Ge

JR

h

•••

a

ARR1A

KGDCCH

1

crTgtgah

0h

Neuroscience Letters 561 (2014) 58– 63

Contents lists available at ScienceDirect

Neuroscience Letters

jou rn al hom epage: www.elsev ier .com/ locate /neule t

lycine receptor in hippocampal neurons as a target for action ofxtracellular cyclic nucleotides

ulia V. Bukanova, Elena I. Solntseva ∗, Rodion V. Kondratenko, Vladimir G. Skrebitskyesearch Center of Neurology, Russian Academy of Medical Sciences, Moscow, Russia

i g h l i g h t s

Isolated rat hippocampal neurons, patch-clamp method, glycine-evoked chloride current.Acceleration of desensitization of glycine-evoked current by extracellular cAMP and cGMP.A novel mode of action of cyclic nucleotides.

r t i c l e i n f o

rticle history:eceived 17 October 2013eceived in revised form0 December 2013ccepted 17 December 2013

eywords:lycine receptoresensitizationyclic AMPyclic GMPippocampus

a b s t r a c t

Cyclic adenosine monophosphate (cAMP) and cyclic guanosine monophosphate (cGMP) are well knownintracellular second messengers. At present study, we describe the effects of extracellularly applied cAMPand cGMP on glycine-induced chloride currents (IGly) in isolated rat hippocampal pyramidal neurons.50 or 500 �M glycine was applied for 600 ms with 1 min intervals. cAMP and cGMP were co-appliedwith glycine. We found that both cAMP and cGMP rapidly, reversibly and in a dose-dependent man-ner accelerated the IGly desensitization. The effect was more prominent on IGly induced by 500 �M thanby 50 �M glycine. Dose–response curves were constructed in the 0.1–100,000 nM range of cAMP andcGMP concentrations. They demonstrate that threshold concentration of both compounds was about1 nM and maximal effect was manifested at 100 nM. When cAMP and cGMP were added to the recordingpipette, their extracellular application caused the effects similar to those obtained with normal intracel-lular medium. The effects of cyclic nucleotides remained unchanged in the presence of the antagonist of

adenosine receptors in extracellular solution, and the agonist of adenosine receptors did not mimic theeffect of cyclic nucleotides. The changes in the decay kinetics were equally pronounced at negative andpositive membrane potentials. When co-administered 1 nM cAMP and 1 nM cGMP caused a weaker effectthan either of the compounds alone which suggests a negative interaction between binding sites for cAMPand cGMP. This work describes a novel mode of action of cyclic nucleotides, namely, the modulation ofGlyRs functions from extracellular side.

© 2013 Elsevier Ireland Ltd. All rights reserved.

. Introduction

Glycine is a crucial inhibitory neurotransmitter acting on spe-ific glycine receptors (GlyRs), which are plentifully expressed inetina, spinal cord, brain stem and cerebellum [for a review, 1,2].heir localization can be both synaptic and extra-synaptic allowinglycine to influence cellular excitability through both fast synapticransmission and tonic extra-synaptic inhibition [3,4]. Although no

lycinergic synaptic currents have been found in the hippocampus,ccumulating evidence indicates the presence of functional GlyR inippocampal CA1, CA3 and dentate gyrus regions [5–7]. It has been

∗ Corresponding author. Tel.: +7 9096427758; fax: +7 4959172382.E-mail addresses: soln@front.ru, synaptology@mail.ru (E.I. Solntseva).

304-3940/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.neulet.2013.12.037

demonstrated that in hippocampal neurons GlyRs can be activatedby a number of endogenous agonists including taurine, �-alanine,and glycine, leading to the opening of strychnine-sensitive chlo-ride channels [6–8]. Extra-synaptic GlyRs in hippocampus providea tonic inhibition [9], which is very important for informationprocessing within a neuronal network and can make a contributionto many pathophysiological processes.

Numerous structurally diverse molecules have been shown tomodulate GlyRs, including anesthetics, alcohols, synthetic neuros-teroids, antagonists of 5-HT3 receptor, cannabinoids, ginkgolide B,cyclothiazide and quercetin [for a review, see 9,10]. GlyRs were

shown to be a subject to phosphorylation and regulation by dif-ferent protein kinases, such as protein kinase A (PKA), proteinkinase C (PKC) and protein tyrosine kinase (PTK) [9]. However,up to date, there have been no reports on the investigations

scienc

omcaaatildloe

2

2

tsle(a(Nadp[

2

bvicwtrt3cCspGtsw

2

msaslT0ksm

J.V. Bukanova et al. / Neuro

f possible direct effects of activator of PKA, cyclic adenosineonophosphate (cAMP), and of activator of protein kinase G,

yclic guanosine monophosphate (cGMP), on GlyRs currents. cAMPnd cGMP are well known intracellular second messengers [for

review, see 11]. Their well established mechanism of action isctivation of corresponding protein kinases with subsequent pro-ein phosphorylation. At the same time, efflux of cAMP and cGMPnto extracellular medium [12–15] and their ability to affect cellu-ar functions from extracellular side have been described for manyifferent cell types, including neurons [16–18]. However, extracel-

ular targets for either cAMP or cGMP are not known yet. The resultsf present study suggest that GlyRs may be considered as possiblextracellular targets for cAMP and cGMP.

. Materials and methods

.1. Cell preparation

All procedures were performed in accordance with the insti-utional guidelines on the care and use of experimental animalset by the Russian Academy of Sciences. The cells were iso-ated from transverse hippocampal slices as described in detaillsewhere [19]. Briefly, the slices (200–500 �m) of Wistar rats11–14 days of age) were incubated at room temperature fort least 2 h in a solution containing the following componentsin mM): 124 NaCl, 3 KCl, 2 CaCl2, 2 MgSO4, 25 NaHCO3, 1.3aH2PO4, 10 d-glucose, pH 7.4. The saline was continuously stirrednd bubbled with carbogen (95% O2 + 5% CO2). Single pyrami-al neurons from CA1 and CA3 were isolated from the stratumyramidale by a vibrating fused glass pipette with a spherical tip19].

.2. Current recordings

Glycine-activated currents in isolated neurons were inducedy a step application of agonist for 600 ms with 1 min inter-als. The application was made through glass capillary, 0.1 mmn diameter, which could be rapidly displaced laterally underontrol of home-made software [20]. Transmembrane currentsere recorded using a conventional patch-clamp technique in

he whole-cell configuration. Patch-clamp electrodes had a tipesistance of ∼2 M�. The solution in the recording pipette con-ained the following (in mM): 40 CsF, 100 CsCl, 0.1 CaCl2, 1 EGTA,

MgCl2, 4 NaATP, 5 HEPES, pH 7.3. The composition of extra-ellular solution was as follows (in mM): 140 NaCl, 3 KCl, 3aCl2, 3 MgCl2, 10 D-glucose, 10 HEPES hemisodium, pH 7.4. Thepeed of perfusion was 0.6 ml/min. Recording of the currents waserformed using EPC7 patch-clamp amplifier (HEKA Electronik,ermany). Unless noted otherwise, the holding potential was main-

ained at −70 mV. Transmembrane currents were filtered at 3 kHz,tored and analyzed with IBM-PC computer, using homemade soft-are.

.3. Reagents

Glycine, adenosine 3′,5′-cyclic monophosphate sodium saltonohydrate (cAMP), guanosine 3′,5′-cyclic monophosphate

odium salt (cGMP), 3-isobutyl-1-methylxanthine (IBMX), �-minobutyric acid (GABA), adenosine 5′-triphosphate disodiumalt hydrate (ATP), as well as all of the chemicals for intracel-ular and extracellular solutions were purchased from “Sigma”.he tested substances were dissolved in distilled water to make

.1–1 mM stock solution, which was divided into daily aliquots andept frozen at −20 ◦C. The substances were dissolved in externalaline to their final concentration immediately before the experi-ents.

e Letters 561 (2014) 58– 63 59

2.4. Data analysis

All statistical analysis was performed with the help of PrismGraphpad software. All comparisons were made with paired two-tailed Student’s t-test and nonparametric Mann–Whitney test ata significance level of P = 0.05. In result descriptions, mean andstandard error of mean (SEM) are specified. In figures, error barsrepresent SEM.

3. Results

3.1. Glycine-activated chloride currents in rat hippocampalpyramidal neurons

Experiments were performed in CA1 and CA3 pyramidal neu-rons isolated from the hippocampus of rats (11–14 d of age).Application of glycine evoked chloride currents (IGly) whichamplitude and kinetics were dependent on glycine concentration(Fig. 1A). The EC50 value of glycine was 90 ± 7 �M (n = 6) whichcorresponds well with the value reported for acutely isolated hip-pocampal neurons by other authors [21]. An average value of thereversal potential of IGly −9.8 ± 0.9 mV matched well the chloridereversal potential calculated for the chloride concentrations used(−9.5 mV). The IGly had low sensitivity to the GABAAR antagonistbicuculline (10 �M), but was completely and reversibly blockedby specific GlyR blocker strychnine (3 �M) (Fig. 1B). No differ-ences between peak amplitude of IGly recorded in neurons isolatedfrom CA1 and CA3 were noticed (3.3 ± 0.9 nA vs 3.8 ± 1.1 nA). Sincethere was also no difference in cyclic nucleotides effects, the resultsobtained from cells of CA1 and CA3 regions were grouped together.

3.2. Modulation of glycine-activated currents by extracellularcyclic nucleotides

The IGly was activated by 50 �M or 500 �M glycine appliedfor 600 ms with 1 min intervals. Different concentrations(0.1–100,000 nM) of cAMP and cGMP were co-applied withglycine. Application of cAMP or cGMP alone did not evoke anydirect membrane response. When co-applied with glycine, bothcAMP and cGMP barely affected the IGly peak amplitude, but rapidly,reversibly and in a dose-dependent manner accelerated the IGlydesensitization (Fig. 1C and E). The effect was well pronouncedin the 1–100,000 nM range of cAMP and cGMP concentrations.To quantitatively assess the alteration of the current kinetics, wemeasured the halftime of a decay (�) of IGly in the absence orpresence of cAMP or cGMP. The IGly desensitization kinetics wasfitted with mono-exponential function (Prism Graphpad software).However, in some experiments, especially with 50 �M glycine, adecay of the control current was not prominent. In such cases, theamplitude of IGly at the end of glycine application (IGly at 600 ms)was measured, and the effect was estimated as a ratio of IGly at600 ms in the presence and absence of the cyclic nucleotide tested.It was found that cAMP and cGMP affected IGly in a similar manner(Fig. 1C–G). The effect was more prominent for IGly induced by500 �M glycine than for IGly induced by 50 �M. Fig. 1D and Fshows dose–effect relationship for a reduction of normalizedIGly at 600 ms evoked by 50 and 500 �M glycine. One can seethat the threshold concentration of cyclic nucleotides is close to1 nM and maximal effect is manifested at 100 nM. The maximaldecrease of IGly evoked by 50 and 500 �M glycine averaged 47 ± 6%(P < 0.0001, n = 8) and 55 ± 3% (P < 0.0001, n = 10), accordingly, in

the presence of 100 nM cAMP, and 43 ± 6% (P < 0.0001, n = 9) and53 ± 5% (P < 0.0001, n = 10), accordingly, in the presence of 100 nMcGMP. The concentrations of cyclic nucleotides above 100 nMcaused weaker effects with large dispersion. It seems that high

60 J.V. Bukanova et al. / Neuroscience Letters 561 (2014) 58– 63

Fig. 1. (A) Agonist concentration dependence of glycine-induced current (IGly). (left) Family of IGly representative traces at increasing glycine concentration ranging from 10to 1000 �M. Holding potential was −70 mV. (right) Normalized averaged concentration–response curve of IGly (n = 5). (B) A blockade of IGly by a GlyR antagonist strychnine(3 �M). (C–G) Modulation of IGly by extracellular cyclic nucleotides. (C) and (E) Representative traces of IGly, induced by 50 (left) and 500 (right) �M glycine, obtained incontrol and in the presence of 100 nM cAMP (C) or 100 nM cGMP (E). (D) and (F) Concentration dependence of cAMP (D) and cGMP (F) effects on the normalized IGly at6 perimw bers uc

cmdtow

3

ti

00 ms evoked by 50 or 500 �M glycine. Vh = −70 mV. (G) The time course of one exere co-applied with glycine. Responses to glycine are shown by open circles. Num

orrespondingly.

oncentrations of cyclic nucleotides trigger an unknown additionalechanism(s) that cancel or masque the acceleration of IGly

esensitization caused by their low concentrations. Fig. 1G showshe time course of one experiment where different concentrationsf cAMP (full circles) and cGMP (full triangles) were co-appliedith glycine.

.3. Voltage-independence of cyclic nucleotides effects on IGly

Cyclic nucleotides significantly accelerated the decay of IGly at allested holding potentials (Fig. 2). The changes in the decay kinetics,nduced by cAMP and cGMP, were equally pronounced at negative

ent where different concentrations of cAMP (full circles) and cGMP (full triangles)nder full circles and under full triangles show concentrations of cAMP and cGMP,

and positive membrane potentials so that the reversal potential ofIGly at 600 ms (−9.6 ± 1.2 mV) remained unchained (n = 6). Resultssuggest that changes in intracellular [Cl−] are unlikely to contributeto the mechanisms of acceleration of IGly decay caused by cyclicnucleotides.

3.4. cAMP and cGMP in the recording pipette

Even though it is well known, that cyclic nucleotides do notreadily penetrate cell membranes, there was still some possibil-ity that they were affecting glycine receptor function by acting atan intracellular site. To examine this possibility, we repeated the

J.V. Bukanova et al. / Neuroscience Letters 561 (2014) 58– 63 61

Fig. 2. Voltage-independence of the IGly modulation by cyclic nucleotides. (A) and (C) The family of representative traces of current induced by 500 �M glycine at holdingpotentials of −90, −70, −30, −10 and +30 mV, obtained in control solution (left) and in the presence of 100 nM cAMP (A, right) and 100 nM cGMP (C, right). (B) and (D)Averaged current–voltage relationships for IGly at 600 ms obtained in control, in the presence of 100 nM cAMP (B) or 100 nM cGMP (D), and after wash-out of the drug. Dataare plotted as a fraction of IGly at 600 ms at −70 mV in control solution.

Fig. 3. (A and B) Persistence of the effects of extracellularly applied cAMP and cGMP with cyclic nucleotides in the recording pipette. (A) Representative traces of IGly, inducedby 500 �M glycine, obtained in control and in the presence of 100 nM cAMP (left) or 100 nM cGMP (right), recorded with the pipette containing 10 �M cAMP and 10 �McGMP. (B) Statistics of the effects of 100 nM cAMP and 100 nM cGMP on the � of IGly in control and in the presence of cAMP and cGMP in the pipette. No differences arenotable between these two groups of data. (C) Persistence of the effects of 10 nM cAMP (left) and 10 nM cGMP (right) when bath solution contained nonselective antagonist ofadenosine receptors IBMX (200 �M). (D) The absence of the effect of adenosine receptors agonist ATP (10 nM) on IGly. (E and F) Chloride current induced by GABA applicationis not changed in the presence of either 10 nM cAMP or 10 nM cGMP. Concentration of GABA is 10 �M (E) and 50 �M (F).

6 science Letters 561 (2014) 58– 63

epcadmc(as

3p

aiImutc(r

3c

d(wcit

3c

waceiocnvRa

4

eiccmnraowe

Fig. 4. Co-administration of cAMP and cGMP caused a weaker effect than either drugalone. (A) Traces of IGly (500 �M) obtained in the control and in the presence of cAMP(left), cGMP (middle), and cAMP together with cGMP (right). The concentration of

2 J.V. Bukanova et al. / Neuro

xperiments adding both cAMP (10 �M) and cGMP (10 �M) to theipette solution. Current responses to glycine recorded under theseonditions were not detectably different from control recordings,nd extracellular applications of cAMP or cGMP accelerated IGlyesensitization the same way as in previously described experi-ents (Fig. 3A and B). Externally applied 100 nM cAMP or 100 nM

GMP reduced � of IGly, evoked by 500 �M glycine, to 44 ± 8%P = 0.0006, n = 4) and 47 ± 4% (P = 0.0006, n = 4) of the control level,ccordingly, which is very similar to the results obtained withtandard intracellular medium.

.5. The maintenance of the effect of cyclic nucleotides in theresence of antagonist of adenosine receptors

To verify whether cyclic nucleotides exert their effects viadenosine receptors, we studied the effects of cAMP and cGMPn the presence of nonselective antagonist of adenosine receptors,BMX [22]. When 200 �M IBMX was added to the extracellular

edium for 20 min, the effect of 10 nM cyclic nucleotides remainednchanged (Fig. 3C). Furthermore, the agonist of adenosine recep-ors ATP did not mimic the effect of cyclic nucleotides: 10 nM ATPo-applied with glycine did not accelerate desensitization of IGlyFig. 3D). The results speak against the involvement of adenosineeceptors in the mechanisms of cyclic nucleotides effect.

.6. Cyclic nucleotides do not accelerate desensitization ofhloride current induced by GABA application

Application of GABA on hippocampal pyramidal neurons evokedose-dependent chloride current with EC50 value of 18 ± 6 �Mn = 6) (not shown). Co-application of 10 nM cAMP or 10 nM cGMPith either 10 �M (Fig. 3E) or 50 �M (Fig. 3F) GABA did not

ause any change in desensitization of GABA-evoked current whichndicates a specificity of cyclic nucleotides action on glycine recep-ors.

.7. Attenuation of the effect in cases of cAMP and cGMPo-administration

To test whether cAMP and cGMP share the same site of action,e compared the effects of cAMP and cGMP during their individual

nd combined application on the same cell (Fig. 4). Unexpectedly,o-administration of cAMP and cGMP caused a weaker effect thanither of the compounds alone. When cAMP and cGMP were appliedn concentration of 1 nM, the values of � of IGly decreased to 60 ± 3%f control in the presence of cAMP, to 68 ± 3% in the presence ofGMP, and to 95 ± 3% in the presence of cAMP + cGMP (P < 0.005,

= 8). At 10 nM concentration of cAMP and cGMP, correspondingalues of � were 50 ± 3%, 53 ± 4%, and 82 ± 3% (P < 0.005, n = 7).esults suggest the existence of separate sites of actions for cAMPnd cGMP with negative cooperativity between them.

. Discussion

In the present work, we have shown for the first time, thatxtracellular application of cAMP and cGMP rapidly, reversibly, andn a dose-dependent manner accelerate decay of GlyRs chlorideurrent. The effect was more prominent with higher glycine con-entration (500 �M vs 50 �M), which indicates a noncompetitiveanner of action of cyclic nucleotides. The effect was equally pro-

ounced at negative and positive membrane potentials so that theeversal potential of IGly at 600 ms did not shift along the voltage

xis. When recording pipette contained cAMP and cGMP, the effectsf extracellular application of these drugs remained unchanged,hich suggests that the sites of cAMP and cGMP action are located

xtracellularly. The involvement of adenosine receptors does not

cAMP and cGMP was 1 nM (upper row) or 10 nM (lower row). (B) Statistics of theeffects of 1 nM and 10 nM of cAMP and cGMP applied separately or together on the� of IGly evoked by 500 �M glycine.

seem plausible because cyclic nucleotides effects persisted in thepresence of a nonselective antagonist of adenosine receptors IBMX[22] and, secondly, adenosine receptors agonist ATP did not mimicthe effect of cyclic nucleotides. Further, this effect seems to bespecific for GlyRs, because GABA-evoked chloride current was notchanged in the presence of 10 nM cAMP or cGMP.

The threshold effective concentrations for both cAMP and cGMPwere about 1 nM, which is close to naturally occurring extracellularcyclic nucleotides concentrations in the brain (0.4 nM) [16], whichindicates some possible physiological implications for the modu-lation of GlyRs by cyclic nucleotides. When cAMP and cGMP wereapplied together, they caused a weaker effect than either of thecompounds alone, which suggests the existence of separate sitesof actions for cAMP and cGMP with negative interaction betweenthem.

There are few reports in literature concerning physiologicaleffects of extracellular cAMP. The depression of population spikeamplitude caused by 50 �M cAMP has been shown in rat hippocam-pal slices. The effect was mimicked by adenosine and was explainedby the interaction of cAMP with membrane-bound adenosinereceptors [23]. A reduction of GABAAR induced chloride current

by extracellular 1–2 mM cAMP was described in rat hippocampalneurons [17] and mice spinal cord neurons [18]. This effect was notmimicked by adenosine, remained unchanged with cAMP addedto recording pipette, and was not reversed by PKA inhibitor H-8.

scienc

AGsoo

iahwiaimnwlmc

nf

C

A

03S

R

[

[

[

[

[

[[

[

[

[

[

[

[

J.V. Bukanova et al. / Neuro

uthors have concluded that high concentrations of cAMP reducedABAAR-activated chloride current by acting at an extracellularite. In our experiments, no noticeable effect of low concentrationsf cAMP and cGMP (10 nM) on GABA-induced chloride current wasbserved.

The level of extracellular cGMP was shown to be an extremelymportant factor for an impairment of cognitive functions in hyper-mmonemia and hepatic encephalopathy [16,24]. In rats withyperammonemia, the disturbance of nitric oxide-cGMP pathwayas found to lead to the reduction of extracellular cGMP level

n brain and to the impairment of learning ability. Intracerebraldministration of phosphodiesterase inhibitor zaprinast or cGMPtself restored the ability of hyperammonemic rats to learn a Y

aze task. An augmentation of extracellular cGMP is proposed as aew therapeutic approach to improve cognitive function in patientsith hepatic encephalopathy [24]. Our results concerning modu-

ation of GlyRs functions by cGMP might help to understand theechanisms underlying the ability of extracellular cGMP to rescue

ognitive functions in hepatic encephalopathy.In conclusion, we have found a novel mode of action for cyclic

ucleotides, namely, their ability to modulate of GlyRs functionrom an extracellular side.

onflict of interest

The authors declare that they have no conflict of interest.

cknowledgments

This work was supported by Grants 13-04-00326 and 11-04-8307 from the Russian Foundation for Basic Research, and Grant598.2012.4 from the Foundation for Support of Russian Scientificchools.

eferences

[1] J.W. Lynch, Molecular structure and function of the glycine receptor chloridechannel, Physiol. Rev. 84 (2004) 1051–1095.

[2] J.W. Lynch, Native glycine receptors subtypes and their physiological roles,Neuropharmacology 56 (2009) 303–309.

[3] G.P. Dugue, A. Dumoulin, A. Triller, S. Dieudonne, Target-dependent use ofcoreleased inhibitory transmitters at central synapses, J. Neurosci. 25 (2005)6490–6498.

[

[

e Letters 561 (2014) 58– 63 63

[4] W. Song, S.C. Chattipakorn, L.L. McMahon, Glycine-gated chloride channelsdepress synaptic transmission in rat hippocampus, J. Neurophysiol. 95 (2006)2366–2379.

[5] H. Betz, B. Laube, Glycine receptors: recent insights into their structural orga-nization and functional diversity, J. Neurochem. 97 (2006) 1600–1610.

[6] C. Chattipakorn, L.L. McMahon, Pharmacological characterization of glycine-gated chloride currents recorded in rat hippocampal slices, J. Neurophysiol. 87(2002) 1515–1525.

[7] L. Danglot, P. Rostaing, A. Triller, A. Bessis, Morphologically identified glycener-gic synapses in the hippocampus, Mol. Cell. Neurosci. 27 (2004) 394–403.

[8] M. Mori, B.H. Gahwiler, U. Gerber, Beta-alanine and taurine as endogenousagonists at glycine receptors in rat hippocampus in vitro, J. Physiol. 539 (2002)191–200.

[9] T.L. Xu, N. Gong, Glycine and glycine receptor signaling in hippocampal neu-rons: diversity, function and regulation, Prog. Neurobiol. 91 (2010) 349–361.

10] G.E. Yevenes, H.U. Zeilhofer, Allosteric modulation of glycine receptors, Br. J.Pharmacol. 164 (2011) 224–236.

11] K. Arora, C. Sinha, W. Zhang, A. Ren, C.S. Moon, S. Yarlagadda, A.P. Naren, Com-partmentalization of cyclic nucleotide signaling: a question of when, where,and why? Pflugers Arch. 465 (2013) 1397–1407.

12] A.K. Cadogan, D.A. Kendall, H. Fink, C.A. Marsden, Social interaction increases5-HT release and cAMP efflux in the rat ventral hippocampus in vivo, Behav.Pharmacol. 5 (1994) 299–305.

13] C.D. Cropp, T. Komori, J.E. Shima, T.J. Urban, S.W. Yee, S.S. More, M.K. Giacomini,Organic anion transporter 2 (SLC22A7) is a facilitative transporter of cGMP, Mol.Pharmacol. 73 (2008) 1151–1158.

14] P.A. Rosenberg, M.A. Dichter, Extracellular cAMP accumulation and degrada-tion in rat cerebral cortex in dissociated cell culture, J. Neurosci. 9 (1989)2654–2663.

15] G. Sager, Cyclic cGMP transporters, Neurochem. Int. 45 (2004) 865–873.16] S. Erceg, P. Monfort, M. Hernandez-Viadel, M. Llansola, C. Montoliu, V. Felipo,

Restoration of learning ability in hyperammonemic rats by increasing extra-cellular cGMP in brain, Brain Res. 1036 (2005) 115–121.

17] N.A. Lambert, N.L. Harrison, Analogs of cyclic AMP decrease gamma-aminobutyric acid A receptor-mediated chloride current in cultured rathippocampal neurons via an extracellular site, J. Pharmacol. Exp. Ther. 255(1990) 90–94.

18] M.K. Ticku, A.K. Mehta, gamma-Aminobutyric acid A receptor desensitizationin mice spinal cord cultured neurons: lack of involvement of protein kinases Aand C, Mol. Pharmacol. 38 (1990) 719–724.

19] V.S. Vorobjev, Vibrodissociation of sliced mammalian nervous tissue, J. Neu-rosci. Methods 38 (1991) 145–150.

20] V.S. Vorobjev, I.N. Sharonova, H.L. Haas, A simple perfusion system for patch-clamp studies, J. Neurosci. Methods 68 (1996) 303–307.

21] N. Lozovaya, N. Yatsenko, A. Beketov, T. Tsintsadze, N. Burnashev, Glycine recep-tors in CNS neurons as a target for nonretrograde action of cannabinoids, J.Neurosci. 25 (2005) 7499–7506.

22] J.W. Daly, K.A. Jacobson, D. Ukena, Adenosine receptors: development of selec-tive agonists and antagonists, Prog. Clin. Biol. Res. 230 (1987) 41–63.

23] T.V. Dunwiddie, B.J. Hoffer, Adenine nucleotides and synaptic transmission inthe in vitro rat hippocampus, Br. J. Pharmacol. 69 (1980) 59–68.

24] C. Montoliu, R. Rodrigo, P. Monfort, M. Llansola, O. Cauli, J. Boix, N. Elmlili, A.Agusti, V. Felipo, Cyclic GMP pathways in hepatic encephalopathy. Neurologicaland therapeutic implications, Metab. Brain Dis. 25 (2010) 39–48.