Forebrain-Specific Inactivation of Gq/G11 Family G Proteins Results in Age-Dependent Epilepsy and Impaired Endocannabinoid Formation

Metabotropic receptors coupled to G_q/G₁₁ family G proteins criticallycontribute to nervous system functions by modulating synaptictransmission, often facilitating excitation. We investigatedthe role of G_q/G₁₁ family G proteins in the regulation of neuronalexcitability in mice that selectively lack the ${alpha}$ -subunits ofG_q and G₁₁, G ${alpha}$ _q and G ${alpha}$ ₁₁, respectively, in forebrain principalneurons. Surprisingly, mutant mice exhibited increased seizuresusceptibility, and the activation of neuroprotective mechanismswas impaired. We found that endocannabinoid levels were reducedunder both basal and excitotoxic conditions and that increasedsusceptibility to kainic acid could be normalized by the enhancementof endocannabinoid levels with an endocannabinoid reuptake inhibitor,while the competitive cannabinoid type 1 receptor antagonistSR141716A did not cause further aggravation. These findingsindicate that G_q/G₁₁ family G proteins negatively regulate neuronalexcitability in vivo and suggest that impaired endocannabinoidformation in the absence of G_q/G₁₁ contributes to this phenotype.

INTRODUCTION

Top

Introduction

Several metabotropic receptors of the central nervous systemsignal through heterotrimeric G proteins of the G_q/G₁₁ family,for example, the metabotropic glutamate receptor subtypes 1and 5, the M₁ muscarinic acetylcholine receptor, and the 5-hydroxytryptamine2 (5-HT₂) serotonin receptor. Stimulation of G_q/G₁₁-coupledreceptors causes the activation of phospholipase C-ß(PLC-ß) isoforms, leading to inositol phospholipidbreakdown, protein kinase C activation, and intracellular calciummobilization (18). G_q/G₁₁-coupled receptors have been shownto contribute to the modulation of synaptic transmission inthe hippocampus and cerebellum (1, 23, 25, 33, 37, 46, 48) andare known to facilitate neuronal activation by opening voltage-dependentcation channels and closing potassium channels (10, 12). Inline with this, agonists at the G_q/G₁₁-coupled metabotropicglutamate receptors or the muscarinic acetylcholine receptorspromote seizures in vitro and in vivo (6, 11, 57), while antagonistsat these receptors have been suggested to have anticonvulsiveeffects (7, 8).

In vitro data, however, indicate that G_q/G₁₁-coupled receptorscan also indirectly exert inhibitory effects through the inhibitionof presynaptic calcium channels (22, 45) or the activation ofthe endocannabinoid system. Endocannabinoids such as anandamideand 2-arachidonoyl glycerol (2-AG) are lipophilic neuromodulatorsthat are synthesized and released "on demand" in postsynapticneurons in response to depolarization-induced calcium influx(14, 34, 50, 69) or activation of G_q/G₁₁-coupled metabotropicreceptors (20, 21, 32, 38, 41, 51, 52, 65). Both transmembraneCa²⁺ influx and G_q/G₁₁-mediated intracellular Ca²⁺ mobilizationenhance anandamide biosynthesis via an N-acyltransferase anda specific phospholipase D-like enzyme (13, 15, 16, 56, 61,62). In parallel, G_q/G₁₁-mediated PLC-ß activationresults in the production of diacylglycerol (DAG), which servesas a precursor for 2-AG via Ca²⁺-sensitive sn-1-DAG lipases(3). PLC-ß, DAG, and sn-1-DAG lipase were suggestedto be of special importance for the production of that poolof 2-AG that functions as an endocannabinoid precursor, especiallyin response to neuronal depolarization and in retrograde signaling(9, 16, 44, 56, 62, 63). Endocannabinoids act mostly via presynapticcannabinoid type 1 receptors to inhibit neurotransmitter releaseat inhibitory as well as excitatory synapses (2, 19, 68), andthey were shown to protect the brain from excitotoxicity andother forms of neuronal damage (16, 43, 56, 62, 63). However,the relative weight and function of G_q/G₁₁-mediated signalingin the regulation of endocannabinoid levels in vivo are unclear.It is also not clear whether the net effect of G_q/G₁₁ signalingin vivo is excitatory or inhibitory.

The two main members of the G_q/G₁₁ family are G_q and G₁₁, whichare almost ubiquitously expressed in the central nervous systemand have highly redundant functions. Since genetic inactivationof both G ${alpha}$ _q and G ${alpha}$ ₁₁ causes embryonic lethality (49), we employedthe Cre/loxP system to generate mice which allow tissue-specificinactivation of G ${alpha}$ _q in constitutively G ${alpha}$ ₁₁-deficient mice (67).These mice were crossed with the Camkcre4 mouse line (39), whichcauses recombination selectively in principal neurons of theforebrain (40). Forebrain-specific G ${alpha}$ _q/G ${alpha}$ ₁₁ double knockout (fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO)mice were viable and fertile and showed normal basic sensoryand motor abilities, but females were unable to raise normalmaternal behavior toward their young (66). We used fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKOmice to investigate the role of G_q/G₁₁ in the regulation ofneuronal excitability, with a special focus on the role of G_q/G₁₁in endocannabinoid synthesis.

MATERIALS AND METHODS

Top

Materials and Methods

Mice. Genotyping for the gnaq^flox allele, the gna11 wild-type andknockout alleles, and the Cre transgene was described previously(67). The primers used for detecting the Camkcre4 transgenewere 5'-TCTCCATTTGCACTCAGGAGC-3' and 5'-AAAACGCCTGGCGATCCCT-G-3'.Mice were housed under specific-pathogen-free conditions, andall animal experiments were performed in accordance with institutionalanimal care and use committee regulations. The genetic backgroundof the mice was predominantly C57BL6/N (fourth-generation backcross),and nonlittermate C57BL6/N mice or littermates with the Camkcre4^–^/^–gnaq^fl/fl gna11^wt/wt genotype were used as controls. No significantdifferences were observed between the two control groups.

Animal experiments. Kainic acid (KA; Sigma, Deisenhofen, Germany) and pentylenetetrazole(PTZ; Sigma) were dissolved in saline and administered intraperitoneally(i.p.) at the indicated dosages. N-Piperidino-5-(4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-3-pyrazolone-carboxamide(SR141716A; Cayman Chemical, Ann Arbor, MI) and OMDM-2 [(R)-N-oleoyl-(1'-hydroxybenzyl)-2'-ethanolamine](synthesized as previously described [(54]) were dissolved inTween 80-ethanol-0.9% saline (1:1:18) and injected i.p. 30 minbefore KA treatment at the concentrations indicated in the figurelegends (for six to eight mice per group). Mice were monitoredfor 60 min, and responses were scored every 5 minutes by anobserver blinded to the genotype, as described previously (24).Briefly, responses were scored as follows: 0, normal behavior;1, rigid or slowed walking; 2, isolated jerks; 3, repetitivejerks; 4, forelimb clonus, isolated clonic jerks; 5, repetitiveclonic jerks, mouse on its side with fore- and hindlimb clonus;6, death. Video and electroencephalography (EEG) monitoringof mice were performed as described previously (35).

Histology. Mice were deeply anesthetized with pentobarbital, 100 mg/kgof body weight i.p., and perfused with 4% paraformaldehyde viathe left cardiac ventricle. Their brains were postfixed overnightand then stored in 0.5% paraformaldehyde at 4°C. Fifty-micrometervibratome sections were cut and incubated at 4°C with antibodiesdirected against the following proteins: c-fos protein (sc-52;Santa Cruz Biotechnology, Santa Cruz, CA; diluted 1:20,000,incubated for 72 h); zif-268 (Santa Cruz Biotechnology; diluted1:200, incubated for 72 h); G ${alpha}$ _q/G ${alpha}$ ₁₁ (sc-392; Santa Cruz Biotechnology;diluted 1:1,000, incubated for 16 h); GFAP (DAKO, Glostrup,Denmark; diluted 1:500, incubated for 16 h); NeuN (Chemicon,Temecula, CA; diluted 1:2,000, incubated for 16 h); GAD67 (Chemicon;diluted 1:2,000, incubated for 16 h); calbindin (Chemicon; diluted1:2,000, incubated for 16 h); calretinin (Chemicon; diluted1:5,000, incubated for 16 h); and MAP2 (Chemicon; diluted 1:1,000,incubated for 16 h). For staining we used the Vectastain EliteABC kits (Vector Labs, Burlingame, CA) and diaminobenzidine(Vector Labs). Nissl staining was performed according to standardprotocols. Timm staining was performed as described previously(5). Sections were photographed using a Leica DMLB microscopeand a Leica DC300F camera (Leica, Bensheim, Germany). The numbersof c-fos- or zif-268-positive cells were counted manually orwith a CellExplorer 2003 program (BioSciTec, Germany) by investigatorsblinded to genotype and treatment. For each group of mice, twosections from each of four animals were analyzed.

Endocannabinoid levels. Mice (six to eight per group) were sacrificed 20 min after injectionof 20 mg/kg kainate, and hippocampi were rapidly dissected,snap-frozen, and stored at –80°C. Endocannabinoidswere extracted, and their levels were measured by isotope dilutionliquid chromatography-mass spectrometry as described previously(40).

Electrophysiology. Hippocampal sections from 3- to 5-week-old forebrain-specificGa_q/Ga₁₁ double knockout mice and wild-type mice were preparedas described previously (53). Recordings were made from CA1pyramidal neurons in slices with patch pipettes pulled from1.5-mm-outer-diameter, thick-walled glass tubing (1511-M; Friedrich& Dimmock). Whole-cell tight seals (>2 G ${Omega}$ ) were made onthe soma under visual control with a 40x water immersion lens.Cells were identified as pyramidal neurons by using both electricaland anatomical criteria. Capacitance was fully compensated forby a patch clamp amplifier (EPC-9; HEKA). The holding potentialwas –60 or –70 mV. The range of series resistanceswe accepted for recordings was 10 to 15 M ${Omega}$ . Bipolar stimulationelectrodes constructed from thin tungsten wires (50-µmouter diameter) were placed on the stratum radiatum in the CA2to CA3 pyramidal regions. For inhibitory postsynaptic current(IPSC) recording, the pipette solution contained 130 mM CsCl,5 mM KCl, 5 mM NaCl, 10 mM HEPES, 0.6 mM EGTA, 4 mM Mg-ATP,and 0.3 mM GTP, with the pH adjusted to 7.3 with NaOH. A non-N-methyl-D-aspartate(non-NMDA) receptor antagonist (6-cyano-7-nitroquinoxalline-2,3-dione,20 µM) and an NMDA antagonist (DL-2-amino-5-phosphonovalericacid, 50 µM) were always added to the perfusing solution.The CB₁ receptor agonist WIN55,212-2 {R-(+)-[2,3-dihydro-5-methyl-3-[(morpholinyl)methyl]pyrrolo[1,2,3-de]-1,4-benzoxazinyl]-(1-naphthalenyl)methanonemesylate} was applied at 5 mM. Single traces are averaged fromfive consecutive IPSCs acquired before and at the end of theWIN55,212-2 application. Chemicals were obtained from Sigma-AldrichCo.

Statistics. All data except electrophysiological data are presented as means± standard error of the means; electrophysiological dataare presented as means ± standard deviations. Data wereanalyzed by analysis of variance followed by Bonferroni's post-hoctest. Statistical analysis was performed by repeated-measuretwo-way analysis of variance for the factors "treatment" and"genotype" and the interaction "treatment/genotype."

RESULTS

Top

Results

Starting at the age of about 3 months, fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice showedspontaneous epileptic seizures, and both the frequency of seizuresin each animal and the number of affected animals increasedwith age (Fig. 1A). The most frequent forms of seizure weremild myoclonic seizures, and combined EEG and video monitoringof mutant mice showed a series of rhythmic spikes (0.2 ±0.03 Hz; 45 s, 7-min duration; 0 to 5 times a day) (Fig. 1B,left panel) concomitant with myocloni of the body. Such episodeswere not observed in wild-type animals (Fig. 1B, right panel).Some animals also experienced clonic-tonic seizures, leadingto an age-dependent reduction of survival (Fig. 1C).

View larger version (29K):
[in this window]
[in a new window]
FIG. 1. Spontaneous epileptic seizures in fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice. (A) Percentages of fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice (KO; n = 30) (triangles) and control mice (WT; n = 30) (squares) with epileptic seizures. (B) Cortical EEGs from an fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mouse during a myoclonic seizure (DKO) and from a wild-type mouse without seizures (WT). (C) Percentages of postnatal survival of fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice (KO; n = 30) (triangles) and control mice (WT; n = 30) (squares).

Histological and immunohistochemical analyses of mice youngerthan 3 months did not show abnormalities (data not shown). However,older mice, especially those with symptomatic epilepsy, showedneuronal degeneration and reactive gliosis in the hippocampalCA1 region, indicative of temporal lobe epilepsy (Fig. 2). Incontrast to other models of temporal lobe epilepsy, we did notobserve mossy fibers sprouting in sections visualized by Timmstaining (data not shown).

View larger version (82K):
[in this window]
[in a new window]
FIG. 2. Histological changes in hippocampi of fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice. Hippocampal sections from 8-month-old control mice (upper row) and fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice (lower row) stained with antibodies against G ${alpha}$ _q/G ${alpha}$ ₁₁, NeuN, and GFAP proteins.

To investigate whether abnormalities in neuronal excitabilityare present before the onset of spontaneous seizures, we studiedthe effects of chemical convulsants in 6- to 10-week-old fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKOmice in vivo. We found that the application of the glutamatereceptor agonist KA or the gamma aminobutyric acid type A (GABA_A)receptor antagonist PTZ induced stronger seizures in fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKOmice than in wild-type mice (Fig. 3A and C) and that mutantmice exhibited seizures at dosages that were already subconvulsivein wild-type mice (Fig. 3A). Latency to seizure onset was decreasedin fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice (Fig. 3B and D), while seizure-associatedlethality was strongly increased (data not shown). Mice witha single deficiency in G ${alpha}$ _q or G ${alpha}$ ₁₁ showed mildly increased seizuresusceptibility when challenged with kainic acid (see Fig. S1in the supplemental material).

View larger version (27K):
[in this window]
[in a new window]
FIG. 3. Increased sensitivities toward chemical convulsants of 6- to 10-week-old fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice compared with those of control mice. Seizure scores (A and C) and latency periods (B and D) to first clonic-tonic (ct) convulsions after intraperitoneal application of kainic acid (A and B) or PTZ (C and D) in control and fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice (five to six mice per group). *, P < 0.05; **, P < 0.01; ***, P < 0.001. Seizures were scored as described previously (28).

KA-induced seizures have been shown to be accompanied by enhancedexpression of immediate early genes like c-fos, fosB, or zif-268in the hippocampi of wild-type mice (30), and especially c-fosexpression appears to be critically involved in the protectionagainst KA-induced excitotoxicity (40, 70). To test whetherthe impaired activation of potentially neuroprotective pathwaysmay contribute to increased seizure susceptibility in fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKOmice, we investigated the expression of c-fos and zif-268 proteinsin untreated and KA-treated wild-type and mutant mice (Fig.4). Compared to the upregulation in wild-type mice, the KA-inducedupregulation of both proteins was almost abrogated in hippocampifrom fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice (Fig. 4), suggesting that neuroprotectivemechanisms are impaired in these mice.

View larger version (25K):
[in this window]
[in a new window]
FIG. 4. Kainate-induced upregulation of immediate early genes is impaired in fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice. (A and B) Staining with antibodies against c-fos (A) and zif-268 (B) proteins in the hippocampi of control and fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice 45 min after intraperitoneal injection of saline or 20 mg/kg KA. (C and D) Statistical evaluation of the number of c-fos-positive (C) or zif-268-positive (D) cells in hippocampal regions CA1 and CA3 and the dentate gyrus (DG) (four mice per group). Black bars, control mice; white bars, fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice; ***, P < 0.0001 versus kainate wild-type mice.

Since increased excitability and impaired immediate early geneactivation are also observed in mice lacking endocannabinoidCB₁ receptors in principal forebrain neurons and since the G_q/G₁₁G proteins have been involved in endocannabinoid synthesis,we determined the levels of anandamide and 2-AG in the hippocampiof wild-type and fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-deficient mice (Fig. 5). Under basalconditions, anandamide levels were significantly lower in mutantmice than in controls (Fig. 5A), while 2-AG levels did not differsignificantly (Fig. 5B). The chemical convulsant KA enhancedboth anandamide and 2-AG levels in control mice (Fig. 5A and B).Interestingly, KA treatment still increased anandamide levelsin fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice but failed to induce 2-AG synthesis (Fig.5A and B). In mice with a single deficiency in G ${alpha}$ _q or G ${alpha}$ ₁₁, KA-inducedsynthesis of 2-AG was also impaired, but the reduction was lessprominent than in fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-deficient mice (see Fig. S1 in thesupplemental material). These data indicate that basal hippocampallevels of anandamide, but not of 2-AG, are under tonic controlof the G_q/G₁₁-mediated signaling pathway. In contrast, G_q/G₁₁-mediatedsignaling is necessary for KA-induced formation of 2-AG butnot of anandamide.

View larger version (16K):
[in this window]
[in a new window]
FIG. 5. Endocannabinoid levels in hippocampi of fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice. (A and B) Anandamide (AEA) (A) and 2-AG (B) levels were determined by isotope dilution liquid chromatography-mass spectrometry in hippocampi from control and fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice after intraperitoneal application of saline or 20 mg/kg KA (six to eight mice per group). **, P < 0.01; ***, P < 0.001 versus saline in wild-type mice; $§$ $§$ $§$ , P < 0.001 versus saline in fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice; ##, P < 0.01 versus KA in wild-type mice.

To further test whether the reduced formation of endocannabinoidsin the absence of Ga_q/G ${alpha}$ ₁₁ is involved in abnormal neuronal excitabilityand impaired neuroprotection, we pretreated mice with the CB₁receptor antagonist SR141716A. As expected, SR141716A stronglyaggravated KA-induced seizures in wild-type mice (Fig. 6A).However, in fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice, a blockade of CB₁ receptorsdid not further aggravate KA-induced seizures (Fig. 6A). Inthis experiment, KA was administered at the reduced dosage of10 mg/kg to detect the potential aggravation of seizure intensityin fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice as well, which have per se higher seizurescores than wild-type mice (see Fig. 3). Since SR141716A isa competitive antagonist at the CB₁ endocannabinoid receptor,the lack of SR141716A aggravation of KA-induced seizures infb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice supports our finding that endogenous levelsof endocannabinoids are significantly decreased in fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKOmice, at least at those synapses involved in CB₁-mediated neuroprotection.To enhance endocannabinoid levels selectively at sites of physiologicalproduction, we treated mice with the endocannabinoid reuptakeinhibitor OMDM-2 (54). OMDM-2 caused a slight but nonsignificantamelioration of seizures induced by 30 mg/kg KA in wild-typemice but a strong amelioration in fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice (Fig. 6B),suggesting that endocannabinoid-mediated effects are basicallyfunctional in mutant mice. To further substantiate this notion,we investigated cannabinoid-mediated retrograde inhibition ofsynaptic transmission by electrophysiological means. In boththe wild-type and the mutant hippocampal slices, the applicationof the CB₁ receptor agonist WIN55,212-2 caused a significantreduction in the amplitude of IPSCs in CA1 hippocampal neurons(see Fig. S2 in the supplemental material). Taken together,both pharmacological and electrophysiological data indicatethat endocannabinoid-mediated protective mechanisms are basicallyintact in fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice but that endogenous cannabinoidproduction is impaired.

View larger version (28K):
[in this window]
[in a new window]
FIG. 6. Pharmacological manipulation of the endocannabinoid system in wild-type and fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice. (A and B) Effect of the competitive CB₁ antagonist SR141716A (A) and the endocannabinoid reuptake inhibitor OMDM-2 (B) on kainic acid-induced seizures. The fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO (KO) and control (WT) mice received saline (Sal), 3 mg/kg SR141716A (SR), or 5 mg/kg OMDM-2 (OM) intraperitoneally 30 min before the application of kainic acid (at 10 [A] or 30 [B] mg/kg KA; six to eight mice per group). *, P < 0.05, saline-treated versus OMDM-2-treated fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice; ***, P < 0.001, saline-treated versus SR-treated wild-type mice.

DISCUSSION

Top

Discussion

In this study we have shown that genetic inactivation of theG_q/G₁₁-mediated signaling pathway in principal forebrain neuronsresults in increased neuronal excitability. This finding wasunexpected since the modulatory effects of G_q/G₁₁-mediated signalingare believed to facilitate rather than dampen neuronal activity.However, increased seizure susceptibility was also found inmice lacking PLC-ß1 (31), the inositol 1,4,5-trisphosphatereceptor (42), or the G_q/G₁₁-coupled 5-HT_2c receptor (60), suggestingthat the G_q/G₁₁-mediated signaling pathway indeed exerts importantinhibitory effects.

Only a few G_q/G₁₁-mediated effects that might promote neuronalinhibition have been described. G_q/G₁₁-coupled receptors arepresent on hippocampal interneurons (26, 64), and impaired activationof these cells might shift the balance between excitation andinhibition. However, since Cre-mediated recombination in theCamkcre4 mouse line is restricted to principal neurons (40),this model obviously does not apply here. Some G_q/G₁₁-coupledreceptors, e.g., the M₁ acetylcholine receptor, have been suggestedto negatively regulate N-type calcium channels through a breakdownof membrane phosphatidylinositol 4,5-biphosphate (29) and therebymight contribute to inhibition. However, given the fact thata majority of inhibitory presynaptic G protein-coupled receptors,such as adenosine A₁, ${alpha}$ ₂ adrenergic µ and ${delta}$ opioid, andGABA_B receptors, signal through G_i/o family G proteins (17,45), the contribution of G_q/G₁₁-coupled receptors to calciumchannel regulation seems to be rather small. Also, the endocannabinoidCB₁ receptor is known to mediate its presynaptic inhibitoryeffects through G_i/o family G proteins (27, 28, 47, 55). Itwas recently reported that the synthetic CB₁ agonist WIN55,212-2induces a calcium response in CB₁-transfected HEK293 cells ina G_q/G₁₁-dependent manner; however, natural CB₁ agonists suchas ${Delta}$ 9-tetrahydrocannabinol, 2-AG, and anandamide analogues didnot elicit such responses (36).

Though not directly coupled to CB₁, the G_q/G₁₁-mediated signalingpathway was suggested to contribute to endocannabinoid functionsthrough its role in endocannabinoid synthesis. We found thathippocampal levels of anandamide, but not of 2-AG, are reducedin the absence of G ${alpha}$ _q/G ${alpha}$ ₁₁, indicating that anandamide formationis under the tonic control of receptors coupled to these G proteins,while 2-AG formation is not. Muscarinic and purinergic receptorsare among the G protein-coupled receptors that might contributeto basal anandamide levels as they are able to induce anandamideformation in the absence of extracellular Ca²⁺ as well (61).On the other hand, in the formation of basal 2-AG levels aspreviously described, PLC-ß-independent biosyntheticroutes (4, 59) seem to predominate.

In contrast to basal 2-AG production, G_q/G₁₁-mediated signalingis indispensable for hippocampal 2-AG formation following KAapplication, and this might be due to the fact that DAG, a fundamentalprecursor of that pool of 2-AG that acts as an endocannabinoid(13, 15, 70), is mostly formed in depolarized neurons followingthe activation of PLC-ß (58). In contrast to 2-AGformation in response to KA, G_q and G₁₁ are not necessary forKA-induced anandamide formation. This observation supports theconcept that anandamide can be produced in response to eithermetabotropic (61) or ionotropic (15) receptor stimulation alone,provided that the intracellular Ca²⁺ concentration is increasedin order to induce the activation of Ca²⁺-sensitive biosynthesizingenzymes and the subsequent processing of non-PLC-ß-derivedbiosynthetic precursors (54). With respect to the differentialrole of the two endocannabinoids in control of excitability,it is important to note that in wild-type mice, KA-induced 2-AGconcentrations are at least 50-fold higher than the anandamideconcentrations, and the loss of KA stimulation therefore resultsin much lower total endocannabinoid concentrations in fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKOmice than in wild-type mice (a net increase of 0.06 plus 3.35or 3.41 nmol/g tissue in wild-type mice compared to a net decreaseof 0.12 to 0.3 or –0.18 nmol/g tissue in fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKOmice) (Fig. 5).

The phenotype of CB₁ receptor-deficient mice resembles thatof fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice in that both mouse strains have increasedseizure susceptibilities in response to KA and inabilities toinduce expression of immediate early genes encoding the c-fosand zif-268 transcription factors, which are believed to counteractexcitotoxic damage (40, 70). This supports the hypothesis thatimpaired endocannabinoid-mediated retrograde signaling underliesincreased neuronal excitability in fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKO mice. However,forebrain-specific CB₁-deficient mice seem to be less proneto spontaneous epilepsy, although seizures have been observedin these animals, especially during the stress of handling (B.Lutz, personal communication). This phenotypic difference mightbe explained by the presence of non-CB₁ (13, 16, 19, 56) receptorson hippocampal principal neurons, which contributes to endocannabinoid-mediatedneuroprotection. In addition, defects in other G_q/G₁₁-mediatedprocesses which occur independently of endocannabinoids andwhich under normal conditions decrease neuronal activity, likethe above-mentioned negative regulation of N-type Ca²⁺ channels,may contribute to the more-severe phenotype observed in fb-G ${alpha}$ _q/G ${alpha}$ ₁₁-DKOmice.

Taken together, our data suggest that the G_q/G₁₁-mediated signalingpathway is critically involved in endocannabinoid synthesisunder both basal and excitotoxic conditions. The phenotypicalsimilarities between forebrain-specific CB₁-deficient mice andpharmacological in vivo data indicate that impaired endocannabinoidformation is associated with increased seizure susceptibilityand impaired neuroprotection. However, defective endocannabinoidsynthesis may not account exclusively for the increased neuronalexcitation in the absence of G_q/G₁₁, and other G_q/G₁₁-mediatedinhibitory mechanisms may contribute to the phenotype.

ACKNOWLEDGMENTS

We thank Barbara Wallenwein, Anke Rogatzki, Marianne Hillesheim,and Marta Valenti for expert technical assistance and MariaPritsch, University of Heidelberg, for help with the statisticalanalysis.

This work was funded by the Volkswagenstiftung (to V.D.M.) andby the Collaborative Research Center 488 of the Deutsche Forschungsgemeinschaft(to S.O.).

FOOTNOTES

* Corresponding author. Mailing address for Vincenzo Di Marzo: Endocannabinoid Research Group, Istituto di Chimica Biomolecolare, Consiglio Nazionale delle Ricerche, Via Campi Flegrei 34, Comprensorio Olivetti, 80078 Pozzuoli (NA), Italy. Phone: 39-081-8675093. Fax: 39-081-8041770. E-mail: vdimarzo@icmib.na.cnr.it. Mailing address for Stefan Offermans: Institute of Pharmacology, University of Heidelberg, Im Neuenheimer Feld 366, 69120 Heidelberg, Germany. Phone: 49-6221-548246. Fax: 49-6221-548549. E-mail: stefan.offermans{at}pharma.uni-heidelberg.de.

Supplemental material for this article may be found at http://mcb.asm.org/.

Present address: N.V. Organon, Molenstraat 110, 5340 BH Oss,The Netherlands.

Present address: Pharmacology and Toxicology, University ofZurich, Winterthurerstr. 190, 8057 Zurich, Switzerland.

REFERENCES

Top