- Email: [email protected]
Immediate and delayed effects of in vitro ischemia on glutamate efflux from guinea-pig cerebral cortex slices
Brain Research 751 Ž1997. 300–306
Research report
Immediate and delayed effects of in vitro ischemia on glutamate efflux from guinea-pig cerebral cortex slices G. Calo, ` S. Sbrenna, C. Bianchi, L. Beani
)
Institute of Pharmacology, UniÕersity of Ferrara, Via Fossato di Mortara 17-19, 44100 Ferrara, Italy Accepted 20 November 1996
Abstract Immediate and delayed effects of glucose deprivation, oxygen deprivation Žhypoxia. and both oxygen and glucose deprivation Žin vitro ischemia. on glutamate efflux from guinea pig cerebral cortex slices were studied. Immediate effects were evaluated by measuring changes of glutamate efflux during the metabolic insults. Delayed effects were evaluated by measuring the response of the tissue to a 50 mM KCl pulse applied 60 min after the metabolic insults. Deprivation of glucose in the medium did not induce either immediate or delayed effects, while hypoxic condition produced an immediate slight stimulation of glutamate efflux without any delayed effect. Conversely, in vitro ischemia produced both immediate and delayed effects on glutamate efflux. During in vitro ischemia glutamate efflux dramatically increased in a calcium-independent and tetrodotoxin-sensitive manner; this effect was potentiated by a low sodium containing medium. The blockade of the sodiumrpotassium ATPase exchanger by ouabain caused a glutamate outflow similar to that induced by in vitro ischemia. On the whole, these data demonstrate the central role played by the sodium electrochemical gradient and by the membrane glutamate uptake system in the glutamate overflow induced by in vitro ischemia. Moreover, in slices previously exposed to both oxygen and glucose deprivation the effect of KCl on glutamate efflux was potentiated. This in vitro ischemia-induced delayed potentiation of neurotransmitter efflux, until now unreported in the literature, was found to be selectively restricted to glutamatergic structures and to be mainly due to an enhancement of the exocytotic component of glutamate release. q 1997 Elsevier Science B.V. All rights reserved. Keywords: Glutamate; In vitro ischemia; Brain slice; NaqrKq ATPase ; Naqrglutamate transporter; Postischemic glutamate release potentiation
1. Introduction Both in vitro w19,28x and in vivo w4,19x studies have shown that extracellular glutamate concentration massively increases during cerebral ischemia. In spite of the numerous studies addressed to this field, the mechanismŽs. leading to accumulation of extracellular glutamate during cerebrovascular insults is not completely understood. In fact, contrasting results are reported in the literature regarding either the Ca2q-dependence w1,13,19,20x or the cellular origin of ischemia-induced glutamate outflow w17,25x. However, it is generally accepted that overactivation of glutamate receptors promotes postischemic neuronal damage w4x.
) Corresponding author. Fax: q39 Ž532. 291205; E-mail: [email protected]
Although the immediate rise of extracellular glutamate has an important role in postischemic neuronal death, other events should be taken into consideration to explain why there is a significant delay between rise in extracellular glutamate concentration and development of postischemic neuronal death and why both glutamate receptor antagonists w14,18x and sodium channel blockers w23,29x Žwhich were found to inhibit ischemia-induced glutamate release w35x. are neuroprotective even if tested after the ischemic period. These results have been recently interpreted by Szatkowski and Attwell w31x and by Obrenovitch and Richards w27x assuming that ischemia induces not only an immediate rise in extracellular glutamate concentration but also a delayed potentiation of the glutamatergic transmission reaching neurotoxic levels. This potentiation could result, theoretically, either from enhanced postsynaptic responses to glutamate or from enhanced glutamate release w31x. The former phenomenon has been experimentally
0006-8993r97r$17.00 Copyright q 1997 Elsevier Science B.V. All rights reserved. PII S 0 0 0 6 - 8 9 9 3 Ž 9 6 . 0 1 4 2 5 - 4
G. Calo` et al.r Brain Research 751 (1997) 300–306
demonstrated. In fact, after an ischemic insult, postsynaptic responses to glutamate mediated by NMDA receptors are potentiated by the following mechanisms: ŽI. long-lasting elevation of glycine extracellular levels w16x; ŽII. up-regulated expression of NMDA receptor subunits w7x; ŽIII. increase of arachidonic acid levels w26x; and ŽIV. activation of protein kinase C w10x. As for the delayed increase of glutamate release induced by ischemia, no data are so far available. Thus, the aims of the present study were to investigate the immediate effects of different metabolic insults Žglucose deprivation, hypoxia and in vitro ischemia. on glutamate efflux in guinea pig cerebral cortex slices and to evaluate the delayed consequences these insults exert on the responsiveness of glutamatergic structures.
2. Materials and methods 2.1. Slice preparation and release studies Guinea pigs Ž350–450 g. of both sexes were decapitated and their brains rapidly dissected. The temporoparietal cortex was used to prepare 400 mm thick slices. The slices were transferred to a beaker containing oxygenated Ž95% O 2 , 5% CO 2 . Krebs solution Žin mM: glucose 10; NaCl 118.5; CaCl 2 2.5; KCl 4.7, MgSO4 1.2; KH 2 PO4 1.2; NaHCO 3 25. for an equilibration period of 20 min. Slices were set up in four perfusion chambers Ž2–3 slices per chamber; 70–100 mg fresh tissue. and superfused with Krebs solution at a flow rate of 0.4 mlrmin for 40 min before starting sample collection. The bath temperature was maintained at 378C Žfor details see Beani et al. w3x.. All the experiments were carried out collecting 10 min samples. Glucose deprivation was obtained removing glucose from the Krebs solution. Hypoxic condition was mimicked by replacing the O 2-equilibrated Krebs solution ŽpO 2 600 mm Hg. with a N2-equilibrated Krebs solution ŽpO 2 - 10 mm Hg.. In vitro ischemia was reproduced replacing O 2 with N2 in a glucose-free solution Žfor details see Badini et al. w2x.. To study the sodium-dependence of glutamate efflux, a low Ž5 mM. sodium solution containing choline instead of sodium and trizma base instead of NaHCO 3 was used. In a separate set of experiments devoted to reveal delayed effects of metabolic insults on depolarizationevoked glutamate efflux, a first KCl 50 mM pulse lasting 10 min was applied at the 60th min ŽS1. of perfusion and a second one was applied at the 210th min ŽS2. of perfusion. The metabolic insult was performed from the 120th to the 150th min Žsee Fig. 4.. Endogenous L-glutamate and GABA concentrations were measured in the samples by HPLC and fluorimetric detection after precolumn derivatization with ophthaldialdehyde. The system consisted of a Chromsep
301
analytical column ŽChrompack, Middelburg, Netherlands. Ž3 mm internal diameter; 10 cm length. containing 5 mm C 18 material, a Triathlon autosampler ŽSpark Holland, Emmen, Netherlands., a fluorescence spectrophotometer RF-551 ŽShimadzu, Kyoto, Japan. Žexcitation and emission wavelengths set at 370 and 450 nm, respectively. and a Beckman 125 pump ŽBeckman Instruments, Fullerton, CA, USA.. Acquisition and analysis of chromatograms was performed by a computer controlled system ŽBeckman Gold System.. The column was perfused with a mobile phase containing 0.1 M sodium acetate, 10% methanol and 2.5% tetrahydrofurane ŽpH 6.5.. Separation was achieved using a two step linear gradient of methanol in aqueous sodium acetate buffer. In these conditions the retention times for glutamate and GABA were f 1.5 and f 9.0 min, respectively. A linear gradient Ž100% eluent changing to 100% methanol over 1 min and returning to 100% eluent 2.5 min later. was adopted to rapidly clean the column after the elution of GABA. Endogenous acetylcholine ŽACh. concentrations in the samples were measured by HPLC and electrochemical detection as described by Giovannini et al. w15x. A cation exchange column was prepared by loading a reverse phase analytical column ŽChrompack, Middelburg, Netherlands. Žinternal diameter 3 mm; length 10 cm. with sodium lauryl sulphate Ž5 mgrml. w11x. Mobile phase consisted of 0.2 M phosphate buffer ŽpH 8.0. containing 5 mM KCl, 1 mM tetramethylammonium and 0.3 mM EDTA. ACh was hydrolysed by acetylcholinesterase in a postcolumn enzyme reactor and hydrogen peroxide was maximally oxidized at q500 mV. The electrochemical detector Žmodel MF 2063, Bioanalytical Systems, West Lafayette, USA. was equipped with a platinum working electrode and an in situ AgrAgCl reference electrode. At a flow rate of 0.7 mlrmin, ACh retention time was f 5.5 min. 2.2. Data presentation and statistical analysis Immediate changes of glutamate spontaneous efflux induced by treatments were expressed as treatmentrbasal ratios, calculated by dividing the outflow of the last treated sample ŽT. with that of the last basal sample ŽB.. Delayed modifications of KCl-induced glutamate overflow caused by treatments were expressed as S2rS1 ratio. Stimulation evoked overflow was calculated by subtracting the presumed basal outflow from the total amount of glutamate found in the stimulated samples. The results are always presented as means " S.E.M. of n experiments. Data were statistically analysed using the Mann-Whitney U-test via a software package w32x; P values lower than 0.05 were considered to be significant. 2.3. Drugs All reagents were purchased from Merck ŽDarmstadt, Germany., except for TTX and ouabain which were from Sigma Chemicals ŽSt. Louis, MO, USA..
G. Calo` et al.r Brain Research 751 (1997) 300–306
302
2.4. Terminology The terms ‘outflow’ or ‘efflux’ were preferred instead of release to indicate that an increase of glutamate levels in the perfusate may depend on either an increase in the release or a reduction in the reuptake.
3. Results 3.1. Spontaneous glutamate efflux After 40 min of perfusion, glutamate concentration in the perfusate was 57 " 4 nM, slowly decreasing Žf 15% for each hour of perfusion. during the experiments. 0.5 mM TTX did not modify glutamate spontaneous efflux, whereas the omission of calcium ions from the medium induced a statistically significant decrease ŽTable 1.. Perfusing slices with a 5 mM Naq containing Krebs solution ŽFig. 3. caused a progressive increase of glutamate levels, which reached after 30 min a peak effect of 8-fold increase. All these effects were rapidly and completely reversible by returning to normal Krebs solution. 3.2. Effect of KCl and ouabain on glutamate efflux A 10 min pulse of 50 mM KCl evoked an immediate f 4-fold increase of glutamate efflux from guinea pig cerebral cortex slices. Under calcium-free condition, KClinduced glutamate efflux was nearly halved. TTX did not modify the effect of KCl ŽTable 1.. Perfusion of the slices with 10 mM ouabain caused a 38-fold increase of glutamate efflux ŽTable 1., reaching the peak after f 20 min of perfusion. Unlike KCl-induced glutamate outflow, ouabain effect was potentiated under calcium-free condition and strongly inhibited by 0.5 mM TTX ŽTable 1..
Fig. 1. Effect of glucose deprivation, hypoxia and in vitro ischemia on glutamate outflow in guinea-pig cerebral cortex slices. Horizontal bar indicates the period of treatment. The data are means"S.E.M. of at least 7 experiments.
3.3. Immediate effects of metabolic insults on glutamate efflux Glucose deprivation for 30 min did not significantly modify glutamate outflow ŽFig. 1., while 30 min of hypoxia evoked a slight but significant increase of glutamate extracellular levels ŽFig. 1.. On the contrary, 30 min of in vitro ischemia strongly stimulated glutamate efflux up to 36-fold the control values. These effects were completely reverted in 20 min after returning to normal Krebs solution ŽFig. 1.. The effect of in vitro ischemia on glutamate efflux was studied in the absence of calcium, in the presence of 0.5
Table 1 Effect of Ca2q-free or tetrodotoxin containing Krebs solution on glutamate spontaneous and stimulated outflow from guinea-pig cerebral cortex slices Standard medium
Ca2 q-free medium
TTX 0.5 mM medium
Spontaneous efflux
0.81" 0.05
0.58" 0.06
a
0.88"0.12
Stimulated efflux KCl Ž50 mM. Ouabain Ž10 mM. In vitro ischemia Ž30 min.
4.27" 0.36 38.47"15.13 36.14" 6.20
2.32" 0.21 a 116.86"27.60 a 77.39"18.58 a
4.02"0.33 12.23"1.27 9.86"3.06
a a
The data, expressed as T r B ratio, are means"S.E.M. of at least 6 experiments. Statistically different from standard medium condition: a P - 0.05 according to the Mann-Whitney U-test.
Fig. 2. Effect of tetrodotoxin against in vitro ischemia-induced glutamate efflux in standard and 5 mM Naq containing Krebs solution in guinea-pig cerebral cortex slices. Horizontal bar indicates the period of in vitro ischemia. The data are means"S.E.M. of at least 6 experiments.
G. Calo` et al.r Brain Research 751 (1997) 300–306
303
efflux. This efflux Žabout 34-fold the control values. was far higher than the sum of glutamate outflow evoked by the single treatments ŽFig. 3.. 3.4. Delayed effects of metabolic insults on KCl-induced glutamate outflow
Fig. 3. Effect of hypoxia, 5 mM Naq containing Krebs solution and hypoxia plus 5 mM Naq on glutamate outflow in guinea-pig cerebral cortex slices. Horizontal bar indicates the period of treatment. The data are means"S.E.M. of at least 6 experiments.
To study delayed effects of metabolic insults, the protocol with two KCl 10-min pulses described in the method section was followed. The S2rS1 ratio in control slices was 0.56 " 0.04. No statistically significant differences in S2rS1 ratios were found between control slices and those exposed to 30 min of deprivation of glucose or hypoxia Ž0.49 " 0.13 and 0.59 " 0.08, respectively.. On the contrary, when 30 min of in vitro ischemia were applied between the two stimulations the S2rS1 ratio was significantly increased Ž1.46 " 0.08. ŽFig. 4.. Furthermore, an in vitro ischemia application period of 15 min was also able to induce a significant increase of S2rS1 ratio Ž1.06 " 0.10..
mM TTX and with a low sodium Ž5 mM. containing Krebs solution. The in vitro ischemia-induced glutamate release was potentiated under Ca2q-free condition, but strongly counteracted by TTX ŽTable 1.. Reducing sodium concentration to 5 mM in the perfusion medium, caused a 7-fold increase of glutamate outflow induced by in vitro ischemia, which became no longer sensitive to TTX ŽFig. 2.. In view of the dramatic effect of the low sodium medium, hypoxia-induced glutamate efflux was re-examined under this condition. Coapplication of hypoxia and low sodium concentrations, caused a massive glutamate
Fig. 4. 30 min of in vitro ischemia performed between the two stimulations significantly increased S 2 rS1 ratio Žcontrol: 0.56"0.04; ischemia: 1.46"0.08.. Closed horizontal bar indicates the period of in vitro ischemia, open horizontal bars indicate the two 50 mM KCl stimulations. The data are means"S.E.M. of 11 experiments. Statistically different from control: ) P - 0.05 according to the Mann-Whitney U-test.
Fig. 5. Panel A: effect of 30 min of in vitro ischemia on KCl-induced glutamate, acetylcholine and GABA overflow in guinea pig cerebral cortex slices. The data are means"S.E.M. of at least 6 experiments. Statistically different from control: ) P - 0.05 according to the MannWhitney U-test. Panel B: under Ca2q-free condition the potentiation induced by 30 min of in vitro ischemia on the stimulated glutamate overflow was more than halved. The data are means"S.E.M. of at least 7 experiments. Statistically different from standard medium condition: ) P - 0.05 according to the Mann-Whitney U-test.
304
G. Calo` et al.r Brain Research 751 (1997) 300–306
The same experimental protocol with two 50 mM KCl stimulations was performed monitoring ACh and GABA levels in the perfusate. The first 10 min pulse of KCl evoked a 24-fold and an 11-fold increase of ACh and GABA efflux, respectively. The S2rS1 ratios for ACh and GABA in control condition were 0.98 " 0.02 and 0.61 " 0.07, respectively. ACh and GABA levels significantly rose during 30 min of in vitro ischemia Ž2-fold and 6-fold increase, respectively. but neither ACh nor GABA S2rS1 ratios in these tissues were significantly different from controls ŽFig. 5, panel A.. Finally, in order to evaluate the calcium-dependence of the potentiation that in vitro ischemia induced on the stimulated glutamate overflow, the second KCl pulse was performed under Ca2q-free condition. Under this experimental condition, glutamate outflow during the second KCl stimulation was dramatically reduced, with a S2rS1 ratio of 0.65 " 0.03 ŽFig. 5, panel B..
4. Discussion 4.1. Spontaneous and stimulated glutamate efflux The first step of the present study was to characterize spontaneous and depolarization-evoked glutamate efflux. Under resting condition, glutamate efflux was completely insensitive to the blockade of voltage-dependent Naq channels; slightly but significantly inhibited by calcium-free medium and greatly facilitated under low sodium condition. All together these results suggested that spontaneous glutamate efflux did not result from conducted neuronal activity. Rather it reflected a minor Ca2q-dependent Žexocytotic?. component and a major component due to leakage from the cytoplasmatic pool mediated by the excitatory amino acid transporter ŽEAAT.. In fact, the direction of net flux of amino acids across the membrane via the EAAT is set by the Naq transmembrane gradient w8,31x. Interestingly, KCl and ouabain appeared to raise extracellular glutamate levels with different mechanisms. Glutamate efflux induced by KCl was unaffected by TTX, but reduced by omitting Ca2q from the perfusion medium. These results were in agreement with other studies Žfor a review see Bernath w6x. indicating that KCl-induced depolarization directly activated voltage-operated calcium channels, causing a rise in intracellular calcium levels and promoting the vesicular exocytosis from glutamatergic terminals. The reduction of Kq electrochemical gradient caused by high Kq-containing medium inhibited the activity of EAAT. Therefore, the Ca2q-independent component of KCl-induced glutamate overflow may be due to the reversal of EAAT. On the contrary, ouabain-induced glutamate efflux was dramatically inhibited by TTX and almost doubled in calcium-free medium. Similar results were obtained in vivo by Westerink et al. w38x with the micro-
dialysis technique. These findings suggest that ouabain Žby inhibiting the NaqrKq ATPase . induced a progressive reduction of the sodium and potassium electrochemical gradients which caused reversed EAA transport. On the whole, our data demonstrated that glutamate efflux could be induced in the same preparation by different stimuli via at least two different mechanisms: a calcium-dependent exocytotic mechanism and a sodium-dependent non-exocytotic mechanism implying the reversal of the glutamate uptake. 4.2. Immediate effects of metabolic insults on glutamate efflux The omission of glucose from the perfusion medium did not modify glutamate efflux. This suggests that this treatment did not deplete the tissue energy stores and consequently did not modify neurotransmitter release. Hypoxic condition caused a slight increase of glutamate efflux. Interestingly, Madl and Burgesser w25x found that in rat hippocampal slices, hypoxia induced a moderate rise in extracellular glutamate concentration due to inhibition of the uptake mechanism: the amount of glutamate released was unchanged while its uptake was decreased down to 10–20%. Instead, in vitro ischemic condition evoked a massive increase of glutamate efflux. The ionic-dependence of this glutamate efflux was strikingly similar to that evoked by ouabain. In fact, in vitro ischemia-induced glutamate outflow was dramatically inhibited in the presence of TTX and almost doubled in the absence of extracellular calcium. Therefore, both ouabain and in vitro ischemia appear to cause glutamate release by the same mechanism: inhibition of the NaqrKq ATPase w22x. As proposed by other studies w1,19,25,34x in vitro ischemiaevoked glutamate efflux originated from the cytosolic pool of neurotransmitter via reversal of the glutamate uptake carrier. This reversal was caused by the increase in intracellular Naq due to the reduction of the NaqrKq ATPase activity which followed the ATP depletion observed under in vitro ischemic condition w21,25x. This hypothesis allows us to interpret several findings of our study. Firstly, the glutamate efflux induced by in vitro ischemia was potentiated by perfusing the slices with a calcium-free medium. This phenomenon was for the first time described by Ikeda et al. w19x and then interpreted by Amoroso et al. w1x as follows: during ischemia, the intracellular sodium overload reversed the direction of operation of both the EAAT and the NaqrCa2q exchanger. Since calcium-free condition blocked the NaqrCa2q exchanger, the only remaining pathway for sodium extrusion was represented by the EAAT. Therefore, under this experimental condition, a larger amount of glutamate was transported outside the cells. Secondly, the presence of TTX in the medium significantly inhibited the effect of in vitro ischemia. This finding was in agreement with other studies w31,34x and demonstrated that the most important pathway
G. Calo` et al.r Brain Research 751 (1997) 300–306
for ischemia-induced sodium overload was represented by voltage-dependent sodium channels Žprobably of non-inactivating nature w33x., which appeared to be blocked by TTX. Thirdly, the pivotal role of Naq rather than Ca2q in the mechanism of the in vitro ischemia-induced EAA efflux, was confirmed by data obtained using low sodium Krebs solution. Under this experimental condition, the stimulatory effect of in vitro ischemia was potentiated and the weak effects of hypoxia were converted in a stimulatory action equal to that induced by in vitro ischemia. These data confirmed the critical role of normal sodium electrochemical gradient as a safety factor in maintaining low levels of extracellular glutamate w12x. Furthermore, the lack of effect of TTX in low sodium conditions demonstrated that the inhibitory action of the toxin on in vitro ischemia-induced glutamate efflux was only due to the blockade of voltage-dependent sodium channels. As for the cellular source Žneurons or glia. of ischemiainduced glutamate efflux, our findings may not give any clear indication. However, it is worthy of note that, under similar experimental conditions, Madl and Burgesser w25x found that neuronal terminals are the main source of in vitro ischemia-induced glutamate overflow. This idea is consistent with previous reports demonstrating that deafferentation decreases the release of glutamate induced by ischemia w5x or hypoglycemia w9x. 4.3. Delayed effects of metabolic insults on KCl-induced glutamate outflow No modifications of KCl effect were found in slices previously exposed to deprivation of glucose or to hypoxia, while after 30 min of in vitro ischemia glutamate release from KCl was more than doubled. Also when ischemic treatment was applied only for 15 min, a statistically significant potentiation of glutamate overflow induced by KCl was evident. Moreover, this delayed potentiation of responsiveness was selectively restricted to the glutamatergic system. In fact, neither GABAergic nor cholinergic structures showed any modification in their response to KCl after 30 min of in vitro ischemia. Furthermore, using a Ca2q-free 50 mM KCl solution during the second stimulation, glutamate efflux potentiation was found to be more than halved. This indicated that postischemic glutamate release potentiation originated mainly from the neuronal pool of the transmitter through exocytosis of vesicles. This is the first evidence of an ischemia-induced selective potentiation of glutamatergic neurosecretory functions. This important finding depicts a postischemic scenario where ischemia-induced super-responsivity of presynaptic structure Ždemonstrated by this study. and super-sensitivity of postsynaptic structures Ždemonstrated by other studies, see Introduction. can generate a mutual potentiating effect. The subsequent overactivation of glutamatergic transmission may be the basis of the postischemic neuronal death.
305
The delayed nature of the proposed mechanisms is consistent with findings demonstrating that both glutamate receptor antagonists w14,18x and sodium channel blockers w23,29x are effective in preventing postischemic neuronal death even if administrated after the ischemic insult. The therapeutic implications of these findings may be of great interest. In fact, according to the above hypothesis, the combination of glutamate receptor antagonists with sodium channel blockers may have greater efficacy in preventing postischemic neuronal death than the single treatment alone. The results recently obtained by Lynch et al. w24x in cortical cell cultures support this view. Further experiments are needed to address the mechanismŽs. by which ischemia induces an abnormal response to depolarization in glutamatergic synapses. As a starting point, the role of different substances such as free-radicals, nitric oxide andror arachidonic acid will be investigated. In fact, these substances are known to be released during ischemia or the subsequent re-oxygenation period and to facilitate glutamate release andror to inhibit glutamate uptake w28,30,36,37x. In conclusion the present study has demonstrated that: ŽI. in vitro ischemia induces an immediate dramatic rise of extracellular glutamate levels; in this process Naq rather than Ca2q plays a pivotal role; ŽII. after an ischemic insult, presynaptic glutamatergic terminals become super-responsive to KCl stimulation releasing larger amounts of glutamate through exocytosis; and ŽIII. this super-responsivity to KCl stimulation induced by in vitro ischemia is selectively restricted to the glutamatergic system. Acknowledgements This work was supported by 40% MURST and Pharmacia Grants. References w1x Amoroso, S., Di Renzo, G.F. and Annunziato, L., Inhibition of the Naq-Ca2q exchanger enhances anoxia and glucopenia-induced w 3 Hx-Aspartate release in hippocampal slices, J. Pharmacol. Exp. Ther., 264 Ž1993. 515–520. w2x Badini, I., Zerbetto, E., Bianchi, C., Beani, L. and Siniscalchi, A., Post-hypoxic recovery of acetylcholine release: different sensitivity of guinea pig neocortical and striatal slices, Neurochem. Int., 29 Ž1996. 477–485. w3x Beani, L., Bianchi, C., Giacomelli, A. and Tamberi, F., Noradrenaline inhibition of acetylcholine release from guinea-pig brain, Eur. J. Pharmacol., 48 Ž1978. 179–193. w4x Benveniste, H., The excytotoxin hypothesis in relation to cerebral ischemia, CerebroÕasc. Brain Metab. ReÕ., 3 Ž1991. 213–245. w5x Benveniste, H., Jorgensen, M.B., Sandberg, M., Christensen, T., Hagberg, H. and Diemer, N.H., Ischemic damage in hippocampal CA1 is dependent on glutamate release and intact innervation from CA3, J. Cereb. Blood Flow Metab., 9 Ž1989. 629–639. w6x Bernath, S., Calcium-independent release of amino acid neurotransmitters: fact or artefact?, Prog. Neurobiol., 38 Ž1992. 57–91. w7x Bessho, Y., Nawa, H. and Nakanishi, S., Selective up-regulation of
306
w8x
w9x
w10x
w11x
w12x w13x
w14x
w15x
w16x
w17x
w18x
w19x
w20x
w21x
w22x
G. Calo` et al.r Brain Research 751 (1997) 300–306 an NMDA receptor subunit mRNA in cultured cerebellar granule cells by Kq induced depolarization and NMDA treatment, Neuron, 12 Ž1994. 87–95. Bouvier, M., Szatkowski, M., Amato, A. and Attwell, D., The glial cell glutamate uptake carrier countertransports pH-changing anions, Nature, 360 Ž1992. 471–474. Butcher, S.P., Sandberg, M., Hagberg, H. and Hamberger, A., Cellular origins of endogenous amino acids released into the extracellular fluid of the rat striatum during severe insulin-induced hypoglycemia, J. Neurochem., 48 Ž1987. 722–728. Chen, L. and Huang, L.Y., Protein kinase C reduces Mg 2q block of NMDA-receptor channels as a mechanism of modulation, Nature, 356 Ž1992. 521–523. Damsma, G., Lammerts van Bueren, D., Westerink, B.H.C. and Horn, A.S., Determination of acetylcholine and choline in the femtomole range by means of HPLC, a post-column enzyme reactor and electrochemical detection, Chromatographia, 24 Ž1987. 827–831. Danbolt, N.C., The high affinity uptake system for excitatory amino acids in the brain, Prog. Neurobiol., 44 Ž1994. 377–396. Drejer, J., Benveniste, H., Diemer, N.H. and Schousboe, A., Cellular origin of ischemia-induced glutamate release from brain tissue in vivo and in vitro, J. Neurochem., 45 Ž1985. 145–151. Gill, R., Foster, A.C. and Woodruff, G.N., MK-801 is neuroprotective in gerbils when administrated during the postischaemic period, Neuroscience, 25 Ž1988. 847–855. Giovannini, M.G., Mutolo, D., Bianchi, L., Michelassi, A. and Pepeu, G., NMDA receptor antagonists decrease GABA outflow from the septum and increase acetylcholine outflow from the hippocampus: a microdialysis study, J. Neurosci., 14 Ž1994. 1358–1365. Globus, M.Y.T., Ginsberg, M.D. and Busto, R., Excytotoxic index – a biochemical marker of selective vulnerability, Neurosci. Lett., 127 Ž1991. 39–42. Gundersen, V., Shuplliakov, O., Brodin, L., Ottersen, O.P. and Storm-Mathisen, J., Quantification of excitatory amino acid uptake at intact glutamatergic synapses by immunocytochemistry of exogenous D-aspartate, J. Neurosci., 15 Ž1995. 4417–4428. Hartley, D.M. and Choi, D.W., Delayed rescue of NMDA receptormediated neuronal injury in cortical culture, J. Pharmacol. Exp. Ther., 250 Ž1989. 752–758. Ikeda, M., Nakazawa, T., Abe, K., Kaneko, T. and Yamatsu, K., Extracellular accumulation of glutamate in the hippocampus induced by ischemia is not calcium dependent – in vitro and in vivo evidence, Neurosci. Lett., 96 Ž1989. 202–206. Katayama, Y., Katayama, T., Tamura, T., Hovda, D.A., Becker, D.P. and Tsubokawa, T., Calcium-dependent glutamate release concomitant with massive potassium flux during cerebral ischemia in vivo, Brain Res., 558 Ž1991. 136–140. Kauppiten, R.A., McMahon, H.T. and Nicholls, D.G., Ca2q-dependent and Ca2q-independent glutamate release, energy status and cytosolic free Ca2q concentrations in isolated nerve terminals following metabolic inhibition: possible relevance to hypoglycaemia and anoxia, Neuroscience, 1 Ž1988. 175–182. Lees, G.J., Inhibition of sodium-potassium-ATPase: a potentially ubiquitous mechanism contributing to central nervous system neu-
ropathology, Brain Res. ReÕ., 16 Ž1991. 283–300. w23x Leikieffre, D. and Meldrum, B.S., The pyrimidine-derivate, BW1003C87, protects CA1 and striatal neurones following transient severe forebrain ischaemia in rats. A microdialysis and histological study, Neuroscience, 56 Ž1993. 93–99. w24x Lynch III, J.J., Yu, S.P., Canzoniero, L.M.T., Sensi, S.L. and Choi, D.W., Sodium channel blockers reduce oxygen-glucose deprivationinduced cortical neuronal injury when combined with glutamate receptor antagonists, J. Pharmacol. Exp. Ther., 273 Ž1995. 554–560. w25x Madl, J.E. and Burgesser, K., Adenosine triphosphate depletion reverses sodium-dependent, neuronal uptake of glutamate in rat hippocampal slices, J. Neurosci., 13 Ž1993. 4429–44. w26x Miller, B., Sarantis, M., Traynelis, S.F. and Attwell, D., Potentiation of NMDA currents by arachidonic acid, Nature, 355 Ž1992. 722– 725. w27x Obrenovitch, T.P. and Richards, D.A., Extracellular Neurotransmitter changes in cerebral ischaemia, CerebroÕasc. Brain Metab. ReÕ., 7 Ž1995. 1–54. w28x Pellegrini-Giampietro, D.E., Cherici, G., Alesiani, M., Carla, V. and Moroni, F., Excitatory amino acid release and free radical formation may cooperate in the genesis of ischemia-induced neuronal damage, J. Neurosci., 10 Ž1990. 1035–1041. w29x Pratt, J., Rataud, J., Bartot, F., Roux, M., Blanchard, J.C., Laduron, P.M. and Stutzmann, J.M., Neuroprotective actions of riluzole in rodent models of global and focal cerebral ischaemia, Neurosci. Lett., 140 Ž1992. 225–230. w30x Simonato, M., Bregola, G., Bianchi, C. and Beani, L., Effect of arachidonic acid on w 3 Hx-D-aspartate outflow in the rat hippocampus, Neurochem. Res., 19 Ž1994. 195–200. w31x Szatkowski, M. and Attwell, D., Triggering and execution of neuronal death in brain ischemia: two phases of glutamate release by different mechanisms, Trends Neurosci., 17 Ž1994. 359–365. w32x Tallarida, R.J. and Murray, R.B., Manual of Pharmacologic Calculations with Computer Programs, Springer-Verlag, 1986. w33x Taylor, C.P., Naq currents that fail to inactivate, Trends Neurosci., 16 Ž1993. 455–460. w34x Taylor, C.P., Burke, S.P. and Weber, M.L., Hippocampal slices: glutamate overflow and cellular damage from ischemia are reduced by sodium-channel blockade, J. Neurosci. Methods, 59 Ž1995. 121– 128. w35x Urenjak, J. and Obrenovitch, T.P., Pharmacological modulation of voltage-gated Naq channels: a rational and effective strategy against ischemic brain damage, Pharmacol. ReÕ., 48 Ž1996. 2167. w36x Vazquez, E., Herrero, I., Miras-Portugal, M.T. and Sanchez-Prieto, J., Role of arachidonic acid in the facilitation of glutamate release from rat cerebrocortical synaptosomes independent of metabotropic glutamate receptor responses, Neurosci. Lett., 174 Ž1994. 9–13. w37x Volterra, A., Trotti, D., Tromba, C., Floridi, S. and Racagni, G., Glutamate uptake inhibition by oxygen free radicals in rat cortical astrocytes, J. Neurosci., 14 Ž1994. 2924–2932. w38x Westerink, B.H.C., Damsma, G. and De Vries, J.B., Effect of ouabain applied by intrastriatal microdialysis on the in vivo release of dopamine, acetylcholine, and amino acids in the brain of conscious rats, J. Neurochem., 52 Ž1989. 705–712.
Research report
Immediate and delayed effects of in vitro ischemia on glutamate efflux from guinea-pig cerebral cortex slices G. Calo, ` S. Sbrenna, C. Bianchi, L. Beani
)
Institute of Pharmacology, UniÕersity of Ferrara, Via Fossato di Mortara 17-19, 44100 Ferrara, Italy Accepted 20 November 1996
Abstract Immediate and delayed effects of glucose deprivation, oxygen deprivation Žhypoxia. and both oxygen and glucose deprivation Žin vitro ischemia. on glutamate efflux from guinea pig cerebral cortex slices were studied. Immediate effects were evaluated by measuring changes of glutamate efflux during the metabolic insults. Delayed effects were evaluated by measuring the response of the tissue to a 50 mM KCl pulse applied 60 min after the metabolic insults. Deprivation of glucose in the medium did not induce either immediate or delayed effects, while hypoxic condition produced an immediate slight stimulation of glutamate efflux without any delayed effect. Conversely, in vitro ischemia produced both immediate and delayed effects on glutamate efflux. During in vitro ischemia glutamate efflux dramatically increased in a calcium-independent and tetrodotoxin-sensitive manner; this effect was potentiated by a low sodium containing medium. The blockade of the sodiumrpotassium ATPase exchanger by ouabain caused a glutamate outflow similar to that induced by in vitro ischemia. On the whole, these data demonstrate the central role played by the sodium electrochemical gradient and by the membrane glutamate uptake system in the glutamate overflow induced by in vitro ischemia. Moreover, in slices previously exposed to both oxygen and glucose deprivation the effect of KCl on glutamate efflux was potentiated. This in vitro ischemia-induced delayed potentiation of neurotransmitter efflux, until now unreported in the literature, was found to be selectively restricted to glutamatergic structures and to be mainly due to an enhancement of the exocytotic component of glutamate release. q 1997 Elsevier Science B.V. All rights reserved. Keywords: Glutamate; In vitro ischemia; Brain slice; NaqrKq ATPase ; Naqrglutamate transporter; Postischemic glutamate release potentiation
1. Introduction Both in vitro w19,28x and in vivo w4,19x studies have shown that extracellular glutamate concentration massively increases during cerebral ischemia. In spite of the numerous studies addressed to this field, the mechanismŽs. leading to accumulation of extracellular glutamate during cerebrovascular insults is not completely understood. In fact, contrasting results are reported in the literature regarding either the Ca2q-dependence w1,13,19,20x or the cellular origin of ischemia-induced glutamate outflow w17,25x. However, it is generally accepted that overactivation of glutamate receptors promotes postischemic neuronal damage w4x.
) Corresponding author. Fax: q39 Ž532. 291205; E-mail: [email protected]
Although the immediate rise of extracellular glutamate has an important role in postischemic neuronal death, other events should be taken into consideration to explain why there is a significant delay between rise in extracellular glutamate concentration and development of postischemic neuronal death and why both glutamate receptor antagonists w14,18x and sodium channel blockers w23,29x Žwhich were found to inhibit ischemia-induced glutamate release w35x. are neuroprotective even if tested after the ischemic period. These results have been recently interpreted by Szatkowski and Attwell w31x and by Obrenovitch and Richards w27x assuming that ischemia induces not only an immediate rise in extracellular glutamate concentration but also a delayed potentiation of the glutamatergic transmission reaching neurotoxic levels. This potentiation could result, theoretically, either from enhanced postsynaptic responses to glutamate or from enhanced glutamate release w31x. The former phenomenon has been experimentally
0006-8993r97r$17.00 Copyright q 1997 Elsevier Science B.V. All rights reserved. PII S 0 0 0 6 - 8 9 9 3 Ž 9 6 . 0 1 4 2 5 - 4
G. Calo` et al.r Brain Research 751 (1997) 300–306
demonstrated. In fact, after an ischemic insult, postsynaptic responses to glutamate mediated by NMDA receptors are potentiated by the following mechanisms: ŽI. long-lasting elevation of glycine extracellular levels w16x; ŽII. up-regulated expression of NMDA receptor subunits w7x; ŽIII. increase of arachidonic acid levels w26x; and ŽIV. activation of protein kinase C w10x. As for the delayed increase of glutamate release induced by ischemia, no data are so far available. Thus, the aims of the present study were to investigate the immediate effects of different metabolic insults Žglucose deprivation, hypoxia and in vitro ischemia. on glutamate efflux in guinea pig cerebral cortex slices and to evaluate the delayed consequences these insults exert on the responsiveness of glutamatergic structures.
2. Materials and methods 2.1. Slice preparation and release studies Guinea pigs Ž350–450 g. of both sexes were decapitated and their brains rapidly dissected. The temporoparietal cortex was used to prepare 400 mm thick slices. The slices were transferred to a beaker containing oxygenated Ž95% O 2 , 5% CO 2 . Krebs solution Žin mM: glucose 10; NaCl 118.5; CaCl 2 2.5; KCl 4.7, MgSO4 1.2; KH 2 PO4 1.2; NaHCO 3 25. for an equilibration period of 20 min. Slices were set up in four perfusion chambers Ž2–3 slices per chamber; 70–100 mg fresh tissue. and superfused with Krebs solution at a flow rate of 0.4 mlrmin for 40 min before starting sample collection. The bath temperature was maintained at 378C Žfor details see Beani et al. w3x.. All the experiments were carried out collecting 10 min samples. Glucose deprivation was obtained removing glucose from the Krebs solution. Hypoxic condition was mimicked by replacing the O 2-equilibrated Krebs solution ŽpO 2 600 mm Hg. with a N2-equilibrated Krebs solution ŽpO 2 - 10 mm Hg.. In vitro ischemia was reproduced replacing O 2 with N2 in a glucose-free solution Žfor details see Badini et al. w2x.. To study the sodium-dependence of glutamate efflux, a low Ž5 mM. sodium solution containing choline instead of sodium and trizma base instead of NaHCO 3 was used. In a separate set of experiments devoted to reveal delayed effects of metabolic insults on depolarizationevoked glutamate efflux, a first KCl 50 mM pulse lasting 10 min was applied at the 60th min ŽS1. of perfusion and a second one was applied at the 210th min ŽS2. of perfusion. The metabolic insult was performed from the 120th to the 150th min Žsee Fig. 4.. Endogenous L-glutamate and GABA concentrations were measured in the samples by HPLC and fluorimetric detection after precolumn derivatization with ophthaldialdehyde. The system consisted of a Chromsep
301
analytical column ŽChrompack, Middelburg, Netherlands. Ž3 mm internal diameter; 10 cm length. containing 5 mm C 18 material, a Triathlon autosampler ŽSpark Holland, Emmen, Netherlands., a fluorescence spectrophotometer RF-551 ŽShimadzu, Kyoto, Japan. Žexcitation and emission wavelengths set at 370 and 450 nm, respectively. and a Beckman 125 pump ŽBeckman Instruments, Fullerton, CA, USA.. Acquisition and analysis of chromatograms was performed by a computer controlled system ŽBeckman Gold System.. The column was perfused with a mobile phase containing 0.1 M sodium acetate, 10% methanol and 2.5% tetrahydrofurane ŽpH 6.5.. Separation was achieved using a two step linear gradient of methanol in aqueous sodium acetate buffer. In these conditions the retention times for glutamate and GABA were f 1.5 and f 9.0 min, respectively. A linear gradient Ž100% eluent changing to 100% methanol over 1 min and returning to 100% eluent 2.5 min later. was adopted to rapidly clean the column after the elution of GABA. Endogenous acetylcholine ŽACh. concentrations in the samples were measured by HPLC and electrochemical detection as described by Giovannini et al. w15x. A cation exchange column was prepared by loading a reverse phase analytical column ŽChrompack, Middelburg, Netherlands. Žinternal diameter 3 mm; length 10 cm. with sodium lauryl sulphate Ž5 mgrml. w11x. Mobile phase consisted of 0.2 M phosphate buffer ŽpH 8.0. containing 5 mM KCl, 1 mM tetramethylammonium and 0.3 mM EDTA. ACh was hydrolysed by acetylcholinesterase in a postcolumn enzyme reactor and hydrogen peroxide was maximally oxidized at q500 mV. The electrochemical detector Žmodel MF 2063, Bioanalytical Systems, West Lafayette, USA. was equipped with a platinum working electrode and an in situ AgrAgCl reference electrode. At a flow rate of 0.7 mlrmin, ACh retention time was f 5.5 min. 2.2. Data presentation and statistical analysis Immediate changes of glutamate spontaneous efflux induced by treatments were expressed as treatmentrbasal ratios, calculated by dividing the outflow of the last treated sample ŽT. with that of the last basal sample ŽB.. Delayed modifications of KCl-induced glutamate overflow caused by treatments were expressed as S2rS1 ratio. Stimulation evoked overflow was calculated by subtracting the presumed basal outflow from the total amount of glutamate found in the stimulated samples. The results are always presented as means " S.E.M. of n experiments. Data were statistically analysed using the Mann-Whitney U-test via a software package w32x; P values lower than 0.05 were considered to be significant. 2.3. Drugs All reagents were purchased from Merck ŽDarmstadt, Germany., except for TTX and ouabain which were from Sigma Chemicals ŽSt. Louis, MO, USA..
G. Calo` et al.r Brain Research 751 (1997) 300–306
302
2.4. Terminology The terms ‘outflow’ or ‘efflux’ were preferred instead of release to indicate that an increase of glutamate levels in the perfusate may depend on either an increase in the release or a reduction in the reuptake.
3. Results 3.1. Spontaneous glutamate efflux After 40 min of perfusion, glutamate concentration in the perfusate was 57 " 4 nM, slowly decreasing Žf 15% for each hour of perfusion. during the experiments. 0.5 mM TTX did not modify glutamate spontaneous efflux, whereas the omission of calcium ions from the medium induced a statistically significant decrease ŽTable 1.. Perfusing slices with a 5 mM Naq containing Krebs solution ŽFig. 3. caused a progressive increase of glutamate levels, which reached after 30 min a peak effect of 8-fold increase. All these effects were rapidly and completely reversible by returning to normal Krebs solution. 3.2. Effect of KCl and ouabain on glutamate efflux A 10 min pulse of 50 mM KCl evoked an immediate f 4-fold increase of glutamate efflux from guinea pig cerebral cortex slices. Under calcium-free condition, KClinduced glutamate efflux was nearly halved. TTX did not modify the effect of KCl ŽTable 1.. Perfusion of the slices with 10 mM ouabain caused a 38-fold increase of glutamate efflux ŽTable 1., reaching the peak after f 20 min of perfusion. Unlike KCl-induced glutamate outflow, ouabain effect was potentiated under calcium-free condition and strongly inhibited by 0.5 mM TTX ŽTable 1..
Fig. 1. Effect of glucose deprivation, hypoxia and in vitro ischemia on glutamate outflow in guinea-pig cerebral cortex slices. Horizontal bar indicates the period of treatment. The data are means"S.E.M. of at least 7 experiments.
3.3. Immediate effects of metabolic insults on glutamate efflux Glucose deprivation for 30 min did not significantly modify glutamate outflow ŽFig. 1., while 30 min of hypoxia evoked a slight but significant increase of glutamate extracellular levels ŽFig. 1.. On the contrary, 30 min of in vitro ischemia strongly stimulated glutamate efflux up to 36-fold the control values. These effects were completely reverted in 20 min after returning to normal Krebs solution ŽFig. 1.. The effect of in vitro ischemia on glutamate efflux was studied in the absence of calcium, in the presence of 0.5
Table 1 Effect of Ca2q-free or tetrodotoxin containing Krebs solution on glutamate spontaneous and stimulated outflow from guinea-pig cerebral cortex slices Standard medium
Ca2 q-free medium
TTX 0.5 mM medium
Spontaneous efflux
0.81" 0.05
0.58" 0.06
a
0.88"0.12
Stimulated efflux KCl Ž50 mM. Ouabain Ž10 mM. In vitro ischemia Ž30 min.
4.27" 0.36 38.47"15.13 36.14" 6.20
2.32" 0.21 a 116.86"27.60 a 77.39"18.58 a
4.02"0.33 12.23"1.27 9.86"3.06
a a
The data, expressed as T r B ratio, are means"S.E.M. of at least 6 experiments. Statistically different from standard medium condition: a P - 0.05 according to the Mann-Whitney U-test.
Fig. 2. Effect of tetrodotoxin against in vitro ischemia-induced glutamate efflux in standard and 5 mM Naq containing Krebs solution in guinea-pig cerebral cortex slices. Horizontal bar indicates the period of in vitro ischemia. The data are means"S.E.M. of at least 6 experiments.
G. Calo` et al.r Brain Research 751 (1997) 300–306
303
efflux. This efflux Žabout 34-fold the control values. was far higher than the sum of glutamate outflow evoked by the single treatments ŽFig. 3.. 3.4. Delayed effects of metabolic insults on KCl-induced glutamate outflow
Fig. 3. Effect of hypoxia, 5 mM Naq containing Krebs solution and hypoxia plus 5 mM Naq on glutamate outflow in guinea-pig cerebral cortex slices. Horizontal bar indicates the period of treatment. The data are means"S.E.M. of at least 6 experiments.
To study delayed effects of metabolic insults, the protocol with two KCl 10-min pulses described in the method section was followed. The S2rS1 ratio in control slices was 0.56 " 0.04. No statistically significant differences in S2rS1 ratios were found between control slices and those exposed to 30 min of deprivation of glucose or hypoxia Ž0.49 " 0.13 and 0.59 " 0.08, respectively.. On the contrary, when 30 min of in vitro ischemia were applied between the two stimulations the S2rS1 ratio was significantly increased Ž1.46 " 0.08. ŽFig. 4.. Furthermore, an in vitro ischemia application period of 15 min was also able to induce a significant increase of S2rS1 ratio Ž1.06 " 0.10..
mM TTX and with a low sodium Ž5 mM. containing Krebs solution. The in vitro ischemia-induced glutamate release was potentiated under Ca2q-free condition, but strongly counteracted by TTX ŽTable 1.. Reducing sodium concentration to 5 mM in the perfusion medium, caused a 7-fold increase of glutamate outflow induced by in vitro ischemia, which became no longer sensitive to TTX ŽFig. 2.. In view of the dramatic effect of the low sodium medium, hypoxia-induced glutamate efflux was re-examined under this condition. Coapplication of hypoxia and low sodium concentrations, caused a massive glutamate
Fig. 4. 30 min of in vitro ischemia performed between the two stimulations significantly increased S 2 rS1 ratio Žcontrol: 0.56"0.04; ischemia: 1.46"0.08.. Closed horizontal bar indicates the period of in vitro ischemia, open horizontal bars indicate the two 50 mM KCl stimulations. The data are means"S.E.M. of 11 experiments. Statistically different from control: ) P - 0.05 according to the Mann-Whitney U-test.
Fig. 5. Panel A: effect of 30 min of in vitro ischemia on KCl-induced glutamate, acetylcholine and GABA overflow in guinea pig cerebral cortex slices. The data are means"S.E.M. of at least 6 experiments. Statistically different from control: ) P - 0.05 according to the MannWhitney U-test. Panel B: under Ca2q-free condition the potentiation induced by 30 min of in vitro ischemia on the stimulated glutamate overflow was more than halved. The data are means"S.E.M. of at least 7 experiments. Statistically different from standard medium condition: ) P - 0.05 according to the Mann-Whitney U-test.
304
G. Calo` et al.r Brain Research 751 (1997) 300–306
The same experimental protocol with two 50 mM KCl stimulations was performed monitoring ACh and GABA levels in the perfusate. The first 10 min pulse of KCl evoked a 24-fold and an 11-fold increase of ACh and GABA efflux, respectively. The S2rS1 ratios for ACh and GABA in control condition were 0.98 " 0.02 and 0.61 " 0.07, respectively. ACh and GABA levels significantly rose during 30 min of in vitro ischemia Ž2-fold and 6-fold increase, respectively. but neither ACh nor GABA S2rS1 ratios in these tissues were significantly different from controls ŽFig. 5, panel A.. Finally, in order to evaluate the calcium-dependence of the potentiation that in vitro ischemia induced on the stimulated glutamate overflow, the second KCl pulse was performed under Ca2q-free condition. Under this experimental condition, glutamate outflow during the second KCl stimulation was dramatically reduced, with a S2rS1 ratio of 0.65 " 0.03 ŽFig. 5, panel B..
4. Discussion 4.1. Spontaneous and stimulated glutamate efflux The first step of the present study was to characterize spontaneous and depolarization-evoked glutamate efflux. Under resting condition, glutamate efflux was completely insensitive to the blockade of voltage-dependent Naq channels; slightly but significantly inhibited by calcium-free medium and greatly facilitated under low sodium condition. All together these results suggested that spontaneous glutamate efflux did not result from conducted neuronal activity. Rather it reflected a minor Ca2q-dependent Žexocytotic?. component and a major component due to leakage from the cytoplasmatic pool mediated by the excitatory amino acid transporter ŽEAAT.. In fact, the direction of net flux of amino acids across the membrane via the EAAT is set by the Naq transmembrane gradient w8,31x. Interestingly, KCl and ouabain appeared to raise extracellular glutamate levels with different mechanisms. Glutamate efflux induced by KCl was unaffected by TTX, but reduced by omitting Ca2q from the perfusion medium. These results were in agreement with other studies Žfor a review see Bernath w6x. indicating that KCl-induced depolarization directly activated voltage-operated calcium channels, causing a rise in intracellular calcium levels and promoting the vesicular exocytosis from glutamatergic terminals. The reduction of Kq electrochemical gradient caused by high Kq-containing medium inhibited the activity of EAAT. Therefore, the Ca2q-independent component of KCl-induced glutamate overflow may be due to the reversal of EAAT. On the contrary, ouabain-induced glutamate efflux was dramatically inhibited by TTX and almost doubled in calcium-free medium. Similar results were obtained in vivo by Westerink et al. w38x with the micro-
dialysis technique. These findings suggest that ouabain Žby inhibiting the NaqrKq ATPase . induced a progressive reduction of the sodium and potassium electrochemical gradients which caused reversed EAA transport. On the whole, our data demonstrated that glutamate efflux could be induced in the same preparation by different stimuli via at least two different mechanisms: a calcium-dependent exocytotic mechanism and a sodium-dependent non-exocytotic mechanism implying the reversal of the glutamate uptake. 4.2. Immediate effects of metabolic insults on glutamate efflux The omission of glucose from the perfusion medium did not modify glutamate efflux. This suggests that this treatment did not deplete the tissue energy stores and consequently did not modify neurotransmitter release. Hypoxic condition caused a slight increase of glutamate efflux. Interestingly, Madl and Burgesser w25x found that in rat hippocampal slices, hypoxia induced a moderate rise in extracellular glutamate concentration due to inhibition of the uptake mechanism: the amount of glutamate released was unchanged while its uptake was decreased down to 10–20%. Instead, in vitro ischemic condition evoked a massive increase of glutamate efflux. The ionic-dependence of this glutamate efflux was strikingly similar to that evoked by ouabain. In fact, in vitro ischemia-induced glutamate outflow was dramatically inhibited in the presence of TTX and almost doubled in the absence of extracellular calcium. Therefore, both ouabain and in vitro ischemia appear to cause glutamate release by the same mechanism: inhibition of the NaqrKq ATPase w22x. As proposed by other studies w1,19,25,34x in vitro ischemiaevoked glutamate efflux originated from the cytosolic pool of neurotransmitter via reversal of the glutamate uptake carrier. This reversal was caused by the increase in intracellular Naq due to the reduction of the NaqrKq ATPase activity which followed the ATP depletion observed under in vitro ischemic condition w21,25x. This hypothesis allows us to interpret several findings of our study. Firstly, the glutamate efflux induced by in vitro ischemia was potentiated by perfusing the slices with a calcium-free medium. This phenomenon was for the first time described by Ikeda et al. w19x and then interpreted by Amoroso et al. w1x as follows: during ischemia, the intracellular sodium overload reversed the direction of operation of both the EAAT and the NaqrCa2q exchanger. Since calcium-free condition blocked the NaqrCa2q exchanger, the only remaining pathway for sodium extrusion was represented by the EAAT. Therefore, under this experimental condition, a larger amount of glutamate was transported outside the cells. Secondly, the presence of TTX in the medium significantly inhibited the effect of in vitro ischemia. This finding was in agreement with other studies w31,34x and demonstrated that the most important pathway
G. Calo` et al.r Brain Research 751 (1997) 300–306
for ischemia-induced sodium overload was represented by voltage-dependent sodium channels Žprobably of non-inactivating nature w33x., which appeared to be blocked by TTX. Thirdly, the pivotal role of Naq rather than Ca2q in the mechanism of the in vitro ischemia-induced EAA efflux, was confirmed by data obtained using low sodium Krebs solution. Under this experimental condition, the stimulatory effect of in vitro ischemia was potentiated and the weak effects of hypoxia were converted in a stimulatory action equal to that induced by in vitro ischemia. These data confirmed the critical role of normal sodium electrochemical gradient as a safety factor in maintaining low levels of extracellular glutamate w12x. Furthermore, the lack of effect of TTX in low sodium conditions demonstrated that the inhibitory action of the toxin on in vitro ischemia-induced glutamate efflux was only due to the blockade of voltage-dependent sodium channels. As for the cellular source Žneurons or glia. of ischemiainduced glutamate efflux, our findings may not give any clear indication. However, it is worthy of note that, under similar experimental conditions, Madl and Burgesser w25x found that neuronal terminals are the main source of in vitro ischemia-induced glutamate overflow. This idea is consistent with previous reports demonstrating that deafferentation decreases the release of glutamate induced by ischemia w5x or hypoglycemia w9x. 4.3. Delayed effects of metabolic insults on KCl-induced glutamate outflow No modifications of KCl effect were found in slices previously exposed to deprivation of glucose or to hypoxia, while after 30 min of in vitro ischemia glutamate release from KCl was more than doubled. Also when ischemic treatment was applied only for 15 min, a statistically significant potentiation of glutamate overflow induced by KCl was evident. Moreover, this delayed potentiation of responsiveness was selectively restricted to the glutamatergic system. In fact, neither GABAergic nor cholinergic structures showed any modification in their response to KCl after 30 min of in vitro ischemia. Furthermore, using a Ca2q-free 50 mM KCl solution during the second stimulation, glutamate efflux potentiation was found to be more than halved. This indicated that postischemic glutamate release potentiation originated mainly from the neuronal pool of the transmitter through exocytosis of vesicles. This is the first evidence of an ischemia-induced selective potentiation of glutamatergic neurosecretory functions. This important finding depicts a postischemic scenario where ischemia-induced super-responsivity of presynaptic structure Ždemonstrated by this study. and super-sensitivity of postsynaptic structures Ždemonstrated by other studies, see Introduction. can generate a mutual potentiating effect. The subsequent overactivation of glutamatergic transmission may be the basis of the postischemic neuronal death.
305
The delayed nature of the proposed mechanisms is consistent with findings demonstrating that both glutamate receptor antagonists w14,18x and sodium channel blockers w23,29x are effective in preventing postischemic neuronal death even if administrated after the ischemic insult. The therapeutic implications of these findings may be of great interest. In fact, according to the above hypothesis, the combination of glutamate receptor antagonists with sodium channel blockers may have greater efficacy in preventing postischemic neuronal death than the single treatment alone. The results recently obtained by Lynch et al. w24x in cortical cell cultures support this view. Further experiments are needed to address the mechanismŽs. by which ischemia induces an abnormal response to depolarization in glutamatergic synapses. As a starting point, the role of different substances such as free-radicals, nitric oxide andror arachidonic acid will be investigated. In fact, these substances are known to be released during ischemia or the subsequent re-oxygenation period and to facilitate glutamate release andror to inhibit glutamate uptake w28,30,36,37x. In conclusion the present study has demonstrated that: ŽI. in vitro ischemia induces an immediate dramatic rise of extracellular glutamate levels; in this process Naq rather than Ca2q plays a pivotal role; ŽII. after an ischemic insult, presynaptic glutamatergic terminals become super-responsive to KCl stimulation releasing larger amounts of glutamate through exocytosis; and ŽIII. this super-responsivity to KCl stimulation induced by in vitro ischemia is selectively restricted to the glutamatergic system. Acknowledgements This work was supported by 40% MURST and Pharmacia Grants. References w1x Amoroso, S., Di Renzo, G.F. and Annunziato, L., Inhibition of the Naq-Ca2q exchanger enhances anoxia and glucopenia-induced w 3 Hx-Aspartate release in hippocampal slices, J. Pharmacol. Exp. Ther., 264 Ž1993. 515–520. w2x Badini, I., Zerbetto, E., Bianchi, C., Beani, L. and Siniscalchi, A., Post-hypoxic recovery of acetylcholine release: different sensitivity of guinea pig neocortical and striatal slices, Neurochem. Int., 29 Ž1996. 477–485. w3x Beani, L., Bianchi, C., Giacomelli, A. and Tamberi, F., Noradrenaline inhibition of acetylcholine release from guinea-pig brain, Eur. J. Pharmacol., 48 Ž1978. 179–193. w4x Benveniste, H., The excytotoxin hypothesis in relation to cerebral ischemia, CerebroÕasc. Brain Metab. ReÕ., 3 Ž1991. 213–245. w5x Benveniste, H., Jorgensen, M.B., Sandberg, M., Christensen, T., Hagberg, H. and Diemer, N.H., Ischemic damage in hippocampal CA1 is dependent on glutamate release and intact innervation from CA3, J. Cereb. Blood Flow Metab., 9 Ž1989. 629–639. w6x Bernath, S., Calcium-independent release of amino acid neurotransmitters: fact or artefact?, Prog. Neurobiol., 38 Ž1992. 57–91. w7x Bessho, Y., Nawa, H. and Nakanishi, S., Selective up-regulation of
306
w8x
w9x
w10x
w11x
w12x w13x
w14x
w15x
w16x
w17x
w18x
w19x
w20x
w21x
w22x
G. Calo` et al.r Brain Research 751 (1997) 300–306 an NMDA receptor subunit mRNA in cultured cerebellar granule cells by Kq induced depolarization and NMDA treatment, Neuron, 12 Ž1994. 87–95. Bouvier, M., Szatkowski, M., Amato, A. and Attwell, D., The glial cell glutamate uptake carrier countertransports pH-changing anions, Nature, 360 Ž1992. 471–474. Butcher, S.P., Sandberg, M., Hagberg, H. and Hamberger, A., Cellular origins of endogenous amino acids released into the extracellular fluid of the rat striatum during severe insulin-induced hypoglycemia, J. Neurochem., 48 Ž1987. 722–728. Chen, L. and Huang, L.Y., Protein kinase C reduces Mg 2q block of NMDA-receptor channels as a mechanism of modulation, Nature, 356 Ž1992. 521–523. Damsma, G., Lammerts van Bueren, D., Westerink, B.H.C. and Horn, A.S., Determination of acetylcholine and choline in the femtomole range by means of HPLC, a post-column enzyme reactor and electrochemical detection, Chromatographia, 24 Ž1987. 827–831. Danbolt, N.C., The high affinity uptake system for excitatory amino acids in the brain, Prog. Neurobiol., 44 Ž1994. 377–396. Drejer, J., Benveniste, H., Diemer, N.H. and Schousboe, A., Cellular origin of ischemia-induced glutamate release from brain tissue in vivo and in vitro, J. Neurochem., 45 Ž1985. 145–151. Gill, R., Foster, A.C. and Woodruff, G.N., MK-801 is neuroprotective in gerbils when administrated during the postischaemic period, Neuroscience, 25 Ž1988. 847–855. Giovannini, M.G., Mutolo, D., Bianchi, L., Michelassi, A. and Pepeu, G., NMDA receptor antagonists decrease GABA outflow from the septum and increase acetylcholine outflow from the hippocampus: a microdialysis study, J. Neurosci., 14 Ž1994. 1358–1365. Globus, M.Y.T., Ginsberg, M.D. and Busto, R., Excytotoxic index – a biochemical marker of selective vulnerability, Neurosci. Lett., 127 Ž1991. 39–42. Gundersen, V., Shuplliakov, O., Brodin, L., Ottersen, O.P. and Storm-Mathisen, J., Quantification of excitatory amino acid uptake at intact glutamatergic synapses by immunocytochemistry of exogenous D-aspartate, J. Neurosci., 15 Ž1995. 4417–4428. Hartley, D.M. and Choi, D.W., Delayed rescue of NMDA receptormediated neuronal injury in cortical culture, J. Pharmacol. Exp. Ther., 250 Ž1989. 752–758. Ikeda, M., Nakazawa, T., Abe, K., Kaneko, T. and Yamatsu, K., Extracellular accumulation of glutamate in the hippocampus induced by ischemia is not calcium dependent – in vitro and in vivo evidence, Neurosci. Lett., 96 Ž1989. 202–206. Katayama, Y., Katayama, T., Tamura, T., Hovda, D.A., Becker, D.P. and Tsubokawa, T., Calcium-dependent glutamate release concomitant with massive potassium flux during cerebral ischemia in vivo, Brain Res., 558 Ž1991. 136–140. Kauppiten, R.A., McMahon, H.T. and Nicholls, D.G., Ca2q-dependent and Ca2q-independent glutamate release, energy status and cytosolic free Ca2q concentrations in isolated nerve terminals following metabolic inhibition: possible relevance to hypoglycaemia and anoxia, Neuroscience, 1 Ž1988. 175–182. Lees, G.J., Inhibition of sodium-potassium-ATPase: a potentially ubiquitous mechanism contributing to central nervous system neu-
ropathology, Brain Res. ReÕ., 16 Ž1991. 283–300. w23x Leikieffre, D. and Meldrum, B.S., The pyrimidine-derivate, BW1003C87, protects CA1 and striatal neurones following transient severe forebrain ischaemia in rats. A microdialysis and histological study, Neuroscience, 56 Ž1993. 93–99. w24x Lynch III, J.J., Yu, S.P., Canzoniero, L.M.T., Sensi, S.L. and Choi, D.W., Sodium channel blockers reduce oxygen-glucose deprivationinduced cortical neuronal injury when combined with glutamate receptor antagonists, J. Pharmacol. Exp. Ther., 273 Ž1995. 554–560. w25x Madl, J.E. and Burgesser, K., Adenosine triphosphate depletion reverses sodium-dependent, neuronal uptake of glutamate in rat hippocampal slices, J. Neurosci., 13 Ž1993. 4429–44. w26x Miller, B., Sarantis, M., Traynelis, S.F. and Attwell, D., Potentiation of NMDA currents by arachidonic acid, Nature, 355 Ž1992. 722– 725. w27x Obrenovitch, T.P. and Richards, D.A., Extracellular Neurotransmitter changes in cerebral ischaemia, CerebroÕasc. Brain Metab. ReÕ., 7 Ž1995. 1–54. w28x Pellegrini-Giampietro, D.E., Cherici, G., Alesiani, M., Carla, V. and Moroni, F., Excitatory amino acid release and free radical formation may cooperate in the genesis of ischemia-induced neuronal damage, J. Neurosci., 10 Ž1990. 1035–1041. w29x Pratt, J., Rataud, J., Bartot, F., Roux, M., Blanchard, J.C., Laduron, P.M. and Stutzmann, J.M., Neuroprotective actions of riluzole in rodent models of global and focal cerebral ischaemia, Neurosci. Lett., 140 Ž1992. 225–230. w30x Simonato, M., Bregola, G., Bianchi, C. and Beani, L., Effect of arachidonic acid on w 3 Hx-D-aspartate outflow in the rat hippocampus, Neurochem. Res., 19 Ž1994. 195–200. w31x Szatkowski, M. and Attwell, D., Triggering and execution of neuronal death in brain ischemia: two phases of glutamate release by different mechanisms, Trends Neurosci., 17 Ž1994. 359–365. w32x Tallarida, R.J. and Murray, R.B., Manual of Pharmacologic Calculations with Computer Programs, Springer-Verlag, 1986. w33x Taylor, C.P., Naq currents that fail to inactivate, Trends Neurosci., 16 Ž1993. 455–460. w34x Taylor, C.P., Burke, S.P. and Weber, M.L., Hippocampal slices: glutamate overflow and cellular damage from ischemia are reduced by sodium-channel blockade, J. Neurosci. Methods, 59 Ž1995. 121– 128. w35x Urenjak, J. and Obrenovitch, T.P., Pharmacological modulation of voltage-gated Naq channels: a rational and effective strategy against ischemic brain damage, Pharmacol. ReÕ., 48 Ž1996. 2167. w36x Vazquez, E., Herrero, I., Miras-Portugal, M.T. and Sanchez-Prieto, J., Role of arachidonic acid in the facilitation of glutamate release from rat cerebrocortical synaptosomes independent of metabotropic glutamate receptor responses, Neurosci. Lett., 174 Ž1994. 9–13. w37x Volterra, A., Trotti, D., Tromba, C., Floridi, S. and Racagni, G., Glutamate uptake inhibition by oxygen free radicals in rat cortical astrocytes, J. Neurosci., 14 Ž1994. 2924–2932. w38x Westerink, B.H.C., Damsma, G. and De Vries, J.B., Effect of ouabain applied by intrastriatal microdialysis on the in vivo release of dopamine, acetylcholine, and amino acids in the brain of conscious rats, J. Neurochem., 52 Ž1989. 705–712.