Autoradiographic studies on the uptake of adenosine and on binding of adenosine analogues in neurons and astrocytes of cultured rat cerebellum and spinal cord

Neuroprotection by adenosine in the brain: From A(1) receptor activation to A (2A) receptor blockade

Adenosine is a neuromodulator that operates via the most abundant inhibitory adenosine A(1) receptors (A(1)Rs) and the less abundant, but widespread, facilitatory A(2A)Rs. It is commonly assumed that A(1)Rs play a key role in neuroprotection since they decrease glutamate release and hyperpolarize neurons. In fact, A(1)R activation at the onset of neuronal injury attenuates brain damage, whereas its blockade exacerbates damage in adult animals. However, there is a down-regulation of central A(1)Rs in chronic noxious situations. In contrast, A(2A)Rs are up-regulated in noxious brain conditions and their blockade confers robust brain neuroprotection in adult animals. The brain neuroprotective effect of A(2A)R antagonists is maintained in chronic noxious brain conditions without observable peripheral effects, thus justifying the interest of A(2A)R antagonists as novel protective agents in neurodegenerative diseases such as Parkinson's and Alzheimer's disease, ischemic brain damage and epilepsy. The greater interest of A(2A)R blockade compared to A(1)R activation does not mean that A(1)R activation is irrelevant for a neuroprotective strategy. In fact, it is proposed that coupling A(2A)R antagonists with strategies aimed at bursting the levels of extracellular adenosine (by inhibiting adenosine kinase) to activate A(1)Rs might constitute the more robust brain neuroprotective strategy based on the adenosine neuromodulatory system. This strategy should be useful in adult animals and especially in the elderly (where brain pathologies are prevalent) but is not valid for fetus or newborns where the impact of adenosine receptors on brain damage is different.

Nucleoside transporter expression and function in cultured mouse astrocytes

Uptake of purine and pyrimidine nucleosides in astrocytes is important for several reasons: (1) uptake of nucleosides contributes to nucleic acid synthesis; (2) astrocytes synthesize AMP, ADP, and ATP from adenosine and GTP from guanosine; and (3) adenosine and guanosine function as neuromodulators, whose effects are partly terminated by cellular uptake. It has previously been shown that adenosine is rapidly accumulated by active uptake in astrocytes (Hertz and Matz, Neurochem Res 14:755-760, 1989), but the ratio between active uptake and metabolism-driven uptake of adenosine is unknown, as are uptake characteristics for guanosine. The present study therefore aims at providing detailed information of nucleoside transport and transporters in primary cultures of mouse astrocytes. Reverse transcription-polymerase chain reaction identified the two equilibrative nucleoside transporters, ENT1 and ENT2, together with the concentrative nucleoside transporter CNT2, whereas CNT3 was absent, and CNT1 expression could not be investigated. Uptake studies of tritiated thymidine, formycin B, guanosine, and adenosine (3-s uptakes at 1-4 degrees C to study diffusional uptake and 1-60-min uptakes at 37 degrees C to study concentrative uptake) demonstrated a fast diffusional uptake of all four nucleosides, a small, Na(+)-independent and probably metabolism-driven uptake of thymidine (consistent with DNA synthesis), larger metabolism-driven uptakes of guanosine (consistent with synthesis of DNA, RNA, and GTP) and especially of adenosine (consistent with rapid nucleotide synthesis), and Na(+)-dependent uptakes of adenosine (consistent with its concentrative uptake) and guanosine, rendering neuromodulator uptake independent of nucleoside metabolism. Astrocytes are accordingly well suited for both intense nucleoside metabolism and metabolism-independent uptake to terminate neuromodulator effects of adenosine and guanosine.

Novel metabotropic glutamate receptor negatively coupled to adenylyl cyclase in cultured rat cerebellar astrocytes

Several excitatory amino acid ligands were found potently to inhibit forskolin-stimulated cAMP accumulation in rat cultured cerebellar astrocytes: L-cysteine sulfinic acid (L-CSA) = L-aspartate > L-glutamate >/= the glutamate uptake inhibitor, L-PDC. This property did not reflect activation of conventional glutamate receptors, since the selective ionotropic glutamate receptor agonists NMDA, AMPA, and kainate, as well as several mGlu receptor agonists [(1S,3R)-ACPD, (S)-DHPG, DCG-IV, L-AP4, L-quisqualate, and L-CCG-I], were without activity. In addition, the mGlu receptor antagonists, L-AP3, (S)-4CPG, Eglu, LY341495, (RS)-CPPG, and (S)-MCPG failed to reverse 30 microM glutamate-mediated inhibitory responses. L-PDC-mediated inhibition was abolished by the addition of the enzyme glutamate-pyruvate transaminase. This finding suggests that the effect of L-PDC is indirect and that it is mediated through endogenously released L-glutamate. Interestingly, L-glutamate-mediated inhibitory responses were resistant to pertussis toxin, suggesting that G(i)/G(o) type G proteins were not involved. However, inhibition of protein kinase C (PKC, either via the selective PKC inhibitor GF109203X or chronic PMA treatment) augmented glutamate-mediated inhibitory responses. Although mGlu3 receptors (which are negatively coupled to adenylyl cyclase) are expressed in astrocyte populations, in our study Western blot analysis indicated that this receptor type was not expressed in cerebellar astrocytes. We therefore suggest that cerebellar astrocytes express a novel mGlu receptor, which is negatively coupled to adenylyl cyclase, and possesses an atypical pharmacological profile.

Purine uptake and release in rat C6 glioma cells: nucleoside transport and purine metabolism under ATP-depleting conditions

Adenosine, through activation of membrane-bound receptors, has been reported to have neuroprotective properties during strokes or seizures. The role of astrocytes in regulating brain interstitial adenosine levels has not been clearly defined. We have determined the nucleoside transporters present in rat C6 glioma cells. RT-PCR analysis, (3)H-nucleoside uptake experiments, and [(3)H]nitrobenzylthioinosine ([(3)H]NBMPR) binding assays indicated that the primary functional nucleoside transporter in C6 cells was rENT2, an equilibrative nucleoside transporter (ENT) that is relatively insensitive to inhibition by NBMPR. [(3)H]Formycin B, a poorly metabolized nucleoside analogue, was used to investigate nucleoside release processes, and rENT2 transporters mediated [(3)H]formycin B release from these cells. Adenosine release was investigated by first loading cells with [(3)H]adenine to label adenine nucleotide pools. Tritium release was initiated by inhibiting glycolytic and oxidative ATP generation and thus depleting ATP levels. Our results indicate that during ATP-depleting conditions, AMP catabolism progressed via the reactions AMP --> IMP --> inosine --> hypoxanthine, which accounted for >90% of the evoked tritium release. It was surprising that adenosine was not released during ATP-depleting conditions unless AMP deaminase and adenosine deaminase were inhibited. Inosine release was enhanced by inhibition of purine nucleoside phosphorylase; ENT2 transporters mediated the release of adenosine or inosine. However, inhibition of AMP deaminase/adenosine deaminase or purine nucleoside phosphorylase during ATP depletion produced release of adenosine or inosine, respectively, via the rENT2 transporter. This indicates that C6 glioma cells possess primarily rENT2 nucleoside transporters that function in adenosine uptake but that intracellular metabolism prevents the release of adenosine from these cells even during ATP-depleting conditions.

Expression of purine metabolism-related enzymes by microglial cells in the developing rat brain

The nucleoside triphosphatase (NTPase), nucleoside diphosphatase (NDPase), 5'-nucleotidase (5'-Nase), and purine nucleoside phosphorylase (PNPase) activity has been examined in the cerebral cortex, subcortical white matter, and hippocampus from embryonic day (E)16 to postnatal day (P)18. Microglia display all four purine-related enzymatic activities, but the expression of these enzymatic activities differed depending on the distinct microglial typologies observed during brain development. We have identified three main morphologic typologies during the process of microglial differentiation: ameboid microglia (parenchymatic precursors), primitive ramified microglia (intermediate forms), and resting microglia (differentiated cells). Ameboid microglia, which were encountered from E16 to P12, displayed the four enzymatic activities. However, some ameboid microglial cells lacked 5'-Nase activity in gray matter, and some were PNPase-negative in both gray and white matter. Primitive ramified microglia were already observed in the embryonic period but mostly distributed during the first 2 postnatal weeks. These cells expressed NTPase, NDPase, 5'-Nase, and PNPase. Similar to ameboid microglia, we found primitive ramified microglia lacking the 5'-Nase and PNPase activities. Resting microglia, which were mostly distinguishable from the third postnatal week, expressed NTPase and NDPase, but they lacked or displayed very low levels of 5'-Nase activity, and only a subpopulation of resting microglia was PNPase-positive. Apart from cells of the microglial lineage, GFAP-positive astrocytes and radial glia cells were also labeled by the PNPase histochemistry. As shown by our results, the differentiation process from cell precursors into mature microglia is accompanied by changes in the expression of purine-related enzymes. We suggest that the enzymatic profile and levels of the different purine-related enzymes may depend not only on the differentiation stage but also on the nature of the cells. The use of purine-related histoenzymatic techniques as a microglial markers and the possible involvement of microglia in the control of extracellular purine levels during development are also discussed.

Adenosine A1 receptor activation preferentially protects cultured cerebellar neurons versus astrocytes against hypoxia-induced death

Administration of adenosine A1 receptor agonists in vivo is neuroprotective in various stroke models. Experiments using either mixed cultures of neurons and astrocytes or brain slices, in which several cell types are present, have demonstrated that activation of A1 receptors also id protective against hypoxia and/or hypoglycemia in vitro. In this study, we have examined the effect of the A1 agonist cyclopentyladenosine (CPA) on cellular damage, measured by efflux of lactate dehydrogenase (LDH), in highly enriched primary cultures of either neurons of astrocytes exposed to different metabolic insults. CPA reduced neuronal LDH release induced by a combination of hypoxia and substrate deprivation ("simulated ischemia"; IC50 = 28 nM) of by hypoxia alone (IC50 = 170 nM). In contrast, CPA had no effect on neuronal damage induced by substrate deprivation alone, not did it affect ischemic death to astrocytes. The neuroprotective effect of CPA during simulated ischemia and hypoxia were reversed by the A1 antagonist 1,3-dipropyl-8-cyclopentylxanthine (DPCPX). These data demonstrate that activation of an adenosine A1 receptor on neurons, but not astrocytes, is protective against cellular damage of death induced specifically by hypoxia as opposed to other metabolic insults such as hypoglycemia.

Potentiated cAMP rise in metabotropically stimulated rat cultured astrocytes by a Ca2+-related A1/A2 adenosine receptor cooperation

Adenosine agonists favoured an intracellular Ca2+ rise in cultured type 1 astrocytes if the metabotropic glutamate receptors were concomitantly stimulated by (2S, 1's, 2's)-2-(carboxycyclopropyl)glycine (L-CCG-I; group II agonist), quisqualate (group I agonist) or 1-aminocyclopentane-trans-1, 3-dicarboxylic acid (t-ACPD; groupI/II agonist). Since the generation of a Ca2+ signal reflected a newly adopted adenosine A1 receptor action, we tested the possible consequence that the established opposing control of the cellular cAMP content by inhibitory A1 and stimulatory A2 receptor activation was also altered. During metabotropic receptor stimulation by LCCG-I, quisqualate or t-ACPD, the non-selective adenosine agonist 2-chloroadenosine (Cl-adenosine) caused a potentiated cAMP increase which markedly exceeded that produced by Cl-adenosine alone. This cAMP potentiation resulted from altered and Ca2+-dependent A1/A2 receptor cooperation. It was abolished by A1 receptor blockade and could not be achieved in the presence of t-ACPD by the A1 agonist R(-)N6-(2-phenylisopropyl)-adenosine or by the A2 agonist 5'-N-ethyl carboxyamidoadenosine alone, but obtained using their combination. The cAMP potentiation was blocked by intracellular Ca2+ chelation and the required A1 receptor action could be mimicked by a Ca2+ signal generated by the P2y receptor agonist adenosine 5'-(beta-thio)diphosphate. The results support the conclusion that nanomolar concentrations of adenosine may influence astrocyte reactions by stimulating Ca2+ and cAMP-dependent signalling cascade.

Morphology and distribution of microglial cells in the young and adult mouse cerebellum

The morphology and distribution of microglial cells were studied in the normal cerebellum of young and adult mice using the histochemical demonstration of nucleoside diphosphatase as a specific microglial marker. Our results showed that microglial cells were present in all cerebellular lobules of both young and adult mice, but their distribution and morphology were not homogeneous throughout the cerebellum. Heterogeneity in microglial cell distribution was exclusively related to their location in the different histological layers, and no significant differences were found either between the different cerebellar lobules or between young and adult mice. Microglial density was higher in the cerebellar nuclei than in the cortex; within the cortex, the molecular layer was less densely populated by microglial cells than the granular layer and the white matter. The morphological study revealed that microglial cells were ramified in all cerebellar lobules of both young and adult mice but showed different sizes and ramification patterns as a function of their specific location in the different histological layers. Several typologies of microglial cells were described on the basis of observations in both horizontal and coronal sections. The specific layer-related pattern of microglial distribution and morphology in mouse cerebellum strongly suggests a physical and functional adaptation of these cells to the characteristics of their microenvironment.

Modulation of nerve and glial function by adenosine--role in the development of ischemic damage

Adenosine is released during brain ischemia and provides neuroprotection by actions on nerve and glial cells. Activation of the adenosine A1 receptor enhances the K+ and Cl- conductance in neurons, leading to membrane hyperpolarization and postsynaptic reduction of neuronal Ca2+ influx through voltage- and NMDA receptor-dependent channels. In addition adenosine A1 receptor activation decreases excitatory amino acid release, possibly via inhibition of N- and P-type Ca2+ channels. The A1 and A2 receptors, coupled to Gi/G(o) and Gs proteins respectively, often co-exist and interact with the phospholipase C-dependent activation of the protein kinase C and the adenylyl cyclase. Activation of the A1 receptor may mimic metabotropic receptor stimulation in activating intracellular Ca2+ mobilization and PKC. A2 receptor mediated cAMP formation is depressed by high intracellular Ca2+ but enhanced by PKC activation. By modulating these metabolic signaling events, adenosine may influence acute cell functions, gene transcription and sustained changes of nerve and glial cells relevant for the development of ischemic damage. The neuroprotective adenosine effect seems to be amplified by treatment with propentofylline, which enhances adenosine release, influences the balance between A1 and A2 receptor mediated actions, depresses the free radical formation in activated microglia and influences astrocyte reactions.

Characterization of adenosine receptors in a model of cultured neurons from rat forebrain

The neuromodulator adenosine is acting through specific receptors coupled to adenylate cyclase via G-proteins. The expression of both adenosine receptors A1 and A2 as well as forskolin binding sites was investigated by radioligand binding techniques in 8-day-old neurons isolated from fetal rat forebrain and cultured in chemically-defined medium. Adenosine A1 receptors were specifically labeled with [3H]chloro-N6-cyclopentyladenosine (CCPA), whereas [3H]CGS 21680 was used for the analysis of A2 receptors. Cultured neurons exhibited high affinity binding sites for CCPA (Bmax = 160 fmol/mg protein; Kd = 2.9 nM), and for CGS 21680 (Bmax = 14 fmol/mg protein; Kd = 1.7 nM). These data correlate well with those obtained in crude membranes isolated from the newborn rat forebrain. The incubation of culture membranes in the additional presence of guanylyl-5'-imidodiphosphate (Gpp(NH)p, a GTP analogue) led to significantly increased Kd-values, suggesting the association of adenosine receptors with G-proteins. Finally, cultured neurons also bound specifically [3H]forskolin with characteristics close to those found in the newborn brain, indicating that cultured neurons appear as an appropriate model for studying the neuromodulatory properties of adenosine.

Expression of adenosine A1 receptors in cultured neurons from fetal rat brain

The expression of adenosine A1 receptors was investigated using [3H]2-chloro-N6-cyclopentyladenosine (CCPA) in 8-day-old cultured neurons from fetal rat forebrain grown in serum-free medium. [3H]CCPA bound specifically and with high affinity (Kd = 2.9 nM) to a homogeneous population of sites. Displacement of CCPA binding by various adenosine derivatives indicated that A1 receptors were selectively labeled. The presence of Gpp(NH)p, a GTP analogue, reduced significantly the binding affinity (Kd = 12.2 nM), suggesting that A1 receptors detected in intact cultured cells are linked to associated G proteins.

Mechanisms of drug actions against neuronal damage caused by ischemia--an overview

1. Oxygen and energy deficits induces a cascade of pathological processes which lead to neuronal dysfunction and cell death. 2. The pathogenesis of ischemia-induced neuronal damage includes a disturbed calcium homeostasis, an excessive release of EAA and an enhanced formation of free oxygen radicals. 3. Calcium antagonists inhibit Ca2+ influx into the neuronal cell via VSCC. 4. Glutamate antagonists reduce intracellular Ca2+ concentration by inactivation of NMDA receptor-associated calcium channels (NMDA antagonists) or AMPA/quisqualate receptor-linked sodium channels (non-NMDA antagonists). 5. Furthermore, oxygen radical scavengers can avoid neuronal damage. 6. Agonists of the adenosinergic and serotonergic transmitter systems contribute to neuroprotection by hyperpolarization of the neuronal membrane due to an increase of K+ permeability.

The distribution of A1 adenosine receptor and 5'-nucleotidase in pig brain cortex subcellular fractions

Pig brain cerebral cortex was subfractionated by isopycnic centrifugation in sucrose gradients. In each subfraction the content of the agonist [3H]R-PIA binding, the activity of adenosine metabolizing enzymes (5'-nucleotidase and adenosine deaminase) and the activity of membrane marker enzymes were determined. The fractions were also examined by electron microscope. In general, the results suggest a widespread distribution of A1 adenosine receptors in membranes from different origins. Marker enzyme profile characterization indicated an enrichment of A1 adenosine receptor in pre-synaptic membranes isolated from the crude synaptosomal fraction (P2B subfraction) as well as in membranes of glial origin such as myelin. The receptor is also present in the endoplasmic reticulum and in membranes isolated from the microsomal fraction that seem to have a post-synaptic origin (P3B). In subfractions having a high content of adenosine receptor the equilibrium binding parameters were obtained as well as the proportion of high- to low-affinity sites. From the values of the equilibrium constants it was not possible to find differences between the receptor in the different subfractions. Analysis of the affinity state distribution showed a diminished percentage of high-affinity sites in fraction P3A, which can be accounted by the existence of myelin membranes; in contrast the percentage of high-affinity states was higher in P2 and P3B, indicating that in these fractions the receptor is present in synaptosomal membranes. The close correlation shown between the enzyme 5'-nucleotidase specific activity and the specific ligand binding distributions led us to postulate an important role for the enzyme in the regulation of adenosine action in pig brain cortex.

Nucleosides influence the myelination of dorsal root ganglion cells in vitro; a quantitative comparison

alpha-Modified minimal essential medium (alpha MEM) is the most useful commercial medium for survival and myelin formation of dorsal root ganglia (DRGs). When components of alpha MEM are compared with those of other commercial media, nucleosides are found only in alpha MEM. We examined the effects of nucleosides on myelin formation in rat DRG cultures. When myelin formation was viewed quantitatively, cultures using alpha MEM with nucleosides yielded 2- to 3-fold more myelin segments than cultures using alpha MEM without nucleosides.

Direct effects of chronic ethanol exposure on beta-adrenergic and adenosine-sensitive adenylate cyclase activities and cyclic AMP content in primary cerebellar cultures

The direct effects of chronic ethanol exposure on adenylate cyclase activity and cyclic AMP content were investigated in primary cerebellar cultures. By morphological criteria these cultures mainly contain granule cells with some astrocytes, and each cell type appears to contain both beta-adrenergic and adenosine-sensitive adenylate cyclase systems. Chronic treatment of the primary cerebellar cultures with 120 mM ethanol for 6 days caused a reduction in the stimulation of cyclic AMP content by isoproterenol and by the adenosine analogue 2-chloroadenosine. Kinetic analysis indicated that the chronic ethanol treatment decreased maximal activation of adenylate cyclase, as well as increased the EC50 values for norepinephrine and 2-chloroadenosine. Activation of norepinephrine-stimulated adenylate cyclase activity by in vitro ethanol was significantly enhanced after the chronic ethanol exposure. However, the chronic treatment did not alter activation of the 2-chloroadenosine-stimulated enzyme by in vitro ethanol. A similar difference in the response to in vitro ethanol after the chronic treatment was observed when cyclic AMP content of the intact cells was measured. The present data indicate that chronic ethanol exposure causes a selective increase in the sensitivity of adenylate cyclase to ethanol in some brain cells and a more generalized desensitization of receptor-stimulated cyclic AMP production.

Pharmacological evidence for the modulation of monoamine release by adenosine in the invertebrate nervous system

An in vitro preparation from the pedal ganglia of the marine bivalve, Mytilus edulis, was used to examine the modulation of transmitter release by adenosine and its analogs from invertebrate nervous tissue. The ganglia of this organism contain the monoamines dopamine (DA), serotonin (5-HT), and norepinephrine (NE), and the presynaptic release of these substances is known to be calcium-dependent. This organism also contains a DA-sensitive adenylate cyclase system which resembles that seen in mammals. Neural tissue from the pedal ganglia was incubated with labeled monoamines, and release studies were then conducted in superfusion chambers; release of monoamines was evoked by the addition of 50 mM KCl. Addition to the superfusion medium of the adenosine analog, 5'-N-ethylcarboxamidoadenosine (NECA; 10 nM), inhibited the release of 5-HT and DA, and to a lesser extent NE, whereas 100-fold higher concentrations of adenosine itself and the adenosine analog, R-N6-phenylisopropyladenosine, were required to achieve comparable levels of inhibition. The inhibitory effects of NECA on neurotransmitter release were blocked by the adenosine receptor antagonist, theophylline (IC50 = 10-14 microM). The results from this study indicate for the first time the possible role of adenosine as a modulator of neurotransmitter release in the invertebrate nervous system.

Receptor alterations associated with spinal motoneuron degeneration in bovine Akabane disease

Akabane disease in cattle is characterized by congenital abnormalities including arthrogryposis, which is characterized by a depletion of spinal ventral horn motoneurons, a loss of axons, and depletion of myelin in the lateral and ventral tracts. These neuropathological changes produced major reductions (70-80%) in the density of muscarinic cholinergic, glycine/strychnine, and central-type benzodiazepine receptors in the ventral horn motor nuclei. The density of peripheral-type benzodiazepine receptors and adenosine A1 receptors was dramatically increased (250-300%) in the lateral and ventral spinal columns, reflecting the proliferation of glial cells. Bovine Akabane disease represents a useful model for assessing the processes and consequences of neuronal degeneration and demyelination and has implications for human diseases such as amyotrophic lateral sclerosis.

ATP-evoked Ca2+ mobilisation and prostanoid release from astrocytes: P2-purinergic receptors linked to phosphoinositide hydrolysis

Astrocyte cultures prelabelled with either [3H]inositol or 45Ca2+ were exposed to ATP and its hydrolysis products. ATP and ADP, but not AMP and adenosine, produced increases in the accumulation of intracellular 3H-labelled inositol phosphates (IP), efflux of 45Ca2+, and release of thromboxane A2 (TXA2). Whereas ATP-stimulated 3H-IP accumulation was unaffected, its ability to promote TXA2 release was markedly reduced by mepacrine, an inhibitor of phospholipase A2 (PLA2). ATP-evoked 3H-IP production was also spared following treatment with the cyclooxygenase inhibitor, indomethacin. We conclude that ATP-induced phosphoinositide (PPI) breakdown and 45 Ca2+ mobilisation occurred in parallel with, if not preceded, the release of TXA2. Following depletion of intracellular Ca2+ with a brief preexposure to ATP in the absence of extracellular Ca2+, the release of TXA2 in response to a subsequent ATP challenge was greatly reduced when compared with control. These results suggest that mobilisation of cytosolic Ca2+ may be the stimulus for PLA2 activation and, thus, TXA2 release. Stimulation of alpha 1-adrenoceptors also caused PPI breakdown and 45 Ca2+ efflux but not TXA2 release. The effects of ATP and noradrenaline (NA) on 3H-IP accumulation were additive, but their combined ability to increase 45Ca2+ efflux was not. Interestingly, in the presence of NA, ATP-stimulated TXA2 release was reduced. Our data provide evidence that functional P2-purinergic receptors are present on astrocytes and that ATP is the first physiologically relevant stimulus found to initiate prostanoid release from these cells.

Autoradiographic localization of binding sites for vasoactive intestinal peptide and angiotensin II on neurons and astrocytes of cultured rat central nervous system

The cellular localization of binding sites for [125I] vasoactive intestinal polypeptide and [3H]angiotensin II was studied in explant cultures of rat spinal cord, brain stem, cerebellum and cortex by means of autoradiography. In spinal cord cultures, interneurons of the dorsal horn and motoneurons of the ventral horn were labelled by [125I]vasoactive intestinal polypeptide and [3H]angiotensin II. In many brain stem cultures, groups of large neurons revealed intense binding of both peptides. In contrast, only few medium-sized cerebellar neurons, probably interneurons, showed binding sites for [125I]vasoactive intestinal polypeptide and [3H]angiotensin II. Furthermore, the intensity of labelling of cerebellar neurons was usually weaker than that of neurons of the brain stem and spinal cord. Many neurons in cultures of neocortex were also labelled by [125I]vasoactive intestinal polypeptide, whereas little binding was found with [3H]angiotensin II. In addition to neurons, binding sites for both peptides were also observed on astrocytes. Labelling of these cells was more intense in spinal cord and brain stem cultures than in cultures of cerebellum and cortex, suggesting that only a certain type or a certain population of astrocytes possesses receptors for vasoactive intestinal polypeptide and angiotensin II, or that glial cells in different parts of the CNS have different physiological and pharmacological properties.

The role of purines in nociception

The preceding review indicates that there is convincing evidence for the presence of adenosine in and release of adenosine from capsaicin-sensitive small diameter primary afferent neurons in the spinal cord (Fig. 1). Within the dorsal spinal cord, adenosine inhibits the transmission of nociceptive information, although details of mechanisms involved in this action remain to be established. In view of the antinociceptive actions of adenosine analogues, there has been some interest in the possibility of developing adenosine analogues as analgesic agents. However, this goal may be frustrated by this concomitant suppression of motor function, as well as the production of other side effects due to the diverse nature of pharmacological effects seen with adenosine analogues. Release of adenosine from small diameter primary afferent nerve terminals and subsequent activation of extracellular adenosine receptors in the dorsal horn of the spinal cord appears to contribute significantly to the spinal action of opioids. An understanding of spinal mechanisms of actions of adenosine therefore is an important prerequisite for our understanding of the action of this clinically important group of drugs. ATP may be a sensory neurotransmitter released from non-nociceptive large diameter primary afferent neurons (Fig. 1). The subsequent extracellular conversion of released ATP to adenosine may produce suppression of the transmission of noxious sensory information via small diameter primary afferent fibres, and contribute to the phenomenon of vibration induced analgesia. Clearly, the role of purines on spinal cord processing of nociceptive information merits considerable attention.