B R A I N R E S E A R C H 1 3 4 9 ( 2 0 1 0 ) 1 6 2 – 1 7 3

ava i l ab l e a t www.sc i enced i r ec t . com

www.e l sev i e r . com/ loca te /b ra i n res

Research Report

Early maternal deprivation induces changes on the expressionof 2-AG biosynthesis and degradation enzymes in neonatalrat hippocampus

Juan Suáreza,⁎, Patricia Riveraa, Ricardo Llorenteb, Silvana Y. Romero-Zerboa,Francisco J. Bermúdez-Silvaa,c, Fernando Rodríguez de Fonsecaa, María-Paz Viverosb,⁎aLaboratorio de Medicina Regenerativa, Fundación IMABIS, Hospital Carlos Haya, 29010 Málaga, SpainbDepartamento de Fisiología (Fisiología Animal II), Facultad de Biología, Universidad Complutense de Madrid, 28040 Madrid, SpaincNeurocentre Magendie, Inserm U862, 33077, Bordeaux, France

A R T I C L E I N F O

⁎ Corresponding authors. M.-P. Viveros is toUniversidad Complutense de Madrid, 28040Carlos Haya 82, 29010 Málaga, Spain.

E-mail addresses: juan.suarez@fundacion

0006-8993/$ – see front matter © 2010 Elsevidoi:10.1016/j.brainres.2010.06.042

A B S T R A C T

Article history:Accepted 17 June 2010Available online 25 June 2010

Early maternal deprivation (MD) in rats (24 h, PND 9–10) is a model for neurodevelopmentalstress. Our previous data showed that MD altered the hippocampal levels of theendocannabinoid 2-AG and the expression of hippocampal cannabinoid receptors in 13-day-old rats, withmales beingmoremarkedly affected. The aim of this study was to analyzethe impact of MD on the enzymes involved in 2-AG biosynthesis (DAGLα and DAGLβ) anddegradation (MAGL) in relevant areas (DG, CA1, CA3) of the hippocampus in 13-day-oldneonatal rats. The expression of the enzymes was evaluated by quantitative RT–PCR,immunohistochemistry, and densitometry. MD induced a significant increase in DAGLαimmunoreactivity in both males and females, which was mainly associated with fibers inthe polymorphic cell layer of the dentate gyrus and in the stratum pyramidale of CA3. Incontrast, the molecular layer of the dentate gyrus showed a significant decrease in DAGLαimmunoreactivity in MD males and females. No changes were observed in DAGLβimmunoreactivity. MD induced a significant decrease in MAGL immunoreactivity inhippocampal CA3 and CA1 areas, more marked in males than in females, and that wasmainly associated with fibers in all strata of CA3 and CA1. The results also showed asignificant decrease of MAGL mRNA levels in MD males. These data support a clearassociation between neurodevelopmental stress and dysregulation of the endocannabinoidsystem. This association may be relevant for schizophrenia and other neurodevelopmentalpsychiatric disorders.

© 2010 Elsevier B.V. All rights reserved.

Keywords:Neonatal ratMaternal deprivationNeurodevelopmentHippocampus2-AG biosynthesis and degradationenzymes

be contacted at Departamento de Fisiología (Fisiología Animal II), Facultad de Biología,Madrid, Spain. Suárez, Laboratorio de Medicina Regenerativa, Fundación IMABIS, Avenida

imabis.org (J. Suárez), pazviver@bio.ucm.es (M.-P. Viveros).

er B.V. All rights reserved.

163B R A I N R E S E A R C H 1 3 4 9 ( 2 0 1 0 ) 1 6 2 – 1 7 3

1. Introduction

Exposing neonatal rats [postnatal day (PND) 9] to a singleprolonged 24-h episode of maternal deprivation (MD) hasbeen shown to induce long-term behavioral alterations thatresemble psychotic-like symptoms. Thus,MD rats showed, inthe adulthood, disturbances in prepulse inhibition, latentinhibition, and auditory sensory gating and startle habitua-tion. Moreover, MD led to a significant reduction in hippo-campal levels of polysialylated-neuronal cell adhesionmolecule (PSA-NCAM) and brain-derived neurotrophic factor(BDNF) as well as a significant decrease in themRNA levels ofglutamate N-methyl-D-aspartate (NMDA) receptor subunits,NR- 2A and NR-2B. These changes are suggestive of a loss ofsynaptic plasticity and hypofunctionality of the glutamater-gic system, as recently postulated for schizophrenia (see forreview, Ellenbroek and Riva, 2003; Ellenbroek et al., 2004).

Several lines of evidence support an association betweenan altered endocannabinoid system (ECS) and the pathogen-esis of schizophrenia. For example, increases in CB1 cannabi-noid receptor expression have been found in the prefrontalcortex (Dean et al., 2001) and cingulate cortex (Zavitsanou etal., 2004) of schizophrenic patients. Also, elevated levels of theendocannabinoid anandamide have been detected in thecerebrospinal fluid of schizophrenics (Leweke et al., 1999,2007; Giuffrida et al., 2004). Furthermore, frequent cannabisuse significantly increases the risk for psychotic symptomsand schizophrenia (see for review, Di Forti et al., 2007; Lewekeand Koethe, 2008). In addition, glutamatergic transmission isregulated by retrograde endocannabinoid signaling, and boththe endocannabinoid and glutamatergic systems are involvedin the pathophysiology of major symptoms of schizophrenia(Leweke and Koethe, 2008). Previous data from our groupindicated that MD resulted in alterations of the hippocampalendogenous cannabinoid system (Llorente et al., 2008; Suárezet al., 2009), as it has been described for the glutamatergictransmission (Roceri et al., 2002; Pickering et al., 2006), furthersupporting the view that MDmodifies neurochemical systemsthat are also altered in schizophrenia. We first reported anincrease of the endocannabinoid 2-arachidonylglycerol (2-AG)in the hippocampus of MD male but not in MD female rats atPND 13 (Llorente et al., 2008). More recently, we demonstratedthat, at this same age, MD male rats showed a significantlydecreased expression of hippocampal CB1 receptor, whereashippocampal CB2 receptor expression was significantly in-creased in MD animals of both genders (Suárez et al., 2009).These findings provided the first direct evidence for a linkbetween MD stress and the ECS that appeared to be expressedin a gender-dependent manner.

One of the most prevalent hypotheses for the pathogenesisof schizophrenia states that the disease is a neurodevelop-mental disorder associated with early brain developmentalabnormalities (Weinberger, 1987; Lewis and Levitt, 2002; Marekand Merchant, 2005). From this perspective, behavioral deficitobserved in mature MD animals might be related to alteredneurodevelopmental processes triggered by stress-inducedincreases in glucocorticoid levels. In fact, MD, acutely, leads tohigh levels of corticosterone that persist elevated at PND 13(Llorente et al., 2008; Viveros et al., 2009) and the developmental

hippocampus that shows a high density of glucocorticoidreceptors and appears to be particularly sensitive to thedetrimental effects of glucocorticoids (Sapolsky et al., 1988;Gould, 1994; Sapolsky, 2000) is clearly affected by MD (bothneurons and glia) (Llorente et al., 2009), supporting thehypothesis that MD might be a model of neuropsychiatricsymptoms with a neurodevelopmental basis. There is evidencefor a crucial implication of the ECS in brain developmentalprocesses such as neural progenitor proliferation, lineagesegregation, migration and phenotypic specification of imma-ture neurons, axonal elongation, and synaptogenesis (see forreviews, Fernández-Ruiz et al., 2000; Berghuis et al., 2007;Harkany et al., 2008). Retrograde signaling involving endocan-nabinoids appears to be responsible for the homeostatic controlof synaptic transmission and the resulting network patterns inthe immature hippocampus (Bernard et al., 2005). Thus, it islikely that there is a functional link between the hippocampalneuronal and glial developmental alterations observed in theMD neonatal rats and their altered ECS.

Studies on the role of the ECS in the control of neurodeve-lopmental processes have mainly focused in CB1 receptors.However, there is scarce information on the presence andfunctions of other components of the ECS, such as enzymesinvolved in 2-AG biosynthesis, diacylglycerol lipase α and β(DAGLα and DAGLβ), and degradation, monoacylglyceridelipase (MAGL), whose functions point to a spatial and temporalregulation of the 2-AG level in the brain (Dihn et al., 2002;Bisogno et al., 2003; Suárez et al., 2008). In addition, DAGLactivity is related to axonal growth and guidance duringdevelopment (Brittis et al., 1996; Williams et al., 2003).Supporting this hypothesis, the expression of DAGL isozymes(α and β) changes during development of the mouse brain,fromaxonal tracts of the embryo to dendritic fields of the adult(Bisogno et al., 2003). In turn, this correlates with thedevelopmental change in the requirement for 2-AG synthesisfromthepre- to thepostsynaptic compartment, for instance, inthe hippocampal pyramidal dendritic field (Yoshida et al.,2006). Moreover, the substantial down-regulation of DAGLβcontrasts with the high DAGLα expression in the adult mousecerebellum (Bisogno et al., 2003). Finally, most brain 2-AGhydrolase activity can be ascribed to MAGL (Blackman et al.,2007), where it localizes to presynaptic terminals (Gulyas et al.,2004) and regulates the inhibitorypostsynaptic currents (IPSCs)by themodulationof thebasal 2-AG tone (Hashimotodani et al.,2007). Its role in development is linked to its focuseddistribution on the presynaptic compartment where it helpsto establish effective retrograde signaling. These findingsprovide additional evidence supporting relevance for theregulation of 2-AG/cannabinoid receptor signaling in neuro-developmental alterations.However, the effect of dysregulatedMAGL activity on development and adult phenotype remainsto be determined conclusively. As above mentioned, we haveshown that precisely this endocannabinoid is significantlyincreased in the developmental hippocampus of MD rats.Thus, in the present study, we have focused on the impact ofMD on the enzymes involved in 2-AG biosynthesis (DAGLα andDAGLβ) and degradation (MAGL), in the hippocampus of 13-day-old rats, the same age at which we have previouslyreported the hippocampal cellular, biochemical (endocanna-binoid levels), and immunohistochemical (cannabinoid CB1

164 B R A I N R E S E A R C H 1 3 4 9 ( 2 0 1 0 ) 1 6 2 – 1 7 3

and CB2 receptors) (Llorente et al., 2008, 2009; Suárez et al.,2009) changes induced by MD stress. Our results provide a fullcharacterizationof the expressionof theseenzymes (evaluatedby quantitative RT–PCR, immunohistochemistry, and densi-tometry) in the developmental hippocampus of control maleand female rats and show that they are affected byMD in awaythat is compatible with the previously reported changes of 2-AG levels in this brain area.

2. Results

In the present study, we mapped and quantified the expres-sion of DAGLα, DAGLβ andMAGL in the dentate gyrus, and thehippocampal CA1 and CA3 areas in 13-day-old neonatal ratsby quantitative RT–PCR, immunohistochemistry, and densi-tometry. Previously, to perform quantitative RT–PCR andimmunohistochemical techniques, we analyzed controls andWestern blot results to demonstrate that DAGLα, DAGLβ, andMAGL antibodies recognized the corresponding antigen in therat hippocampus.Western blot analysis of membrane extractsfrom Wistar male rat hippocampus revealed prominentimmunoreactive bands of expected molecular masses of120 kDa for DAGLα, 76 kDa for DAGLβ, and 37 and 35 kDa forMAGL (Fig. 1). These data were corroborated by DAGLα,DAGLβ, and MAGL immunoreactivity in the hippocampus ofadult Wistar rats (Fig. 2).

2.1. DAGLα immunoreactivity

Representative microphotographs of DAGLα immunoreactiv-ity in DG, CA3, and CA1 corresponding to the four experimen-tal groups are shown in Fig. 3. DAGLα immunoreactivity was

Fig. 1 – Western blots of membrane extracts from rathippocampus show prominent immunoreactive bands ofexpectedmolecular masses of 120 kDa for DAGLα, 76 kDa forDAGLβ, and 37 and 35 kDa for MAGL. Positions of molecularmarkers are indicated on the left.

Fig. 2 –DAGLα (A), DAGLβ (B), andMAGL (C) immunoreactivityin the hippocampus of adult Wistar rats. We observed anintense immunostaining of DAGLα and MAGL but a weakerimmunoreactivity for DAGLβ. Inset in A shows that DAGLαimmunoreactivity was present in fibers, crossing thepyramidal cell layer and puncta that formed two bandsadjacent to the pyramidal cell layer in both strata oriensand radiatum of CA3. Inset in C shows that MAGLimmunoreactivity was located surrounding pyramidalsomata. CA1/CA3, cornu ammonis field 1/field 3; DG,dentate gyrus; S, subiculum; SO, stratum oriens; SP,stratum pyramidale; SR, stratum radiatum. Scale bars: A,B, C=500 μm; inset in A=40 μm; inset in C=20 μm.

very prominent in neuropil of the molecular layer of DG (Figs.3A–D), the stratum lucidum and stratum oriens of CA3 (Figs.3E–H), and the stratum lacunosum-moleculare of CA1 (Figs.3I–L). On the basis of previous neuroanatomical studies byhigh resolution images and electron microscopy (see Discus-sion), DAGLα-immunopositive (DAGLα+) neuropil and punctaprobably represent the dendritic spines of the granular cells in

165B R A I N R E S E A R C H 1 3 4 9 ( 2 0 1 0 ) 1 6 2 – 1 7 3

DG, and the apical dendritic tree (in the stratum lucidum) andthe basal dendritic tree (in the stratum oriens) of thepyramidal cells in CA3. In the stratum lacunosum-moleculareof CA1, immunoreactivity probably stains the distal apicaldendritic terminals of pyramidal and nonpyramidal neurons.In the molecular layer of DG and the strata lacunosum-moleculare of CA1, the staining for DAGLα is very intense,suggesting that these structures may be closely related to theentorhinal afferents (perforant path). Weakly stained neuropilwas evident in the polymorphic cell layer of DG and in thestrata oriens and radiatum of CA1. In these strata of CA1,DAGLα staining may localize to the dendrites of pyramidalneurons (Figs. 3I–L). Regarding to hippocampal principal celllayers, very low DAGLα immunoreactivity was detected in the

granular cell layer of DG, but a number of fibers can beobserved surrounding pyramidal cells.

2.2. DAGLβ immunoreactivity

Fig. 4 shows representative microphotographs of DAGLβimmunoreactivity. Very weak DAGLβ immunoreactivity wasobserved in all fields of the hippocampus at PND 13 incomparison with adulthood.

2.3. MAGL immunoreactivity

Fig. 5 shows representative microphotographs of MAGLimmunoreactivity of DG, CA3, and CA1 areas correspondingto the four experimental groups. The distribution of MAGLimmunoreactivity in the hippocampus was characterized byan intense network of MAGL-immunopositive (MAGL+) fibers,which possibly represent axon terminals on the basis ofprevious neuroanatomical studies by high resolution imagesand electron microscopy (see discussion). Staining was morepronounced in the stratum radiatum and stratum oriens ofCA3 (Figs. 5E–H) and CA1 (Figs. 5I–L). Numerous MAGL-containing pericellular basket terminals surrounded numer-ous unstained pyramidal cell somata, mostly in the CA3 field(Figs. 5E and G), suggesting that, at least, axon terminals ofinhibitory interneuronswere stained. However, in the stratumradiatum of CA1 and CA3, part of MAGL+ fibers may representexcitatory Schaffer collaterals, and probably also recurrentcollaterals. In addition, in the stratum oriens of CA1 and CA3,part ofMAGL+ fibersmay also constitute excitatory fibers fromthe contralateral hippocampus and medial septal nucleus. Inthe molecular layer of DG and the stratum lacunosum-moleculare of CA1, the staining is much weaker for MAGL,suggesting that entorhinal afferents (perforant path) lack thisenzyme. However, someMAGL+ fibers could be detected in thegranular and polymorphic cell layers of DG.

2.4. Quantification of DAGLα, DAGLβ, and MAGLimmunoreactivity

Results corresponding to the quantification of DAGLα immuno-reactivity are shown in Fig. 6A. In general, we detected

Fig. 3 – Representative microphotographs of DAGLαimmunoreactivity in relevant subregions (DG, CA1, and CA3)of the hippocampus in 13-day-old rats; effects of maternaldeprivation. The intense DAGLα immunoreactivity in thehippocampus was mainly associated with a dense networkof fibers, possibly dendrites, in the molecular andpolymorphic cell layers of DG (A–D), in the stratumoriens andthe suprapyramidal sublayer of stratum radiatum (stratumlucidum) of CA3 (E-H), and in the stratumlacunosum-moleculare of CA1 (I–L). CA, cornu ammonis; Co,control; DG, dentate gyrus; gcl, granular cell layer; MD,maternal deprivation; ml, molecular layer; pcl, polymorphiccell layer; SL, stratum lucidum; SL-M, stratumlacunosum-moleculare; SO, stratum oriens; SP, stratumpyramidale; SR, stratum radiatum. Scale bars are included inimages A, E, and I. A, E, I: male Co; B, F, J: male MD; C, G, K:female Co; D, H, L: female MD.

Fig. 4 – Representative microphotographs of DAGLβimmunoreactivity showing DG, CA1, and CA3 of 13-day-old rathippocampus. DAGLβ immunoreactivity was very weak, andno differences between experimental groups were found. CA,cornu ammonis; Co, control; DG, dentate gyrus; MD, maternaldeprivation. Scale bars are included in each image. A:male Co;B: male MD; C: female Co; D: female MD.

166 B R A I N R E S E A R C H 1 3 4 9 ( 2 0 1 0 ) 1 6 2 – 1 7 3

significant differences between the hippocampal areas ana-lyzed (DG, CA3, and CA1) [F=80.34, P<0.0001], and significantinteraction between the brain area and treatment factors[F=3.16, P<0.01], that is, MD produced different effects depend-ing on the hippocampal area. The MD procedure induced anincrease inDAGLα-positive fibers in several hippocampal areas,an effect that was significant in SP of CA3 and the polymorphiccell layer of DG in bothmales and females: [SP CA3,MD: F(1,24)=6.09, P<0.05; pcl DG,MD: F(1,24)=4.85, P<0.05]. Interestingly, themolecular layer of DG showed a significant decrease in DAGLαimmunoreactivity in bothMDmales and females: [ml DG,MD: F(1,23)=12.56,P<0.01].TheANOVAtestalso revealedasignificantinteraction between sex and treatment factors for DAGLαimmunoreactivity in CA1: [SP CA1, MD×sex: F(1,24)=6.26,P=0.0196], that is, sex does not show the same effect in bothcontrol and MD procedures.While the level of DAGLα protein ishigher in MD males than control males, the level of DAGLαprotein is lower in MD females than in control females. Nodifferences were found among the four experimental groupsanalyzed in DAGLβ immunoreactivity.

Results corresponding to the quantification of MAGLimmunoreactivity are shown in Fig. 6B. We also detectedsignificant differences between the hippocampal areasanalyzed (DG, CA3, and CA1) [F=48.63, P<0.0001] andbetween treatment (Co and MD) [F=43.687, P<0.0001]. Inaddition, a significant interaction was observed between thebrain area and treatment factors [F=2.87, P<0.01], that is, MDproduced different effects depending on the hippocampalarea and between sex and treatment factors [F=9.57, P<0.01],that is, sex does not show the same effect in both control andMD procedures. In contrast to the findings in relation toDAGLα, the MD procedure induced very significant decreasesin MAGL immunoreactivity mainly in the hippocampal CA1and CA3 areas, with this effect being more marked in malesthan in females. Statistical analyses rendered significant

effects of MD: SO CA3, [F(1,23)=9.36, P<0.01]; SR CA3, [F(1,23)=10.98, P<0.01]; SP CA3, [F(1,23)=6.79, P<0.05]; SO CA1, [F(1,22)=7.03, P<0.05]; SP CA1, [F(1,23)=10.46, P<0.01]; SR CA1, [F(1,23)=7.65, P<0.05]; SL-M CA1, [F(1,23)=10.53, P<0.01]. The ANOVAtest also revealed a significant interaction between MD and sexwhen analyzing MAGL immunoreactivity in CA1: SO CA1, [F(1,22)=4.67, P<0.05], that is, sex doesnot show thesameeffect inboth control and MD procedures.

2.5. DAGLα, DAGLβ, and MAGL mRNA expression

The relative abundance of the gene transcripts is shown inFig. 7. No significant effects of MD, sex, or interaction betweenfactorswere found for the relative levels of DAGLα, DAGLβ andMAGLmRNA expression. However, a strong trend to increasedDAGLβmRNA expression inMD females was observed (P=0.06by Student's t test; t=1.995; df=12). In contrast, we found asignificant decrease of MAGL mRNA in MD males (P<0.05 byStudent's t test; t=2,128; df=13). In addition, a marked sexualdimorphism was found for MAGL mRNA expression, withcontrol males showing a significantly higher expression thancontrol females (P=0.01 by Student's t test; t=2,987; df=10).

3. Discussion

The present maternal deprivation procedure includes a combi-nation of three stressors, i.e., (1) lack ofmaternal care during thematernal separation episode (although the maternal behavioris increased thereafter; unpublished data), (2) lack of food, and(3) decrease in body temperature. It is very likely that it is thecombination of three stressors that will determine the long-term outcome of maternal deprivation. From our own previousresults, we can deduce that a lack of food might be especiallyrelevant in the final outcome. Thus, maternal deprivationinduces a clear increase in plasma corticosterone levelstogether with a significant decrease in leptin levels (Viveros etal 2009, 2010). Changes in body temperature also appear to beimportant. For example, Zimmerberg and Shartrand (1992)specifically investigated the effects of temperature (althoughthey used a different deprivation method) and they found thatdeprivation at room temperature led to a significant increase insensitivity to the dopamine releasing agent amphetamine,whereas deprivation at 35–37 °C had no effect on the sensitivityto amphetamine. In viewof the above data, we propose that it isthe combination of three stressors that will determine most ofthe effects observed in maternal deprivation, including effectson the endocannabinoid system.

On the basis of diverse long-term behavioral and neuro-chemical alterations induced by early MD in rats (24 h, PND 9–10), this procedure was proposed as a potential animal modelfor certain psychotic symptoms (see for reviews, Ellenbroekand Riva, 2003; Ellenbroek et al., 2004). Based on theneurodevelopmental theory, we hypothesised that at leastsome of those long-term abnormalities might be related to analtered neurodevelopment in MD rats, and in fact, we did finda series of endocrine (increased plasma corticosterone) andhippocampal neuronal and glial alterations that appeared tosupport the hypothesis (Llorente et al., 2009; De la Fuente et al.,2009). Moreover, on the basis of (1) the well-known

167B R A I N R E S E A R C H 1 3 4 9 ( 2 0 1 0 ) 1 6 2 – 1 7 3

implications of the endocannabinoid system in both, braindevelopment and schizophrenia and (2) our findings of alteredresponses of MD animals to cannabinoid agonists (see forreview, Marco et al., 2009), we hypothesized that MD couldhave an impact on the developing ECS. The first directevidence came when we found that the endocannabinoid 2-AG was significantly increased in the hippocampus of 13-day-old MD male rats (Llorente et al., 2008), and more recently, wehave shown that MD alters hippocampal cannabinoid recep-tors in these animals (Suárez et al., 2009). In line with ourresults, it has been reported that serum 2-AG concentrationwas increased after exposure to social stress in adult animals(Hill et al., 2009). Moreover, it has been recently suggested thatglucocorticoids shifts arachinonic acid to the biosyntheticpathways of endocannabinoids (Malcher-Lopes et al., 2008),

further supporting our findings. To further investigate possi-ble effects of MD stress on the 2-AG signaling pathway, wehave now evaluated the expression of enzymes involving 2-AG biosynthesis (DAGLα and DAGLβ) and degradation (MAGL)in relevant hippocampal subregions of neonatal rats of thesame age. 2-AG is one of the two major endocannabinoids inthe brain (Devane et al., 1992; Mechoulam et al., 1995; Sugiuraet al., 1995), and it is synthesized on demand from diacylgly-cerol by sn-1-specific diacylglycerol lipase (DAGL) (Bisogno etal., 2003). The fact that DAGL activity is required for axonalgrowth during development and for retrograde synapticsignaling at mature synapses suggests a 2-AG requirementto manage the changes during brain development (Brittis etal., 1996; Bisogno et al., 2003; Williams et al., 2003; Yoshida etal., 2006). Despite the growing interest in 2-AG/CB1 signaling,there are few studies in terms of pathophysiologic conditionsincluding stress or neurodegenerative diseases (Di Marzo etal., 2000; Panikashvili et al., 2001). The present results showthat MD stress produced an increase in the expression ofDAGLα in some hippocampal areas. However, DAGLα expres-sion levels were specifically reduced in the molecular layer ofDG, a region that directly receives entorhinal afferents.Paradoxically, we did not detect changes in DAGLα mRNAexpression after MD episode neither in male nor in femaleanimals. However, although protein levels point out to ageneral increase in DAGLα, two complementary events couldexplain this apparent contradiction: (1) the mRNA quantifica-tion represents a measurement from the whole hippocampusand thus regional specific decreases could be counteractingincreases in other regions, e.g., DG molecular layer decreasescould be counteracting increases in polymorphic cell layer ofDG and SP from CA3; and (2) in addition, a dilution effect couldbe masking a general tendency since the majority of thestudied sublayer displayed no changes, i.e., granular cell layerfrom DG, SO and SR from CA3 and SO, SP and SL-M from CA1showed no protein level changes. Interestingly, we detected adecrease in the MAGL mRNA and protein levels in thehippocampus of MD rats, being this effect more marked inmales than in females. We did not detect changes in mRNAexpression and in protein levels of DAGLβ. A regional analysis

Fig. 5 – Representative microphotographs of MAGLimmunoreactivity in relevant subregions (DG, CA1, and CA3)of 13-day-old rat hippocampus; effects of MD. The intenseMAGL immunoreactivity in the hippocampus was mainlyassociated with a dense network of fibers, possibly axonterminals, in the polymorphic cell layer of DG (A–D) and in thestrata oriens and radiatum of CA3 (E–H) and CA1 (I–L). Note adense, punctate, axon terminal-associated neuropil labelingforming pericellular baskets around the negative pyramidalcell somata (E and G). Note also the very low staining in themolecular layer of DG and stratum lacunosum-moleculare ofCA1 and CA3. CA, cornu ammonis; Co, control; DG, dentategyrus; gcl, granular cell layer; MD, maternal deprivation; ml,molecular layer; pcl, polymorphic cell layer; SL-M, stratumlacunosum-moleculare; SO, stratum oriens; SP, stratumpyramidale; SR, stratum radiatum. Scale bars are included inimages A, E, and I. A, E, I: male Co; B, F, J: male MD; C, G, K:female Co; D, H, L: female MD.

Fig. 6 – Quantification of DAGLα andMAGL immunoreactivity in relevant subregions (DG, CA1, and CA3) of the hippocampus in13-day-old rats; effects of MD. (A) The MD stress induced significant increases in DAGLα immunoreactivity, which was mainlyassociated with a denser network of fibers (possibly dendrites) in the polymorphic cell layer of DG, in the stratum pyramidale ofCA3, and in the stratum radiatum of CA1. Interestingly, the molecular layer of DG showed a significant decrease in DAGLαimmunoreactivity. Differences of DAGLα immunoreactivity in the remaining hippocampal layers and strata were notsignificant. (B) MD induced a significant decrease in the MAGL immunoreactivity in most strata of CA1 and CA3, and this effectwas more marked in males than in females. No differences in MAGL immunoreactivity were detected in the layers of DG. CA,cornu ammonis; Co, control; DG, dentate gyrus; gcl, granular cell layer; MD, maternal deprivation; ml, molecular layer; pcl,polymorphic cell layer; SL, stratum lucidum; SL-M, stratum lacunosum-moleculare; SO, stratum oriens; SP, stratumpyramidale; SR, stratum radiatum. Histograms represent the mean±SEM (seven animals per experimental group). Two-wayANOVA and Bonferroni post hoc test: *P<0.05; **P < 0.01; ns, no significance.

168 B R A I N R E S E A R C H 1 3 4 9 ( 2 0 1 0 ) 1 6 2 – 1 7 3

of DAGLα, DAGLβ and MAGL mRNA expression by in situhybridization could confirm the mRNA expression by RT–PCRand corroborate immunohistochemical data, so it could be aninteresting proximity to future studies. The general increase inthe enzyme of 2-AG synthesis together with the decrease of itsdegradation enzyme appear to fit well with the increase ofhippocampal 2-AG levels that we have previously found in MDanimals of the same age (Llorente et al., 2008).

Identifying the sites of biosynthesis and elimination of 2-AG and its associated enzymes may shed light on theirfunctional roles and help to understand their mechanisms ofregulation in the brain. This study provides an original (nopreviously reported) description of DAGLα and MAGL distri-bution within the hippocampus of neonatal rats (PND 13).Instead of the evident location of these enzymes in neuronalfibers, we cannot exclude their presence in glial processes. TheDAGLα staining pattern showed prominent DAGLα+dendriticterminals in the molecular layer of DG, in the strata lucidumand oriens of CA3, and in the stratum lacunosum-moleculareof CA. These data agreewith previous descriptions in the adultmice (Katona et al., 2006; Yoshida et al., 2006) which

demonstrated by high resolution images and electron micros-copy that DAGLα is mainly associated to the cell membrane ofpostsynaptic spines of the hippocampal pyramidal cells. So, atPND 13, the hippocampal DAGLα distribution might be similarto the adult pattern. The immunostaining pattern for MAGLsuggests that this enzyme, at this postnatal age, is associatedwith hippocampal presynaptic axon terminals as was previ-ously described in adult rat (Dihn et al., 2002; Gulyas et al.,2004). A dense, punctate, axon terminal-associated neuropillabeling is present in most hippocampal areas, except themolecular layer of the DG, the stratum lacunosum-moleculareof the CA3 and CA1, and the subiculum. Intensively stainedvaricosities can be found around pyramidal cell somata thatresemble inhibitory cell terminals forming pericellular bas-kets. Confirming these results, Gulyas et al. (2004) describedMAGL-containing inhibitory terminals on somata and axoninitial segments of pyramidal neurons that correlate with theCCK-containing perisomatic terminals of an inhibitory celltype. In addition, CB1 receptors were also present in the CCK-containing subset (Katona et al., 1999), suggesting that MAGLis present in the CCK/CB1-immunoreactive basket cell

Fig. 7 – Quantification of DAGLα (Α), DAGLβ (B) and MAGL (C)mRNA expression in the hippocampus of 13-day-old maleand female rats; effects of maternal deprivation (MD).β-ActinmRNA expressionwas used as constitutive gene. Histogramsrepresent the mean±SEM (seven animals per experimentalgroup). Student's t test: *P<0.05 versus control males.

169B R A I N R E S E A R C H 1 3 4 9 ( 2 0 1 0 ) 1 6 2 – 1 7 3

terminals. Axo-axonic (chandelier) cell terminals, whichlocalized in the stratum oriens and innervated the axon initialsegments of pyramidal neurons, were also immunoreactivefor MAGL (Gulyas et al., 2004). Most axon terminals in thestratum radiatum of CA1 and CA3 may constitute Schaffercollaterals, but also recurrent collaterals from CA3 pyramidalneurons andmossy fibers fromDG granular cells, respectively.We also observed a lack of neuropil staining for MAGL at thesubiculum confirming that Schaffer collaterals, but not CA1pyramidal cell axons, contain this enzyme. These resultsappear to accord with electron microscopy studies showingasymmetrical synapses on inhibitory cell dendrites in allhippocampal layers (Gulyas et al., 2004).

Considering the detailed distribution and the majorchanges of DAGLα and MAGL expression induced by MD, wepropose the following working model that lead us tohypothesize that the glutamatergic tone might be enhancedin the hippocampus of MD rats at the age of this study, i.e., atPND 13 (to more readily follow the description and interpre-tation detailed below see Fig. 8). The low expression of DAGLαin the molecular layer of DG, a region that directly receivesprojections from the pyramidal neurons of the entorhinalcortex, suggests a decrease of retrograde 2-AG inhibitorysignaling on glutamatergic transmission. In contrast, weobserved a high expression of DAGLα or a low expression of

MAGL in the stratumpyramidale of CA1 andCA3 respectively,a region that mainly contains perisomatic inhibitory term-inals from basket cells, suggesting an increase of retrograde2-AG inhibitory signaling and, therefore, a blockade ofGABAergic transmission to pyramidal neurons. In addition,we detected a high expression of DAGLα in the polymorphiccell layer of DG, and a low expression of MAGL in the stratumoriens of CA1 and CA3, whose regions mainly contain axo-axonic cell terminals that innervate the axon initial segmentsof pyramidal neurons, suggesting again an increase ofretrograde 2-AG inhibitory signaling on GABAergic transmis-sion. Finally, we also observed a low expression of MAGL inthe stratum radiatum of CA3 and CA1, whose fibers (mossyfibers and Schaffer collaterals respectively) constitute asym-metrical synapses on inhibitory cell dendrites, suggesting anincrease of retrograde 2-AG inhibitory signaling on glutama-tergic transmission to inhibitory interneurons. The presentworkingmodel points to the hypothesis of amultidirectional,functional relationship between glucocorticoids, endocanna-binoids, and glutamate in the hippocampus (Mailleux andVanderhaeghen, 1993, 1994; Di et al., 2009) that may beessential for a homeostatic control of behavior (Rodríguez deFonseca et al., 1997; Cota et al., 2007; Viveros et al., 2007) andthat may be altered by early life stress.

In conclusion, we provide a detailed description of theexpression of hippocampal DAGLα and MAGL in neonatal ratsof both sexes and the first evidence for the effects of MD onthese enzymes, suggesting that their functions may beaffected by early developmental stressful experiences. Ingeneral, the observed changes induced by MD in theseenzymes are compatible with the previously reported increaseof hippocampal 2-AG levels in MD rats of the same age. Thepresent MD procedure may provide relevant informationabout the specific role of 2-AG signaling in neurodevelopmentand the consequences of its disruption in relation to neuro-psychiatric diseases with a developmental origin.

4. Experimental procedures

4.1. Animals

We used Wistar albino rats of both sexes. Subjects were theoffspring of rats purchased fromHarlan Interfauna Ibérica S.A.(Barcelona, Spain) which were mated (onemale×two females)in our laboratory approximately 2 weeks after their arrival. Allanimals were maintained at a constant temperature (22±1 °C)and humidity (50±1%) in a reverse 12 h dark–light cycle (lightson at 20:00 h), with free access to food (commercial diet forrodents A04/A03; Panlab, Barcelona, Spain) and water. On theday of birth (PND 0), eight litters were sex-balanced, weightedand culled to four pups per litter (twomales and two females).So, we selected an animal per litter for each experimentalgroup, to get eight animals per group, to reduce litter effect.

The experiments performed in this studywere approved byour institutional committee and were in compliance with theRoyal Decree 1201/2005, October 21, 2005 (BOE no. 252), aboutthe protection of experimental animals, as well as with theEuropean Communities Council Directive of 24 November1986 (86/609/EEC).

Fig. 8 – Schematic representation of the hippocampal local circuits that resumes the changes of DAGLα and MAGL expressionlevels and hypothesizes the glutamatergic tone in relevant regions of the rat hippocampus at PND 13.

170 B R A I N R E S E A R C H 1 3 4 9 ( 2 0 1 0 ) 1 6 2 – 1 7 3

4.2. Maternal deprivation

MD took place on PND 9 according to Ellenbroek procedure(Ellenbroek et al., 1998). In summary, mothers from thedeprived group were removed early in themorning (beginningat 9:00 h) and pups were weighed and kept undisturbed intheir home cages (in the same room) for 24 h. On PND 10, pupswere weighed again and mothers returned to theircorresponding cage. In the control group, the mothers werebriefly removed on PND 9 and 10 to weigh the pups.

4.3. Tissue collection

Animals (n=28; seven per group) were sacrificed by quickdecapitation during the dark phase of the cycle (9:00 to 14:00 h)at PND 13. Brains were immediately isolated, snap frozen inliquid nitrogen, and stored at −80 °C until use for Western blotand quantitative real-time PCR analyses, or fixed with 4%paraformaldehyde (Merck) in 0.1 M phosphate buffer (PB), pH

Table 1 – Oligonucleotide primers used in this study.

Gene GenBank accessionnumber

AnnealingT (°C)

β-Actin NM031144.2 51,4 5´CADAGLα NM001005886.1 60.5 5′GGDAGLβ NM001107120.1 58.9 5′ TGMAGL NM138502.1 58.1 5′CA

7.3, during 48 h at 4 °C, washed three times in PB 0.1 M, pH 7.3,and stored at 4 °C in PB with 0.002% (wt./vol.) of sodium azideuntil used for immunostaining. The latter brains were thenequilibrated in PBS containing 30% sucrose and 0.01% (wt./vol.)of sodium azide during 24 h, and cut into 30-μm-thicktransverse sections using a sliding microtome. Sections werecollected from −4.16 to −4.80 bregma levels to perform eachantibody and Nissl staining (Paxinos and Watson, 1998).

4.4. Quantitative real-time PCR

Quantitative real-time PCR was used to measure DAGLα,DAGLβ and MAGL mRNA levels. Briefly, total RNA wasextracted from the hippocampus of each animal by using theTrizol method, according the manufacturer's instructions(Gibco BRL Life Technologies, Baltimore, MD, USA). All RNAsamples had A260/280 ratios of 1.8 to 2.0 after purification withRNeasy micro- or minelute cleanup kit (Quiagen, Hilden,Germany). Purified RNA from each sample and random

Oligonucleotide primers

Forward Reverse

GGCTGTGTTGTCCCTGTA 5´GCTGTGGTGGTCAAGCTGTAGTACCTAATGGCTGCTCA 5´AGGACTGACCATCCAACCTGATAGGCCCAAAGATGCTG 5′AAGTCCATTGTGCTCGTCAGTGGAGCTGGGGAACACTG 5′GGAGATGGCACCGCCATGGAG

171B R A I N R E S E A R C H 1 3 4 9 ( 2 0 1 0 ) 1 6 2 – 1 7 3

hexamers were used to generate first strand cDNA usingtranscriptor reverse transcriptase (Roche Applied Science,Indianapolis, IN, USA). Negative controls included reversetranscription reactions omitting reverse transcriptase. Theobtained cDNAwas used as the template for quantitative real-time PCR, which was performed in an iCycler system (Bio-Rad,Hercules, CA, USA) using the SYBR Green I detection format(FastStart Universal Master kit; Roche). Each reaction was runin duplicate and contained 2.5 μl of cDNA template, 5 mMCl2Mg and 0.4–0.5 μM of primers in a final reaction volume of15 μl. Cycling parameters were 95 °C for 5 min to activate DNApolymerase, then 35–45 cycles of 95 °C for 10 s, annealingtemperature for 15 s (see Table 1) and a final extension step of72 °C for 15 s in which fluorescence was acquired. Meltingcurves analysis were performed to ensure only a singleproduct was amplified. Gene expression was normalized tothe expression of the housekeeping gene actin beta. Primersfor PCR reaction (Table 1) were designed based on NCBIdatabase sequences of rat reference mRNA and checked forspecificity with BLAST software from NCBI Webs ite (http://www.blast.ncbi.nlm.nih.gov/Blast.cgi). Quantification wascarried out with a standard curve run at the same time asthe samples with each reaction run in duplicate. Absolutevalues from each sample were normalized with regard to thehousekeeping gene actin beta.

4.5. Western blot analysis

e performed Western blot analyses to demonstrate thatDAGLα, DAGLβ, and MAGL antibodies recognized thecorresponding antigen in the hippocampus of Wistar malerats (Fig. 2). Rat brain samples were homogenized in HEPES50 mM (pH 8)–sucrose 0.32 M buffer to obtain membraneprotein extracts. The homogenate was centrifuged at 800×gfor 10 min at 4 °C and the supernatant centrifuged at 40,000×gfor 30 min. The pellets were resuspended in HEPES 50 mMbuffer and potterized using a homogenizer. For inmunoblot-ting, equivalents amounts of membrane proteins (40 μg) fromrat brain were separated by 10% (wt./vol.) sodium dodecylsulfate–polyacrylamide gel electrophoresis (SDS–PAGE), trans-ferred on nitrocellulose membranes (Bio-Rad), and controlledby Ponceau red staining. We preincubated blots with ablocking buffer containing PBS, 0.1% (vol./vol.) Tween 20, and2% (wt./vol.) albumin fraction V from bovine serum (Merck) atroom temperature for 1 h. For protein detection, each blottedmembrane lane was incubated separately with the specificDAGLα (1:100), DAGLβ (1:100) and MAGL (1:200) antibodies anddiluted in PBS containing 0.1% Tween 20 and 2% albuminfraction V from bovine serum at room temperature overnight.After extensive washing in PBS containing 1% Tween 20 (PBS-T), a peroxidase-conjugated goat anti-rabbit antibody (Pro-mega, Madison, WI) was added (1:2500) for 1 h at roomtemperature. Biotinylated marker proteins with definedmolecular weights were used for molecular weight determi-nation in Western blots (ECL Western Blotting MolecularWeight Markers ; Amersham). We incubated thecorresponding markers lane with ExtrAvidin peroxidase(Sigma). Membranes were then subjected to repeated washingin PBS-T, and the specific protein bands were visualized byusing the enhanced chemiluminescence technique (ECL;

Amersham) and Auto-Biochemi Imaging System (LTF Labor-technik GmbH, Wasserburg/Bodensee, Germany). Westernblots showed that each primary antibody detected a proteinof the expected molecular size.

4.6. Immunohistochemistry

One immunostaining batch containing 28 slides (four groups,seven animals each) was stained simultaneously to avoidvariations in the intensity of staining due to the procedure.Two or three different batches were run for each antibody andfor each animal. For each staining batch, free-floating sectionswere first washed several times with PBS (0,1 M; pH 7,3) toremove sodium azide, then incubated in sodium citrate(50 mM in dH2O, pH 6) for 30 min at 80 °C, followed by threewashes in PBS for 10 min each time, and incubated in 3%hydrogen peroxide in PBS for 20 min in darkness at roomtemperature to inactivate endogenous peroxidase. Then,sections were incubated in blocking solution containing 10%donkey serum, 0.3% Triton X-100, and 0.1% sodium azide inPBS for 1 h.

In this experiment, we have evaluated whether maternaldeprivation alters the enzymes involved in the route ofsynthesis–degradation of the endocannabinoid 2-AG, such asdiacylglycerol lipase α (DAGLα), diacylglycerol lipase β (DAGLβ),and monoacylglycerol lipase (MAGL). Polyclonal antibodiesDAGLα and DAGLβ were generated in our laboratory byimmunizing rabbits following the process outlined in Suárezet al. (2008), and rabbit polyclonal antibody MAGL waspurchased commercially (Cayman Chemical, Ann Harbor, MI;cat. no. 100035; lot. no. 202410). For each experiment, weincluded slides containing adult rat hippocampus (PND 70;250 g) to be compared with those corresponding to PND 13 (Fig.3). Sectionwere incubated in thediluted primary antibody (anti-DAGLα, diluted1:250; anti-DAGLβ, diluted1:200;andanti-MAGL,diluted 1:100) overnight at roomtemperature. The followingdaysections were washed three times with PBS, incubated in abiotinylated donkey anti-rabbit immunoglobulin (Amersham,Little Chalfont, England) diluted 1:500 for 1 h, washed again inPBS, and incubated in ExtrAvidin peroxidase (Sigma, St. Louis,MO) diluted 1:2,000 in darkness at room temperature for 1 h.After three washes in PBS in darkness, we revealed immuno-labeling with 0.05% diaminobenzidine (DAB; Sigma), 0.05%nickel ammonium sulfate, and 0.03% H2O2 in PBS. All stepswere carried out by gentle agitation at room temperature. Afterseveral washes in PBS, sections were mounted in slides treatedwith poly-L-lysine solution (Sigma), air-dried, dehydrated inethanol, cleared with xylene, and coverslipped with Eukittmounting medium (Kindler GmBH & Co, Freiburg, Germany).

4.7. Quantification of immunostaining on hippocampal slices

Digital high-resolutionmicrophotographs of the hippocampuswere taken with a 10× objective under the same conditions oflight and brightness/contrast by an Olympus BX41microscopeequipped with an Olympus DP70 digital camera. Densitomet-ric quantification of the immunoreactivity of selected areaswas made using the analysis software ImageJ 1.38X (NIH,USA). On each tissue section, we focused on CA1 and CA3areas of Ammon's horn and dentate gyrus (DG). For both CA

172 B R A I N R E S E A R C H 1 3 4 9 ( 2 0 1 0 ) 1 6 2 – 1 7 3

subfields, we carried out separate densitometrical analysiscorresponding to stratum oriens (SO), stratum pyramidale(SP), stratum radiatum (SR), and stratum lacunosum-molecu-lare. For dentate gyrus, we carried out separate densitometricanalysis corresponding to the molecular layer (ml), thegranular cell layer (gcl), and the polymorphic cell layer (pl).Representative digital images were mounted and labeledusing Adobe PageMaker (San Jose, CA).

4.8. Statistical analysis

Immunohistochemical data were analyzed by three-wayanalysis of variance (ANOVA), with the three factors beinghippocampal areas (DG, CA3, and CA1) and subregions, sex(males and females), and MD (non deprived and deprivedanimals), followed by Bonferroni test for post hoc compar-isons. RT–PCR data were analyzed by two-way ANOVA withthe two factors being sex and MD. Results from RT–PCR wereindeed evaluated by Student's t test (two-tailed, pairedgroups). The level of significance was set at P<0.05 in all cases.

Acknowledgments

Grant sponsors: Consejería de Salud (Junta de Andalucía), PI-0220/2008, PI-0232/2008; Consejería de Innovación, Ciencia yEmpresa (Junta de Andalucía), P05-CV1-1038; Ministerio deEducación y Ciencia, SAF2006-07523; Ministerio de Sanidad yConsumo, FIS 07/1226 and FIS 07/0880; Red de TrastornosAdictivos, RD06/0001/1013. Ricardo Llorente is a predoctoralfellow of the Universidad Complutense of Madrid. Francisco J.Bermúdez-Silva is recipient of a research contract from theNational System of Health (Instituto de Salud Carlos III; CP07/00283). Juan Suárez is recipient of a postdoctoral contract fromthe National System of Health (Instituto de Salud Carlos III,grant number CD08/00203).The authors deeply thank to Dr.Elke Binder for the statistical support.

Appendix A. Supplementary data

Supplementary data associated with this article can be found,in the online version, at doi:10.1016/j.brainres.2010.06.042.

R E F E R E N C E S

Berghuis, P., Rajnicek, A.M., Morozov, Y.M., Ross, R.A., Mulder, J.,Urban, G.M., Monory, K., Marsicano, G., Matteoli, M., Canty, A.,et al., 2007. Hardwiring the brain: endocannabinoids shapeneuronal connectivity. Science 316, 1212–1216.

Bernard, C., Milh, M., Morozov, Y.M., Ben Ari, Y., Freund, T.F.,Gozlan, H., 2005. Altering cannabinoid signaling duringdevelopment disrupts neuronal activity. Proc. Natl. Acad. Sci.U.S.A. 102, 9388–9393.

Bisogno, T., Howell, F., Williams, G., Minassi, A., Cascio, M.G.,Ligresti, A., Matias, I., Schiano-Moriello, A., Paul, P.,Williams, E.J.,Gangadharan, U., Hobbs, C., Di Marzo, V., Doherty, P., 2003.Cloning of the first sn1-DAG lipases points to the spatial andtemporal regulation of endocannabinoid signaling in the brain. J.Cell. Biol. 163, 463–468.

Blackman, J.L., Simon, G.M., Cravatt, B.F., 2007. A comprehensiveprofile of brain enzymes that hydrolyze the endocannabinoid2-arachidonoylglycerol. Chem. Biol. 14, 1347–1356.

Brittis, P.A., Silver, J., Walsh, F.S., Doherty, P., 1996. Fibroblastgrowth factor receptor function is required for the orderlyprojection of ganglion cell axons in the developingmammalianretina. Mol. Cell. Neurosci 8, 120–128.

Cota, D., Steiner, M.A., Marsicano, G., Cervino, C., Herman, J.P.,Grübler, Y., Stalla, J., Pasquali, R., Lutz, B., Stalla, G.K., Pagotto,U., 2007. Requirement of cannabinoid receptor type 1 for thebasal modulation of hypothalamic–pituitary–adrenal axisfunction. Endocrinology. 148, 1574–1581.

De la Fuente, M., Llorente, R., Baeza, I., De Castro, N.M., Arranz, L.,Cruces, J., Viveros, M.P., 2009. Early maternal deprivation inrats: a proposed animal model for the study of developmentalneuroimmunoendocrine interactions. Ann. N. Y. Acad. Sci.1153, 176–183.

Dean, B., Sundram, S., Bradbury, R., Scarr, E., Copolov, D., 2001.Studies on [3H] CP-55940 binding in the human central nervoussystem: regional specific changes in density of cannabinoid-1receptors associated with schizophrenia and cannabis use.Neuroscience. 103, 9–15.

Devane, W.A., Hanus, L., Breuer, A., Pertwee, R.G., Stevenson, L.A., Griffin, G., Gibson, D., Mandelbaum, A., Etinger, A.,Mechoulam, R., 1992. Isolation and structure of a brainconstituent that binds to the cannabinoid receptor. Science.258, 1946–1949.

Di, S., Maxson, M.M., Franco, A., Tasker, J.G., 2009. Glucocorticoidsregulate glutamate and GABA synapse-specific retrogradetransmission via divergent nongenomic signaling pathways. J.Neurosci. 29, 393–401.

Dihn, T.P., Carpenter, D., Leslie, F.M., Freund, T.F., Katona, I., Sensi,L., Kathuria, S., Piomelli, D., 2002. Brain monoglyceride lipaseparticipating in endocannabinoid inactivation. Proc. Natl.Acad. Sci. U.S.A. 99, 10819–10824.

Di Forti, M., Morrison, P.D., Butt, A., Murray, R.M., 2007. Cannabisuse and psychiatric and cognitive disorders: the chicken or theegg? Curr. Opin. Psychiatry. 20, 228–234.

Di Marzo, V., Hill, M.P., Bisogno, T., Crossman, A.R., Brotchie, J.M.,2000. Enhanced levels of endogenous cannabinoids in theglobus pallidus are associated with reduction in movement inan animal model of Parkinson's disease. FASEB. J. 14,1432–1438.

Ellenbroek, B.A., de Bruin, N.M., van Den Kroonenburg, P.T., vanLuijtelaar, E.L., Cools, A.R., 2004. The effects of early maternaldeprivation on auditory information processing in adultWistarrats. Biol. Psychiatry. 55, 701–707.

Ellenbroek, B.A., Riva, M.A., 2003. Early maternal deprivation as ananimal model for schizophrenia. Clin. Neurosci. Res. 3, 297–302.

Ellenbroek, B.A., van den Kroonenberg, P.T., Cools, A.R., 1998. Theeffects of an early stressful life event on sensorimotor gating inadult rats. Schizophr. Res. 30, 251–260.

Fernández-Ruiz, J.J., Berrendero, F., Hernández, M.L., Ramos, J.A.,2000. The endogenous cannabinoid system and braindevelopment. Trends. Neurosci. 23, 14–20.

Giuffrida, A., Leweke, F.M., Gerth, C.W., Schreiber, D., Koethe, D.,Faulhaber, J., Klosterkotter, J., Piomelli, D., 2004. Cerebrospinalanandamide levels are elevated in acute schizophrenia and areinversely correlated with psychotic symptomsNeuropsychopharmacology. 29, 2108–2114.

Gould, E., 1994. The effects of adrenal steroids and excitatory inputon neuronal birth and survival. Ann. N. Y. Acad. Sci. 743, 73–92.

Gulyas, A.I., Cravatt, B.F., Bracey,M.H., Dihn, T.P., Piomelli, D., Boscia,F., Freund, T.F., 2004. Segregation of twoendocannabinoid-hydrolyzing enzymes into pre- andpostsynaptic compartments in the rat hippocampus, cerebellumand amygdale. Eur. J. Neurosci. 20, 441–458.

Harkany, T., Keimpema, E., Barabás, K., Mulder, J., 2008.Endocannabinoid functions controlling neuronal specification

173B R A I N R E S E A R C H 1 3 4 9 ( 2 0 1 0 ) 1 6 2 – 1 7 3

during brain development. Mol. Cell. Endocrinol. 286 (1–2Suppl 1), S84–90.

Hashimotodani, Y., Ohno-Shosaku, T., Kano, M., 2007. Presynapticmonoacylglycerol lipase activity determines basalendocannabinoid tone and terminates retrogradeendocannabinoid signaling in the hippocampus. J. Neurosci. 27,1211–1219.

Hill, M.N., Miller, G.E., Carrier, E.J., Gorzalka, B.B., Hillard, C.J., 2009.Circulating endocannabinoids and N-acyl ethanolamines aredifferentially regulated in major depression and followingexposure to social stress. Psychoneuroendocrinology. 34 (8),1257–1262.

Katona, I., Sperlagh, B., Sik, A., Kafalvi, A., Vizi, E.S., Mackie, K.,Freund, T.F., 1999. Presynaptically located CB1 cannabinoidreceptors regulate GABA release from axon terminals ofspecific hippocampal interneurons. J. Neurosci. 19, 4544–4558.

Katona, I., Urban, G.M., Wallace, M., Ledent, C., Jung, K.M., Piomelli,D., Mackie, K., Freund, T.F., 2006. Molecular composition of theendocannabinoid systemat glutamatergic synapses. J. Neurosci.26, 5628–5637.

Leweke, F.M., Giuffrida, A., Koethe, D., Schreiber, D., Nolden, B.M.,Kranaster, L., Neatby, M.A., Schneider, M., Gerth, C.W.,Hellmich, M., Klosterkotter, J., Piomelli, D., 2007. Anandamidelevels in cerebrospinal fluid of first-episode schizophrenicpatients: impact of cannabis use. Schizophr. Res. 94, 29–36.

Leweke, F.M., Giuffrida, A., Wurster, U., Emrich, H.M., Piomelli, D.,1999. Elevated endogenous cannabinoids in schizophrenia.Neuroreport. 10, 1665–1669.

Leweke, F.M., Koethe, D., 2008. Cannabis and psychiatric disorders:it is not only addiction. Addict. Biol. 13, 264–275.

Lewis, D.A., Levitt, P., 2002. Schizophrenia as a disorder ofneurodevelopment. Annu. Rev. Neurosci. 25, 409–432.

Llorente, R., Llorente-Berzal, A., Petrosino, S., Marco, E.M., Guaza,C., Prada, C., López-Gallardo, M., Di Marzo, V., Viveros, M.P.,2008. Gender dependent cellular and biochemical effects ofmaternal deprivation on the hippocampus of neonatal rats; apossible role for the endocannabinoid system. Dev. Neurobiol.68, 1334–1347.

Llorente, R., López-Gallardo, M., Llorente-Berzal, A., Prada, C.,Garcia-Segura, L.M., Viveros, M.P., 2009. Early maternaldeprivation in rats induces gender-dependent effects ondeveloping hippocampal and cerebellar cells. Int. J. Dev.Neurosci. 27, 233–241.

Mailleux, P., Vanderhaeghen, J.J., 1993. Glucocorticoid regulationof cannabinoid receptor messenger RNA levels in the ratcaudate-putamen. An in situ hybridization study. Neurosci.Lett. 156, 51–53.

Mailleux, P., Vanderhaeghen, J.J., 1994. Glutamatergic regulation ofcannabinoid receptor gene expression in the caudate-putamen.Eur. J. Pharmacol. 266, 193–196.

Malcher-Lopes, R., Franco, A., Tasker, J.G., 2008. Glucocorticoidsshift arachidonic acid metabolism toward endocannabinoidsynthesis: a non-genomic anti-inflammatory switch. Eur. J.Pharmacol. 583, 322–339.

Marco, E.M., Adriani, W., Llorente, R., Laviola, G., Viveros, M.P., 2009.Detrimental psychophysiological effects of early maternaldeprivation inadolescent andadult rodents: altered responses tocannabinoid exposure. Neurosci. Biobehav. Rev. 33, 498–507.

Marek, G., Merchant, K., 2005. Developing therapeutics forschizophrenia and other psychotic disorders. NeuroRx. 2, 579–589.

Mechoulam, R., Ben-Shabat, S., Hanus, L., Ligumsky, M., Kaminski,N.E., Schatz, A.R., Gopher, A., Almog, S., Martin, B.R., Compton,D.R., Pertwee, R.G., Griffin, G., Bayerwitch, M., Barg, J., Vogel, Z.,1995. Identification of an endogenous 2-monoglyceride, presentin canine gut, that binds to cannabinoid receptors. Biochem.Pharmacol. 50, 83–90.

Panikashvili, D., Simeonidou,C., Ben-Shabat, S., Hanus, L., Breuer, R.,Mechoulam, R., Shohami, E., 2001. An endogenous cannabinoid

(2-AG) is neuroprotective after brain injury. Nature. 413,527–531.

Paxinos, G., Watson, C., 1998. The Rat Brain in StereotaxicCoordinates24th ed. Academic Press, London, UK.

Pickering, C., Gustafsson, L., Cebere, A., Nylander, I., Liljequist, S.,2006. Repeated maternal separation of male Wistar rats altersglutamate receptor expression in the hippocampus but not theprefrontal cortex. Brain. Res. 1099, 101–108.

Roceri, M., Hendriks, W., Racagni, G., Ellenbroek, B.A., Riva, M.A.,2002. Early maternal deprivation reduces the expression ofBDNF and NMDA receptor subunits in rat hippocampus. Mol.Psychiatry. 7, 609–616.

Rodríguez de Fonseca, F., Carrera, M.R., Navarro, M., Koob, G.F.,Weiss, F., 1997. Activation of corticotropin-releasing factor inthe limbic system during cannabinoid withdrawal. Science.276, 2050–2054.

Sapolsky, R.M., Packan, D.R., Vale, W.W., 1988. Glucocorticoidtoxicity in the hippocampus: in vitro demonstration. Brain.Res. 453, 367–371.

Sapolsky, R.M., 2000. The possibility of neurotoxicity in thehippocampus in major depression: a primer on neuron death.Biol. Psychiatry. 48, 755–765.

Suárez, J., Bermúdez-Silva, F.J., Mackie, K., Ledent, C., Zimmer, A.,Cravatt, B.F., de Fonseca, F.R., 2008. Immunohistochemicaldescription of the endogenous cannabinoid system in the ratcerebellum and functionally related nuclei. J. Comp. Neurol.509, 400–421.

Suárez, J., Llorente,R., Romero-Zerbo, S.Y.,Mateos, B., Bermúdez-Silva,F.J., Rodríguez de Fonseca, F., Viveros, M.P., 2009. Early maternaldeprivation induces gender-dependent changeson the expressionof hippocampal CB1 and CB2 cannabinoid receptors of neonatalrats. Hippocampus. 19, 623–632.

Sugiura, T., Kondo, S., Sukagawa, A., Nakane, S., Shinoda, A., Itoh,K., Yamashita, A., Waku, K., 1995. 2-Arachidonoylglycerol: apossible endogenous cannabinoid receptor ligand in brain.Biochem. Biophys. Res. Commum. 215, 89–97.

Viveros, M.P., Marco, E.M., Llorente, R., López-Gallardo, M., 2007.Endocannabinoid system and synaptic plasticity: implicationsfor emotional responses. Neural. Plast. 2007, 52908.

Viveros, M.P., Llorente, R., López-Gallardo, M., Suárez, J.,Bermúdez-Silva, F., De la Fuente, M., Rodríguez de Fonseca,F., García-Segura, L.M., 2009. Sex-dependent alterations inresponse to maternal deprivation in rats.Psychoneuroendocrinology 34 (Suppl 1), S217–26.

Viveros, M.P., Díaz, F., Mateos, B., Rodríguez, N., Chowen, J.A., 2010.Maternal deprivation induces a rapiddecline in circulating leptinlevels and sexually dimorphic modifications in hypothalamictrophic factors and cell turnover. Horm Behav. 57, 405-14.

Weinberger, D.R., 1987. Implications of normal brain developmentfor the pathogenesis of schizophrenia. Arch. Gen. Psychiatry.44, 660–669.

Williams, E.J., Walsh, F.S., Doherty, P., 2003. The FGF receptor usesthe endocannabinoid signaling system to couple to an axonalgrowth response. J. Cell. Biol. 160, 481–486.

Yoshida, T., Fukaya, M., Uchigashima, M., Miura, E., Kamiya, H., Kano,M., Watanabe, M., 2006. Localization of diacylglycerol lipase-αaround postsynaptic spine suggests close proximity betweenproduction site of an endocannabinoid, 2-arachidonoyl-glycerol,and presynaptic cannabinoid CB1 receptor. J. Neurosci. 26,4740–4751.

Zavitsanou, K., Garrick, T., Huang, X.F., 2004. Selectiveantagonist [3H] SR141716A binding to cannabinoid CB1receptors is increased in the anterior cingulate cortex inschizophrenia. Prog. Neuropsychopharmacol. Biol.Psychiatry. 28, 355–360.

Zimmerberg, B., Shartrand, A.M., 1992. Temperature-dependenteffects of maternal separation on growth, activity, andamphetamine sensitivity in the rat. Dev. Psychobiol. 25, 213–226.