Modulation of GABA transport by adenosine A1R-A2AR heteromers, which are coupled to both Gs- and G(i/o)-proteins.

Astrocytes play a key role in modulating synaptic transmission by controlling the available extracellular GABA via the GAT-1 and GAT-3 GABA transporters (GATs). Using primary cultures of rat astrocytes, we show here that an additional level of regulation of GABA uptake occurs via modulation of the GATs by the adenosine A(1) (A(1)R) and A(2A) (A(2A)R) receptors. This regulation occurs through a complex of heterotetramers (two interacting homodimers) of A(1)R-A(2A)R that signal via two different G-proteins, G(s) and G(i/o), and either enhances (A(2A)R) or inhibits (A(1)R) GABA uptake. These results provide novel mechanistic insight into how G-protein-coupled receptor heteromers signal. Furthermore, we uncover a previously unknown mechanism in which adenosine, in a concentration-dependent manner, acts via a heterocomplex of adenosine receptors in astrocytes to significantly contribute to neurotransmission at the tripartite (neuron-glia-neuron) synapse.

The heat shock cognate protein hsc73 assembles with A(1) adenosine receptors to form functional modules in the cell membrane.

A(1) adenosine receptors (A(1)Rs) are G protein-coupled heptaspanning receptors that interact at the outer face of the plasma membrane with cell surface ecto-adenosine deaminase (ecto-ADA). By affinity chromatography the heat shock cognate protein hsc73 was identified as a cytosolic component able to interact with the third intracellular loop of the receptor. As demonstrated by surface plasmon resonance, purified A(1)Rs interact specifically with hsc73 with a dissociation constant in the nanomolar range (0.5 +/- 0.1 nM). The interaction between hsc73 and A(1)R led to a marked reduction in the binding of the ligands and prevented activation of G proteins, as deduced from (35)S-labeled guanosine-5'-O-(3-thio)triphosphate binding assays. Interestingly this effect was stronger than that exerted by guanine nucleotide analogs, which uncouple receptors from G proteins, and was completely prevented by ADA. As assessed by immunoprecipitation a high percentage of A(1)Rs in cell lysates are coupled to hsc73. A relatively high level of colocalization between A(1)R and hsc73 was detected in DDT(1)MF-2 cells by means of confocal microscopy, and no similar results were obtained for other G protein-coupled receptors. Colocalization between hsc73 and A(1)R was detected in specific regions of rat cerebellum and in the body of cortical neurons but not in dendrites or synapses. Remarkably, agonist-induced receptor internalization leads to the endocytosis of A(1)Rs by two qualitatively different vesicle types, one in which A(1)R and hsc73 colocalize and another in which hsc73 is absent. These results open the interesting possibility that signaling via G protein-coupled receptors may be regulated by heat shock proteins.

Diabetes decreases sensitivity of adipocyte lipolysis to inhibition by Gi-linked receptor agonists.

(1) Streptozotocin-diabetes decreased the responsiveness of noradrenaline- or forskolin-stimulated lipolysis to inhibition by phenylisopropyladenosine (PIA), prostaglandin E1 (PGE1) and nicotinate in rat adipocytes. (2) Diabetes had no effect on high affinity binding of [3H]PIA to adipocyte plasma membranes. (3) Plasma membranes from diabetic animals had increased abundance of alpha-subunits of Gi1 and Gi2. The effect of pertussis toxin in overcoming inhibition of lipolysis by PIA was delayed in adipocytes from diabetic rats. (4) Diabetes decreased the GTP-dependent right-wards shift in the dose-curve for displacement of the antagonist [3H]DPCPX by PIA in adipocyte plasma membranes. (5) It is concluded that, despite increased abundance of Gi in diabetic adipocytes, less of this is functional. This may contribute to reduced sensitivity to PIA, PGE1 and nicotinate and explains some of the loss of control of lipolysis in insulin-dependent diabetes.

Role of adenosine receptor activation in myocardial infarct size limitation by ischaemic preconditioning.

OBJECTIVE: The aims were to examine the role of adenosine receptors in the mechanism of preconditioning in a chronic rabbit model of myocardial infarction; to assess whether the preconditioning effect is blocked by an adenosine receptor antagonist, 8-phenyltheophylline; and to determine whether an adenosine A1 receptor agonist, R(-)N6-2-phenylisopropyl adenosine (R-PIA), mimics infarct size limitation by preconditioning. METHODS: Myocardial infarction was induced in male rabbits by occlusion of the left coronary artery for 30 min, which was followed by 72 h reperfusion. Before the 30 min ischaemia, rabbits were subjected to one of the following six protocols: (1) untreated control; (2) intravenous injection of 8-phenyltheophylline; (3) preconditioning with 5 min ischaemia; (4) pretreatment with 8-phenyltheophylline plus preconditioning; (5) intravenous injection of R-PIA; or (6) R-PIA plus atrial pacing (240.min-1). Infarct size and area at risk were determined by histology and fluorescent particles, respectively. RESULTS: Preconditioning significantly limited infarct size, normalised as a percent of area at risk (%IS/AR), to 19.2 (SEM 2.3)% v control value of 46.5(2.8)%. 8-Phenyltheophylline alone did not modify the %IS/AR, but its injection before preconditioning attenuated the preconditioning effect such that IS/AR = 34.4(2.5)%. While R-PIA did not achieve statistically significant myocardial salvage, R-PIA plus atrial pacing limited infarct size to 33.7(3.0)% (p<0.05 v control). The R-PIA group had severe hypotension and their infarct sizes were inversely correlated with diastolic blood pressure at reperfusion. There was no such correlation in the R-PIA plus pacing group in which bradycardia and hypotension induced by R-PIA were attenuated by atrial pacing. CONCLUSIONS: The infarct size limiting effect of preconditioning was attenuated by 8-phenyltheophylline, and pretreatment with R-PIA was able to limit myocardial infarct size when severe hypotension was avoided by atrial pacing. These findings suggest that adenosine receptor activation plays a crucial role in the mechanism of preconditioning.

The rat testicular A1 adenosine receptor-adenylate cyclase system.

When adenosine interacts with membrane-bound A1 receptors, it is capable of inhibiting the enzyme adenylate cyclase in brain and fat tissue. In this paper we characterize the A1 adenosine receptor-adenylate cyclase system of the rat testes. The agonist radioligand (-)-N-[3-[125I]iodo-4-hydroxyphenyl-(isopropyl)adenosine binds with high affinity (Kd, congruent to 1 nM) in a saturable manner (maximum binding, congruent to 600 fmol/mg protein). The A1 adenosine receptor binding displays the appropriate pharmacology, stereospecificity, and sensitivity to guanine nucleotides. The testicular A1 receptor is also coupled in an inhibitory manner to the enzyme adenylate cyclase, as demonstrated by the ability of N6-R-phenyl-2-propyladenosine to inhibit isoproterenol- and forskolin-stimulated cAMP accumulation. The testicular A1 receptor can be solubilized in high yield and in an active form with the detergent digitonin. The solubilized receptor retains all of the pharmacological properties of the membrane-bound receptor. Although there are many similarities among the A1 receptor from testes, brain, and fat tissue, the testicular A1 receptor displays a larger apparent mol wt. By photoaffinity labeling, the A1 adenosine receptor-binding subunit of fat and brain are 38,000 mol wt proteins, while the testicular A1 receptor binding subunit is a 42,000 mol wt protein.

Adenosine receptors of cerebral microvessels and choroid plexus.

We studied, by ligand binding methods, the two adenosine receptors, A1 and A2, in rat and pig cerebral microvessels and pig choroid plexus. Ligand binding to cerebral microvessels was compared with that to membranes of the cerebral cortex. [3H]Cyclohexyladenosine and [3H]L-phenylisopropyladenosine were the ligands used for A1-receptors, and [3H]5'-N-ethylcarboxamide adenosine ([3H]NECA) was used to assess A2-receptors. We report that cerebral microvessels and choroid plexus exhibit specific [3H]NECA binding, but have no appreciable A1-receptor ligand binding sites. Specific binding of [3H]NECA to cerebral microvessels, choroid plexus, and cerebral cortex was saturable and suggested the existence of two classes of A2-receptor sites: high-affinity (Kd approximately 250 nM) and low-affinity (Kd approximately 1-2 microM) sites. The Kd and Bmax of NECA binding to cerebral microvessels and cerebral cortex were similar within each species. Our results, indicating the existence of A2-receptors in cerebral microvessels, are consistent with results of increased adenylate cyclase activity by adenosine and some of its analogues in these micro-vessels.

[3H]N6-(L-Phenylisopropyl) adenosine binding in brains from young and old rats.

The binding of [3H]N6-(L-Phenylisopropyl) adenosine (L-PIA) to membrane preparations of whole brains from normal male Sprague-Dawley rats 12 and 84 weeks of age, respectively, was examined. Two populations of binding sites, probably corresponding to A1 and A2 adenosine receptors, were detected in both young and old rats. No statistically significant differences between young and old rats were detected but both the numbers of binding sites (Bmax) and dissociation constants (KD) for both high and low affinity binding sites were greater in 84 week old rats. These results were compared to earlier studies of adenosine receptors and related to previously reported changes in sleep with aging in rats.

The binding of [3H]R-PIA to A1 adenosine receptors produces a conversion of the high- to the low-affinity state.

Kinetic evidence for negative cooperativity on the binding of [3H]R-PIA to A1 adenosine receptors was obtained from dissociation experiments at different ligand concentrations and from the equilibrium isotherm. The dissociation curves indicate that there is an apparent ligand-induced transformation of high- to low-affinity states of the receptor. At concentrations of 18.2 nM R-PIA or higher there was only found the low-affinity state of the receptor. In view of these results equilibrium binding data were analyzed by the usual two-state model (assuming that there is an interconversion between them) and by the negative cooperativity model employing the Hill equation.

Photochemical cross-linking of 125I-hydroxyphenylisopropyl adenosine to the A1 adenosine receptor of rat adipocytes.

Rat adipocyte plasma membranes were incubated with the A1 adenosine receptor agonist, 125I-hydroxyphenylisopropyl adenosine (1 nM) and then treated with the photoactive cross-linking agent, ANB-NOS. The membranes were solubilized and analyzed by SDS-PAGE and autoradiography. A single protein, Mr approx. 38,000, was specifically labeled. Reduction with 2-mercaptoethanol did not affect the apparent Mr of the labeled protein. Labeling was inhibited by the adenosine receptor agonists, HPIA, PIA and NECA, and by the antagonist, theophylline, but was unaffected by inosine. We conclude that the A1 adenosine receptor is a single protein of Mr approx. 38,000.

Adenosine deaminase affects ligand-induced signalling by interacting with cell surface adenosine receptors.

Adenosine deaminase (ADA) is not only a cytosolic enzyme but can be found as an ecto-enzyme. At the plasma membrane, an adenosine deaminase binding protein (CD26, also known as dipeptidylpeptidase IV) has been identified but the functional role of this ADA/CD26 complex is unclear. Here by confocal microscopy, affinity chromatography and coprecipitation experiments we show that A1 adenosine receptor (A1R) is a second ecto-ADA binding protein. Binding of ADA to A1R increased its affinity for the ligand thus suggesting that ADA was needed for an effective coupling between A1R and heterotrimeric G proteins. This was confirmed by the fact that ASA, independently of its catalytic behaviour, enhanced the ligand-induced second messenger production via A1R. These findings demonstrate that, apart from the cleavage of adenosine, a further role of ecto-adenosine deaminase on the cell surface is to facilitate the signal transduction via A1R.

Solubilization and partial characterization of adenosine binding sites from rat brainstem.

Binding sites for adenosine were solubilized from rat brainstem membranes with either sodium cholate, sodium deoxycholate or 3-[(3-cholamidopropyl)dimethyl-ammonio]-1-propanesulfonate. About 30% of the binding activity were released by [3H]phenylisopropyladenosine (PIA) binding. Specific [3H]PIA binding to the solubilized fraction was saturable and was found to be a monophasic saturation profile. In contrast, [3H]PIA binding to the brainstem membranes exhibited a biphasic profile suggesting the presence of two binding sites. By gel filtration on a Sepharose CL-6B column, the adenosine binding site--detergent complex was estimated to have app. Mr 280000 and rs = 5.4 nm.

Species differences in structure-activity relationships of adenosine agonists and xanthine antagonists at brain A1 adenosine receptors.

A series of 28 adenosine analogs and 17 xanthines has been assessed as inhibitors of binding of N6-R-[3H]phenylisopropyladenosine binding to A1 adenosine receptors in membranes from rat, calf, and guinea pig brain. Potencies of N6-alkyl- and N6-cycloalkyladenosines are similar in the different species. However, the presence of an aryl or heteroaryl moiety in the N6 substituent results in marked species differences with certain such analogs being about 30-fold more potent at receptors in calf than in guinea pig brain. Potencies at receptors in rat brain are intermediate. Conversely, 2-chloroadenosine and 5'-N-ethylcarboxamidoadenosine are about 10-fold less potent at receptors in calf brain than in guinea pig brain. Potencies of xanthines, such as theophylline, caffeine and 1,3-dipropylxanthine are similar in the different species. However, the presence of an 8-phenyl or 8-cycloalkyl substituent results in marked species differences. For example, a xanthine amine conjugate of 1,3-dipropyl-8-phenylxanthine is 9-fold more potent at receptors in calf than in rat brain and 110-fold more potent in calf than in guinea pig brain. Such differences indicate that brain A1 adenosine receptors are not identical in recognition sites for either agonists or antagonists in different mammalian species.

Adenosine analogs with covalently attached lipids have enhanced potency at A1-adenosine receptors.

Chemically functionalized congeners of N6-phenyladenosine and 1,3-dipropyl-8-phenylxanthine have been covalently coupled to fatty acids, diglycerides, and a phospholipid. The lipid-drug conjugates inhibit R-[3H]-phenylisopropyladenosine binding to A1-adenosine receptors in rat cerebral cortex membranes. A xanthine-phosphatidylethanolamine conjugate bound with a Ki value of 19 nM. Various xanthine esters of low potency are potential prodrugs. Amides of an adenosine amine congener (ADAC) with 18-carbon fatty acids exhibited Ki values at A1-adenosine receptors of 70 pM, representing a 130-fold enhancement over the affinity of the corresponding acetyl amide. The very high affinity of adenosine-lipid conjugates may be due to stabilization of these adducts in the phospholipid microenvironment of the receptor protein.

Chronic theophylline exposure increases agonist and antagonist binding to A1 adenosine receptors in rat brain.

Chronic treatment of rats with theophylline (75mg/kg twice-daily for 21 consecutive days) significantly increased the specific binding of [3H]CHA and [3H]DPCPX in cerebral cortical membranes. The absolute increase in the number of binding sites following theophylline treatment was approximately the same for each ligand, although this number represents a larger percentage of the total sites available to [3H]CHA. Saturation analysis of [3H]DPCPX binding indicated that theophylline treatment increased the maximum number of binding sites from 799 +/- 13 to 920 +/- 22 fmol/mg protein, while the affinity of [3H]DPCPX for A1 receptors was unaltered. These results suggest that chronic theophylline exposure produces both an increase in the number of A1 adenosine receptors and an enhancement of coupling of these receptors to G proteins.

The effect of soluflazine on adenosine receptors in the rat brain.

Soluflazine, a potent adenosine transport inhibitor, was intracerebroventricularly administered to rats via ALZET mini osmotic pumps (4nmole, 0.5 L/hr) for 14 days and the effect on adenosine receptors was determined in specific brain areas. Soluflazine decreased adenosine A1 radioligand binding in the hippocampus as measured by [3H]R-PIA, and lowered adenosine A2 binding sites in the striatum, as estimated by the "NECA minus R-PIA" assay. Previous work from our lab has shown the ability of diazepam and triazolam to decrease adenosine binding in the same brain areas. The data show that a specific adenosine transport inhibitor produces the same effect on adenosine receptors as benzodiazepines, and suggest a role for adenosine in the CNS effects of benzodiazepines.

Structure-activity relationships of 1,3-dialkylxanthine derivatives at rat A3 adenosine receptors.

1,3-Dialkylxanthine analogues containing carboxylic acid and other charged groups on 8-position substituents were synthesized. These derivatives were examined for affinity in radioligand binding assays at rat brain A3 adenosine receptors stably expressed in CHO cells using the new radioligand [125I]AB-MECA (N6-(4-amino-3-iodobenzyl)adenosine-5'-N-methyluronamide), and at rat brain A1 and A2a receptors using [3H]PIA and [3H]CGS 21680, respectively. A synthetic strategy for introducing multiple carboxylate groups at the 8-position using iminodiacetic acid derivatives was explored. The presence of a sulfonate, a carboxylate, or multiple carboxylate groups did not result in a significant enhancement of affinity at rat A3 receptors, although as previously observed an anionic group tended to diminish potency at A1 and A2a receptors. The rat A3 receptor affinity was not highly dependent on the distance of a carboxylate group from the xanthine pharmacophore. 2-Thio vs 2-oxo substitution favored A3 potency, and 8-alkyl vs 8-aryl substitution favored A3 selectivity, although few derivatives were truly selective for rat A3 receptors. 1,3-Dimethyl-8-(3-carboxypropyl)-2-thioxanthine was 7-fold selective for A3 vs A2a receptors. 1,3,7-Trimethyl-8-(trans-2-carboxyvinyl)xanthine was somewhat selective for A3 vs A1 receptors. For 8-arylxanthines affinity at A3 receptors was enhanced by 1,3-dialkyl substituents, in the order dibutyl > dipropyl > diallyl.

Synthesis and structure-activity relationship of pyrazolo[3,4-d]pyrimidines: potent and selective adenosine A1 receptor antagonists.

A series of 12 substituted 1-phenylpyrazolo[3,4-d]pyrimidines were synthesized and evaluated for rat brain adenosine A1 and A2a receptor binding affinity. Substituents at C-4 and C-6 were varied in order to define these regions in terms of molecular recognition by the receptor subtypes. At C-4, the effects of a mercapto, methylthio, and amino substituent were evaluated, while at C-6, amides with varying alkyl groups extending from the alpha-carbon were examined. This study identified both potent and selective adenosine A1 receptor antagonists. The most potent of the 12 compounds was alpha-[(4-amino-1-phenylpyrazolo[3,4-d]pyrimidin-6-yl)thio]hexanamide (14); with an A1 Ki of 0.939 nM and an A2a Ki of 88.3 nM, this compound is 94-fold A1 selective. The most selective of the 12 compounds was alpha-[[4-(methylthio)-1-phenylpyrazolo[3,4-d]pyrimidin-6-yl]thio]hex anamide (10); with an A1 Ki of 6.81 nM and an A2a Ki > 40 000 nM, this compound is > 5900-fold A1 selective. The structure-activity relationships for the complete series has identified discrete structural differences between the A1 and A2a receptors with respect to the binding of pyrazolo[3,4-d]pyrimidines. This study resulted in prediction that increased A1 affinity could be achieved by incorporation of NH-alkyl substituents at C-4. This was confirmed by synthesis of alpha-[[4-(methylamino)-1-phenylpyrazolo[3,4-d]pyrimidin-6-yl]thiol] hexanamide (15) which was found to have an A1 Ki of 0.745 nM.

8-Aryl-and 8-cycloalkyl-1,3-dipropylxanthines: further potent and selective antagonists for A1-adenosine receptors.

A series of 1,3-dipropylxanthines were prepared with a variety of substituents at the 8-position. These included 8-aryl and 8-cycloalkyl groups. Polar carboxylate and carboxamide moieties were introduced as aryl substituents to increase water solubility. 1,3-Dipropyl-8-[2-hydroxy-4-[(carboxymethyl)oxy]phenyl]xanthine provided a functionalized congener with high potency (Ki = 37 nM) and selectivity (54-fold) for A1-adenosine receptors. This congener was used for preparation of a series of other analogues, some with higher potency and some with higher selectivity. 8-Cyclopentyl- and 8-cyclohexyl-1,3-dipropylxanthines were both very potent (Ki = 1-1.5 nM) and selective for A1 receptors, while 8-cycloalkylmethyl analogues were 10-fold less potent, but still very selective for A1 receptors. 8-Piperidinyl and 8-pyrazinyl analogues had very low activities as adenosine receptor antagonists.

Imidazodiazepinediones: a new class of adenosine receptor antagonists.

A series of imidazo[4,5-e][1-4]diazepine-5,8-diones were synthesized from hypoxanthines. Certain of these cyclic homologues of caffeine, theophylline, theobromine, 3-isobutyl-1-methylxanthine, and enprofylline were inhibitors of binding of adenosine analogues to rat brain A1 and A2 adenosine receptors and were antagonists of A2 adenosine receptors stimulatory to adenylate cyclase in rat PC12 cell membranes. Activity at adenosine receptors was lower than the corresponding xanthines, perhaps because imidazodiazepinediones contain a boat-shaped seven-membered ring rather than the planar heteroaryl ring system of the xanthines. The imidazodiazepinediones had low affinity for brain benzodiazepine sites.

7-Deaza-2-phenyladenines: structure-activity relationships of potent A1 selective adenosine receptor antagonists.

A series of derivatives of 7-deazapurines with varying substituents in the 2-, 6-, and 9-position was synthesized in an attempt to improve the adenosine receptor affinity and A1 or A2 selectivity. The adenosine receptor affinities were assessed by measuring the inhibition of [3H]-(R)-N6-(phenylisopropyl) adenosine (R-PIA) binding to rat brain A1 and inhibition of [3H]-5'-(N-ethylcarboxamido)adenosine (NECA) binding to rat striatum A2 adenosine receptors. A selected set of compounds representing the main structural variations was further examined in adenosine receptor coupled adenylate cyclase assays. All tested compounds antagonized the inhibition of adenylate cyclase elicited by interaction of R-PIA with A1 receptors in rat fat cell membranes and the activation of adenylate cyclase elicited by interaction of NECA with A2 receptors of pheochromocytoma PC12 cell membranes. The results indicate that 7-deazahypoxanthines have a potential for A2 selectivity, while all 7-deazaadenines are A1 selective. Introduction of a phenyl residue in the 2-position of 7-deazaadenines increases A1 activity tremendously. 2-(p-Chlorophenyl)-7,8-dimethyl-9-phenyl-7-deazaadenine (29) is potent and specific for the A1 receptors of rat brain (Ki = 122 nM), having no affinity for the A2 receptors of rat striatum. The compound has low activity at the A2 receptors of rat PC12 cell membranes where it appears to act as a noncompetitive inhibitor. A 1-phenylethyl substituent at the 9-position was found to be superior to a phenyl residue in terms of A1 affinity. The most potent A1 antagonist in the present series is the highly A1 selective (790-fold) (R)-7,8-dimethyl-2-phenyl-9-(1-phenylethyl)-7-deazaadenine (31, Ki = 4.7 nM), which is 30-35 times more potent at A1 receptors than its S enantiomer. The solubility of six of the potent 7-deaza-2-phenyladenines was determined by means of an A1 binding assay. Chloro substitution of the 2-phenyl ring appeared to improve the solubility as well as the solubility over A1 affinity ratio of 9-phenyl- and 9-(1-phenylethyl)-substituted 7-deazadenines.