1521-0111/83/1/142–156$25.00 http://dx.doi.org/10.1124/mol.112.080986MOLECULAR PHARMACOLOGY Mol Pharmacol 83:142–156, January 2013Copyright ª 2013 by The American Society for Pharmacology and Experimental Therapeutics

Neurounina-1, a Novel Compound That Increases Na+/Ca2+

Exchanger Activity, Effectively Protects against Stroke Damage

Pasquale Molinaro, Maria Cantile, Ornella Cuomo, Agnese Secondo, Anna Pannaccione,Paolo Ambrosino, Giuseppe Pignataro, Ferdinando Fiorino, Beatrice Severino, Elena Gatta,Maria José Sisalli, Marco Milanese, Antonella Scorziello, Giambattista Bonanno, Mauro Robello,Vincenzo Santagada, Giuseppe Caliendo, Gianfranco Di Renzo, and Lucio AnnunziatoDivision of Pharmacology, Department of Neuroscience, School of Medicine, “Federico II” University of Naples-National Instituteof Neuroscience, Naples, Italy (P.M., M.C., O.C., A.Se., A.P., P.A., G.P., M.J.S., A.Sc., G.D.R., L.A.); Department ofPharmaceutical and Toxicological Chemistry, Faculty of Pharmacy, “Federico II” University of Naples, Naples, Italy(F.F., B.S., V.S., G.C.); and Department of Physics (E.G., M.R.) and Department of Experimental Medicine, Section ofPharmacology and Toxicology, University of Genoa, Genoa, Italy (M.M., G.B.)

Received July 5, 2012; accepted October 11, 2012

ABSTRACTPrevious studies have demonstrated that the knockdown orknockout of the three Na1/Ca21 exchanger (NCX) isoforms, NCX1,NCX2, and NCX3, worsens ischemic brain damage. This suggeststhat the activation of these antiporters exerts a neuroprotective ac-tion against stroke damage. However, drugs able to increase theactivity of NCXs are not yet available. We have here succeeded insynthesizing a new compound, named neurounina-1 (7-nitro-5-phenyl-1-(pyrrolidin-1-ylmethyl)-1H-benzo[e][1,4]diazepin-2(3H)-one), providedwith an high lipophilicity index and able to increaseNCX activity. Ca21 radiotracer, Fura-2 microfluorimetry, and patch-clamp techniques revealed that neurounina-1 stimulatedNCX1 andNCX2 activities with an EC50 in the picomolar to low nanomolarrange, whereas it did not affect NCX3 activity. Furthermore, by usingchimera strategy and site-directed mutagenesis, three specificmolecular determinants of NCX1 responsible for neurounina-1

activitywere identified in thea-repeats. Interestingly,NCX3becameresponsive to neurounina-1 when both a-repeats were replacedwith the corresponding regions of NCX1. In vitro studies showedthat 10 nM neurounina-1 reduced cell death of primary corticalneurons exposed to oxygen-glucose deprivation followed by re-oxygenation. Moreover, in vitro, neurounina-1 also reduced g-aminobutyric acid (GABA) release, enhanced GABAA currents,and inhibited both glutamate release and N-methyl-D-aspartatereceptors. More important, neurounina-1 proved to have a widetherapeutic window in vivo. Indeed, when administered at dosesof 0.003 to 30 mg/kg i.p., it was able to reduce the infarct volumeof mice subjected to transient middle cerebral artery occlusioneven up to 3 to 5 hours after stroke onset. Collectively, the presentstudy shows that neurounina-1 exerts a remarkable neuroprotec-tive effect during stroke and increases NCX1 and NCX2 activities.

IntroductionThe Na1/Ca21 exchanger (NCX) is an important bidirec-

tional plasma membrane transporter that is driven by

electrochemical gradient. Under physiologic conditions, itmainly works in the forward mode by coupling the extrusionof one Ca21 ion with the influx of three Na1 ions into the cells.Under certain physiologic and pathophysiologic conditions,when the electrochemical gradient reverts, NCX works in thereverse mode, thus coupling the extrusion of three Na1 ionswith the influx of one Ca21 ion (Blaustein and Lederer, 1999;Philipson and Nicoll, 2000; Annunziato et al., 2004).NCX belongs to a multigene family comprising three

isoforms, named NCX1, NCX2, and NCX3, which are differen-tially distributed through the body. NCX1 is ubiquitouslyexpressed in all tissues, NCX2 is mainly restricted to the brain,

This work was supported by Ministero della Salute Ricerca Finalizzata 2006,Ricerca Oncologica 2006, Progetto Strategico 2007, Progetto Ordinario 2007, andRicerca Finalizzata 2009 [Grants RF-FSL-352059, RF-IDI-2006-367185, RF-06711, RF-SDN-2007-666932, and RF-0832] and by Ministero dell’Istruzione,dell’Università e della Ricerca Programma Operativo Nazionale and COFIN2008[Grants PON_01602 and 20089BARSR_002], all to L.A.

P.M., M.C., and O.C. contributed equally to this work.Neurounina-1 is under patent protection (FI2010A000233) obtained from

Italian Office for Patents and Trademarks of the Ministry of Industry andTrade.

dx.doi.org/10.1124/mol.112.080986.

ABBREVIATIONS: AM, acetoxymethyl ester; BHK, baby hamster kidney; CBF, cerebral blood flow; DL-TBOA, DL-threo-b-benzyloxyaspartic acid;DMSO, dimethyl sulfoxide; ER, endoplasmic reticulum; FDA, fluorescein diacetate; GABAA, g-aminobutyric acid receptor A; HPLC, high-performance liquid chromatography KB-R7943, 2-[2-[4-(4-nitrobenzyloxyl)phenyl]ethyl]isothiourea methanesulfonate; MCA, middle cerebral artery;MK-801, (5S,10R)-(1)-5-Methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-imine maleate; MTT, 3[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazoliumbromide; NCX, Na1/Ca21 exchanger; NMDA, N-methyl-D-aspartate; NMDG, N-methyl-D-glucamine; OGD, oxygen and glucose deprivation; PI, propidiumiodide; SEA0400, 2-[4-[(2,5-difluorophenyl)methoxy]phenoxy]-5-ethoxyaniline; SKF89976A, 1-(4,4-Diphenyl-3-butenyl)-3-piperidinecarboxylic acid; SM-15811, 4-phenyl-3-[(N-benzyl)-4-pyperidine]-3,4-dihydro-2(1H)-quinazolinone; SN-6, 2-[4-(4-nitrobenzyloxy)benzyl]thiazolidine-4-carboxylic acid ethylester; TEA, tetraethylammonium; TFA, trifluoroacetic acid; tMCAO, transient middle cerebral artery occlusion; TTX, tetrodotoxin; YM-244769, N-(3-aminobenzyl)-6-[4-[(3-fluorobenzyl)oxy]-phenoxy]nicotinamide.

142

at ASPE

T Journals on M

ay 13, 2018m

olpharm.aspetjournals.org

Dow

nloaded from

andNCX3 is expressed exclusively in brain and skeletalmuscles(Quednau et al., 1997). These NCX isoforms have similar mo-lecular topologies comprising nine transmembrane segmentsand a large cytoplasmic loop that regulates their activity (Nicollet al., 1999). All NCX isoforms share two highly conservedrepeat sequences, named a1 and a2, that are involved in theionic transport process (Nicoll et al., 1996).The primary function of NCX is to extrude Ca21 from the

cell, in the forward mode, using [Na1] gradient. However,under some physiologic conditions, such as the occurrence ofan action potential or of spontaneous [Ca21]i oscillations,NCX could revert its mode of operation, thus participating inendoplasmic reticulum (ER)-Ca21 refilling. In fact, it has beenrecently demonstrated that spontaneous [Ca21]i oscillationsinduced by inositol 1,4,5-triphosphate receptor stimulationmight lead to the activation of nonselective cation channels,which causes Na1 influx into the junctional cytosol micro-domains. This influx facilitates the entrance of Ca21 throughNCXworking in the reverse mode (Fameli et al., 2007). On theother hand, under pathologic conditions, the function of NCXismore complex.More specifically, in the early phase of neuronalanoxic insult, the Na1/K1-ATPase blockade increases [Na1]i,which in turn induces NCX to reverse its mode of operation.Although during this phase NCX causes an increase in [Ca21]i,its effect on neurons appears beneficial for two reasons. First, bypromoting Ca21 influx, NCX favors Ca21 refilling into the ER,which is depleted by anoxia followed by reoxygenation, thusallowing neurons to delay ER stress (Sirabella et al., 2009).Second, by eliciting the decrease in [Na1]i overload, NCXprevents cell swelling and death (Annunziato et al., 2007).Conversely, in the later phase of neuronal anoxia, when [Ca21

]i overload takes place, NCX forward mode of operationcontributes to the lowering of [Ca21]i, thus protecting neuronsfrom [Ca21]i-induced neurotoxicity (Annunziato et al., 2004).Furthermore, NCX is involved in some serious diseasescharacterized by a loss of ion homeostasis control, includingAlzheimer’s disease, multiple sclerosis, and epilepsy (Craneret al., 2004; Ketelaars et al., 2004; Pannaccione et al., 2012).Recently, experiments in ischemic rats treated with inhibitorsof NCX activity or synthesis, together with experiments inknockout mice for the ncx2 or ncx3 gene, have demonstratedthat the blockade of NCX protein synthesis or activityworsens ischemic brain damage by dysregulating Na1 andCa21 homeostasis (Pignataro et al., 2004a,b; Jeon et al., 2008;Molinaro et al., 2008). Therefore, in recent years there hasbeen great interest in the identification of new compoundscapable of increasing NCX activity to limit the extension ofischemic brain damage (Annunziato et al., 2009). To date,only NCX inhibitors are available including 3-amino-6-chloro-5-[(4-chloro-benzyl)amino]-N-[[(2,4-dimethylbenzyl)amino]iminomethyl]-pyrazinecarboxamide (CB-DMB) (Se-condo et al., 2009), KB-R7943 (Watano et al., 1999), SEA0400(Matsuda et al., 2001), SM-15811 (Hasegawa et al., 2003), SN-6 (Iwamoto et al., 2004a), and YM-244769 (Iwamoto and Kita,2006). These inhibitors have a number of interesting features.In particular, KB-R7943 preferentially inhibits NCX3 morethan NCX1 and NCX2 (Iwamoto et al., 2001), whereasSEA0400 and SN-6 preferentially block NCX1 rather thanNCX2 and NCX3 (Iwamoto et al., 2004a,b). However, despitetheir potency, these compounds possess some nonspecific actionsagainst other ion channels and receptors (Pintado et al., 2000;Reuter et al., 2002).

To obtain an activator of the NCX isoforms, we modified thestructure of one of the most potent NCX inhibitors, SM-15811,thereby synthesizing a new compound, 7-nitro-5-phenyl-1-(pyrrolidin-1-ylmethyl)-1H-benzo[e][1,4]diazepin-2(3H)-one,that we named neurounina-1. In this study, we investigatedthe effect of this newly synthesized compound on NCX1,NCX2, and NCX3 activity in the forward and reverse modes ofoperation by means of Ca21 radiotracer, Fura-2 microfluorim-etry, and patch-clamp techniques, and thereafter, with thehelp of chimera strategy, deletion, and site-directed muta-genesis, we identified the molecular determinants of thiscompound on NCX structure.More important, since NCX activity is involved in stroke

pathophysiology, we examined the putative protective effectsof neurounina-1 on in vitro and in vivo models of cerebralischemia.

Materials and MethodsProcedures for the Synthesis of Neurounina-1

Formaldehyde (0.1 ml, 3.5 mmol) was added to pyrrolidine (0.3 ml,3.5 mmol) at 0°C. The obtained solution was then added tonitrazepam (0.1 g, 0.35 mmol), previously dissolved in glacial aceticacid (5 ml). Next, the reaction mixture was placed in a closed reactionvessel, equipped with a temperature control unit, and irradiatedaccording to the following parameters: initial power, 500 W; initialtime, 1 minute (ramping); final power, 500 W; temperature, 80°C;reaction time, 15 minutes. Afterward, the reaction mixture wasextracted with 2 N NaOH (three times with 40 ml) and isopropanol/chloroform (1:1 [v/v]). The organic phase was then washed with brine,dried over anhydrous Na2SO4, filtered, and concentrated in vacuo.The obtained residue was purified on a silica gel column eluted withdichloromethane/methanol (8:2 [v/v]), affording 61 mg of purifiedproduct (yield, 48%), which was converted to the correspondingtrifluoroacetate salt by dissolving in 0.1% trifluoroacetic acid (TFA) inH2O/acetonitrile (60:40 [v/v]). Finally, the obtained solution wasfrozen and lyophilized to yield the desired salt.

The final product was analyzed by NMR spectroscopy. In particular,results showed 1H NMR (400 MHz, CDCl3) d 1.20–1.30 (m, 4H),1.90–2.10 (m, 4H), 3.20 (s, 2H), 4.20 (s, 2H), 7.33–7.51 (m, 6H), 8.24 (s,1H), and 8.35 (d, 1H); electrospray ionizationmass spectroscopy analysiscalculated for C20H20N4O3364.4 found [M 1 H]1 365.1, where "s" wasused for singlet and "m" was used for multiplet. The 1H and 13C NMRspectra were recorded on a Varian Mercury Plus 400 MHz instrument(Varian Inc., Palo Alto, CA).

Purity of the product was assessed by analytical reversed-phasehigh-performance liquid chromatography (HPLC) using a C18 column(5 mm, 4.6 � 250 mm, Beckman Coulter Inc, Brea, CA) using thefollowing conditions: eluent A, 0.05% TFA (v/v) in water; eluent B,0.05% TFA (v/v) in acetonitrile; gradient, 10–40% B over 25 minutes;UV detection at 254 nm; flow rate, 1 ml/min. The column wasconnected to a Rheodyne model 7725 injector, a Waters 600 HPLCsystem, a Waters 486 tunable absorbance detector, and a Waters 746chart recorder (Waters Corporation, Milford, MA). Neurounina-1 wasdissolved in DMSO and diluted in the medium to the final concen-trations used. DMSO at the final concentration did not modify ioniccurrents and Na1-dependent Ca21 transport.

Cell Cultures

Baby hamster kidney (BHK) cells stably transfected with caninecardiac NCX1.1, rat brain NCX2.1, and NCX3.3 were a generous giftfrom Dr. Kenneth D. Philipson (University of California, Los Angeles,CA). All BHK cell lines were grown on plastic dishes in a mix ofDulbecco’s modified Eagle’s medium and Ham’s F-12 medium (1:1)(Life Technologies, San Giuliano Milanese, Italy) supplemented with

Neurounina-1 Increases NCX Activity and Protects against Stroke 143

at ASPE

T Journals on M

ay 13, 2018m

olpharm.aspetjournals.org

Dow

nloaded from

5% fetal bovine serum, 100 U/ml penicillin, and 100 mg/mlstreptomycin (Sigma-Aldrich, St. Louis, MO). Cells were cultured ina humidified 5% CO2 atmosphere, and the culture medium waschanged every 2 days. For microfluorimetric and electrophysiologicstudies, cells were seeded on glass coverslips (Thermo FisherScientific, Springfield, NJ) coated with poly-L-lysine (30 mg/ml)(Sigma-Aldrich) and used at least 12 hours after seeding. Granulecells were prepared from 8-day-old rats as previously described(Robello et al., 1993) and studied from the 6th to the 10th day in vitro.Cortical neurons were prepared from brains of Wistar rat pups 2–4days old. The tissue was minced and trypsinized (0.1% for 15 minutesat 37°C), triturated, and plated on poly-D-lysine-coated coverslipsand cultured in Neurobasal medium (Invitrogen) supplemented withB-27 (Invitrogen) and 2 mM L-glutamine. Cells were plated atconcentration of 1.8 � 106 on 25-mm glass coverslips. Cultures weremaintained at 37°C in a humidified atmosphere of 5% CO2 and 95%air, fed twice a week, and maintained for a minimum of 10 days beforeexperimental use (Scorziello et al., 2007).

Generation and Stable Expression of Wild-Type and MutantExchangers

All NCX1/NCX3 chimeras shown in Fig. 3 were constructed byIwamoto’s group, as previously thoroughly described (Iwamoto et al.,2004a). Deletion mutant of NCX1, NCX1Df, was obtained by deleting theamino acid region 241–680 of canineNCX1.1 cDNAwith theQuikChangeSite-Directed Mutagenesis kit (Stratagene, La Jolla, CA). Substitution ofthe amino acid residues in NCX1 was performed by site-directedmutagenesis using the same polymerase chain reaction–based strategy(Stratagene). All mutant exchangers obtained were verified by sequenc-ing both DNA strands (Primm, Milan, Italy).

To stably express chimeric and mutant exchangers, pKCRH orpCEFL plasmids carrying cDNAs were transfected into wild-type BHKcells by the Lipofectamine 2000 (Life Technologies) protocol. Cell cloneshighly expressing Na1/Ca21 exchange activity were selected by a Ca21

-killing procedure with the Ca21 ionophore ionomycin. In the presenceof this ionophore, cells not expressing the exchanger died from Ca21

overload (Iwamoto et al., 1998).

Measurement of Na1-Dependent 45Ca21 Uptake and 45Ca21

Efflux45Ca21 influx into the cells was measured as previously described

(Secondo et al., 2007). After treatments, cells cultured in 24-welldishes were incubated in normal Krebs (in mM): 5.5 KCl, 145 NaCl,1.2 MgCl2, 1.5 CaCl2, 10 glucose, and 10 Hepes-NaOH (pH 7.4),containing 1 mM ouabain and 10 mM monensin at 37°C for 10minutes. Then, 45Ca21 uptake was initiated by switching the normalKrebs medium to Na1-free N-methyl-D-glucamine (NMDG) (in mM):5.5 KCl, 147 NMDG, 1.2 MgCl2, 1.5 CaCl2, 10 glucose, and 10 Hepes-NaOH (pH 7.4), containing 10 mM 45Ca21 (800 MBq/ml) (PerkinElmerLife and Analytical Sciences, Monza, Italy) and 1 mM ouabain. After30 seconds of incubation, cells were washed with an ice-cold solutioncontaining 2 mMLa31 to stop 45Ca21 uptake. Cells were subsequentlylysed with 0.1 N NaOH, and aliquots were taken to determineradioactivity and protein content.

To measure 45Ca21 efflux, cells were loaded with 1 mM 45Ca21 (80MBq/ml) together with 1 mM ionomycin for 60 seconds in normalKrebs. Next, cells were exposed either to a Ca21- and Na1-freesolution—a condition that blocks both intracellular 45Ca21 efflux andextracellular Ca21 influx— or to a Ca21-free plus 2 mM EGTAcontaining 145 mM Na1, a condition that promotes 45Ca21 efflux.Thapsigargin (1 mM) was present in both solutions. 45Ca21 efflux wasstarted by using Ca21-free, Na1-containing solution plus 2 mMEGTA. Cells were exposed to this solution, which promotes 45Ca21

efflux, for 10 seconds. At the time chosen (10 seconds), a very lowefflux was detected in wild-type BHK cells. Na10-dependent

45Ca21

efflux was estimated by subtracting 45Ca21 efflux in Ca21- and Na1

-free solution from that in Ca21-free solution. Cells were subsequentlylysed with 0.1 N NaOH, and aliquots were taken to determineradioactivity and protein content by the Bradford method. The EC50

values for neurounina-1 were obtained by fitting the data with theequation a 1 b � exp(2x/t), where a is the maximal response, b is thebasal response, x is the drug concentration, and t is the EC50.

[Ca21]i Measurement

[Ca21]i was measured by single-cell computer-assisted videoimaging (Secondo et al., 2007). Briefly, BHK cells, grown on glasscoverslips, were loadedwith 6 mMFura-2 acetoxymethyl ester (Fura-2AM) for 30 minutes at 37°C in normal Krebs solution containing thefollowing (in mM): 5.5 KCl, 160 NaCl, 1.2 MgCl2, 1.5 CaCl2, 10glucose, and 10 Hepes-NaOH (pH 7.4). At the end of the Fura-2 AMloading period, the coverslips were placed into a perfusion chamber(Medical Systems, Greenvale, NY) mounted onto the stage of aninverted Zeiss Axiovert 200 microscope (Carl Zeiss, Oberkochen,Germany) equipped with a FLUAR 40� oil objective lens. Theexperiments were carried out with a digital imaging system composedof MicroMax 512BFT cooled CCD camera (Princeton Instruments,Trenton, NJ), LAMBDA 10-2 filter wheeler (Sutter Instruments,Novato, CA), and Meta-Morph/MetaFluor Imaging System software(Universal Imaging, West Chester, PA). After loading, cells werealternatively illuminated at wavelengths of 340 nm and 380 nm bya xenon lamp. The emitted light was passed through a 512-nm barrierfilter. Fura-2 fluorescence intensity was measured every 3 seconds.Forty to sixty-five individual cells were selected and monitoredsimultaneously from each coverslip. All the results are presented ascytosolic Ca21 concentration. Since KD for Fura-2 was assumed to be224 nm, the equation of Grynkiewicz (Grynkiewicz et al., 1985) wasused for calibration. NCX activity was evaluated as Ca21 uptakethrough the reverse mode by switching the normal Krebs medium toNa1-deficient NMDG1 medium named Na1-free (in mM): 5.5 KCl,147 NMDG, 1.2 MgCl2, 1.5 CaCl2, 10 glucose, and 10 Hepes-Trizma(pH 7.4), in the presence of thapsigargin, the irreversible and selectiveinhibitor of the sarco(endo)plasmic reticulum Ca21-ATPase (SERCA)(Secondo et al., 2007).

Neurounina-1 was incubated with cells for 30 minutes before NCXactivity was studied. Each EC50 of neurounina-1was obtained by fittingthe data with the equation a 1 b � exp(2x/t), where a is the maximalresponse, b is the basal response, x is the drug concentration, and t isthe EC50. The activity of the NMDA receptor was studied as [Ca21]iincrease detected in cortical neurons (10 DIV) when rapidly exposed toNMDA (100 mM) and glycine (10 mM) in a Mg21-free solution.

Electrophysiology

NCX Currents. NCX currents (INCX) were recorded, by patch-clamp technique in whole-cell configuration (Secondo et al., 2007;Molinaro et al., 2008, 2011), in the following groups: 1) wild-typeBHK, 2) BHK stably transfected with NCX1, 3) NCX2, and 4) NCX3.Each experimental group was exposed to neurounina-1 or its vehicle.Currents were filtered at 5 kHz and digitized using a Digidata 1322Ainterface (Molecular Devices, Sunnyvale, CA). Data were acquiredand analyzed using the pClamp software (version 9.0; MolecularDevices). In brief, INCX was recorded starting from a holding potentialof 270 mV up to a short-step depolarization at 160 mV (60 ms)(Secondo et al., 2009). Then, a descending voltage ramp from160 mVto 2120 mV was applied. The current recorded in the descendingportion of the ramp (from 160 mV to 2120 mV) was used to plot thecurrent-voltage (I-V) relation curve. The magnitudes of INCX weremeasured at the end of160mV (reversemode) and at the end of2120mV(forward mode), respectively. To isolate INCX, the same cells of allexperimental groups were recorded first for total currents and thenfor currents in the presence of Ni21 (5 mM), a selective blocker of INCX.To obtain the isolated INCX, the Ni21-insensitive unspecific currentswere subtracted from the total currents (INCX 5 IT 2 INiResistant)

144 Molinaro et al.

at ASPE

T Journals on M

ay 13, 2018m

olpharm.aspetjournals.org

Dow

nloaded from

(Molinaro et al., 2008, 2011; Secondo et al., 2011). In all experimentalgroups, both neurounina-1-treated and nontreated cells were recordedbefore and after NiCl2 exposure. Neurounina-1-induced INCX increasewas calculated as follows: (INCXneurounina-1/INCXcontrol) � 100.

External Ringer solution contained the following (in mM unlessotherwise specified): 126 NaCl, 1.2 NaHPO4, 2.4 KCl, 2.4 CaCl2, 1.2MgCl2, 10 glucose, and 18 NaHCO3, 20 TEA, 10 nM TTX, and 10 mMnimodipine (pH 7.4). The dialyzing pipette solution contained thefollowing (in mM): 100 potassium gluconate, 10 TEA, 20 NaCl, 1 Mg-ATP, 0.1 CaCl2, 2 MgCl2, 0.75 EGTA, and 10 Hepes, adjusted to pH 7.2with CsOH. TEA (20 mM) and Cs were included to block delayedoutward rectifier K1 components; nimodipine (10mM) and TTX (50 nM)were added to external solution to blockL-type Ca21 channels andTTX-sensitive Na1 channels, respectively. The quantifications of INCX werenormalized for membrane capacitance and expressed as a percentage ofcontrols as previously reported (Molinaro et al., 2008, 2011), whereasthe representative traces of INCX are expressed as picoamps permillivolt (pA/mV). The reversibility of neurounina-1’s effect on INCX wasmeasured in the same BHK-NCX1 or BHK-NCX2 cells, exposing themto 10 nM neurounina-1 and after 5 minutes of drug washout.

To obtain the EC50 of neurounina-1 on each aforementionedcurrent, all data were fitted to the following binding isotherm: y 5max/[1 1 (X/EC50)]

n, where X is the drug concentration and n is theHill coefficient.

GABAA Currents. In these experiments, membrane currentswere measured with the standard whole-cell patch-clamp techniqueusing an Axopatch 200B amplifier (Axon Instruments, Burlingame,CA) on rat cerebellar granule cells. Patch pipettes were preparedfrom borosilicate glass capillaries (Type 1304336 TW 150-3; WorldPrecision Instruments, Sarasota, FL) with the model P-30 puller(Sutter Instruments). In all the experiments reported, the holdingpotential was set at 280 mV, since it resulted in the most suitablecondition for recording the total chloride current elicited by GABA.Ionic currents were recorded with a Labmaster D/A, A/D converterdriven by pClamp software (Axon Instruments). Analysis wasperformed with pClamp 7 and SigmaPlot software 9.0 (JandelScientific, Ekrath, Germany). The standard external solutioncontained the following (in mM): 135 NaCl, 5.4 KCl, 1.8 CaCl2, 1MgCl2, 5 Hepes, and 10 glucose; pH 7.4 was adjusted with NaOH.The internal solution contained the following (in mM): 142 KCl, 10Hepes, 2 EGTA, 1 MgCl2, and 3 ATP; pH was adjusted to 7.3 withTris base. GABA and neurounina-1 were diluted with the externalsolution at the desired final concentration just before experimentswere performed. The external solution was applied to the cell bath bysteady perfusion (around 3 ml/min gravity flow). In all the experi-ments, the peak chloride current evoked by GABA in the presence ofneurounina-1 was referred to those activated by plain GABA in thesame cell. In particular, neurounina-1-induced current increase wasexpressed as E% 5 100 � (IN1GABA 2 IGABA)/IGABA, where IN1GABA

represents the current in the presence of neurounina-1 plus GABA,whereas IGABA represents the current with GABA alone. Theconcentration-response curves for the compound can be fitted by Hillequation E% 5 E%max Cn/Cn 1 Kn, where E%max is the maximalpercentage of enhancement, C is the drug concentration, n is the Hillnumber, and K is the concentration where E% 5 E%max/2.

In Vitro Toxicity

BHK-NCX1, BHK-NCX2, and BHK-NCX3 cell lines were exposedto 10 mM neurounina-1 at 37°C in a humidified 5% CO2 atmosphere.After 24 hours of incubation, cell injury was assessed using a mixtureof the fluorescent dyes propidium iodide (PI) and fluoresceindiacetate (FDA) (Life Technologies) at the final concentrations of7 and 36 mM, respectively. PI- and FDA-positive cells were countedin three representative fields of independent cultures obtained witha 40� objective, and cell death was determined by the ratio of thenumber of PI-positive cells to PI- and FDA-positive cells (Secondoet al., 2007).

Determination of Mitochondrial Activity

Mitochondrial dysfunction was evaluated with the MTT method(Secondo et al., 2007). In brief, after the experimental procedures,BHK-NCX1, BHK-NCX2, and BHK-NCX3 cells were washed withnormal Krebs solution and incubated with 1 ml of MTT solution (0.5mg/ml in phosphate-buffered saline). This yellow water-solubletetrazolium salt is converted into a water-insoluble purple formazanby the succinate dehydrogenase system of the active mitochondria.Therefore, the amount of formazan produced is proportional to thenumber of cells with mitochondria that are still alive. After 1 hour ofincubation at 37°C, cells were dissolved in 1 ml of DMSO, in which therate of MTT reduction was measured by a spectrophotometer ata wavelength of 540 nm. Data are expressed as percentage ofmitochondrial dysfunction versus sham-treated cultures.

In Vitro Model of Ischemia

Hypoxic conditions were induced by exposing cortical neurons tooxygen- and glucose-free medium in a humidified atmospherecontaining 95% nitrogen and 5% CO2 (Scorziello et al., 2007). After3 hours of oxygen and glucose deprivation (OGD), reoxygenation wasinduced by exposing cortical neurons to the previous Neurobasalmedium containing glucose and oxygen. After 21 hours of reoxygena-tion, cell injury was assessed using the fluorescent dyes PI (7 mM) andFDA (36 mM) (Life Technologies). PI- and FDA-positive cells werecounted in three representative fields of independent cultures, andcell death was determined by the ratio of the number of PI-positivecells to PI- and FDA-positive cells.

Experimental Animal Groups

Wild-type male C57/BL6 mice (Charles River Laboratories, Calco,Lecco, Italy) aged 6–8 weeks andweighing 27–30 g were housed underdiurnal lighting conditions. Each experimental group included atleast five animals. Experiments were performed according to theinternational guidelines for animal research and approved by theAnimal Care Committee of “Federico II”, University of Naples, Italy.

Transient Middle Cerebral Artery Occlusion Model

Mice were anesthetized using a mixture of 5% sevoflurane in 70%nitrous oxide/30% oxygen, and anesthesia was maintained during thesurgical procedure with 2% sevoflurane. Transient middle cerebralartery (MCA) occlusion (tMCAO) was induced as previously described(Cuomo et al., 2008; Molinaro et al., 2008). In brief, a 5-0 nylonfilament was inserted through the external carotid artery stump andadvanced into the left internal carotid artery until it blocked theorigin of theMCA. After a 60-minuteMCA occlusion, the filament waswithdrawn to restore blood flow. Body temperature was monitoredthroughout the entire duration of the surgical procedure andmaintained at 37.5°C with a thermostatic blanket.

Monitoring of Blood Gas Concentration and Cerebral BloodFlow with Laser-Doppler Flowmetry

A catheter was inserted into the femoral artery to measure arterialblood gases before and after ischemia (Rapidlab 860; Chiron Diagnostic,Burladingen, Germany). Cerebral blood flow (CBF)wasmonitored in thecerebral cortex ipsilateral to the occluded MCA with a laser-Dopplerflowmeter (PeriFlux System 5000; Perimed, Järfälla, Stockholm,Sweden) (Cuomo et al., 2007). Once a stable CBF signal was obtained,theMCAwas occluded. CBFmonitoring was continued up to 30minutesafter the end of the surgical procedure once the reperfusion was verified.Only the animals showing a reduction in CBF of at least 70% wereincluded in the experimental groups, and no difference in reduction inCBF was present among the experimental groups. The percentage ofCBF reduction in the different experimental groups was: 77.1 6 1.07%in the vehicle-treated group; 77.5 6 0.5% with 3 ng/kg neurounina-1 3hours after ischemia; 77.86 0.93% with 10 ng/kg neurounina-1 3 hours

Neurounina-1 Increases NCX Activity and Protects against Stroke 145

at ASPE

T Journals on M

ay 13, 2018m

olpharm.aspetjournals.org

Dow

nloaded from

after ischemia; 76.3 6 0.9% with 30 ng/kg neurounina-1 3 hours afterischemia; 77.36 0.6%with 3mg/kg neurounina-1 3 hours after ischemia;77.56 0.5% with 30 mg/kg neurounina-1 3 hours after ischemia; 76.860.6% with 30 ng/kg neurounina-1 5 hours after ischemia; 76.7 6 1.45%with 3 mg/kg neurounina-1 5 hours after ischemia; 77.3 6 0.6% with30 mg/kg neurounina-1 5 hours after ischemia; 76.8 6 0.6% with10 mg/kg flumazenil; and 77.8 6 0.7% with 30 ng/kg neurounina-1 110 mg/kg flumazenil.

Evaluation of Ischemic Volume

The ischemic area was evaluated by 2,3,5-triphenyltetrazoliumchloride staining (Pignataro et al., 2004a). In brief, mice weredecapitated 24 hours after tMCAO. The brains were quickly removedand placed in ice-cold saline for 5 minutes and then cut into 500 mMcoronal slices with a vibratome (752M Vibroslice Tissue Cutter;Campden Instruments, Loughborough, UK). Sections were incubatedin 2% 2,3,5-triphenyltetrazolium chloride containing saline solution for20 minutes and in 10% formalin overnight. The infarcted area wascalculated by image analysis software (Image-Pro Plus, MediaCybernetics inc, Rockville, MD) (Pignataro et al., 2004a). Total infarctvolume was expressed as percentage of the volume of the hemisphereipsilateral to the lesion to eliminate the effect of brain swelling on thetotal calculated infarct volume (Pignataro et al., 2004a).

Experimental Protocol for Drug Administration

Neurounina-1 saline solution was administered at doses of 0.003,0.01, 0.03, 3, and 30 mg/kg i.p. 3 hours after tMCAO or at doses of 0.03,3, and 30 mg/kg i.p. 5 hours after ischemia induction. Flumazenil wasadministered at a dose of 10 mg/kg i.p. three times, 3, 4, and 5 hoursafter tMCAO. All experiments were carried out in blind manner, thepeople who performed the experiments and analyzed the data werenot aware of the pharmacological treatment. The body temperature ofmice was monitored every hour starting 1 hour after neurounina-1administration until 6 hours. The vehicle-treated group displayeda temperature of 36.93 6 0.03°C, whereas the temperature for the0.03 mg/kg neurounina-1-treated group was 36.8760.04°C.

Synaptosome Purification

Rats were sacrificed, the brains were rapidly removed, and thecerebral cortices were dissected at 4°C. Synaptosomes were preparedessentially as previously described (Stigliani et al., 2006). The tissuewas homogenized in 10 volumes (1 g in 10 ml) of 0.32M sucrose andbuffered at pH 7.4 with Tris-HCl using a glass-Teflon tissue grinder(clearance, 0.25 mm). The homogenate was centrifuged (5 minutes,1000� g) to remove nuclei and debris, and the supernatant was gentlystratified on a discontinuous Percoll gradient (2, 6, 10, and 20% [v/v] inTris-buffered sucrose). After centrifugation at 33,500� g for 5 minutes,the layer between 10 and 20% Percoll (synaptosomal fraction) wascollected, washed by centrifugation at 20,000g for 15minutes, and thenresuspended in physiologic medium having the following composition(in mM): 140 NaCl, 3 KCl, 1.2 MgSO4, 1.2 NaH2PO4, 5 NaHCO3, 1.2CaCl2; 10 Hepes, and 10 glucose (pH 7.4). Proteins were measuredaccording to the Bradford method using bovine serum albumin asa standard. All the above procedures were performed at 4°C.

Release Experiments

Synaptosomes were incubated at 37°C for 15 minutes; aliquots ofsynaptosomal suspension (~300-mg protein) were equally layered onmicroporous filters placed at the bottom of a set of parallel superfusionchambers maintained at 37°C (Superfusion System; Ugo Basile,Comerio, Italy) (Tardito et al., 2010). Superfusion was then startedwith standard medium at a rate of 0.5 ml/min and continued for 48minutes. After 36 minutes of superfusion to equilibrate the system,samples were collected according to the following scheme: two 3-minutesamples (t 5 36–39 minutes and t 5 45–48 minutes; basal outflow)

before and after one 6-minute sample (t 5 39–45 minutes; stimulus-evoked release). A 90-second period of stimulation was applied at t5 39minutes, after the collection of the first sample.

Stimulation of synaptosomes was performed with 15 mM KClsubstituting for equimolar concentration of NaCl. In the firstexperimental group, in which the calcium dependence was assessed,a Ca21-free medium was added at 20 minutes and maintained duringthe KCl pulse period; in the second and third groups, in which thecarrier-mediated component of glutamate or GABA release wasevaluated, the broad-spectrum glutamate uptake inhibitor DL-threo-b-benzyloxyaspartic acid (DL-TBOA) (10 mM) or the selective GAT1transporter inhibitor SKF89976A (10 mM), respectively, was added at30 minutes and maintained during the KCl pulse period; in the fourthgroup, in which the effect of neurounina-1 (10 nM, 30 nM, 100 nM) wasmeasured on neurotransmitter release, the drug was added at 30minutes and maintained during the 15 mM KCl stimulation pulseperiod.

Collected samples were analyzed for endogenous glutamate andGABA content. Amino acid release was expressed in picomoles permilligram of protein. The stimulus-evoked overflow was estimated bysubtracting the transmitter content of the two 3-minute samples (basaloutflow) from the release evoked in the 6-minute sample collectedduring and after the depolarization pulse (stimulus-evoked release).

Neurotransmitter Release Determination

Endogenous glutamate and GABA content was measured by HPLCanalysis following precolumn derivatization with o-phthalaldehydeand gradient separation on a C18 reverse-phase chromatographiccolumn (10 � 4.6 mm, 3 mm; at 30°C; Chrompack, Middleburg, TheNetherlands) coupled with fluorometric detection (excitation wave-length, 350 nm; emission wavelength, 450 nm). Homoserine wasused as an internal standard. For more details, see Tardito et al.(2010).

Statistical Analysis

Data are expressed as mean (6S.E.M.) of six independentdeterminations. Log P value of neurounina-1 was calculated on thebasis of ChemDraw 11.0 software (CambridgeSoft, Cambridge, MA).Statistical comparisons between controls and treated experimentalgroups were performed using the one-way analysis of variance,followed by Newman Keuls test. P, 0.05 was considered statisticallysignificant.

ResultsNeurounina-1 Increases NCX1 and NCX2 Activity in StablyTransfected BHK Cells, as Revealed by Ca21 RadiotracerInflux/Efflux and Fura-2 Microfluorimetry Techniques

The newly synthesized compound neurounina-1, whosestructure is illustrated in the inset of Fig. 1A, increased NCX1and NCX2 activity in a concentration-dependent manner. It isremarkable that it displayed a high potency, with anEC50 in thelow nanomolar range (1.1–2.7 nM), as detected by Ca21radiotracer method, and in the picomolar range (34–87 pM),as detected by Fura-2-monitored Ca21 influx technique.Furthermore, these changes occurred in both forward andreverse modes of operation, as revealed by Na1-dependent45Ca21 influx/efflux methods (Fig. 1, A and B) and by Na1-free-induced [Ca21]i increases in single cells (Fig. 1C). By contrast,neurounina-1 was substantially ineffective on NCX3-mediatedinflux and efflux even at concentrations up to 10 mM (Fig. 1, Aand B).

146 Molinaro et al.

at ASPE

T Journals on M

ay 13, 2018m

olpharm.aspetjournals.org

Dow

nloaded from

Neurounina-1 Increases NCX1 and NCX2 Activity in StablyTransfected BHK Cells, as Revealed by Patch-ClampTechnique

In stably transfected BHK cells, 10 nM neurounina-1significantly increased NCX1 and NCX2 currents, recorded bypatch-clamp technique, in both forward and reverse modes ofoperation (Fig. 2, A and B). In particular, NCX1 activity wasincreased by ∼60% and that of NCX2 by ∼40%. By contrast,neurounina-1 did not affect INCX in BHK-NCX3 nor unspecificcurrents in BHK wild-type cell lines (Fig. 2, C and D). Fiveminutes of washout completely reverted neurounina-1-inducedincrease in both NCX1 and NCX2 activities (Fig. 2, A and B).

a1- and a2-Repeats Are the Molecular Determinants of theNeurounina-1 Stimulatory Effect on NCX1

To investigate the molecular determinants of neurounina-1responsible for NCX1-stimulatory activity, the intracellularf-loop, a region mainly involved in the regulation of NCXfunction, was deleted. In particular, despite the f-loopdeletion, 10 nM neurounina-1 was still able to increase theactivity of the mutant NCX1D241–680 (NCX1Df) in bothforward and reverse modes of operation (Fig. 3, A and B),suggesting that the f-loop is not a critical region forneurounina-1 binding and activity. Taking advantage of the

fact that NCX1 was sensitive to neurounina-1, whereas NCX3was insensitive despite the high sequence homology betweenthe two isoforms, we carried out experiments with chimerasobtained between these two NCX isoforms to identify theregion(s) responsible for the stimulatory effect of neurounina-1 in the NCX1 molecule. In particular, we used a series ofchimeras in which a segment from NCX3.3 was transferredinto NCX1.1 in exchange for the homologous segment(s)(Iwamoto et al., 2004b) and vice versa (Fig. 3). All chimerasused exhibited basal exchange activities similar to that of thewild-type NCX1 (Iwamoto et al., 2004b). Figure 3 shows theeffects of 10 nM neurounina-1 on the rates of Na1-dependent45Ca21 efflux/influx into BHK cells expressing wild-type orchimeric exchangers. Neurounina-1 (10 nM) increased the45Ca21 efflux/influx rates of the wild-type NCX1 to approxi-mately 70 and 35% of the respective controls. NCX1193/445chimera, containing the fifth transmembrane domain and theN-terminal portion of the cytosolic f loop of NCX3, showeda significant increased activity in the presence of 10 nMneurounina-1, in both forward and reverse modes of opera-tion. A similar increase was observed in NCX1718/787 chimera,containing the C-terminal portion of the cytosolic f loop andthe sixth transmembrane domain of NCX3 (Fig. 3, A and B).By contrast, NCX1109/133 chimera, which contained the

a1-repeat of NCX3, and NCX1788/829 chimera, which contained

Fig. 1. Effect of neurounina-1 on NCX1, NCX2, and NCX3 activity measured by Na+-dependent 45Ca2+ influx/efflux and Fura-2-monitored Ca2+

techniques. Inset shows the chemical structure of neurounina-1. (A and B) concentration-response curves of the effects of neurounina-1 on Nao-dependent 45Ca2+ efflux and Nao-dependent

45Ca2+ uptake, respectively, in BHK cells expressing NCX1, NCX2, or NCX3. (C) effect of neurounina-1(0.0001–0.01 mM) on NCX1, NCX2, and NCX3 activity measured by single-cell Fura-2 AM microfluorimetry (n = 60 cells for each group). Data arecalculated as D% of plateau/basal [Ca2+]i values after Na+-free addition. Under control conditions, basal values of [Ca2+]i were (in nM) 726 4.8 in BHK-NCX1 cells, 79 6 3.5 in BHK-NCX2 cells, and 73 6 4.2 in BHK-NCX3 cells. Data are means 6 S.E.M. of three independent experiments. *P , 0.05 vs.NCX3 and control groups. **P , 0.05 vs. NCX2 and NCX3 groups.

Neurounina-1 Increases NCX Activity and Protects against Stroke 147

at ASPE

T Journals on M

ay 13, 2018m

olpharm.aspetjournals.org

Dow

nloaded from

the a2-repeat of NCX3, displayed no significant increasedactivity in the presence of 10 nM neurounina-1 in both modesof operation (Fig. 3, A and B), thereby indicating that thesetwo segments are necessary for drug activity.Interestingly, a chimera of the neurounina-1-insensitive

NCX3 isoform, in which the a1- and a2-regions were substi-tuted with the corresponding sequences of the neurounina-1-sensitive NCX1 isoform, became responsive to the drug (Fig.3, A and B). This finding further indicates that the a1- and a2-segments are exclusively responsible for the different drugresponses between NCX1 and NCX3.

Val118, Asn125, and Leu808 Are Responsible for NCX1Sensitivity to Neurounina-1

To establish the sites responsible for the stimulatory actionof neurounina-1 on NCX1 activity, we substituted individual

critical residues of the a1- and a2-regions of NCX1.1 with thecorresponding amino acid present on the neurounina-1-insensitive NCX3 isoform (Fig. 4A). Each NCX1 mutant wassingly and stably transfected in wild-type BHK cells. InNCX1V118L, NCX1N125G (a1-repeat), and NCX1L808F (a2-repeat) mutants, neurounina-1 was unable to elicit anincrease in the antiporter activity, as revealed by the Ca21

radiotracer (Fig. 4B) and single-cell Fura-2-monitored Ca21

microfluorimetry techniques (Fig. 4C). These data suggestthat these three residues are critical for NCX1 sensitivity tothe drug. In addition, in two of the three mutants,i.e., NCXV118L and NCX1N125G, the stimulatory effect ofneurounina-1 was converted into an inhibitory effect. Since theinhibition of NCX1 by several drugs, such as KB-R7943,SEA0400, SN-6, and YM-244769, depends on Gly833 (Iwamotoet al., 2001, 2004b,c), we tested whether neurounina-1 was stillactive on the mutant NCX1G833C lacking this apolar amino

Fig. 2. Effect of neurounina-1 on NCX1, NCX2, and NCX3 currents measured by patch-clamp technique in whole-cell configuration. (A–C) the leftpanels represent superimposed traces of total currents recorded from BHK cells stably transfected with NCX1, NCX2, or NCX3, respectively, without(gray trace) or with (black trace) Ni2+; the middle panels show representative traces of NCX1, NCX2, and NCX3 isolated Ni2+-subtracted currents undercontrol conditions, in the presence of neurounina-1 or after wash; the right panels show concentration-response curves of the neurounina-1 effect onNCX1 (top) and NCX2 (bottom) currents. (D) representative superimposed traces of the currents recorded from BHK wild type under control conditions(gray traces) and after 10 nM neurounina-1 exposure (black traces). (E) quantification of neurounina-1’s effect (10 nM) on NCX1, NCX2, and NCX3currents in the reverse and forward modes of operation. The values are expressed as percentage mean 6 S.E.M. of three independent experimentalsessions (n = 20 for each group). *P , 0.05 vs. the respective control groups.

148 Molinaro et al.

at ASPE

T Journals on M

ay 13, 2018m

olpharm.aspetjournals.org

Dow

nloaded from

acid. It is noteworthy that the substitution of the criticalresidue Gly833 with cysteine on NCX1 did not prevent thestimulatory effect of neurounina-1 (Fig. 4, B and C).

Neurounina-1 Does Not Induce Cell Death andMitochondrial Dysfunction in Stably TransfectedBHK-NCX1, BHK-NCX2, and BHK-NCX3 Cells

The toxicity of neurounina-1 was evaluated in vitro byexposing BHK-NCX1, BHK-NCX2, and BHK-NCX3 cells to 10mM neurounina-1, a concentration 104 times higher than thecalculated EC50 values. Neurounina-1 incubation for 24 hoursdid not cause any cell death measured by the PI/FDA1PImethod (Fig. 5A) or any significant decrease in mitochondrialactivity as revealed by MTT method (Fig. 5B).

Neurounina-1 Reduces Cell Death in Primary CorticalNeurons Exposed to OGD Plus Reoxygenation

Primary cortical neurons exposed to 3 hours of OGD followedby 21 hours of reoxygenation were used to investigate the

neuroprotective effects of neurounina-1 in an in vitro model ofanoxia. Neurounina-1 (10 nM) exerted a remarkable neuro-protective effect by reducing cell death induced by OGD andOGD plus reoxygenation (Fig. 5C).

Neurounina-1 Reduces Ischemic Brain Damage in MiceSubjected to tMCAO

Remarkably, prediction analysis revealed that neurounina-1possesses a high lipophilicity and, thus, a high capability ofcrossing the blood-brain barrier, with an estimated log P of 0.87(n-octanol/water).Consistent with in vitro results, the intraperitoneal in-

jection of neurounina-1 in a single dose ranging from 0 to 30mg/kg in C57B male mice, previously subjected to 60 minutesof MCA occlusion followed by 23 hours of reperfusion, causeda dose-dependent reduction in the ischemic volume comparedwith vehicle-treated mice (Fig. 6, A and B), even whenadministered in the low nanograms per kilogram range. Inparticular, 3 ng/kg neurounina-1 administered 3 hours after

Fig. 3. Effect of neurounina-1 on NCX1/NCX3 chimeras measured by Na+o-dependent45Ca2+ uptake and 45Ca2+ efflux techniques. At left, all chimeras

have been reported as cartoons. (A and B) effect of 10 nM neurounina-1 on NCX1/NCX3 chimeras measured by Na+o-dependent45Ca2+ efflux (A) and

uptake (B). Values are expressed as percentage versus the respective control. Data aremeans6 S.E.M. of six independent experiments. *P, 0.05 versusthe respective control group. **P , 0.05 vs. the respective control and NCX1WT groups.

Neurounina-1 Increases NCX Activity and Protects against Stroke 149

at ASPE

T Journals on M

ay 13, 2018m

olpharm.aspetjournals.org

Dow

nloaded from

stroke onset produced a mild but nonetheless significantdecrease (∼20%) of the ischemic damage. Higher doses ofneurounina-1, from 30 ng/kg up to 30 mg/kg, produced a moreremarkable neuroprotective effect, causing a 60% reduction ofthe ischemic brain damage. More important, at the dose of30 mg/kg, neurounina-1 displayed a wide therapeutic window,as it was still effective even 5 hours after stroke induction(Fig. 6B). Themortality rate in all neurounina-1-treated groupswas around 5%, as observed for the control groups.

Neurounina-1 Increases the GABAA-Mediated Currents ofChloride Ions

Since neurounina-1 contains a benzodiazepine-like struc-ture, we tested whether it could increase the GABAA-mediated currents. Neurounina-1 at 10 nM did not affectthe chloride currents per se (data not shown); however, in thepresence of 10 mMGABA, neurounina-1 increased the peak ofchloride current in a concentration-dependent manner withan EC50 of 3.6 nM. The neurounina-1 effect on GABAA

currents was reverted by 10 mM flumazenil (Fig. 7, A and B).The maximum increase of the GABAA currents was observedat neurounina-1 concentrations higher than 10 nM (Fig. 7C).

The Benzodiazepine Antagonist Flumazenil Does NotRevert the Neuroprotective Effect of Neurounina-1 in MiceSubjected to tMCAO

To verify whether the effect of neurounina-1 during cerebralischemia was mediated by the activation of GABAA receptors,we evaluated the effect of this compound on the ischemic braindamage in the presence of the benzodiazepine-antagonistflumazenil. Mice were treated with three injections of 10mg/kgflumazenil (Brogden and Goa, 1988) at 3, 4, and 5 hours afterischemia induction and with neurounina-1, at the dose of0.03 mg/kg, 3 hours after ischemia induction. Under theseconditions, flumazenil did not revert the neuroprotectioninduced in vivo by neurounina-1 after tMCAO and flumazenilalone did not influence the extent of the infarct volume ascompared with vehicle-treated ischemic mice (Fig. 7D).

Neurounina-1 Inhibits 15 mM KCl-Evoked EndogenousGlutamate and GABA Releases in Synaptosomes andNMDA Receptor Activation

Since neurounina-1 has a neuroprotective effect on an in vitromodel of cerebral ischemia, we tested whether it could interfere

Fig. 4. Effect of neurounina-1 on NCX1 mutants measured by Na+o-dependent45Ca2+ uptake and Fura-2 single-cell microfluorimetry. (A) amino acid

alignment of a1- and a2-regions of NCX1,NCX2, and NCX3. Conserved amino acids among exchangers are indicatedwith dots. Asterisks indicate the positionof mutations that altered neurounina-1 sensitivity. (B) effect of neurounina-1 on NCX1 mutants, measured by Na+o-dependent

45Ca2+ uptake, in which thesingle residues in the regions indicated in (A) were exchanged with the corresponding residue of NCX3. Values are represented as percentage ratios of thevalues obtained in the presence of 10 nMneurounina-1 versus nontreated control cells. Data are presented asmeans6S.E.M. of six independent experiments.*P, 0.05 vs. the respective control group. **P, 0.05 vs. NCX1WT and NCX3WT groups. (C) Na+-free-induced [Ca2+]i increase in cells expressing the above-described NCX1 mutants and treated with 10 nM neurounina-1. Each value is reported as percentage of its nontreated controls. Data are presented asmeans6 S.E.M. of three independent experiments. *P, 0.05 vs. the respective control group. **P, 0.05 vs. NCX1WT and NCX3WT groups. WT, wild type.

150 Molinaro et al.

at ASPE

T Journals on M

ay 13, 2018m

olpharm.aspetjournals.org

Dow

nloaded from

with glutamate/GABA release and/or NMDA receptor activa-tion. DL-TBOA (10 mM), a broad-spectrum glutamate uptakeinhibitor, and 1-(4,4-Diphenyl-3-butenyl)-3-piperidinecarboxylicacid (SKF89976A) (10 mM), a selective GAT1 transporter

inhibitor, did not interfere with KCl-evoked glutamate andGABA release, respectively.Neurounina-1, at concentrations ranging from 10 to 100 nM,

did not affect basal release of endogenous glutamate and

Fig. 5. Effect of neurounina-1 on cell survival. (A) cell death percentage of BHK cells stably transfected with NCX1, NCX2, or NCX3 exposed to 10 mMneurounina-1 for 24 hours. (B) mitochondrial activity of BHK cells exposed to 10 mMneurounina-1 for 24 hours. (C) representative images of the effect ofneurounina-1 on cell survival in primary cortical neurons subjected to OGD plus reoxygenation (left). Quantification of cell death in normoxia, after 3hours of OGD and after 3 hours of OGD followed by 21 hours of reoxygenation, in the presence or in the absence of 10 nM neurounina-1 (right). Data arepresented as means 6 S.E.M. *P , 0.05 vs. the control group. **P , 0.05 vs. the respective vehicle-treated group.

Fig. 6. Dose-dependent effect of neurounina-1 on ische-mic volume in mice subjected to tMCAO. Neurounina-1administered i.p. 3 hours (A) or 5 hours (B) after strokeonset. Each column represents the mean 6 S.E.M. of thepercentage of the infarct volume compared with theipsilateral hemisphere. Each dot represents the singlevalue measured. *P , 0.05 vs. control group. **P , 0.05vs. 0.01 mg/kg treated group.

Neurounina-1 Increases NCX Activity and Protects against Stroke 151

at ASPE

T Journals on M

ay 13, 2018m

olpharm.aspetjournals.org

Dow

nloaded from

GABA neurotransmitters (Table 1), but decreased only 15 mMKCl–evoked endogenous glutamate andGABA release from ratbrain cortical synaptosomes (Fig. 8). On the other hand, 10 nMneurounina-1 inhibited NMDA-induced [Ca21]i increase by∼50% (Fig. 8C), whereas the NMDA receptor blocker (5S,10R)-(1)-5-Methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-iminemaleate (MK-801) totally prevented Ca21 influx inducedby NMDA perfusion.

DISCUSSIONThe present study demonstrated that neurounina-1, a com-

pound synthesized in our laboratories and obtained bymodifying the structure of the powerful NCX1 inhibitor SM-15811 (Hasegawa et al., 2003), is the first molecule thatenhancesNCX1andNCX2activity and also exerts a remarkableneuroprotective effect in stroke.It is noteworthy that neurounina-1 displayed a potent and

reversible stimulatory effect on NCX1 and NCX2 in bothforward and reverse modes of operation, with an estimatedEC50 in the low nanomolar range. By contrast, this compounddid not affect the activity of the NCX3 isoform in the range of

0.001 to 10 mM. These pharmacodynamic properties wereconfirmed by using three different methods to evaluate NCXfunction: Ca21 radiotracer influx/efflux, Fura-2 microfluorim-etry, and patch-clamp electrophysiology in whole-cell config-uration. To characterize themolecular determinants responsiblefor neurounina-1 activity on NCX1, we used chimera strategy,deletion, and site-directed mutagenesis. The removal of theNCX1 f-loop (amino acids 241–680), a region involved in NCXactivity regulation, did not prevent the stimulatory effect ofneurounina-1, suggesting that this region is not involved inthe binding and stimulatory activity of the drug on NCX1. Bycontrast, the use of NCX1/NCX3 chimeras allowed us toidentify the two amino acid sequences required for thebiochemical interaction between NCX1 and neurounina-1,i.e., 109–133 and 788–829, the former corresponding to thea1-repeat and the latter to the a2-repeat. Site-directedmutagenesis, instead, allowed us to identify the moleculardeterminants for neurounina-1 pharmacological action onNCX1. In particular, three amino acids, Val118 and Asn125,present in the a1-repeat, and Leu808, present in the a2-repeat, were essential to elicit the effect of the drug on theexchanger. It is noteworthy that these three amino acids,

Fig. 7. Effect of neurounina-1 on GABAA-mediated chloride currents in rat cerebellar granule cells and effect of flumazenil on neurounina-1-inducedneuroprotection in ischemic mice. (A and B) representative traces and quantification of chloride currents elicited by 10 mM GABA under controlconditions, in the presence of neurounina-1 and in the presence of both neurounina-1 and flumazenil. *P , 0.05 vs. GABA and GABA + neurounina-1 +flumazenil groups. (C) concentration-dependent effect of neurounina-1 on chloride currents evoked by GABAA receptor exposed to 10 mM GABA. (D)evaluation of the ischemic volume in mice 24 hours after tMCAO and treated with three injections of 10 mg/kg flumazenil 3, 4, and 5 hours after tMCAOand one injection of 30 ng/kg neurounina-1 3 hours after tMCAO. Each dot represents the single value measured. *P, 0.05 vs. vehicle- and flumazenil-treated groups.

152 Molinaro et al.

at ASPE

T Journals on M

ay 13, 2018m

olpharm.aspetjournals.org

Dow

nloaded from

which specifically determined NCX1.1 sensitivity to this newcompound, were also present in all NCX1 splice variants,including the neuronal forms NCX1.4, NCX1.5, and NCX1.12,and in the corresponding sites of the only known splicevariant of NCX2, NCX2.1. The relevance of a1- and a2-regionsin determining NCX1 and NCX2 sensitivity to neurounina-1is further supported by evidence that the neurounina-1-insensitive NCX3 isoform became sensitive when its a1- anda2-regions were substituted with the corresponding regionsof NCX1. Noticeably, the substitution of Gly833, a commonmolecular determinant for severalNCX1 inhibitors, with cysteinedid not prevent the stimulatory effect of neurounina-1, thus

suggesting that the sites of action of some NCX inhibitors maydiffer from those of neurounina-1. Another peculiar aspect thatneeds to be underlined is the high potency of neurounina-1 onNCX1 and NCX2 in both forward and reverse modes. In fact,most of drugs that inhibit NCX, including KB-R7943, YM-244769, and SN-6, display an activity in the micromolar range,whereas neurounina-1 exerts its effect onNCX1 andNCX2 in thepicomolar and low nanomolar range of concentrations, with anEC50 of 1 to 0.1 nM. Moreover, at variance with most NCXinhibitors (Iwamoto et al., 1996; Annunziato et al., 2004; Iwamotoand Kita, 2006), neurounina-1 did not display a significantdifference in potency between the forward and reverse modes of

TABLE 1Effects of neurounina-1 on endogenous glutamate and GABA basal release. Neurounina-1 atconcentrations ranging from 10 to 100 nM did not affect basal release of the endogenous glutamate andGABA neurotransmitters.

Neurounina-1 Endogenous Glutamate BasalRelease [Mean (6S.E.M.)]

Endogenous GABA BasalRelease [Mean (6S.E.M.)] n

pmol/mg of proteins

0 (nM) 92.2 6 10.78 42.50 6 4.66 1010 (nM) 80.1 6 15.21 45.41 6 6.43 1030 (nM) 83.2 6 8.06 44.50 6 4.90 10

100 (nM) 90.1 6 13.63 45.00 6 4.58 10

Fig. 8. Effect of neurounina-1 on endogenous GABA and glutamate release from cortical synaptosomes and on [Ca2+]i increase elicited by NMDA receptoractivation in primary cortical neurons. (A) endogenous glutamate release from cortical synaptosomes under basal conditions, after 15 mMKCl, in Ca2+-freemedium, and in the presence of the glutamate uptake inhibitor DL-TBOA. (B) effect of different concentrations of neurounina-1 on endogenous glutamaterelease induced by 15mMKCl (data were subtracted for basal release values). (C) [Ca2+]i increase induced by 100mMNMDAunder control conditions, in thepresence of 10 nM neurounina-1 or in the presence of the NMDA receptor blocker MK-801 (100 mM). (D) endogenous GABA release from corticalsynaptosomes under basal conditions, after 15 mM KCl, in Ca2+-free medium or in the presence of the GABA uptake inhibitor SKF89976A. (E) effect ofdifferent concentrations of neurounina-1 on endogenous GABA release induced by 15 mMKCl (data were subtracted for basal release values). Prt was usedas abbreviation for proteins. *P , 0.05 vs. control group. **P , 0.05 vs. 10 nM neurounina-1. #P , 0.05 vs. control and Ca2+-free groups.

Neurounina-1 Increases NCX Activity and Protects against Stroke 153

at ASPE

T Journals on M

ay 13, 2018m

olpharm.aspetjournals.org

Dow

nloaded from

operation. Finally, it should be mentioned that neurounina-1,similarly to NCX inhibitors, displayed a reversibility of thestimulatory effect as demonstrated by patch-clamp evaluation.Because the specific knocking down or knocking out of the

three NCX isoforms worsens ischemic brain damage in in vitroand in vivomodels of cerebral ischemia (Pignataro et al.,2004a;Jeon et al., 2008; Molinaro et al., 2008), it was conceivableto hypothesize that the activation of this exchanger could beneuroprotective in stroke. Indeed, the stimulation of theantiporter, by modifying the dysregulation of intracellularNa1 and Ca21 ion homeostasis, could help the rescue of injuredneurons in the ischemic and peri-ischemic areas of the brain.However, to date, only nonselective NCX activators have beenreported to stimulateNCXactivity, including lithium (Iwamotoet al., 1999), redox agents (Reeves et al., 1986; Secondo et al.,2011), agonists of G-protein-coupled receptors (Stengl et al.,1998; Eriksson et al., 2001;Woo andMorad, 2001; Annunziatoet al., 2004), diethylpyrocarbonate (Ottolia et al., 2002),concanavalin A, nerve growth factor, and insulin (Guptaet al., 1986; Makino et al., 1988; Formisano et al., 2008).However, all of these compounds do not possess the pharma-cological properties necessary to become valid active drugs.Neurounina-1, on the other hand, does display some in-teresting pharmacological properties for a number of reasons.First, its lipophilicity index, determined by log P prediction, isindicative of having a good blood-brain barrier permeability.Second, being water-soluble, it is injectable by parenteral ad-ministration. Finally, as demonstrated in our in vitro experi-ments, it did not cause cell toxicity at up to 10 mM, aconcentration ∼104-fold higher than the calculated EC50.In addition to these pharmacological properties, neurounina-1

exerted a remarkable neuroprotective effect in both in vitro andin vivo experimental models of cerebral ischemia, possibly byenhancing NCX activity. In particular, cortical neurons exposedto neurounina-1 displayed a higher resistance to neuronaldamage induced by 3 hours of OGD or 3 hours of OGD followedby 21 hours of reoxygenation as compared with vehicle-treatedneurons. It is possible to hypothesize that part of this neuro-protective effect of neurounina-1 during OGD followed byreoxygenation is due to the increase in NCX activity that, inturn, counteracts ER-Ca21 depletion occurring in neuronsduring OGD followed by reoxygenation, a mechanism mediatedby the NCX1 isoform (Sirabella et al., 2009).It is noteworthy that the results obtained in vivo showed that

the intraperitoneal administration of neurounina-1 in singledoses ranging from 0.003 to 30 mg/kg significantly reduced theinfarct volume in a mouse model of tMCAO. Remarkably,neurounina-1 was effective even when administered up to 5hours after ischemia induction, a therapeutic window that,along with its viable route of administration, may haveinteresting clinical perspectives.Since neurounina-1 contains a benzodiazepine-like struc-

ture, we tested whether it activates the GABAA-mediatedcurrents, whose increase might interfere with the ischemicprocess. Indeed, neurounina-1 in the presence of GABAincreased chloride currents in a concentration-dependentmanner, with an EC50 of 3.6 nM. In addition, the effect ofneurounina-1 on GABA-induced chloride currents was com-pletely reverted by flumazenil, the competitive antagonist ofthe benzodiazepine recognition site on the GABAA receptor.However, although neurounina-1 reinforced the action ofGABA on the GABAA receptor, at the same time it reduced its

release. In consideration of this dual and opposite action ofneurounina-1 on GABA neurotransmission, we can hypothe-size that the modulation of GABA transmission might playa less relevant role in the neuroprotective effect exerted byneurounina-1 after cerebral ischemia. In accordance with thishypothesis, the administration of flumazenil in ischemic micedid not revert the neuroprotective effect of neurounina-1.Another aspect that deserves consideration is the effect of

neurounina-1 on glutamate release from cortical synapto-somes, which suggested that part of the neuroprotective effectof the newly synthesized compound in stroke could be due toa partial inhibition of both endogenous glutamate release andNMDA receptor activity. Although this hypothesis cannot beruled out by the results of in vitro experiments, it should beconsidered that a large body of evidence produced in the last20 years showed that glutamate receptor antagonists may beeffective as neuroprotectants only if administered before orimmediately after MCA occlusion in rodents. In fact, theefficacy of these compounds is lost when administered 1 hourafter stroke onset (Buchan et al., 1991; Hossmann, 1996;Gladstone et al., 2002; Di Renzo et al., 2009). By contrast, inour study neurounina-1 is also effective when administered 5hours after stroke onset. It should be also considered thatNMDA receptor antagonists exert part of their neuroprotec-tive effect in stroke animal models by their hypothermic effect(Buchan and Pulsinelli, 1990), whereas in the present studyneurounina-1 did not affect body temperature.At variance with the results of the present study, some data

presented in the literature showed that pharmacologicalinhibition of NCX activity protects the brain against stroke(Jeffs et al., 2007). However, it has also been demonstrated thatthese drugs are not selective for NCX (Pintado et al., 2000;Reuter et al., 2002; Annunziato et al., 2004), and some of thempossess a remarkable and long-lasting hypothermic effect(Pignataro et al., 2004b) that is neuroprotective by itself.As concerns the possibility that neurounina-1 administra-

tion might affect cardiac function, it is difficult to predict theeffect of neurounina-1 on the cardiovascular system because ofthe controversial role played by NCX1.1 in the heart underdifferent pathophysiologic conditions. In fact, some reportssuggest that elevated NCX1 activity might be associated withheart failure and arrhythmia, whereas other reports indicatethat the overexpression of NCX1 or an increase in its activityresults in an attenuation of postinfarction myocardial dysfunc-tion and preserved diastolic function (Litwin and Bridge, 1997;Hasenfuss et al., 1999; Min et al., 2002; Sipido et al., 2002).Although more specific studies are necessary to evaluate theeffects of acute, subacute, and chronic regimens of neurounina-1 on all cardiovascular functions, we found that after 1 and 7days of administration of this NCX1 and NCX2 activator, noincrease in mortality rate was observed, either in mice thatunderwent cerebral ischemia or in sham-operated animals.Collectively, our results showed that neurounina-1 is provided

with high potency for NCX1 and NCX2, a high lipophilicityindex, low toxicity, and a remarkable neuroprotective effect inexperimental model of cerebral ischemia, with a wide thera-peutic window and easy route of administration.

Acknowledgments

The authors thank Dr. Paola Merolla for editorial revision, Dr.Takahiro Iwamoto for providing us the cDNA of NCX1/NCX3

154 Molinaro et al.

at ASPE

T Journals on M

ay 13, 2018m

olpharm.aspetjournals.org

Dow

nloaded from

chimeras, and Vincenzo Grillo and Carmine Capitale for technicalassistance.

Authorship Contributions

Participated in research design: Molinaro, Cantile, Cuomo, Secondo,Pannaccione, Pignataro, Fiorino, Severino, Gatta, Milanese, Scorziello,Bonanno, Robello, Santagada, Caliendo, Di Renzo, Annunziato.

Conducted experiments: Molinaro, Cantile, Cuomo, Secondo,Pannaccione, Ambrosino, Fiorino, Severino, Gatta, Sisalli, Milanese,Scorziello.

Performed data analysis: Molinaro, Cantile, Cuomo, Secondo,Pannaccione, Ambrosino, Fiorino, Severino, Gatta, Sisalli, Milanese.

Wrote or contributed to the writing of the manuscript: Molinaro,Cantile, Cuomo, Secondo, Pannaccione, Pignataro, Scorziello, DiRenzo, Annunziato.

REFERENCES

Annunziato L, Cataldi M, Pignataro G, Secondo A, and Molinaro P (2007) Glutamate-independent calcium toxicity: introduction. Stroke 38 (Suppl 2):661–664.

Annunziato L, Molinaro P, Secondo A, Pannaccione A, Scorziello A, Pignataro G,Cuomo O, Sirabella R, Boscia F, Spinali A et al. (2009) The Na/Ca21 Exchanger: ATarget for Therapeutic Intervention in Cerebral Ischemia, in New Strategies in StrokeIntervention, ed. by L. Annunziato (Springer, Totowa, 2009), pp. 65–87.

Annunziato L, Pignataro G, and Di Renzo GF (2004) Pharmacology of brain Na1/Ca21 exchanger: from molecular biology to therapeutic perspectives. PharmacolRev 56:633–654.

Blaustein MP and Lederer WJ (1999) Sodium/calcium exchange: its physiologicalimplications. Physiol Rev 79:763–854.

Brogden RN and Goa KL (1988) Flumazenil. A preliminary review of its benzodiaz-epine antagonist properties, intrinsic activity and therapeutic use. Drugs 35:448–467.

Buchan A, Li H, and Pulsinelli WA (1991) The N-methyl-D-aspartate antagonist,MK-801, fails to protect against neuronal damage caused by transient, severeforebrain ischemia in adult rats. J Neurosci 11:1049–1056.

Buchan A and Pulsinelli WA (1990) Hypothermia but not the N-methyl-D-aspartateantagonist, MK-801, attenuates neuronal damage in gerbils subjected to transientglobal ischemia. J Neurosci 10:311–316.

Craner MJ, Newcombe J, Black JA, Hartle C, Cuzner ML, and Waxman SG (2004)Molecular changes in neurons in multiple sclerosis: altered axonal expression ofNav1.2 and Nav1.6 sodium channels and Na1/Ca21 exchanger. Proc Natl AcadSci USA 101:8168–8173.

Cuomo O, Gala R, Pignataro G, Boscia F, Secondo A, Scorziello A, Pannaccione A,Viggiano D, Adornetto A, Molinaro P et al. (2008) A critical role for the potassium-dependent sodium-calcium exchanger NCKX2 in protection against focal ischemicbrain damage. J Neurosci 28:2053–2063.

Cuomo O, Pignataro G, Gala R, Scorziello A, Gravino E, Piazza O, Tufano R, Di RenzoG, and Annunziato L (2007) Antithrombin reduces ischemic volume, amelioratesneurologic deficits, and prolongs animal survival in both transient and permanentfocal ischemia. Stroke 38:3272–3279.

Di Renzo G, Pignataro G, and Annunziato L (2009) Why Have Ionotropic andMetabotropic Glutamate Antagonists Failed in Stroke Therapy? , in New Strategiesin Stroke Intervention, ed. by L. Annunziato (Springer, Totowa, 2009), pp. 13–25.

Eriksson KS, Stevens DR, and Haas HL (2001) Serotonin excites tuberomammillaryneurons by activation of Na(1)/Ca(21)-exchange. Neuropharmacology 40:345–351.

Fameli N, van Breemen C, and Kuo KH (2007) A quantitative model for linking Na1/Ca21 exchanger to SERCA during refilling of the sarcoplasmic reticulum to sus-tain [Ca21] oscillations in vascular smooth muscle. Cell Calcium 42:565–575.

Formisano L, Saggese M, Secondo A, Sirabella R, Vito P, Valsecchi V, Molinaro P, DiRenzo G, and Annunziato L (2008) The two isoforms of the Na1/Ca21 exchanger,NCX1 and NCX3, constitute novel additional targets for the prosurvival action ofAkt/protein kinase B pathway. Mol Pharmacol 73:727–737.

Gladstone DJ, Black SE, and Hakim AM; Heart and Stroke Foundation of OntarioCentre of Excellence in Stroke Recovery (2002) Toward wisdom from failure: les-sons from neuroprotective stroke trials and new therapeutic directions. Stroke 33:2123–2136.

Grynkiewicz G, Poenie M, and Tsien RY (1985) A new generation of Ca21 indicatorswith greatly improved fluorescence properties. J Biol Chem 260:3440–3450.

Gupta MP, Makino N, Khatter K, and Dhalla NS (1986) Stimulation of Na1-Ca21exchange in heart sarcolemma by insulin. Life Sci 39:1077–1083.

Hasegawa H, Muraoka M, Matsui K, and Kojima A (2003) Discovery of a novel potentNa1/Ca21 exchanger inhibitor: design, synthesis and structure-activity relation-ships of 3,4-dihydro-2(1H)-quinazolinone derivatives. Bioorg Med Chem Lett 13:3471–3475.

Hasenfuss G, Schillinger W, Lehnart SE, Preuss M, Pieske B, Maier LS, Prestle J,Minami K, and Just H (1999) Relationship between Na1-Ca21-exchanger proteinlevels and diastolic function of failing human myocardium. Circulation 99:641–648.

Hossmann KA (1996) Excitotoxic mechanisms in focal ischemia. Adv Neurol 71:69–74.

Iwamoto T, Inoue Y, Ito K, Sakaue T, Kita S, and Katsuragi T (2004a) The exchangerinhibitory peptide region-dependent inhibition of Na1/Ca21 exchange by SN-6 [2-[4-(4-nitrobenzyloxy)benzyl]thiazolidine-4-carboxylic acid ethyl ester], a novelbenzyloxyphenyl derivative. Mol Pharmacol 66:45–55.

Iwamoto T and Kita S (2006) YM-244769, a novel Na1/Ca21 exchange inhibitor thatpreferentially inhibits NCX3, efficiently protects against hypoxia/reoxygenation-induced SH-SY5Y neuronal cell damage. Mol Pharmacol 70:2075–2083.

Iwamoto T, Kita S, Uehara A, Imanaga I, Matsuda T, Baba A, and Katsuragi T(2004b) Molecular determinants of Na1/Ca21 exchange (NCX1) inhibition bySEA0400. J Biol Chem 279:7544–7553.

Iwamoto T, Kita S, Uehara A, Inoue Y, Taniguchi Y, Imanaga I, and Shigekawa M(2001) Structural domains influencing sensitivity to isothiourea derivative in-hibitor KB-R7943 in cardiac Na(1)/Ca(21) exchanger. Mol Pharmacol 59:524–531.

Iwamoto T, Uehara A, Nakamura TY, Imanaga I, and Shigekawa M (1999) Chimericanalysis of Na(1)/Ca(21) exchangers NCX1 and NCX3 reveals structural domainsimportant for differential sensitivity to external Ni(21) or Li(1). J Biol Chem 274:23094–23102.

Iwamoto T, Wakabayashi S, Imagawa T, and Shigekawa M (1998) Na1/Ca21 exchangeroverexpression impairs calcium signaling in fibroblasts: inhibition of the [Ca21] in-crease at the cell periphery and retardation of cell adhesion. Eur J Cell Biol 76:228–236.

Iwamoto T, Watano T, and Shigekawa M (1996) A novel isothiourea derivative se-lectively inhibits the reverse mode of Na1/Ca21 exchange in cells expressingNCX1. J Biol Chem 271:22391–22397.

Jeffs GJ, Meloni BP, Bakker AJ, and Knuckey NW (2007) The role of the Na(1)/Ca(21) exchanger (NCX) in neurons following ischaemia. J Clin Neurosci 14:507–514.

Jeon D, Chu K, Jung KH, KimM, Yoon BW, Lee CJ, Oh U, and Shin HS (2008) Na(1)/Ca(21) exchanger 2 is neuroprotective by exporting Ca(21) during a transient focalcerebral ischemia in the mouse. Cell Calcium 43:482–491.

Ketelaars SO, Gorter JA, Aronica E, and Wadman WJ (2004) Calcium extrusionprotein expression in the hippocampal formation of chronic epileptic rats afterkainate-induced status epilepticus. Epilepsia 45:1189–1201.

Litwin SE and Bridge JH (1997) Enhanced Na(1)-Ca21 exchange in the infarctedheart. Implications for excitation-contraction coupling. Circ Res 81:1083–1093.

Makino N, Zhao D, and Dhalla NS (1988) Stimulation of heart sarcolemmal Na1-Ca21 exchange by concanavalin A. Biochem Biophys Res Commun 154:245–251.

Matsuda T, Arakawa N, Takuma K, Kishida Y, Kawasaki Y, Sakaue M, Takahashi K,Takahashi T, Suzuki T, Ota T et al. (2001) SEA0400, a novel and selective inhibitorof the Na1-Ca21 exchanger, attenuates reperfusion injury in the in vitro and invivo cerebral ischemic models. J Pharmacol Exp Ther 298:249–256.

Min JY, Sullivan MF, Yan X, Feng X, Chu V, Wang JF, Amende I, Morgan JP,Philipson KD, and Hampton TG (2002) Overexpression of Na1/Ca21 exchangergene attenuates postinfarction myocardial dysfunction. Am J Physiol Heart CircPhysiol 283:H2466–H2471.

Molinaro P, Cuomo O, Pignataro G, Boscia F, Sirabella R, Pannaccione A, Secondo A,Scorziello A, Adornetto A, Gala R et al. (2008) Targeted disruption of Na1/Ca21exchanger 3 (NCX3) gene leads to a worsening of ischemic brain damage. J Neu-rosci 28:1179–1184.

Molinaro P, Viggiano D, Nisticò R, Sirabella R, Secondo A, Boscia F, Pannaccione A,Scorziello A, Mehdawy B, Sokolow S et al. (2011) Na1 -Ca21 exchanger (NCX3)knock-out mice display an impairment in hippocampal long-term potentiation andspatial learning and memory. J Neurosci 31:7312–7321.

Nicoll DA, Hryshko LV, Matsuoka S, Frank JS, and Philipson KD (1996) Mutation ofamino acid residues in the putative transmembrane segments of the cardiac sar-colemmal Na1-Ca21 exchanger. J Biol Chem 271:13385–13391.

Nicoll DA, Ottolia M, Lu L, Lu Y, and Philipson KD (1999) A new topological model ofthe cardiac sarcolemmal Na1-Ca21 exchanger. J Biol Chem 274:910–917.

Ottolia M, Schumann S, Nicoll DA, and Philipson KD (2002) Activation of the cardiacNa1/Ca21 exchanger by DEPC. Ann N Y Acad Sci 976:85–88.

Pannaccione A, Secondo A, Molinaro P, D’Avanzo C, Cantile M, Esposito A, Boscia F,Scorziello A, Sirabella R, Di Renzo G et al. (2012) A new concept: Ab1-42 generatesa hyperfunctional proteolytic NCX3 fragment that delays caspase-12 activationand neuronal death. J Neurosci 32:10609–10617.

Philipson KD and Nicoll DA (2000) Sodium-calcium exchange: a molecular perspec-tive. Annu Rev Physiol 62:111–133.

Pignataro G, Gala R, Cuomo O, Tortiglione A, Giaccio L, Castaldo P, Sirabella R,Matrone C, Canitano A, Amoroso S et al. (2004a) Two sodium/calcium exchangergene products, NCX1 and NCX3, play a major role in the development of perma-nent focal cerebral ischemia. Stroke 35:2566–2570.

Pignataro G, Tortiglione A, Scorziello A, Giaccio L, Secondo A, Severino B, SantagadaV, Caliendo G, Amoroso S, Di Renzo G et al. (2004b) Evidence for a protective roleplayed by the Na1/Ca21 exchanger in cerebral ischemia induced by middle ce-rebral artery occlusion in male rats. Neuropharmacology 46:439–448.

Pintado AJ, Herrero CJ, García AG, and Montiel C (2000) The novel Na(1)/Ca(21)exchange inhibitor KB-R7943 also blocks native and expressed neuronal nicotinicreceptors. Br J Pharmacol 130:1893–1902.

Quednau BD, Nicoll DA, and Philipson KD (1997) Tissue specificity and alternativesplicing of the Na1/Ca21 exchanger isoforms NCX1, NCX2, and NCX3 in rat. AmJ Physiol 272:C1250–C1261.

Reeves JP, Bailey CA, and Hale CC (1986) Redox modification of sodium-calciumexchange activity in cardiac sarcolemmal vesicles. J Biol Chem 261:4948–4955.

Reuter H, Henderson SA, Han T, Matsuda T, Baba A, Ross RS, Goldhaber JI,and Philipson KD (2002) Knockout mice for pharmacological screening: testing thespecificity of Na1-Ca21 exchange inhibitors. Circ Res 91:90–92.

Robello M, Amico C, and Cupello A (1993) Regulation of GABAA receptor in cere-bellar granule cells in culture: differential involvement of kinase activities. Neu-roscience 53:131–138.

Scorziello A, Santillo M, Adornetto A, Dell’aversano C, Sirabella R, Damiano S,Canzoniero LM, Renzo GF, and Annunziato L (2007) NO-induced neuroprotectionin ischemic preconditioning stimulates mitochondrial Mn-SOD activity and ex-pression via Ras/ERK1/2 pathway. J Neurochem 103:1472–1480.

Secondo A, Molinaro P, Pannaccione A, Esposito A, Cantile M, Lippiello P, SirabellaR, Iwamoto T, Di Renzo G, and Annunziato L (2011) Nitric oxide stimulates NCX1and NCX2 but inhibits NCX3 isoform by three distinct molecular determinants.Mol Pharmacol 79:558–568.

Secondo A, Pannaccione A, Molinaro P, Ambrosino P, Lippiello P, Esposito A, Cantile M,Khatri PR, Melisi D, Di Renzo G et al. (2009) Molecular pharmacology of the amilorideanalog 3-amino-6-chloro-5-[(4-chloro-benzyl)amino]-n-[[(2,4-dimethylbenzyl)-amino]

Neurounina-1 Increases NCX Activity and Protects against Stroke 155

at ASPE

T Journals on M

ay 13, 2018m

olpharm.aspetjournals.org

Dow

nloaded from

iminomethyl]-pyrazinecarboxamide (CB-DMB) as a pan inhibitor of the Na1-Ca21exchanger isoforms NCX1, NCX2, and NCX3 in stably transfected cells. J PharmacolExp Ther 331:212–221.

Secondo A, Staiano RI, Scorziello A, Sirabella R, Boscia F, Adornetto A, Valsecchi V,Molinaro P, Canzoniero LM, Di Renzo G et al. (2007) BHK cells transfected withNCX3 are more resistant to hypoxia followed by reoxygenation than those trans-fected with NCX1 and NCX2: possible relationship with mitochondrial membranepotential. Cell Calcium 42:521–535.

Sipido KR, Volders PG, Vos MA, and Verdonck F (2002) Altered Na/Ca exchangeactivity in cardiac hypertrophy and heart failure: a new target for therapy? Car-diovasc Res 53:782–805.

Sirabella R, Secondo A, Pannaccione A, Scorziello A, Valsecchi V, Adornetto A, Bilo L,Di Renzo G, and Annunziato L (2009) Anoxia-induced NF-kappaB-dependentupregulation of NCX1 contributes to Ca21 refilling into endoplasmic reticulum incortical neurons. Stroke 40:922–929.

Stengl M, Mubagwa K, Carmeliet E, and Flameng W (1998) Phenylephrine-inducedstimulation of Na1/Ca21 exchange in rat ventricular myocytes. Cardiovasc Res38:703–710.

Stigliani S, Zappettini S, Raiteri L, Passalacqua M, Melloni E, Venturi C, TacchettiC, Diaspro A, Usai C, and Bonanno G (2006) Glia re-sealed particles freshly

prepared from adult rat brain are competent for exocytotic release of glutamate.J Neurochem 96:656–668.

Tardito D, Milanese M, Bonifacino T, Musazzi L, Grilli M, Mallei A, Mocaer E,Gabriel-Gracia C, Racagni G, Popoli M et al. (2010) Blockade of stress-inducedincrease of glutamate release in the rat prefrontal/frontal cortex by agomelatineinvolves synergy between melatonergic and 5-HT2C receptor-dependent pathways.BMC Neurosci 11:68.

Watano T, Harada Y, Harada K, and Nishimura N (1999) Effect of Na1/Ca21 ex-change inhibitor, KB-R7943 on ouabain-induced arrhythmias in guinea-pigs. Br JPharmacol 127:1846–1850.

Woo SH and Morad M (2001) Bimodal regulation of Na(1)–Ca(21) exchanger bybeta-adrenergic signaling pathway in shark ventricular myocytes. Proc Natl AcadSci USA 98:2023–2028.

Address correspondence to: Lucio Annunziato, Division of Pharmacology,Department of Neuroscience, School of Medicine, “Federico II” University ofNaples, Via Pansini 5, 80131 Naples, Italy. E-mail: lannunzi@unina.it

156 Molinaro et al.

at ASPE

T Journals on M

ay 13, 2018m

olpharm.aspetjournals.org

Dow

nloaded from