Full text of DOI:10.1038/sj.bjp.0703598

British Journal of Phar
m
acology (2000) 131, 617 ± 623
ã
2000 Macmillan Publishers Ltd All rights reserved 0007 ± 1188/00 $15.00
www.nature.com/bjp
The e€ect of nucleotides and adenosine on stimulus-evoked
glutamate release from rat brain cortical slices
1,2Gillian C. Bennett & *,1Michael R. Boarder
1Department of Cell Physiology and Pharmacology, University of Leicester, Leicester LE1 9HN
1
Evidence has previously been presented that P1 receptors for adenosine, and P2 receptors for
nucleotides such as ATP, regulate stimulus-evoked release of biogenic amines from nerve terminals
in the brain. Here we investigated whether adenosine and nucleotides exert presynaptic control over
depolarisation-elicited glutamate release.
2
Slices of rat brain cortex were perfused and stimulated with pulses of 46 mM K+ in the presence
of the glutamate uptake inhibitor L-trans-pyrrolidine-2,4-dicarboxylic acid (0.2 mM). High K+
substantially increased e‚ux of glutamate from the slices.
3
Basal glutamate release was unchanged by the presence of nucleotides or adenosine at
concentrations of 300 mM.
Adenosine, ATP, ADP and adenosine 5'-O-(3-thiotriphoshate) at 300 mM attenuated depolarisa-
4
tion-evoked release of glutamate. However UTP, 2-methylthio ATP, 2-methylthio ADP, and a,b-
methylene ATP at 300 mM had no e€ect on stimulated glutamate e‚ux.
5
Adenosine deaminase blocked the e€ect of adenosine, but left the response to ATP unchanged.
6
The A1 antagonist 8-cyclopentyl-1,3-dipropylxanthine antagonised the inhibitory e€ect of both
adenosine and ATP.
Cibacron blue 3GA inhibited stimulus-evoked glutamate release when applied alone. When
7
cibacron blue 3GA was present with ATP, stimulus-evoked glutamate release was almost eliminated.
However, this P2 antagonist had no e€ect on the inhibition by adenosine.
8
These results show that the release of glutamate from depolarised nerve terminals of the rat
cerebral cortex is inhibited by adenosine and ATP. ATP appears to act directly and not through
conversion to adenosine.
British Journal of Pharmacology (2000) 131, 617 ± 623
Keywords: P2 receptors; P2Y receptors; adenosine receptors; glutamate release; ATP; presynaptic inhibition
Abbreviations: ATPgS, adenosine 5'-O-(3-thiotriphoshate); DPCPX, 8-cyclopentyl-1,3-dipropylxanthine; 2MeSADP, 2-
methylthio ADP; 2MeSATP, 2-methylthio ATP; OPA, o-opthaldehyde; PDC, L-trans-pyrrolidine-2,4-
dicarboxylic acid
Introduction
It has been accepted for many years that ATP acts as a
neurotransmitter (Burnstock, 1972; 1976). This is largely a
result of studies on the autonomic neuroe€ector junction,
where ATP is a co-transmitter, and may in some cases act as
the primary transmitter (Burnstock et al., 1972; Meldrum &
Burnstock, 1983; von Kugelgen et al., 1994a). Our under-
standing of the neurotransmitter role for ATP developed
further with evidence for prejunctional modulation of
neurotransmitter release from autonomic terminals by adenine
nucleotides and adenosine acting on presynaptic receptors
(Paton, 1981; Wicklund et al., 1985). UTP has also been shown
to directly regulate noradrenaline release from sympathetic
neurons (Boehm, 1998; Norenberg et al., 2000).
On release into the synaptic cleft ATP is rapidly broken
down to adenosine. A major issue in these studies is whether
ATP acts directly at presynaptic P2 receptors, or is converted
to adenosine, which then acts on presynaptic P1 receptors.
While inhibitory P1 receptors are well known (Fredholm &
Dunwiddie, 1988), there is evidence that ATP acts directly at
presynaptic P2 receptors in the autonomic nervous system (e.g.
Shinozuka et al., 1988; Kurz et al., 1993; Allgaier et al., 1994;
von Kugelgen et al., 1995).
P2 receptors for nucleotides are divided into two families:
ionotropic P2X receptors, and
G protein-coupled P2Y
receptors (Fredholm et al., 1997; Boarder & Hourani, 1998).
P2X receptors are essentially receptors for ATP, whereas the
P2Y family includes receptors for both purine and pyrimidine
nucleotides (ATP, ADP, UTP and UDP). In the brain ATP is
known to be a fast excitatory neurotransmitter, acting directly
at postsynaptic P2X receptors (e.g. Edwards et al., 1992). P2Y
receptors are believed to be widespread in the brain (Webb et
al., 1994), where they may have an excitatory role, acting by
reducing K+ channel opening (Shen & North, 1993; Nieber et
al., 1997). Additionally, in rat brain preparations P2Y
receptors have been shown to play a role in presynaptic
inhibition of release of noradrenaline, 5-hydroxytryptamine
and dopamine (von Kugelgen et al., 1994b; Koch et al., 1995;
1997a,b). There is also a report of enhanced dopamine release
on activation of P2Y1 receptors in the rat brain striatum
(Zhang et al., 1995).
*Author for correspondence at: Cell Signalling Laboratory, Depart-
ment of Biological Sciences, De Montfort University, Leicester LE1
9BH; E-mail: mboarder@dmu.ac.uk
Glutamate is the major excitatory neurotransmitter in the
central nervous system, and there is emerging evidence that its
release is subject to presynaptic regulation by P2 receptors.
Activation of P2X receptors has been shown to elicit glutamate
2 Current address: Department of Pharmacology, Cambridge
University, Tennis Court Road, Cambridge, CB2 1QJ

618
G.C. Bennett & M.R. Boarder
P2Y regulation of glutamate release
release from terminals of dorsal root ganglia in the dorsal horn
of the spinal cord (Gu & MacDermott, 1997). An electro-
physiological study on hippocampal neurons (Koizumi &
Inoue, 1997) provided evidence suggesting an inhibitory
in¯uence of presynaptic P2 receptors on glutamate release.
These ®ndings follow previous electrophysiological studies
showing attenuation of central excitatory neurotransmission
by adenosine acting at presynaptic P1 receptors (e.g. Silinsky &
Ginsborg, 1983; Manzoni et al., 1994). Therefore it is possible
that ATP acting at the presynaptic site may modulate brain
glutamate release by direct action, and also by conversion to
adenosine. In the present study we set out to directly
investigate whether glutamate release from rat brain cortical
slices is subject to regulation by adenosine and by nucleotides
such as ATP.
samples onto a reverse phase (C18) analytical column and
elution with an acetonitrile/methanol gradient. Detection was
with a CMA 280 ¯uorometric detector. The amino acid peaks
were identi®ed using standards.
A standard curve was
constructed using known concentrations of glutamate. A
straight-line relationship between glutamate concentration and
¯uorometric signal was observed. The two methods gave
essentially the same results as shown in Figure 1, in which S1
and S2 peaks of glutamate release were analysed by both
methods. The apparent basal levels varied depending on the
method used, but the area under each of the two peaks was the
same for each method. Data analysis included deduction of
basal release, so the outcome of the experiments was the same
regardless of the method used.
Data analysis
Methods
Control and treatment pairs were drawn from the same
experiment i.e. from the same group of slices. Areas under
curves and determinations based on standard curves were
calculated with Graphpad Prism. Each experiment was
performed at least four times, and statistical analysis was by
analysis of variance with Bonferroni's correction for multiple
comparisons.
Preparation and perfusion of brain slices
Adult Wistar rats were killed, the cortices dissected on ice and
cross-chopped into 3506350 mm prisms which were washed
three times in 20 ml Krebs gassed with 95% O2/5% CO2. Slices
were preincubated with gentle shaking at 378C for 80 min with
four changes of gassed medium. One ml of gravity packed
slices was then placed in each superfusion chamber and
perfused with pre-gassed Krebs (378C) at 0.7 ml min71 for
45 min before beginning drug treatments. Two-minute
fractions of the perfusate were then collected.
Results
E€ect of inhibiting glutamate re-uptake with
L-trans-pyrrolidine-2,4-dicarboxylic acid (PDC)
Stimulation of perfused slices
Various concentrations of PDC were tested in the superfusion
system for an e€ect on basal levels of glutamate release, as well as
glutamate e‚ux in response to 46 mM K+ (Figure 2). The left-
hand traces in each panel of Figure 2 show a small
depolarisation-elicited increase in glutamate e‚ux in the
absence of PDC; depolarisation was then repeated for the same
slice in the presence of PDC. At 0.7 or 0.5 mM, PDC caused an
increase in basal glutamate e‚ux, whereas 0.2 or 0.09 mM PDC
was without e€ect (consistent with Volterra et al., 1996). The
slices were then exposed to a 2 min pulse of K+ in the presence of
A double stimulation protocol (S1/S2) was adopted. In a typical
experiment 0.2 mM of the glutamate uptake inhibitor, L-trans-
pyrrolidine-2,4-dicarboxylic acid (PDC) was present through-
out. Three 2 min samples were collected and their glutamate
concentration averaged to give an estimation of basal e‚ux,
followed by a 2 min pulse of 46 mM KCl (S1), a recovery
period of 22 min, and a second KCl pulse (S2) of 2 min.
Agonists (adenosine or nucleotides) were included for 2 min
prior to and during S2; antagonists and adenosine deaminase
were introduced, as indicated, 4 min prior to and during S2.
Two-minute fractions were collected during and either side of
both S1 and S2. Glutamate concentrations in the fractions were
determined, and S2/S1 ratios calculated from the area under the
S1 and S2 peaks after deduction of basal release. Conclusions
were drawn by comparing S2/S1 ratios in the presence of test
compounds during S2, with the ratios obtained in controls run
in the same experiment in which no test compounds were
introduced.
Measurement of glutamate
Two methods were used to measure glutamate. The ®rst
procedure measures the glutamate-dependent conversion of
NADP to NADPH. A 475 ml aliquot of each fraction was
incubated with 30 units of glutamate dehydrogenase (10 ml)
and 15 ml NADP (1 mM ®nal concentration) for 10 min at
378C. Fluorescence was measured in a SPEX ¯uorimeter
(lEx=366, lEm 432 nm). In the second method, glutamate
concentration was determined by HPLC with pre column o-
opthaldehyde (OPA) derivatisation. Fifty-ml aliquots of
perfusate fractions were reacted with an equal volume of OPA
(50 mg in 20 ml of 5% methanol, 0.5% b-mercaptoethanol in
1 M boric acid, pH 9.5) for 2 min before injection of the
Figure 1 Stimulation of glutamate release by 2 consecutive pulses of
46 mM K+. Three basal samples were collected before slices were
depolarised using a 2 min pulse of 46 mM K+ as indicated by the
®rst bar. This depolarisation was repeated 30 min later. PDC
(0.2 mM) was present throughout the experiment. The concentration
of glutamate released from the slices upon depolarisation was
calculated using the ¯uorometric and the HPLC methods for the
same samples and plotted against time. These are typical examples of
results from a single experiment.
British Journal of Pharmacology, vol 131 (3)

G.C. Bennett & M.R. Boarder
P2Y regulation of glutamate release
619
PDC to determine the e€ect of the uptake inhibitor on
depolarisation-elicited glutamate e‚ux. At each concentration
of PDC tested, the stimulus-evoked glutamate release was
greatly enhanced. A concentration of 0.2 mM PDC was chosen
for all subsequent experiments, since this enhanced stimulus-
evoked glutamate release without a€ecting basal e‚ux (Figure
2C). Typical pro®les of stimulus-evoked glutamate release with
two sequential pulses of high K+ (S1 and S2), in the presence of
0.2 mM PDC are shown in Figure 1.
(Figure 3B, Table 1). Like ATP, ADP (300 mM) signi®cantly
reduced glutamate release during S2 (Figure 4A, Table 1),
whereas a,b-methylene ATP, 2MeSATP, or 2MeSADP (each
at 300 mM) were without e€ect (Figure 4B and Table 1).
Adenosine also e€ectively inhibited K+-evoked glutamate
release (Table 1), reducing the S2/S1 ratio by 28%. It is
therefore possible that ATP and ADP could in¯uence
glutamate release by extracellular breakdown to adenosine
and action at P1 (adenosine) receptors. To test this we
investigated the e€ect of the ATP analogue, ATPgS which is
relatively resistant to breakdown by ectonucleotidases. ATPgS
(300 mM) exerted a similar inhibitory e€ect to ATP (Table 1),
consistent with the notion that breakdown of ATP to
adenosine was not required to inhibit glutamate release.
To con®rm that the detected increase in ¯uorescence was not
due to release of ¯uorescent material or to cell disruption, caused
by a combination of KCl and PDC, two sets of controls were
carried out. In one series of experiments glutamate dehydro-
genase was omitted from the assay system; as shown in Figure
2A this eliminated the stimulus-evoked increase in ¯uorescence.
In another series, slices were stimulated with 46 mM KCl and
0.2 mM PDC in the absence of added Ca2+ and in the presence of
0.5 mM EGTA. Under these conditions stimulus-evoked release
E€ect of adenosine deaminase
To further investigate this, the enzyme adenosine deaminase was
introduced into the superfusate 2 min prior to the inclusion of
adenosine or ATP in S2. If formation of adenosine was required
for the action of ATP, the response should be inhibited by the
presence of adenosine deaminase in the superfusion bu€er.
Adenosine deaminase alone (2 u ml71) increased stimulus-
evoked release in each experiment (i.e. higher S2/S1 ratio), but
when results were pooled across experiments this e€ect was not
signi®cant. In the presence of the enzyme the ability of adenosine
(300 mM) to inhibit glutamate release during S2 was eliminated.
was reduced to 7.2+4.2% of that seen in the presence of Ca2+
.
E€ect of P2 receptor agonists and adenosine on
depolarization-evoked glutamate release
We tested the e€ect of a range of nucleotides and adenosine on
depolarization-evoked glutamate release. None of these
compounds had an e€ect on basal glutamate release. This can
be seen for ATP, UTP, ADP and a,b-methylene ATP in the
lower panels of the traces in Figures 3 and 4, where the
compounds were included alone in the superfusate for 2 min
prior to depolarisation.
Addition of ATP (300 mM) lowered the depolarisation-
evoked release of glutamate by 43% (Figure 3A, Table 1). In
contrast, 300 mM UTP had no e€ect on glutamate release
However, in the presence of adenosine deaminase (2 u ml71
)
ATP (300 mM) was still able to attenuate glutamate release.
Expressed as S2/S1 ratios pooled across experiments (n=4 ± 10)
the results were: control, 0.77+0.05; adenosine deaminase
alone, 1.06+0.12+; adenosine deaminase plus adenosine,
1.10+0.07; adenosine deaminase plus ATP, 0.65+0.09*; (+not
Figure 2 E€ect of PDC on stimulus-evoked release of glutamate. Slices were depolarised with two consecutive pulses of 46 mM
KCl, with PDC present during the second pulse, as indicated by the bars. Di€erent ®nal concentrations of PDC were used in each of
the four experiments shown: (A) 0.7 mM; (B) 0.5 mM; (C) 0.2 mM; (D) 0.09 mM. Glutamate was assayed by the glutamate
dehydrogenase method in this study, and in (A) a single set of results are shown when glutamate dehydrogenase was omitted from
the assay, and aliquots of fractions were incubated with 1 mM NADP alone. The results are from a single representative experiment.
British Journal of Pharmacology, vol 131 (3)

620
G.C. Bennett & M.R. Boarder
P2Y regulation of glutamate release
Figure 3 E€ect of ATP and UTP on stimulus-evoked glutamate release. In each case two consecutive pulses of 46 mM KCl enabled
an S2/S1 ratio of areas under peaks to be calculated, in the absence of any other agent (control ratio, upper panels) and in the
presence of nucleotides during S2 (treatment ratio, lower panels). (A) ATP (300 mM), (B) UTP (300 mM). A representative experiment
is shown here, in which each panel is data collected from aliquots from the same brain slice preparation perfused simultaneously.
Results of ratios pooled across experiments are presented in Table 1.
Figure 4 E€ect of ADP and a,b-methylene ATP on stimulus-evoked glutamate release. In each case two consecutive pulses of
46 mM KCl enabled an S2/S1 ratio to be calculated in the absence of any other agent (control ratio, upper panels), and in the
presence of nucleotides during S2 (treatment ratio, lower panels). (A) ADP (300 mM), (B) a,b-methylene ATP (300 mM). A
representative experiment is shown here, in which each panel is data collected from aliquots from the same brain slice preparation
perfused simultaneously. Results of ratios pooled across experiments are presented in Table 1.
British Journal of Pharmacology, vol 131 (3)

G.C. Bennett & M.R. Boarder
Table 1 S2/S1 ratios for nucleotides and adenosine
S2/S1 ratio
P2Y regulation of glutamate release
621
Compound
Control ratio
Treatment ratio
% of control
signi®cance
ATP
ATPgS
ADP
2MeSATP
2MeSADP
a,b-meATP
UTP
0.72+0.02
0.86+0.15
0.76+0.02
0.81+0.02
0.67+0.04
0.65+0.04
0.81+0.09
0.78+0.05
0.42+0.03
0.54+0.07
0.52+0.02
0.76+0.04
0.68+0.04
0.68+0.11
0.95+0.04
0.56+0.04
57
63
68
94
101
104
117
72
50.01
50.05
50.01
±
±
±
±
Adenosine
50.05
Rat brain cortical slices were depolarised in the presence of 0.2 mM PDC by two 2 min pulses of K+. No other pharmacological
intervention was used during either S1 or S2 in the control condition. In the treatment condition, the second depolarisation (S2) was in
the presence of the compounds shown in the table, each at a concentration of 300 mM. Data were pooled from at least four separate
experiments.
signi®cantly di€erent from control; *signi®cantly di€erent from
adenosine deaminase alone, P=0.02).
E€ect of DPCPX
To further evaluate the role of adenosine receptors in the
response to ATP we used DPCPX, an A1 antagonist. DPCPX
alone had no e€ect on glutamate release (data not shown). As
shown in Figure 5, DPCPX (10 ± 100 mM) reversed the e€ect of
adenosine (300 mM) in a concentration dependent manner.
However, DPCPX (10 and 100 mM) also reversed the inhibition
of K+-evoked glutamate release by 300 mM ATP.
E€ect of cibacron blue 3GA
von Kugelgen et al. (1994b) showed that cibacron blue 3GA
antagonised presynaptic inhibition of noradrenaline release in
the rat cortex by ATP but not by adenosine. We therefore used
this P2 antagonist to investigate further the hypothesis that
ATP and adenosine act on di€erent receptors when causing the
inhibition of stimulus-evoked glutamate release. The results
were surprising (Figure 6). Cibacron blue 3GA (30 mM) alone
signi®cantly reduced the 46 mM K+-evoked glutamate e‚ux.
Inclusion of both 300 mM ATP and 30 mM cibacron blue 3GA
had an additive inhibitory e€ect on the recorded glutamate
output. In three out of four experiments, the depolarisation-
evoked glutamate e‚ux was completely abolished by this
treatment. In contrast cibacron blue 3GA had no e€ect on the
inhibitory response to 300 mM adenosine.
Figure 5 E€ect of DPCPX on inhibition of glutamate release by
adenosine and ATP. Rat brain cortical slices were depolarised twice
using 2 min pulses of 46 mM K+. The second depolarisation was in
the presence of the compounds indicated. DPCPX was introduced
4 min prior to S2, and 300 mM adenosine or 300 mM ATP, 2 min prior
to S2. S2/S1 ratios were calculated from the areas under the peaks and
plotted as a bar chart (mean+s.e.mean, n=4). By one-way ANOVA
the e€ect of DPCPX in the presence of adenosine was signi®cant at
P50.001 and in the presence of ATP at P50.05. Using Neumann-
Keul's post-test, adenosine and ATP alone were di€erent from
controls (*P50.05) and 100 mM DPCPX in the presence of adenosine
or ATP were di€erent from adenosine or ATP alone (**P50.001).
Discussion
inhibitor (Swanson et al., 1997). The observation that PDC
greatly enhances the e‚ux of glutamate in response to 46 mM
K+ indicates that the uptake mechanisms within the slices
remain e€ective under the experimental conditions used. The
results described here established a concentration of PDC
which enhanced stimulus-evoked glutamate e‚ux, with no
e€ect on basal levels of release, enabling PDC to be used in
Glutamate is the dominant excitatory neurotransmitter in the
brain, a€ecting all aspects of brain function. It seems likely
that presynaptic modulation of glutamate release may di€er
between pathways or brain regions, providing a route for
speci®c drug manipulation of glutamatergic in¯uence. In this
study we set out to ask whether adenosine and nucleotides
regulate glutamate release from rat brain cortex, following
reports that presynaptic P2Y and adenosine receptors regulate
biogenic amine release from the brain (von Kugelgen et al.,
1994b; Koch et al., 1995; 1997a,b).
The increase in glutamate e‚ux in response to depolarisa-
tion with 46 mM K+ was modest in the absence of inhibition of
uptake. In rat brain, three glutamate transporter proteins have
been isolated; GLAST and GLT-1 are mainly glial and
EAAC1 is mainly neuronal. Under the conditions used in
these experiments, PDC should not discriminate between these
proteins and therefore acts as a non-speci®c glutamate uptake
subsequent experiments in
a manner analogous to the
widespread use of uptake inhibition in the study of biogenic
amine release from perfused slices.
The abolition of K+-evoked glutamate e‚ux in the absence
of extracellular Ca2+ in this system is consistent with this being
an exocytotic process. We con®rmed that high K+ treatment
did not cause release as a result of cell disruption by analysis of
the serine content of the superfusate samples. Serine is a
marker of cell lysis and therefore would not be present at
appreciable levels in the superfusate unless cell disruption was
occurring. Serine levels did not rise signi®cantly above basal
levels during treatment with 46 mM K+ (not shown).
British Journal of Pharmacology, vol 131 (3)

622
G.C. Bennett & M.R. Boarder
P2Y regulation of glutamate release
Regulation of glutamate e‚ux from brain slices may be at
the site of release in the nerve terminal. However, if intact
neurons are found within the slice it is conceivable that
regulation is by action remote from the terminal, at the cell
body. In these experiments the stimulus for release was
depolarisation with 46 mM K+ in the medium, e€ectively
depolarising the membrane potential at the nerve terminal for
the duration of the high K+ pulse. This means that the site of
action of adenosine and ATP in regulation of glutamate release
in these experiments is likely to be at the nerve terminal close
to the site of release.
Both ATP and adenosine, at 300 mM, inhibited stimulus-
evoked glutamate e‚ux. Attempts to generate concentration
response curves did not yield suitably reproducible e€ects
below 100 mM when data were pooled across experiments (data
not shown). Consequently we were unable to derive EC50
values for agonists.
As ATP and adenosine exerted similar e€ects, several lines
of investigation were pursued to ascertain whether the e€ect of
ATP was direct, or via breakdown by ectonucleotidases to
adenosine. Ectonucleotidases have been shown to be abundant
and highly active in brain tissue (James & Richardson, 1993;
Cunha et al., 1996). Several ®ndings here provide evidence that
ATP is acting directly, and not by conversion to adenosine.
Firstly the stable ATP analogue ATPgS, which is relatively
resistant to breakdown, had the same e€ect as ATP. Secondly,
whilst the addition of adenosine deaminase completely
abolished the inhibition of release by adenosine, it did not
attenuate the inhibitory e€ect of ATP. Thirdly, cibacron blue
3GA has di€erent e€ects on the responses to adenosine and
ATP (see below).
It seems likely, therefore, that ATP is acting directly on a
receptor which inhibits glutamate release, so we investigated
whether a previously characterized P2 receptor is responsible.
UTP had no e€ect on glutamate release, e€ectively eliminating
a role for P2Y2, P2Y4 and P2Y6 receptors at which UTP is an
agonist. Similarly the lack of e€ect of 2MeSATP and
2MeSADP excludes P2Y1 receptors. The agonist potency
order of P2Y11 receptors is ATP42-MeSATP44ADP. As
ATP and ADP gave a similar magnitude of inhibition, and
2MeSATP was ine€ective, it seems unlikely that P2Y11
receptors mediate the nucleotide-induced inhibition of
depolarisation-evoked glutamate e‚ux. The failure of both
2MeSATP and a,b-meATP to produce an inhibitory e€ect
rules out all but the P2X7 member of the P2X receptor family.
Furthermore, activation of any P2X receptor would be
expected to increase neurotransmitter release, in contrast to
the inhibition seen here. The results presented therefore are not
consistent with the involvement of any of the currently cloned
and characterized P2 receptor types. However, as full curves to
the agonists could not be constructed, it is impossible to
produce an agonist potency order for this system.
The A1 receptor antagonist, DPCPX was found to reverse
the e€ects of both adenosine and ATP in the regulation of
glutamate release. This might be taken as evidence that ATP is
acting via conversion to adenosine and stimulation of A1
receptors. However, an alternative explanation was o€ered for
an apparently similar ®nding reported by von Kugelgen et al.
(1994b), who showed a presynaptic inhibition of noradrenaline
release in the rat cortex by ATP and adenosine. The e€ects of
both adenosine and ATP were blocked by DPCPX, while the P2
antagonist, cibacron blue 3GA was only e€ective against ATP.
This led to the suggestion that two distinct receptors were
present; an atypical A1 receptor that responds to both adenosine
and ATP, and a P2 receptor at which ATP could act but not
adenosine. However, it should be noted that our results di€er in
several important respects from those of von Kugelgen (1994b).
Firstly, cibacron blue 3GA antagonized the response to ATP in
their report but had the opposite e€ect in ours; secondly, they
found DPCPX to be e€ective at much lower concentrations
than those used here; thirdly, 2MeSATP mimicked ATP in their
study but was without e€ect in the present study.
Cunha et al. (1994) reported relevant ®ndings in synapto-
somes isolated from rat cerebral cortex. They found that the
e€ect of ATP on [3H]-acetylcholine release was not prevented
by adenosine deaminase. The e€ect of both ATPgS and
adenosine in the cerebral cortex was prevented by DPCPX.
This report also suggests the presence of presynaptic receptors
for ATP which are sensitive to DPCPX.
In view of these studies we decided to investigate the e€ect
of cibacron blue 3GA in the stimulus-evoked glutamate release
system. We anticipated that it would either inhibit the e€ect of
ATP but not adenosine (as in von Kugelgen et al., 1994b), or
do nothing. Surprisingly it did neither. Alone, cibacron blue
3GA inhibited depolarisation-evoked glutamate e‚ux. The
response to adenosine was not a€ected by the presence of
cibacron blue 3GA. When this compound was added together
with ATP their inhibitory e€ects were essentially additive,
almost abolishing the K+-evoked e‚ux of glutamate. This
type of e€ect of cibacron blue 3GA, or of its parent compound
reactive blue 2, has not previously been reported. It may be the
result of inhibition by cibacron blue 3GA of ectonucleotidases
(Koch et al., 1997a), as reported for reactive blue 2 (Bultmann
et al. 1999), leading to an increase in activation of inhibitory
presynaptic P2 receptors by endogenous ATP. This could
account for the inhibitory e€ect of cibacron blue 3GA alone. It
may also reduce the breakdown of exogenous ATP, thereby
contributing to a more e€ective delivery to the site of the
receptor. If ATP which is degraded by ectonucleotidases in
brain slices is rapidly converted to adenosine, and ATP and
adenosine are equally e€ective at a common receptor in
attenuating glutamate release, then inhibition of ectonucleoti-
dases by cibacron blue 3GA would have no e€ect on the
release inhibitory e€ect of ATP. This is clearly not the case.
While the mechanism of action of cibacron blue 3GA remains
Figure 6 E€ect of cibacron blue 3GA (Ciba) on adenosine and ATP
inhibition of glutamate release. Rat brain cortical slices were
depolarised twice using 2 min pulses of 46 mM K+. The second
depolarisation was in the presence of the compounds shown on the x-
axis of the graph (ATP and adenosine, 300 mM; cibacron blue 3GA,
30 mM). Glutamate concentrations were determined using the HPLC
technique and the areas under the peaks calculated using GraphPad
Prism. S2/S1 ratios were calculated and plotted as a bar chart
(mean+s.e.mean, n=4 separate experiments, *P50.05, ***P50.001).
British Journal of Pharmacology, vol 131 (3)

G.C. Bennett & M.R. Boarder
P2Y regulation of glutamate release
623
to be con®rmed, it does suggest that ATP and adenosine act at
di€erent receptors in modulating depolarisation-evoked
glutamate release.
In conclusion we have provided direct evidence that
glutamate release from depolarised nerve terminals is subject
to inhibitory control by both ATP and adenosine, and that the
e€ect of ATP is not dependent on conversion to adenosine.
This study establishes that depolarisation-elicited glutamate
release in the brain, like biogenic amine release, is subject to
regulation by the presynaptic action of ATP and adenosine.
The results may be interpreted as suggesting that ATP and
adenosine are acting at separate receptors. It seems likely that
presynaptic adenosine and P2 receptors exert a widespread
in¯uence on brain function.
This work was supported by The BBSRC and Roche Bioscience,
Palo Alto, California. We thank Dr Andrew Young (Department of
Psychology, University of Leicester) for help establishing the HPLC
analysis, Dr C.J. Dixon for help with manuscript preparation and
Dr Anthony Ford (Roche Bioscience) for support and discussion.
References
ALLGAIER, C., PULLMANN, F., SCHOBERT, A., VON KUGELGEN, I.
& HERTTING, G. (1994). P2 purinoceptors modulating noradre-
naline release from sympathetic neurons in culture. Eur. J.
Pharmacol., 252, R7 ± R8.
MANZONI, O.J., TOSHIYA, M. & NICOLL, R.A. (1994). Release of
adenosine by activation of NMDA receptors in the hippocampus.
Science, 265, 2098 ± 2101.
MELDRUM, L.A. & BURNSTOCK, G. (1983). Evidence that ATP acts
as a cotransmitter with noradrenaline in sympathetic nerves
supplying the guinea-pig vas deferens. Eur. J. Pharmacol., 92,
161 ± 163.
BOARDER, M.R.
& HOURANI, S.M.O. (1998). The regulation of
vascular function by P2 receptors: multiple sites and multiple
receptors. Trends Pharmacol. Sci., 19, 99 ± 107.
BOEHM, S. (1998). Selective inhibition of M-type potassium channels
in rat sympathetic neurons by uridine nucleotide preferring
receptors. Br. J. Pharmacol., 116, 2341 ± 2343.
NIEBER, K., POELCHEN, W. & ILLES, P. (1997). Role of ATP in fast
excitatory synaptic potential in locus coeruleus neurons of the
rat. Br. J. Pharmacol., 122, 423 ± 430.
BULTMANN, R., TRENDELENBURG, M., TULUC, F., WITTENBURG,
H. & STARKE, K. (1999). Concomitant blockade of P2X-receptors
and ecto-nucleotidases by P2-receptor antagonists: functional
consequences in rat vas deferens. Naunyn-Schmiedeberg's Arch.
Pharmacol., 359, 339 ± 344.
NORENBERG, W., VON KUGELGEN, I., MEYER, A., ILLES, P. &
STARKE, K. (2000). M-type K+ currents in rat cultured
thoracolumbar sympathetic neurons and their role in uracil
nucleotide-evoked noradrenaline release. Br. J. Pharmacol., 129,
709 ± 731.
BURNSTOCK, G. (1972). Purinergic nerves. Pharmacol. Rev., 24,
509 ± 581.
BURNSTOCK, G. (1976). Do some nerve cells release more than one
neurotransmitter? Neuroscience, 1, 239 ± 248.
BURNSTOCK, G., DUMSDAY, B. & SMYTHE, A. (1972). Atropine-
resitant excitation of the urinary bladder: The possibility of
transmission via nerves releasing a purine nucleotide. Br. J.
Pharmacol., 44, 451 ± 461.
CUHNA, R.A., SEBASTIAO, A.M. & RIBEIRO, J.A. (1994). Purinergic
modulation of the evoked release of [3H]acteylcholine from the
hippocampus and cerebral corte of the rat: role of the
ectonucleotidases. Eur. J. Neurosci., 6, 33 ± 42.
PATON, D.M. (1981). Presynaptic neuromodulation mediated by
purinergic receptors. In Purinergic receptors. ed. Burnstock, G.
pp. 199 ± 219. Chapman and Hall: London.
SHEN, K.-Z. & NORTH, R.A. (1993). Excitation of rat locus coeruleus
neurons by adenosine 5'-triphosphate: ionic mechanism and
receptor characterisation. J. Neurosci., 13, 894 ± 899.
SHINOZUKA, K., BJUR, R.A. & WESTFALL, D.P. (1988). Character-
isation of prejunctional purinoceptors on adrenergic nerves of
the rat caudal artery. Naunyn-Schmiedeberg's Arch. Pharmacol.,
338, 221 ± 227.
SILINSKY, E.M. & GINSBORG, B.L. (1983). Inhibition of acetylcho-
line release from preganglionic frog nerves by ATP but not
adenosine. Nature, 305, 327 ± 328.
CUHNA, R.A., VIZI, E.S., RIBIERO, J.A. & SEBASTIAO, A. (1996).
Preferential release of ATP and its extracellular catabolism as a
source of adenosine upon high- but not low-frequency stimula-
tion of rat hippocampal slices. J. Neurochem., 67, 2180 ± 2187.
SWANSON, R.A., LIU, J., MILLER, J.W., ROTHSTEIN, J.D., FARRELL,
K., STEIN, B.A.
& LONGUEMARE, M.C. (1997). Neuronal
regulation of glutamate transporter subtype expression in
astrocytes. J. Neurosci., 17, 932 ± 940.
VOLTERRA, A., BEZZI, P., RIZZINI, B.L., TROTTI, D., ULLENSVANG,
EDWARDS, F.A., GIBB, A.J.
& COLQUHOUN, D. (1992). ATP
receptor-mediated synaptic currents in the central nervous
system. Nature, 359, 144 ± 147.
K., DANBOLT, N.
& RACAGNI, G. (1996). The competitive
FREDHOLM, B., ABBRACHIO, M.P., BURNSTOCK, G., DUBYAK,
G.R., HARDEN, T.K., JACOBSEN, K.A., SCHWABE, U. & WIL-
LIAMS, M. (1997). Towards a revised nomenclature of P1 and P2
receptors. Trends Pharmacol. Sci., 18, 79 ± 82.
transport inhibitor L-trans-pyrrolidine-dicarboxylate triggers
excitotoxicity in rat cortical-astrocyte co-cultures via glutamate
release rather than uptake inhibition. Eur. J. Neurosci., 8, 2019 ±
2029.
FREDHOLM, B.B. & DUNWIDDIE, T.V. (1988). How does adenosine
inhibit transmitter release? Trends Pharmacol. Sci., 9, 130 ± 134.
GU, J.G. & MACDERMOTT, A.B. (1997). Activation of ATP P2X
receptors elicits glutamate release from sensory neuron synapses.
Nature, 389, 749 ± 753.
JAMES, S. & RICHARDSON, P.J. (1993). Production of adenosine
from extracellular ATP at striatal cholinergic synapse. J.
Neurochem., 60, 219 ± 227.
VON KUGELGEN, I., ALLGAIER, C., SCHOBERT, A. & STARKE, K.
(1994a). Corelease of noradrenaline and ATP from cultured
sypathetic neurons. Neuroscience, 61, 199 ± 202.
VON KUGELGEN, I., SPATH, L. & STARKE, K. (1994b). Evidence for
P2-purinoceptor-mediated inhibition of noradrenaline release in
rat brain cortex. Br. J. Pharmacol., 113, 815 ± 822.
VON KUGELGEN, I., STOFFEL, D.
& STARKE, K. (1995). P2-
purinoceptor mediated inhibition of noradrenaline release in
rat atria. Br. J. Pharmacol., 115, 247 ± 254.
KOCH, H., KUGELGEN, I.
& STARKE, K. (1997a). P2-receptor-
mediated inhibition of noradrenaline release in the rat hippo-
campus. Naunyn-Schmiedeberg's. Arch. Pharmacol., 355, 707 ±
715.
WEBB, T.E., SIMON, J., BATESON, A.N. & BARNARD, E.A. (1994).
Transient expression of the recombinant chick brain P2Y1
receptor and localisation of the corresponding mRNA. Mol.
Cell Biol., 40, 437 ± 442.
WICKLUND, N.P., GUSTAFSSON, L.E. & LUNDIN, J. (1985). Pre- and
postjunctional modulation of cholinergic neuroe€ector transmis-
sion by adenine nucelotides. Experiments with agonist and
antagonist. Acta Physiologica Scand., 125, 681 ± 691.
ZHANG, Y.X., YAMASHITA, H., OHSHITA, T., SAWAMOTO, N. &
NAKAMURA, S. (1995). ATP increases extracellular dopamine
level through stimulation of P2Y purinoceptors in the rat
striatum. Brain Res., 691, 205 ± 212.
KOCH, H., TRENDELENBERG, A.U., BULTMANN, R., VON KUGEL-
GEN, I. & STARKE, K. (1997b). P2-purinoceptors inhibiting the
release of endogenous dopamine in rat neostriatum. Naunyn-
Schmiedeberg's Arch. Pharmacol., 355, R33.
KOCH, H., VON KUGELGEN, I. & STARKE, K. (1995). Presynaptic P2-
purinoceptors at serotininergic axons in rat brain cortex.
Naunyn-Schmiedeberg's Arch. Pharmacol., 352, R21.
KOIZUMI, S. & INOUE, K. (1997). Inhibition by ATP of calcium
oscillations in rat cultured hippocampal neurones. Br. J.
Pharmacol., 122, 51 ± 58.
KURZ, K., VON KUGELGEN, I. & STARKE, K. (1993). Prejunctional
modulation of noradrenaline release in mouse and rat vas
deferens: Contribution of P1 and P2-purinoceptors. Br. J.
Pharmacol., 110, 1465 ± 1472.
(Received April 4, 2000
Revised June 16, 2000
Accepted July 12, 2000)
British Journal of Pharmacology, vol 131 (3)