Memantine Inhibits ATP-Dependent K� Conductances inDopamine Neurons of the Rat Substantia Nigra Pars Compacta

Michela Giustizieri, Maria Letizia Cucchiaroni, Ezia Guatteo, Giorgio Bernardi,Nicola B. Mercuri, and Nicola BerrettaFondazione Santa Lucia Istituto di Ricovero e Cura a Carattere Scientifico, Laboratory of Experimental Neurology(M.G., M.L.C., E.G., G.B., N.B.M., N.B.) and Department of Neuroscience, University of Rome “Tor Vergata” (G.B., N.B.M.),Rome, Italy

Received March 5, 2007; accepted May 10, 2007

ABSTRACT1-Amino-3,5-dimethyl-adamantane (memantine) is a noncom-petitive N-methyl-D-aspartate (NMDA) receptor antagonistused in clinical practice to treat neurodegenerative disordersthat could be associated with excitotoxic cell death. Be-cause memantine reduces the loss of dopamine neurons ofthe substantia nigra pars compacta (SNc) in animal models ofParkinson’s disease, we examined the effects of this drug ondopamine cells of the SNc. Besides inhibition of NMDAreceptor-mediated currents, memantine (30 and 100 �M)increased the spontaneous firing rate of whole-cell recordeddopamine neurons in a midbrain slice preparation. Occasion-ally, a bursting activity was observed. These effects wereindependent from the block of NMDA receptors and wereprevented in neurons dialyzed with a high concentration ofATP (10 mM). An increase in firing rate was also induced bythe ATP-sensitive potassium (KATP) channel antagonist tol-

butamide (300 �M), and this increase occluded further ef-fects of memantine. In addition, KATP channel-mediated out-ward currents, induced by hypoxia, were inhibited bymemantine (30 and 100 �M) in the presence of the NMDAreceptor antagonist (5S,10R)-(�)-5-methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-imine maleate (MK-801)(10 �M). An increase in the spontaneous firing rate by me-mantine was observed in dopamine neurons recorded withextracellular planar 8 � 8 multielectrodes in conditions ofhypoglycemia. These results highlight KATP channels as pos-sible relevant targets of memantine effects in the brain.Moreover, in view of a proposed role of KATP conductancesin dopamine neuron degeneration, they suggest anothermechanism of action underlying the protective role of me-mantine in Parkinson’s disease.

Memantine is a derivative of amantadine used in clinicalpractice to treat several neurological disorders associatedwith excitotoxic cell death, including Parkinson’s disease,amyotrophic lateral sclerosis, Alzheimer’s disease, epilepsy,stroke, spasticity, convulsions, and vascular dementia (Fleis-chhacker et al., 1986; Ditzler, 1991; Sonkusare et al., 2005;Chen and Lipton, 2006; Volbracht et al., 2006). Its mecha-nism of action resides on the ability of memantine to bindN-methyl-D-aspartate (NMDA) receptors in a noncompetitivemanner, acting as a low-affinity open-channel blocker (Lip-

ton, 2004, 2006). Because of its peculiar pharmacologicalproperties, memantine is proposed to be beneficial because itblocks excessive NMDA receptor activation without interfer-ing with its physiological activity. In doing so, it is a welltolerated NMDA receptor antagonist able to reduce or pre-vent excitotoxic damage without producing undesired sideeffects, such as hallucination, agitation, catatonia, centrallymediated increase in blood pressure, and anesthesia, typicalof other NMDA receptor antagonists.

Although NMDA receptors constitute the main target ofmemantine, other mechanisms of action have been reported,including reduced action potential firing in cultured neurons(Netzer and Bigalke, 1990) and block of 5-HT3 receptors(Reiser et al., 1988; Rammes et al., 2001) and nicotinic re-ceptors (Maskell et al., 2003; Aracava et al., 2005). These

This work has been supported by Grant RF04.125 from the Italian Ministryof Health and by Lundbeck Italia SpA.

Article, publication date, and citation information can be found athttp://jpet.aspetjournals.org.

doi:10.1124/jpet.107.122036.

ABBREVIATIONS: memantine, 1-amino-3,5-dimethyl-adamantane; NMDA, N-methyl-D-aspartate; SNc, substantia nigra pars compacta; KATP,ATP-sensitive potassium; ACSF, artificial cerebrospinal fluid; dopamine, 3,4-dihydroxyphenethylamine; AP5, D-(�)-2-amino-5-phosphono-eptanoic acid; MK-801, (5S,10R)-(�)-5-methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-imine maleate; tolbutamide, 1-buthyl-3-(4-meth-ylbenzenesulfonyl)urea; mirtazapine, 1,2,3,4,10,14b-hexahydro-2-methylpyrazino[2,1-a]pyrido[2,3-c][2]benzazepine; methyllycaconitine,[1�,4(S),6�,14�,16�]-20-ethyl-1,6,14,16-tetramethohy-4-[[[2-(3-methyl-2,5-dioxo-1-pyrrolidinyl) benzoyl]oxy]methyl]aconitane-7,8-diol citrate;INMDA, NMDA-mediated inward current; IHYPO, hypoxia-induced outward current; 5-HT, serotonin.

0022-3565/07/3222-721–729$20.00THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 322, No. 2Copyright © 2007 by The American Society for Pharmacology and Experimental Therapeutics 122036/3231817JPET 322:721–729, 2007 Printed in U.S.A.

721

at Fondazione S

anta Lucia - Biblioteca on N

ovember 19, 2012

jpet.aspetjournals.orgD

ownloaded from

results suggest the presence of additional mechanisms ofaction whose relative importance may be dependent on thebrain area under investigation.

An antiparkinsonian activity has been described for me-mantine in animal models of Parkinson’s disease and inparkinsonian patients (Danysz et al., 1997; Merello et al.,1999). In addition, memantine prevents cell death induced by1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (Lange et al.,1997; Kucheryanu and Kryzhanovskii, 2000). Therefore, wefocused our attention on the action of this drug on the dopa-mine neurons of the substantia nigra pars compacta (SNc),whose progressive degeneration is a hallmark of Parkinson’sdisease. Indeed, several lines of evidence indicate an over-stimulation of glutamate receptors, especially of the NMDAsubtype, as the main cause of the progressive loss of thisneuronal population (Gardoni and Di Luca, 2006); thus,NMDA receptor antagonism by memantine in the SNc isexpected to represent the main mechanism underlying itsneuroprotective action. Accordingly, we now present evi-dence that memantine is effective in reducing NMDA-medi-ated currents of the dopamine neurons; however, we alsoshow a novel mechanism of action of this drug, consisting ofthe ability to reduce ATP-sensitive potassium (KATP) conduc-tances, opened in conditions of metabolic stress.

Materials and MethodsSlice Preparation. Wistar rats (18–25 days old) were anesthe-

tized with halothane and killed by decapitation. All the experimentswere carried out in accordance with the Declaration of Helsinki andfollowed international guidelines on the ethical use of animals fromthe European Communities Council Directive of November 24, 1986(86/609/EEC). The brain was rapidly removed from the skull, andhorizontal midbrain slices (250 �m) were cut in cold (8–12°C) arti-ficial cerebrospinal fluid (ACSF) and left to recover at 33.5°C for atleast 1 h. ACSF composition was the following: 126 mM NaCl, 2.5mM KCl, 1.2 mM MgCl2, 2.4 mM CaCl2, 1.2 mM NaH2PO4, 24 mMNaHCO3, and 11 mM glucose, saturated with 95% O2/5% CO2, pH7.4.

Patch-Clamp Recordings. An individual slice was placed in arecording chamber on the stage of an upright microscope (AxioscopeFS, Carl Zeiss Spa, Arese, Italy) and submerged in a continuouslyflowing (2.5 ml/min) solution at 33°C (�0.2°C). Neurons were visu-alized with infrared video microscopy (Hamamatsu Photonics ItaliaSrl, Milan, Italy). Borosilicate glass electrodes (3–4 M�) were filledwith 135 mM potassium gluconate, 10 mM KCl, 2 mM MgCl2, 0.045mM CaCl2, 0.1 mM EGTA, 10 mM HEPES, 2 mM ATP, and 0.3 mMGTP (pH 7.3, with KOH). In experiments with high calcium buffer,EGTA concentration was increased to 10 mM and CaCl2 was in-creased to 4 mM, whereas potassium gluconate was reduced to 115mM. In experiments with high ATP, the filling solution was the sameas the control solution but with 10 mM ATP. Whole-cell voltage-clamp (�60 mV holding potential) or current-clamp experimentswere carried out with a MultiClamp 700A amplifier (Molecular De-vices, Sunnyvale, CA), filtered at 1 kHz, and digitized at 10 kHz.Dopamine neurons were identified electrophysiologically based on aprominent Ih in response to hyperpolarizing voltage steps and avoltage sag when negative current steps were applied in current-clamp mode (Grillner and Mercuri, 2002). Data are expressed asmean � S.E.M. and compared using the Student’s t test, with p �0.05 as minimal level for significance.

Pressure-applied NMDA (10 psi, 0.5–1.0 s) was used to obtainNMDA-mediated inward currents (INMDA) through a patch electrodefilled with NMDA (100 �M) dissolved in ACSF and connected to aPneumatic Pico-pump PV 800 (WPI, Berlin, Germany). The puff

electrode was positioned above the slice in close proximity to therecorded neuron. Hypoxic insults were obtained by exposing theslices for 2 to 3 min to ACSF saturated with 95% N2/5% CO2.

Multielectrode Recordings. Individual slices were placed overan 8 � 8 array of planar microelectrodes, each 20 � 20 �m in size,with an interpolar distance of 100 �m (MED-P2105; MatsushitaElectric Industrial Co., Ltd., Kadoma, Japan). Slices were sub-merged under a nylon mesh and positioned over the multielectrodearray under visual control through an upright microscope (LeicaDM-LFS; Leica Microsystems, Wetzlar, Germany) in such a way thatthe area close to the medial terminal nucleus of the accessory optictract covered most of the electrodes (Geracitano et al., 2005). Extra-cellular signals were acquired using the Panasonic MED64 System(Alpha MED Sciences, Kadoma, Japan). Signals were low-cut filteredat 100 Hz and digitized at 20 kHz with a 6071E Data AcquisitionCard (National Instruments, Austin, TX) using the MED64 Conduc-tor Software (Alpha MED Sciences). The frequency of the fast tran-sients corresponding to spontaneous action potentials was calculatedoff-line with Spike2 software (Cambridge Electronic Design Ltd,Cambridge, UK) using an amplitude threshold adjusted by visualinspection in each individual active channel. Spikes recorded by asingle channel could differ in shape and amplitude, reflecting spon-taneous action potentials arising from more than one neuron; there-fore, spike-sorting discrimination of multiunit responses wasachieved by generating spike templates with Spike2 (CambridgeElectronic Design Ltd) sorted with a normal mixtures algorithm onindependent clusters obtained from principal component data.

On average, we detected spikes from 34.6 � 1.5 active channels ineach midbrain slice (n � 14); however, following the spike-sortingprocedure, activity arising from 82.7 � 3.6 cells/slice could be ob-tained. Because dopamine neurons of the SNc are selectively hyper-polarized by dopamine (Grillner and Mercuri, 2002), we used sensi-tivity to dopamine (30 �M) as a criterion to discriminate spikesoriginating from this neuronal population. Of the total number ofcells, 43.4 � 2.0 cells/slice met this pharmacological criterion andwas considered for further analysis. Hypoglycemic conditions wereobtained by perfusing the slices in ACSF containing 1 mM glucose for25 to 30 min before challenge with memantine or tolbutamide (Figs.6 and 7). This procedure caused an overall reduction of firing fre-quency in the dopamine cell population, and a complete loss ofactivity occurred in some of them. Only the dopamine cells whosefiring could still be detected during the experimental session inhypoglycemia were considered to evaluate the effects of memantineor tolbutamide.

Drugs. All the drugs, with the exception of NMDA, which waspressure-applied, were applied in the bath. They included dopamine,AP5, MK-801, and tolbutamide from Sigma-Aldrich (Milan, Italy);mirtazapine and methyllycaconitine from Tocris Cookson Ltd (Bris-tol, UK); and memantine from Lundbeck Italia SpA (Milan, Italy).

ResultsMemantine Inhibits NMDA-Mediated Currents. Do-

pamine neurons of the SNc were identified based on electro-physiological criteria (see under Materials and Methods) andrecorded in voltage-clamp mode at �60 mV holding potential.We investigated the effects of memantine on NMDA-medi-ated inward currents (INMDA) evoked by pressure applicationof NMDA (100 �M, 10 psi, 0.5–1.0 s) at 3-min intervalsthrough a glass pipette positioned in close proximity to therecorded neuron. INMDA (190.5 � 18.4 pA, n � 26) remainedconstant on repeated application of NMDA and was sensitiveto the NMDA receptor antagonist MK-801 (10 �M, n � 4)(Fig. 1). Perfusion with 30 �M memantine reversibly reducedINMDA amplitude to 48.8 � 2.9% of control (n � 15). Increas-ing memantine concentration to 100 �M produced further

722 Giustizieri et al.

at Fondazione S

anta Lucia - Biblioteca on N

ovember 19, 2012

jpet.aspetjournals.orgD

ownloaded from

inhibition of INMDA amplitude (26.7 � 3.0% of control, n � 15)(Fig. 1B). These results confirmed the ability of memantine toinhibit NMDA receptor-mediated responses in dopamineneurons of the SNc.

NMDA Receptor-Independent Increase of DopamineNeuron Firing Rate by Memantine. We then investigatedthe effects of memantine on dopamine neuron firing proper-

ties. Neurons were recorded in current-clamp mode to detecttheir typical spontaneous tonic action potential discharge(Grillner and Mercuri, 2002). We found that memantine onits own was capable of increasing the basal firing frequencyof the recorded neurons. At 30 �M, memantine reversiblyincreased the firing frequency from 1.1 � 0.2 to 2.0 � 0.3 Hz(p � 0.002 paired t test, n � 11) (Fig. 2A). At higher concen-

Fig. 1. Dose-dependent inhibition ofthe NMDA receptor-mediated inwardcurrent (INMDA) evoked on dopamineneurons by pressure application ofNMDA (100 �M). A, bottom, plot ofINMDA amplitude against time, in re-sponse to perfusion with 30 �M me-mantine and 10 �M MK-801. On top,raw traces of INMDA acquired at thetimes indicated by the correspondingletters in the plot. B, histogram ofINMDA amplitude expressed as per-centage of control (mean � S.E.M.)following exposure to 30 and 100 �Mmemantine and 10 �M MK-801.

Fig. 2. NMDA receptor-independent effects of memantineon dopamine neuron spontaneous firing rate. A, plot of theinstantaneous firing frequency against time from a singledopamine neuron, showing the reversible increase in ac-tion potential discharge induced by perfusion of 30 and 100�M memantine. B, trace records from a dopamine neuronexposed to 100 �M memantine, showing a reversiblechange in the firing mode of action potential discharge,from tonic to bursting activity. C, trace records from adopamine neuron exposed to 100 �M memantine in thecontinuous presence of 10 �M MK-801. D, histogram of theincrease in firing rate (mean � S.E.M.) induced by 100 �Mmemantine, in control conditions, in the presence of 10 �MMK-801, 50 �M AP5, and 10 �M mirtazapine � 10 nMmethyllycaconitine. �, p � 0.05, ���, p � 0.001; paired ttest.

Memantine Inhibition of KATP Channels 723

at Fondazione S

anta Lucia - Biblioteca on N

ovember 19, 2012

jpet.aspetjournals.orgD

ownloaded from

tration (100 �M), memantine increased the firing rate from1.4 � 0.1 to 3.2 � 0.3 Hz (p � 0.001 paired t test, n � 21) (Fig.2, A and D), and in 12 cells it transformed their typical tonicfiring into rhythmic bursts of action potentials (Fig. 2B).

To examine whether this effect of memantine could bebecause of reduction of tonic NMDA receptor stimulationwithin the neuronal circuitry, we repeated the same experi-ments in the presence of 10 �M MK-801 or 50 �M AP5;however, both of these NMDA receptor antagonists did notmimic or prevent the effects of 100 �M memantine on dopa-mine neuron firing (Fig. 2, C and D).

Memantine has also been reported to act as an antagonistof 5-HT3 and �7-containing nicotinic receptors (Reiser et al.,1988; Rammes et al., 2001; Maskell et al., 2003; Aracava etal., 2005); however, in the presence of the 5-HT3 receptorantagonist mirtazapine (100 �M) and the �7-containing nic-otinic receptor antagonist methyllycaconitine (10 nM), me-mantine (100 �M) still produced a significant increase indopamine neuron firing rate (Fig. 2D).

No Effect of Memantine in Multielectrode Record-ings of Dopamine Neuron Firing Rate. We then tried tocorroborate these results using a less invasive technique ofspontaneous firing recording in the same midbrain slicepreparation, consisting of single unit’s detection from an 8 �8 array of planar electrodes (Geracitano et al., 2005). Becausedopamine neurons of the SNc are selectively hyperpolarizedby dopamine (Grillner and Mercuri, 2002), spikes originatingfrom dopamine neurons were identified based on their sen-sitivity to brief (30–60 s) exposure to 30 �M dopamine. Usingthis pharmacological criterion, we recorded the spontaneousfiring arising from 302 presumed dopamine neurons from sixmidbrain slices. Their basal firing rate was 2.08 � 0.10 Hz,and no significant alteration in spontaneous firing was ob-served following 100 �M memantine perfusion (2.05 � 0.01Hz in memantine, p � 0.27 paired t test) (Fig. 3).

High Intracellular ATP Prevents the Effects of Me-mantine on Dopamine Neuron Firing. To elucidate whymemantine increased dopamine neuron firing rate only when

recorded with patch-clamp electrodes, we reasoned that in-tracellular dialysis associated with this technique might beresponsible for some form of alteration in the cellular phys-iology, unmasking a novel property of the drug, independentof its ability to antagonize NMDA receptor-mediated re-sponses.

We first increased the calcium buffering properties of thepatch-clamp electrode filling solution using high (10 mM)EGTA; however, 100 �M memantine still increased the neu-ronal firing (n � 3; data not shown). In contrast, when thedopamine neurons were recorded with high (10 mM) intra-cellular ATP, no significant change in their firing rate wasobserved in 100 �M memantine (from 1.6 � 0.2 to 1.7 � 0.2Hz in memantine; n � 8, p � 0.2 paired t test) (Fig. 4, A andC).

Block of KATP Channels Occludes the Effects of Me-mantine on Dopamine Neuron Firing. Dopamine neu-rons are highly enriched in KATP channels, opened whenintracellular ATP levels are reduced under conditions of met-abolic stress (Mercuri et al., 1994; Guatteo et al., 1998;Marinelli et al., 2000; Liss et al., 2005). Thus, we hypothe-sized that memantine depolarized the dopamine neurons byblocking KATP channels, partially opened in neurons re-corded using a standard 2 mM ATP filling solution.

In agreement with this hypothesis, we found that perfu-sion with the KATP channel antagonist tolbutamide (300 �M)did not change the firing rate of dopamine neurons recordedwith 10 mM intracellular ATP (from 1.7 � 0.1 to 1.6 � 0.1 Hzin tolbutamide; n � 9, p � 0.3 paired t test), whereas itincreased the firing rate of neurons recorded using 2 mMATP (from 1.3 � 0.2 to 2.5 � 0.3 Hz in tolbutamide, p � 0.01paired t test, n � 6) (Fig. 4C). Moreover, in the same neurons,no further increase in the frequency of action potential dis-charge was observed when 100 �M memantine was added inthe continuous presence of tolbutamide (2.5 � 0.3 Hz intolbutamide and memantine; p � 0.54 paired t test, n � 6)(Fig. 4, B and C).

Fig. 3. Lack of effect of memantine on dopamine neuron firing rate using the extracellular multielectrode array technique. Plot of the averagespontaneous firing frequency (mean � S.E.M.) against time (10-s bin size) obtained from the raster plots shown above. On top, raster plots are shownof single unit action potentials obtained from 44 neurons sensitive to 30 �M dopamine (left) in a single midbrain slice. The same neurons wereinsensitive to 100 �M memantine (right).

724 Giustizieri et al.

at Fondazione S

anta Lucia - Biblioteca on N

ovember 19, 2012

jpet.aspetjournals.orgD

ownloaded from

Memantine Reduces KATP-Mediated Outward Cur-rents Induced by Hypoxia. The previous experimentsstrongly suggest that memantine reduces the opening of tol-butamide-sensitive KATP channels. To further confirm thishypothesis, we directly tested the involvement of KATP con-ductances by briefly exposing the slice (2–3 min) to a hypoxicmedium. These experiments were performed in the continu-ous presence of 10 �M MK-801 to rule out indirect actionsthrough NMDA receptors. As previously shown (Guatteo etal., 1998), this experimental protocol induces the develop-ment of a KATP-dependent outward current in dopamineneurons recorded in voltage-clamp mode. As shown in Fig. 5,bath perfusion of 30 �M memantine reversibly reduced hy-

poxia-induced outward current (IHYPO) to 73.6 � 3.8% ofcontrol (n � 12). Increasing the memantine concentration to100 �M produced further inhibition of IHYPO to 47.9 � 3.6%of control (n � 11). Moreover, IHYPO was abolished by tolbu-tamide (300 �M, n � 4), thus confirming an effect of meman-tine on KATP conductances.

Multielectrode Recordings of Firing Rate Increasein Hypoglycemic Conditions. Exposure of the dopamineneurons to a hypoglycemic medium inhibits SNc dopamineneurons through activation of KATP channels (Marinelli etal., 2000). Therefore, we tested whether the spontaneousfiring rate of dopamine neurons recorded with the multielec-trode system could be increased by memantine in hypoglyce-

Fig. 4. Involvement of KATP conduc-tances in the effects of memantine ondopamine neuron firing rate. A, tracerecords from a dopamine neuron re-corded with a patch pipette filling so-lution containing 10 mM ATP. Expo-sure to 100 �M memantine did notresult in significant alteration of thespontaneous firing rate. B, tracerecords from a dopamine neuron re-corded with a patch pipette filling so-lution containing 2 mM ATP. Perfu-sion with the KATP channel antagonist300 �M tolbutamide produced an in-crease in the firing rate, and no fur-ther increase resulted from perfusionof 100 �M memantine in the continu-ous presence of tolbutamide. C, histo-gram of dopamine neuron firing rate(mean � S.E.M.) recorded using 10mM ATP (left) or 2 mM ATP (right) inthe patch pipette filling solution. Thespontaneous firing rate did not changesignificantly in neurons recorded in10 mM ATP exposed to 100 �M me-mantine or 300 �M tolbutamide. Con-versely, a significant increase was in-duced by 300 �M tolbutamide inneurons recorded with 2 mM ATP, butno additional increase in the firingrate was observed in the same neu-rons when 100 �M memantine wasadded in the continuous presence of300 �M tolbutamide. ��, p � 0.01;paired t test.

Fig. 5. Inhibition by memantine ofKATP-dependent hypoxia-induced out-ward current (IHYPO). All the experi-ments were conducted in the continu-ous presence of 10 �M MK-801. A,trace record obtained from a singledopamine neuron showing IHYPO gen-erated in response to repeated 2-minexposures (vertical arrows) to ACSFsaturated with 95% N2/5% CO2. Me-mantine, 30 and 100 �M, reversiblyinhibited IHYPO in a dose-dependentmanner. Perfusion with the KATPchannel blocker 300 �M tolbutamidecompletely abolished IHYPO, unmask-ing a small inward current followedby an electrogenic posthypoxic out-ward response (Mercuri et al., 1994;Guatteo et al., 1998). B, histogram ofIHYPO amplitude expressed as per-centage of control (mean � S.E.M.),following exposure to 30 and 100 �Mmemantine and 300 �M tolbutamide.

Memantine Inhibition of KATP Channels 725

at Fondazione S

anta Lucia - Biblioteca on N

ovember 19, 2012

jpet.aspetjournals.orgD

ownloaded from

mic conditions. To avoid an excessive metabolic stress, weperfused the slices in 1 mM glucose, a threshold dose fordopamine neuron hyperpolarization (Marinelli et al., 2000).As shown in Fig. 6, when the recordings were obtained inACSF containing 1 mM glucose, a low overall firing rate wasdetected (compare with Fig. 3). In these experimental condi-tions, 100 �M memantine reversibly increased the firing rateof presumed dopamine neurons. When glucose was increasedto 2 mM in the same slice, the overall firing rate increased,and 100 �M memantine became ineffective, in accordancewith what was previously observed in normoglycemic condi-tions (Fig. 3). Similar results were obtained from 47 pre-sumed dopamine neurons from four midbrain slices perfusedin 1 mM glucose. Their basal firing rate increased from0.37 � 0.09 to 1.51 � 0.28 Hz (p � 0.0001 paired t test) in 100�M memantine.

In accordance with the hypothesis of an involvement ofKATP channels, we found that 300 �M tolbutamide increasedthe spontaneous firing rate in hypoglycemic conditions, andno further increase was induced by 100 �M memantine (Fig.7). When glucose was increased to 10 mM in the same slice,the overall firing rate increased, and 300 �M tolbutamidebecame ineffective (Fig. 7). In 74 presumed dopamine neu-rons recorded from four midbrain slices perfused in 1 mMglucose, 300 �M tolbutamide increased their firing rate from0.51 � 0.05 to 1.72 � 0.12 Hz (p � 0.0001 paired t test).Exposure to 100 �M memantine in hypoglycemia and tolbu-tamide did not produce a further increase in firing rate(1.56 � 0.11 Hz).

DiscussionOur results provide a direct electrophysiological demon-

stration of a novel mechanism of action of the amantadinederivative memantine. We have shown that, in addition to itsexpected property of NMDA receptor antagonist, memantinereduces neuronal hyperpolarization mediated by the openingof KATP channels in dopamine neurons of the SNc in a dose-dependent manner.

Three lines of evidence point at this conclusion. First, wehave shown that action potential discharge of the dopamineneurons increased in frequency on memantine perfusion indopamine neurons recorded using the patch-clamp whole-celltechnique. This effect was evident when the internal patch-clamp electrode solution contained 2 mM ATP, whereas un-der high (10 mM) internal ATP, memantine was ineffective.During whole-cell recordings, ATP internal concentration islargely imposed by the filling solution composition; thus, theconcentration of ATP in the patch-clamp electrode greatlyaffects the degree of KATP channel opening in response tometabolic stress. For example, previous experiments fromour laboratory have shown that dopamine neurons exposedto a hypoxic medium develop a KATP-dependent current onlywhen they are recorded with 2 mM ATP in the patch-clampelectrode solution, whereas no KATP outward current is in-duced by hypoxia if internal ATP is set to 10 mM (Guatteo etal., 1998). On this basis, we hypothesized that memantineincreased the firing rate of dopamine neurons recorded inwhole-cell conditions by closing KATP channels partially openunder 2 mM internal ATP. This hypothesis was confirmed bythe observation that the KATP channel antagonist tolbut-amide produced an analogous increase in the firing rate ofdopamine neurons recorded with 2 mM internal ATP,whereas the same antagonist was ineffective when intracel-lular ATP was set to 10 mM. Moreover, the frequency in-crease induced by tolbutamide in 2 mM ATP occluded anyadditional increase by memantine.

The second point supporting our hypothesis derives fromexperiments conducted using the noninvasive, extracellular8 � 8 planar multielectrode device. Presumed dopamine neu-rons, identified based on their sensitivity to exogenouslyapplied dopamine, were insensitive to memantine in controlexperimental conditions. However, memantine markedly in-creased the dopamine neuron firing rate in slices perfused in1 mM glucose ACSF instead of 10 mM. Previous experimentsfrom our laboratory showed that 1 mM glucose is a thresholdhypoglycemic condition for the development of a KATP-depen-

Fig. 6. Increase of dopamine neuronfiring rate by memantine using theextracellular multielectrode arraytechnique in hypoglycemic conditions.Plot of the average spontaneous firingfrequency (mean � S.E.M.) againsttime (10-s bin size) obtained from theraster plots shown above. On top, ras-ter plots are shown of single unit ac-tion potentials obtained from 22 neu-rons sensitive to 30 �M dopamine (notshown) in a single midbrain slice. Per-fusion of 100 �M memantine in ACSFcontaining 1 mM glucose caused amarked, reversible increase in the fir-ing rate (left). Increasing the glucoseconcentration to 2 mM caused anoverall recovery of the firing fre-quency, and no effect by memantinewas observed in this experimentalcondition (right).

726 Giustizieri et al.

at Fondazione S

anta Lucia - Biblioteca on N

ovember 19, 2012

jpet.aspetjournals.orgD

ownloaded from

dent hyperpolarization in dopamine neurons of the SNc(Marinelli et al., 2000). Conceivably, the reduced basal firingrate in 1 mM glucose, compared with controls, was caused bya more hyperpolarized state of the dopamine neuronal pop-ulation because of the opening of KATP conductances in 1 mMglucose. Thus, memantine increased dopamine neuron firingrate by closing these constitutively open KATP channels. Thishypothesis was confirmed by the observation that the KATP

channel antagonist tolbutamide mimicked the effects of me-mantine and occluded additional increase in the firing rateby memantine in hypoglycemia. Both memantine and tolbu-tamide were ineffective under higher glucose concentrationbecause no tonic KATP-dependent current was present inthese conditions.

Indeed, these experiments also suggest that the tonic ac-tivation KATP conductance observed in our patch-clampwhole-cell recordings, using 2 mM internal ATP, is a non-physiological phenomenon because no such tonic KATP cur-rent seems to be evident using the less invasive multielec-trode extracellular technique.

Finally, the third, more direct, piece of evidence of meman-tine effects on KATP conductances derives from our experi-ments using brief exposures of the dopamine neurons tohypoxia. Dopamine neurons respond to oxygen deprivationwith an early hyperpolarization, largely because of the open-ing of KATP channels (Mercuri et al., 1994; Guatteo et al.,1998; Geracitano et al., 2005). Memantine reversibly reducedthis tolbutamide-sensitive outward current, even in the pres-ence the open-channel NMDA receptor antagonist MK-801,thus ruling out any possible indirect effect through inhibitionof NMDA receptors.

As shown in Fig. 2, memantine not only increased thefiring rate of the dopamine neurons but also changed occa-sionally their firing mode from tonic to bursting behavior.This change in the firing pattern may have important func-tional implications because burst firing of the dopamine neu-rons has been associated with increased release of dopamine

in the areas of nigral projection (Gonon and Buda, 1985;Suaud-Chagny et al., 1992; Floresco et al., 2003; Phillips etal., 2003). However, such a change was observed using highdoses of memantine, and more importantly we never ob-served burst firing in neurons recorded using the extracellu-lar multielectrode technique. Probably burst generation bymemantine is linked to some unspecific effect, amplified by amore invasive technique like whole-cell patch-clamp record-ing. However, we cannot exclude the possibility that thisburst-inducing effect of memantine emerges under morecomplex pathological alterations of the intracellular compo-sition, somehow mimicked in our patch-clamp recording con-ditions.

Clinical Relevance. A question arising from our experi-mental observation is whether the above effects, obtainedfrom dopamine neurons recorded in an in vitro slice prepa-ration, also occur in patients under memantine treatment.According to in vivo measurements, treatment with a stan-dard 20-mg daily dose of memantine should result in asteady-state brain level of this drug approaching the lowmicromolar range (Hesselink et al., 1999; Danysz et al.,2000); therefore, it is suggested that results obtained usinghigher concentrations may reflect unspecific and nonclini-cally relevant effects (Chen and Lipton, 2006). However, at-tention should be given to the limiting factor offered in a slicepreparation by the need of the drug to diffuse within thebrain tissue during a relatively short time of bath perfusion.Consequently, the effective concentration of the drug reach-ing the neuronal membranes is not the same as that of theextracellular medium. Indeed, an almost complete block ofNMDA receptor-mediated responses can be achieved in iso-lated cell cultures at concentrations of memantine close to 10�M (Chen et al., 1992), whereas in our brain slice prepara-tion, inhibition of NMDA receptor-mediated responseshardly exceeded 50% at 30 �M memantine or 75% at aconcentration of 100 �M (Fig. 1B). This difference supportsthe clinical relevance of our present results, as they should be

Fig. 7. Increase of dopamine neuronfiring rate by tolbutamide using theextracellular multielectrode arraytechnique in hypoglycemic conditions.Plot of the average spontaneous firingfrequency (mean � S.E.M.) againsttime (10-s bin size) obtained from theraster plots shown above. On top, ras-ter plots are shown of single unit ac-tion potentials obtained from 26 neu-rons sensitive to a 1-min exposure todopamine (30 �M, not shown) in asingle midbrain slice. Perfusion of 300�M tolbutamide in ACSF containing 1mM glucose caused a marked increasein the firing rate. In this condition,100 �M memantine did not producefurther increase in action potentialfiring (left). Restoring normoglycemicconditions caused an overall recoveryof the firing frequency, and no effectby tolbutamide was observed in thisexperimental condition (right).

Memantine Inhibition of KATP Channels 727

at Fondazione S

anta Lucia - Biblioteca on N

ovember 19, 2012

jpet.aspetjournals.orgD

ownloaded from

compared with those obtained on isolated neurons usinglower concentrations of memantine. In addition, we have alsoshown that the effects of 100 �M memantine were completelyprevented in patch-clamp recordings obtained with a 10 mMATP filling solution, and the same concentration of meman-tine did not modify the firing rate of the dopamine neuronsusing the multielectrode device in normoglycemic conditions.Both of these observations rule out unspecific effects simplycaused by a high dose of memantine.

Memantine and Parkinson’s Disease. According to ourobservations, memantine does not affect the basal firing ac-tivity of the dopamine neurons in physiological conditions,whereas in conditions of metabolic stress, a significant effectof memantine emerges, resulting in recovery of firing activityof previously silenced dopamine neurons. This property maybe particularly relevant in terms of firing dependent dopa-mine release and in relation to prevention of neuronal loss inParkinson’s disease. The activity of complex I of the mito-chondrial respiratory chain is reduced in dopamine neuronsof parkinsonian patients (Schapira, 2001), and drugs actingas inhibitors of complex I, like rotenone or 1-methyl-4-phe-nyl-1,2,3,6-tetrahydropyridine, induce dopamine neuron de-generation (Betarbet et al., 2000; Przedborski and Vila,2003). Inhibition of mitochondrial complex I may lead toreduction of ATP and increased production of reactive oxygenspecies (Hoglinger et al., 2003; Testa et al., 2005), causing theopening of KATP channels and dopamine neuron silencing(Avshalumov et al., 2005; Berretta et al., 2005; Geracitano etal., 2005). A recent report by Liss et al. (2005) proposed thatthe selective vulnerability of the SNc dopamine neurons inParkinson’s disease is casually correlated with the opening ofKATP conductances in these neurons; thus, the presence offunctional KATP channels promotes the selective loss of SNcdopamine neurons in both a genetic model of Parkinson’sdisease and in response to mitochondrial complex I inhibi-tion. At present, the mechanism through which KATP channelopening contributes to dopamine neuron degeneration is stillunclear. There is evidence that increasing neuronal excitabil-ity protects dopamine neurons from degeneration (Salthun-Lassalle et al., 2004); for this reason, Liss et al. (2005) pro-posed that drugs acting at KATP channels of SNc dopamineneurons should cause a recovery from their functional silenc-ing, thus providing a clinical benefit in the treatment ofParkinson’s disease. Indeed, memantine has been shown toprevent cell death associated with Parkinson’s disease(Danysz et al., 1997; Lange et al., 1997; Merello et al., 1999;Kucheryanu and Kryzhanovskii, 2000), although preventionof excitotoxic neuronal damage through an uncompetitiveinhibition of NMDA receptors has been proposed as its un-derlying mechanism of action. Our results show that meman-tine does inhibit NMDA responses in the SNc; however, me-mantine may also have beneficial results in Parkinson’sdisease patients because it reduces dopamine neuron silenc-ing through closure of KATP conductances.

Acknowledgments

We thank Lundbeck Italia SpA for the gift of memantine.

ReferencesAracava Y, Pereira EF, Maelicke A, and Albuquerque EX (2005) Memantine blocks

�7 nicotinic acetylcholine receptors more potently than N-methyl-D-aspartate re-ceptors in rat hippocampal neurons. J Pharmacol Exp Ther 312:1195–1205.

Avshalumov MV, Chen BT, Koos T, Tepper JM, and Rice ME (2005) Endogenoushydrogen peroxide regulates the excitability of midbrain dopamine neurons viaATP-sensitive potassium channels. J Neurosci 25:4222–4231.

Berretta N, Freestone PS, Guatteo E, de Castro D, Geracitano R, Bernardi G,Mercuri NB, and Lipski J (2005) Acute effects of 6-hydroxydopamine on dopami-nergic neurons of the rat substantia nigra pars compacta in vitro. Neurotoxicology26:869–881.

Betarbet R, Sherer TB, MacKenzie G, Garcia-Osuna M, Panov AV, and GreenamyreJT (2000) Chronic systemic pesticide exposure reproduces features of Parkinson’sdisease. Nat Neurosci 3:1301–1306.

Chen HS and Lipton SA (2006) The chemical biology of clinically tolerated NMDAreceptor antagonists. J Neurochem 97:1611–1626.

Chen HS, Pellegrini JW, Aggarwal SK, Lei SZ, Warach S, Jensen FE, and Lipton SA(1992) Open-channel block of N-methyl-D-aspartate (NMDA) responses by meman-tine: therapeutic advantage against NMDA receptor-mediated neurotoxicity.J Neurosci 12:4427–4436.

Danysz W, Parsons CG, Kornhuber J, Schmidt WJ, and Quack G (1997) Aminoada-mantanes as NMDA receptor antagonists and antiparkinsonian agents—preclinical studies. Neurosci Biobehav Rev 21:455–468.

Danysz W, Parsons CG, and Quack G (2000) NMDA channel blockers: memantineand amino-aklylcyclohexanes–in vivo characterization. Amino Acids 19:167–172.

Ditzler K (1991) Efficacy and tolerability of memantine in patients with dementiasyndrome. A double-blind, placebo controlled trial. Arzneimittelforschung 41:773–780.

Fleischhacker WW, Buchgeher A, and Schubert H (1986) Memantine in the treat-ment of senile dementia of the Alzheimer type. Prog Neuropsychopharmacol BiolPsychiatry 10:87–93.

Floresco SB, West AR, Ash B, Moore H, and Grace AA (2003) Afferent modulation ofdopamine neuron firing differentially regulates tonic and phasic dopamine trans-mission. Nat Neurosci 6:968–973.

Gardoni F and Di Luca M (2006) New targets for pharmacological intervention in theglutamatergic synapse. Eur J Pharmacol 545:2–10.

Geracitano R, Tozzi A, Berretta N, Florenzano F, Guatteo E, Viscomi MT, Chiolo B,Molinari M, Bernardi G, and Mercuri NB (2005) Protective role of hydrogenperoxide in oxygen-deprived dopaminergic neurones of the rat substantia nigra.J Physiol 568:97–110.

Gonon FG and Buda MJ (1985) Regulation of dopamine release by impulse flow andby autoreceptors as studied by in vivo voltammetry in the rat striatum. Neuro-science 14:765–774.

Grillner P and Mercuri NB (2002) Intrinsic membrane properties and synapticinputs regulating the firing activity of the dopamine neurons. Behav Brain Res130:149–169.

Guatteo E, Mercuri NB, Bernardi G, and Knopfel T (1998) Intracellular sodium andcalcium homeostasis during hypoxia in dopamine neurons of rat substantia nigrapars compacta. J Neurophysiol 80:2237–2243.

Hesselink MB, De Boer BG, Breimer DD, and Danysz W (1999) Brain penetrationand in vivo recovery of NMDA receptor antagonists amantadine and memantine:a quantitative microdialysis study. Pharm Res 16:637–642.

Hoglinger GU, Carrard G, Michel PP, Medja F, Lombes A, Ruberg M, Friguet B, andHirsch EC (2003) Dysfunction of mitochondrial complex I and the proteasome:interactions between two biochemical deficits in a cellular model of Parkinson’sdisease. J Neurochem 86:1297–1307.

Kucheryanu VG and Kryzhanovskii GN (2000) Effect of glutamate and antagonistsof N-methyl-D-aspartate receptors on experimental parkinsonian syndrome inrats. Bull Exp Biol Med 130:629–632.

Lange KW, Kornhuber J, and Riederer P (1997) Dopamine/glutamate interactions inParkinson’s disease. Neurosci Biobehav Rev 21:393–400.

Lipton SA (2004) Failures and successes of NMDA receptor antagonists: molecularbasis for the use of open-channel blockers like memantine in the treatment of acuteand chronic neurologic insults. NeuroRx 1:101–110.

Lipton SA (2006) Paradigm shift in neuroprotection by NMDA receptor blockade:memantine and beyond. Nat Rev Drug Discov 5:160–170.

Liss B, Haeckel O, Wildmann J, Miki T, Seino S, and Roeper J (2005) K-ATPchannels promote the differential degeneration of dopaminergic midbrain neurons.Nat Neurosci 8:1742–1751.

Marinelli S, Bernardi G, Giacomini P, and Mercuri NB (2000) Pharmacologicalidentification of the K� currents mediating the hypoglycemic hyperpolarization ofrat midbrain dopaminergic neurones. Neuropharmacology 39:1021–1028.

Maskell PD, Speder P, Newberry NR, and Bermudez I (2003) Inhibition of human �7nicotinic acetylcholine receptors by open channel blockers of N-methyl-D-aspartatereceptors. Br J Pharmacol 140:1313–1319.

Mercuri NB, Bonci A, Johnson SW, Stratta F, Calabresi P, and Bernardi G (1994)Effects of anoxia on rat midbrain dopamine neurons. J Neurophysiol 71:1165–1173.

Merello M, Nouzeilles MI, Cammarota A, and Leiguarda R (1999) Effect of meman-tine (NMDA antagonist) on Parkinson’s disease: a double-blind crossover random-ized study. Clin Neuropharmacol 22:273–276.

Netzer R and Bigalke H (1990) Memantine reduces repetitive action potential firingin spinal cord nerve cell cultures. Eur J Pharmacol 186:149–155.

Phillips PE, Stuber GD, Heien ML, Wightman RM, and Carelli RM (2003) Subseconddopamine release promotes cocaine seeking. Nature 422:614–618.

Przedborski S and Vila M (2003) The 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridinemouse model: a tool to explore the pathogenesis of Parkinson’s disease. Ann N YAcad Sci 991:189–198.

Rammes G, Rupprecht R, Ferrari U, Zieglgansberger W, and Parsons CG (2001) TheN-methyl-D-aspartate receptor channel blockers memantine, MRZ 2/579 and otheramino-alkyl-cyclohexanes antagonise 5-HT3 receptor currents in cultured HEK-293 and N1E-115 cell systems in a non-competitive manner. Neurosci Lett 306:81–84.

728 Giustizieri et al.

at Fondazione S

anta Lucia - Biblioteca on N

ovember 19, 2012

jpet.aspetjournals.orgD

ownloaded from

Reiser G, Binmoller FJ, and Koch R (1988) Memantine (L-amino-3, 5-dimethylada-mantane) blocks the serotonin-induced depolarization response in a neuronal cellline. Brain Res 443:338–344.

Salthun-Lassalle B, Hirsch EC, Wolfart J, Ruberg M, and Michel PP (2004) Rescueof mesencephalic dopaminergic neurons in culture by low-level stimulation ofvoltage-gated sodium channels. J Neurosci 24:5922–5930.

Schapira AH (2001) Causes of neuronal death in Parkinson’s disease. Adv Neurol86:155–162.

Sonkusare SK, Kaul CL, and Ramarao P (2005) Dementia of Alzheimer’s disease andother neurodegenerative disorders—memantine, a new hope. Pharmacol Res 51:1–17.

Suaud-Chagny MF, Chergui K, Chouvet G, and Gonon F (1992) Relationship be-tween dopamine release in the rat nucleus accumbens and the discharge activity

of dopaminergic neurons during local in vivo application of amino acids in theventral tegmental area. Neuroscience 49:63–72.

Testa CM, Sherer TB, and Greenamyre JT (2005) Rotenone induces oxidative stressand dopaminergic neuron damage in organotypic substantia nigra cultures. BrainRes Mol Brain Res 134:109–118.

Volbracht C, van Beek J, Zhu C, Blomgren K, and Leist M (2006) Neuroprotectiveproperties of memantine in different in vitro and in vivo models of excitotoxicity.Eur J Neurosci 23:2611–2622.

Address correspondence to: Nicola Berretta, Fondazione Santa LuciaI.R.C.C.S., Via del Fosso di Fiorano, 64, 00143 Rome, Italy. E-mail:n.berretta@hsantalucia.it

Memantine Inhibition of KATP Channels 729

at Fondazione S

anta Lucia - Biblioteca on N

ovember 19, 2012

jpet.aspetjournals.orgD

ownloaded from