Inhibition by ascorbic acid of NMDA-evoked acetylcholine release in rabbit caudate nucleus*

Received July 23, 1993/Accepted August 31, 1993




The interaction of ascorbic acid and of dehydroascorbic acid with acetylcholine (ACh) release in rabbit caudate nucleus was investigated. The presence of ascorbic acid in the superfusion medium decreased the release of ACh evoked by N-methyl-D-aspartate (NMDA), but not by electrical stimulation. The pH of the buffer was always maintained at 7.4. Inhibition occurred even at 570 μmol/l ascorbic acid, a concentration which is widely employed in transmitter release experiments. In vivo this concentration may be reached extracellularly in brain tissue. Both ascorbic acid and dehydroascorbic acid inhibited the NMDA, ;-evoked ACh release to the same degree in a non-competitive manner. The nearly identical action of ascorbic acid and dehydroascorbic acid makes a mode of action by lipid peroxidation or by redox phenomena unlikely. The mechanism of action underlying the described effects is unknown.


Key words: Ascorbic acid - Dehydroascorbic acid - Acetylcholine release - NMDA receptor - Rabbit caudate nucleus




Relatively high concentrations of ascorbic acid have been found in the mammalian brain (Schenk et al. 1982). The whole tissue concentration of ascorbic acid in the rat brain is about 1 -2 mmol/1, whereas the extracellular concentration is usually maintained in the range of 200-400 μmol/1. Under pathological conditions, like ischemia or trauma, the extracellular concentration of ascorbic acid is increased (Schenk et al. 1982). Trauma and ischemia are also associated with an excessive N-methyl-D-aspartate (NMDA) receptor stimulation (O'Neill et al. 1984). Therefore, a pathophysiological interaction of ascorbic acid and NMDA receptor complex seems possible. Ascorbic acid (500 μmol/l) inhibited both the binding of glutamate to the NMDA binding site and that of thienylcyclohexylpiperidine to the NMDA receptor gated channel (Majewska et al. 1990). Thus, a depression by ascorbic acid of the function of NMDA receptors should reflect both a competitive and a non-competitive antagonism. Majweska et al. (1990), however, did not discriminate between these types of antagonism when they observed a decrease of NMDA-induced inward currents by ascorbic acid. In their model, the lack of effect of dehydroascorbic acid suggested that redox mechanisms may comprise NMDA receptor regulatory systems.


In transmitter release experiments buffers containing up to 570 μmol/l ascorbic acid are commonly employed, but the effects of ascorbic acid especially on NMDA-induced release have not been adequately considered until now. The purpose of the present study was to evaluate the effects of ascorbic acid and of dehydroascorbic acid on NMDA-stimulated acetylcholine (ACh) release in slices of rabbit caudate nucleus.




Fresh slices of the caudate nucleus of rabbits (0.3 mm thick, 2-3 mm diameter, 2.5-4 mg wet weight) were incubated at 37°C for 30 min with [3H]-choline (specific activity 85.1 Ci/mmol; NEN, Dreieich, Germany) at a concentration of 0.1 μmol/l. They were then superfused at 37 °C with a Mg++-free buffer (mmol/l: NaCI 118; KCI14.8; CaCI2, 2.6; NaHC03 25; KH2PO4 1.2; glucose 11), saturated with 5% CO2/95% O2 and adjusted to pH 7.4 with NaOH (1 mol/l) if necessary. For instance, saturation with 5% CO2/95% O2 decreased the original pH from 8.40 to 7.38 which was not further changed by addition of ascorbic acid er dehydroascorbic acid up to 570 μmol/l. With 5.7 mmol/l ascorbic acid, however, the pH was decreased to 7.23 and, therefore, was readjusted to 7.40 with NaOH (1 mol/1). The superfusate was collected in 5-min samples after a 50 min equilibration period. Each slice was stimulated twice, either by addition of NMDA (Sigma, München, Germany) or electrically (2 min, rectangular pulses, 2 ms, 3 Hz, 5 V/cm, 24 mA) after 60 min (S1) and 100 min (S2) of superfusion. The fractional rate of [3H]-outflow and stimulation-evoked [3H]-overflow were expressedd as percentage of the [3H]-tissue content at the onset of the respective collection period. The effects of ascorbic acid, dehydroascorbic acid (Sigma, München, Germany) or AP-5 ((±)-2-amino-5-phos- phopentanoic acid; RBI, Natick, Mass., USA), added before S2 were evaluated by calculating the ratio S2/S1 of the [3H]-overflow evoked during the two stimulation periods. At the end of the experiments the tritium content of the superfusates and of the slices was determined by liquid scintillation counting. Since the stimulation-induced [3H]-overflow from brain slices prelabelled with [3H]-choline ist referred to as release of ACh (Richardson and Szerb) 1974) [3H]-overflow is referred to as ACh release in the following.




In slices superfused with ascorbic acid-free buffer and stimulated with NMDA (10 μmol/l) tritium overflow was about 2.24±0.06% (mean±SEM) at S1 and 1.66±0.07% at S2 (S2/S1: 0.74±0.02, n = 15). Both ascorbic acid (17 μmol/l-5.7 mmol/l) and dehydroascorbic acid (57 μmol/I-5.7 mmol/l) added 35 min before S2 produced a concentration-dependent inhibition of ACh release (Table 1). ACh release evoked by electrical stimulation was not affected by 570 μmol/l ascorbic acid (data not shown).

In order to find out whether ascorbic acid or dehydroascorbic acid inhibited the NMDA-evoked ACh release in a competitive and/or in an non-competitive manner, concentration-response curves for NMDA at different concentrations during S2 were determined in the presence or absence of 570 μmol/l asorbic acid or 570 μmol/l dehydroascorbic acid, respectively, added 35 min before S2. The results were evaluated by nonlinear regression analysis using the function



L: log mol/l of the NMDA concentration at S2; KD: dissociation constant between NMDA and receptor; (S2/S1)max: maximal effect of NMDA on ACh release (Lupp et al. 1992). The application of this mathematical model was in line with the most simple assumption of direct proportionality between occupation of receptors by NMDA and evoked release of ACh, which was justified by a so-called slope parameter around unity (not shown).



Release of ACh (% of Controls) in the presence of

ascorbic acid dehydroascorbic acid

  17 μmol/l



  57 μmol/l



570 μmol/l



    5.7 mmol/l




Table 1. Inhibition by ascorbic acid or by dehydroascorbic acid of NMDA-evoked release of ACh in rabbit caudate nucleus tissue


The slices were stimulated twice with NMDA (10 μmol/I), added to the medium for 2 min after 60 min (S1) and after 100 min (S2). Ascorbic acid or dehydroascorbic acid at the concentrations indicated was present from 35 min before S2 onwards. The effects of ascorbic acid and dehydroascorbic acid were evaluated as S2/S1 and expressed as % of the mean of controls (n = 4-13). *P<0.05; **P< 0.01 compared to the control mean.


Fig. 1. Effects of ascorbic acid and of AP-5 on the concentration-response curve of NMDA-evoked release of ACh in rabbit caudate nucleus tissue. NMDA was added to the medium for 2 min after 60 min (10 μmol/l, S1) and after 100 min (various concentrations, S2). AP-5 was added from 30 min and ascorbic acid from 35 min before S2 onwards. The values represent means±SEMs from 6-12 experiments at each concentration of NMDA. The sigmoidal curves represent the best nonlinear fit to all individual values of S2/S1


The condition of a slope parameter around unity may exclude spare receptors in the system under investigation and, therefore, suggested to use the term KD instead of EC50 (see Feuerstein et al. 1993). In the presence of ascorbic acid, the maximal effect of NMDA was significantly (P< 0.05) decreased by 25.1±5.7% as compared to the absence of ascorbic acid (Fig. 1), whereas the estimates of KD, or of the negative logarithm pKD, were very similar: pKD = 4.03±0.08, presence of ascorbic acid; pKD = 3.94±0.06; absence of ascorbic acid. Dehydroascorbic acid (570 μmol/l) produced a nearly identical "non-competitive" flattening out of the NMDA concentration-response curve, compared to that obtained in the presence of ascorbic acid (570 μmol/l), with a significant (P<0.05) decrease by 18.3±5% of the maximal effect (graph not shown). Again, no change of the pKD (3.95±0.06), was observed. When the inhibition by ascorbic acid or by dehydroascorbic acid of the NMDA-evoked ACh release was expressed in percent of the effect obtained at each concentration of NMDA given alone, no increase in the inhibitory effect by 570 μmol/l of ascorbic acid or dehydroascorbic acid with increasing NMDA concentrations was observed (data not shown). Correspondingly, the percentage compression in the direction of the y-axis (by 25.1%) of the fitted concentration-response sigmoid by ascorbic acid remained identical at each point of the curve (Fig. 1). The same was true for the curve obtained in the presence of dehydroascorbic acid (not shown).


For comparison, the competitive NMDA antagonist AP-5 (100 μmol/l) shifted the concentration response curve of NMDA to the right in a parallel fashion, but did not attenuate its maximal effect (Fig. 1, pKD = 3.94±0.06, absence of AP-5; pKD = 3.09±0.09, presence of AP-5; pA2 = 4.78±0.13). The Law of Error Propagation was considered when the standard deviation of the pA2 or of a percentage value was calculated, see above.




The presence of NMDA receptors on rat and rabbit striatal cholinergic interneurons and their characterization has been reported by Scatton and Lehmann 1982, and Lupp et al. 1992, respectively. Ascorbic acid inhibited the NMDA-evoked ACh release in a concentration-dependent manner with a significant inhibition even at 570 μmol/l, a concentration commonly employed in experiments on transmitter release from brain slices. The extent of this inhibition at 570 μmol/l (22.9±3.5%) corresponded well with the estimate of the reduction (25.1±5.7%) of (S2/S1)max in the nonlinear regression analysis. The inhibitory effect of ascorbic acid seems to be specific for the NMDA receptor, since ACh release evoked by electrical stimulation was not affected by ascorbic acid. It has been reported that acidosis attenuates NMDA receptor activation (Giffard et al. 1990). The pH of our buffer solution, however, was significantly dereased by ascorbic acid or dehydroascorbic acid only at the highest concentration used (5.7 mmol/l, see methods). Since we scrutinized or readjusted the pH to the initial value of 7.4 we can reject lowering of the pH as the explanation for inhibition of NMDA response by ascorbic acid.


Our data do not support inhibition of binding of NMDA by ascorbic acid (Majewska et al. 1990) as mechanism of inhibition of ACh release. If ascorbic acid reduced the affinity of NMDA to the receptor, as it is the case with AP-5, this interaction would represent a competitive antagonism of ACh release with a typical increase of the KD value, to be observed possibly together with an additional non-competitive mechanism. Both ascorbic acid and dehydroascorbic acid, however, only depressed the maximum effect, but left the KD virtually unchanged. In contrast to typical "NMDA channel blockers", as dizolcipine or memantine, the use-dependent action of which could be clearly recognized in our experimental model (Lupp et al. 1992), both ascorbic acid and dehydroascorbic acid acted as pure non-competitive antagonists, i.e. their antagonist effects did not increase in parallel with increasing NMDA concentrations. This finding argues against an action at the dizolcipine binding site within the NMDA-gated ion channel. Our results are rather compatible with a mechanism of action unrelated to the binding of NMDA or dizolcipine. Since both the reduced and the oxidized form of ascorbic acid acted very similarly, lipid peroxidation or other redox phenomena may not explain the observed effects. The exact mechanism of action underlying these effects, however, remains unknown.

The present findings demonstrate that ACh release in rabbit caudate nucleus induced by NMDA is inhibited both by ascorbic acid and by dehydroascorbic acid. Inhibition occurred even at concentrations of ascorbic acid which are commonly employed in transmitter release experiments. Apart from this experimental condition which may diminish NMDA receptor-mediated effects, concentrations of ascorbic acid in the range of 570 μmol/l may prevail during ischemia or trauma in the extracellular space of brain tissue. It may be speculated that in such a pathophysiological condition ascorbic acid may attenuate glutamate neurotoxicity mediated by NMDA receptors.




Feuerstein TJ, Sauermann W, Allgaier C, Singer CA (1993)

Mathematical modelling and quantification of the autoinhibitory feedback control of noradrenaline release in brain slices.

Naunyn-Schmiedeberg's Arch Pharmacol 347:171-179


Giffard RG, Monyer H, Christine CW, Choi DW (1990)

Acidosis reduces NMDA receptor activation, glutamate neurotoxicity, and oxygenglucose deprivation neuronal injury in cortical cultures.

Brain Res 506:339-342


Lupp A, Lücking CH, Koch R, Jackisch R, Feuerstein TJ (1992)

Inhibitory effects of the antiparkinsonian drugs memantine and amantadine on N-methyl-D-aspartate-evoked acetylcholine release in the rabbit caudate nucleus in vitro.

J Pharmacol Exp Ther 263:777-724


Majweska MD, Bell JA, London ED (7990)

Regulation of the NMDA receptor by redox phenomena: inhibitory rote of ascorbate.

Brain Res 537:328-332


O'Neill RD, Fillenz M, Sundstrom L, Rawlins JNP (1984)

Voltammetrically monitored brain ascorbate as an index of excitatory amino acid release in the unrestrained rat.

Neurosci Lett 52:227-233


Richardson IW, Szerb JC (1974)

The release of labelled acetlycholine and choline from cerebral cortical slices stimulated electrically.

Br J Pharmacol 52:499-507


Scatton B, Lehmann J (1982)

N-methyl-D-aspartate-type receptors mediate striatal 3H-acetlycholine release evoked by excitatory amino acids.

Nature 297:422-424  


Schenk JO, Miller E, Gaddis R, Adams RN (1982)

Homeostatic control of ascorbate concentration in CNS extracellular fluid.

Brain Res 253:353-356  



*Erschienen im "Naunyn-Schmiedeberg's Arch Pharmacol", 1993, 348:549-551

Autoren: Dr. med. Günter Lang1, Dr. med. Thomas J. Feuerstein1, Dr. med. Günter Weinheimer1, Dr. med. Thomas Ginap1

Dr. med. G. Reinhard Rossner2

1Neuropharmakologisches Labor der Neurologischen Universitätsklinik, c/o Gödecke AG, Mooswaldallee 1 - 9, D-79090 Freiburg

2Institut für Medizinische Biometrie und Medizinische Informatik, Stefan-Meier-Strasse 26, D-79104 Freiburg