Effect of Gambierol and Its Tetracyclic and Heptacyclic Analogues in Cultured Cerebellar Neurons: A Structure–Activity Relationships Study

Effect of Gambierol and Its Tetracyclic and Heptacyclic Analogues inCultured Cerebellar Neurons: A Structure−Activity RelationshipsStudySheila Perez,† Carmen Vale,† Eva Alonso,† Haruhiko Fuwa,‡ Makoto Sasaki,‡ Yu Konno,‡ Tomomi Goto,‡

Yuto Suga,‡ Mercedes R. Vieytes,§ and Luis M. Botana*,†

†Departamento de Farmacología, Facultad de Veterinaria, Universidad de Santiago de Compostela, Lugo, Spain‡Graduate School of Life Sciences, Tohoku University, 2-1-1 Katahira, Aoba-ku, Sendai, Japan§Departamento de Fisiología, Facultad de Veterinaria, Universidad de Santiago de Compostela, Lugo, Spain

ABSTRACT: The polycyclic ether class of marine naturalproducts has attracted the attention of researchers due to theircomplex and large chemical structures and diverse biologicalactivities. Gambierol is a marine polycyclic ether toxin, firstisolated along with ciguatoxin congeners from the dino-flagellate Gambierdiscus toxicus. The parent compoundgambierol and the analogues evaluated in this work share themain crucial elements for biological activity, previouslydescribed to be the C28C29 double bond within the Hring and the unsaturated side chain [Fuwa, H., Kainuma, N.,Tachibana, K., Tsukano, C., Satake, M., and Sasaki, M. (2004) Diverted total synthesis and biological evaluation of gambierolanalogues: Elucidation of crucial structural elements for potent toxicity. Chem. Eur. J. 10, 4894−4909]. With the aim to gain adeeper understanding of the cellular mechanisms involved in the biological activity of these compounds, we compared its activityin primary cultured neurons. The three compounds inhibited voltage-gated potassium channels (Kv) in a concentration-dependent manner and with similar potency, caused a small inhibition of voltage-gated sodium channels (Nav), and evokedcytosolic calcium oscillations. Moreover, the three compounds elicited a “loss of function” effect on Kv channels atconcentrations of 0.1 nM. Additionally, both the tetracyclic and the heptacyclic derivatives of gambierol elicited synchronouscalcium oscillations similar to those previously described for gambierol in cultured cerebellar neurons. Neither gambierol nor itstetracyclic derivative elicited cell toxicity, while the heptacyclic analogue caused a time-dependent decrease in cell viability.Neither the tetracyclic nor the heptacyclic analogues of gambierol exhibited lethality in mice after ip injection of 50 or 80 μg/kgof each compound. Altogether, the results presented in this work support an identical mechanism of action for gambierol and itstetracyclic and heptacyclic analogues and indicate a “loss of function” effect on potassium channels even after administration ofthe three compounds at subnanomolar concentrations. In addition, because gambierol is known to stabilize the closed state ofKv3 channels, the results presented in this paper may have implications for understanding of channel functions and for futuredevelopment of therapies against ciguatera poisoning and potassium channel-related diseases.

■ INTRODUCTION

Gambierol is a trans-fused octacyclic polyether first isolatedwith ciguatoxins from the dinoflagellate Gambierdiscus toxicus.2

Because of the common biogenetic origin of both gambieroland ciguatoxins and their polycyclic structure, it has long beenspeculated that gambierol may contribute to the symptoms ofciguatera.3 Ciguatera fish poisoning (CFP) is a major economicand social problem worldwide, with more than 25000 personspoisoned annually.4 The consumption of these toxins throughcontaminated fish causes human illness whose clinicalmanifestation is characterized by gastrointestinal, neurological,and cardiac symptoms. The neurological alterations includesensory abnormalities, such as tingling lips, hands, or feet,unusual temperature sensation, ataxia, hallucination, andparesthesia. The neurological symptoms, caused by ciguatoxin(CTX) in mice, are similar to the neurological disturbances

caused by gambierol.5 Gambierol exhibits toxicity in mice witha minimal lethal dose ranging from 50 to 75 μg/kg byintraperitoneal injection.5 By oral administration, the LD100 ofthe toxin was 150 μg/kg, while an intraperitoneal dose of 80μg/kg killed seven animals out of a group of eight mice.5 Inhumans, the minimum toxicity level for CTXs is estimated at0.5 ng/g.6 There are a few cases of fatality symptoms associatedwith CTXs, with paralysis, coma, and even death.7

CTXs are known to act on voltage-gated sodium channels(Nav) with affinities in the nanomolar range. Their mechanismof action includes a hyperpolarizing shift in the voltagedependence of channel activation, causing channel opening atresting membrane potentials and disruption of channel

Received: June 1, 2012Published: August 15, 2012

Article

pubs.acs.org/crt

© 2012 American Chemical Society 1929 dx.doi.org/10.1021/tx300242m | Chem. Res. Toxicol. 2012, 25, 1929−1937

inactivation leading to persistent activation.8 However, adifferent potency of CTXs to affect both Nav and voltage-gated potassium channels (Kv), in cultured cerebellar neurons,recently has been described.9

The accomplishment of the chemical synthesis of gambier-ol10 has made it possible to begin detailed biological andpharmacological analysis of its effects on living tissue.1,5

However, the actions of gambierol on voltage-gated channelsare not yet completely known. Thus, it has been reported thatthe toxin did not affect Nav channels in mouse taste cells12 norindividual sodium channels in Xenopus,13 but it inhibited theintracellular calcium increase caused by the VGSC activatorbrevetoxin 2.14 In addition, gambierol shows a potent blockingeffect on Kv channels in skeletal muscle cells15 and isolated Kvchannels expressed in Xenopus laevis oocytes.13 In a recentwork, we have shown that gambierol and its analoguesproduced a concentration-dependent inhibition of potassiumcurrent densities in primary cortical neurons.16 Moreover, wehave recently shown that different purified CTXs affectedsodium and potassium channels in cultured cerebellar neurons,although with potencies in the low nanomolar range,promoting a decrease in Kv current amplitude and ahyperpolarizing shift in Nav channel activation.9 In addition,in the same neuronal system, gambierol, at concentrationsranging from 0.1 to 30 μM, elicited calcium oscillationssuggested to be due to its inhibitory action on Kv channels andhyperpolaryzation of sodium channel activation.17 In this work,we extended the analysis of the effects of gambierol incerebellar neuron by evaluating the effect of nanomolar toxinconcentrations on Nav and Kv channels using electro-physiological recording techniques. Cultured cerebellar granulecells were chosen to perform this analysis because theyconstitute a reliable neuronal model for the study of neuralfunction and pathology18,19 and have been shown previously tobe sensitive to both CTXs and gambierol.9,17

Previous work has identified different structural elements ofgambierol indispensable for exhibiting potent toxicity1 andrevealed that the C1 modified analogues have similar toxicity tothe parent compound, but they probably have relatively lowabsorption and/or distribution properties. The same studydescribed that the structural elements of gambierol indispen-sable for exhibiting potent toxicity are the C28C29 doublebond and the unsaturated side chain. In this context, we haveevaluated the structure−activity relationship of gambierol andits heptacyclic and tetracyclic derivatives containing, respec-tively, the B−H rings and the E−H rings of the parentcompound, together with the C28C29 double bond, and theunsaturated side chain but lacking the C1 and C6 hydroxygroup of gambierol, to establish the structural elements that areimportant for the cellular effect of gambierol and gain a deeperunderstanding of the cellular mechanisms involved in thebiological activity of these compounds.

■ MATERIAL AND METHODSChemicals and Solutions. Seven day old Swiss mice were

obtained from the animal care facilities of the University of Santiago deCompostela. Plastic tissue culture dishes were purchased from Falcon(Madrid, Spain). Fetal calf serum was obtained from Gibco (Glasgow,United Kingdom), and Dulbecco's modified Eagle's medium (DMEM)was from Biochrom (Berlin, Germany). Saxitoxin (STX) was purifiedin our laboratory. Gambierol and its tetracyclic and heptacyclicanalogues were obtained by chemical synthesis.11 All other chemicalswere reagent grade and purchased from Sigma-Aldrich (Madrid,Spain).

Cell Cultures. Primary cultures of cerebellar granule cells wereobtained from the cerebella of 7 day old mice as previouslydescribed.19,20 In brief, cells were dissociated by mild trypsinizationat 37 °C, followed by trituration in a DNase solution (0.004% w/v)containing a soybean trypsin inhibitor (0.05% w/v). The cells weresuspended in DMEM containing 25 mM KCl, 31 mM glucose, and 0.2mM glutamine supplemented with p-amino benzoate, insulin,penicillin, and 10% fetal calf serum. The cell suspension was seededin 18 mm glass coverslips precoated with poly-D-lysine and incubatedin 12 multiwell plates for 6−11 days in vitro (div) in a humidified 5%CO2/95% air atmosphere at 37 °C. Cytosine arabinoside, 20 μM, wasadded before 48 h in culture to prevent glial proliferation.

Electrophysiology. Membrane currents from single cells werestudied at room temperature (22−25 °C) by whole-cell patchrecordings in voltage-clamp mode21,22 using a computer-controlledcurrent and voltage-clamp amplifier (Multiclamp 700B, MolecularDevices). Signals were recorded and analyzed using a Pentiumcomputer equipped with Digidata 1440 data acquisition system andpClamp10 software (Molecular Devices, Sunnyvale, CA). pClamp10was used to generate current and voltage-clamp commands and torecord the resulting data. Signals were prefiltered at 10 kHz anddigitized at 20 μs intervals.

Recording electrodes were fabricated from borosilicate glassmicrocapillaries (outer diameter, 1.5 mm), and the tip resistance was5−10 MΩ. The internal pipet solution contained (in mM): 108 Csgluconate, 1.7 NaCl, 0.9 EGTA, 9 HEPES, 1.8 MgCl2, 4 Na2ATP, and0.3 NaGTP, pH 7.2.23 After the membrane resistance had stabilized(usually between 5 and 20 min after obtaining the GΩ seal), data wereobtained. For the whole-cell patch-clamp recordings, the extracellularmedium was Locke's buffer containing (in mM): 154 NaCl, 5.6 KCl,3.6 NaHCO3, 1.3 CaCl2, 1 MgCl2, 5 glucose, and 10 HEPES, pH 7.4.

For voltage-dependent sodium channels, voltage-gated ion currentswere elicited in cerebellar granule cells (CGCs) by applying a series of25 ms depolarizing pulses (voltage steps) in 5 mV increments from aholding potential of −100 mV;24 moreover, 20 mM tetraethylammo-nium (TEA) and 1 mM 4-aminopyridine (4-AP) were added to thebathing solution. Voltage-gated sodium currents were obtained bymeasuring the maximum peak amplitude of the current in the absenceand in the presence of gambierol or its heptacyclic and tetracyclicanalogues. As granule cell neurons display two main voltage-dependent outward K+ currents, fast transient outward potassiumcurrent (IA), and delayed rectifier outward potassium currents (IK), theeffect of gambierol and analogues on these currents was also examined.To evaluate the effect of gambierol and its tetracyclic and heptacyclicanalogues on K+ channels, cesium was replaced by K+ in theintracellular pipet solution and contained (in mM) 129 potassiumgluconate, 5 KCl, 0.6 EGTA, 9 HEPES, 2 MgCl2, 1.2 Na2ATP, 3.3NaGTP, pH 7.2, and STX, at a final concentration of 50 nM, wasadded to the bathing solution. In these experiments, transient outwardIA and delayed rectifier IK currents were elicited by two sequential 200ms depolarizing pulses to 40 mV at 1 s interval. The holding potentialswere set to −100 mV (first pulse) for activation of the global K+

current (IA plus IK current) and at −40 mV (second pulse) foractivation of IK currents. IA currents were obtained by subtraction ofthe outward K+ currents elicited at a holding potential of −40 mVfrom the global K+ current evoked at a holding potential of −100 mV.25

Determination of the Cytosolic Calcium Concentration[Ca2+]c. CGCs cultured from 6 to 11 days in vitro (div) were loadedwith the Ca2+-sensitive fluorescent dye Fura-2 acetoxymethyl ester(Fura-2AM; 2.5 μM) for 10 min at 37 °C. After incubation, the loadedcells were washed three times with cold buffer. The glass coverslipswere inserted into a thermostatted chamber at 37 °C (Life ScienceResources, Royston, Herts, United Kingdom), and cells were viewedwith a Nikon Diaphot 200 microscope, equipped with epifluorescenceoptics (Nikon 40× immersion UV-Fluor objective). The thermo-statted chamber was used in the open bath configuration, andadditions were made by removal and addition of fresh bathingsolution.

Chemical Research in Toxicology Article

dx.doi.org/10.1021/tx300242m | Chem. Res. Toxicol. 2012, 25, 1929−19371930

The [Ca2+]c was obtained from the images collected by doubleexcitation fluorescence with a Life Science Resources equipment. Thelight source was a 175 W xenon lamp, and light reached the objectivewith an optical fiber. The excitation wavelengths for Fura-2AM were340 and 380 nm, with emission at 505 nm. The calibration of thefluorescence versus intracellular calcium was made by using themethod described by Grynkiewicz.26 In these experiments, the mediaused was ACSF solution containing (in mM): 123 NaCl, 4 KCl, 1.2KH2PO4, 1.3 MgSO4, 28 NaHCO3, 15 glucose, and 2.4 CaCl2. In all ofthe assays, the medium was equilibrated with CO2 prior to use, toadjust the final pH to 7.4. The pH was maintained constant bybubbling CO2 during the experiment. All experiments were carried outin duplicate.Cell Viability Assay. The cytotoxic action of gambierol and its

tetracyclic and heptacyclic analogues was studied in cultured CGC bythe 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide(MTT) test. The MTT assay was performed as describedpreviously.20,27,28 This test, which measures mitochondrial function,was used to assess cell viability, as it has been shown that in neuronalcells there is a good correlation between a drug-induced decrease inmitochondrial activity and its cytotoxicity.29 Briefly, after 4, 8, 24, 48,and 72 h of exposure to different concentrations of gambierol and itsanalogues, cells were rinsed and incubated for 30 min with a solutionof MTT (500 μg/mL) dissolved in Locke's buffer. After excess MTTwas washed off, the cells were disaggregated with 5% sodium dodecylsulfate, and the colored formazan salt was measured at 590 nM in aspectrophotometer plate reader.Determination of the Toxicity of the Compounds. The

toxicity of the tetracyclic and heptacyclic analogues of gambierol wasevaluated by the mouse bioassay. Stock solutions of the compoundswere made in DMSO, and aliquots of the stock solution were made in0.2 mL of saline as previously described for gambierol.5 Swiss mice(three for each dose) were injected intraperitoneally with the twoanalogues of gambierol at doses of 50 and 80 μg/kg.Statistical Method. All data are expressed as means ± SEMs of n

experiments (each performed in duplicate). Statistical comparison wasby nonpaired Student's t test. P values <0.05 were consideredstatistically significant.

■ RESULTS

In this study, we have analyzed the effect of gambierol and itsanalogues on Kv and Nav channels, cytosolic calciumconcentration, and cell viability of cultured cerebellar neurons.The chemical structure of gambierol and its tetracyclic andheptacyclic analogues is shown in Figure 1. CGCs expresstetrodotoxin-sensitive sodium channels mainly attributable toNav1.6, Naβ1, and Naβ2 sodium channels in the soma.30

Potassium currents identified in these neurons include thetransient IA current due to the expression of Kv4.2 channels orKv4.2/4.3 heteromers and the α-subunits of the Kv2 family,which encode the neuronal delayed-rectifier current IK,

31

although this later current has as many as six different potentialcontributors from the Kv family of α-subunits.32 Because wehave recently described that gambierol and its tetracyclic andheptacyclic analogues inhibited Kv channels in corticalneurons,16 we first compared the activity of the threecompounds on both Nav and Kv channels in CGCs.Effect of Nanomolar Concentrations of Gambierol on

Sodium and Potassium Currents in CGCs. Gambierol isknown to inhibit voltage-gated potassium currents in thenanomolar range in different biological preparations.12,13,15,16,33

Therefore, we first evaluated the concentration−response effectof gambierol on Kv channels employing gambierol concen-trations ranging from 1 to 100 nM and different cells for eachconcentration. Figure 2a shows representative traces for theeffect of 10 nM gambierol on voltage-gated potassium currents.

As shown in Figure 2b, gambierol caused a concentration-dependent inhibition of the total potassium current (IA+ IK)and both the fast inactivating current component (IA) and thedelayed rectifier current component (IK). A nonlinear fit of thedata shown in Figure 2b yielded an estimated IC50 (95%confidence intervals) for the gambierol inhibition of Kvchannels of 8.9 nM (2.5−32.2 nM) on the total current, of9.3 nM (3.4−25.9 nM) on the IA current, and of 9.1 nM (1.47−56.3 nM) on the IK current with complete inhibition of thecurrents at the highest concentration evaluated (100 nM).These results showed a higher inhibition of potassium currentsby gambierol than previously reported by us in cortical neuronswhere gambierol inhibited IK currents about 30% after additionof consecutive doses of the toxin to the same cell.16 Therefore,and because of the previously described properties of gambierolto stabilize the closed state of the potassium channel,33 thesame experiment was performed by adding all doses of thetoxin at the same cell. In this case, as shown in Figure 2c, theIC50 (95% confidence intervals) for the gambierol inhibition ofKv channels administering consecutive concentrations of toxinto the same cell at 5 min intervals were as follows: 629 nM (199nM to 1.98 μM) for the total current, 104 nM (51.5−212 nM)for the IA current, and 688 nM (212 nM to 2.22 μM) for the IKcurrent.Previous studies on the effect of gambierol on Nav have

revealed controversial results. Thus, the toxin has been foundnot to affect sodium currents at nanomolar concentration inmouse taste cells,12 to act as a partial agonist of sodiumchannels in human neuroblastoma cells,34 and to decrease theamplitude of sodium currents in primary cerebellar neurons atmicromolar concentrations.17 To evaluate the effect ofgambierol on sodium currents, we analyzed the effect of thetoxin at nanomolar concentrations. As shown in Figure 3, peakinward sodium current decreased by 37.8 ± 16.5 (p = 0.07) and33.6 ± 16.7% (p = 0.06) in the presence of 10 and 100 nMgambierol, respectively, in the extracellular solution. None ofthe gambierol concentrations evaluated in this work modifiedthe activation potential of sodium currents.

Figure 1. Chemical structures of gambierol and its heptacyclic andtetracyclic analogues.

Chemical Research in Toxicology Article

dx.doi.org/10.1021/tx300242m | Chem. Res. Toxicol. 2012, 25, 1929−19371931

Effect of Tetracyclic Analogue of Gambierol onSodium and Potassium Currents in CGCs. To comparethe relative potencies of gambierol and its tetracyclic analogue,which contains the E−H rings of gambierol, we first analyzedthe effect of the tetracyclic compound on both Kv and Navchannels. Figure 4 shows the analysis of the effect of tetracyclicanalogue on Kv channel amplitude. As reported above forgambierol, the effect of the tetracyclic analogue was evaluatedfirst by adding each concentration into one single cell and afterby adding all concentrations in the same cell. Figure 4a showsrepresentative traces for the effect of a concentration of 10 nMtetracyclic analogue on voltage-gated potassium currents. Asshown in Figure 4b, the tetracyclic compound inhibited both IAand IK in a concentration-dependent manner yielding anestimated IC50 (95% CI) for the inhibition of the totalpotassium current of 12.4 nM (7.9−19.3 nM), while the IC50for inhibition of the IA current was 6.8 (4.4−10.6 nM) and 19.3nM (9.7−38.6 nM) for inhibition of the IK current with fullinhibition of all of the currents after adding the compound at100 nM. Similarly to gambierol, the inhibition of potassiumcurrent by the tetracyclic analogue was much smaller whensuccessive applications of the compound were realized in thesame cell. As shown in Figure 4c, in this case, the IC50 forinhibition of the total potassium current was 477 nM (95% CI,175 nM to 1.3 μM), 30.6 nM (95% CI, 9.21−101 nM) for

inhibition of the IA current, and 895 nM (95% CI, 408 nM to1.96 μM) for inhibition of the IK current. Figure 4d shows theeffect of the compound on the peak amplitude of inwardsodium currents. Bath application of the compound caused aconcentration-dependent decrease on Nav current amplitude.Thus, inward sodium current decreased by 7.5 ± 21.8% (p =0.14) in the presence of this analogue at 10 nM and by 38.9 ±11.9% (p = 0.017) after bath application of the compound at100 nM. This compound did not modify the activationpotential of sodium currents at 10 nM, while it significantlyhyperpolarized sodium channel activation at 100 nM (control,−26.7 ± 1.7 mV; tetracyclic, −38.3 ± 3.3 mV; n = 3; p =0.017).

Effect of the Heptacyclic Analogue of Gambierol onSodium and Potassium Currents in CGCs. To completethe evaluation of the structure−activity relationships ofgambierol and its analogues, we also analyzed the effect ofthe heptacyclic analogue containing the B−H rings ofgambierol, on voltage-gated potassium and sodium currents.Figure 5 shows the effect of this compound on potassiumcurrents. Figure 5a shows representative traces for the effect ofthe heptacyclic analogue at 10 nM on voltage-gated potassiumcurrents after administering each concentration in each cellseparately. In this case, as shown in Figure 5b, the heptacyclicanalogue of gambierol also affected Kv channels in aconcentration-dependent manner. Thus, the compound yieldedan estimated IC50 (95% CI) for inhibition of the totalpotassium current of 6.4 nM (2.6−16.2 nM), while its IC50for inhibition of the IA and IK currents was 7.2 (2.1−24.6 nM)and 8.7 nM (3.6−21.1 nM), respectively, applying oneconcentration per cell and providing complete inhibition afteraddition of the compound at 100 nM. As shown in Figure 5c,the addition of successive compound concentrations to thesame cell decreased Kv inhibition to 50−70%. In this case, theestimated IC50 (95% CI) for inhibition of the total potassiumcurrent was 195 nM (107−356 nM), while the IC50 forinhibition of the IA current was 472 (0.2−0.009 M) and 121nM (67.7−218 nM) for inhibition of the IK current. The effect

Figure 2. Effect of gambierol on potassium currents in CGCs. (a) Representative K+ currents elicited in the presence of 50 nM STX in theextracellular solution in the absence (black trace) and in the presence of 10 nM gambierol (gray trace). (b) Concentration−response curves showingthe effect of gambierol on the total potassium current (IA + IK), fast inactivating (IA), and delayed rectifier (IK) potassium currents in cerebellarneurons adding one concentration to a single cell. (c) Concentration−response curves showing the effect of gambierol on the total potassiumcurrent (IA + IK), fast inactivating (IA), and delayed rectifier (IK) potassium currents in cerebellar neurons after sequential addition of the differentconcentrations of gambierol to the same cell.

Figure 3. Effect of gambierol on Nav current amplitude in cerebellarneurons. The peak amplitude of the sodium currents was measured inthe presence of gambierol at concentrations of 10 and 100 nM.

Chemical Research in Toxicology Article

dx.doi.org/10.1021/tx300242m | Chem. Res. Toxicol. 2012, 25, 1929−19371932

of the compound on Nav is shown in Figure 5d. Bathapplication of this gambierol analogue caused a concentration-dependent decrease on Nav current amplitude. Thus, peakinward sodium current decreased by 18.1 ± 0.6% (p ≤ 0.5) inthe presence of the heptacyclic compound at 10 nM and by45.9 ± 11.2% (p ≤ 0.05) in the presence of 100 nM of theheptacyclic analogue in the extracellular solution. In addition,this analogue caused a significant hyperpolarizing shift on thesodium channel activation at both concentrations (control,−32.5 ± 2.5 mV; 10 nM heptacyclic, −40.0 ± 0 mV; and 100nM heptacyclic, −41.7 ± 1.7 mV).Effect of Gambierol and Its Tetracyclic and Heptacy-

clic Analogues on [Ca2+]c. We have previously reported thatgambierol at concentrations ranging from 0.1 to 30 μM evokescalcium oscillations in CGCs.17 Therefore, to compare thebiological activity of gambierol and its tetracyclic andheptacyclic analogues, we compared their effect on the cytosoliccalcium concentration in cerebellar neurons. As shown inFigure 6, gambierol and both the tetracyclic and the heptacyclicderivatives of gambierol at 10 μM elicited cytosolic calciumoscillations that were highly synchronous, as those previouslyreported for the parent compound gambierol.17

Cytotoxic Effect of Gambierol and Its Tetracyclic andHeptacyclic Analogues. To analyze the cytotoxic effect ofthe tetracyclic and heptacyclic analogues of gambierol incerebellar neurons and because no reports in this sense havebeen reported, the cultures were incubated with compoundconcentrations ranging from 1 nM to 10 μM for 4, 8, 24, 48,and 72 h in culture, and the cytotoxicity was evaluated with theMTT assay as illustrated in Figure 7. None of the compoundselicited cytotoxic effects after 4 h of exposure. Moreover,neither gambierol nor its tetracyclic analogue elicitedcytotoxicity at any of the incubation times evaluated asshown in Figure 7a. In contrast, the heptacyclic compoundelicited a cytotoxic effect that started after 8 h of incubation ofthe cells with the toxin and increased progressively with theexposure time as shown in Figure 7b, reaching an IC50 value of26.7 nM (95% confidence intervals: 13.1−54.3 nM) after 72 hof exposure of the neurons to the heptacyclic analogue ofgambierol.

In Vivo Effects of the Tetracyclic and HeptacyclicAnalogues of Gambierol. Because no reports on the lethalityof the tetracyclic and heptacyclic analogues of gambierol havebeen released so far, both analogues were administered to miceby intraperitoneal injection at doses of 50 and 80 μg/kg of each

Figure 4. Effect of the tetracyclic analogue of gambierol on sodium and potassium currents in CGCs. (a) Representative K+ currents elicited in thepresence of 50 nM STX in the extracellular solution in the absence (black trace) and in the presence of the tetracyclic analogue at 10 nM (graytrace). (b) Concentration−response curves showing the effect of the compound on the total potassium current (IA + IK), fast inactivating (IA), anddelayed rectifier (IK) potassium currents in cerebellar neurons when each concentration was administered to a single cell separately. (c)Concentration−response curves showing the effect of the compound on the total potassium current (IA + IK), fast inactivating (IA), and delayedrectifier (IK) potassium currents in cerebellar neurons, applying all concentrations to the same cell. (d) Effect of the analogue on the peak amplitudeof Nav currents in cerebellar neurons. *p < 0.05 vs control in the absence of the compound.

Chemical Research in Toxicology Article

dx.doi.org/10.1021/tx300242m | Chem. Res. Toxicol. 2012, 25, 1929−19371933

compound to evaluate the in vivo toxicity of the two analoguesof gambierol. None of the mice died within 2 h of the trial, but45 min after administration of the compound, mice remainedvery quiet, with respiratory distress. In the case of thetetracyclic compound, two of the three mice showed trobbingrespiration 45 min after ip injection of the compound at 50 μg/kg, while after ip administration of the compound at 80 μg/kg,the three mice evaluated showed respiratory distress thatstarted also after 45 min. Similar symptoms were observed afterip administration of the heptacyclic compound at 50 and 80μg/kg in the three mice tested at each dose, although in thiscase, frequent twitches were observed in one of the mice at the

highest dose tested, which started 1 h after injection anddisappeared after 15 min.

■ DISCUSSIONThe polycyclic ether class of marine natural products hasattracted the attention of researchers due to their complex andlarge chemical structure and their diverse biological activities.Gambierol is one of the marine polycyclic ether toxins, whichwas first isolated along with CTX congeners from thedinoflagellate G. toxicus.2,4 Because very few reports concerningthe structure−activity relationships of marine polycycliccompounds have been performed so far,1,36,37 the mainobjective of this study was to deepen in the analysis of thebiological effects of gambierol and its tetracyclic and heptacyclicanalogues. Primary cultures of CGCs were used as the cellularmodel because these cells are one of the most reliable neuronalmodels to analyze neuronal function and pathology.18,19

Gambierol is biosynthesized by G. toxicus dinoflagellates, theproducer of CTX, and the symptoms induced by gambierol aresimilar to those induced by CTXs, which are characterized bygastrointestinal and neurological symptoms; for this reason,gambierol is included in the ciguatera group of marine toxins.CTXs are sodium channel activator toxins, and recently,potassium and sodium channels have been proposed as possibletargets of gambierol.3,12,34 In this work, we have carried out acomparative study of the effect of gambierol and its tetracyclicand heptacyclic analogues on voltage-gated potassium and

Figure 5. Effect of the heptacyclic analogue of gambierol on sodium and potassium currents in CGCs. (a) Representative K+ currents elicited in thepresence of 50 nM STX in the extracellular solution in the absence (black trace) and in the presence of the heptacyclic analogue at 10 nM (graytrace). (b) Concentration−response curves showing the effect of the compound, applying one concentration to a single cell on the total potassiumcurrent (IA + IK), fast inactivating (IA), and delayed rectifier (IK) potassium currents in cerebellar neurons. (c) Concentration−response curvesshowing the effect of the compound, applying all concentrations to a single cell on the total potassium current (IA + IK), fast inactivating (IA), anddelayed rectifier (IK) potassium currents in cerebellar neurons. (d) Effect of the analogue on the peak amplitude of Nav currents in cerebellarneurons. *p < 0.05 vs control in the absence of the compound.

Figure 6. Cytosolic calcium oscillations generated by exposure ofCGCs of 7 days in culture to gambierol and its tetracyclic andheptacyclic analogues at 10 μM. Compound addition is indicated bythe arrow.

Chemical Research in Toxicology Article

dx.doi.org/10.1021/tx300242m | Chem. Res. Toxicol. 2012, 25, 1929−19371934

sodium channels, to complete a previous work, whichdemonstrated that these compounds are active in mice corticalneurons.16 Previous reports indicated that gambierol is a potentblocker of Kv channels in different cellular models,12,13,33 butthe effect of gambierol on sodium channels is still contradictory.Thus, gambierol, at micromolar concentrations, had no effect indifferent sodium channels subtypes except the Nav1.3subtype,13,37 while in CGCs, the toxin decreased Nav currentamplitude and shifted Nav current activation in a negativedirection at micromolar concentrations.17 Here, we found thatthe toxin at nanomolar concentrations did not modify theactivation potential of sodium currents, but it decreased sodiumcurrent amplitude. Because granule cells in the cerebellummainly express Nav1.2 (in the axon) and Nav1.6, Naβ1 andNaβ2 sodium channels in the soma30 and gambierol have beenshown not to affect either isolated Nav1.2 nor Nav1.6 channels,the effect of gambierol in this neuronal system can be attributedto the action of the toxin on Naβ1 and Naβ2 sodium channels.In addition, we reported here that the compound inhibited Kvchannels in a concentration-dependent manner, showingaffinities in the low nanomolar range as previously shown inother cellular systems,13,33 and suggested also for cerebellarneurons.17 The effect of gambierol in Kv channels in cerebellarneurons is in agreement with previous data indicating that thetoxin inhibited Kv3.1 potassium channels at very lownanomolar concentrations, while Kv2 and Kv4 channels wereinsensitive to micromolar concentrations of gambierol33 andwith the high diversity of Kv subtypes present in culturedcerebellar neurons.32

Both the gambierol and the tetracyclic and heptacycliccompounds affected Kv channels at very low nanomolarconcentrations. Although there is very little information aboutthe activity of gambierol analogues,16 previous structure−activity studies indicated that the C28C29 double bondwithin the H ring of gambierol and the unsaturated side chainare the crucial structural elements for the biological activity ofgambierol and analogues.1 The results reported in this paperare in agreement with the previous observations on thestructure−activity relationships of gambierol, since we foundthat gambierol and its tetracyclic and heptacyclic analoguesdisplayed a similar potency on both Navs and Kvs. However,the unexpected difference between the three compounds wasthe cytotoxicity of the heptacyclic analogue. Long-termexposure of cerebellar neurons to gambierol or its tetracyclicanalogues did not cause cell death, while the heptacycliccompound at nanomolar concentrations produced cellcytotoxicity that increased with the time of treatment. Thelack of cytotoxicity of gambierol in primary neuronal cultures isin agreement with a recent report indicating that the toxin didnot alter cell viability in human neuroblastoma cells.35

However, the surprising finding was the higher cytotoxicity ofthe heptacyclic compound. To confirm this finding, bothanalogues of gambierol were also injected intraperitoneally tomice. In this case, both compounds caused some symptomssimilar to those elicited by gambierol after ip injection of 50 and80 μg/kg of each compound, but none of the mice died duringthe 2 h observation period. This observation suggests that bothanalogues are active in vivo, but they seem to be less toxic thangambierol, since previous reports have reported a minimallethal dose of gambierol of 50−75 μg/kg after ip injection.5

Although differences in mouse strains can account for this effectand more detailed studies of the in vivo effects of thecompounds must be performed, previous reports on the acutetoxicity of gambierol employing different mouse strains showedlittle difference in the toxic response of gambierol.1,5 Thedecreased toxicity of the tetracyclic and heptacyclic analoguesof gambierol is consistent with the findings of Fuwa et al.,1

indicating that the loss of the ring A hydroxyl group or the C-1hydroxyl group decreases its toxicity to mice. Becausegambierol and its tetracyclic and heptacyclic analogues sharethe crucial structures for exerting potent biological activitypreviously reported to be the C28C29 double bond withinthe H ring and the unsaturated side chain,1 the highercytotoxicity of the heptacyclic compound found in this workcould be due to a higher potency of this analogue on differentpotassium channel subtypes. In this sense, Kv1 channels areknown to be involved in the apoptosis of cultured cerebellarneurons;38 therefore, further studies would be needed to fullyelucidate the Kv subtype affinities of gambierol and analoguesas well as the effect of the long-term exposure of the neurons tothese compounds on Kv and Nav expression. Moreover, theresults presented here indicate that the chemical modificationsin the gambierol analogues had little effect on their potency toinhibit Kv in vitro, suggesting that while the ring A hydroxylgroup and the C1-hydroxyl group of gambierol are importantfor in vivo toxicity, they are not important with regard to theirin vitro effects on potassium channels.Gambierol is known to inhibit different Kv1 potassium

channels subtypes with different potencies and an unknownmechanism13 and to stabilize the closed state of Kv3 channels.33

Moreover, during the course of our experiments, we observed amuch higher inhibition of Kv currents by the compounds when

Figure 7. Effect of different concentrations of gambierol and itstetracyclic and heptacyclic analogues on mitochondrial function asassessed with the MTT assay. Cerebellar neurons were exposed todifferent compound concentrations from 4 to 72 h in culture. (a)None of the concentrations of gambierol and the tetracycliccompound elicited cell toxicity even after 72 h of exposure of theneurons to the compounds. (b) Time dependence of the cytotoxicityof the heptacyclic analogue of gambierol. Means are presented aspercentages of control nontreated neurons and include results fromthree separate experiments, each performed in triplicate.

Chemical Research in Toxicology Article

dx.doi.org/10.1021/tx300242m | Chem. Res. Toxicol. 2012, 25, 1929−19371935

a single concentration was added to each cell. This observationprompted us to analyze the effect of gambierol and itsanalogues in Kv by adding different concentrations of thesetoxins in the same cell. In this case, the inhibition of Kvchannels by the compounds was significantly decreased andsimilar to those previously described by us in primary corticalneurons.16 Altogether, the results presented in this worksupport a similar mechanism of action for gambierol and itstetracyclic and heptacyclic analogues and indicate a “loss offunction” effect of potassium channels even after administrationof the three compounds at concentrations of 0.1 nM. This fact,as well as the exact mechanism beyond it and the Kv channelsinvolved, should be taken into account for future studiesaddressing the therapeutic effect of gambierol and its analogues.

■ AUTHOR INFORMATIONCorresponding Author*Tel/Fax: +34982822233. E-mail: [email protected] work was funded with the following FEDER cofunded-grants: From Ministerio de Ciencia y Tecnologia, Spain:SAF2009-12581 (subprograma NEF), AGL2009-13581-CO2-01, TRA2009-0189, and AGL2010-17875. From Xunta deGalicia, Spain: GRC 2010/10, and PGDIT 07MMA006261PR,PGIDIT (INCITE) 09MMA003261PR, PGDIT (INCITE)09261080PR, 2009/XA044, and 10PXIB261254 PR. From EUVIIth Frame Program: 211326-CP (CONffIDENCE), 265896BAMMBO, 265409 μAQUA, and 262649 BEADS, 312184PharmaSea. From the Atlantic Area Programme (Interreg IVBTrans-national): 2009-1/117 Pharmatlantic. From the Ministryof Education, Culture, Sports, Science and Technology(MEXT), Japan: Grants-in-Aid for Scientific Research onInnovative Areas (Nos. 24102507 and 23102016). From theJapan Society for Promotion of Science (JSPS): Grant-in-Aidfor Young Scientists (A) (No. 23681045) and Grant-in-Aid forScientific Research (A) (No. 21241050).NotesThe authors declare no competing financial interest.

■ ABBREVIATIONSCGCs, cerebellar granule cells; DMEM, Dulbecco's modifiedEagle's medium; STX, saxitoxin; CTX, ciguatoxin; CFP,ciguatera fish poisoning; 4-AP, 4-aminopyridine; TEA,tetraethylammonium; Nav, voltage-gated sodium channel; Kv,voltage-gated potassium channel

■ REFERENCES(1) Fuwa, H., Kainuma, N., Tachibana, K., Tsukano, C., Satake, M.,and Sasaki, M. (2004) Diverted total synthesis and biologicalevaluation of gambierol analogues: Elucidation of crucial structuralelements for potent toxicity. Chem. Eur. J. 10, 4894−4909.(2) Satake, M., Murata, M., and Yasumoto, T. (1993) Gambierol: Anew toxic polyether compound isolated from the marine dinoflagellateGambierdiscus toxicus. J. Am. Chem. Soc. 115, 361−362.(3) Inoue, M., Hirama, M., Satake, M., Sugiyama, K., and Yasumoto,T. (2003) Inhibition of brevetoxin binding to the voltage-gatedsodium channel by gambierol and gambieric acid-A. Toxicon 41, 469−474.(4) Lewis, R. J. (2001) The changing face of ciguatera. Toxicon 39,97−106.(5) Ito, E., Suzuki-Toyota, F., Toshimori, K., Fuwa, H., Tachibana, K.,Satake, M., and Sasaki, M. (2003) Pathological effects on mice bygambierol, possibly one of the ciguatera toxins. Toxicon 42, 733−740.

(6) Legrand, A. M. (1998) Ciguatera toxins: Origin, transfer throughfood chain and toxicity to humans. In Harmful Algae (Reguera, B. B. J.,and Fernandez, M. T. W., Eds.) pp 39−43, Santiago de Compostela:Xunta de Galicia and Intergovernmental Oceanographic Commission,Santiago de Compostela.(7) Otero, P., Perez, S., Alfonso, A., Vale, C., Rodriguez, P., Gouveia,N. N., Gouveia, N., Delgado, J., Vale, P., Hirama, M., Ishihara, Y.,Molgo, J., and Botana, L. M. (2010) First toxin profile of ciguateric fishin Madeira Arquipelago (Europe). Anal. Chem. 82, 6032−6039.(8) Catterall, W. A., Cestele, S., Yarov-Yarovoy, V., Yu, F. H., Konoki,K., and Scheuer, T. (2007) Voltage-gated ion channels and gatingmodifier toxins. Toxicon 49, 124−141.(9) Perez, S., Vale, C., Alonso, E., Rodriguez, P., Otero, P., Alfonso,A., Vale, P., Hirama, M., Vieytes, M. R., and Botana, L. M. (2010) Acomparative study of the effect of ciguatoxins on voltage-dependentNa+ and K+ channels in cerebellar neurons. Chem. Res. Toxicol. 24,587−596.(10) Fuwa, H., Sasaki, M., Satake, M., and Tachibana, K. (2002)Total synthesis of gambierol. Org. Lett. 4, 2981−2984.(11) Fuwa, H., Kainuma, N., Tachibana, K., and Sasaki, M. (2002)Total synthesis of (−)-gambierol. J. Am. Chem. Soc. 124, 14983−14992.(12) Ghiaroni, V., Sasaki, M., Fuwa, H., Rossini, G. P., Scalera, G.,Yasumoto, T., Pietra, P., and Bigiani, A. (2005) Inhibition of voltage-gated potassium currents by gambierol in mouse taste cells. Toxicol.Sci. 85, 657−665.(13) Cuypers, E., Abdel-Mottaleb, Y., Kopljar, I., Rainier, J. D., Raes,A. L., Snyders, D. J., and Tytgat, J. (2008) Gambierol, a toxin producedby the dinoflagellate Gambierdiscus toxicus, is a potent blocker ofvoltage-gated potassium channels. Toxicon 51, 974−983.(14) LePage, K. T., Rainier, J. D., Johnson, H. W., Baden, D. G., andMurray, T. F. (2007) Gambierol acts as a functional antagonist ofneurotoxin site 5 on voltage-gated sodium channels in cerebellargranule neurons. J. Pharmacol. Exp. Ther. 323, 174−179.(15) Schlumberger, S., Ouanounou, G., Girard, E., Sasaki, M., Fuwa,H., Louzao, M. C., Botana, L. M., Benoit, E., and Molgo, J. (2010) Themarine polyether gambierol enhances muscle contraction and blocks atransient K+ current in skeletal muscle cells. Toxicon 56, 785−791.(16) Alonso, E., Fuwa, H., Vale, C., Suga, Y., Goto, T., Konno, Y.,Sasaki, M., Laferla, F. M., Vieytes, M. R., Gimenez-Llort, L., andBotana, L. M. (2012) Design and Synthesis of Skeletal Analogues ofGambierol: Attenuation of Amyloid-beta and Tau Pathology withVoltage-Gated Potassium Channel and N-Methyl-d-aspartate ReceptorImplications. J. Am. Chem. Soc. 134, 7467−7479.(17) Alonso, E., Vale, C., Sasaki, M., Fuwa, H., Konno, Y., Perez, S.,Vieytes, M. R., and Botana, L. M. (2010) Calcium oscillations inducedby gambierol in cerebellar granule cells. J. Cell. Biochem. 110, 497−508.(18) Contestabile, A. (2002) Cerebellar granule cells as a model tostudy mechanisms of neuronal apoptosis or survival in vivo and invitro. Cerebellum 1, 41−55.(19) Schousboe, A., Drejer, J., Hansen, G. H., and Meier, E. (1985)Cultured neurons as model systems for biochemical and pharmaco-logical studies on receptors for neurotransmitter amino acids. Dev.Neurosci. 7, 252−262.(20) Vale, C., Nicolaou, K. C., Frederick, M. O., Gomez-Limia, B.,Alfonso, A., Vieytes, M. R., and Botana, L. M. (2007) Effects ofazaspiracid-1, a potent cytotoxic agent, on primary neuronal cultures. Astructure-activity relationship study. J. Med. Chem. 50, 356−363.(21) Vale, C., and Sanes, D. H. (2000) Afferent regulation ofinhibitory synaptic transmission in the developing auditory midbrain. J.Neurosci. 20, 1912−1921.(22) Vale, C., and Sanes, D. H. (2002) The effect of bilateral deafnesson excitatory and inhibitory synaptic strength in the inferior colliculus.Eur. J. Neurosci. 16, 2394−2404.(23) Afshari, F. S., Ptak, K., Khaliq, Z. M., Grieco, T. M., Slater, N. T.,McCrimmon, D. R., and Raman, I. M. (2004) Resurgent Na currentsin four classes of neurons of the cerebellum. J. Neurophysiol. 92, 2831−2843.

Chemical Research in Toxicology Article

dx.doi.org/10.1021/tx300242m | Chem. Res. Toxicol. 2012, 25, 1929−19371936

(24) Osorio, N., Alcaraz, G., Padilla, F., Couraud, F., Delmas, P., andCrest, M. (2005) Differential targeting and functional specialization ofsodium channels in cultured cerebellar granule cells. J. Physiol. 569,801−816.(25) Hu, C. L., Liu, Z., Zeng, X. M., Liu, Z. Q., Chen, X. H., Zhang,Z. H., and Mei, Y. A. (2006) 4-aminopyridine, a Kv channel antagonist,prevents apoptosis of rat cerebellar granule neurons. Neuropharmacol-ogy 51, 737−746.(26) Grynkiewicz, G., Poenie, M., and Tsien, R. Y. (1985) A newgeneration of Ca2+ indicators with greatly improved fluorescenceproperties. J. Biol. Chem. 260, 3440−3450.(27) Vale, C., Alfonso, A., Sunol, C., Vieytes, M. R., and Botana, L. M.(2006) Modulation of calcium entry and glutamate release in culturedcerebellar granule cells by palytoxin. J. Neurosci. Res. 83, 1393−1406.(28) Vale, C., Fonfria, E., Bujons, J., Messeguer, A., Rodriguez-Farre,E., and Sunol, C. (2003) The organochlorine pesticides gamma-hexachlorocyclohexane (lindane), alpha-endosulfan and sieldrin differ-entially interact with GABAA and glycine-gated chloride channels inprimary cultures of cerebellar granule cells. Neuroscience 117, 397−403.(29) Varming, T., Drejer, J., Frandsen, A., and Schousboe, A. (1996)characterization of a chemical anoxia model in cerebellar granuleneurons using sodium azide: Protection by nifedipine and MK-801. J.Neurosci. Res. 44, 40−46.(30) Schaller, K. L., and Caldwell, J. H. (2003) Expression anddistribution of voltage-gated sodium channels in the cerebellum.Cerebellum 2, 2−9.(31) Blaine, J. T., and Ribera, A. B. (2001) Kv2 channels formdelayed-rectifier potassium channels in situ. J. Neurosci. 21, 1473−1480.(32) Mathie, A., Clarke, C. E., Ranatunga, K. M., and Veale, E. L.(2003) What are the roles of the many different types of potassiumchannel expressed in cerebellar granule cells? Cerebellum 2, 11−25.(33) Kopljar, I., Labro, A. J., Cuypers, E., Johnson, H. W., Rainier, J.D., Tytgat, J., and Snyders, D. (2009) A polyether biotoxin binding siteon the lipid-exposed face of the pore domain of Kv channels revealedby the marine toxin gambierol. Proc. Natl. Acad. Sci. U.S.A. 106, 9896−9901.(34) Louzao, M. C., Cagide, E., Vieytes, M. R., Sasaki, M., Fuwa, H.,Yasumoto, T., and Botana, L. M. (2006) The sodium channel ofhuman excitable cells is a target for gambierol. Cell. Physiol. Biochem.17, 257−268.(35) Cagide, E., Louzao, M. C., Espina, B., Ares, I. R., Vieytes, M. R.,Sasaki, M., Fuwa, H., Tsukano, C., Konno, Y., Yotsu-Yamashita, M.,Paquette, L. A., Yasumoto, T., and Botana, L. M. (2011) ComparativeCytotoxicity of Gambierol versus Other Marine Neurotoxins. Chem.Res. Toxicol. 24, 835−842.(36) Gawley, R. E., Rein, K. S., Jeglitsch, G., Adams, D. J.,Theodorakis, E. A., Tiebes, J., Nicolaou, K. C., and Baden, D. G.(1995) The relationship of brevetoxin 'length' and A-ring functionalityto binding and activity in neuronal sodium channels. Chem. Biol. 2,533−541.(37) Ghiaroni, V., Fuwa, H., Inoue, M., Mizyazaki, K., Hirama, M.,Yasumoto, T., Rossini, G. P., Scalera, G., and Bigiani, A. (2006) Effectof ciguatoxin 3C on voltage gated Na+ and K+ currents in mouse tastecells. Chem. Senses 31, 673−680.(38) Hu, C. L., Zeng, X. M., Zhou, M. H., Shi, Y. T., Cao, H., andMei, Y. A. (2008) Kv 1.1 is associated with neuronal apoptosis andmodulated by protein kinase C in the rat cerebellar granule cell. J.Neurochem. 106, 1125−1137.

Chemical Research in Toxicology Article

dx.doi.org/10.1021/tx300242m | Chem. Res. Toxicol. 2012, 25, 1929−19371937