas the uptake slightly but significantly (P,0.04) enhanced to 1816652 fmol/mg protein (n = 4) in P2rx72/2 mice. The [ 3 H]Citalopram-labeled 5-HT transporter demonstrated saturable binding with no significant changes in the K d value in the hippocampal membranes of either genotype (Table 2); however, a significant increase was observed in the B max of P2rx72/2 mice compared with P2rx7+/+ littermates (Fig. 6C, Table 2). To assess, whether the deletion of P2rx7 affected the NA transporter, the binding of the specific ligand 3 H-Nisoxetine was examined. No difference was observed in 3 H-Nisoxetine binding parameters between P2rx7+/+ and P2rx72/2 mice in the hippocampus ( Table 2). As b-adrenergic receptors are downregulated after chronic antidepressant treatment [54], we also studied the binding of 3 H-dihydroalprenolol. There was no difference in the affinity of the receptors or the number of binding sites in the hippocampal membranes of the two genotypes (Table 2).
In our behavior experiments, we found that P2rx72/2 mice displayed an antidepressant-like phenotype in two tests, used to assess mood-related behavioral modifications pertinent to depression. In the TST model, an increased time of basal immobility was observed in the deficiency of P2rx7 consistent with the previous results obtained by our group [25] and of others using similar behavioral protocols [23,24]. In our previous study, we also showed that similar changes could be reproduced by the pharmacological blockade of P2rx7 using the selective and in vivo active [58] P2rx7 antagonist Brilliant blue G. In the present study, we extended these data, showing that Brilliant blue G also counteracts bacterial endotoxin-induced depressive behavior, which suggests that P2rx7 regulates stress-induced depressive behavior.
Whereas the increased immobility in TST reflects the deficit in active coping strategy against negative environmental stimuli, another important aspect of depression is anhedonia, which is modeled in animal experiments by a decreased preference for sucrose in the SPT test. We have precipitated anhedonia in our experiments by a relatively mild endotoxin challenge, which elicited a selective decrease in sucrose, but not water consumption. Following the protocol, described by a previous study [41], the depression-like behavior can be temporally separated from the early symptoms of sickness behavior, such as the robust depression of motor activity and food intake [59]. Nevertheless, it should be noted that some behavioral, biochemical and electrophysiological changes persist days after LPS exposure and that even with this clear distinction LPS induced anhedonia response is only similar but not identical to depressive behavior characteristic to human major depression: whereas the former represents an adaptive response to infection, the latter occurs in the absence of immune challenge.
In our experiments, a significantly attenuated anhedonia response was observed in the deficiency of P2rx7, which could be pharmacologically reproduced using Brilliant blue G and another potent P2rx7 antagonist, AZ-10606120 [60]. Notably, AZ-10606120 was active upon acute application and displayed a higher potency than the potent antidepressant compound citalopram. Moreover, the effect of AZ-10606120, but not citalopram, was significantly attenuated in P2rx72/2 mice, confirming the involvement of P2rx7 in this effect. Interestingly, both AZ-10606120 and BBG was more effective in this test than the deletion of P2rx7. The compensatory upregulation of other P2X receptors, or the C-terminal truncated variant of P2rx7, which is not inactivated in the Pfizer-type knockouts [61] could be responsible for the discrepancy. In fact, we have previously showed the upregulation of P2rx4 in the striatum of the same P2rx7deficient mouse line [62]. Another potential explanation is a P2rx7-independent component in the effect of antagonists. On the other hand, the relatively weak effect of acute citalopram treatment is consistent with previous literature data showing that anhedonia is largely sensitive to subacute or chronic, but not acute antidepressant treatment [63].
Overall, the behavioral data obtained in this study, together with previous results [23,24,25], provides further support for the hypothesis that P2X7 receptors regulate emotional behavior in animal models of depression.
To identify cellular actions that mediate the action of P2rx7 on mood-related behavior, we have chosen the hippocampus as the target area of our present studies. The rationale of the choice of hippocampus was two-fold: 1) a previous study [24] found that in parallel with a decreased response to repeated forced swimming test in P2rx72/2 animals a reduction in c-fos immunoreactivity was detected in the amygdala and hippocampus. 2) In our previous study P2rx7-dependent neurochemical alterations have already explored in the amygdala [25]. We have to note, however, that P2rx7-dependent changes in hippocampal signaling pathways found by the present study might also underlie other hippocampal functions such as memory formation. In fact, it has been described that there is a spatial memory deficit in the deficiency of P2rx7 [55], which was found to be correlated with impaired interleukin-1b and c-Fos expression in the hippocampus. On the other hand, non-hippocampal mechanisms might also participate directly or indirectly in the action of P2rx7 on mood-related behavior.
We examined glutamatergic transmission because one important function of P2rx7 activation in the brain is to promote the release of glutamate [19,20] and previous observations from animal and human studies strongly indicate that excessive glutamatergic transmission might be involved in the pathophysiology of depressive disorders [64]. These findings include abnormalities in glutamate levels, e.g., [65], genetic polymorphism [66] and the differential expression of NMDA and AMPA receptor subunits in BPD and MDD patients [67,68]; antidepressant properties of glutamate receptor antagonists and modulators [69]; and the regulation of glutamate receptor subunits by clinically used antidepressants [70,71]. Moreover, in our previous study, consistent with the presumed dysfunction of glutamatergic transmission and consequent neuroplasticity changes in depression [69,72], the downregulation of different AMPA and metabotropic glutamate receptor subunits and the upregulation of the NR2B subunit of NMDA receptors was detected in the amygdala of P2rx72/2 mice [25]. The real-time PCR analysis conducted in the present study revealed that the upregulation of the NR2B subunit of NMDA receptors also occurs in the hippocampus, which is another component of the limbic system. In addition, the results of the present study demonstrated that ATP elicits concentration-dependent tritiated and endogenous glutamate efflux in the hippocampus, and P2X7 receptors mediate the majority of this glutamate efflux, whereas a minor, residual portion of the glutamate release detected in the P2rx72/2 mice is mediated through P2X1 receptors. We also show that the basal [ 3 H]Glu efflux is lower in the hippocampi of P2rx7 deficient animals, which is indicative for a lowered basal extracellular glutamate levels under the conditions of behavior experiments. These data are consistent with previous results showing that the activation of P2rx7 in the brain leads to increased glutamate release from the nerve terminals [19,20,22] and astrocytes [73], and both P2X7 and P2X1 receptors are expressed in the hippocampus [74]. Therefore the source of P2rx7 mediated glutamate release could be either neuronal or glial. Another possibility is that P2rx7 controls astrocytic glutamate uptake in the hippocampus; however this assumption needs further investigation.
A potential pathway, whereby increased glutamate release may lead to a change in mood is the alteration of the level of neurotrophic factors. Among them, BDNF is the most widely distributed neurotrophin in the CNS, and it plays several roles in synaptic plasticity and neuronal survival [75,76]. The role of hippocampal BDNF and subsequent neurogenesis in depression and in the therapeutic action of antidepressants is an emerging hypothesis, which is supported by many experimental data [26,77,78,79]. As a consequence, BDNF has become a key target in the pathology of several neurological and psychiatric diseases [80], and clinical studies have shown that BDNF protein expression is significantly decreased in both the serum and brain of depressed patients [29,81,82]. Therefore, we characterized the potential changes in BDNF expression in the hippocampus of wild type and P2rx72/2 mice. In our ex vivo study, we observed elevated basal BDNF levels in the hippocampus of P2rx72/2 mice compared with the corresponding saline-treated P2rx7+/+ controls. After systemic LPS challenge, we detected a decrease in BDNF protein expression in both genotypes. These results are consistent with previous data showing that hippocampal BDNF expression is significantly decreased at both the mRNA and protein levels in response to systemic LPS treatment [83,84]. However, our results also imply that whereas an alteration in BDNF levels might be a mediator of P2rx7-dependent changes in behavior in the absence of LPS, e.g., the decreased basal immobility in the TST test (Fig. 1A), P2rx7-mediated glutamate efflux and subsequent changes in BDNF levels play only a minor role in the P2rx7-dependent regulation of LPS-induced depressive behavior ( Fig. 1 B, C).
To explore the local regulatory role of P2X7 receptors in the modulation of basal BDNF production, we conducted an in vitro study. In support of the ex vivo results, the basal BDNF expression in the hippocampal slices of P2rx72/2 mice was significantly higher than in P2rx7+/+ mice, indicating a tonic inhibitory regulation of BDNF production through P2rx7. This was further confirmed using BBG, the selective P2rx7 antagonist; although BBG significantly increased the level of BDNF expression in both P2rx7+/+ and P2rx72/2 mice, its facilitatory effect was significantly attenuated in the deficiency of the P2X7 receptor.
Interestingly, similar to our in vivo behavior studies, the pharmacological blockade of P2rx7 elicited a more pronounced effect than the genetic deletion, which could be explained by compensatory gene expression changes in case of genetic deletion.
When hippocampal P2rx7 was stimulated through the P2X receptor agonist BzATP, a significant decrease in BDNF protein expression was detected in P2rx7+/+ mice, which was absent in P2rx72/2 mice and reversed using BBG. These data indicate that BDNF levels in the hippocampus are under the local regulatory influence of P2rx7. To our knowledge, our study is the first to demonstrate a role for P2rx7 in the regulation of BDNF production in the central nervous system. In contrast, in the periphery, P2X7 receptor activation could induce the vesicular release of BDNF from the Schwann cells, which in turn might play a trophic role on neurons [85].
Because our previous data indicated that P2rx7 activation primarily releases glutamate [19,20], we explored the role of glutamate receptors in the inhibitory action of BzATP on hippocampal BDNF protein expression. The inhibitory effect of BzATP was completely reversed by CNQX, the non-NMDA receptor antagonist TCN-201, the NR1/NR2A selective glutamate receptor antagonist and by the NR2B receptor selective antagonist RO-256981. These results indicate that the activation of both non-NMDA and NMDA receptors are necessary conditions for the P2rx7-mediated inhibitory regulation of BDNF level. Taken into account that glutamate receptor antagonists by themselves decreased BDNF production, consistently with the key role of synaptic glutamate receptor activation in the induction of BDNF [86,87], a possible interpretation of our results is that glutamate released by P2rx7 activation primarily acted on extrasynaptic NMDA receptors, which are able to shut-off the induction of BDNF [86], but only if synaptic glutamate receptors are also co-activated. Because among NMDA receptor subunits, NR2B are primarily localized to extrasynaptic sites in the hippocampus [88,89,90] we propose that P2X7 receptor activation leads to increased glutamate release and to subsequent overactivation of extrasynaptic NR2B receptors. NR2B activation, in turn, downregulates BDNF expression and thereby causes longlasting changes in neuronal plasticity, which might underlie pathological changes in behavior. The upregulation of hippocampal NR2B subunits after the genetic deletion of P2X7 receptors observed in these experiments also supports this proposed mechanism, consistent with the presumed dysfunction of glutamatergic transmission [91,92] and consequent changes in neuroplasticity during depression [69,72].
On the other hand, in the absence of exogenous activation of P2rx7 by BzATP, BDNF levels were decreased (CNQX, RO-258981) or unaffected (TCN-201) by ionotropic glutamate receptor antagonists, which implies that the net effect of endogenous P2rx7 activation and other signaling pathways converging on ionotropic glutamate receptors on BDNF level is stimulatory.
In contrast, effect of BzATP was not alleviated through the mGluR 1,5 selective receptor antagonist, MCPG, whereas the administration of mGluR 1,5 selective agonist DHPG, significantly increased BDNF expression in the presence and absence of BzATP. Therefore, mGluR 1,5 -mediated facilitatory modulation of hippocampal BDNF expression seems to be independent from the activation of P2rx7, consistently with the idea that BzATPmediated glutamate release preferentially activates extrasynaptic NMDA receptors. Treatment with MCPG alone resulted in decreased BDNF expression, whereas DHPG treatment increased BDNF production, consistent with the well-known stimulatory role of mGluR 1,5 receptors in BDNF production e.g. [93]. Interest-ingly, we also observed that BzATP paradoxically increased BDNF expression in the presence of the mGluR 1,5 agonist, DHPG. This effect, which requires further investigation, is most likely independent from the activation of P2rx7.
In addition to the regulation of BDNF production, previous studies have associated neurogenesis with the beneficial actions of specific antidepressants, suggesting a connection between decreased hippocampal neurogenesis and depression [94,95]. Others have hypothesized that neurogenesis might promote neuroplasticity [96,97]. Our study revealed that basal level of neurogenesis detected using BrdU staining is higher in the deficiency of the P2X7 receptor in the dentate gyrus, suggesting that the endogenous activation of P2rx7 inhibits adult neurogenesis in the hippocampus through the regulation of BDNF levels or independently from it. Thus we cannot exclude the possibility that P2rx7 directly regulates hippocampal neuronal survival/neurogenesis without the involvement of glutamate or neurotrophic factors. Consistent with this latter assumption, the results of electrophysiological studies have shown that neuronal progenitor cells (NPCs) from the adult rat hippocampus [98,99] and embryonic mouse striatum [100] express functional P2rx7. Activation of P2rx7 elicits necrotic cell death in the latter [100] and the inhibition of P2X7 receptors promotes axonal growth in cultured hippocampal neurons [101]. Therefore, further in situ analysis on the effect of the genetic deletion and pharmacological antagonism of P2rx7 on NPC survival and neurogenesis is of potential interest.
Taken together, the results obtained in this study provide several possible mechanisms to explain the antidepressant phenotype observed in the deficiency of P2rx7 [23,24]. However, the signaling pathways mediating the effect of P2rx7 activation on stress-induced depressive behavior, i.e., those evoked through bacterial endotoxins, remains to be established. Thus, we determined whether neurochemical markers of monoaminergic transmission are changed in the hippocampus after genetic deletion of P2rx7 in saline and LPS-treated animals. Under basal conditions, we detected a dysregulation of monoaminergic transmission, and these results are consistent with our previous findings in the amygdala [25]. We observed elevated 5-HT levels and a decreased 5HIAA/5-HT ratio, reflecting the increased bioavailability of 5-HT, together with decreased basal NA levels. As an explanation for the elevated 5-HT levels, the number of [ 3 H]Citalopram binding sites and the [ 3 H]5-HT uptake was significantly higher in the hippocampus in the deficiency of P2rx7, whilst there was no difference in basal and stimulation-evoked [ 3 H]5-HT release in the two genotypes. These findings are reminiscent of the effects of chronic antidepressant treatment, which increases the uptake of 5-HT in hippocampal slices [102]. However, we did not observe any change in NA transporter binding sites, and the ß-receptors were not downregulated, although it was previously published that treatment of rats with antidepressants reduced the number of ß-receptors, while serotonergic terminals remained intact [54]. The question arises whether the detected changes in serotonergic signaling in the hippocampus P2rx7 deficient mice are the cause or consequences of the alteration in glutamate and BDNF levels and neurogenesis, or parallel changes. The relationship between hippocampal serotonergic transmission, BDNF, neurogenesis and depressive behavior has been explored in numerous studies and there are several links between them. BDNF is critical for the normal development and function of central 5-HT neurons and the elaboration of behaviors that depend on the activity of these neurons [103,104,105]. As an example, elevation of extracellular 5-HT and antidepressant-like behavioral effects in the FST test by SSRI antidepressants are completely eliminated in mice deficient in BDNF (BDNF+/2) [104,105] and a decrease in the [ 3 H]5-HT uptake and [ 3 H]citalopram binding densities are found in these mice [106], the opposite that is detected in P2rx7 deficient mice.
Interestingly, LPS treatment caused a significant elevation in 5-HT expression in the hippocampus of P2rx7+/+ mice, which was completely absent in P2rx72/2 mice; moreover, the 5-HT levels were depleted in these mice. Therefore, the direct or indirect modulation of serotonergic transmission is a potential mechanism that might underlie the action of P2rx7 to regulate stress-induced changes in mood-related behavior, although the possibility that alterations in 5-HT levels occur as a compensatory change in response to changes in behavior cannot be excluded.
Finally, it should be noted that although the present study focused on P2rx7, this is not the only pathway, whereby endogenous ATP or other purines could affect depressive behavior. ATP is a ubiquitous signaling molecule and the majority of its ionotropic (P2X1-7) and metabotropic (P2Y 1,2,4,6,11,12,13,14 ) receptors are also expressed in the hippocampus and other brain regions involved in the processing of depressive behavior. The identification of the role of these receptors, however, awaits further investigation.
In conclusion, our data show that the genetic deletion of P2rx7 leads to an antidepressant phenotype associated with changes in hippocampal monoaminergic transmission, neurotrophin protein expression and adult neurogenesis in the dentate gyrus. Moreover, these data support the view that P2rx7 antagonists might have therapeutic potential in mood-related disorders.