Adenosine and Purines

Trevor W. Stone, M-R. Nikbakht and E. Martin O'Kane Abstract

Adenosine can act on four subtypes of receptor, of which the Ai and A2A subtypes have received the most attention experimentally. The Ai receptors are primarily inhibitory by depressing transmitter release or causing hyperpolarisation, while the A2A receptors often cause overall excitation by direct depolarisation or the facilitation of transmitter release. Activation of these receptors can also modulate neuronal sensitivity to classical transmitters by altering receptor function, especially of acetylcholine and glutamate receptors, two of the transmitters most closely involved in processess of learning and memory. Both the Ai and A2A receptors have been shown to modulate synaptic plasticity in areas such as the hippocampus, although the relationship between these effects and the influence on individual classical transmitters remains unclear at present. Adenine nucleotides are also known to be active at receptors in the brain, and some forms of long-term potentiation may be in part attributable to the local release ofATP Together, the purine nucleosides and nucleotides represent strong candidates for major physiological regulators of the cellular processes underlying neuronal excitability and synaptic plasticity.

Origin of Adenosine in the Extracellular Fluid

Adenosine is normally present in the extracellular fluid at a concentration of around i^M or less.13,28,29,137,185,264 The origin of this adenosine remains unclear with some authors supporting the view that the nucleoside is transported out of cells by bi-directional membrane transporters when the intracellular level of free adenosine exceeds a threshold level while others argue that enzymes such as adenosine deaminase and adenosine kinase maintain intracellular free adenosine at a low concentration, and that extracellular nucleoside is primarily the consequence of metabolism of ATP which has been released from cells as a neurotransmitter, cotransmitter or trophic factor, for example. Release can be stimulated by cellular depolarisation produced by transmitters such as glutamate and acetylcholine.44,185

Adenosine Receptors

To date four types of adenosine receptor have been cloned, namely adenosine A1, A2A, A2B, and A3 receptors.83,149,151,266 Adenosine A1 and A2 receptors occur widely distributed throughout the CNS, with the heaviest density of A1 receptors in the hippocampus and of 'classical' A2A receptors in the striatum and limbic areas such as the nucleus accumbens and olfactory tubercle.117,138,165 A1 receptors in the hippocampus have been localised to granule cell bodies and dendrites and to pyramidal neurons, but do also occur on glial cells.

There is, however, uncertainty as to whether the A2A receptors found throughout the CNS are homogeneous. Molecular biology has revealed only a single population of sites, but there are significant pharmacological differences, especially in the binding affinities of 2-[4-(2-carboxyethyl)-phenylethylamino]-5'N-ethyl-carboxamido-adenosine (CGS 21680) and 4-(2-[7-amino-2-{2-furyl}{1,2,4}-triazolo{2,3-a}-(1,3,5}triazin-5-yl-amino]ethyl)phenol (ZM

From Messengers to Molecules: Memories Are Made of These, edited by Gernot Riedel and Bettina Platt. ©2004 Eurekah.com and Kluwer Academic / Plenum Publishers.

241385), which suggest functional differences, probably attributable to other cellular or membrane components linked to, and modifying responsiveness of, the receptors themselves. There is certainly a high affinity binding site for the prototypical (striatal) A2A receptor agonist CGS 21680, but the detailed characterisation of CGS 21680 binding sites in the hippocampus reveals that they are not identical in their properties and pharmacology to those present (in much greater abundance) in the striatum.49,118,119,143

Adenosine and Learning

Adenosine analogues exhibit a range of behavioural effects (see ref. 227), which include sedation,14,45,74,218,220 anticonvulsant activity,15,65,74 anti-nocisponsive effects,3,110,190,259 inhibition of aggression181 and suppression of operant responding. 35

Surprisingly, however, relatively little attention has been paid to purine modulation of memory processes. Several studies have indicated that adenosine analogues can suppress aspects oflearning such as the acquisition of conditioned reflexes254,255 and conditioned avoidance respond-ing.153,172,262 Depressant effects on working memory177 and specifically tests of spatial memory247 have also been reported. These generally inhibitory actions have formed the basis of current interest in the potential use of xanthine derivatives as cognition enhancers.203,221,233

Adenosine Receptor Subtypes and Learning

Few of these studies have been designed specifically to clarify the relative importance of the different adenosine receptor subtypes. Hooper et al111 addressed this question using one of the simplest tests of memory function—spontaneous alternation in a Y-maze. Tests involving spontaneous alternation became widely used after the classical studies ofDennis,59 Douglas and Isaacson63 and Anisman,10 and are based on the tendency of rodents to enter that arm of a Y-maze least recently visited. Alternation scores by definition are significantly greater than 0.5, the proportion of alternations expected if the animal was selecting arms purely by chance. Some authors have attempted to interpret spontaneous alternation behaviour in terms of habituation to the most recently explored arm of the Y-maze,10,97,132 but it is now generally accepted that spontaneous alternation behaviour reflects spatial working memory19,184200,231,248 as originally proposed by Dennis.59

This system was used to examine the effects of purine receptor ligands with some selectivity for acting at A1 and A2 adenosine receptors. The A1 receptor selective agonist N6-cyclopentyladenosine (CPA) did not change spontaneous alternation behaviour alone, but it prevented the decrease of spontaneous alternation scores produced by scopolamine. The A1 receptor selective antagonist 1,3-dipropyl-8-cyclopentylxanthine (DPCPX) prevented this scopolamine reversal by CPA although it had no effect when administered alone. The nonselective adenosine receptor antagonist 8-(p-sulphophenyl)theophylline (8PST), which does not cross the blood-brain barrier, had no effect upon alternation behaviour or arm entries. The A2 receptor selective agonist (N6-[2-(3,5-dimethoxyphenyl)-2-(2-methylphenyl)ethyl]adenosine (DPMA), and the A2 receptor selective antagonist 1,3-dimethyl-1-propargylxanthine (DMPX) had no effect on alternation behaviour alone and did not modify the effect of scopolamine.

These results contrast with previous studies using more complex experimental paradigms. Normile and Barraco,172 for example, observed that CPA attenuated retention in a passive avoidance test. Winsky and Harvey254 reported that R-phenylisopropyl-adenosine (R-PIA) reduced the acquisition of a conditioned avoidance response and similar results were claimed by Martin et al.153 Conversely, acute administration of an A1 receptor antagonist has been claimed to facilitate learning.2 3 The explanation for the ability of these groups to find effects of the purines tested alone, and the results of Hooper et al111 is not clear, although different behavioural tests were used in each case, and the adenosine receptor ligands were also different. In the case of R-PIA, only very low doses can be used, less than 1 mg/kg, if depression of overall motor activity is to be avoided. In the work of Normile and Barraco172 the doses of CPA found to be effective were over 0.5 mg/kg - doses over ten-fold greater than the doses used by Hooper et al.111 Similar results were obtained by Zarrindast and Bijan262 who only obtained effects on passive avoidance learning at R-PIA doses of 0.125 mg/kg or above. In the study by Martin et al153 the ED50 doses of R-PIA and CPA were found to be 10 mg/kg and 1.5 mg/kg respectively, doses far in excess of those used by Hooper et al.111 While it remains possible that inhibition of learning does occur at these high doses, the work of Hooper et al111 clearly indicates that at low doses, a reversal of scopolamine-induced memory deficits can be obtained.

When purine receptor ligands were combined with scopolamine, however, it was clear that A1 but not A2 receptor activation could modify working memory deficits induced by scopola-mine. Perhaps even more significantly, however, the fact that neither the A1 receptor selective antagonist DPCPX25,144 nor the A2 receptor antagonist DMPX had any effects themselves upon spontaneous alternation behaviour or arm entries, and did not modify scopolamine's elimination of spontaneous alternation behaviour suggests that activation of A1 or A2 receptors by endogenous adenosine is not normally involved in spatial working memory.

Of great interest is the later finding that blockade of A2A receptors by DMPX could reverse the detriment to learning caused by the NMDA receptor channel blocker dizocilpine,86 possibly implying that different neural mechanisms and/or pathways were involved in the disruption of learning produced by scopolamine and dizocilpine. A facilitatory effect of a more selective A2A receptor antagonist—7-( 2-phenylethyl)-5-amino-2 -(2-furyl)-pyrazolo- [4,3e] -1,2,4-triazolo [1,5-c]-pyrimidine (SCH 58261)—was also reported by Kopf et al133 using an inhibitory avoidance test.

A potential confounding factor in studies with purines is the influence of locomotor depression. Importantly, however, locomotor activity was unchanged by CPA at any of the doses used in the analysis of alternation behaviour. Furthermore, it would be expected that a decrease in total entries, which would imply an increased time between successive entries, might allow greater time for forgetting the previous arm, thus hindering spontaneous alternation behaviour. Anisman10 has investigated specifically the relationship between arm entries and spontaneous alternation behaviour using three strains of mice with different degrees of locomotor activity. Despite this difference, all strains showed the same level of spontaneous alternation behaviour. In the same study, it was noted that scopolamine increased arm entries in two strains but not the third, whereas it eliminated spontaneous alternation behaviour in all mice. In addition, Drew et al66 observed no correlation between arm entries and spontaneous alternation behaviour. Nevertheless, the motor effects of modulating purine receptor function may contribute to some instances of apparent changes of learning behaviour. Blockade of A2 receptors, for example, increases motor activity234 and still needs to be excluded as a factor in the memory-enhancing effects of A2 antagonists.

Cellular Actions of Adenosine

Given this evidence for adenosine receptor modulation of learning, what are the cellular processes which may underlie the behavioural effect? This section will be approached by dealing firstly with studies of the effects of purines and their antagonists on synaptic plasticity, and then considering in more detail the various sites and mechanisms of action of adenosine and its analogues which could underlie the change of plasticity.

Once in the extracellular space adenosine is able to act on its receptors to modulate neuronal activity in the central nervous system by a variety of actions including inhibition of the release of neurotransmitters such as glutamate,33,38,78,191,201 acetylcholine,51,125,135,154,225 dopamine,160,263 serotonin82 and noradrenaline,120,130 by acting at the A1 receptor. The A2 receptors on the other hand tend to increase the release of some of these transmitters40,126,225 including the all-important glutamate.42,176,189 Depression of release is directed largely against excitatory transmitters, with little influence on the release of inhibitory transmitters.109, 36,261

While there is general agreement that A1 receptors depress, and A2A receptors increase, the release of acetylcholine,51,1 5,135,154,225 studies of GABA release have proved more controversial. It has been reported, for example, that A2A receptors can increase41,155 or decrease125,134,166 GABA release. It is almost certain that methodological differences, especially the use of radiolabelled versus endogenous material, account for some of these differences, but clarification would be valuable.

The inhibitory effects of adenosine are often mediated by an inhibition of calcium influx into synaptic terminals5'6'87'205'258 or by a decreased availability of calcium within the terminals to the active site for release.214 Aj receptors have been shown to inhibit presynaptic ro—conotoxin sensitive calcium channels,260 while several authors have shown inhibitory actions on N, P and Q-type channels.6'98'244 In dissociated hippocampal neurons, Mogul et al164 described the inhibition by Aj receptors of N-type currents' while A2B receptors appeared to increase P-type channels. Synaptic currents can still be inhibited by Aj receptors in the presence of calcium channel blockers' suggesting that there is an additional component to the presynaptic activity of adenosine which is independent of calcium movement.204

In postsynaptic cell somata' adenosine can alter neuronal polarisation' Aj receptors often inducing hyperpolarisation' while A2 receptors often cause hyperpolarisations have usually been ascribed to the opening of potassium channels in the hippocampus and elsewhere99'100'182'196'241 or of chloride channels.150 The channels on hippocampal pyramidal neurons are sensitive to blockade by barium21 which not only prevents the direct hyperpolarisation by adenosine' but also prevents the changes by adenosine of spike activation threshold and EPSP / spike coupling.

There is some evidence that adenosine can act partly via ATP-sensitive potassium channels. For example' tolbutamide and glibenclamide' blockers of those channels' can reduce the postsynaptic actions of adenosine' including the changes of EPSP / spike coupling' at concentrations which do not alter the presynaptic actions.112' 74

Distinguishing the site of action of adenosine—presynaptic or postsynaptic—is difficult to achieve in the mammalian hippocampus in view of the difficulty of recording directly from synaptic terminals. The use of paired-pulse stimuli is far from ideal' but does provide a window on that distinction which lends a different view from simple measures of spike or postsynaptic potential size' or the demands of quantal analysis. Using the paired-pulse approach Higgins and Stone108 attempted to examine the effects of adenosine specifically on the presynaptic terminals' as described below.

Adenosine and Synaptic Plasticity

The activation of A1 receptors was shown to suppress the induction of LTP' provided that adenosine (the agonist used in those experiments) was applied within one minute of the inducing tetanus; there was no effect of adenosine if applied 5 minutes after stimulation.11 A1 receptors may even respond to endogenous levels of adenosine sufficiently to regulate the degree of LTP and LTD induced by electrical stimulation' since antagonists at these receptors increase the amplitude of both these phenomena.55'56 However' while Fuji et al91-93 confirmed the ability of adenosine to restrain the size of LTP' they reported that the presence of the nonselective adenosine receptor antagonist 8-cyclopentyl-theophylline (CPT) also decreased the size of a subsequent depotentiation' implying that in their system' endogenous adenosine was contributing to' or facilitating' the extent of depotentiation. Whether this is simply a species difference between rats and guinea-pigs would be interesting to establish.

There are several reports that adenosine' released spontaneously or as the result of neuronal activation' can participate in the physiological regulation of synaptic transmission. Thus' even low frequency stimulation of hippocampal axons can apparently release enough adenosine to inhibit synaptic transmission.163 Given the numerous physiological factors which can in turn modulate adenosine levels extracellularly' or can modify the results of activating adenosine receptors' this modulatory role of adenosine could play a pivotal role in many aspects of hip-pocampal function' including those related to learning and memory. A recently described example of this has been reported by Huang et al.113 This group studied the depotentiation of hippocampal potentials following the enhancement by LTP. When a period of low frequency stimulation was initiated within a few minutes of the initial LTP' stable depotentiation was obtained which could be mimicked by bath application of adenosine' and was prevented by the A1 receptor antagonist DPCPX.

Interestingly, the LTP obtained by stimulation in the presence of an Ai receptor blocker was dependent on the activation of NMDA receptors, whereas LTD induced under similar conditions was not.54 The intriguing result was that of a dissociation between the effects of adenosine receptor activation (Ai or A2) on synaptic potentials and on EPSP-spike coupling91,92 providing some of the clearest evidence for nonparallel changes of presynaptic and postsynaptic actions of adenosine which in turn imply a complex, state- and environment-sensitive modulation by adenosine of synaptic plasticity.

A2 receptors also do appear to contribute to classical, NMDA-dependent LTP, since agonists increase56 and most importantly antagonists reduce the amplitude of tetanus-induced potentiated potentials91,122,123,212 Antagonists were only effective when applied within a relatively short time window after an inducing tetanus; application after 45 minutes, for example, failed to modify the potentiated potential size, suggesting that the A2 receptors are more important for the induction of LTP than for its maintenance.122 The A2A antagonist produced a substantially greater facilitation of depotentiation when studied using evoked excitatory postsyn-aptic potentials compared with postsynaptic population spikes, suggesting that the effect is preferentially expressed presynaptically rather than postsynaptically.9

The A2A receptor population in the hippocampus was reported not to greatly influence electrophysiological activity, possibly because these receptors show low affinity for agonists.147 This lack of effect of the A2A receptor agonist CGS 21680 was later confirmed by Kessey and Mogul,122 although less selective agonists could increase synaptic potentials, while antagonists reduced them, leading to the proposal that it is the A2B population which can most readily modulate transmission. However, Sebastiao and Ribeiro209 used concentrations of CGS 21680 which they believed to be more selective for A2A receptors, and without the complicating activation of A1 receptors which was noted by Lupica et al.147 This, and later work from the same laboratory50 demonstrated that A2A receptors could enhance transmission, an effect which probably stems inpart from the ability of A2 receptor agonists to increase presynaptic calcium conductances.98,1 The similar enhancement of transmission recorded by Li and Henry141 was accompanied by a slowly developing depolarisation which was responsible for a post-inhibitory excitatory action of adenosine. Paradoxically, however, the absence of any change of paired-pulse facilitation in response to A2 receptor activation would seem to exclude a presynaptic site of action.122 While this paradox has not yet been fully resolved, part of the explanation is that A2A receptors can facilitate postsynaptic responses to AMPA, allowing the emergence of an NMDA receptor-independent form of LTP.122

Paired-Pulse Inhibition

Endogenous adenosine may play a significant part in other aspects of synaptic transmission in addition to LTP and LTD. The phenomenon of paired-pulse inhibition is believed to reflect the depletion of presynaptic stores of transmitter and any decrease of that inhibition should indicate a specifically presynaptic inhibitory site of action of an agent. Higgins and Stone108 concluded that adenosine probably contributed to that fraction of paired-pulse inhibition which was not blocked by bicuculline and was not therefore mediated by GABA release from local interneurones. In the same study it was revealed that adenosine contributed also to the inhibition produced by twin stimuli separated by only 30ms, implying that a rapid release of adenos-ine might allow this substance to function as a classical neurotransmitter. The reduction of paired-pulse inhibition produced by CPT in the presence of bicuculline was only partly reversible, raising the possibility that endogenous adenosine itself may play a role in long-term changes of neuronal excitability. In terms of understanding fully the relationship between adenosine receptors and learning, it would be valuable to have a clearer view than is available at present on the relative magnitudes of the various actions of adenosine on hippocampal transmission and, in particular, whether all those actions are optimally expressed under the same or different environmental conditions existing under varying physiological and pathological conditions.

Several groups have previously reported an inhibitory effect of adenosine Ai receptors on population excitatory postsynaptic potentials (popEPSP), population spikes (PS) and the relationship between the two i.e., EPSP-spike (E-S) coupling in the CA1 area of rat hippocampus.173 The popEPSP gives primarily a measure of membrane potential changes generated by excitatory synapses on the apical dendrites of CA1 pyramidal neurones. The population spike reflects the summated firing of CA1 pyramidal neurones8 and gives a measure of the excitability of the postsynaptic neurone. EPSP-spike coupling gives an indication of the ability of a given level of synaptic depolarisation to induce the postsynaptic cell to fire an action potential. Intracellularly, changes in excitability can be measured as a change in EPSP slope or as a change in the firing probability of the cell.2,9,206,237

Interactions between Adenosine and Cholinergic Neurotransmission

Central cholinergic systems have been widely implicated in learning and memory processes (see ref. 101 for an excellent review; and also the relevant chapters in this book). The postnatal time course of development of cholinergic neurones parallels closely the development of spontaneous alternation behaviour.76 As more recently observed by Dunbar et al,68 specific cholinergic markers such as choline acetyltransferase, in areas of the brain believed to be associated with learning such as the hippocampus, correlate with spatial learning ability. The selective block of hippocampal muscarinic Mi receptors has also been shown to impair working memory in rats.178

Consistent with this view, cholinergic antagonists have been canvassed as a means of inducing a pharmacological model of the memory disturbance encountered in Alzheimer's and other degenerative disorders.64,222,223,240 The alkaloid scopolamine has been shown to impair learning in a variety of paradigms and in a range of species including humans64,222,240 and accordingly reduces alternation scores in rodents and other species.19,76,116,178,200,226,231,235,240,248

Cellular Mechanisms of Adenosine / Acetylcholine Interactions

Both adenosine148 and acetylcholine192,210 act on presynaptic receptors to regulate glutamate release from synaptic terminals, including those of the CA1 Schaffer collateral and commis-sural axons. Again, paired-pulse inhibition was used as a sensitive indicator of presynaptic terminal function.10,29,58,159,252 Nikbakht and Stone169 demonstrated that, using this protocol, both adenosine, acting at A1 receptors and oxotremorine-M acting at M2 receptors2 ,105 were able to depress transmitter release at short interpulse intervals (10ms), and facilitate release at longer intervals (20 and 50ms) as shown by others.72

There is long-standing evidence that the activation of adenosine receptors can suppress responses to muscarinic receptors43 and similar data have been collected from experiments on sensory ganglia36 as well as the hippocampus.23 A more recent study sought to establish whether the interactions between adenosine and muscarinic receptors were apparent specifically on presynaptic terminals.169 The presynaptic inhibitory effects of A1 and M2 receptors are occlusive: the combination of agonists at these sites has a less than additive effect upon transmitter release from CA1 terminals (Fig. 1). This suggests that they are acting via a common mechanism. Previous work has suggested that the suppression of transmitter release is mediated by a reduction of calcium influx or calcium availability to the release process.214 The blockade of calcium channels by adenosine and muscarinic receptors exhibits occlusion193 and might, therefore, underlie their occlusive interaction on transmitter release. However, several groups have reported that presynaptic cholinomimetic effects in the hippocampus are not mediated by a suppression of calcium channels,202 so that potassium conductances may be more relevant. Raising extracellular potassium levels or adding 4-aminopyridine to block potassium channels suppressed the responses to both CPA and oxotremorine-M, suggesting that both receptor types are operating by increasing potassium conductance in the axon terminals. These channels may, therefore, represent a common site of action. A similar convergence was reported by McCormick and Williamson156 on postsynaptic sites. These effects could be secondary to the

Figure 1. Histograms summarising the effect on the EPSP slope in rat hippocampal slices of two combinations of CPA and oxotremorine-M. In both cases, the effects of CPA and oxotremorine-M produce comparable degrees of inhibition of the response, but their combined addition produces an effect which is not significantly greater than either alone. Both were able to produce 100% inhibition at sufficiently high concentrations. The final column indicates the predicted effect if responses to the two agents had been additive. The actual combined response (ACT) was significantly different from the predicted additive response (p < 0.001, n = 5 for (A), n = 3 for (B)) even when the addition in B was limited to the theoretical maximum of 100%. (Reproduced with permission from ref. 169).

Figure 1. Histograms summarising the effect on the EPSP slope in rat hippocampal slices of two combinations of CPA and oxotremorine-M. In both cases, the effects of CPA and oxotremorine-M produce comparable degrees of inhibition of the response, but their combined addition produces an effect which is not significantly greater than either alone. Both were able to produce 100% inhibition at sufficiently high concentrations. The final column indicates the predicted effect if responses to the two agents had been additive. The actual combined response (ACT) was significantly different from the predicted additive response (p < 0.001, n = 5 for (A), n = 3 for (B)) even when the addition in B was limited to the theoretical maximum of 100%. (Reproduced with permission from ref. 169).

reported effects on calcium conductances, or the calcium changes could be secondary to the changes of potassium movements altering the polarisation state of the presynaptic terminals.

Adenosine and Acetylcholine Release

Acetylcholine is one of the transmitters whose release is modulated by the activation of adenosine receptors. However, while several groups have demonstrated an overall inhibitory effect of adenosine A1 receptors or stimulation by A2a receptors in the CNS,50,51,125,135 Cunha et al48,51 uncovered regional variations within the hippocampus such that release was modulated only by inhibitory A1 receptors in the CA1 area, whereas it was inhibited by A1 and increased by A2A receptors in CA3. If subtle differences such as these occur in other areas of brain, and if different regions of hippocampus are involved in different aspects of learning, then dissecting out the relative roles of Ai and A2 receptors in learning will not be simple.

The effects of purines on acetylcholine release in the hippocampus in vivo need to be examined however, since Materi et al154 have demonstrated recently that Ai receptors suppress evoked but not spontaneous release of acetylcholine from the rat neocortex and, more surprisingly, that A2A receptors did not modify spontaneous or evoked release. It is, therefore, important to establish whether the effects of A2A receptors seen in vitro do not represent an experimental artefact and that they do also occur in the intact, conscious animal, in which the release of acetylcholine is known to be a critical factor in wakefulness. Some differences between neocor-tex and hippocampus may in fact reflect differences in the source of extracellular adenosine (adenosine efflux or nucleotide metabolism, see ref. 47) and differences in the accessibility of adenosine to Ai and A2A receptors which they could produce.47

Interactions between Purines and Glutamate Receptors

The activation of glutamate receptor subtypes is now known to be important for several aspects of long-term plastic changes including LTP and LTD, and there is now evidence for a variety of ways in which the activation of adenosine receptors can modify the presence or actions of glutamate.

Adenosine and Glutamate Release

As in the case of acetylcholine, there is evidence for a dual modulation by purines of glutamate release in the brain, Ai receptors inhibiting and A2 receptors increasing release.176,189,216

Adenosine and Glutamate Receptor Interactions

A close relationship may exist between the presence of adenosine receptors and the extent to which NMDA receptors can participate in plastic changes of neurotransmission. Klishin et al128 noticed that in the presence of an increased ratio of extracellular calcium to magnesium in hippocampal slices, blockade of A1 adenosine receptors induced a long-lasting increase in the NMDA receptor-mediated component of excitatory postsynaptic currents relative to the nonNMDA component. The authors proposed that a proportion of NMDA receptors may normally be functionally masked by A1 receptors, and it is these which are made available to the transmission process after A1 blockade. This would certainly account for the facilitation of learning reported by some groups using A1 receptor antagonists (e.g., ref. 203), and could be highly relevant to physiological learning (in the absence of pharmacological agents) if the A1 receptors are inactivated by other transmitters or receptors. The conclusion that there may be a population of 'latent' NMDA receptors suppressed under resting conditions by endogenous levels of adenosine was supported by the demonstration that, after blocking all NMDA receptor function with the channel blocker dizocilpine, an NMDA receptor-mediated component of transmission could be restored by perfusing slices with 8-cyclopentyltheophylline. One explanation of this finding is that removing the influence of endogenous adenosine had again revealed a population of NMDA receptors which had not previously contributed to glutamate sensitivity and which had therefore escaped blockade by the use-dependent agent dizocilpine.129

In dissociated hippocampal pyramidal neurons, de Mendonca et al57 found that A1 receptor activation would suppress ionic conductances induced by NMDA. This raises the possibility that intense stimulation of neurons, whether pathologically by hypoxia-ischaemia or physiologically during memory formation, might lead to a degree of NMDA receptor activation which is limited by local increases in adenosine concentration. Certainly, Mitchell et al163 have concluded that adenosine can be released by quite low levels of hippocampal fibre stimulation, reaching local levels high enough to inhibit further transmitter release. This report requires reexamination, however, to assess whether the cells studied were exhibiting a homogeneous response to A1 receptors, since it has been reported that on a sub-population of striatal neurons

Ai receptors do not modify NMDA receptor activation, whereas both A2A receptors and A3 receptors were able to inhibit NMDA-induced currents.171,256

As long ago as 1988, it was reported that the presynaptic inhibitory effects of adenosine on glutamate release in the hippocampal CA1 region were dependent on the presence of magnesium, since removal of this ion from the superfusing medium prevented responses to adenosine.16 This change was later shown to be reproduced by superfusing N-methyl-D-aspartate (NMDA), and prevented by including blockers of the NMDA-sensitive receptors (such as dizocilpine or 2-amino-5-phosphono-pentanoic acid) before the removal of magnesium.17 This result suggested that activation of NMDA receptors was involved in the suppression of adenos-ine sensitivity. Also consistent with this view was the weaker ability of adenosine receptor activation to suppress neuronal firing induced by microiontophoretically applied NMDA compared with firing induced by acetylcholine or quisqualate.18 Although interpreted as consistent with a postsynaptic locus for the interaction between NMDA and adenosine receptors, it is difficult to be certain of the site of action of agents applied by microiontophoresis,228 and attempts to do so by, for example, lowering extracellular calcium, complicate interpretation by modifying neuronal excitability and receptor function.

There remains a major question as to the site of the adenosine / NMDA interaction— presynaptic or postsynaptic. The interaction has therefore been reexamined using the paired-pulse paradigm, which is widely accepted as providing a more accurate indication of presynaptic events than the study of population spikes and postsynaptic potentials.106,107,252,258 Paired-pulse inhibition at interpulse intervals of around 10 ms reflects the depletion of transmitter from presynaptic stores, 6,106,252 and is reduced by agents or procedures which decrease transmitter release. Paired-pulse facilitation, on the other hand, at longer interpulse intervals, results from the residual intraterminal calcium which increases transmitter is al ready ample evidence for the existence of presynaptic glutamate receptors85 and especially presyn-aptic NMDA receptors27,83,121,152,180 on terminals in the hippocampus and other regions of CNS.

Data showed that NMDA receptor activation suppresses the inhibitory effects of adenosine on transmitter release assessed using paired-pulse interactions both with population spikes and population EPSPs. This interaction occurs at levels of NMDA receptor activation which are not themselves sufficient to alter paired-pulse inhibition and strongly suggests that the primary site of the interaction is presynaptic. The fact that the interaction can also be observed in the presence of bicuculline suggests that the receptors involved are likely to be located on the main terminals of the Schaffer collateral fibres, and not on inhibitory interneurones. In addition, the suppression of adenosine sensitivity can be produced by methods other than the direct activation of NMDA receptors. Thus, the induction of LTP, which is known to involve the activation of NMDA receptors by synaptically released glutamate, or the application of exogenous glycine which can enhance the activation of NMDA receptors162 and induce or facilitate LTP in regions such as the hippocampus213 and superior colliculus1,188,249 also reduced adenosine responses. Responses to baclofen were unaffected.

One explanation for some of the earlier data of Bartrup and Stone17 was proposed by Smith and Dunwiddie,217 who argued that the effects of magnesium removal could simply reflect the altered balance between calcium and magnesium in determining the amount of transmitter release and thus account for the loss of sensitivity to adenosine. However, the finding that application of NMDA itself mimicked the effects of low magnesium, while NMDA antagonists prevented it, indicates that this cannot represent the whole explanation and that amino acid receptors probably contribute to the phenomenon. Of course, it is still possible that the activation of NMDA receptors changes sensitivity to adenosine by way of an alteration of intracellular calcium levels or availability to the transduction mechanism.

The inhibitory effect of adenosine on population spikes, and the decrease of paired-pulse inhibition assessed using either population spikes or population excitatory postsynaptic potentials, were suppressed by performing the experiments in magnesium-free medium, or by superfusion of the slices with NMDA at a concentration (4 ^M) which did not itself affect potential size. The suppressant effect of NMDA was prevented by 2-amino-5-phosphonopentanoic acid. All these interactions were still seen in the presence of bicuculline methobromide, 30 ^M. Neither a-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid (AMPA) nor kainate produced a suppression of adenosine responses. The presence of NMDA did not modify the effects of baclofen on population potentials or paired-pulse inhibition. Activating NMDA receptors by the induction of LTP or by superfusion with glycine also reduced significantly the effects of adenosine on population spikes and paired-pulse interactions. Increasing population potential size by a mechanism which did not involve the activation of NMDA receptors (increasing stimulus strength) did not change sensitivity to adenosine. When adenosine receptor-selective agonists were tested, it was found that NMDA did not modify the inhibitory effect of the Ai receptor agonist N6-cyclopentyladenosine (CPA), but did enhance the excitatory effect of the A2A receptor agonist 2-[ p- (2-carboxyethyl)phenylethylamino] -5-N-ethylcarboxamidoadenosine (CGS 21680). The combined response to NMDA and CGS21680 was prevented by the A2A receptor selective antagonist 4- (2- 7-amino-2-(2-furyl)[1,2,4] triazolo [2,3a][1,3,5]triazin-5-ylamino]ethyl)phenol (ZM 241385). It was concluded that NMDA receptor activation can suppress neuronal sensitivity to adenosine by acting at presynaptic sites, and that this interaction results from an increase in the excitatory action of A2A receptors, rather than a depression of Ai receptor function.170

This result appears particularly surprising in view of demonstrations that the activation of A2A receptors can suppress neuronal responses to NMDA in slices and patch-clamp experiments57, 71,256 (Fig. 2). It should be emphasised, however, that the interactions described in the present study involved a concentration of NMDA which was not active when tested alone. It therefore seems that the simultaneous activation of A2A and NMDA receptors at low (sub-threshold) concentrations produce an increase of glutamate release and neuronal excitability, whereas their combined activation at higher concentrations - which are themselves depolarising - results in antagonism.

Overall, therefore, NMDA receptor activation seems able to modify selectively the presynaptic responses to activation of A2A adenosine receptors, leading to the masking of adenosine's inhibitory activity on transmitter release. The physiological significance of this is potentially interesting. Craig and White44 have proposed that adenosine A1 receptors present a barrier to the actions of NMDA receptors which must be overcome if the full effects of NMDA receptor activation are to be observed in phenomena such as LTP. The present work suggests that part of the mechanism of overcoming this barrier may be that, under conditions in which the amount of adenosine released by neurons and glia is greatly increased so that the relatively low affinity A2A receptors are activated, the inhibitory A1 receptors effects are overcome. Such a sequence provides at least one rationale for the otherwise curious coexistence of inhibitory A1 and facili-tatory A2A receptors on the same population of glutamatergic terminals, and is consistent with earlier proposals that A2A receptor activation can suppress responses mediated by A1 recep-

By affecting NMDA receptors adenosine may have a fundamental role to play in controlling the dynamics of neuronal interactions. The nonspecific blockade of adenosine receptors has been claimed to block what may be a crucial role for adenosine of preventing dendritic spikes generated by NMDA receptors.142 Such a blockade releases the tendency, described by many authors in the presence of adenosine antagonists and with specific experimental conditions (e.g., ref. 32) for neurons to generate spontaneous bursts of action potentials.

Other Receptor Interactions

In addition to the potential interactions between purine receptors and acetylcholine and glutamate receptors—two of the neurotransmitters most clearly and consistently related to learning and memory—there is also evidence for interaction between adenosine receptors themselves, between adenosine and dopamine receptors, and between adenosine and peptide receptors.

Figure 2. CGS21680 inhibited the conductance ot NMDA receptor channels. The holding potential was —80mV in this and all subsequent experiments. (A) Experimental procedure to assess possible ettects ot CGS21680 on the conductance ofNMDA-activated channels. The input resistance ot striatal mediumspiny neuronswas monitored by applying hyperpolarising voltage steps, 10 to 20mV in amplitude, and 100ms in duration every 10 seconds. The input resistance was measured 4 times: immediately before the first application of NMDA (10|iM; R1 beforeTl), during the maximum response to NMDA (10|iM; R2 during Tl), immediately before the second challenge with NMDA (10|iM) in the presence of CGS21680 (0.1|iM;R3 beforeT2), and duringthe maximum response to NMDA (1 un M j in the presence OÍCGS21680 (0. l|iM; R4 duringT2). (B) Input resistance values in 8 neurons sensitive to CGS21680. Means of 3 current responses at Rl, R2, R3, and R4 respectively were obtained according to the scheme in (A), either in the absence (open columns) or in the presence of CGS21680 (0. l|iM; solid columns)." I' < 0.001 significant differences from the respective controls (C) in the absence of NMDA (R2 compared with Rl, and R4 compared with R3, respectively); **P < 0.001, significant difference between NMDA alone and NMDA plus CGS21680 (R2 and R4). (C) Input resistance in 4 medium spiny neurons which did not respond to CGS21680. Here, the NMDA (10|iM)-evoked increase in membrane conductance was uninfluenced by CGS21680 0.1|iM; compare R2 with R4). *P0.001, significant differences from the respective controls in the absence of NMDA (10|iM). (Reproduced with permission from ref. 171).

A NMDA 10 fiM

1500

Ci 1250 I

g 1000

T NMDA

NMDAIOfiM T,

Kj O

CGS 21680 0.1 ]iu

50 pA

1 min

NMDA

NMDA

C 1500 n

3 1250

g 1000

"NMDA

"NMDA

Adenosine Receptor Interactions

The activation of both A2a50'137'173 and A3 receptors69 has been shown to suppress the activation of Ai receptors. There is evidence that these interactions may involve the enhanced desensitisation of the Aj receptors.62'69

One form of interaction between A2A and Aj receptors is illustrated in Figs. 3 and 4 Activation of adenosine Aj receptors with the specific agonist CPA caused a greater inhibition of PS amplitudes than popEPSP slopes indicating that the resulting E-S dissociation is due mainly to postsynaptic effects ofAj receptor activation.173 When adenosine A2A receptors are coactivated with A1 receptors, using CGS 21680 along with CPA, the postsynaptic inhibitory effects of A1 receptor activation are significantly attenuated showing that cross-talk exists between the two types of receptor. The attenuation of Aj receptor-mediated inhibitory responses by adenosine A2A receptor activation is not due to a functional antagonism between excitatory versus inhibitory effects of the two receptor types, since activation of A2A receptors by CGS21680 does not cause a significant degree of excitation.173

Cunha et al50 have previously shown an attenuation of adenosine Aj receptor responses on PS amplitude by A2A receptor activation in the rat hippocampus, while Dixon et al6 reported a desensitisation of adenosine Aj receptors by A2A receptors in the rat striatum, an effect mediated by protein kinase C. A reduction of Aj receptor binding can also be demonstrated in the presence of A2A receptor agonists, and this is also mediated via protein kinase C.146 In the hippocampus, however, we found no evidence that the Aj receptor effects could be modified by inhibitors of protein kinases (A or C). Previous investigators have also demonstrated a lack of relationship between cAMP levels and the electrophysiological effects of adenosine70,73,195,257 except in forskolin treated hippocampal slices.88 There is growing evidence for an interaction between adenosine and nitric oxide systems.61,157,186 However, neither the competitive nitric oxide synthase inhibitor L-nitroarginine methylester (L-NAME) nor the brain specific inhibitor 7-nitroindazole (7NI) showed any effects on the inhibition obtained upon addition of CPA, suggesting that nitric oxide does not play a significant role in the inhibition seen with Aj activation.

In contrast, blockade of potassium channels with barium attenuated the postsynaptic actions of adenosine Aj receptor activation. It is well established that, postsynaptically, adenosine increases potassium conductance,179,210,241 and it has been shown that barium will selectively block the postsynaptic hyperpolarising effects of adenosine.4,96,103,238 The E-S dissociation caused by adenosine Aj receptor activation studied extracellularly, and the directly measured effect on spike threshold recorded intracellularly, are also prevented by barium. The possibility exists that a similar suppression of a potassium current may be the mechanism by which ad-enosine A2A receptor activation causes inhibition of Aj receptor-mediated changes of spike threshold. Since barium can block several potassium currents,3 ,102 it is not clear which of these might be involved in the A1/A2A receptor interaction. It is unlikely that the IA current is involved, however, as Pan et al182 have shown that, whereas barium blocks the postsynaptic hyperpolarisation induced by adenosine, it does not prevent adenosine activation of the A-current.

Adenosine and Dopamine

There are several reports of dopamine receptors exerting a modulatory influence on synap-tic plasticity, and there has been much interest in the receptor-receptor interaction between D2 dopamine and A2A adenosine receptors79,80,83 and between D1 and Aj receptors.81 It is not yet clear whether D2 receptors are involved in the physiological regulation oflearning and memory, although dopamine itself does suppress spontaneous alternation253 and D2 dopamine agonists reverse scopolamine-induced memory is possible, therefore, that this represents another site at which endogenous purines could interact to modify learning in a psychologically state-dependent manner.

Figure 3. Intracellular recordings from pyramidal neurons in the hippocampus. Record A (a) illustrates the membrane potential of a neurone which is stimulated by a depolarising current pulse just sufficient to induce the production of an action potential on most occasions (0.2 nA delivered every 30s). During the period indicated by the bar below the record, N6-cyclopentyladenosine (CPA) was perfused at 100nM, and causes failure of action potential initiation with no accompanying change of membrane potential. In record A(b), the cell is superfused with CPA 100nM plus CGS 21680 at 30nM. The latter compound was perfused for 15 min before the addition of CPA. The elevation of spike threshold is now blocked and there is some evidence of increased synaptic activity and spontaneous action potentials, with little overall change of membrane potential. Records in B show this effect on a more expanded time scale from a different cell. B(a) and (d) represent responses of the cell to pulses of 0.2 and 0.4nA in the control state. Record (b) shows the failure of spike initiation and (e) a reduced number of spikes produced during superfusion with CPA 100nM. The latter record (e) also shows an increase in the degree of after-hyperpolarisation which probably contributed to the reduced spike number. Records (c) and (f) show the prevention of the CPA effect when coperfused with CGS21680 at 30nM. Calibrations: 20mV and 10 min in A; 50mV, 0.2nA and 300ms for B. (Reproduced with permission from ref. 175).

Figure 3. Intracellular recordings from pyramidal neurons in the hippocampus. Record A (a) illustrates the membrane potential of a neurone which is stimulated by a depolarising current pulse just sufficient to induce the production of an action potential on most occasions (0.2 nA delivered every 30s). During the period indicated by the bar below the record, N6-cyclopentyladenosine (CPA) was perfused at 100nM, and causes failure of action potential initiation with no accompanying change of membrane potential. In record A(b), the cell is superfused with CPA 100nM plus CGS 21680 at 30nM. The latter compound was perfused for 15 min before the addition of CPA. The elevation of spike threshold is now blocked and there is some evidence of increased synaptic activity and spontaneous action potentials, with little overall change of membrane potential. Records in B show this effect on a more expanded time scale from a different cell. B(a) and (d) represent responses of the cell to pulses of 0.2 and 0.4nA in the control state. Record (b) shows the failure of spike initiation and (e) a reduced number of spikes produced during superfusion with CPA 100nM. The latter record (e) also shows an increase in the degree of after-hyperpolarisation which probably contributed to the reduced spike number. Records (c) and (f) show the prevention of the CPA effect when coperfused with CGS21680 at 30nM. Calibrations: 20mV and 10 min in A; 50mV, 0.2nA and 300ms for B. (Reproduced with permission from ref. 175).

Figure 4. (A) Effect of N6-cyclopentyladenosine (CPA) on the population spikes (PS) and population excitatory postsynaptic potentials (popEPSP) ratio in rat hippocampal slices. While 50nM CPA decreased this by 80.1 ± 6.41% compared with control values, barium or CGS21680 greatly reduced this effect. (B) summarises the number of spikes evoked by intracellular pulses of varying amplitude and shows the depressant effect of CPA, the block of this effect by CGS21680, and the ability of ZM241385 to prevent the effect of CGS21680.(mean ± s.e.mean for n = 5). The insets show representative records at threshold and 5 x threshold for a typical cell. (C)illustrates the use of a depolarising current ramp to determine the threshold for spike initiation, and (D) summarises the results showing the elevation of threshold by CPA, the nonsignificant increase of threshold by barium and the blockade by barium of the CPA effect. (E) shows representative records of membrane voltage in response to hyperpolarising current pulses in the presence of CPA alone or in CPA plus CGS21680, and (F) summarises the pooled data for CPA (open circles) and CPA with CGS21680 (closed circles) indicating the absence of any changes at the concentrations used here.

Figure 4. (A) Effect of N6-cyclopentyladenosine (CPA) on the population spikes (PS) and population excitatory postsynaptic potentials (popEPSP) ratio in rat hippocampal slices. While 50nM CPA decreased this by 80.1 ± 6.41% compared with control values, barium or CGS21680 greatly reduced this effect. (B) summarises the number of spikes evoked by intracellular pulses of varying amplitude and shows the depressant effect of CPA, the block of this effect by CGS21680, and the ability of ZM241385 to prevent the effect of CGS21680.(mean ± s.e.mean for n = 5). The insets show representative records at threshold and 5 x threshold for a typical cell. (C)illustrates the use of a depolarising current ramp to determine the threshold for spike initiation, and (D) summarises the results showing the elevation of threshold by CPA, the nonsignificant increase of threshold by barium and the blockade by barium of the CPA effect. (E) shows representative records of membrane voltage in response to hyperpolarising current pulses in the presence of CPA alone or in CPA plus CGS21680, and (F) summarises the pooled data for CPA (open circles) and CPA with CGS21680 (closed circles) indicating the absence of any changes at the concentrations used here.

Adenosine and Peptides

There is growing evidence for an ability of adenosine receptors to modify responses to certain neuropeptide hormones and neurotransmitters. The targets studied to date include calcitonin gene-related peptide (CGRP) and vasoactive intestinal peptide (VIP). Correia-de-Sa and Ribeiro39 demonstrated that A2A receptors could facilitate the actions of CGRP at motor nerve terminals, and more recently have shown complex interactions in the hippocampus. Ai receptors exert a restricting effect on responses to CGRP in the hippocampus, so that effects of the peptide can only be observed when the influence of endogenous adenosine is removed by a suitable antagonist.208 Conversely, A2A receptors enhance the response to CGRP The studies on purine/peptide interactions to date have been well reviewed by Ribeiro.197

The Effects of Ageing on Adenosine Receptors

Intriguingly for any consideration of purines and learning, Corsi et al42 have reported that A2A receptors increase spontaneous glutamate release only in young, not old, rats. Although this work was performed in the striatum (in vivo), could a similar change in the hippocampus help to account for the declining memory so often associated with ageing? Interestingly, 8-cyclopentyltheophylline (CPT) has been shown to increase acetylcholine release from hip-pocampal slices only from relatively young (4 and 12 month) rats, but not from 24-month rats.22 The change was apparently attributable to a reduced Ai receptor density rather than affinity. The authors noted that in the older animals, there was a parallel increase in the level of extracellular adenosine, with the possibility that this has induced a down-regulation of Ai receptors and thus reduced Ai sensitivity. This in turn could mean that the physiological regulation of synaptic plasticity by adenosine is less effective with ageing.

The decline in the density, though not the affinity, of Ai receptors in the hippocampus and other regions of brain has also been noted by Cunha et al46,4 in old animals (24 months) compared with young ones (6 weeks). Conversely, there is an increase in the density of A2A receptors in the hippocampus46,48 and cortex. These binding studies were supported by electrophysiological data showing a parallel decrease in the efficacy of CPA to decrease neurotransmission by activating Ai receptors208 (Fig. 5). However, CPX generated a greater increase of potential size, implying that there was a greater concentration of functionally active Ai receptors despite the reduced apparent density. The authors attempted to explain this seeming paradox by suggesting that the reduced number of receptors is accompanied by an increased relative activation by endogenous adenosine. This explanation would certainly be consistent with the numerous reports of an increased level of extracellular adenosine in the brain of aged rats, but is difficult to reconcile with the studies of acetylcholine release.224

At the junior end of the ageing spectrum, the presynaptic inhibitory effects of Ai receptors are poorly developed shortly after birth and become increasingly apparent only over the first few weeks of life. This is probably related to the similar time course of development of those adenosine-sensitive processes relevant to transmitter release.67

Trophic Functions of Nucleosides

Long-term memory formation is usually assumed to involve some form of permanent or semi-permanent structural change in cells, whether neurons, glia or both. It is therefore pertinent to any discussion of learning to note that a number of purines and pyrimidines have been shown to have marked trophic effects on cells, altering neuronal growth or viability and glial proliferation. While this review is concentrated on the events surrounding the initial establishment of a memory trace, these long-term changes are clearly important and, as more is learnt about their cellular mechanisms and the nature of the receptors involved, these could form new targets for future generation drugs intended to reverse memory and cognitive decline. The trophic actions of purines and related compounds have been the subject of excellent and detailed review by Rathbone et al.i94

Figure 5. Comparison between the effects of the adenosine Ai receptor agonist CPA on field excitatory postsynaptic potentials (fEPSPs) recorded from the CA1 area of hippocampal slices taken from young adult (6 weeks) and and aged (24 months) rats. (a) shows trace recordings of averaged fEPSPs obtained in one experiment with an aged rat (left), and in another experiment with a young rat (right); in each panel the fEPSP obtained in the same slice in control conditions (C) and 30-34 mins after the application of CPA (40nM) are superimposed; calibration bars 500 ^V, 10ms. (b) shows the log-concentration response curves for the inhibitory effects of CPA on the slope of fEPSPs in aged and young adult rats; on the ordinate 0% corresponds to the fEPSP slope before CPA application (0.42 ± 0.06 mV/ms in young and 0.42 ± 0.10 mV/ ms in aged rats) and 100% represents the complete inhibition of fEPSPs. The data for each curve were obtained from 4-5 experiments, except for saturating concentrations ofCPA (60-100nM) in young animals, which represent results from 2 experiments; the s.e.mean are shown when they exceed the symbols in size. (Reproduced with permission from ref. 208)

Figure 5. Comparison between the effects of the adenosine Ai receptor agonist CPA on field excitatory postsynaptic potentials (fEPSPs) recorded from the CA1 area of hippocampal slices taken from young adult (6 weeks) and and aged (24 months) rats. (a) shows trace recordings of averaged fEPSPs obtained in one experiment with an aged rat (left), and in another experiment with a young rat (right); in each panel the fEPSP obtained in the same slice in control conditions (C) and 30-34 mins after the application of CPA (40nM) are superimposed; calibration bars 500 ^V, 10ms. (b) shows the log-concentration response curves for the inhibitory effects of CPA on the slope of fEPSPs in aged and young adult rats; on the ordinate 0% corresponds to the fEPSP slope before CPA application (0.42 ± 0.06 mV/ms in young and 0.42 ± 0.10 mV/ ms in aged rats) and 100% represents the complete inhibition of fEPSPs. The data for each curve were obtained from 4-5 experiments, except for saturating concentrations ofCPA (60-100nM) in young animals, which represent results from 2 experiments; the s.e.mean are shown when they exceed the symbols in size. (Reproduced with permission from ref. 208)

Nucleotides and Synaptic Plasticity

ATP has been shown to produce excitation of neurones in several regions of the central nervous system89,104,232,242 and to modulate membrane potassium114,167,198,243 or Ca2+ conductances.31,53,131 In addition, it is now recognised that ATP can function as a fast excitatory neurotransmitter in the locus coeruleus,75,168 peripheral ganglia95,215 and between cultured neurones77 often with a pharmacology suggestive of a P2 purinoceptor rather than an indirect effect such as ion chelation or metabolic modification.

Binding and molecular biology data suggest the presence of P2x3, P2x4 and P2x6 receptor subunits and their messenger RNAs in the hippocampus.12,22,124,139,161,211,219,236 Homomeric assemblies of P2x4 subunits respond poorly to aP-methyleneATP and are relatively insensitive to suramin. Combinations of P2x4 and P2x6 subunits, however, have been shown recently to be sensitive to the agonist effects of ap-methyleneATP and blockade by suramin.139

Despite this evidence, several studies have failed to detect any consistent functional responses to adenine nucleotides on neuronal networks and synaptic transmission in the mammalian hippocampus which cannot be explained by metabolism to adenosine.52,73,140,187,229,230 Only depressant effects were noted when ATP was applied by microiontophoresis to single neurones which were spontaneously active or excited by glutamate.60,229 Furthermore, ATP and derivatives depressed evoked potentials52,73,230 even when analogues were used which were resistant to hydrolysis and had selective actions on P2x and P2y receptors respectively.230 Cunha et al52 have recently performed a careful analysis suggesting that ATP must first be metabolised by ecto-nucleotidas

Was this article helpful?

0 0
Unraveling Alzheimers Disease

Unraveling Alzheimers Disease

I leave absolutely nothing out! Everything that I learned about Alzheimer’s I share with you. This is the most comprehensive report on Alzheimer’s you will ever read. No stone is left unturned in this comprehensive report.

Get My Free Ebook


Post a comment