Valproate acidifies hippocampal CA3-neurons—a novel mode of action

European Neuropsychopharmacology 12 (2002) 279–285 www.elsevier.com / locate / euroneuro

Valproate acidifies hippocampal CA3-neurons—a novel mode of action a, b c a c Udo Bonnet *, Dieter Bingmann , Tobias Leniger , Norbert Scherbaum , Guido Widman , c b Andreas Hufnagel , Martin Wiemann a

¨ ¨ Psychiatrie und Psychotherapie, Virchowstrasse 174, D-45147 Essen, Germany f ur Rheinische Kliniken, Universitatsklinik b ¨ Physiologie der Universitat-GH ¨ Institut f ur Essen, Hufelandstrasse 55, D-45122 Essen, Germany c ¨ ¨ Neurologie der Universitat-GH ¨ Universitatsklinik und Poliklinik f ur Essen, Hufelandstrasse 55, D-45122 Essen, Germany Received 23 January 2001; accepted 14 February 2002

Abstract Various hypotheses try to explain the anticonvulsive and mood stabilizing effects of valproate. Among them, amplification of GABAergic inhibition and reduction of membrane excitability is favored. Here we show that superfusion with 0.1–1 mM valproate induced a moderate intracellular acidification of BCECF-AM-loaded CA3-neurons (hippocampal slices, guinea pig) which was measured as the difference between intracellular pH before (baseline pH i ) and during valproate treatment (DpH i ). In two groups of neurons treated with 1 mM and 0.1–0.5 mM, DpH i values amounted to 0.2060.10 and 0.1060.04 (DpH i 6S.D.), respectively, suggesting a dependence on the used valproate-concentration. DpH i did not correlate with the baseline pH i . Furthermore, the acidification seems to be independent from an activation of postsynaptic GABA-A receptors, as it was not influenced by 0.1 mM picrotoxin. Since our previous studies clearly demonstrated a reduction of membrane excitability during moderate intracellular acidification, we suggest that the valproate-mediated intracellular acidification may substantially contribute to its anticonvulsive and mood stabilizing properties.  2002 Elsevier Science B.V. All rights reserved. Keywords: Valproate; Intracellular pH; Epilepsy; Bipolar affective disorders

1. Introduction Valproate is effective in the treatment of bipolar affective disorders and several types of epilepsies (Post and Siberstein, 1994; Keck et al., 1998; Loscher, 1999; Grunze et al., 1999). Various hypotheses have been put forward to explain its beneficial mode of action in these disorders, among which enhancement of central GABAergic neurotransmission as well as direct effects on excitable membranes via enhanced Na 1 -channel inactivation and inhibition of Ca 21 -channels are favored (Chapman et al., 1982; ¨ Preisendorfer et al., 1987; Olpe et al., 1988; Kito et al., 1994; Tunnicliff, 1999; Loscher, 1999). We have recently

*Corresponding author. Tel.: 149-201-7227-150; fax: 149-201-7227303. E-mail address: [email protected] (U. Bonnet).

described that a moderate decrease of intracellular pH (pH i ) of less than 0.25 pH units exerts inhibitory and anticonvulsive effects on hippocampal pyramidal cells (Bonnet et al., 1998a, 2000a; Bonnet and Wiemann, 1999a). Moreover, 31 P-MRS measurements revealed that patients suffering from temporal lobe epilepsy or affective disorders display abnormal pH i values of cortical cells (Kato et al., 1998; van der Grond et al., 1998). Since these patients often benefit from valproate treatment, we were interested in the effect of the weak acid valproate on pH i of cortical neurons. We chose hippocampal CA3-neurons in slices because this preparation is well established to investigate the action of anticonvulsant drugs (Preisen¨ dorfer et al., 1987; Sagratella, 1998; Bonnet and Bingmann, 1998b) and is also used to study changes of the neuronal pH i (Bonnet et al., 1998a; Bonnet and Wiemann, 1999a). Preliminary parts of this work have appeared in abstract form (Bonnet and Wiemann, 1999b; Bonnet et al., 2001).

0924-977X / 02 / $ – see front matter  2002 Elsevier Science B.V. All rights reserved. PII: S0924-977X( 02 )00023-8

280

U. Bonnet et al. / European Neuropsychopharmacology 12 (2002) 279 – 285

2. Materials and methods Transverse hippocampal slices (200–400 mm thick) were prepared from brains obtained from ether anesthetized adult guinea pigs (300–400 g). Slices were preincubated for 2 h in 28 8C warm saline equilibrated with 5% CO 2 in O 2 . This saline contained in mM: NaCl 124, KCl 3, CaCl 2 0.75, MgSO 4 1.3, KH 2 PO 4 1.25, NaHCO 3 26 and glucose 10. After pre-incubation, slices were transferred to a perspex recording chamber (volume, 4 ml) which was mounted on an inverted microscope. In this chamber submerged slices were continuously superfused at a rate of 4.5 ml / min by 32 8C warm saline. The composition of this control solution was the same as during the pre-incubation period except for the calcium concentration which was raised to 1.75 mM. Valproate sodium (0.07–2 mM, pKa 54.6, Desitin) and picrotoxin (0.1 mM, Sigma) were added to the saline. The pH of all solutions was 7.35–7.4. To analyze pH i changes, hippocampal slices were loaded with 0.5–1 mM 29,7-bis(2-carboxyethyl)-5(6)-carboxyfluorescein-acetoxymetyl ester (BCECF-AM, Molecular Probes) for 3–5 min in the pre-incubation saline. After at least a 30 min wash superficially located neurons of the stratum pyramidale of the CA3 region were identified by stained apical dendrites (Bonnet et al., 1998a; Bonnet and Wiemann, 1999a) and examined under 403 or 603 water immersion objectives. Slices were intermittently excited at 440 and 490 nm using a computer operated filter wheel (Sutter Instruments) equipped with a 100 W halogen lamp. Fluorescence images were captured (0.05 Hz) by an intensified CCD camera (PTI, Surbiton, UK). Background correction and ratio imaging was performed with a CARAT system (Dr. O. Ahrens, Bargteheide, Germany). At the end of an experiment the ratio 440 / 490 (R) was calibrated by a standard curve which was obtained by the in vitro calibration method adapted to an upright microscope (Bonnet et al., 1998a; Bonnet and Wiemann, 1999a). Background light or autofluorescence of the slice were included into the standard curve and, thus, eliminated from the measurement. Neurons with a high loss of fluorescent dye within the first 15 min of observation were excluded from further measurements (cf. Bevensee et al., 1995). Neither valproate nor picrotoxin contributed to the fluorescence signal when excited at 440 or 490 nm, or changed the pH of the superfusate. Spontaneous pH i deflections (observed to be in the range of 60.02 pH units) and noise were eliminated from the curves by calculating sliding averages of 3–6 values. All pH i values were taken from single neuronal somata. Changes of the baseline pH i (averaged from a 5 min lasting period immediately before drug application) were regarded as valproate-mediated (‘responders’) if they: (i) exceeded 0.03 pH i units, (ii) occurred upon drug application, and (iii) were at least partly reversible after washout. The respective DpH i values were measured when a more acidic steady state pH i had

been reached (c.f. Fig. 1A2, A3) or, if no steady state occurred, after 20 min of drug administration (cf. Fig. 1B). In cases of missing valproate effect on pH i , drug-application was extended to 30 min.

2.1. Statistics Values were expressed as mean6standard deviation (S.D.). In Section 3.1, the neurons not responsive to valproate were excluded from averaging DpH i values. A t-test for impaired samples was employed to test for

Fig. 1. Effect of valproate on pH i of BCECF-AM loaded hippocampal CA3-neurons. Application of the respective valproate concentration is indicated by bars (A1–A3). Different concentrations of valproate consecutively applied to one neuron evoked reversible shifts of the neuronal pH i . Time between the measurements A1–3 was 1–5 min. While with 0.07 mM (A1) a pH i response was missing, 0.15 mM (A2) and 0.3 mM (A3) shifted pH i to more acidic steady states. (B) In this neuron pH i continued to fall during application of 1 mM valproate. No stable steady state was reached within the application period. Acidification was partially reversible upon washout of the drug.

U. Bonnet et al. / European Neuropsychopharmacology 12 (2002) 279 – 285

281

significant differences (P#0.01). Linear correlation-analysis and calculation of R 2 values was carried out to test for possible interdependence of baseline pH i and DpH i .

3. Results Experiments were carried out on 53 BCECF-AM loaded CA3 neurons in 46 slices from 12 animals. The mean baseline pH i of CA3-neurons was 7.1360.17 during superfusion with control solution (n553). We tested concentrations between 0.07 and 2 mM valproate while changes of pH i were monitored simultaneously.

3.1. pHi responses during valproate treatment In a first step five neurons (n55) were subjected to increasing concentrations of valproate with intermittent washing to analyze pH i responses and to find the most effective concentration of valproate. Fig. 1A shows that increasing concentrations led to an augmented acidification in the particular neuron. Acidification was sustained and started to occur within the first 10 min after valproate application ($0.15 mM). Within the next 10–20 min a more acidic steady state was reached (Fig. 1A2,3). In some cases, however, no new steady state was reached within 30 min and the pH i decreased until valproate was removed from the bath (Fig. 1B). Valproate induced acidification was completely reversible within 20 min when small valproate concentrations had been used (Fig. 1A2). At higher concentrations however, pH i recovery was often incomplete within this period (Fig. 1A3, 1B). In no case we observed a counter-regulation of the acidified pH i neither an alkalotic overshoot upon removal of valproate. In a next step we exposed a larger number of neurons to 1 mM valproate for 30 min (n535) and measured the DpH i obtained upon valproate treatment as described in the Materials and methods section. Within this collective 30 neurons responded with an acidification while 5 neurons did not change their baseline pH i values (DpH i ,0.03 pH units, cf. Materials and methods section). Among the 30 responding neurons 21 reached a more acidic steady state whereas in nine neurons pH i continued to slightly decline; we nevertheless included these nine neurons into evaluations of responders (DpH i taken also after 20 min) because the effect on pH i was at least partially reversible when valproate was removed. Fig. 2A shows that the 30 responding neurons of the ‘1 mM group’ were acidified by 0.2060.10 pH units. Due to this considerable variability we did not perform dose–response curves. Instead, we compared that mean DpH i value to one obtained with 0.07–0.5 mM valproate (‘#0.5 mM-group’, n521). In contrast to the former group 12 neurons were not affected by #0.5 mM valproate and only nine neurons were slightly acidified by 0.1060.04 pH units, which was significantly different from the mean value of the ‘1 mM group’

Fig. 2. Effects of baseline pH i and valproate concentrations on pH i responses. DpH i values of the ‘1 mM-group’ (A) and of the ‘#0.5 mM-group’ (B) were plotted against the baseline pH i . The gray area in both panels marks the ,0.03 pH-unit zone containing missing responses (c.f. Materials and methods section). (A) The dotted line shows the calculated linear regression curve of 35 measurements. The R 2 value indicates missing correlation between baseline pH i and DpH i induced by 1 mM valproate. The mean value calculates to 0.2060.10 if missing pH i responses (5 of 35, gray area) were omitted. (B) In the ‘#0.5 mM-group’ experiments with 0.07–0.5 mM valproate were gathered. Concentrations used for responses are given above each data point. Non-responses (12 of 21, gray area) were much more frequent as with 1 mM and received (in mM) 0.07 (n52), 0.1 (n52), 0.2 (n53), 0.35 (n52) and 0.5 (n53).

(P,0.001, t-test). The lowest effective concentration of valproate found in the ‘#0.5 mM-group’ was 0.1 mM (2 out of four cases, DpH i : 0.13, cf. Fig. 2B). These data point to a concentration dependence of the valproatemediated acidification which was reflected by both the different degree of acidification (DpH i ) and the different numbers of non-responding cases. In Fig. 2A DpH i values were also plotted against the baseline pH i of the respective neuron. This was done to test whether the valproate-mediated acidification was, e.g., increased in more acidic neurons. There was, however, no linear correlation (R 2 50.0121, Fig. 2A, dotted line) between baseline pH i and the valproate induced DpH i values. Thus, more acidic or more basic neurons appear equally prone to the acidifying action of valproate.

282

U. Bonnet et al. / European Neuropsychopharmacology 12 (2002) 279 – 285

3.2. Neurons not responding to valproate As outlined above, valproate failed to induce pH i responses in some neurons (Fig. 2, gray areas). To test whether this was just a matter of valproate concentration we selected four neurons not responding to 1 mM valproate, and two neurons not responding to 0.5 mM valproate, and equally exposed them to 2 mM valproate. Also under this condition these neurons failed to acidify. These findings let us assume that there exists a subset of hippocampal neurons which is rather resistant to the acidifying action of valproate.

3.3. Missing interference with GABA-mediated acidification We furthermore investigated the possible contribution of GABA-A receptors to the acidification because: (i) an activation of postsynaptic GABA receptors is known to acidify neurons (Pasternack et al., 1993; Kaila, 1994), and (ii) valproate is known to increase synaptic GABA-levels ¨ (Loscher, 1999; Tunnicliff, 1999). We found that 0.1 mM of the GABA-A antagonist picrotoxin which is sufficient to clearly diminish GABAergic synaptic events of hippocampal pyramidal cells (Thalmann et al., 1981; Hablitz, 1984;

Fig. 3. Missing effect of picrotoxin on baseline pH i of a BCECF-AM loaded CA3-neuron and on valproate-mediated pH i response. Application of the drugs is indicated by bars. (A) Baseline pH i was not affected within 30 min of picrotoxin treatment. (B) pH i response upon valproate application was also not influenced by picrotoxin. After stopping the application of the drugs, acidification was at least partially reversible (not illustrated). Both curves were from the same neuron (3 min time between both measurements). The figure is representative for four experiments in different neurons.

¨ Muller and Misgeld, 1991) did not induce pH i responses per se (n56, Fig. 3A). Also we observed no interference with valproate-mediated pH i responses (n54, Fig. 3B). From this we suggest that an activation of postsynaptic GABA-A receptors seems to be not involved in the acidification mediated by valproate.

4. Discussion The main finding of our study was that valproate acidifies the majority of hippocampal CA3-neurons and that this action seems to be not attributable to valproate’s known effect on GABAergic neurotransmission. To our knowledge this is the first report of an intracellular acidification of neurons by valproate, although there was some speculation about teratogenic effects mediated by a valproate-induced drop in pH i of embryonic cells (Scott et al., 1997). In our experiments on CA3-neurons of guinea pig hippocampal slices intracellular acidification started at 0.1 mM valproate in the superfusate (Fig. 2B) which is in the upper range of assumed therapeutical levels in human brain (Cotariu et al., 1990; Olpe et al., 1988). However, it is generally accepted that rodents require larger concentrations of valproate to limit epileptic activity (Chapman et al., 1982) and that especially in slice preparations, the anticonvulsive concentration of valproate ranged between ¨ one and several millimoles (Preisendorfer et al., 1987; Sokolova et al., 1998). Similar observations were reported using large neurons from buccal ganglia of Helix pomatia allowing even an intracellular application of valproate (Altrup et al., 1992; Walden et al., 1993). We found that administration of 1 mM valproate decreased the neuronal pH i by 0.2060.10 pH units, which is a sufficient intracellular acidification to diminish epileptiform activity of CA3-neurons as we have recently shown by employing several methods of intracellular acidification (Bonnet et al., 1998a, 2000a,b; Bonnet and Wiemann, 1999a). Taking into account that the anticonvulsant valproate action could result from its potency to inhibit transmembrane Na 1 - and Ca 21 -fluxes (Chapman et al., 1982; Cotariu et al., 1990; Kito et al., 1994; Mcdonald and ¨ Kelly, 1995; Tunnicliff, 1999; Loscher, 1999) it appears likely that these effects are secondary to an increase in intracellular free protons affecting certain pH i -sensitive Na 1 - and Ca 21 -channels (Hille, 1968; Takahashi and Copenhagen, 1996; Tombaugh, 1998; Tombaugh and Somjen, 1998; Kiss and Korn, 1999). Intracellular acidification also has an impact on energy-metabolism (Nicholls and Ward, 2000) and transmitter release (Trudeau et al., 1999), both of which possibly contribute to the pathophysiology of affective disorders (Post and Siberstein, 1994; Stoll and Severus, 1996; Kato et al., 1998) and several types of epilepsies (Dichter, 1989; Mcdonald and Kelly, 1995; van der Grond et al., 1998). It is thinkable that valproate can restore such disturbances to non-

U. Bonnet et al. / European Neuropsychopharmacology 12 (2002) 279 – 285

pathological conditions by a proton-mediated decrease in membrane excitability. In addition, a moderate pH i decrease could lower aberrant intracellular signaling (Vignes et al., 1996; Stoll and Severus, 1996), e.g., in critical regions of the limbic forebrain where hyperexcitability may occur in mania (Kato et al., 1998) or in the temporal lobe where epileptic seizures can originate (Koehling et al., 1998). These speculations are further encouraged by the fact that antimanic and anticonvulsive drugs, such as acetazolamide, influence the pH i or its regulation (Bickler et al., 1988; Brooks and Bachelard, 1992; Hayes, 1994; Leniger et al., 2000; Sappey-Marinier et al., 2000). Also a reduction of pH i of cortical neurons was demonstrated for MAO-A inhibitors such as moclobemide and clorgyline, which shared anticonvulsive properties in pre-clinical investigations (Goldina and Valdman, 1994; Bonnet et al., 2000c) and at least clorgyline appeared beneficial in rapidcycling of bipolar affective disorder (Potter et al., 1982; Piletz and Halaris, 1995). Furthermore, some organic calcium-antagonists such as verapamil and nimodipin which are anecdotally reported to be beneficial in the treatment of rapid-cyclers (Keck et al., 1998) and which increased the seizure-threshold in preclinical investigations (Bingmann and Speckmann, 1989; Straub et al., 1992) reduced pH i of CA3-neurons (Bonnet et al., 2001). It should be outlined that intracellular protons have properties of a second-messenger (Takahashi and Copenhagen, 1996; Vignes et al., 1996). Therefore, a pH i modulation due to valproate could contribute to delayed adaptive neuronal changes underlying especially its effects on mood—taking several days to weeks as a rule. In contrast, anticonvulsive effects appear after a shorter time taking into account its relatively fast action in status epilepticus. This may result from more direct effects of valproate-mediated acidification on excitibility of membranes (e.g., due to a proton-mediated inhibition of voltage-gated and receptor-mediated Na 1 - and Ca 21 -channels, see above). In this context, valproate-mediated acidification started in vitro within the first 10 min of application which may be attributed to the fact that valproate is a weak acid (pKa 54.6). In principle, weak acids should be able to shuttle protons through membranes as is known for propionate (Roos and Boron, 1981). In our hands, propionate only at far higher concentrations of $5 mM induced a similar intracellular acidosis also sufficient to reduce excitability of hippocampal CA3-neurons (Bonnet et al., 2000a). In contrast to propionate treatment, the valproateinduced acidosis was without counter regulation. It is therefore reasonable to assume additional effects of valproate on pH i regulation. These effects could consist in a perturbed cellular acid extrusion or enhanced acid loading (Orlowski and Grinstein, 1997; Bevensee and Boron, 1998; Bonnet et al., 2000b). However, we found no correlation of the baseline pH i and the DpH i induced by valproate such that more acid and more basic neurons appeared equally prone to the influence of valproate on

283

pH i . Nevertheless, there are many different acid extruders in brain neurons (Orlowski and Grinstein, 1997; Bevensee and Boron, 1998; Bonnet et al., 2000b) and nearly all hitherto known subtypes of Na 1 / H 1 exchangers are expressed in the hippocampus formation (Ma and Haddad, 1997; Baird et al., 1999). At present it remains to be determined whether valproate can selectively inhibit such membrane transporters. As lactate may be a source of intracellular protons for neurons in nervous tissue (Pellerin et al., 1998) and valproate can interfere with monocarboxylate transport (Bolanos and Medina, 1994; Tamai et al., 1995; Mac et al., 2000) a disturbed lactate flux may contribute to the valproate induced acidification. Moreover, the large disparity of the DpH i values could be due to differences in expression of monocarboxylate transporters, cell metabolism, and / or oxygen supply of individual neurons. It is well known that valproate can enhance GABAergic ¨ neurotransmission (Preisendorfer et al., 1987; Olpe et al., ¨ 1988; Loscher, 1999; Tunnicliff, 1999) and that an activation of postsynaptic GABA-A receptors acidifies cells (Pasternack et al., 1993; Kaila, 1994). This simple mechanism, however, may not account for the acidification observed, because picrotoxin—an antagonist at GABA-A receptors—failed to affect pH i in our study and, moreover, did not diminish the extent of the acidification mediated by valproate. This observation possibly reflects experiences from valproate studies with electroshock-induced convulsions in vivo (Kerwin et al., 1980; Kerwin and Taberner, 1981). These experiments clearly showed that the anticonvulsive action of valproate was fast (within 10 min after intraperitoneal application) whereas GABA-levels in the brain increased not before 20 min after valproate application. The rapid fall in neuronal pH i upon valproate application observed here would provide a reasonable explanation for rapid anticonvulsive effects in vivo, which may also contribute to limitation of status epilepticus in humans (Vankataraman and Wheless, 1999). To summarize, we suggest that anticonvulsive concentrations of valproate used in humans are sufficient to moderately acidify at least a subset of cortical neurons. This action of valproate may contribute to its efficacy in the treatment of affective and convulsive disorders.

References Altrup, U., Gerlach, G., Reith, H., Said, M.N., Speckmann, E.-J., 1992. Effects of valproate in a model nervous system (buccal ganglia of Helix pomatia): I antiepileptic actions. Epilepsia 33, 743–752. Baird, N.R., Orlowski, J., Szabo, E.Z., Zaun, H.C., Schultheis, P.J., Menon, A.G., Shull, G.E., 1999. Molecular cloning, genomic organization, and functional expression of Na 1 / H 1 exchanger isoform 5 (NHE5) from human brain. J. Biol. Chem. 12, 4377–4382. Bevensee, M.O., Schwiening, C.J., Boron, W.F., 1995. Use of BCECF and propidium iodide to assess membrane integrity of acutely isolated CA1 neurons from rat hippocampus. J. Neurosci. Methods 58, 61–75.

284

U. Bonnet et al. / European Neuropsychopharmacology 12 (2002) 279 – 285

Bevensee, M.O., Boron, W.F., 1998. pH regulation in mammalian neurones. In: Kaila, K., Ransom, B. (Eds.), pH and Brain Function. Wiley Liss, New York, pp. 373–394. Bickler, P.E., Litt, I., Banville, D.L., Severinghaus, J.W., 1988. Effects of acetazolamide on cerebral acid–base balance. J. Appl. Physiol. 65, 422–427. Bingmann, D., Speckmann, E.-J., 1989. Specific suppression of pentylenetetrazol-induced epileptiform discharges in CA3 neurons (hippocampal slice, guinea pig) by the organic calcium antagonists flunarizine and verapamil. Exp. Brain Res. 74, 239–248. Bolanos, J.P., Medina, J.M., 1994. Inhibition of neonatal brain fuel utilization by valproate and E-delta 2-valproate is not a consequence of the stimulation of the gamma-amino acid shunt. Life Sci. 55, PL397– 402. Bonnet, U., Wiemann, M., Bingmann, D., 1998a. CO 2 / HCO 3 -withdrawal from the bath medium of hippocampal slices: biphasic effect on intracellular pH and bioelectric activity of CA3-neurons. Brain Res. 796, 161–170. Bonnet, U., Bingmann, D., 1998b. Missing action of carbamazepine on postsynaptic GABA-responses of hippocampal CA3-neurons (slice, guinea pig). Eur. Neuropsychopharmacol. 8, 353–356. Bonnet, U., Wiemann, M., 1999a. Ammonium prepulse: effects on intracellular pH and bioelectric activity of CA3-neurones in guinea pig hippocampal slices. Brain Res. 840, 16–22. Bonnet, U., Wiemann, M., 1999b. Intracellular free protons: relevance for neuropsychopharmacology? Pharmacopsychiatry 32, 172. Bonnet, U., Bingmann, D., Wiemann, M., 2000a. Intracellular pH modulates spontaneous and epileptiform bioelectric activity of hippocampal CA3-neurones. Eur. Neuropsychopharmacol. 10, 97–103. Bonnet, U., Leniger, T., Wiemann, M., 2000b. Alteration of intracellular pH and activity of CA3-pyramidal cells in guinea pig hippocampal slices by inhibition of transmembrane acid extrusion. Brain Res. 872, 116–124. Bonnet, U., Leniger, T., Wiemann, M., 2000c. Moclobemide reduces intracellular pH and neuronal activity of CA3 neurones in guinea pig hippocampal slices—implication for its neuroprotective properties. Neuropharmacology 39, 2067–2074. Bonnet, U., Leniger, T., Wiemann, M., 2001. Neuropsychopharmaca and intracellular pH. Pharmacopsychiatry 34, 165. Brooks, K.J., Bachelard, H.S., 1992. The regulation of intracellular pH studied by 31P- and 1H-NMR spectroscopy in superfused guinea pig cortex slices. Neurochem. Int. 21, 375–379. Chapman, A., Keane, P.E., Meldrum, B.S., Simiand, J., Vernieres, J.C., 1982. Mechanism of anticonvulsant action of valproate. Prog. Neurobiol. 19, 315–359. Cotariu, D., Zaidman, J.L., Evans, S., 1990. Neurophysiological and biochemical changes evoked by valproic acid in the central nervous system. Prog. Neurobiol. 34, 343–354. Dichter, M.A., 1989. Cellular mechanisms of epilepsy and potential new treatment strategies. Epilepsia 30 (Suppl. 1), S3–S12. Goldina, O.A., Valdman, A.V., 1994. Anticonvulsive effect of some psychotropic drugs and benzamide derivates with regard to antimonoamine oxidase activity. Eur. Neuropsychopharmacol. 4, 446– 447. Grunze, H., Erfurth, A., Amann, B., Giupponi, G., Kammerer, C., Walden, J., 1999. Intravenous valproate loading in acutely manic and depressed bipolar I patients. J. Clin. Psychopharmacol. 19, 303–309. Hablitz, J.J., 1984. Picrotoxin-induced epileptiform activity in hippocampus: role of endogenous versus synaptic factors. J. Neurophysiol. 51, 1011–1027. Hayes, S.G., 1994. Acetazolamide in bipolar affective disorders. Ann. Clin. Psychiatry 6, 91–98. Hille, B., 1968. Charges and potentials at the nerve surface. Divalent ions and pH. J. Gen. Physiol. 51, 221–236. Kaila, K., 1994. Ionic basis of GABAA receptor channel function in the nervous system. Prog. Neurobiol. 42, 489–537. Kato, T., Inubushi, T., Kato, N., 1998. Magnetic resonance spectroscopy in affective disorders. J. Neuropsychiatry Clin. Neurosci. 10, 133–147.

Keck, P.E., McElroy, S.L., Strakowski, S.M., 1998. Anticonvulsants and antipsychotics in the treatment of bipolar disorder. J. Clin. Psychiatry 59 (Suppl. 6), 74–81. Kerwin, R.W., Olpe, H.R., Schmutz, M., 1980. The effect of sodium-ndipropyl acetate on g-aminobutyric acid-dependent inhibition in the rat cortex and substantia nigra in relation to its anticonvulsant activity. Br. J. Pharm. 71, 545–551. Kerwin, R.W., Taberner, P.V., 1981. The mechanism of action of sodium valproate. Gen. Pharmacol. 12, 71–75. Kiss, L., Korn, S.J., 1999. Modulation of N-type Ca channels by intracellular pH in chick sensory neurons. J. Neurophysiol. 81, 1839– 1847. Kito, M., Maehara, M., Watanabe, K., 1994. Antiepileptic drugs—calcium current interaction in cultured human neuroblastoma cells. Seizure 3, 141–149. Koehling, R., Luecke, A., Straub, H., Speckmann, E.-J., Tuxhorn, I., Wolf, P., Pannek, H., Oppel, F., 1998. Spontaneous sharp waves in human neocortical slices excised from epileptic patients. Brain 121, 1073–1087. Leniger, T., Wiemann, M., Hufnagel, A., Bonnet, U., 2000. Carbonic anhydrase inhibitors sulthiame and acetazolamide: effects on neuronal activity and intracellular pH of CA3 neurones. Epilepsia 41 (Suppl. 7), 34. ¨ Loscher, W., 1999. Valproate: a reappraisal of its pharmacodynamic properties and mechanisms of action. Prog. Neurobiol. 58, 31–59. Ma, E., Haddad, G.G., 1997. Expression and localization of Na 1 / H 1 exchangers in rat central nervous system. Neuroscience 79, 591–603. Mac, M., Nehling, A., Nalecz, M.J., Nalecz, K.A., 2000. Transport of alpha-ketoisocaproate in the rat cerebral cortical neurons. Arch. Biochem. Biophys. 376, 347–353. Mcdonald, R.L., Kelly, K.M., 1995. Antiepileptic drug mechanism of action. Epilepsia 36 (Suppl. 2), S2–S12. ¨ Muller, W., Misgeld, U., 1991. Picrotoxin- and 4-aminopyridine-induced activity in hilar neurons in the guinea pig hippocampal slice. J. Neurophysiol. 65, 141–147. Nicholls, D.G., Ward, M.W., 2000. Mitochondrial membrane potential and neuronal glutamate excitotoxicity: mortality and millivolts. Trends Neurosci. 23, 166–174. Olpe, H.-R., Steinmann, M.W., Pozza, M.F., Brugger, F., Schmutz, M., 1988. Valproate enhances GABA-A mediated inhibition of locus coeruleus neurones in vitro. Naunyn-Schmiedebergs Arch. Pharmacol. 338, 655–657. Orlowski, J., Grinstein, S., 1997. Na 1 / H 1 exchangers of mammalian cells. J. Biol. Chem. 272, 22373–22376. Pasternack, M., Voipio, L., Kaila, K., 1993. Intracellular carbonic anhydrase activity and its role in GABA-induced acidosis in isolated rat hippocampal pyramidal neurones. Acta Physiol. Scand. 148, 229–231. Pellerin, L., Pellegri, G., Bittar, P.G., Charnay, Y., Bouras, C., Martin, J.L., Stella, N., Magistretti, P.J., 1998. Evidence supporting the existence of an activity-dependent astrocyteneuron lactate shuttle. Dev. Neurosci. 20, 291–299. Piletz, J.E., Halaris, A., 1995. Involvement of 1 1 -imidazoline receptors in mood disorders. Ann. N.Y. Acad. Sci. 763, 510–519. Post, R.M., Siberstein, S.D., 1994. Shared mechanisms in affective illness, epilepsy, and migraine. Neurology 44 (Suppl. 7), S37–S47. Potter, W.Z., Murphy, D.L., Wehr, T.A., Linnoila, M., Goodwin, F.K., 1982. Clorgyline. A new treatment for patients with refractory rapidcycling disorder. Arch. Gen. Psychiatry 39, 505–510. ¨ Preisendorfer, U., Zeise, M.L., Klee, M.R., 1987. Valproate enhances inhibitory postsynaptic potentials in hippocampal neurons in vitro. Brain Res. 435, 213–219. Roos, A., Boron, W.F., 1981. Intracellular pH. Physiol. Rev. 61, 296–434. Sagratella, S., 1998. Characterization of the in vitro antiepileptic activity of new and old anticonvulsant drugs. Gen. Pharmacol. 30, 153–160. Sappey-Marinier, D., Vighetto, A., Peyren, R., Broussolle, E., Bonmartin, A., 2000. Phosphorus and proton magnetic resonance spectroscopy in episodic ataxia type 2. Ann. Neurol. 47, 838–839.

U. Bonnet et al. / European Neuropsychopharmacology 12 (2002) 279 – 285 Scott, Jr. W.J., Schreiner, C.M., Nau, H., Vorhees, C.V., Beliles, R.P., Colvin, J., McCandless, D., 1997. Valproate-induced limb malformations in mice associated with reduction of intracellular pH. Reprod. Toxicol. 11, 483–493. ¨ Sokolova, S., Schmitz, D., Zhang, C.L., Loscher, W., Heinemann, U., 1998. Comparison of effects of valproate and trans-2-en-valproate on different forms of epileptiform activity in rat hippocampal and temporal cortex slices. Epilepsia 39, 251–258. Stoll, A.L., Severus, W.E., 1996. Mood stabilizers: shared mechanisms of action at postsynaptic signal-transduction and kindling processes. Harvard Rev. Psychiatry 4, 77–89. Straub, H., Danz, C., Speckmann, E.-J., 1992. Depressive effects of organic calcium-antagonists on bicuculline induced epileptic activity in hippocampal and neocortical neurons. In: Speckmann, E.-J., Gutnick, M.J. (Eds.), Epilepsy and Inhibition. Urban & Schwarzenberg, Munich, pp. 255–270. Takahashi, K.-I., Copenhagen, D.R., 1996. Modulation of neuronal function by intracellular pH. Neurosci. Res. 24, 109–116. Tamai, I., Takanaga, H., Maeda, H., Sai, Y., Ogihara, T., Higashida, H., Tsuji, A., 1995. Participation of a proton-cotransporter, MCT1, in the intestinal transport of monocarboxylic acids. Biochem. Biophys. Res. Commun. 214, 482–489. Thalmann, R.H., Peck, E.J., Ayala, G.F., 1981. Biphasic response of hippocampal pyramidal neurons to GABA. Neurosci. Lett. 21, 319– 324.

285

Tombaugh, G.C., Somjen, G.G., 1998. pH modulation of voltage-gated ion channels. In: Kaila, K., Ransom, B.R. (Eds.), pH and Brain Function. Wiley Liss, New York, pp. 395–416. Tombaugh, G.C., 1998. Intracellular pH buffering shapes activity-depen21 dent Ca dynamics in dendrites of CA1 interneurons. J. Neurophysiol. 80, 1702–1712. Trudeau, L.-E., Parpura, V., Haydon, P.G., 1999. Activation of neurotransmitter release in hippocampal nerve terminals during recovery from intracellular acidification. J. Neurophysiol. 81, 2627–2635. Tunnicliff, G., 1999. Actions of sodium valproate on the central nervous system. J. Physiol. Pharmacol. 50, 347–365. van der Grond, J., Gerson, J.R., Latxer, K.D., Hugg, J.W., Matson, G.B., Weiner, M.W., 1998. Regional distribution of interictal 31P metabolic changes in patients with temporal lobe epilepsy. Epilepsia 39, 527– 536. Vankataraman, V., Wheless, J.W., 1999. Safety of rapid intravenous infusion of valproate loading doses in epilepsy patients. Epilepsy Res. 35, 147–153. ` Vignes, M., Blanc, E., Guiramand, J., Gonzalez, E., Sassetti, I., Recasens, M., 1996. A modulation of glutamate-induced phosphoinositide breakdown by intracellular pH changes. Neuropharmacology 35, 1595– 1604. Walden, J., Altrup, U., Reith, H., Speckmann, E.-J., 1993. Effects of valproate on early and late potassium currents of single neurons. Eur. Neuropsychopharmacol. 3, 137–141.