Neurochemistry International 60 (2012) 1–8

Contents lists available at SciVerse ScienceDirect

Neurochemistry International

journal homepage: www.elsevier .com/locate /nci

CB1 receptor activation inhibits neuronal and astrocytic intermediary metabolismin the rat hippocampus

João M.N. Duarte a,1, Samira G. Ferreira a, Rui A. Carvalho a,b, Rodrigo A. Cunha a,c, Attila Köfalvi a,⇑a Center for Neurosciences and Cell Biology of Coimbra, University of Coimbra, Coimbra, Portugalb Department of Life Sciences, Faculty of Sciences and Technology, University of Coimbra, Coimbra, Portugalc Institute of Biochemistry, Faculty of Medicine, University of Coimbra, Coimbra, Portugal

a r t i c l e i n f o

Article history:Received 17 September 2011Received in revised form 21 October 2011Accepted 31 October 2011Available online 9 November 2011

Keywords:CB1 cannabinoid receptorIntermediary metabolismHippocampusAstrocyteNeuronNuclear magnetic resonance

0197-0186/$ - see front matter � 2011 Elsevier Ltd. Adoi:10.1016/j.neuint.2011.10.019

Abbreviations: 2-AG, 2-arachidonoyl-glycerol; 4cannabinoid receptor type 1; D9-THC, D9-tetrahydrocsulfoxide; GABA, c-aminobutyrate; NMR, nucleartricarboxylic acid; WIN, WIN55212-2.⇑ Corresponding author. Address: Center for Neuro

Coimbra, Faculty of Medicine, University of Coimbra,Tel.: +351 96 8802309; fax: +351 239 822776.

E-mail address: akofalvi@yahoo.com (A. Köfalvi).1 Present address: Center for Biomedical Imaging, La

a b s t r a c t

Cannabinoid CB1 receptor (CB1R) activation decreases synaptic GABAergic and glutamatergic transmis-sion and it also controls peripheral metabolism. Here we aimed at testing with 13C NMR isotopomer anal-ysis whether CB1Rs could have a local metabolic role in brain areas having high CB1R density, such as thehippocampus. We labelled hippocampal slices with the tracers [2-13C]acetate, which is oxidized in glialcells, and [U–13C]glucose, which is metabolized both in glia and neurons, to evaluate metabolic compart-mentation between glia and neurons. The synthetic CB1R agonist WIN55212-2 (1 lM) significantlydecreased the metabolism of both [2-13C]acetate (�11.6 ± 2.0%) and [U–13C]glucose (�11.2 ± 3.4%) inthe tricarboxylic acid cycle that contributes to the glutamate pool. WIN55212-2 also significantlydecreased the metabolism of [U–13C]glucose (�11.7 ± 4.0%) but not that of [2-13C]acetate contributingto the pool of GABA. These effects of WIN55212-2 were prevented by the CB1R antagonist AM251(500 nM). These results thus suggest that CB1Rs might be present also in hippocampal astrocytes besidestheir well-known neuronal localization. Indeed, confocal microscopy analysis revealed the presence ofspecific CB1R immunoreactivity in astrocytes and pericytes throughout the hippocampus.

In conclusion, CB1Rs are able to control hippocampal intermediary metabolism in both neuronal andglial compartments, which suggests new alternative mechanisms by which CB1Rs control cell physiologyand afford neuroprotection.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Cannabinoid CB1 receptors (CB1Rs) are metabotropic receptorswith a high presynaptic density in the hippocampus (Katonaet al., 1999), where they can control the release of various neuro-transmitters (Katona et al., 1999; Balázsa et al., 2003; Freundet al., 2003; Kawamura et al., 2006; Mulder et al., 2011). CB1Rsare endogenously activated by the endocannabinoid molecules,anandamide and 2-arachidonoyl-glycerol (2-AG), which arereleased in an activity-dependent manner from the post-synapticmembrane, serving as retrograde signalling molecules (Freundet al., 2003). Both central and peripheral CB1Rs also play an

ll rights reserved.

AP, 4-aminopyridine; CB1R,annabinol; DMSO, dimethyl-

magnetic resonance; TCA,

sciences and Cell Biology of3004-517 Coimbra, Portugal.

usanne, Switzerland.

important role in systemic glucose homeostasis. Centrally (e.g. inthe hypothalamus, the limbic system and sympathetic neurons),CB1R activation intensifies food seeking behavior and hunger, andincreases anabolic hormonal tone resulting in a decreased cellularglucose oxidation (Matias et al., 2008; Quarta et al., 2010; Di Marzoet al., 2011). Peripherally, CB1Rs are bona fide regulators of insulinand glucagon secretion at the pancreatic islets of Langerhans (Liet al., 2011), and by being linked to the insulin signalling pathwaycan influence insulin sensitivity and glucose uptake (Pagano et al.,2007; Song et al., 2011), or glucose metabolism, e.g. via the AMP-activated protein kinase (Hardie, 2008; Matias et al., 2008). How-ever, in contrast to this role of CB1Rs in systemic glucoregulation,little is known about such a possible function of CB1R as metaboliccontrollers of brain metabolism.

We previously observed that hippocampal CB1R density is in-creased in the streptozotocin model of type-1 diabetes (Duarteet al., 2007b), suggesting a possible association of CB1Rs with hip-pocampal glucoregulation. Whether these latter CB1Rs are neuro-nal is yet to be determined. Indeed, there is some evidence forastrocytic expression of CB1Rs: astrocytic CB1Rs can induce synap-tic potentiation in the hippocampus (Navarrete and Araque, 2010),

2 J.M.N. Duarte et al. / Neurochemistry International 60 (2012) 1–8

and a possible glucoregulation by astrocytic CB1Rs has also beenproposed (Sánchez et al., 1998; Stella, 2010). Hypothetically, endo-cannabinoids, being a measure of neuronal activity, could signal toastrocytes the need for more energy. Increased neuronal activitydemands more energy, which is supported principally by astrocyticglucose metabolism (Öz et al., 2004; Magistretti, 2006; Pellerin,2008).

Finally, glucose facilitates the overall performance in variousmemory tasks (Messier, 2004). Thus, if CB1Rs inhibit hippocampalglucose metabolism they could contribute to impaired memoryconsolidation and retrieval in the healthy hippocampus (Hampsonand Deadwyler, 1999) and in Alzheimer’s disorder where we re-cently found that the levels of 2-AG were elevated due to its im-paired clearance (Mulder et al., 2011). Additionally, Suzuki et al.(2011) reported that lactate is necessary for long-term potentia-tion, which is a prerequisite of memory formation. Thus, if CB1Ractivation alters lactate levels it could alter memory formation thisway, besides controlling the bioavailability of GABA and glutamatederived from the Szent-Györgyi – Krebs (tricraboxylic acid, TCA)cycle.

The incorporation of 13C atoms from tracers into intermediarymetabolites can be followed by NMR spectroscopy and has beenwidely used to monitor brain metabolism both in vitro andin vivo (Cerdán et al., 1990; Duarte et al., 2007a, 2011). 13C-glucoseis widely used as a non-selective tracer to monitor glial and neuro-nal metabolism, while 13C-acetate can also be used as a selectiveglial tracer (Cerdán et al., 1990; Waniewski and Martin, 1998).Thus, we now investigated the effect of selective CB1R activationon hippocampal intermediary metabolism, using 13C NMR isotopo-mer analysis of hippocampal slices superfused with the 13C en-riched substrates [2-13C]acetate and [U–13C]glucose.

2. Materials and methods

2.1. Reagents

[U–13C]glucose (99%), sodium [2-13C]acetate (99%) and 2H2O(99.9%) were purchased from Isotec (Miamisburg, USA). 4-Amino-pyridine (4AP), solutions of 20% (w/w) 2HCl and 40% (w/w) NaO2H,and other common reagents (highest purity available) were ob-tained from Sigma–Aldrich (Sintra, Portugal). The synthetic can-nabinoid receptor agonist, WIN55212-2 (WIN) and the CB1Rantagonist, AM251 were obtained from Tocris (Bristol, UK). WINand AM251stock solutions were prepared in dimethylsulfoxide(DMSO) and the same volume of DMSO was included in the respec-tive paired control experiments.

2.2. Preparation and superfusion of hippocampal slices

All studies were conducted in accordance with the principlesand procedures outlined in the EU guidelines (86/609/EEC) andby FELASA, with particular care to minimize animal suffering andthe number of animals used. Male Wistar rats (8 weeks old;Charles-Rivers, Barcelona) were housed in an SPF facility, with12 h light on/off cycles and ad libitum access to food and water un-til use, when they were anaesthetized with halothane and decapi-tated. The brain was rapidly removed and the isolated hippocampiwere transversely cut in 400 lm slices using a McIlwain tissuechopper. The hippocampal slices were allowed to recover in a mod-ified Krebs solution (composition in mM: 115 NaCl, 25 NaHCO3, 3KCl, 1.2 KH2PO4, 2 CaCl2, 1.2 MgSO4, 5.5 glucose and 5.5 sodiumacetate, previously and continuously gassed with carbogen, pH7.4), at room temperature, during at least 45 min. Hippocampalslices were then transferred to a submerged chamber and super-fused (3 mL/min) with the same gassed solution at 37 �C, during

an initial 60-min period of stabilization. The hippocampal sliceswere then superfused in the presence of the 13C tracers [U–13C]glu-cose and sodium [2-13C]acetate (equimolar concentration of5.5 mM), and 50 lM 4AP to allow intermittent burst-like stimula-tion (Tibbs et al., 1989). A superfusion period of 3 h was necessaryto reach isotopic steady-state in these experimental conditions(Duarte et al., 2007a), and WIN (1 lM) and/or AM251 (500 nM)were included in the superfusion medium during this period.

2.3. Metabolite extraction and NMR spectroscopy

After each superfusion protocol, the hippocampal slices wereimmediately transferred to liquid nitrogen and water-solublemetabolites were extracted with 7% (v/v) perchloric acid (PCA).Briefly, the grinded tissue was mixed with PCA and centrifugedat 21,000g during 15 min, at 4 �C. The supernatant was neutralizedwith KOH and lyophilised. The pellet was re-dissolved in a solutioncontaining 6 M urea and 5% sodium dodecyl-sulphate, and theamount of protein was determined by the bicinchoninic acid meth-od (Smith et al., 1985) using a kit from Pierce (Rockford, USA).

The lyophilised extracts were re-dissolved in 600 lL 2H2O andthe p2H was re-adjusted to 7.0 with 2HCl or NaO2H solutions. So-dium fumarate (2 mM) was added as internal standard for quanti-fication of absolute concentrations by 1H NMR spectroscopy. NMRspectra were acquired at 25 �C on an 11.7 T Varian Unity-500 spec-trometer (Varian, Palo Alto, USA), using a 5 mm broadband probe.Solvent-suppressed 1H NMR spectra were acquired with 8 kHzsweep width, using 3 s water pre-saturation, 70� pulse angle, 3 sacquisition time, 14 s delay to allow total proton relaxation, andat least 200 scans per tissue extract. 1H-decoupled 13C NMR spectrawere acquired using 30 kHz sweep width, 45� pulse angle, 1.5 sacquisition time, 1.5 s relaxation delay and a Waltz proton decou-pling sequence. The total repetition time allowed full relaxation ofthe aliphatic carbons. To achieve adequate signal to noise ratio, thenumber of scans recorded was at least 10,000. The acquired spectrawere processed using the PC-based NUTS™ software (Acorn NMR,Fremont, USA). Free induction decays were baseline corrected,zero-filled and multiplied by a 0.5 Hz exponential function, priorto Fourier transformation. Areas of both 1H and 13C NMR spectrawere quantified by curve fitting.

The concentrations of metabolites were determined by measur-ing the peak areas in the 1H NMR spectra, using sodium fumarateas internal standard, and normalized to the protein content. Frac-tional enrichments for lactate and alanine were determined in 1HNMR spectra by calculating the ratio of 1H signal coupled to 13Cover the area of the total resonance (12C + 13C). 13C isotopomerpopulations of glutamate and GABA were determined from theanalysis of 13C NMR spectra by deconvolution of the multipletspresent in the resonance of each carbon.

2.4. Statistical analysis

Results are presented as means ± SEM. Since all experimentshad a paired control condition, paired Student’s t test was usedto compare measured data.

2.5. Microscopy sections

Under deep sodium pentobarbital anesthesia (100 mg/kg bodyweight, i.p.), male Wistar rats, CB1R null-mutant mice, kindly giftedby Dr. Catherine Ledent, Belgium (Ledent et al., 1999) of the CD-1strain and their wild-type littermates were fixed transcardiallywith a fixative (4% paraformaldehyde in 0.1 M phosphate buffer(PB), pH 7.4). The brains were removed and immersed in the fixa-tive overnight and then kept in 30% sucrose in physiological saline(0.9% NaCl) for at least 48 h before sectioning. Forty micron-thick

J.M.N. Duarte et al. / Neurochemistry International 60 (2012) 1–8 3

sections from the mouse brains and 30-lm-thick sections from therat brains were cut using a cryostate microtome (Leica) and col-lected into 0.1 M PB containing 0.1% sodium azide.

2.6. Immunohistochemistry

Free floating sections were blocked in a solution of 10% normalgoat serum (Vector Laboratories, CA, USA), 5% BSA, and 0.3% TritonX-100 for 40 min and incubated overnight in a primary antibodycocktail of guinea pig anti-CB1R (1:1000; Frontier Science,Hokkaido, Japan; licensed by Dr. Masahiko Watanabe; see Uchiga-shima et al., 2007) and mouse anti-GFAP (1:200; Calbiochem, Ger-many). Sections were then washed in phosphate buffer (PB; 0.1 M)and incubated with a secondary antibody cocktail containing dy-light 594 goat anti-guinea pig and dylight 488 goat anti-mouse(both at 1:200; Kirkegaard & Perry Laboratories, Inc., USA) for2 h. After washing in PB 0.1 M, the sections were mounted and cov-erslipped using Vectashield Hardest Mounting Medium (VectorLaboratories). Low magnification images where taken on a Zeissaxiovert 200 microscope equipped with AxioVision software andMosaiX module. Confocal images were taken using a ZeissLSM510 META confocal microscope. While the CB1R staining wasessentially similar to previous publications in the rat and thewild-type mouse brain, no immunostaining was detected in theCB1R knockout mouse brain sections at any resolution (Fig. 5E1–3) or with the secondary antibodies in the absence of the primaryantibodies (figure not shown).

3. Results

3.1. Incorporation of 13C labelling into NMR-detectable metabolites

The oxidation of [U–13C]glucose and [2-13C]acetate leads toenrichment of carbons from the tricarboxylic acid (TCA) cycleintermediates and related metabolites, like glutamate, glutamine,GABA, aspartate, lactate and alanine. Incorporation of the 13Catoms in different carbon positions of metabolites originate differ-ent isotopomer populations (molecules with different 12C and 13Clabelling patterns), which are easily distinguishable by 13C NMRspectroscopy (Duarte et al., 2007a). Fig. 1 shows a typical 13CNMR spectrum of a PCA extract from hippocampal slices super-fused for 3 h with 5.5 mM [U–13C]glucose and [2-13C]acetate in

Fig. 1. Representative 13C NMR spectrum of PCA extracts from hippocampal slices supspectrum expansion of glutamate C4 and GABA C2, Q and D indicate quartets and doub

the presence of 4AP (50 lM), and the expansion of the spectrumshows the resonances of glutamate C4 and GABA C2 at 34.2 and35.1 ppm, respectively. Based on the 13C labelling patterns of theadjacent carbons, the carbon resonance is divided in 9 linesgrouped in 4 sets, due to 13C–13C coupling. For instance, the gluta-mate C4 resonance in the 13C NMR spectra allows distinguishingC4S (C3 and C5 are unenriched), C4D34 (C3 is enriched but C5 isnot), C4D45 (C5 is enriched but C3 is not) and C4Q (C3 and C5are both enriched). Glutamate C4 originates GABA C2 by decarbox-ylation, and the labelling pattern is similar to that described forglutamate but now the C1 of GABA behaves as the C5 of glutamate.Thus, for GABA C2, the multiplets C2S, C2D23, C2D12 and C2Q referto isotopomers having both C1 and C3 unenriched, C3 enriched butnot C1, C1 enriched but not C3 and C1 and C3 both enriched,respectively.

Fig. 2 delineates the metabolism of [U–13C]glucose and[2-13C]acetate leading to the enrichment of glutamate and GABA,which are highly abundant in PCA extracts of hippocampal slices.[U–13C]glucose and [2-13C]acetate originate [1,2-13C]acetyl-CoAand [2-13C]acetyl-CoA, respectively, which enter the TCA cycle bycondensation with oxaloacetate and lead to the formation of gluta-mate C4D45 and C4S, respectively. In the next turns of the TCA cy-cle, metabolic intermediates become enriched and the oxidation of[U–13C]glucose and [2-13C]acetate originates glutamate C4Q andC4D34, respectively. Thus, the ratios of glutamate C4Q to C4D45(GluC4Q/D45) and glutamate C4D34 to C4S (GluC4D34/S) are indi-cators of the flux of the TCA cycle(s) that oxidize each substrate(Duarte et al., 2007a). The multiplets of GABA C2 can also be usedto evaluate the flux through the TCA cycle: the ratios of GABA C2Qto GABA C2D12 (GABAC2Q/D12) and GABA C2D23 to GABA C2S(GABAC2D23/S) reflect the TCA cycle flux oxidizing [U–13C]glucoseand [2-13C]acetate, respectively. Glucose is oxidized in both neu-rons and glia while acetate is mainly metabolized in glial cells(Waniewski and Martin, 1998; Wyss et al., 2011). Thus, we can in-fer about the glial TCA cycle flux by evaluating glutamate andGABA isotopomer populations originated by 13C-enriched acetate,namely GluC4D34/S and GABAC2D23/S (see Duarte et al., 2007a).

[1,2,3-13C3]pyruvate originated from [U–13C]glucose can enterthe TCA cycle by carboxylation to oxaloacetate in glial cells andoriginate glutamate labelled in C2 and C3 (e.g. Duarte et al.,2011). Since pyruvate is uniformly labelled, pyruvate carboxylationcannot be directly distinguished and can interfere with the

erfused for 3 h with medium containing [2-13C]acetate and [U–13C]glucose. In thelets, respectively.

Fig. 2. Schematic model of the fate of 13C labelling in atoms of glutamate and GABA upon metabolism of [U–13C]glucose (A) and [2-13C]acetate (B). [2-13C]acetate yields[2-13C]acetyl-CoA and [U–13C]glucose originates [1,2-13C]acetyl-CoA. Both acetyl-CoA molecules enter the TCA cycle by condensation with oxaloacetate, leading to theformation of glutamate enriched in the carbons 4 and 4 + 5, respectively. In the next turns of the TCA cycle, when oxaloacetate becomes enriched, molecules of glutamatelabelled in the carbons 3 + 4 and 3 + 4 + 5 are produced. This scheme was simplified to highlight the source of the labelling pattern observed in glutamate C4 and GABA C2.

4 J.M.N. Duarte et al. / Neurochemistry International 60 (2012) 1–8

evaluation of the TCA cycle by the aforementioned method. Thiswas solved by us using [2-13C]acetate that enters the astrocyticTCA cycle through acetyl-CoA.

Fig. 3. Fractional enrichment of lactate and alanine as calculated from 1H NMRspectra from hippocampal slice extracts after a 3 h superfusion period with[U–13C]glucose and [2-13C]acetate (both at 5.5 mM) in the presence of 4AP (50 lM)and CB1R ligands (the agonist WIN55212-2 [WIN], 1 lM, and the antagonist AM251,0.5 lM). White, gray, black and striped bars represent control, WIN, WIN + AM251and AM251 conditions. Results are presented as means ± SEM of five experiments.

3.2. [U–13C]glucose and [2-13C]acetate metabolism upon CB1Ractivation

As in previous experiments (Duarte et al., 2007a), in the presentexperimental design, the 13C enrichment of both [U–13C]glucoseand [2-13C]acetate was always 98 ± 1% (n = 5) in the superfusionmedium, as measured in 1H NMR spectra of superfusate samplescollected after superfusing the hippocampal slices. The metabolismof [U–13C]glucose through glycolysis leads to the enrichment oflactate and alanine and, as shown in Fig. 3, this enrichment wasnot changed by the pharmacological manipulation of CB1Rs.WIN55212-2 and AM251 also failed to alter lactate and alanineconcentration in hippocampal slices (Table 1). The concentrationof the neuronal marker NAA and the metabolic energy indicatorPCre/Cre remained unaltered upon CB1R activation or blockade(Table 1).

Metabolism of [U–13C]glucose and [2-13C]acetate in hippocam-pal slices (Fig. 2) leads to the enrichment of intermediary metabo-lites that are related to the TCA cycle in neurons and glia. In

Table 1The level of other important energy sources in the hippocampal slices are not affected by pharmacological manipulation of the CB1R. The concentrations of agonist WIN55212-2and antagonist AM251 were 1 and 0.5 lM, respectively, and the effect of these CB1R ligands is reported in relation to the respective control condition (in the presence of DMSO).Values are mean ± SEM of 5 experiments. NAA, N-acetylaspartate; PCre/Cre, phosphocreatine/creatine ratio.

Control (lmol/g protein) WIN55212-2 (% of control) WIN55212-2 + AM251 (% of control) AM251 (% of control)

NAA 16.7 ± 1.9 �0.7 ± 5.3 �5.7 ± 3.3 10.6 ± 5.7PCre/Cre 0.63 ± 0.05 �6.3 ± 7.1 1.3 ± 6.6 �4.4 ± 6.9Lactate 75.6 ± 9.7 �2.1 ± 9.2 7.3 ± 7.7 2.7 ± 7.9Alanine 4.0 ± 0.4 9.0 ± 27.6 5.8 ± 8.3 �14.4 ± 7.0

J.M.N. Duarte et al. / Neurochemistry International 60 (2012) 1–8 5

particular, as described above, the incorporation of 13C atoms intoglutamate and GABA allowed inferring about neuronal and astro-cytic TCA cycles. As shown in Fig. 4, selective activation of CB1Rswith WIN55212-2 (1 lM) during the superfusion of the hippocam-pal slices with the 13C-tracers decreased the ratios GluC4Q/D45and GABAC2Q/D12 in the 13C NMR spectra, suggesting a reductionof [U–13C]glucose oxidation in the TCA cycles that contribute to theformation of both glutamate and GABA. CB1R activation also inhib-ited the oxidation of [2-13C]acetate but only in the TCA cycle con-tributing to glutamate production, as suggested by the decreasedratio GluC4D34/S. These effects were indeed mediated by CB1Rssince AM251, while having no effect alone, prevented the effectsof WIN55212-2 on the metabolism of [U–13C]glucose and[2-13C]acetate.

3.3. CB1 cannabinoid receptor immunoreactivity in astrocytes

The previous results indicate that CB1R activation alters astrog-lial intermediary metabolism. Thus, astrocytes are expected to

Fig. 4. Representative resonances of glutamate C4 and GABA C2 in the 13C NMRspectra (A; see Fig. 1 for resonance assignment) and multiplet ratios calculated fromthese resonances (B). 13C NMR spectra were acquired from extracts from hippo-campal slices superfused with [U–13C]glucose and [2-13C]acetate (both at 5.5 mM)in the presence of 4AP (50 lM) and CB1R ligands WIN55212-2 (1 lM) and AM251(0.5 lM). White, gray and black bars represent WIN, WIN + AM251 and AM251conditions. All the experiments are paired experiments with a test condition (drug)versus control (containing DMSO), considered 100%. Results in the histogram aremean ± SEM of 5 experiments, comparison was made with a paired Student’s-t test:⁄P < 0.05, different from control.

express CB1Rs, although neuronal CB1Rs may also indirectly controlastrocyte metabolism. To map the putative presence of CB1Rs inastrocytes, we evaluated GFAP and CB1R immunofluorescent stain-ing in rat hippocampal slices by confocal microscopy. We observedthe strongest co-localization between CB1R and GFAP immunoflu-orescence in the superficial and the deep layers of the cortex, in theglobus pallidus (not shown) and in the hippocampus where thestrata pyramidale, radiatum and moleculare displayed the strongestoverlapping staining (Fig. 5A). At the highest resolution in 380 nm-thin optical sections, the CB1R immunoreactivity confined mostlyto fibres corresponding to putative axons and nerve terminals.Astrocytic CB1R immunoreactivity appeared weaker in intensity,but uniformly distributed among GFAP-positive cell bodies andfilopodes of pericapillary and parenchymal astrocytes in the den-tate gyrus (Fig. 5B1–3), stratum lacunosum moleculare (Fig. 5C1–3) and radiatum (Fig. 5D1–3).

4. Discussion

In contrast to peripheral and systemic effects, little is knownabout the local cerebral role of CB1 cannabinoid receptors to mod-ulate intermediary metabolism. The present results demonstratethat CB1Rs activation diminishes the oxidative metabolism of glu-cose in the acute hippocampal slices of the rat. The relative incor-poration of 13C atoms from [U–13C]glucose and [2-13C]acetate intothe carbon position 3 of glutamate and GABA suggests a reductionin the flux through the TCA cycle of both neuronal and astrocyticcompartments.

CB1R activation inhibited TCA cycle flux in neurons participat-ing in both glutamatergic and GABAergic neurotransmission, asindicated by reduced 13C label incorporation from [U–13C]glucoseinto glutamate and GABA C3, compared to glutamate C4 and GABAC2, respectively. However, the incorporation of labelling from[2-13C]acetate was only reduced in glutamate and not in GABA,suggesting that metabolism in the astrocytic compartment sup-porting and feeding GABAergic neurons is not significantly affectedby CB1R receptor modulation.

One should note that, in contrast to glutamate, GABA releasedinto the synaptic cleft is mostly cleared by the GABAergic nerveterminal rather than surrounding astrocytes (Schousboe andWaagepetersen, 2003). This suggests that, in comparison to gluta-mate, GABA synthesis is less dependent from glial glutamine pro-vision and, therefore, less sensitive to probing acetate oxidationin astrocytes. This may be the reason for not detecting reductionof acetate oxidation using GABA multiplets.

The presented data do not exclude the possibility of reducedanaplerotic fluxes upon CB1R activation, which would lead to labelincorporation in glutamate carbons other than C4. However, ana-plerosis in the brain mainly occurs through pyruvate carboxylationin astrocytes (see Duarte et al., 2011) and, by using the astrocyte-specific [2-13C]acetate, we found a reduction in astrocytic TCA cy-cle fluxes that replenish glutamate pools.

Our results suggesting decreased oxidative fluxes upon CB1Ractivation are consistent with previous studies with 14C-2-deoxy-glucose autoradiography indicating that both WIN55212-2 and

Fig. 5. CB1R immunoreactivity is present in hippocampal astrocytes. (A) Low magnification (5�) fluorescent microscope image showing the distribution of fluorescentimmunostaining of the astrocytic marker, glial fibrillar acidic protein (GFAP; green) and the CB1R (red) in labelled rat hippocampal subregions. (B–D) CB1R immunoreactivity(B2, C2, D2; in red) is widely distributed in GFAP-positive cells (B1, C1, D1), i.e. in astrocytes, (marked by arrows), in the hippocampus including the regions of the dentategyrus (orthogonal projections, (B), and in the strata lacunosum-moleculare (C) and radiatum (D). Regularly, perycytes also appeared with CB1R immunoreactivity (arrowheadsin B). All confocal images seen in B–D represent 380 nm-thick optical sections extracted from Z-stacks from coronal slices (63�/1.40 oil DIC M27, 1 � zoom). (E) Typicalimmunofluorescent staining with the CB1R antibody (Uchigashima et al., 2007) in the rat (E1) and the WT mouse (E2) brain. In the sagittal slices, the globus pallidus externaand interna, the substantia nigra as well as the stratum pyramidale of the hippocampus give strong signal. The antibody does not stain in the CB1R knockout mouse (Ledentet al., 1999), e.g. in the representative sagittal section (E3), or at any resolution in confocal microscopy (not shown).

6 J.M.N. Duarte et al. / Neurochemistry International 60 (2012) 1–8

D9-THC decrease in vivo rates of glucose uptake in the hippocam-pal tissue (Freedland et al., 2002; Pontieri et al., 1999; Whitlowet al., 2002; Matias et al., 2008). In contrast, it was reported thatCB1R activation stimulates glucose utilization in other brain re-gions (Freedland et al., 2002; Pontieri et al., 1999; Whitlow et al.,2002; Matias et al., 2008), and that D9-THC increases glucose oxi-dation to CO2 and glucose incorporation into glycogen and phos-pholipids in primary cultures of cortical and hippocampalastrocytes (Sánchez et al., 1998) and C6 glioma cells (Sánchezet al., 1997), which was suggested to occur through activation ofMAPK (Guzmán and Sánchez, 1999). Moreover, the in vivo effectof CB1R activation on the uptake of glucose derivatives was shownto be biphasic and dose-dependent (Margulies and Hammer,1991), which could explain different results obtained by differentauthors. Together with these glucose uptake studies, reported de-creases in cerebral blood flow by cannabinoids (Bloom et al.,1997; Stein et al., 1998) are in agreement with reductions in hippo-campal intermediary metabolism.

Although cannabinoids may modulate glucose uptake andmetabolism in the hippocampus, we found that the fractionalenrichments of lactate and alanine, as well as their absoluteamount, remained unchanged upon the manipulation of CB1Rs.Since the concentration of glucose in the brain at normoglycaemiais around 1 mM (Duarte et al., 2009), the lack of effect of CB1R onthese glycolytic-related measures in superfused hippocampalslices may result from a certain degree of saturation of the glyco-lytic pathway by the amount of glucose provided in the superfu-sion medium (5.5 mM), which is larger than the Km ofhexokinase (Barros et al., 2007).

It should be emphasized that astrocytes are involved in the up-take of glucose from the blood stream that is metabolized to lactate

and putatively shuttled to neurons (Magistretti, 2006; Pellerin,2008). Since the superfusion of hippocampal slices provides sub-strates without using the vascular system, there is a metaboliccomponent in astrocytic metabolism that may be modulated byCB1R activation being absent in our experimental system. In anycase this should not undermine the main conclusion of this studybecause all the cellular elements have access to the provided sub-strates. Thus, we now directly show that the global effect of CB1Ractivation in hippocampal slices results in a reduction of neuronaland glial metabolism through the TCA cycle.

The activation of presynaptic CB1Rs inhibits the stimulation-evoked release of GABA and glutamate (Katona et al., 1999; Wang,2003; Martire et al., 2011). It is feasible thus that the activation ofthese presynaptic CB1Rs diminishes the demand for newly syn-thesised GABA and glutamate from the TCA cycle. However, func-tional CB1Rs were found in astrocytes around CA3–CA1 synapses.The activation of these astrocytic CB1Rs increases astrocytic gluta-mate release and induce synaptic potentiation (Navarrete and Ara-que, 2010), which may eventually demand more glutamate fromthe TCA cycle. It has been also shown that in cultured astrocytesand astrocytomas, CB1Rs are capable of controlling glucose metab-olism (Sánchez et al., 1998; Stella, 2010), suggesting that cannabi-noids can also interact with the TCA cycle independently fromsynaptic activity.

If the synthesis of glutamate and GABA was reduced withoutreduction of TCA cycle flux, less carbon skeletons would lead toglutamate synthesis from 2-oxoglutarate. Hence, 13C would con-tinue to cycle in the TCA cycle leading to higher enrichment ofC2 and C3 versus C4 in molecules of 2-oxoglutarate (and conse-quently, of glutamate). This would eventually lead to increased ra-tios GluC4D34/S and GluC4DQ/D45, as well as the equivalent ratios

J.M.N. Duarte et al. / Neurochemistry International 60 (2012) 1–8 7

in GABA upon treatment with WIN55212-2, compared to control.Therefore, our observation is consistent with a simultaneousreduction both in glutamate synthesis and in the TCA cycle flux.De novo synthesis of glutamate requires pyruvate carboxylationin glia (Hassel, 2000). A reduction of pyruvate carboxylation wouldlead to lower enrichment of C3 versus C4 in glutamate (and of C2in GABA). Further studies are invited to investigate this possibilityby using glucose labelled in carbons C1 and/or C6.

Since the TCA cycle takes place in the matrix of the mitochon-drion, CB1R-triggered signals from the cytoplasm to the mitochon-dria or a previously undetected location of CB1Rs in themitochondrion itself may be responsible for the observed effects.Indeed, CB1R activation can affect oxidative phosphorylation inbrain mitochondria (Costa and Colleoni, 2000). Additionally, CB1Rblockade with the antiobesity molecule, SR141716A (Rimonabant,Acomplia) increases fatty acid oxidation in liver mitochondria(Flamment et al., in press), and D9-THC inhibits respiration of oralcancer cells and of isolated bovine heart mitochondria (Whyteet al., 2010). Last but not least, CB1R activation impairs mitochon-drial biogenesis in cultured mouse and human white adipocytes(Tedesco et al., 2010).

5. Conclusions

The inhibition of mitochondrial intermediary metabolism canhypothetically diminish releasable GABA and glutamate pools, thuscontributing to changes in synaptic plasticity (Freund et al., 2003).In diseases with increased endocannabinoid levels and impairedsynaptic plasticity and memory such as Alzheimer’s disease (Mul-der et al., 2011), suppression of glucose metabolism as an early sig-nal is indeed detected (Mosconi, 2005). Taking into account thatneuroprotection can be achieved via modulating glucose oxida-tion; the complex pathological and therapeutic role of CB1Rs inthe function of the hippocampus (Campbell and Gowran, 2007)now appears to involve one more mechanism.

Acknowledgements

This work was supported by the research grants of the Portu-guese Government, PTDC/SAU-OSM/105663/2008 (A.K.), PDTC/EBB-EBI/115810/2009 (R.A. Carvalho), PTDC/SAU-NEU/108668/2008 (R.A. Cunha) and Fundação Oriente (R.A. Cunha). J.M.N.D.and S.G.F. acknowledge the Ph.D. fellowships from FCT, SFRH/BD/17795/2004 and SFRH/BD/33467/2008. A.K. acknowledges the helpof the European Neuroscience Institute Network (ENI-Net).

References

Balázsa, T., Bíró, J., Gullai, N., Ledent, C., Sperlágh, B., 2003. CB1-cannabinoidreceptors are involved in the modulation of non-synaptic [3H]serotonin releasefrom the rat hippocampus. Neurochem. Int. 52, 95–102.

Barros, L.F., Bittner, C.X., Loaiza, A., Porras, O.H., 2007. A quantitative overview ofglucose dynamics in the gliovascular unit. Glia 55, 1222–1237.

Bloom, A.S., Tershner, S., Fuller, S.A., Stein, E.A., 1997. Cannabinoid-inducedalterations in regional cerebral blood flow in the rat. Pharmacol. Biochem.Behav. 57, 625–631.

Campbell, V.A., Gowran, A., 2007. Alzheimer’s disease; taking the edge off withcannabinoids? Br. J. Pharmacol. 152, 655–662.

Cerdán, S., Künnecke, B., Seelig, J., 1990. Cerebral metabolism of [1,2-13C2]acetate asdetected by in vivo and in vitro 13C NMR. J. Biol. Chem. 265, 12916–12926.

Costa, B., Colleoni, M., 2000. Changes in rat brain energetic metabolism afterexposure to anandamide or Delta9-tetrahydrocannabinol. Eur. J. Pharmacol.395, 1–7.

Di Marzo, V., Piscitelli, F., Mechoulam, R., 2011. Cannabinoids andendocannabinoids in metabolic disorders with focus on diabetes. Handb. Exp.Pharmacol. 203, 75–104.

Duarte, J.M.N., Cunha, R.A., Carvalho, R.A., 2007a. Different metabolism ofGlutamatergic and GABAergic compartments in superfused hippocampalslices characterized by nuclear magnetic resonance spectroscopy.Neuroscience 144, 1305–1313.

Duarte, J.M.N., Nogueira, C., Mackie, K., Oliveira, C.R., Cunha, R.A., Köfalvi, A., 2007b.Increase of cannabinoid CB1 receptor density in the hippocampus ofstreptozotocin-induced diabetic rats. Exp. Neurol. 204, 479–484.

Duarte, J.M.N., Lanz, B., Gruetter, R., 2011. Compartmentalised cerebral metabolismof [1,6-13C]glucose determined by in vivo 13C NMR spectroscopy at 14.1 T.Front. Neuroenergetics 3, 3.

Duarte, J.M.N., Morgenthaler, F.D., Lei, H., Poitry-Yamate, C., Gruetter, R., 2009.Steady-state brain glucose transport kinetics re-evaluated with a four-stateconformational model. Front. Neuroenergetics 1, 6.

Flamment, M., Gueguen, N., Wetterwald, C., Simard, G., Malthièry, Y., Ducluzeau,P.H., in press. Effects of the cannabinoid CB1 antagonist, rimonabant, on hepaticmitochondrial function in rats fed a high fat diet. Am. J. Physiol. Endocrinol.Metab.

Freedland, C.S., Whitlow, C.T., Miller, M.D., Porrino, L.J., 2002. Dose-dependenteffects of Delta9-tetrahydrocannabinol on rates of local cerebral glucoseutilization in rat. Synapse 45, 134–142.

Freund, T.F., Katona, I., Piomelli, D., 2003. Role of endogenous cannabinoids insynaptic signaling. Physiol. Rev. 83, 1017–1066.

Guzmán, M., Sánchez, C., 1999. Effects of cannabinoids on energy metabolism. LifeSci. 65, 657–664.

Hampson, R.E., Deadwyler, S.A., 1999. Cannabinoids, hippocampal function andmemory. Life Sci. 65, 715–723.

Hardie, D.G., 2008. AMPK: a key regulator of energy balance in the single cell andthe whole organism. Int. J. Obes. (Lond.) 32, S7–12.

Hassel, B., 2000. Carboxylation and anaplerosis in neurons and glia. Mol. Neurobiol.22, 21–40.

Katona, I., Sperlágh, B., Sík, A., Köfalvi, A., Vizi, E.S., Mackie, K., Freund, T.F., 1999.Presynaptically located CB1 cannabinoid receptors regulate GABA release fromaxon terminals of specific hippocampal interneurons. J. Neurosci. 19, 4544–4558.

Kawamura, Y., Fukaya, M., Maejima, T., Yoshida, T., Miura, E., Watanabe, M., Ohno-Shosaku, T., Kano, M., 2006. The CB1 cannabinoid receptor is the majorcannabinoid receptor at excitatory presynaptic sites in the hippocampus andcerebellum. J. Neurosci. 26, 2991–3001.

Ledent, C., Valverde, O., Cossu, G., Petitet, F., Aubert, J.F., Beslot, F., Böhme, G.A.,Imperato, A., Pedrazzini, T., Roques, B.P., et al., 1999. Unresponsiveness tocannabinoids and reduced addictive effects of opiates in CB1 receptor knockoutmice. Science 283, 401–404.

Li, C., Jones, P.M., Persaud, S.J., 2011. Role of the endocannabinoid system in foodintake, energy homeostasis and regulation of the endocrine pancreas.Pharmacol. Ther. 129, 307–320.

Magistretti, P.J., 2006. Neuron-glia metabolic coupling and plasticity. J. Exp. Biol.209, 2304–2311.

Margulies, J.E., Hammer Jr., R.P., 1991. D9-Tetrahydrocannabinol alters cerebralmetabolism in a biphasic, dose-dependent manner in rat brain. Eur. J.Pharmacol. 202, 373–378.

Martire, A., Tebano, M.T., Chiodi, V., Ferreira, S.G., Cunha, R.A., Köfalvi, A., Popoli, P.,2011. Pre-synaptic adenosine A2A receptors control cannabinoid CB1 receptor-mediated inhibition of striatal glutamatergic neurotransmission. J. Neurochem.116, 273–280.

Matias, I., Di Marzo, V., Köfalvi, A., 2008. Endocannabinoids in energy homeostasisand metabolic disorders. In: Köfalvi, A. (Ed.), Cannabinoids and the Brain.Springer, USA, pp. 131–160.

Messier, C., 2004. Glucose improvement of memory: a review. Eur. J. Pharmacol.490, 33–57.

Mosconi, L., 2005. Brain glucose metabolism in the early and specific diagnosis ofAlzheimer’s disease. FDG-PET studies in MCI and AD. Eur. J. Nucl. Med. Mol.Imaging 32, 486–510.

Mulder, J., Zilberter, M., Pasquaré, S., Alpár, A., Schulte, G., Ferreira, S.G., Köfalvi, A.,Martín-Moreno, A.M., Keimpema, E., Tanila, H., Watanabe, M., Mackie, K.,Hortobágyi, T., de Ceballos, M.L., Harkány, T., 2011. Molecular reorganization ofendocannabinoid signalling in Alzheimer’s disease. Brain 134, 1041–1060.

Navarrete, M., Araque, A., 2010. Endocannabinoids potentiate synaptic transmissionthrough stimulation of astrocytes. Neuron 68, 113–126.

Öz, G., Berkich, D.A., Henry, P.G., Xu, Y., LaNoue, K., Hutson, S.M., Gruetter, R., 2004.Neuroglial metabolism in the awake rat brain: CO2 fixation increases with brainactivity. J. Neurosci. 24, 11273–11279.

Pagano, C., Pilon, C., Calcagno, A., Urbanet, R., Rossato, M., Milan, G., Bianchi, K.,Rizzuto, R., Bernante, P., Federspil, G., Vettor, R., 2007. The endogenouscannabinoid system stimulates glucose uptake in human fat cells viaphosphatidylinositol 3-kinase and calcium-dependent mechanisms. J. Clin.Endocrinol. Metab. 92, 4810–4819.

Pellerin, L., 2008. Brain energetics (thought needs food). Curr. Opin. Clin. Nutr.Metab. Care 11, 701–705.

Pontieri, F.E., Conti, G., Zocchi, A., Fieschi, C., Orzi, F., 1999. Metabolic mapping of theeffects of WIN 55212-2 intravenous administration in the rat.Neuropsychopharmacology 21, 773–776.

Quarta, C., Bellocchio, L., Mancini, G., Mazza, R., Cervino, C., Braulke, L.J., Fekete, C.,Latorre, R., Nanni, C., Bucci, M., et al., 2010. CB1 signaling in forebrain andsympathetic neurons is a key determinant of endocannabinoid actions onenergy balance. Cell Metab. 11, 273–285.

Sánchez, C., Galve-Roperh, I., Rueda, D., Guzmán, M., 1998. Involvement ofsphingomyelin hydrolysis and the mitogen-activated protein kinase cascadein the Delta9-tetrahydrocannabinol-induced stimulation of glucose metabolismin primary astrocytes. Mol. Pharmacol. 54, 834–843.

8 J.M.N. Duarte et al. / Neurochemistry International 60 (2012) 1–8

Sánchez, C., Velasco, G., Guzmán, M., 1997. Delta9-tetrahydrocannabinol stimulatesglucose utilization in C6 glioma cells. Brain Res. 767, 64–71.

Schousboe, A., Waagepetersen, H.S., 2003. Role of astrocytes in homeostasis ofglutamate and GABA during physiological and pathophysiological conditions.Adv. Mol. Cell Biol. 31, 461–474.

Smith, P.K., Krohn, R.I., Hermanson, G.T., Mallia, A.K., Gartner, F.H., Provenzano,M.D., Fujimoto, E.K., Goeke, N.M., Olson, B.J., Klenk, D.C., 1985. Measurement ofprotein using bicinchoninic acid. Anal. Biochem. 150, 76–85.

Song, D., Bandsma, R.H., Xiao, C., Xi, L., Shao, W., Jin, T., Lewis, G.F., 2011. Acutecannabinoid receptor type 1 (CB1R) modulation influences insulin sensitivity byan effect outside the central nervous system in mice. Diabetologia 54, 1181–1189.

Stein, E.A., Fuller, S.A., Edgemond, W.S., Campbell, W.B., 1998. Selective effects ofthe endogenous cannabinoid arachidonylethanolamide (anandamide) onregional cerebral blood flow in the rat. Neuropsychopharmacology 19, 481–491.

Stella, N., 2010. Cannabinoid and cannabinoid-like receptors in microglia,astrocytes, and astrocytomas. Glia 58, 1017–1030.

Suzuki, A., Stern, S.A., Bozdagi, O., Huntley, G.W., Walker, R.H., Magistretti, P.J.,Alberini, C.M., 2011. Astrocyte-neuron lactate transport is required for long-term memory formation. Cell 144, 810–823.

Tedesco, L., Valerio, A., Dossena, M., Cardile, A., Ragni, M., Pagano, C., Pagotto, U.,Carruba, M.O., Vettor, R., Nisoli, E., 2010. Cannabinoid receptor stimulationimpairs mitochondrial biogenesis in mouse white adipose tissue, muscle, and

liver: the role of eNOS, p38 MAPK, and AMPK pathways. Diabetes 59, 2826–2836.

Tibbs, G.R., Barrie, A.P., van Mieghem, F.J., McMahon, H.T., Nicholls, D.G., 1989.Repetitive action potentials in isolated nerve terminals in the presence of 4-aminopyridine: effects on cytosolic free Ca2+ and glutamate release. J.Neurochem. 53, 1693–1699.

Uchigashima, M., Narushima, M., Fukaya, M., Katona, I., Kano, M., Watanabe, M.,2007. Subcellular arrangement of molecules for 2-arachidonoyl-glycerol-mediated retrograde signaling and its physiological contribution to synapticmodulation in the striatum. J. Neurosci. 27, 3663–3676.

Wang, S.J., 2003. Cannabinoid CB1 receptor-mediated inhibition of glutamaterelease from rat hippocampal synaptosomes. Eur. J. Pharmacol. 469, 47–55.

Waniewski, R.A., Martin, D.L., 1998. Preferential utilization of acetate by astrocytesis attributable to transport. J. Neurosci. 18, 5225–5233.

Whitlow, C.T., Freedland, C.S., Porrino, L.J., 2002. Metabolic mapping of the time-dependent effects of delta9-tetrahydrocannabinol administration in the rat.Psychopharmacology (Berl) 161, 129–136.

Whyte, D.A., Al-Hammadi, S., Balhaj, G., Brown, O.M., Penefsky, H.S., Souid, A.K.,2010. Cannabinoids inhibit cellular respiration of human oral cancer cells.Pharmacology 85, 328–335.

Wyss, M.T., Magistretti, P.J., Buck, A., Weber, B., 2011. Labeled acetate as a marker ofastrocytic metabolism. J. Cereb. Blood Flow. Metab. 31, 1668–1674.