endocannabinoids produced upon action potential firing evoke a cl− current via type-2 cannabinoid...

ION CHANNELS, RECEPTORS AND TRANSPORTERS

Endocannabinoids produced upon action potential firing evokea Cl− current via type-2 cannabinoid receptors in the medialprefrontal cortex

Femke S. den Boon & Pascal Chameau & Kas Houthuijs & Simone Bolijn &

Nicolina Mastrangelo & Chris G. Kruse & Mauro Maccarrone & Wytse J. Wadman &

Taco R. Werkman

Received: 19 December 2013 /Revised: 6 March 2014 /Accepted: 12 March 2014# Springer-Verlag Berlin Heidelberg 2014

Abstract The functional presence of type-2 cannabinoid re-ceptors (CB2Rs) in layer II/III pyramidal neurons of the ratmedial prefrontal cortex (mPFC) was recently demonstrated.In the present study, we show that the application of theendocannabinoids (eCBs) 2-arachidonoylglycerol (2-AG)and methanandamide [a stable analog of the eCB anandamide(AEA)] can activate CB2Rs of mPFC layer II/III pyramidalneurons, which subsequently induces a Cl− current. In addi-tion, we show that action potential (AP) firing evoked by 20-Hz current injections results in an eCB-mediated opening ofCl− channels via CB2R activation. This AP-evoked synthesisof eCBs is dependent on the Ca2+ influx through N-typevoltage-gated calcium channels. Our results indicate that 2-AG is the main eCB involved in this process. Finally, wedemonstrate that under physiologically relevant intracellularCl− conditions, 20-Hz AP firing leads to a CB2R-dependentreduction in neuronal excitability. Altogether, our data

indicate that eCBs released upon action potential firing canmodulate, through CB2R activation, neuronal activity in themPFC. We discuss how this may be a mechanism to preventexcessive neuronal firing.

Keywords AEA . 2-AG .Ca2+-activatedCl− current . In vitrobrain slice preparation .Whole-cell recording

Introduction

The endocannabinoid (eCB) system comprises at least two Gprotein-coupled receptors (GPCRs), type 1 and 2 cannabinoidreceptors (CB1Rs and CB2Rs), various lipid endogenous can-nabinoids (eCBs), eCB transporters, and enzymes responsiblefor the synthesis and degradation of eCBs [26]. Cannabinoidreceptors are activated by phytocannabinoids found inmarijuana, synthetic cannabinoids, and eCBs, such as2-arachidonoylglycerol (2-AG) and anandamide (AEA).AEA was the first eCB to be isolated [13], and it is thoughtto be synthesized from its membrane precursor N-acylphosphatidylethanolamine (NAPE) by a specific phos-pholipase D (NAPE-PLD) and, for the most part, hydrolyzedby fatty acid amide hydrolase (FAAH). 2-AG, identified a fewyears later, is synthesized by diacylglycerol lipase α (DLG α)and β (DGL β) from DAG-containing phospholipids and isprincipally hydrolyzed by monoacylglycerol lipase (MGL)[15, 16, 39, 48]. The eCB system is involved in many cogni-tive functions that are associated with the medial prefrontalcortex (mPFC) [42]. The expression pattern and functionalrole of CB1Rs as presynaptic receptors that modulate neuro-transmitter release in the central nervous system (CNS) havebeen well described [8, 23, 27, 51]. However, there is rela-tively little knowledge about the functionality of CB2Rs in the

W. J. Wadman and T. R. Werkman share last authorship.

F. S. den Boon : P. Chameau :K. Houthuijs : S. Bolijn :C. G. Kruse :W. J. Wadman : T. R. Werkman (*)Center for Neuroscience, Swammerdam Institute for Life Sciences,University of Amsterdam, Science Park 904, 1098 XH Amsterdam,The Netherlandse-mail: [email protected]

N. MastrangeloDepartment of Experimental Medicine and Surgery,Tor Vergata University, 00133 Rome, Italy

N. Mastrangelo :M. MaccarroneEuropean Center for Brain Research (CERC)/Santa LuciaFoundation, 00179 Rome, Italy

M. MaccarroneCenter of Integrated Research, Campus Bio-MedicoUniversity of Rome, 00128 Rome, Italy

Pflugers Arch - Eur J PhysiolDOI 10.1007/s00424-014-1502-6

CNS. Initially, CB2Rs were not detected in the brain, but theirpresence in microglia cells, other brain immune cells, andneurons is now generally accepted [45, 50]. We have recentlyshown, using a combination of biochemical and electrophys-iological techniques, that CB2Rs are present intracellularly inlayer II/III pyramidal neurons of the rat mPFC [12].Furthermore, we demonstrated that their activation with syn-thetic, selective agonists leads to an IP3R-dependent openingof Ca2+-activated Cl− channels (CaCCs) located in the plasmamembrane. Under physiologically relevant intracellularCl− conditions, the opening of Cl− channels will stabi-lize the membrane potential around resting membranepotential, thereby reducing neuronal excitability [20,22]. Although we have shown that CB2Rs are involvedin the modulation of mPFC neuronal excitability, pre-sumably through the opening of CaCCs, these findingswere obtained with synthetic cannabinoids [12]. Thequestions how and when CB2Rs are activated by en-dogenously released eCBs in mPFC layer II/III neuronsremain to be answered. The downstream signaling path-way following CB2R activation suggests a role forCB2Rs in the modulation of neuronal excitability. Wehypothesize that in response to increased AP firing,mPFC layer II/III neurons synthesize and release eCBsto prevent excessive neuronal activity via a CB2R-mediatedfeedback system.

Endogenous ligands for cannabinoid receptors are synthe-sized ‘upon demand’ from membrane phospholipids follow-ing an increase in [Ca2+]i, and/or activation of GPCRs such asmetabotropic glutamate receptors or M1/M3 muscarinic ace-tylcholine receptors [26]. Aside from the ‘on demand’ pool,eCBs are believed to be present in a ‘basal pool’, beingconstantly produced, accumulated and released [2, 14, 38]. Itis known that adequate stimulation (10-50 Hz) can lead toeCB synthesis and subsequent CB1R-mediated plasticity [5,30]. Furthermore, in the mouse PFC, repetitive AP firing at anintermediate frequency (20 Hz) was also reported to evokeeCB synthesis and resulted in 2-AG-dependent CB1R-medi-ated depression of synaptic transmission [52]. Other studieshave also demonstrated CB1R-mediated effects in corticalareas following brief trains of APs [18, 31]. We propose thatstimulation protocols that are able to induce synthesis andrelease of eCBs which induce CB1R-mediated effects, couldalso lead to CB2R-mediated effects.

To assess the physiological role of CB2Rs, we proceeded toinvestigate the eCB-mediated activation of CB2Rs and thesubsequent opening of CaCCs, which in turn will modulateneuronal excitability. We demonstrate how the exogenousapplication of 2-AG and mAEA concentration dependentlyevokes a Cl− current in layer II/III pyramidal neurons of the ratmPFC. Furthermore, we reveal the activation of CB2Rs byeCBs following 20-Hz suprathreshold current injections, me-diated—at least for the most part—by 2-AG. Finally, we show

that 20-Hz AP firing leads to eCB production, which canreduce neuronal excitability through the activation of CB2Rs.

Materials and methods

Electrophysiology

Coronal slices (300 μm) of the mPFC were obtained frommale Wistar rats (Harlan, The Netherlands) aged 14–19 dayspostnatal andmale C57BL/6Wtmice ormale C57BL/6 CB2RKO mice (The Jackson Laboratory, USA) aged 14–19 dayspostnatal. Experiments were approved by the animal welfarecommittee of the University of Amsterdam. Animals werekilled by decapitation, their brains rapidly removed and placedin oxygenated (95 % O2–5 % CO2) ice-cold (4 °C) adaptedartificial cerebrospinal fluid (aACSF, Table 1). Slices were cutin aACSF on a vibratome (VT1200S, Leica, Germany) andplaced for 30 min in regular ACSF (Table 1) at 32 °C. Sliceswere kept at room temperature for at least 1 h prior to record-ing. Whole-cell current and voltage clamp recordings weremade at 32 °C from the soma of layer II/III pyramidal neurons.Glass recording pipettes were pulled from borosilicate glass(Science Products, Germany) and had a resistance of 2–3 MΩwhen filled with (high Cl−) pipette solution (Table 1). Thispipette solution was used to record Cl− currents and mem-brane depolarizations. A modified ACSF (mACSF) contain-ing K+-, Ca2+-, and Na+-channel blockers and a pipette solu-tion with high [Cl-]in (30 mM) and no K+ (Table 1) were usedin the experiments with 2-AG and mAEA. Another pipettesolution with low [Cl−]in (8.75 mM; Table 1) was used forrecording Cl− currents following synaptic stimulation and fordetecting effects on the firing rate.

For recordings of Ca2+ currents carried by voltage-dependent Ca2+ channels (VDCCs), a VDCC ACSF and aVDCC pipette solution were used (Table 1). Ca2+ currentswere evoked by 200-ms voltage steps (2-s interval) from ahyperpolarizing prepulse potential (400 ms, −120 mV) to testpotentials ranging from –90 to 60 mV. Sustained Ca2+ currentamplitudes were measured as the average current during thelast 30 ms of the current.

To allow for the detection of CB2R-mediated Cl− currents,currents were evoked in the voltage clamp configuration every15 s by series of rectangular voltage steps (100-ms duration;Fig. 1a) from a holding potential of –80 mV to voltagepotentials ranging from −90 to −20 mV (Figs. 1 and 2b, c)or −105 to −55 mV (Fig. 2d) in 10-mV increments. CB2R-mediated currents were estimated by evoking membrane cur-rents with the rectangular voltage steps before and after bathapplication of cannabinoid ligands (Fig. 1) or a stimulationprotocol (in the current clamp configuration) that evoked APs(Fig. 2a, see below).

Pflugers Arch - Eur J Physiol

Due to difficulties with washing out bath-appliedcannabinoid ligands, every cell was treated with oneconcentration. Concentration–response curves for canna-binoid ligands were fitted with the following logisticequation:

Current density ¼ Maximal current density

1 þ EC50= Ligand½ � ð1Þ

Under current clamp conditions, a stimulation protocol wasused to evoke APs in layer II/III neurons with 60suprathreshold current injections (1 nA, 5 ms duration, 45-ms interval) into the soma via the patch pipette (Fig 2a). Forexperiments with synaptically evoked APs and/or EPSPs,currents ranging from 300 to 600 μA (0.1 ms) were applied(at 20 Hz) via a separate stimulation electrode that was placedin layer I (Fig. 2d, inset). For the current clamp experiments, aslow feedback system was used that guaranteed that currentclamp recordings started at a defined membrane voltage of–80 mV (Fig. 3) or −70 mV (Fig. 4). eCB-mediated effects oncell excitability were performed by injecting frozen filtered

Gaussian noise (time constant=10 ms) via the patch pipettewith a variance adjusted for each neuron to result in a meanspike frequency of ~0.95 Hz. Firing frequency was calculatedusing 1-min bins. Prior to the stimulation protocol, 5-minrecordings were used as control. Maximal effects of 20-HzAP firing were observed within 2–6 min after the stimulationprotocol was applied.

Recordings were made using an Axopatch 200b (Axon,USA) and in-house software running under Matlab(MathWorks, USA). Signals were filtered at 2.9 kHz andsampled at 10 kHz. Series resistance ranged from 5 to 15 MΩand was compensated to ~65 %. Signals were corrected forliquid junction potential. Current densities were calculatedusing cell capacitance (read from the amplifier dials) andexpressed in pA/pF.

Data analysis

Data were compared with t tests. When data was notnormally distributed, a Mann–Whitney test was used. In

Table 1 Composition of extra-cellular and pipette solutions

Concentrations are in mM, unlessstated otherwise. If relevant, theCl− concentrations are given. It isindicated in which figures the(combinations of the) solutionswere used. Extracellular and pi-pette solutions had a pH of 7.4 or7.3, respectively. The pH of thepipette solutions was set withKOH or CsOH (for the solutionwithout K+ )

ACSF artificial cerebral spinal flu-id, aACSF adapted ACSF,mACSF modified ACSF, 4-AP4-aminopyridine, BAPTA 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid,EGTA ethylene glycol-bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid, HEPES 104 - ( 2 - h y d r o x y e t h y l ) - 1 -piperazineethanesulfonic acid,TEA tetraethyl ammonium,TTX tetrodotoxin, VDCC voltage-dependent Ca2+ channel

Extracellular solutions Pipette solutions

Solution → aACSF ACSF mACSF VDCC

ACSF

High Cl− High Cl−

no K+

Low Cl− VDCC

[Cl−] (mM) 128.5 131.7 30 30 8.75

Figure 2, 3, 4 1 2c inset 2a–c, 3 1 2d, 4 2c inset

Concentration

Choline-Cl 120

NaCl 120 98 105

KCl 3.5 3.5 3.5 3.5

CaCl2 0.5 2.5 2.5 2.5

MgSO4 6 1.3 1.3 1.3

NaH2PO4 1.25 1.25 1.25 1.25

NaHCO3 25 25 25 25

D-glucose 25 25 25 25

TEA-Cl 25 10

4-AP 5 5

CdCl2 0.2

TTX 0.5 μM 0.5 μM

K-Gluconate 110 131.25

KCl 30 8.75

CsCl 30

CsMeSO4 135

CH3O3SCs 110

HEPES 10 10 10

BAPTA 10

EGTA 0.5 0.5 0.5

Mg-ATP 4 4 4 4

Na-GTP 0.5 0.5 0.5 0.5


the figures, the significance is indicated with asterisks(*P<0.05, **P<0.01, and ***P<0.001), and the number ofobservations are given between brackets (Figs. 1 and 2)or in the bar charts (Figs. 3 and 4). Results are given asmean±standard error of the mean.

Drugs

mAEA, 2-AG, JZL184, RHC-80267, and ω-conotoxinwere purchased from Tocris (The Netherlands).Sch.356036 and JWH-133 were generous gifts from

Fig. 1 2-AG and mAEA induce a Cl− current in layer II/III mPFCpyramidal neurons. a Representative traces of voltage steps (a1), evokingcurrents during baseline (a2), and currents including Cl− currents duringcannabinoid–ligand application (a3); the isolated Cl− currents (a4) areobtained by subtracting the traces of a2 from the traces of a3. b I/Vrelationships of 2-AG- and mAEA-induced Cl− currents at their maximaleffective concentrations (30 and 10 μM, respectively). Erev for 2-AG(−38.7±1 mV) and mAEA (−42.2±2.4 mV) is close to ECl−(–38.9 mV). When 2-AG (30 μM) and mAEA (10 μM) were appliedsimultaneously, current densities were not different than for mAEA(10 μM) alone (Erev=−40.4±1.2 mV). In the presence of the CB2Rantagonist Sch.356036 (5 μM), 2-AG (30 μM) evoked a much smaller

current. Application of the CB2R-selective agonist JWH-133 (1 μM)induced a current (Erev=–41.1±2.1 mV) that was comparable to thecurrent induced by 2-AG (30 μM). c Concentration–response curvesfor 2-AG and mAEA, constructed from the current densities recordedat –90 mV (4–10 cells per concentration). The data points were fittedwith Eq. 1. The estimated maximal current density at −90 mV waslarger for 2-AG (3.1±0.3 pA/pF, 30 μM) than for mAEA (1.3±0.3 pA/pF, 10 μM) (P<0.01). The estimated EC50 value for mAEA(1.2±1 μM) was smaller than that for 2-AG (4.9±1.4 μM) (P<0.01).In the presence of the CB2R antagonist Sch.356036 (5 μM), a largelyreduced current (recorded at −90 mV) was evoked with 30 μM 2-AG(Mann–Whitney, P<0.05)


Abbott Healthcare Laboratories (The Netherlands). 4,4′-Diisothiocyanatostilbene-2,2′-disulfonic acid disodiumsalt hydrate (DIDS) and nifedipine were purchased fromSigma-Aldrich (The Netherlands); tetrodotoxin (TTX)was purchased from Latoxan (France). URB602 andURB597 were purchased from Cayman Chemical

(USA). Most cannabinoid ligands were dissolved indimethyl sulfoxide (DMSO) or ethanol to 50 mM, withthe exception of URB602 and RHC-80267 (dissolved to100 mM). Stock solutions were diluted in ACSF with afinal concentration of DMSO or ethanol that was alwayslower than 0.1 %. For experiments with drugs dissolved

Fig. 2 60 APs at 20 Hz result in a CB2R-mediated Cl− current in layer II/III mPFC neurons. a Cropped traces of 60 evoked APs (a1) and injectedcurrents (a2) at 20 Hz. b I/V relationships showing that AP firingevoked with 60 current injections at 20 Hz induced a CB2R-mediatedCl− current (determined as described in Fig. 1a), Erev=–42.4±2.3 mV(ECl−=–38.3 mV). Such current injections could not evoke a currentafter preincubation with and in the continuous presence of the CB2Rantagonist Sch.356036 (5 μM). After preincubation with and in thecontinuous presence of the CB1R antagonist rimonabant (5 μM), asimilar current, Erev=–38.7±4.9 mV, could be detected following 60current injections at 20 Hz. c The N-type VDCC blocker ω-conotoxin,but not the L-type blocker nifedipine, markedly reduced theamplitude of CB2R-mediated Cl− currents evoked with 60 currentinjections at 20 Hz. Erev for currents recorded with ω-conotoxin

(–41.4±2.3 mV) and with nifedipine (–42.9±1.4 mV) are close to ECl−(−38.3 mV). The current in the absence of VDCC blockers is the same asthe control current depicted in Fig. 2b (closed circles). Inset shows the I/Vrelationship for Ca2+ currents recorded in the absence and presence ofω-conotoxin (0.25 μM) and nifedipine (10 μM). VDCC-mediatedcurrent densities were clearly reduced in the presence of either blocker.d I/V relationship of CB2R-mediated current (determined as described inFig. 1a) following 60 synaptically evoked APs at 20 Hz reversed at −71.7±1.2 mV, which is close to ECl− (–70 mV). The Cl− current could not beobserved if no APs were evoked or after the preincubation with and in thecontinuous presence of Sch.356036 (5 μM). Inset shows a biocytin-filledlayer II/III neuron and the schematic representation of a stimulationelectrode in layer I and a recording electrode for registering APs or EPSPsin layer II/III neurons


in DMSO or ethanol, the same DMSO or ethanol con-centration was present under control conditions. Somecells were filled with biocytin (Sigma, 1 mg/ml dis-solved in pipette solution) during whole-cell patch re-cordings. Biocytin was visualized using previously de-scribed methods [9].

Results

2-AG and mAEA induce a Cl− current

To investigate whether eCBs can induce the previously de-scribed Ca2+-activated Cl− current [12], we first performed

Fig. 3 Stimulation with 60 AP-evoking current injections at 20 Hzinduces eCB synthesis and subsequent CB2R activation in layer II/IIImPFC pyramidal neurons. a Representative trace recorded from a neuronthat depolarizes following 60 current injections at 20 Hz (arrow); thedepolarization generated AP firing. b In current clamp recordings withECl=–38.3 mV, 60 current injections at 20 Hz resulted in a depolarization(24.3±2.2 mV), which was reduced after preincubation with and in thecontinuous presence of the CB2R antagonist Sch.356036 (5 μM) (2.2±0.8 mV, P<0.001) and in the presence of the Cl− channel blocker DIDS(10.8±1.3 mV, P<0.001). The same stimulation protocol depolarizedlayer II/III mPFC neurons of WT mice, but not of CB2R KO mice(16.8±2.4 mV and 1±1.1 mV, respectively, P<0.001). c The magnitudeof the depolarization following 60 current injections at 20 Hz was not

different in the presence of the MGL inhibitors URB602 (100 μM) (17.7±2.4 mV) and JZL184 (100 nM) (16.7±2.4 mV) or the FAAH inhibitorURB597 (1 μM) (24.6±2.6 mV). Incubation with the DGL inhibitorRHC-80267 (100 μM) reduced the depolarization (5.6±2.6 mV, Mann–Whitney, P<0.001). The average control depolarization following 60current injections at 20 Hz is the same as the one depicted in Fig. 3b. dThe duration of the depolarization was increased—compared to the controlstimulation—in the presence of URB602 (245±68 s and 550±109 s,respectively, Mann–Whitney, P<0.05) and JZL184 (500±111 s, Mann–Whitney, P<0.05), but not with URB597 (289±71 s). e The delay afterstimulation until depolarization—compared to control—was not differentin the presence of either URB602, JZL184, or URB597 (147±45 s, 211±64 s, 250±57 s, and 132±35 s, respectively)


whole-cell voltage clamp recordings of layer II/III pyramidalneurons. These experiments were done in the presence ofK+, Ca2+, and Na+ channel blockers (mACSF) and with apipette solution containing a high Cl− concentration andno K+ (Table 1). Under these recording conditions, Cl− isthe predominant ion carrying membrane currents, also atrest (Fig. 1a2). Using a series of step potentials (Fig. 1a)before and during the bath application of 2-AG ormAEA (the stable analog of AEA) allowed the construc-tion of a current–voltage relationship that reversed (Erev)close to the calculated reversal potential for Cl− (ECl−)(Fig. 1b). With 2-AG, in 25 out of 35 recorded cells aCl− current was evoked; with mAEA, in 24 out of 33cells, a Cl− current was detected. In the responding cells,an effect was usually observed 2–6 min after eCB appli-cation. Both 2-AG (0.1-50 μM) and mAEA (1-30 μM)concentration dependently evoked a Cl− current. Thecurrent density values determined at –90 mV (at whichvoltage the largest current amplitudes were evoked) wereplotted as a function of ligand concentration and fittedwith a logistic equation (Eq. 1), yielding EC50 values of1.2±1.0 μM and 4.9±1.4 μM for mAEA and 2-AG,respectively (Fig. 1c).

2-AG- and mAEA-mediated effects occur through CB2Ractivation

The maximal current density recorded at –90 mV for 2-AGexceeded that for mAEA (Fig. 1b, c), which demonstratesthat 2-AG is a more efficacious agonist than mAEA.However, the estimated EC50 values indicate that mAEAis a slightly more potent agonist than 2-AG (Fig. 1c). When2-AG (30 μM) and mAEA (10 μM) were applied simulta-neously, the maximal current density was not different fromthat evoked by mAEA (10 μM) alone, which suggests thatboth ligands compete for the same binding site (Fig. 1b).When 2-AG (30 μM) was tested after a preincubation (for atleast 10 min) and in the continuous presence of the selectiveCB2R antagonist Sch.356036 (5 μM), the current was large-ly reduced (measured 5 min after 2-AG application,Fig. 1b, c), indicating that the 2-AG-evoked current ismediated by CB2R activation. The selective CB2R agonistJWH-133 (1 μM) induced a Cl− current that was comparableto that elicited by 2-AG (Fig. 1b), which also points to aCB2R-mediated mechanism evoking the Cl− current.Together, these experiments show that 2-AG and mAEAcan induce a Cl− conductance, presumably via the activation

Fig. 4 Endogenous CB2R activation decreases firing activity of ratmPFC layer II/III pyramidal neurons. a Stimulation with 60 currentinjections at 20 Hz led to a reduction of the neuronal firing rate. Firingwas induced by a Gaussian current input into the soma. Meanbaseline firing frequency of 0.95±0.09 Hz was normalized overslices to 100±3 % and reduced to 72±6 % 2–6 min after 60current injections at 20 Hz (P<0.01). After preincubation with

and in the continuous presence of the CB2R antagonistSch.356036 (5 μM) (baseline firing frequency 0.93±0.08 Hz,normalized to 100±2 %), 60 current injections at 20 Hz couldnot induce a response (99±9 %). b Representative traces ofcurrent clamp recordings showing AP firing of layer II/III neuronsbefore and after 60 current injections at 20 Hz, in the absenceand presence of Sch.356036 (5 μM)


of CB2Rs, and that 2-AG is a more efficacious, yet slightlyless potent agonist for the CB2R.

AP firing at 20 Hz results in a CB2R-mediated Cl− current

We hypothesized that a stimulation protocol that induceseCB synthesis and CB1R activation may also lead toCB2R activation. We tested whether 60 APs fired at20 Hz (Fig. 2a) could lead to a CB2R-mediated activationof the Cl− current. After stimulation (in the current clampconfiguration), we switched to the voltage clamp configu-ration and recorded currents evoked by 60 APs firing at20 Hz in 10 out of 14 tested neurons. These experimentswere performed with the high Cl− pipette solution(Table 1), and the evoked currents reversed close to ECl−

(Fig. 2b). The stimulation protocol did not evoke a currentwhen neurons were preincubated (for at least 10 min) withand in the continuous presence of the CB2R antagonistSch.356036, which indicates that the current is evokedfollowing CB2R activation (Fig. 2b). To exclude the in-volvement of CB1Rs, we repeated the experiment in thepresence of the selective CB1R antagonist rimonabant(5 μM) and detected a current that was comparable tothe one recorded in the absence of rimonabant (Fig. 2b).In a separate experiment (data not shown), we decreased[Cl−]i to 8.75 mM (ECl−=–70.7 mV) and observed that thereversal potential of the current (Erev, –68.8±1.6 mV, n=7)shifted toward the newly established ECl−. This experimentconfirmed the identity of the CB2R-mediated current as aCl− current.

AP firing at 20 Hz results in a Ca2+ influx through N-typeVDCCs leading to the CB2R-mediated activationof the Cl− current

Since the synthesis of eCBs is sensitive to the influx of Ca2+

[14], we next examined the possible involvement of N- and L-type voltage-dependent Ca2+ channels (VDCCs). VDCC-mediated currents were recorded in the absence and presenceofω-conotoxin (0.25 μM) or nifedipine (10 μM) to block N-or L-type VDCCs, respectively. Ca2+ current densities wereclearly reduced in the presence of either blocker, which showsthe presence of these two channels in layer II/III pyramidalneurons (Fig. 2c, inset). In the presence of the N-type VDCCblocker ω-conotoxin (0.25 μM), 60 current injections at20 Hz resulted in a markedly reduced Cl− current (Fig. 2c).When the L-type VDCC blocker nifedipine (10 μM) waspresent, the Cl− current evoked with the stimulation protocolwas not affected (Fig. 2c). This indicates that Ca2+ influxthrough N-type VDCCs following AP spiking at 20 Hz resultsin a CB2R-mediated activation of the Cl− current.

Synaptically evoked AP firing induces the CB2R-mediatedactivation of the Cl− current

We next applied synaptic stimulation to evoke APs, instead ofcurrent injections into the soma. These experiments wereperformed with normal ACSF and with a pipette solutioncontaining a physiologically relevant Cl− concentration([Cl−]i=8.75 mM) (Table 1). With electrical synaptic stimula-tion in layer I, excitatory postsynaptic potentials (EPSPs) largeenough to induce APs could be generated in layer II/IIIneurons. Following 60 synaptically evoked APs at 20 Hz,we recorded a current that reversed close to ECl− (Fig. 2d).This current was not induced by synaptic stimulation if onlyEPSPs were evoked (Fig. 2d), indicating that AP firing in-duced the response. If APs were evoked, but recordings wereperformed after preincubation with (for at least 10 min) and inthe continuous presence of the CB2R antagonist Sch.356036(5 μM), we did not detect the current (Fig. 2d). These resultsshow that CB2R activation (ultimately leading to the openingof CaCCs) depends on AP firing.

At high intracellular Cl− concentrations, AP firing at 20 Hzevokes a CB2R-mediated depolarization

To elucidate whether 2-AG and/or AEA are involved in theactivation of CB2Rs following the firing of 60 APs at 20 Hz,we performed experiments in the current clamp configuration.In this configuration, at an experimental membrane potentialof –80 mVand with [Cl−]out=128.5 mM and [Cl−]in=30 mM(ECl-=–38.3 mV) (see Table 1), opening of CaCCs resulted ina transient depolarization of the neurons (Fig. 3a). Theseexperimental conditions allow a more precise determinationof the delay, duration, and amplitude of the depolarization ofCB2R-mediated responses and are similar to previous exper-iments with synthetic CB2R ligands [12]. In a first set ofexperiments, we investigated whether the depolarizationcould be evoked repeatedly. In cells where the stimulationprotocol evoked a depolarization, another depolarizationcould be induced a second time (25-min interval betweenstimulations; first depolarization, 25.7±11.1 mV; second de-polarization, 23.7±5.2 mV; n=6, data not shown). The secondstimulation (after 25 min) did not induce a depolarization(P<0.05) when after the first stimulation—evoking a depo-larization of 19.2±7.1 mV (n=5, data not shown)—the sliceswere perfused with the selective CB2R antagonist Sch.356036(5μM). The transient depolarization of the membrane, evokedby the 20-Hz current injection protocol, was also absent inneurons that were preincubated (for at least 10 min) with andin the continuous presence of Sch.356036 (5 μM, Fig. 3b). Tocorroborate the involvement of Cl− channels, we performed anexperiment in the presence of the Cl− channel blocker DIDS(0.2 mM) and observed a reduced depolarization of the mem-brane potential (~60 % reduction, Fig. 3b). Finally, we


confirmed the pharmacological evidence that the depolariza-tion is mediated via CB2R activation using brain slices fromCB2R knockout (KO) and wild-type (WT) mice. Following60 suprathreshold current injections at 20 Hz, we observed adepolarization of WT layer II/III mPFC neurons, but not ofneurons recorded in slices of KO animals (Fig. 3b).

2-AG is the main eCB involved in CB2R-mediated effectsfollowing AP firing at 20 Hz

In order to differentiate between the involvement of 2-AG andAEA, we blocked hydrolysis of these eCBs by their mostimportant catabolic enzymes, MGL and FAAH, respectively.We used URB602 and JZL184 to inhibit MGL activity andURB597 to inhibit FAAH activity [10, 17, 25, 29, 41]. In thepresence of either the MGL inhibitors URB602 (100 μM) andJZL184 (100 nM) or the FAAH inhibitor URB597 (1 μM), 60APs at 20 Hz resulted in depolarizations that were not differ-ent in amplitude from control recordings (Fig. 3c). However,when the MGL inhibitors URB602 or JZL184 were present,the duration of the depolarization was larger than in controlconditions, whereas the FAAH inhibitor URB597 did notaffect the duration of the response (Fig. 3d). This suggeststhat the stimulation protocol mainly produces 2-AG (and notAEA), which activates CB2Rs to induce the depolarization.Under control conditions, the depolarization occurred with adelay of ~150 s after 60 current injections at 20 Hz. Theduration of this delay was in the same range as observed inexperiments with a selective CB2R agonist present in therecording pipette [12]. Delays were not different when neu-rons were tested in the presence of either URB602, JZL184, orURB597 (Fig. 3e). We further implicated 2-AG as the maineCB involved in the CB2R-mediated depolarization by show-ing a decreased depolarization in the presence of the diacyl-glycerol lipase inhibitor RHC-80267 (100 μM), which isknown to disrupt 2-AG synthesis [43] (Fig. 3c). Taken togeth-er, these data indicate that 2-AG is the main eCB involved inCB2R activation following 60 APs at 20 Hz.

CB2R activation following AP firing at 20 Hz decreasesneuronal excitability

We have shown that activation of CB2Rs with the syntheticagonist JWH-133 reduces neuronal excitability under physio-logically relevant intracellular Cl− conditions (ECl−=–70 mV)[12], most likely through the opening of CaCCs. We nexttested whether endogenous activation of CB2Rs has a similareffect in a series of current clamp experiments. In theseexperiments, the pipette solution with [Cl-]in=8.75 mM wasused (Table 1). Neuronal firing of layer II/III mPFC neuronswas evoked by current injection with Gaussian noise, whichled to fluctuations around resting membrane potential,resulting in the occasional firing of APs (Fig. 4b). For each

neuron, the variance of the injected current was adjusted tocause a stable firing rate of ~0.95 Hz. Two to 6 min afterstimulation with 60 APs at 20 Hz, the normalized firing ratewas transiently reduced by ~30 %. This reduction could beprevented by preincubation (for at least 10 min) with and inthe continuous presence of the CB2R antagonist Sch.356036(5 μM) (Fig. 4). These results indicate that, in mPFC neurons,CB2R activation—following 60 APs at 20 Hz—modulatesregular AP firing evoked with an input that could resemblespontaneous background synaptic activity.

Discussion

We have previously demonstrated that the activation of intra-cellular CB2Rs by synthetic cannabinoid ligands in layer II/IIIpyramidal neurons of the mPFC leads to the IP3R-dependentopening of CaCCs [12]. Here, we show that bath applicationof 2-AG and mAEA (the stable analog of AEA) can, concen-tration dependently, induce a similar Cl− current, albeit that2-AG induced a larger current than mAEA. The 2-AG-evokedCl− current could partly be blocked by the selective CB2Rantagonist Sch.356036. Furthermore, 2-AG and mAEAevoked a Cl− current that was comparable to the one evokedby the selective CB2R agonist JWH-133, suggesting that theeffects seen with 2-AG and mAEA are the result of activationof CB2Rs. The estimated EC50 values indicate that 2-AG,although more efficacious, has a slightly lower affinity forCB2Rs than mAEA. Future studies on the binding of eCBs atthe CB2R in the rodent mPFC could show how 2-AG andAEA behave at their binding sites. When both agonists arebath applied simultaneously at their respective maximal effec-tive concentrations, the resulting current density does notdiffer from experiments in which mAEA was applied alone.These data suggest that 2-AG and mAEA compete for thesame CB2R binding site and that they are able to activate acommon downstream signaling cascade. This is in line withearlier findings that AEA can attenuate the agonistic activityof 2-AG [21, 47]. 2-AG and AEA are considered as the twomajor eCBs, and they have been shown to be involved inseveral neuronal processes via abundantly expressed CB1Rs[26, 30, 46, 49, 51]. Despite difficulties with the precisequantification of brain eCB levels, it is clear that 2-AG levelsexceed those for AEA [6, 46]. This also holds for the mPFCslice preparation used in our study, where we found that 2-AGlevels are ~800-fold higher than those of AEA (unpublishedresults). Since the discovery of 2-AG and AEA and thedevelopment of selective pharmacological tools and transgen-ic mice strains, the roles of 2-AG and AEA have been increas-ingly disentangled [36]. So far, collected data suggests that2-AG is the principal eCB that modulates synaptic plasticity,with a smaller role for AEA [26, 28]. In the PFC specifically,2-AG, but not AEA, was reported to mediate long-term


depression (LTD) and depolarization-induced suppression ofinhibition (DSI) [30, 52]. However, AEA has also been shownto be involved in some forms of synaptic plasticity in therodent PFC [35]. In this study, we confirm earlier findingsthat 2-AG is a full agonist at the CB2R, whereas mAEAbehaves as a partial agonist [21, 47]. This means that CB2R-mediated effects depend on the local concentrations of both2-AG and AEA.

The synthesis and release of eCBs that lead to CB1R-mediated effects are described to follow a wide range ofstimulation protocols, such as voltage steps and AP firing [5,30, 34]. In particular, eCBs can be synthesized upon Ca2+

entry through VDCCs after repetitive firing of the postsynap-tic cell and then travel to presynaptic CB1Rs. Alternatively,activation of GPCRs coupled to Gq/11, such as group ImGluRs and M1/M3 muscarinic receptors, can enhance orby itself lead to eCB production [26]. So far, research hasfocused on CB1R-mediated effects following eCB synthesis.We hypothesized that similar protocols that lead to the pro-duction of eCBs to activate CB1Rs activation may also resultin CB2R activation. We tested this by administering 60 AP-evoking current injections at 20 Hz that were described to leadto CB1R-mediated effects in the mouse PFC [52]. PFC neu-rons have been described to fire APs at frequencies in the 10–20 Hz range [4], making this a physiologically relevant stim-ulation paradigm. We demonstrated that with this stimulationprotocol CB2Rs are activated and that this induced a Cl−

conductance. To show that CB2Rs mediated this effect, weused the selective CB2R antagonist Sch.356036 [37, 44] toblock the Cl− current following 60 APs at 20 Hz. We excludedthe involvement of CB1Rs by performing similar experimentsin which we observed the Cl− current in the presence of theselective CB1R antagonist rimonabant. Both the Erev of theCB2R-mediated current and the reduction in the membranedepolarization in the presence of the Cl− channel blockerDIDS [1, 24] identified this current as a Cl− current. In thepresent study, we did not find evidence for the modulation ofother ion channels following CB2R activation. This is incontrast with a recent study, which reported a CB2R-mediatedreduction of neurotransmission in cultured autaptic hippocam-pal neurons of CB1R null mice [3]. The authors determinedthat the CB2R exerted its function from a presynaptic locali-zation, presumably through the inhibition of VDCCs. It ispossible that this discrepancy can be explained by differencesin preparation and brain region.

Here, we show that blocking N-type VDCCs reduced theamplitude of the Cl− current elicited with our stimulationprotocol (repetitive neuronal firing), whereas blockingL-type VDCCs did not. This indicates that the Ca2+ influxthrough N-type VDCCs following 20-Hz AP firing mainlyinduces eCB synthesis. It has to be kept in mind that, since itwas not possible to measure eCB levels following stimulation,we do not provide direct evidence that eCB synthesis was

increased. However, interfering with eCB synthesis and deg-radation modulated the responses, indicating that (newlysynthetized) eCBs do play a role in the CB2R-mediated re-sponses. Blocking 2-AG synthesis reduced the amplitude ofthe CB2R-mediated effect and inhibiting its degradationprolonged it. This suggests that due to a reduction in thedegradation of 2-AG, local 2-AG levels are elevated for alonger period of time, resulting in a longer lasting depolariza-tion of the membrane potential. In addition, this points to aCB2R-evoked effect that is most likely mediated by 2-AG.Interfering with AEA degradation did not affect the response.However, a minor role for AEA cannot be excluded underthese experimental conditions. Although the main route ofAEA synthesis involves the conversion of NAPE into AEA,alternative pathways have been described [27, 33, 36]. A lackof specific pharmacological tools to inhibit AEA synthesismakes it difficult to exclude AEA from playing a role inCB2R-mediated effects. In the striatum, both 2-AG andAEA were reported to induce eCB-mediated LTD throughCB1Rs [32]. Interestingly, different stimulation protocols,high- and low-frequency stimulation, induced the synthesisof AEA and 2-AG, respectively. This is in line with the notionthat 2-AG and AEA play specialized roles in the modulationof synaptic plasticity [36].

In a previous study, we observed a reduction in neuronalexcitability of layer II/III mPFC neurons when CB2Rs wereactivated with a selective, synthetic cannabinoid [12]. In thepresent study, we demonstrate that CB2R activation followingeCB release (evoked with 20-Hz AP firing) results, with adelay of 2–6 min, in a decrease in neuronal firing rate. Thispresumably occurs via the opening of CaCCs. If ECl− is at orclose to the resting membrane potential of neurons (as was thecase in these experiments), the opening of CaCCs followingactivation of CB2Rs will stabilize or even clamp the mem-brane potential around that value. These findings suggest that,in the mPFC, the eCB system plays a slowmodulatory role viathe activation of CB2Rs, conceivably preventing excessivefiring after an initial brief period of high activity. We havepreviously shown that the application of the CB2R antagonistSch.356036 increased the regular firing rate in a similar ex-periment, which points to a basal pool of eCBs and/or consti-tutive activity of CB2Rs [2, 12, 14]. The activity-induced self-inhibition via CB2Rs that we demonstrate here is likely todepend on the on-demand pool. The time scale of the onset ofthis feedback mechanism (minutes) is different from well-described and generally much faster modulatory eCB-mediated processes via presynaptic CB1Rs (seconds), suchas depolarization-induced suppression of excitation (DSE)and DSI. The duration of the delay before the effect occurredwas comparable to that observed in experiments with a syn-thetic CB2R agonist applied via the recording pipette [12].This suggests that CB2R ligands need some time to get accessto their binding sites. It is possible that the intracellular


location of the CB2Rs [12], which was also described in other(peripheral and CNS) tissues [7, 11, 40], contributes to thisdelay. We speculate that the relative slow onset of the CB2R-mediated response involves—after eCB synthesis has beeninduced by increased neuronal activity—pharmacokineticprocesses like sequestration, metabolism, diffusion, and mem-brane penetration of eCBs to the intracellularly located CB2Rs[19, 38]. In the case of whole-cell experiments, the recordingconditions (dilution of intracellular eCBs due to dialysis of thecytosol by the pipette solution) may also be a factor influenc-ing the time course of CB2R-mediated effects. Nevertheless,both cannabinoid receptors can modulate neuronal activity inthe mPFC, albeit through different mechanisms and on differ-ent time scales.

Acknowledgments This work was supported by Dutch Top InstitutePharma (grant number T5-107-1). M.M. was partly supported byFondazione Italiana Sclerosi Multipla (grant 2010) and by Ministerodell’Istruzione, dell’Università e della Ricerca (grant PRIN 2011–2012).

References

1. Akasu T, Nishimura T, Tokimasa T (1990) Calcium-dependent chlo-ride current in neurones of the rabbit pelvic parasympathetic ganglia.J Physiol 422:303–320

2. Alger BE, Kim J (2011) Supply and demand for endocannabinoids.Trends Neurosci 34(6):304–315

3. Atwood BK, Straiker A, Mackie K (2012) CB(2) cannabinoid recep-tors inhibit synaptic transmission when expressed in cultured autapticneurons. Neuropharmacology 63(4):514–523

4. Bandyopadhyay S, Hablitz JJ (2007) Dopaminergic modulation oflocal network activity in rat prefrontal cortex. J Neurophysiol 97(6):4120–4128

5. Brown SP, Brenowitz SD, Regehr WG (2003) Brief presynapticbursts evoke synapse-specific retrograde inhibition mediated by en-dogenous cannabinoids. Nat Neurosci 6(10):1048–1057

6. Buczynski MW, Parsons LH (2010) Quantification of brainendocannabinoid levels: methods, interpretations and pitfalls. Br JPharmacol 160(3):423–442

7. Castaneda JT, Harui A, Kiertscher SM, Roth JD, Roth MD (2013)Differential expression of intracellular and extracellular CB(2) can-nabinoid receptor protein by human peripheral blood leukocytes. JNeuroImmune Pharmacol 8(1):323–332

8. Castillo PE, Younts TJ, Chavez AE, Hashimotodani Y (2012)Endocannabinoid signaling and synaptic function. Neuron 76(1):70–81

9. Chameau P, Inta D, Vitalis T, Monyer H, WadmanWJ, van Hooft JA(2009) The N-terminal region of reelin regulates postnatal dendriticmaturation of cortical pyramidal neurons. Proc Natl Acad Sci U S A106(17):7227–7232

10. Chavez AE, Chiu CQ, Castillo PE (2010) TRPV1 activation byendogenous anandamide triggers postsynaptic long-term depressionin dentate gyrus. Nat Neurosci 13(12):1511–1518

11. Currie S, Rainbow RD, Ewart MA, Kitson S, Pliego EH, Kane KA,McCarron JG (2008) IP(3)R-mediated Ca(2+) release is modulatedby anandamide in isolated cardiac nuclei. J Mol Cell Cardiol 45(6):804–811

12. den Boon FS, Chameau P, Schaafsma-Zhao Q, van Aken W, Bari M,Oddi S, Kruse CG, Maccarrone M, Wadman WJ, Werkman TR

(2012) Excitability of prefrontal cortical pyramidal neurons is mod-ulated by activation of intracellular type-2 cannabinoid receptors.Proc Natl Acad Sci U S A 109(9):3534–3539

13. DevaneWA, Hanus L, Breuer A, Pertwee RG, Stevenson LA, GriffinG, Gibson D, Mandelbaum A, Etinger A, Mechoulam R (1992)Isolation and structure of a brain constituent that binds to the canna-binoid receptor. Science 258(5090):1946–1949

14. Di Marzo V (2011) Endocannabinoid signaling in the brain: biosyn-thetic mechanisms in the limelight. Nat Neurosci 14(1):9–15

15. Dinh TP, Carpenter D, Leslie FM, Freund TF, Katona I, Sensi SL,Kathuria S, Piomelli D (2002) Brain monoglyceride lipase partici-pating in endocannabinoid inactivation. Proc Natl Acad Sci U S A99(16):10819–10824

16. Dinh TP, Kathuria S, Piomelli D (2004) RNA interference suggests aprimary role for monoacylglycerol lipase in the degradation of theendocannabinoid 2-arachidonoylglycerol. Mol Pharmacol 66(5):1260–1264

17. Fegley D, Gaetani S, Duranti A, Tontini A, Mor M, Tarzia G,Piomelli D (2005) Characterization of the fatty acid amide hydrolaseinhibitor cyclohexyl carbamic acid 3′-carbamoyl-biphenyl-3-yl ester(URB597): effects on anandamide and oleoylethanolamide deactiva-tion. J Pharmacol Exp Ther 313(1):352–358

18. Fortin DA, Trettel J, Levine ES (2004) Brief trains of action poten-tials enhance pyramidal neuron excitability via endocannabinoid-mediated suppression of inhibition. J Neurophysiol 92(4):2105–2112

19. Fowler CJ (2013) Transport of endocannabinoids across the plasmamembrane and within the cell. FEBS J 280(9):1895–1904

20. Frings S, Reuter D, Kleene SJ (2000) Neuronal Ca2+-activated Cl−

channels—homing in on an elusive channel species. Prog Neurobiol60(3):247–289

21. Gonsiorek W, Lunn C, Fan X, Narula S, Lundell D, Hipkin RW(2000) Endocannabinoid 2-arachidonyl glycerol is a full agonistthrough human type 2 cannabinoid receptor: antagonism by ananda-mide. Mol Pharmacol 57(5):1045–1050

22. Hartzell C, Putzier I, Arreola J (2005) Calcium-activated chloridechannels. Annu Rev Physiol 67:719–758

23. Herkenham M, Lynn AB, Johnson MR, Melvin LS, de Costa BR,Rice KC (1991) Characterization and localization of cannabinoidreceptors in rat brain: a quantitative in vitro autoradiographic study.J Neurosci 11(2):563–583

24. Hogg RC, Wang Q, Large WA (1994) Effects of Cl channel blockerson Ca-activated chloride and potassium currents in smooth musclecells from rabbit portal vein. Br J Pharmacol 111(4):1333–1341

25. Hohmann AG, Suplita RL, Bolton NM, Neely MH, Fegley D,Mangieri R, Krey JF, Walker JM, Holmes PV, Crystal JD, DurantiA, Tontini A, Mor M, Tarzia G, Piomelli D (2005) Anendocannabinoid mechanism for stress-induced analgesia. Nature435(7045):1108–1112

26. Kano M, Ohno-Shosaku T, Hashimotodani Y, Uchigashima M,Watanabe M (2009) Endocannabinoid-mediated control of synaptictransmission. Physiol Rev 89(1):309–380

27. Katona I, Freund TF (2012) Multiple functions of endocannabinoidsignaling in the brain. Annu Rev Neurosci 35:529–558

28. Kim J, Alger BE (2010) Reduction in endocannabinoid tone is ahomeostatic mechanism for specific inhibitory synapses. NatNeurosci 13(5):592–600

29. King AR, Duranti A, Tontini A, Rivara S, Rosengarth A, Clapper JR,Astarita G, Geaga JA, LueckeH,MorM, Tarzia G, Piomelli D (2007)URB602 inhibits monoacylglycerol lipase and selectively blocks 2-arachidonoylglycerol degradation in intact brain slices. Chem Biol14(12):1357–1365

30. Lafourcade M, Elezgarai I, Mato S, Bakiri Y, Grandes P, Manzoni OJ(2007) Molecular components and functions of the endocannabinoidsystem in mouse prefrontal cortex. PLoS ONE 2(8):e709

31. Lemtiri-Chlieh F, Levine ES (2007) Lack of depolarization-inducedsuppression of inhibition (DSI) in layer 2/3 interneurons that receive


cannabinoid-sensitive inhibitory inputs. J Neurophysiol 98(5):2517–2524

32. Lerner TN, Kreitzer AC (2012) RGS4 is required for dopaminergiccontrol of striatal LTD and susceptibility to parkinsonian motordeficits. Neuron 73(2):347–359

33. Leung LS, Peloquin P, Canning KJ (2008) Paired-pulse depression ofexcitatory postsynaptic current sinks in hippocampal CA1 in vivo.Hippocampus 18(10):1008–1020

34. Llano I, Leresche N, Marty A (1991) Calcium entry increases thesensitivity of cerebellar Purkinje cells to applied GABA and de-creases inhibitory synaptic currents. Neuron 6(4):565–574

35. Lourenco J, Matias I, Marsicano G, Mulle C (2011) Pharmacologicalactivation of kainate receptors drives endocannabinoid mobilization.J Neurosci 31(9):3243–3248

36. Luch i cch i A , P i s t i s M (2012 ) Anandamide and 2 -arachidonoylglycerol: pharmacological properties, functional fea-tures, and emerging specificities of the two major endocannabinoids.Mol Neurobiol 46(2):374–392

37. Lunn CA, Reich EP, Fine JS, Lavey B, Kozlowski JA, Hipkin RW,Lundell DJ, Bober L (2008) Biology and therapeutic potential ofcannabinoid CB2 receptor inverse agonists. Br J Pharmacol 153(2):226–239

38. Maccarrone M, Dainese E, Oddi S (2010) Intracellular trafficking ofanandamide: new concepts for signaling. Trends Biochem Sci35(11):601–608

39. Mechoulam R, Ben-Shabat S, Hanus L, Ligumsky M, Kaminski NE,Schatz AR, Gopher A, Almog S, Martin BR, Compton DR et al(1995) Identification of an endogenous 2-monoglyceride, present incanine gut, that binds to cannabinoid receptors. Biochem Pharmacol50(1):83–90

40. Onaivi ES, Ishiguro H, Gu S, Liu QR (2012) CNS effects of CB2cannabinoid receptors: beyond neuro-immuno-cannabinoid activity.J Psychopharmacol 26(1):92–103

41. Pan B, Wang W, Long JZ, Sun D, Hillard CJ, Cravatt BF, Liu QS(2009) Blockade of 2-arachidonoylglycerol hydrolysis by selectivemonoacylglycerol lipase inhibitor 4-nitrophenyl 4-(dibenzo[d][1,3]dioxol-5-yl(hydroxy)methyl)piperidine-1-carboxylate (JZL184)Enhances retrograde endocannabinoid signaling. J Pharmacol ExpTher 331(2):591–597

42. Pattij T, Wiskerke J, Schoffelmeer AN (2008) Cannabinoid modula-tion of executive functions. Eur J Pharmacol 585(2–3):458–463

43. Puente N, Cui Y, Lassalle O, Lafourcade M, Georges F, Venance L,Grandes P, Manzoni OJ (2011) Polymodal activation of theendocannabinoid system in the extended amygdala. Nat Neurosci14(12):1542–1547

44. Shankar BB, Lavey BJ, Zhou G, Spitler JA, Tong L, Rizvi R, YangDY, Wolin R, Kozlowski JA, Shih NY, Wu J, Hipkin RW, GonsiorekW, Lunn CA (2005) Triaryl bis-sulfones as cannabinoid-2 receptorligands: SAR studies. Bioorg Med Chem Lett 15(20):4417–4420

45. Stella N (2010) Cannabinoid and cannabinoid-like receptors in mi-croglia, astrocytes, and astrocytomas. Glia 58(9):1017–1030

46. Stella N, Schweitzer P, Piomelli D (1997) A second endogenouscannabinoid that modulates long-term potentiation. Nature388(6644):773–778

47. Sugiura T (2009) Physiological roles of 2-arachidonoylglycerol, anendogenous cannabinoid receptor ligand. Biofactors 35(1):88–97

48. Sugiura T, Kondo S, Sukagawa A, Nakane S, Shinoda A, Itoh K,Yamashita A, Waku K (1995) 2-Arachidonoylglycerol: a possibleendogenous cannabinoid receptor ligand in brain. Biochem BiophysRes Commun 215(1):89–97

49. Terranova JP, Michaud JC, Le Fur G, Soubrie P (1995) Inhibition oflong-term potentiation in rat hippocampal slices by anandamide andWIN55212-2: reversal by SR141716 A, a selective antagonist ofCB1 cannabinoid receptors. Naunyn Schmiedeberg’s ArchPharmacol 352(5):576–579

50. Van Sickle MD, Duncan M, Kingsley PJ, Mouihate A, Urbani P,Mackie K, Stella N,Makriyannis A, Piomelli D, Davison JS, MarnettLJ, Di Marzo V, Pittman QJ, Patel KD, Sharkey KA (2005)Identification and functional characterization of brainstem cannabi-noid CB2 receptors. Science 310(5746):329–332

51. Wilson RI, Nicoll RA (2001) Endogenous cannabinoids mediateretrograde signalling at hippocampal synapses. Nature 410(6828):588–592

52. Yoshino H, Miyamae T, Hansen G, Zambrowicz B, Flynn M,Pedicord D, Blat Y, Westphal RS, Zaczek R, Lewis DA, Gonzalez-Burgos G (2011) Postsynaptic diacylglycerol lipase mediates retro-grade endocannabinoid suppression of inhibition in mouse prefrontalcortex. J Physiol 589(Pt 20):4857–4884


endocannabinoids produced upon action potential firing evoke a cl− current via type-2 cannabinoid...

Documents