Role of Excitatory Amino Acid Receptors in K+-and Glutamate-Evoked Release of Endogenous Adenosine from Rat Cortical Slices

K+ and glutamate released endogenous adenosine from superfused slices of rat parietal cortex. The absence of Ca2+ markedly diminished K+- but not glutamate-evoked adenosine release. Tetrodotoxin decreased K+- and glutamate-evoked adenosine release by 40 and 20%, respectively, indicating that release was mediated in part by propagated action potentials in the slices. Inhibition of ecto-5'-nucleotidase by alpha,beta-methylene ADP and GMP decreased basal release of adenosine by 40%, indicating that part of the adenosine was derived from the extracellular metabolism of released nucleotide. In contrast, inhibition of ecto-5'-nucleotidase did not affect release evoked by K+ or glutamate, suggesting that adenosine was released as such. Inhibition of glutamate uptake by dihydrokainate potentiated glutamate-evoked release of adenosine. Glutamate-evoked adenosine release was diminished 50 and 55% by the N-methyl-D-aspartate (NMDA) receptor antagonists, DL-2-amino-5-phosphonovaleric acid and (+)-5-methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-imine hydrogen maleate (MK-801), respectively. The remaining release in the presence of MK-801 was diminished a further 66% by the non-NMDA receptor antagonist, 6,7-dinitroquinoxaline-2,3-dione, suggesting that both NMDA and non-NMDA receptors were involved in glutamate-evoked adenosine release. Surprisingly, K+-evoked adenosine release was also diminished about 30% by NMDA antagonists, suggesting that K+-evoked adenosine release may be partly mediated indirectly through the release of an excitatory amino acid acting at NMDA receptors.     

Adenosine modulation of D-[3H]aspartate release in cultured retina cells exposed to oxidative stress

In this study we evaluated the role of adenosine receptor activation on the K+-evoked D-[3H]aspartate release in cultured chick retina cells exposed to oxidant conditions. Oxidative stress, induced by ascorbate (3.5 mM)/Fe2+ (100 microM), increased by about fourfold the release of D-[3H]aspartate, evoked by KCl 35 mM in the presence and in the absence of Ca2+. The agonist of A1 adenosine receptors, N6-cyclopentyladenosine (CPA; 10 nM), inhibited the K+-evoked D-[3H]aspartate release in control in oxidized cells. The antagonist of A1 adenosine receptor, 1,3-dipropyl-8-cyclopentylxanthine (DPCPX; 50 nM), potentiated the release of D-[3H]aspartate in oxidized cells, and reverted the effect observed in the presence of CPA 10 nM. However, in oxidized cells, when DPCPX was tested together with CPA 100 nM the total release of D-[3H]aspartate increased from 5.1 +/- 0.4% to 11.4 +/- 1.0%, this increase being reverted by 3,7-dimethyl-1-propargylxanthine (DMPX; 100 nM), an antagonist of A2A adenosine receptors. In cells of both experimental conditions, the K+-evoked release of D-[3H]aspartate was potentiated by the selective agonist of A2A adenosine receptors, 2-[4-(2-carboxyethyl)phenethylamino]-5'-N-ethylcarboxamidoadenosin e (CGS 21680; 10 nM), whereas the antagonist of these receptors, DMPX (100 nM), inhibited the release of D-[3H]aspartate in oxidized cells, but not in control cells. Adenosine deaminase (ADA; 1 U/ml), which is able to remove adenosine from the synaptic space, reduced the K+-evoked D-[3H]aspartate release, from 5.1 +/- 0.4% to 3.1 +/- 0.3% in oxidized cells, and had no significant effect in control cells. The extracellular accumulation of endogenous adenosine, upon K+-depolarization, was higher in oxidized cells than in control cells, and was reduced by the inhibitors of adenosine transporter (NBTI) and of ecto-5'-nucleotidase (AOPCP). This suggests that adenosine accumulation resulted from the outflow of adenosine mediated by the transporter, and from extracellular degradation of adenine nucleotide. Our data show that both inhibitory A1 and excitatory A2A adenosine receptors are present in cultured retina cells, and that the K+-evoked D-[3H]aspartate release is modulated by the balance between inhibitory and excitatory responses. Under oxidative stress conditions, the extracellular accumulation of endogenous adenosine seems to reach levels enough to potentiate the release of D-[3H]aspartate by the tonic activation of A2A adenosine receptors.     

Adenosine production and its interaction with protection of ischemic and reperfusion injury of the myocardium

Adenosine exerts cardioprotective effects on the ischemic myocardium. A flexibly mounted microdialysis probe was used to measure the concentration of interstitial adenosine and to assess the activity of ecto-5'-nucleotidase (a key enzyme responsible for adenosine production) in in vivo rat hearts. The level of adenosine during perfusion of adenosine 5'-adenosine monophosphate (AMP) was given as an index of the activity of ecto-5'-nucleotidase in the tissue. Endogenous norepinephrine (NE) activates both alpha(1)-adrenoceptors and protein kinase C (PKC), which, in turn, activates ecto-5'-nucleotidase via phosphorylation thereby enhancing the production of interstitial adenosine. Histamine-release NE activates PKC, which increased ecto-5'-nucleotidase activity and augmented release of adenosine. Opening of cardiac ATP sensitive K(+) (K(ATP)) channels may cause hydroxyl radical (.OH) generation through NE release. Lysophosphatidylcholine (LPC), an endogenous amphiphiphilic lipid metabolite, also increases the concentration of interstitial adenosine in rat hearts, through the PKC-mediated activation of endogenous ecto-5'-nucleotidase. Nitric oxide (NO) facilitates the production of interstitial adenosine, via guanosine 3',5'-cyclic monophosphate (cGMP)-mediated activation of ecto-5'-nucleotidase as another pathway. These mechanisms play an important role in high sensitivity of the cardiac adenosine system. Adenosine plays an important role as a modulator of ischemic reperfusion injury, and that the production and mechanism of action of adenosine are linked with NE release.     

Forskolin and phosphodiesterase inhibitors release adenosine but inhibit morphine-evoked release of adenosine from spinal cord synaptosomes.

The effects of forskolin, Ro 20-1724, rolipram, and 3-isobutyl-1-methylxanthine (IBMX) on morphine-evoked release of adenosine from dorsal spinal cord synaptosomes were evaluated to examine the potential involvement of cyclic AMP in this action of morphine. Ro 20-1724 (1-100 microM), rolipram (1-100 microM), and forskolin (1-10 microM) increased basal release of adenosine, and at 1 microM inhibited morphine-evoked release of adenosine. Release of adenosine by Ro 20-1724, rolipram, and forskolin was reduced 42-77% in the presence of alpha,beta-methylene ADP and GMP, which inhibits ecto-5'-nucleotidase activity by 81%, indicating that this adenosine originated predominantly as nucleotide(s). Significant amounts of adenosine also were released from the ventral spinal cord by these agents. Ro 20-1724 and rolipram did not significantly alter the uptake of adenosine into synaptosomes. Although Ro 20-1724 and rolipram had only limited effects on the extrasynaptosomal conversion of added cyclic AMP to adenosine, IBMX, a phosphodiesterase inhibitor with a broader spectrum of inhibitory activity for phosphodiesterase isoenzymes, significantly inhibited the conversion of cyclic AMP to adenosine and resulted in recovery of a substantial amount of cyclic AMP. As with the non-xanthine phosphodiesterase inhibitors, IBMX increased basal release of adenosine and reduced morphine-evoked release of adenosine. Adenosine released by IBMX was reduced 70% in the presence of alpha,beta-methylene ADP and GMP, and release from the ventral spinal cord was 61% of that from the dorsal spinal cord. Collectively, these results indicate that forskolin and phosphodiesterase inhibitors release nucleotide(s) which is (are) converted extrasynaptosomally to adenosine.(ABSTRACT TRUNCATED AT 250 WORDS)     

Parallel modification of adenosine extracellular metabolism and modulatory action in the hippocampus of aged rats

The neuromodulator adenosine can be released as such, mainly activating inhibitory A1 receptors, or formed from released ATP, preferentially activating facilitatory A2A receptors. We tested if changes in extracellular adenosine metabolism paralleled changes in A1/A2A receptor neuromodulation in the aged rat hippocampus. The evoked release and extracellular catabolism of ATP were 49-55% lower in aged rats, but ecto-5'-nucleotidase activity, which forms adenosine, was 5-fold higher whereas adenosine uptake was decreased by 50% in aged rats. The evoked extracellular adenosine accumulation was 30% greater in aged rats and there was a greater contribution of the ecto-nucleotidase pathway and a lower contribution of adenosine transporters for extracellular adenosine formation in nerve terminals. Interestingly, a supramaximal concentration of an A1 receptor agonist, N6-cyclopentyladenosine (250 nM) was less efficient in inhibiting (17% in old versus 34% in young) and A2A receptor activation with 30 nM CGS21680 was more efficient in facilitating (63% in old versus no effect in young) acetylcholine release from hippocampal slices of aged compared with young rats. The parallel changes in the metabolic sources of extracellular adenosine and A1/A2A receptor neuromodulation in aged rats further strengthens the idea that different metabolic sources of extracellular adenosine are designed to preferentially activate different adenosine receptor subtypes.     

Immunologically distinct isoforms of ecto-5'-nucleotidase in nerve terminals of different areas of the rat hippocampus

Ecto-5'-nucleotidase is regarded as being the key enzyme in the formation of the neuromodulator adenosine from released ATP. However, the association of ecto-5'-nucleotidase with nerve terminals is not consensual. Only enzyme histochemical and biochemical studies, but not immunocytochemical studies, agree on a general synaptic location of the enzyme. To clarify this issue further we tested the effect of an antibody against ecto-5'-nucleotidase, previously used in immunocytochemical studies, on the activity of ecto-5'-nucleotidase in fractions of nerve terminals isolated from different areas of rat hippocampus. The specific activity of extracellular AMP catabolism was higher in synaptosomes from the CA3 area (0.81+/-0.06 nmol/min/mg of protein) than from synaptosomes from the CA1 area or the dentate gyrus or from the whole hippocampus (0.49-0.68 nmol/ min/mg of protein). The catabolism of AMP (10 microM) was equally inhibited (85-92%) in synaptosomes from whole hippocampus, CA1, CA3, or dentate gyrus by alpha,beta-methylene-ADP (100 microM) and equally unaffected by p-nitrophenyl phosphate (0.5 mM) or rabbit IgGs (100 microg/ml). However, the antiserum against ecto-5'-nucleotidase (100 microg/ml) inhibited extracellular AMP catabolism by 44% in CA3 synaptosomes but had little or no effect in synaptosomes from CA1, dentate gyrus, or whole hippocampus. A similar difference in the inhibitory potential of the antibody was observed between fractions of isolated 5'-nucleotidase binding to concanavalin A-Sepharose (70%) and fractions not retained by the lectin column (18%). Taken together, these results suggest that immunological isoforms of ecto-5'-nucleotidase exist in the rat hippocampal nerve terminals, with predominance in the CA3 area.     

Production of Adenosine from Extracellular ATP at the Striatal Cholinergic Synapse

Abstract: The components of the ectonucleotidase pathway at the immunoaffinity-purified striatal cholinergic synapse have been studied. The ecto-ATPase (EC 3.6.1.15) had a K _m of 131 γ M , whereas the ecto-ADPase (EC 3.6.1.6) had a K _m of 58 γ M , was Ca²⁺-dependent, and was inhibited by the ATP analogue 5'-adenylylimidodiphosphate (AMPPNP). The ecto-5'-nucleotidase (EC 3.1.3.5) had a K _m of 21 γ M , was inhibited by AMPPNP and α,β-methylene ADP, and by a specific antiserum. The V _max values of the ATPase, ADPase, and 5'-nucleotidase enzymes present at this synapse were in a ratio of 30:14:1. Very little ecto-adenylate kinase activity was detected on these purified synapses. The intraterminal 5'-nucleotidase enzyme, which amounted to 40% of the total 5'-nucleotidase activity, was inhibited by AMPPNP, α,β-methylene ADP, and the antiserum, and also had the same kinetic properties as the ectoenzyme. The time course of ATP degradation to adenosine outside the nerve terminals showed a delay, followed by a period of sustained adenosine production. The delay in adenosine production was proportional to the initial ATP concentration, was a consequence of feedforward inhibition of the ADPase and 5'-nucleotidase, and was inversely proportional to the ecto-5'-nucleotidase activity. The function and characteristics of this pathway and the central role of 5'-nucleotidase in the regulation of extraterminal adenosine concentrations are discussed.     

Modification by arachidonic acid of extracellular adenosine metabolism and neuromodulatory action in the rat hippocampus

Adenosine and arachidonate (AA) fulfil opposite modulatory roles, arachidonate facilitating and adenosine inhibiting cellular responses. To understand if there is an inter-play between these two neuromodulatory systems, we investigated the effect of AA on extracellular adenosine metabolism in hippocampal nerve terminals. AA (30 microm) facilitated by 67% adenosine evoked release and by 45% ATP evoked release. These effects were not significantly modified upon blockade of lipooxygenase or cyclooxygenase and were attenuated (52-61%) by the protein kinase C inhibitor, chelerythrine (6 microm). The ecto-5'-nucleotidase inhibitor, alpha,beta-methylene ADP (100 microm), caused a larger inhibition (54%) of adenosine release in the presence of AA (30 microm) compared with control (37% inhibition) indicating that the AA-induced extracellular adenosine accumulation is mostly originated from an increased release and extracellular catabolism of ATP. This AA-induced extracellular adenosine accumulation is further potentiated by an AA-induced decrease (48%) of adenosine transporters capacity. AA (30 microm) increased by 36-42% the tonic inhibition by endogenous extracellular adenosine of adenosine A(1) receptors in the modulation of acetylcholine release and of CA1 hippocampal synaptic transmission in hippocampal slices. These results indicate that AA increases tonic adenosine modulation as a possible feedback loop to limit AA facilitation of neuronal excitability.     

Effect of Acute Ethanol on Release of Endogenous Adenosine from Rat Cerebellar Synaptosomes

The effects of pharmacologically relevant concentrations of ethanol on the release of endogenous adenosine from rat cerebellar synaptosomes were investigated. Release was conducted for 5, 10, 30, or 60 s after which time the incubation medium (containing the released adenosine) was rapidly separated from the synaptosomal membranes by vacuum filtration. The adenosine content of the filtrate was measured by HPLC-fluorescence detection. Both basal and KCl-stimulated adenosine release consisted of an initial rapid phase, for the first 10 s, that was followed by a relatively slower phase. Basal endogenous adenosine release was estimated as 199 +/- 14 pmol/mg protein/5 s. Potassium (chloride) increased adenosine release from the basal level to 433 +/- 83 pmol/mg protein/5 s. Ethanol caused a dose-dependent increase of adenosine release. The interaction between dilazep and ethanol indicates that ethanol-stimulated release does not involve the dilazep-sensitive transport system. The results support previous findings that indicate that cerebellar adenosine is involved in the mediation of ethanol-induced motor disturbances in the rat.     

Adenosine Formation and Release by Embryonic Chick Neurons and Glia in Cell Culture

Adenosine formation and release were studied in 48-h-old cultured ciliary ganglia and confluent peripheral and CNS glial cultures from embryonic chicks. Metabolic poisoning induced by 30 mM 2-deoxyglucose and 2 micrograms/ml oligomycin reduced ATP concentration by 90%. An increase in adenosine accounted for 15-40% of the fall in ATP. Dilazep (3 X 10(-6) M), a nucleoside transport inhibitor, decreased both incorporation of adenosine (an index of nucleoside transport) and release of adenosine by 80-90%. Dilazep trapped the newly formed adenosine intracellularly. A concentration of alpha, beta-methylene ADP that inhibited ecto-5'-nucleotidase by 80-90% did not alter the concentration of adenosine or AMP in neurons, glia, or medium. The results demonstrate that adenosine is formed intracellularly and exported out of the cell via the nucleoside transporter. The participation of ecto-5'-nucleotidase was excluded.     

Absolute rates of adenosine formation during ischaemia in rat and pigeon hearts.

1. The activities of ecto- and cytosolic 5'-nucleotidase (EC 3.1.3.5), adenosine kinase (EC 2.7.1.20), adenosine deaminase (EC 3.5.4.4) and AMP deaminase (EC 3.5.4.6) were compared in ventricular myocardium from man, rats, rabbits, guinea pigs, pigeons and turtles. The most striking variation was in the activity of the ecto-5'-nucleotidase, which was 20 times less active in rabbit heart and 300 times less active in pigeon heart than in rat heart. The cytochemical distribution of ecto-5'-nucleotidase was also highly variable between species. 2. Adenosine formation was quantified in pigeon and rat ventricular myocardium in the presence of inhibitors of adenosine kinase and adenosine deaminase. 3. Both adenosine formation rates and the proportion of ATP catabolized to adenosine were greatest during the first 2 min of total ischaemia at 37 degrees C. Adenosine formation rates were 410 +/- 40 nmol/min per g wet wt. in pigeon hearts and 470 +/- 60 nmol/min per g wet wt. in rat hearts. Formation of adenosine accounted for 46% of ATP plus ADP broken down in pigeon hearts and 88% in rat hearts. 4. The data show that, in both pigeon and rat hearts, adenosine is the major catabolite of ATP in the early stages of normothermic myocardial ischaemia. The activity of ecto-5'-nucleotidase in pigeon ventricle (16 +/- 4 nmol/min per g wet wt.) was insufficient to account for adenosine formation, indicating the existence of an alternative catabolic pathway.     

Isolation of Nerve Terminals from Crustacean Muscle

A method for the isolation of gamma-aminobutyric acidergic (GABAergic) and glutamatergic terminals from crustacean muscle was developed, using differential centrifugation and sucrose density gradient centrifugation. Individual fractions were assessed using a variety of markers. One fraction was isolated which showed 40-fold purification of glutamate decarboxylase with a yield of 12%. This fraction was enriched in GABA, glutamate, glutamate dehydrogenase, and 5'-nucleotidase, but not in NADPH cytochrome c reductase. This fraction possessed an uptake system for GABA and glutamate with apparent kinetic constants of Km = 50 microM, Vmax = 250 pmol/min/mg of protein and Km = 183 microM, Vmax = 219 pmol/min/mg of protein, respectively. Electron microscopy showed nerve terminal profiles and a heterogeneous population of membrane vesicles. This fraction contained 3.4 nmol ATP/mg of protein which was stable for 30 min at 12 degrees C, and was also able to synthesise ATP from exogenous adenosine. The terminals released labelled GABA and glutamate in a Ca2+-dependent fashion on depolarisation. No release of ATP was detected. It is concluded that viable nerve terminals have been isolated which could be used as model systems for the study of GABAergic and glutamatergic neurochemistry.     

Glutamate-Evoked Release of Endogenous Adenosine from Rat Cortical Synaptosomes Is Mediated by Glutamate Uptake and Not by Receptors

L-Glutamate (10 microM-1 mM) released endogenous adenosine from rat cortical synaptosomes. Studies with excitatory amino acid antagonists, (+)-5-methyl-16,11,dihydro-5H- dibenzo[a,d]cyclohepten-5,10-imine maleate (MK-801), 6,7-dinitroquinoxaline-2,3-dione (DNQX), Mg2+, and agonists N-methyl-D-aspartate (NMDA), kainate, and quisqualate, indicated that this release was not receptor mediated. D,L-2-Amino-4-phosphonobutanoic acid (APB) also did not affect glutamate-evoked adenosine release. Inhibition of glutamate uptake by dihydrokainate or replacement of extracellular Na+ blocked glutamate-evoked adenosine release. D-aspartate, which is a substrate for the glutamate transporter but is not metabolized, also released adenosine, suggesting that release was due to amino acid transport and not to its subsequent metabolism. D-Glutamate, a relatively poor substrate for the transporter, was correspondingly less potent than L-glutamate at releasing adenosine. Glutamate-evoked adenosine release was not Ca2+ dependent or tetrodotoxin sensitive and did not appear to occur on the bidirectional nucleoside transporter. Inhibition of ecto-5'-nucleotidase virtually abolished glutamate-evoked adenosine release, indicating that adenosine was derived from extracellular metabolism of released nucleotide(s). However, L-glutamate did not release ATP and did not appear to release cyclic AMP. Therefore, transport of glutamate into presynaptic terminals releases some other nucleotide which is converted extracellularly to adenosine. This adenosine could act at P1-purinoceptors to modulate glutamatergic neurotransmission.     

Ecto-5'-nucleotidase of cultured rat mesangial cells

Because adenosine plays a role in the regulation of glomerular filtration rate and of the release of renin, we examined the possibility of a local source for this mediator. We found that rat cultured glomerular mesangial cells converted 5'-AMP into adenosine. The properties of the enzyme involved in the reaction were those of an ecto-5' nucleotidase: (1) the products of the reaction were generated in the extracellular fluid although no 5'-nucleotidase was released by the cells into the medium; (2) identical activities were found for cultured cells in situ and sonicated cells; (3) the diazonium salt of sulfanilic acid which is a nonpenetrating reagent inhibited up to 75% of the enzyme activity. Ecto-5'-nucleotidase activity of intact cells obeyed Michaelis-Menten kinetics. Apparent Km for 5'-AMP was 0.32 mM. 5'-UMP was a strictly competitive inhibitor. ADP exerted a very powerful inhibitory effect and behaved also as a competitive inhibitor. ATP was inhibitory both by increasing Km and by decreasing Vmax. Ecto-5'-nucleotidase was active in the absence of divalent cations. However, Mg2+, Ca2+, Co2+ and Mn2+ were stimulatory. Zn2+ and Cu2+ suppressed the activity. Concanavalin A, a plant lectin, was markedly inhibitory, suggesting that a glycoprotein moiety was necessary to express enzyme activity. Ecto-5'-nucleotidase activity was not modified during phagocytosis of serum-treated zymosan by mesangial cells. Rat cultured glomerular epithelial cells exhibited a 5'-nucleotidase activity which was 4 times lower than that of the mesangial cells in primary culture.(ABSTRACT TRUNCATED AT 250 WORDS)     

ATP Release, Adenosine Formation, and Modulation of Dynorphin and Glutamic Acid Release by Adenosine Analogues in Rat Hippocampal Mossy Fiber Synaptosomes

Using a hippocampal subcellular fraction enriched in mossy fiber synaptosomes, evidence was obtained indicating that adenosine derived from a presynaptic pool of ATP may modulate the release of prodynorphin-derived peptides. and glutamic acid from mossy fiber terminals. Synaptosomal ATP was released in a Ca2+-dependent manner by K+-induced depolarization. The rapid hydrolysis of extracellular [14C]ATP in the presence of intact mossy fiber synaptosomes resulted in the production of [14C]adenosine. Micromolar concentrations of a stable adenosine analogue, 2-chloroadenosine, inhibited the K+-stimulated release of both dynorphin B and dynorphin A(1-8). 2-Chloroadenosine failed to suppress the evoked release of glutamic acid, measured in these same superfusates, unless the mossy fiber synaptosomes were pretreated with D-aspartic acid to deplete the cytosolic, Ca2+-independent, pool of this acidic amino acid. In synaptosomes pretreated in this manner, release of the remaining Ca2+-dependent pool of glutamic acid was significantly inhibited by NiCl2, 2-chloroadenosine, 5'-N-ethylcarboxamidoadenosine, cyclohexyladenosine, and R(-)-N6(2-phenylisopropyl)adenosine, but not by ATP. 2-Chloroadenosine-induced inhibition was reversed when the external CaCl2 concentration was raised from 1.8 mM to 6 mM. 8-Phenyltheophylline, an adenosine receptor antagonist, effectively blocked the inhibitory effects of 2-chloroadenosine on mossy fiber synaptosomes and significantly enhanced the K+-evoked release of both glutamic acid and dynorphin A(1-8) when added alone to the superfusion medium. These results support the proposition that depolarized hippocampal mossy fiber synaptosomes release endogenous ATP and are capable of forming adenosine from extracellular ATP, and that endogenous adenosine may act at a presynaptic site to inhibit the further release of glutamic acid and the prodynorphin-derived peptides.     

Adenosine formation and release from neonatal-rat heart cells in culture.

The incorporation of [3H]adenosine (10 microM) into neonatal-rat heart cell nucleotides was inhibited in a concentration-dependent manner, such that 50% inhibition was obtained with 0.75 microM-dipyridamole, 0.26 microM-hexobendine or 0.22 microM-dilazep. Adenosine formation was accelerated 2.5-fold to 2.1 +/- 0.3 nmol/10(7) cells in 10 min when cells were incubated with a combination of 30 mM-2-deoxyglucose and 2 micrograms of oligomycin/ml. Of the newly formed adenosine, 6 +/- 2% was in the cells. Dipyridamole, hexobendine or dilazep (10 microM) increased the amount of adenosine in the cells and decreased that in the medium such that 45-50% of the newly formed adenosine was in the cells. Antibodies which inhibited ecto-5'-nucleotidase by 98.7 +/- 0.3% did not alter the rate of adenosine formation or its distribution between cells and medium. We conclude that adenosine was formed in the cytoplasm during catabolism of cellular ATP and was released via the dipyridamole-sensitive symmetric nucleoside transporter.