The CB1 Cannabinoid Receptor Mediates Excitotoxicity-induced Neural Progenitor Proliferation and Neurogenesis*

Endocannabinoids are lipid signaling mediators that exert an important neuromodulatory role and confer neuroprotection in several types of brain injury. Excitotoxicity and stroke can induce neural progenitor (NP) proliferation and differentiation as an attempt of neuroregeneration after damage. Here we investigated the mechanism of hippocampal progenitor cell engagement upon excitotoxicity induced by kainic acid administration and the putative involvement of the CB1 cannabinoid receptor in this process. Adult NPs express kainate receptors that mediate proliferation and neurosphere generation in vitro via CB1 cannabinoid receptors. Similarly, in vivo studies showed that excitotoxicity-induced hippocampal NPs proliferation and neurogenesis are abrogated in CB1-deficient mice and in wild-type mice administered with the selective CB1 antagonist rimonabant (N-piperidino-5-(4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-3-pyrazolecarboxamide; SR141716). Kainate stimulation increased basic fibroblast growth factor (bFGF) expression in cultured NPs in a CB1-dependent manner as this response was prevented by rimonabant and mimicked by endocannabinoids. Likewise, in vivo analyses showed that increased hippocampal expression of bFGF, as well as of brain-derived neurotrophic factor and epidermal growth factor, occurs upon excitotoxicity and that CB1 receptor ablation prevents this induction. Moreover, excitotoxicity increased the number of CB +1bFGF+ cells, and this up-regulation preceded NP proliferation. In summary, our results show the involvement of the CB1 cannabinoid receptor in NP proliferation and neurogenesis induced by excitotoxic injury and support a role for bFGF signaling in this process.

In the adult brain, generation of new neurons is restricted to discrete areas including the subventricular zone and the subgranular zone of the dentate gyrus (1,2). Newly generated neurons have the ability to become functional and integrate into established brain circuits (3), and thus, they may contribute to cognitive functions. In addition, the observation of increased neurogenesis after brain injury has been proposed to constitute an endogenous neuroprotective response aimed at reducing brain damage (2). Thus, ischemia induces neurogenesis in the hippocampus and the subventricular zone (4 -6), and similarly, neuronal replacement from endogenous subventricular zone progenitors occurs in the striatum after brain stroke (7). Despite the described mobilization from neurogenic areas of endogenous neural progenitors (NPs) 4 after brain injury, the signaling factors involved in this protective response are still unclear. Therefore, the characterization of the endogenous factors that positively regulate this process is a matter of intense research (2).
Besides classical cell fate signaling systems such as growth factors and cytokines, neuronal activity plays a significant role in neurogenesis (3). Likewise, different neurotransmitters (e.g. Glu, ␥-aminobutyric acid, and dopamine) and neuromodulators such as opioids regulate progenitor cell proliferation and differentiation (3,8). The CB 1 cannabinoid receptor exerts an important neuromodulatory action by regulating synaptic transmission in different brain areas (9,10). Moreover, CB 1 receptor activation is also involved in the control of neural cell fate (11,12) and exerts a neuroprotective action in different in vivo models of brain injury, including excitotoxicity and ischemia (13)(14)(15). Recent studies have shown that the endocannabinoid (eCB) signaling system is expressed in NPs (16,17) and participates in the control of progenitor cell proliferation (17)(18)(19) and differentiation (20,21) in the normal brain. However, the role of the eCB system in the response of NPs that ensues upon brain injury is as yet unknown. In the present study, we therefore investigated the possible involvement of the CB 1 can-nabinoid receptor in NP proliferation in the context of the neuroregenerative response that follows excitotoxic brain injury.
Neural Progenitor Cell Culture-Multipotent hippocampal adult NPs were obtained from 8-week-old mice after dissociation of the tissue located between 1.2 and 3.0 mm, relative to bregma, from which the midbrain was removed, and then the hippocampus, as a whole, was separated gently from the corpus callosum. The tissue was digested by papain and DNase I treatment and further cultured in chemically defined medium consisting of Dulbecco's modified Eagle's and F12 media supplemented with N2 (Invitrogen), 0.6% glucose, non-essential amino acids, 50 mM Hepes, 2 g/ml heparin, 20 ng/ml EGF, and 20 ng/ml bFGF. Primary neurospheres were cultured at clonal density (1,000 cells/ml) and employed for in vitro experiments. Neurosphere generation experiments (see Fig. 2A) were performed in 96-well dishes with 100 l of medium, and the number of wells with neurospheres was quantified. Cell proliferation (see Fig. 2B) was determined by quantifying 5-bro-mo-2Ј-deoxyuridine (BrdUrd)positive cells. Neurospheres were dissociated by incubation with Accutase for 10 min and subsequent mild mechanical dissociation, treated for 16 h with 10 g/ml BrdUrd and, after plating in polyornithine-coated wells, immunostained with rat monoclonal anti-BrdUrd antibody from AbCam (Cambridge, MA). NPs were characterized for neural markers (see Fig. 1) in proliferating conditions (as above), and after differentiation as ensued by growth factor deprivation of adherent NPs in chemically defined medium, supplemented with 1% fetal calf serum and 20 ng/ml brain-derived neurotrophic factor (BDNF). Analyses of multipotentiality (supplemental Table 1) were performed in 35 single cell-derived neurospheres after differentiation in growth factor-deprived chemically defined medium supplemented with 1% fetal calf serum and the action of 500 nM KA or 20 ng/ml BDNF when compared with vehicle.
Animals and Drug Treatment-CB 1 knock-out mice generation has been described previously (22). CB 1 Ϫ/Ϫ mice (8 weeks old) and their respective wild-type littermates were treated with vehicle or KA (15 mg/kg, intraperitoneal), injected with 50 mg/kg of BrdUrd daily for 5 days, and perfused 1, 7, or 30 days later (n ϭ 4, n ϭ 4, n ϭ 6 for each group, respectively). In other experiments, rimonabant (1 mg/kg, intraperitoneal) or the corresponding vehicle (100 l of phosphate-buffered saline supplemented with 0.5 mg of fatty acid-free bovine serum albumin and 4% dimethyl sulfoxide) was co-administered with BrdUrd to wild-type mice (n ϭ 3 for each group). Animal procedures were performed according to the European Union guidelines (86/609/EU) for the use of laboratory animals.
Immunostaining and Confocal Microscopy-Brains were perfused, and immunostaining was performed in 30-m coronal free-floating sections (16,21). Sections were incubated with rat monoclonal anti-BrdUrd, mouse monoclonal anti-NeuN (Sigma), and rabbit polyclonal anti-S100␤ (Swant, Bellinzona, Switzerland) antibodies followed by secondary staining for rat, mouse, and rabbit IgGs with highly cross-adsorbed Alexa Fluor 594, Alexa Fluor 488, and Alexa Fluor 647 secondary antibodies (Invitrogen), respectively. A minimum of five coronal sections per animal was examined. A 1-in-6 series of hippocampal sections located between 1.3 and 2.1 mm posterior to bregma was analyzed using the TCS-SP2 Leica software (Wetzlar, Germany) and SP2 AOBS microscope with ϫ63 objective, two passes with a Kalman filter, and a 1,024 ϫ 1,024 collection box. Positive cells were normalized to the dentate gyrus area determined with ϫ10 objective. The absolute number of positive cells was considering the total hippocampal volume as determined by the sum of the areas of the sampled sections multiplied by the distances between them. Additional immunofluorescence experiments were performed with rabbit monoclonal anti-Ki-67 (clone SP6) antibody (LabVision, Fremont, CA) and monoclonal anti-bFGF clone bFM2 antibody (Upstate Biotechnology -Millipore, Dundee UK). Immun-ofluorescences were also performed with polyclonal EGF antibody (Upstate Biotechnology -Millipore) and rabbit antiserum against recombinant mouse BDNF. Sections were co-stained with DNA-intercalating agent TOTO-3 (Invitrogen). Growth factor immunoreactivity was determined with Metamorph-Offline software (Universal Imaging, Downingtown, PA). In vitro immunofluorescence was performed with antibodies against neural progenitor (nestin, Sox2, musashi-1), neuronal (␤-tubulin III), and the corresponding secondary fluorescent antibodies (as above). In addition, Cy3-conjugated glial (GFAP) antibody was employed. Total cells were counterstained with Hoechst 33258. Confocal immunofluorescence and image analyses were performed with the support and assistance of the Microscopy Imaging Research Unit (Complutense University, Madrid, Spain) and the Confocal Microscopy Unit (Centro Biología Molecular Severo Ochoa, Universidad Autónoma, Madrid, Spain).
mRNA Extraction, Reverse Transcription-PCR, and Quantification-mRNA was extracted with the RNeasy Protect kit (Qiagen) using the RNase-free DNase kit. cDNA was subsequently obtained using the SuperScript first-strand cDNA synthesis kit (Roche Applied Science, Welwyn Garden City, UK), and amplification of cDNA was performed with primers for mouse neural genes and KA receptor primers (supplemental Table 2). The following PCR conditions were used for most of the analyzed genes: 35 cycles (30 s at 94°C, 30 s at 58°C, and 75 s at 72°C). ␤-Tubulin III and S100␤ were performed with an extension temperature of 55 and 65°C, respectively, whereas glyceraldehyde-3-phosphate dehydrogenase was determined with 20 cycles (as above). Finally, after a final extension step at 72°C for 5 min, PCR products were separated on 1.5% agarose gels. Real-time quantitative PCR was performed with TaqMan probes for bFGF, EGF, and BDNF obtained from Applied Biosystems (Foster City, CA). Ampli-

The CB 1 Cannabinoid Receptor Is Involved in Kainate-induced Neural Progenitor Proliferation-
To investigate the mechanism involved in the regulation of NPs upon excitotoxicity, we cultured and characterized adult NPs, cells known to express cannabinoid receptors (16 -21, 23). Hippocampal progenitors formed primary neurospheres and expressed a variety of progenitor markers including nestin, Sox2, and musashi-1 (Fig. 1,  A and B). After differentiation in the presence of serum and BDNF, NPs gave rise to the different neural cell lineages and expressed neuronal (␤-tubulin III and MAP2) and glial markers (GFAP, S100␤, and myelin basic protein) (Fig. 1, A and B), whereas neurotrophin deprivation abolished their neuronal potential (supplemental Table 1). Next, we characterized the expression of KA receptors in NPs, which showed the presence of KA1R, KA2R, GluR6, and GluR7 but not GluR5 (Fig. 1C). Neurosphere generation assays were then performed in the presence of KA, and the role of the CB 1 receptor was investigated. KA increased neurosphere generation in a dose-dependent manner, and blockade of CB 1 receptors with the selective antagonist rimonabant prevented this action ( Fig. 2A). NPs were subsequently incubated in the presence of the eCBs AEA and 2AG, as well as the selective inhibitors of their degradation URB597 and URB754, respectively (24,25). eCB stimulation increased neurosphere generation in a CB 1 -dependent manner ( Fig. 2A). Results derived from neurosphere generation assays were confirmed in cell proliferation experiments by the quantification of BrdUrd incorporation. Thus, both KA and AEA increased NP proliferation (Fig. 2B).
The CB 1 Cannabinoid Receptor Is Involved in Excitotoxicityinduced Neurogenesis-As the aforementioned data point to the involvement of CB 1 receptor activation in KA-induced NP proliferation in vitro, we investigated the involvement of the   The response of hippocampal NPs was analyzed at different time points by quantifying the number of cells in the subgranular zone of the dentate gyrus expressing the endogenous proliferative marker Ki-67. Increased NP proliferation was observed in wild-type mice 24 h and 7 days after KA injection (Fig. 3, A and B), and this increase was, however, no longer detectable at day 30 after excitotoxicity (data not shown). The excitotoxicity-induced proliferative response was severely impaired in CB 1 -deficient mice (Fig. 3, A and B). This observation was corroborated by quantification of BrdUrd incorporation, which further supported that CB 1 Ϫ/Ϫ mice have lost injury-induced hippocampal proliferation (Fig. 3C). Consistently, the induction of nestin, a characteristic neuroepithe-lial progenitor cell marker, was blunted in CB 1 -deficient mice after KA treatment (Fig. 3D).
To determine whether CB 1dependent hippocampal proliferation results in effective neurogenesis, immunostaining for the presence of BrdUrd-labeled cells that co-express the phenotypic markers NeuN and S100␤ for mature neurons and astroglia, respectively, was performed one month after injury. Newly generated cells induced by excitotoxicity were detectable at this time point (Fig. 4A), and increased neurogenesis (NeuN-Br-dUrd double-positive cells) was observed in KA-treated wild-type animals (Fig. 4, B and C). In contrast, this neurogenic response was mostly lost in CB 1 -deficient mice. Excitotoxicity also induced astrogliogenesis (S100␤-BrdUrd doublepositive cells), and CB 1 deletion prevented this response (data not shown). Further support for the role of the CB 1 receptor in KA-induced neurogenesis was obtained by pharmacological regulation with the administration of rimonabant to wild-type mice. In line with the data derived from genetically modified mice, KA-induced hippocampal NP proliferation (Fig. 5A) and neurogenesis (Fig. 5B) were strongly inhibited by CB 1 pharmacological blockade. Overall, these findings support the existence of a CB 1 -mediated regulatory process that promotes neurogenesis after brain excitotoxicity. CB 1 Cannabinoid Receptor-mediated Cell Proliferation Involves bFGF Production-To investigate the mechanism of the CB 1 receptor proliferative action after excitotoxicity, we determined the production of different growth factors (bFGF, EGF, and BDNF) that are involved in the regulation of NP cell proliferation (1,2) and that have been shown to cross-talk with the eCB system (15,26,27). Real-time PCR analyses showed that bFGF was produced at significant levels in adult NP clonal cultures, whereas BDNF and EGF were not present at detectable amounts (Fig. 6A). Likewise, KA and eCBs increased bFGF transcript levels in a CB 1 receptor-dependent manner (Fig. 6B), whereas BDNF and EGF remained undetectable. The involvement of bFGF in the CB 1 -mediated response of NPs to excitotoxicity was analyzed in vehicle and KA-treated mice. Under basal conditions, bFGF expression was lower in CB 1 Ϫ/Ϫ mice than in their wild-type littermates, and excitotoxicity increased bFGF expression in wild-type animals FIGURE 7. Increased bFGF expression after excitotoxicity involves the CB 1 cannabinoid receptor. A, quantification of bFGF immunofluorescence in the hippocampus 1 day after vehicle (Veh) or KA treatment in wild-type (WT) (white bars) and CB 1 Ϫ/Ϫ (black bars) mice. B and C, EGF (B) and BDNF (C) expression was determined under the same conditions. Significantly different from vehicle-treated cells: *, p Ͻ 0.05, **, p Ͻ 0.01; #, p Ͻ 0.05 versus vehicle-treated CB 1 Ϫ/Ϫ mice. Representative images of growth factor expression (green) in wild-type and CB 1 -deficient mice are shown. Total cells were co-stained with TOTO-3 (blue). Scale bar: 45 m.
but not in CB 1 -deficient mice (Fig. 7A). EGF and BDNF expression in vivo was also up-regulated after excitotoxicity, but the reduction of their protein levels in CB 1 -deficient mice was minor when compared with bFGF repression (Fig. 7, B and C, and supplemental Fig. 1). These observations suggest that bFGF is involved in CB 1 -dependent cell proliferation upon excitotoxicity. We quantified CB 1 expression after excitotoxicity and observed a significant up-regulation (Fig. 8A) that correlated with an increased number of double-labeled CB 1 ϩ -bFGF ϩ hippocampal cells after KA administration (Fig. 8, B and C). The quantification of bFGF ϩ -proliferating (Ki-67 ϩ ) progenitor cells supported that bFGF expression precedes NP proliferation (Figs. 3A and 8D). In addition, the involvement of the CB 1 receptor in bFGF induction in progenitor cells was revealed by the analysis of CB 1 Ϫ/Ϫ mice in which KA-triggered increase of bFGF ϩ -Ki-67 ϩ cells was impaired (Fig. 8D).

DISCUSSION
Brain injury-induced neurogenesis requires the existence of complex regulatory mechanisms in which different signaling factors are required to coordinate NP cell proliferation, differentiation, survival, and migration (1,2). Fostering cell proliferation after injury requires the activation of progenitor cells in a process that, besides endogenous progenitor features, relies on extracellular signaling systems (1)(2)(3). Here we addressed the question of whether the CB 1 cannabinoid receptor might be involved in this process by regulating hippocampal progenitor cell proliferation and neurogenesis. For this purpose, we employed an excitotoxicity injury model in which the CB 1 receptor is known to be neuroprotective (15) and in which increased eCB ligands are produced (15,28). CB 1 receptor-deficient mice exhibited impaired hippocampal NP proliferation and neurogenesis after excitotoxicity. Likewise, CB 1 receptor blockade by rimonabant administration to wild-type mice effectively blocked excitotoxicity-induced neurogenesis. eCBs are produced by NPs upon intracellular calcium increase (9), and via CB 1 receptor activation, they promote hippocampal NP proliferation (17,18) and neurogenesis (20). Our results support that the eCB system is involved in excitotoxicity-induced neurogenesis, in line with its participation in the control of neural cell fate and the modulation of the balance between cell proliferation, death, and survival (11,12,23). Thus, the CB 1 cannabinoid receptor plays a neuroprotective role against brain excitotoxicity and stroke (13)(14)(15) and in different neurodegenerative disorders (11).
The existence of an injury-induced neurogenic process is considered to constitute an attempt for neuroregeneration that is well documented in ischemia and excitotoxicity models (4 -7) but, however, usually fails to assess full recovery of brain functionality (2). bFGF is known to play a crucial role in progenitor proliferation in normal (29,30) and excitotoxicity-damaged brain (5). Nonetheless, other growth factors as EGF (30,31), vascular endothelial growth factor (32), and stem-derived supporting factor (33) have also been implicated in NP proliferation in both normal and pathological brain conditions. In the present study, bFGF was up-regulated after kainate treatment of cultured progenitor cells, whereas we did not observe EGF and BDNF expression under the same conditions. In vivo analyses showed that bFGF expression was lower in CB 1 Ϫ/Ϫ than in wild-type mice and that excitotoxicity increased bFGF expression in wild-type mice but not in CB 1 -deficient littermates. EGF and BDNF expression also varied upon excitotoxicity, but the magnitude of the changes was lower when compared with those of bFGF. Our findings point to the involvement of the CB 1 receptor in hippocampal NP proliferation by increasing bFGF production, although the contribution of additional factors cannot be excluded. Previous evidences support a positive cross-talk between the eCB system and growth factor-mediated signaling pathways (15,26,27). BDNF expression is also regulated by the CB 1 receptor (34,35) and may be indeed important for appropriate neuronal maturation and positioning of newly generated cells. In addition, the CB 1 receptor antagonist rimonabant has been reported to inhibit axonal growth stimulated by bFGF, and 2AG production was implicated in this bFGF action (26). Finally, cannabinoid regulation of neurite outgrowth can be mediated by intracellular signaling pathways coupled directly to CB 1 receptor activation (16,36).
The CB 1 receptor is ideally located to regulate neural cell fate decisions as it is present along the neural lineage from progenitor to differentiated cells (9,23), its expression increases after injury in various in vivo models (37,38), and its activation regulates neural cell survival and proliferation (11,12), migration, and axonal growth (23). In addition, the synthesis of the endogenous ligands AEA and 2AG is increased after different types of Ϫ/Ϫ mice.
brain insults including excitotoxicity and during neuronal activity (9,10). Although the neurogenic response to injury may be of benefit for neural activity (2), seizure-induced neurogenesis may contribute as well to aberrant hippocampal reorganization (39). In this context, chronic pharmacological administration of the synthetic cannabinoid HU-210 induces neurogenesis, and this has been associated with improvement in anxiety and depression tests (20). Likewise, selective activation of CB 2 cannabinoid receptors, which is devoid of undesirable psychoactive actions (9,40,41), results in increased progenitor proliferation (19), and this might be of benefit for therapeutic manipulation of neural stem cells after brain injury.
Our results indicate that similarly to other extracellular lipid signaling systems (e.g. sphingosine-1-phosphate and lysophosphatidic acid) that act via G-protein-coupled Edg receptors (42,43), eCBs via the CB 1 receptor exert a regulatory role on NP cell proliferation and differentiation. Overall, the functionality of the eCB system in adult brain neurogenesis, together with the ability of cannabinergic drugs to modulate neurogenesis after brain damage, may open new strategies aimed at improving the insufficient neurogenic response of endogenous NPs to excitotoxicity or ischemia.