Adenosine A, Receptor-Mediated Inhibition of Adenylate ... · cellular adenosine receptors. There have been relatively few reports of adenosine effects on visual biochemistry or physiology

The Journal of Neuroscience, August 1987, 7(8): 2522-2528

Adenosine A, Receptor-Mediated Inhibition of Adenylate Cyclase in Rabbit Retina

Christine Blazynski

Department of Ophthalmology, Washington University School of Medicine, St. Louis, Missouri 63110

The purine adenosine has been postulated as playing a role in CNS neurotransmission or modulation. Evidence is now provided that inhibition of adenylate cyclase in rabbit retinal homogenates is mediated vfa adenosine A, receptors. Nano- molar concentrations of the A, receptor agonists, cyclohexyladenosine (CHA) and phenylisopropyladenosine (PIA), significantly inhibited the activity of forskolin-stimulated adenylate cyclase in preparations in which endogenous adenosine was destroyed by pretreatment with adenosine deaminase. With increasing concentrations of either agonist, biphasic effects on enzyme activity were observed. The effect of the mixed A,-A, agonist, Kethyl-carboxamidoadenosine (NECA), on forskolin-stimulated, as well as basal adenylate cyclase, was also investigated. At micromolar concentrations, NECA significantly increased the activity of adenylate cyclase. lsobutylmethylxanthine, a potent antagonist at extracellular adenosine receptors, blocked the effects observed with PIA, CHA, and NECA. The uptake of both 3H-CHA and 3H-adenosine into retinal cells was demonstrated autoradiographically. Both agonists labeled ganglion cell bodies and certain cell bodies located in the proximal portion of the inner nuclear layer.

Evidence that purines, particularly adenosine, play a role in neurotransmission or modulation in the CNS continues to accumulate. Specific uptake systems (Shimizu et al., 1972; Bender et al., 198 1; Marangos, 1984; Geiger et al., 1985), as well as endogenous release of adenosine, have been demonstrated (Pull and McIlwain, 1972; Fredholm and Hedqvist, 1980). Autora- diographic localization of specific binding sites has been reported for a variety of preparations (Goodman and Snyder, 1982; Goodman et al., 1983; Marangos, 1984; Geiger, 1986). Specific localizations of both degradative and synthetic enzymes related to adenosine metabolism-adenosine deaminase-like (Geiger and Nagy, 1984; Nagy et al., 1984; Nagy et al., 1985) and 5’-nucleotidase-like immunoreactive neurons, respectively-have been reported (Marani, 1977; Schubert et al., 1979). Using an antibody that appears to be highly specific for adenosine, Braas et al. (1986) demonstrated specific localization to

Received Dec. 8, 1986; revised Feb. 11, 1987; accepted Feb. 20, 1987. This work was supported by N.I.H. Grants EY-00258 (Adolph I. Cohen), and

EY-02294 (James A. Ferrendelli). I wish to thank Prof. Adolph I. Cohen for his generous support and critically reading this manuscript. I wish also to thank Prof. James A. Ferrendelli for helpful comments. The technical assistance of Mrs. C. A. Beaty and Mrs. Z. J. Jones is gratefully acknowledged.

Correspondence should be addressed to Dr. Christine Blazynski, Department of Ophthalmology, Box 8096, Washington University School of Medicine, 660 South Euclid Ave., St. Louis, MO 63 110. Copyright 0 1987 Society for Neuroscience 0270-6474/87/082522-07$02.00/O

both the cytoplasm of perikarya and fibers in specific neuronal cell groups of discrete rat brain regions. Adenosine elicits elec- trophysiological effects such as inhibition of both evoked and spontaneous neuronal firing in brain slices (Phillis and Wu, 198 1) and, as demonstrated recently, activates a potassium conductance in cultured striatal neurons (Trussell and Jackson, 1985). Presynaptic inhibition of neurotransmitter release by adenosine has been demonstrated for many in vitro preparations (for review, see Phillis and Barraco, 1985).

Several types of adenosine receptors have been identified: an intracellular receptor, the P receptor, and 2 extracellular types include a high-affinity receptor (A,) and a low-affinity receptor (A,). Recently, it has been suggested that an additional receptor exists, the A, receptor (for review, see Ribeiro and Sebastiao, 1986). Binding of adenosine to the A, receptor stimulates adenylate cyclase, while A, binding inhibits the cyclase (Cooper et al., 1980; Londos et al., 1980). As is the case for hormone- sensitive adenylate cyclases, activation or inhibition by adenosine is mediated via the guanine nucleotide binding proteins, G, or G,, respectively (for review, see Cooper et al., 1985). Identification of receptors has been greatly aided by the development of relatively specific radiolabeled agonists that prefer- entially bind to the A, receptor with k,‘s in the low nanomolar range: phenylisopropyladenosine (PIA) and cyclohexyladenosine (CHA). Other agonists bind to both A, and A, receptors: N-ethyl-carboxamidoadenosine (NECA), 2-chloroadenosine, and adenosine itself. Xanthines are potent antagonists at all extracellular adenosine receptors.

There have been relatively few reports of adenosine effects on visual biochemistry or physiology. Adenosine receptors, localized to terminals of excitatory neurons projecting to the mouse superior colliculus, appear to be located on ganglion cell processes (Goodman et al., 1983). In retina, the presence of adenosine receptors is suggested by several reports. Stimulation of adenylate cyclase by adenosine in embryonic chick retinas (Paes de Carvalho and de Mello, 1982), as well as inhibition of a dopamine-stimulated adenylate cyclase by adenosine (Paes de Carvalho and de Mello, 1985), has been reported. In rabbit retina, a potent A, system has been identified (Blazynski et al., 1986); adenosine-elicited increases in CAMP levels were greater than that seen for any other putative neurotransmitter agent(s) tested. The increase in CAMP was completely blocked by 3-isobutylmethylxanthine (IBMX). Autoradiographic localization of adenosine uptake has been demonstrated in rabbit and carp retinas (Ehinger and Perez, 1984). In rabbit, ganglion cells and their axons, as well as Miiller cells, and certain cell bodies located in the inner nuclear layer accumulated 3H-adenosine, 3H-PIA, and XH-CHA. No specific labeling was observed with

The Journal of Neuroscience, August 1987, 7(8) 2523

Pretreatment

I I I I I I J 0 -10 -9 -8 -7 -6

LOG CCHAI

Figure 1. Inhibition of adenylate cyclase activity by cyclohexyladenosine. Homogenates of rabbit retina were assayed for cyclase activity in the presence of increasing concentrations of CHA. Basal levels of activity were stimulated with 5 PM forskolin. Retinal homogenates were either used directly (0) or incubated at room temperature with adenosine deaminase (ADA), 1 U/cc, for 20-30 min prior to initiation of the assay (A). Data represent means of 3 experiments run in triplicate f SEM.

either 3H-NECA or 3H- 1,3-diethyl-8-phenylxanthine, an antagonist. In carp retina, photoreceptor cell bodies were heavily labeled, while horizontal cell bodies, some amacrine cells, and ganglion cells were lightly labeled but only when 3H-adenosine was used. Adenosine-like immunoreactive sites have been reported for rat retina (Zarbin et al., 1986); the highest labeling was reported for ganglion cells and nerve fiber layer, and in the photoreceptor outer segments. Moderate labeling was seen in the proximal inner nuclear layer and retinal pigment epithelium. Specific A, binding sites were also reported for rat retina in this study. Recently, release of endogenous and radiolabeled purines by potassium depolarization has been demonstrated for rabbit retina (Perez et al., 1986).

Evidence is now provided that adenosine-mediated inhibition of adenylate cyclase via A, receptors is present in rabbit retina. Using homogenates of rabbit retina, in which endogenous adenosine has been removed by treatment with adenosine deaminase, inhibition of forskolin-stimulated adenylate cyclase has been elicited by nanomolar quantities of PIA and CHA. As the concentrations of both agonists are increased, enzyme activity also increases. The effect of NECA on forskolin-stimulated and basal adenylate cyclase activity was also investigated. Modu- lation of adenylate cyclase activity by GTP is also reported.

Uptake of 3H-adenosine and 3H-CHA by rabbit retina in vitro

was also demonstrated. Label was accumulated primarily in ganglion cell bodies and certain cell bodies located in the proximal portion of the inner nuclear layer, in agreement with the studies of Ehinger and Perez (1984) and Zarbin et al. (1986).

Materials and Methods Tissue preparation. Pigmented rabbits, 4-6 lb, obtained locally were decapitated and both eyes removed and placed in ice-cold oxygenated Earl& solution (120 AM NaCl, 5 mM KCl, 1 mM MgSO,, 25 mM NaHCO,. 0.8 mM Na,HPO,. 0.1 mM NaH,PO,. 10 mM glucose. 2 mM CaCl,). The globes were hemisected -1 mm posterior to the limbus and the anterior segment and vitreous discarded. The eyecups were hemisected across the retinal myelinated streak, and retinal pieces were gently removed by carefully brushing them from the sclera-pigment epithelium. All manipulations were done under the surface of the oxygenated cold solution. Isolated retinal pieces were placed in acid-washed tubes, frozen in liquid nitrogen, and stored at -70°C.

Adenylate cyclase assay. Enzyme activity was assayed in homogenates of rabbit retina as described previously (De Vries et al., 1982; Blazynski

160 1

1 I I

-6 -5 -4 LOG [GTP]

Figure 2. Effect of GTP on activity of adenylate cyclase. Homogenates of rabbit retina, pretreated with adenosine deaminase, were assayed for adenylate cyclase activity in the presence of varying concentrations of GTP (0). Adenylate cyclase activity was stimulated by including forskolin (5 PM) in the assay reagent. Data represent means of4 experiments run in triplicate. In parallel experiments, homogenates were assayed in the presence of 1 nM CHA (A) at the GTP concentration indicated. Inhibition of enzyme activity by CHA was observed only with GTP concentrations lower than 25 PM.

et al., 1985). In brief, frozen isolated retinas were homogenized in 2.5 cc of 50 mM Tris (pH 7.8). Homogenates were then incubated for 20 min at room temperature with 1 U/cc of adenosine deaminase. Aliquots (2 ~1) containing approximately 2 pg protein were added to tubes containing 50 ~1 of 50 mM Tris. The reaction was started by adding 50 ~1 of substrate reagent to each tube, yielding the following final concentrations: 60 mM phosphocreatine, 1 mg/cc creatine kinase, 0.05% BSA, 5 mM MgCl,, 1.25 mM ATP, 5 WM GTP, and 1 mM rolipram. Adenosine agonists and forskolin (5 FM) were also present as indicated. All samples were run in triplicate. The reaction proceeded at 25°C for 20 min and was quenched by the addition of 20 ~1 of 100 mM ethylenediaminetetra- acetic acid. The samples were boiled for several min and then centrifuged at 3000 rpm for 20 min. An aliquot from each tube was removed and acetylated according to the method ofHarper and Brooker (1975). Cyclic AMP was assayed by radioimmunoassay, and proteins were assayed using the Lowry method (Lowry et al., 1951).

Uptake autoradiography. Pigmented rabbits (4-5 lb) were anesthe- tized with xylazine (5 mg/kg, IM), followed in 5 min by ketamine (60 mg/kg). After approximately 20 min, the normal eye-blink reflex was absent. The eye to be injected was treated topically with a drop ofalcaine, a local anesthetic. Fifty microliters of either 3H-CHA or 3H-adenosine (approximately 50 PCi) were injected into the vitreal cavity with a narrow-gauge needle attached to a Hamilton syringe. The animal was allowed to revive and returned to its cage. Four hours after injection, the animal was sacrificed by air embolism and the injected eye removed.

The eye was hemisected, anterior segment and vitreous removed, and the eyecup placed overnight in 2.5% glutaraldehyde in 0.16 M cacodylate buffer (pH 7.4). The eyecup was stained with osmium, dehydrated in ethanol, and embedded in Araldite. Sections, 1 pm, were placed on clean slides, dipped in Kodak NTB-2 emulsion, and air-dried. The slides were stored in light-tight boxes in the cold (3°C). Autoradiograms were developed in Kodak D- 19 and lightly stained with toluidine blue.

Materials. Glutaraldehyde was purchased from TAAB Lab. (Reading,

2524 Blazynski * A, Receptor Inhibition of Adenylate Cyclase

IIO- IF z IOO- .E 2 0 a F

90- 2 z 80- a ” 2 70- 2 I I i= 0” -10 -9 -8 -7 -6 -5

LOG [PIA]

Y -,b -b -c -; -k -; LOG [CHAI

Figure 3. Response of forskolin-stimulated adenylate cyclase to various doses of the A, agonists PIA and CHA. Homogenates of rabbit retinas were treated with adenosine deaminase. Data represent means of 6 experiments f SEM.

England). Osmium was obtained from EM Sciences (Ft. Washington, PA). IBMX was purchased from Calbiochem-Behring (San Diego, CA); creatine kinase from Boehringer Mannheim (Indianapolis, IN); radio- nuclides from Amersham (Arlington, IL); NTB-2 emulsion and D-19 developer from Kodak (Rochester, NY). Rolipram was generously provided by Schering AG (Berlin). All other chemicals were obtained from Sigma (St. Louis) and were of the highest grade possible.

IBMX CHA CHA+ PIA PIA+ IBMX IBMX

Figure 4. Antagonism of A, inhibition of adenylate cyclase by IBMX. Compared to basal conditions, or in the presence of 10 PM IBMX, the activity of adenylate cyclase was significantly inhibited (**p < 0.0 1) by either 10 nM CHA or PIA. When either CHA or PIA was combined with IBMX, the inhibition of adenylate cyclase by the A, agonists was eliminated. Data represent means k SEM of 6 experiments. The activity of adenylate cyclase (in the presence of forskolin) in the presence of IBMX was not statistically different from basal levels; enzyme activity in the presence of both IBMX plus PIA and IBMX plus CHA were also not statistically different from basal levels.

5pM Forskolin

0 -10 -9 -8 -7 -6 -5 -4 -3

LOG [ NECA 1

Figure 5. Effect of the mixed A,-A, agonist NECA on the activity of adenylate cyclase. Homogenates of rabbit retina were exposed to increasing doses of NECA in the presence (A) or absence (0) of forskolin. Data represent means ? SEM of 6 experiments, except at 500 PM (2 experiments).

Results Assay conditions In order to assay for adenylate cyclase activity in tissue homogenates, it is necessary to inhibit phosphodiesterase. For most assays, the PDE inhibitor IBMX is preferred. However, since xanthines are potent antagonists at extracellular adenosine receptors, the use of IBMX was precluded. Rolipram has proved to be useful for the assay of adenosine-sensitive adenylate cyclase (Fredholm et al., 1983, 1986). When the basal activities of adenylate cyclase in rabbit retinal homogenates were compared using 1 mM rolipram and 1 mM IBMX, very similar activities were measured (45.4 vs 49.2 pmol cAMP/mg protein/ min, respectively). The use of rolipram in the assay was further justified by comparing the extent of cyclase stimulation by dopamine (De Vries et al., 1982). For both conditions, addition of 50 FM dopamine to the assay reagent resulted in a 30% increase in enzyme activity. As reported by De Vries et al. (1982), IBMX effectively inhibits the hydrolysis of CAMP during the reaction period, and the similarity of results for the 2 PDE inhibitors indicated that rolipram effectively inhibited the hydrolysis of CAMP.

Basal activity of adenylate cyclase was stimulated using forskolin (Fredholm et al., 1983, 1986; Li and Fredholm, 1985). Activity increased from 49.4 -t 4.2 to 106.3 + 3.3 pmol CAMP/ mg proteimmin in the presence of 5 PM forskolin. This concentration significantly increased CAMP in both homogenates and incubated retinas. In the absence of forskolin, inhibition of adenylate cyclase by adenosine analogs was not detected except at high (> 100 PM) concentrations. Under conditions of the enzyme assay, activity was linear over time (25°C) for a period of 25 min; reactions were thus quenched after 15-20 min.

Low concentrations of CHA (0.1-10 nM) inhibited the activity of forskolin-activated adenylate cyclase in retinal homogenates (Fig. 1). This effect was biphasic: At concentrations below 1 nM enzyme activity decreased, whereas higher concentrations resulted in a return to basal levels of activity. If the homogenates were pretreated for 20 min (25°C) with adenosine deaminase (1

The Journal of Neuroscience, August 1997, 7(E) 2525

Figure 6. Uptake of adenosine by rabbit retina. Autoradiograph of a rabbit retina from an eye injected intravitreally with ‘H-adenosine (50 NCi), followed by fixation at 4 hr postinjection. Sub- stantial labeling occurred in ganglion cell bodies (G) and cell bodies located in the middle of the inner nuclear layer (1). Autoradiograph was developed at 2 weeks.

CHA or PIA, the level of adenylate cyclase activity was not significantly different from either basal conditions or with IBMX alone. Thus, the action of PIA and CHA is specific for adenosine receptors.

EJtict of Al-A, agonists The effect of including the mixed A,-A, agonist NECA in the assay for adenylate cyclase is presented in Figure 5. In the absence of forskolin, there was no change from basal levels of activity over a wide range of concentrations. However, NECA elicited a biphasic effect on the activity of forskolin-stimulated adenylate cyclase. Inhibition was seen with nanomolar levels of NECA, while potent stimulation was demonstrated when the levels were increased above 1 MM. Maximal inhibition (35%) was elicited with 10 nM NECA. Enzyme activity continued to increase at higher concentrations; at 1 PM, enzyme activity was 118% of basal levels, at 100 PM the activity was 147% of basal level, and at 500 FM the activity was 156% of basal. IBMX (10 PM) blocked the effect of NECA (data not shown).

Uptake autoradiography The uptake of )H-adenosine and 3H-CHA into retinal cells is illustrated in Figures 6 and 7. These autoradiograms are rep- resentative of injections done in a total of 6 animals (3 with adenosine, 3 with CHA), and results for each agonist and each animal were quite similar. For all experiments, uptake was most evident close to the injection site. However, as the length of development of the autoradiograms increased, more extensive regions of uptake throughout the retina became apparent. For both agonists, uptake was evident in ganglion cell bodies and also in certain cell bodies located in the middle to upper third of the inner nuclear layer. Some of these cell bodies may possibly be Muller cells, but there was no evidence of labeling of Muller cell end feet at the vitreal surface. Moreover, when the autoradiograms were developed after periods of 3-4 months, there was no labeling of Muller cell processes, nor was there any evidence of labeling in the photoreceptor layers.

U/ml) to destroy endogenous adenosine, slightly greater inhibition by CHA was observed. This proved to be true for inhibition of the cyclase by PIA and NECA (data not shown). For all subsequent experiments, homogenates were pretreated with adenosine deaminase.

Both activation and inhibition of adenylate cyclase requires GTP (for review, see Jakobs et al., 1985). As demonstrated in Figure 2, GTP inhibits the activity of forskolin-stimulated adenylate cyclase in a dose-dependent manner. Inhibition of the enzyme by CHA was observed at GTP concentrations less than 10 MM. For all experiments, GTP was included in the reaction media at a concentration of 5 FM.

Effects of A, agonists The results of experiments designed to test the effects of the A, agonists PIA and CHA on the activity of adenylate cyclase are presented in Figure 3. For 6 experiments, both agonists elicited biphasic effects on the forskolin-stimulated activities of the enzyme. CHA appeared to be slightly more potent in inhibiting enzyme activity; maximal inhibition (25%) was demonstrated with 1 nM CHA. With PIA, the maximal inhibition was observed at a concentration of 10 nM. However, PIA inhibited enzyme activity to a slightly greater extent than did CHA. Half- maximal inhibition of enzyme activity by PIA was observed with approximately 0.5 nM, while an estimate of the CHA concentration that elicited half-maximal inhibition was less than 0.1 nM. For both agonists, adenylate cyclase activity was less inhibited as the concentrations were increased toward the micromolar range.

The inhibition of cyclase activity by the A, agonists was spe- cifically blocked by including 10 nM IBMX in the assay (Fig. 4). Both CHA and PIA (10 PM) significantly inhibited enzyme activity (p < 0.01). When IBMX was included in the assay media, the activity was not different from the activity measured in its absence (145.3 f 5.8 vs 139.9 f 4.4, respectively), indicating that IBMX did not augment inhibition of PDE during the reaction period. When IBMX was combined with either

2526 Blazynski l A, Receptor Inhibition of Adenylate Cyclase

Figure 7. Uptake of CHA by rabbit retina. Autoradiograph of a rabbit retina from an eye injected intravitreally with 3H-CHA (50 Ki), followed by fixation at 4 hr postinjection. The distribution of labeling was identical with that observed for adenosine injections. Autoradiograph was developed at 5 weeks.

Discussion The inhibition of adenylate cyclase by A, adenosine agonists demonstrates that receptors of this subtype are present in rabbit retina. Inhibition could be seen only when basal activity of cyclase was elevated with the aid of the diterpene forskolin, an approach first suggested by Seamon and Daly (198 1). Specificity of the action of PIA and CHA on rabbit adenylate cyclase was demonstrated by the ability of low concentration of IBMX to block the inhibition. A direct demonstration of A, receptor- mediated increases in CAMP levels was reported previously for incubated rabbit retinas (Blazynski et al., 1986). The marked stimulation of enzyme activity when homogenates were incubated with micromolar levels of NECA confirms this original observation.

The concentration dependence of the inhibitory process in retina is quite similar to that reported for various regions of mammalian brain (for review, see Phillis and Barraco, 1985). Half-maximal inhibition of adenylate cyclase in rat hippocampus (Wojcik and Neff, 1982) and mouse striatum and cortex (Ebersolt et al., 1983) were reported to be between 10-30 nM for PIA and CHA. A biphasic response was also reported by Ebersolt et al. (1983) for striatal homogenate but not for cortical homogenates. Stimulation of cyclase was observed at 10 and 100 FM PIA. Biphasic dose-response effects on the levels of CAMP were measured in slices of rat hippocampus incubated with various doses of PIA and CHA (Fredholm et al., 1983). CHA, in doses of between lo-* and 1O-6 M, inhibited CAMP accumulation. A significant inhibition of CAMP accumulation was seen with a lower concentration of PIA (10 nM) than of CHA (50 nM). In the absence of forskolin (1 PM), only stimu- latory effects were elicited.

Guanine nucleotides decrease agonist binding to many types of receptors, including those that mediate changes in the activity

of adenylate cyclase (Cooper, 1982; Goodman et al., 1982; Green, 1984; Snyder, 1986); cyclase activation or inhibition is mediated by guanine nucleotide binding proteins. Thus, GTP, or some nonhydrolyzable analog, is required to modulate adenylate cyclase activity but, by inhibiting agonist binding to receptor, may mask any response. It was reported that GTP, GDP, and Gpp(NH)p decrease binding of radiolabeled CHA in homogenates of bovine brain by 50% at a concentration of l-3 PM.

Similar effects have been reported for homogenates of other regions of brain (Yeung and Green1983; Green, 1984). Inhi- bition of agonist binding can be diminished by the addition of sodium or divalent cation, such as magnesium (Goodman et al., 1982; Green, 1984; Jakobs et al., 1985). The enzyme assay media in this study included 0.5 mM MgCl,. Preliminary experiments investigating the effect of GTP on the activity of forskolin-stimulated adenylate cyclase demonstrated an inhibitory effect by GTP (Fig. 2). If the concentration of guanine nucleotide was increased above 10 PM, further inhibition by CHA could not be detected. These results are consistent with GTP inhibiting or antagonizing binding of agonist to receptor, as reported for brain homogenates.

Both inhibition and stimulation of cyclase were observed for low and high doses of NECA, respectively. Previous studies had demonstrated that micromolar quantities of adenosine elicited potent increases in the accumulation of CAMP in incubated rabbit retinas and that the increase was blocked by IBMX (Bla- zynski et al., 1986). Stimulation of adenylate cyclase activity by NECA reconfirms the finding that A, receptors mediate this increase in CAMP levels. For isolated retinas, a significant increase in retinal levels of CAMP were elicited by the addition of 100 PM adenosine to the incubation medium. For homogenates, in the absence of forskolin treatment, a barely significant increase in basal activity was elicited by 100 PM NECA. On the other hand, for homogenates that had been exposed to forskolin,

The Journal of Neuroscience, August 1987, i’(8) 2527

stimulation of cyclase was detected with 1 PM NECA. These differences in the A, response between intact retinas and homogenates are not easily reconciled. The levels of GTP and ionic milieu of isolated, incubated retinas is not easily con- trolled, and this may account for some of these discrepancies among preparations.

Modulation of adenylate cyclase activity by PIA has also been reported for primary cultures of mouse astrocytes (Ebersolt et al., 1983). That all or any of the adenosine or adenosine agonist effects elicited in retina are localized to neurons or glial cells of Muller cannot be determined. However, the pattern of adenosine and CHA uptake in retina suggests that both neurons and glia readily accumulate nucleosides, i.e., adenosine, cyclohexyladenosine, methylphenylethyl-adenosine, inosine, and gua- nosine. That the pattern of accumulation is very similar for each nucleoside indicates that uptake occurs via a nonspecific mech- anism rather than by metabolism to some common compound that is then accumulated. Evidence that adenosine is accumulated by retinal cells and then further metabolized to ATP has recently been reported by Perez et al. (1986), indicating that injected adenosine is transported into the cells (Figs. 6, 7; Ehin- ger and Perez, 1984). I f it is assumed that uptake sites are col- ocalized with receptor sites, then the results of Figures 6 and 7 suggest that biochemical effects on adenylate cyclase occur in ganglion cells and perhaps Muller cells. The distribution ofaden- osine-like immunoreactivity in rat retinas revealed strong labeling of ganglion cells (Zarbin et al., 1986), and if rat and rabbit retina are similar, then both receptor and uptake sites are indeed present in this cell type. Moderate labeling was also seen in the inner nuclear layer, again consistent with the distribution of uptake sites. However, adenosine-like immunoreactivity was also reported for the photoreceptor outer segment layer; uptake was not detected in this layer in rabbit retina, indicating different localizations of the 2 sites. However, uptake of purines into outer segments has been reported with in vitro studies (Ehinger and Perez, 1984). Such uptake may merely reflect an increase in metabolism during experimental manipulation.

Both inhibition and stimulation of adenosine-sensitive adenylate cyclase have been determined for the rabbit retina, indicating that both A, and A, receptors are present in this species. Adenosine-mediated increases in retinal CAMP levels were observed only in rabbit retina; mouse, rat, and guinea pig retinas were unaffected by exogenously supplied adenosine (Blazynski et al., 1986), suggesting that the presence of an A, system may vary from species to species. Specific A, receptor binding has been reported for 2 mammalian retinas-rat and human (Zarbin et al., 1986)-and this report, in combination with A, agonist inhibition of adenylate cyclase and the recent demonstration of adenosine release in retina (Perez et al., 1986), indicates that an A, system most probably is present in many mammalian retinas.

References Bender, A. S., P. H. Wu, and J. W. Phillis (1981) The rapid uptake

and release of [)H]adenosine by rat cerebral cortical synaptosomes. J. Neurochem. 36: 65 l-666.

Blazynski, C., J. A. Ferrendelli, and A. I. Cohen (1985) Indoleamine- sensitive adenylate cyclase in rabbit retina: Characterization and distribution. J. Neurochem. 45: 440-447.

Blazynski, C., D. A. Kinscherf, K. M. Geary, and J. A. Ferrendelli (1986) Adenosine-mediated regulation of cyclic AMP levels in isolated incubated retinas. Brain Res. 366: 224-229.

Braas, K. M., A. C. Newby, V. S. Wilson, and S. H. Snyder (1986) Adenosine-containing neurons in the brain localized by immunocy- tochemistry. J. Neurosci. 6: 1952-l 96 1.

Cooper, D. M. F. (1982) Bimodal regulation of adenylate cyclase. FEBS Lett. 138: 157-163.

Cooper, D. M. F., C. Londos, and M. Rodbell (1980) Adenosine receptor-mediated inhibition of rat cerebral cortical adenylate cyclase bv a GTP-denendent nrocess. Mol. Pharmacol. 18: 598-60 1.

Cooper, D. M. F., S. M.-H. Young, E. Perez-Reyes, J. R. Owens, L. H. Fossom, and D. L. Gill (1985) Properties required of a functional N,, the GTP regulatory complex that mediates the inhibitory actions of neurotransmitters on adenylate cyclase. Adv. Cyclic Nucleotide Prot. Phos. Res. 9: 75-86.

De Vries, G. W., K. M. Campau, and J. A. Ferrendelli (1982) Ade- nylate cyclases in the vertebrate retina: Distribution and characteristics in rabbit and ground squirrel. J. Neurochem. 38: 759-765.

Ebersolt, C., J. Premont, A. Prochiantz, M. Perez, and J. Bockaert (1983) Inhibition of brain adenylate cyclase by A, adenosine receptors: Pharmacological characteristics and locations. Brain Res. 267: 123-129.

Ehinger, B., and M. T. R. Perez (1984) Autoradiography of nucleoside uptake into the retina. Neurochem. Int. 6: 369-38 1.

Fredholm, B. B., and P. Hedqvist (1980) Modulation of neurotransmission by purine nucleotides and nucleosides. Biochem. Pharmacol. 29: 1635-1643.

Fredholm, B. B., B. Jonzon, and K. Lindstrom (1983) Adenosine receptor mediated increases and decreases in cyclic AMP in hippocampal slices treated with forskolin. Acta Physiol. Stand. I 17: 46 l- 463.

Fredholm, B. B., J. Fastbom, and E. Lindgren (1986) Effects of N-ethylmaleimide and forskolin on glutamate release from rat hippocampal slices. Evidence that prejunctional adenosine receptors are linked to N-proteins, but not to adenylate cyclase. Acta Physiol. Stand. 127: 38 l-386.

Geiger, J. D. (1986) Localization of [3H]cyclohexyladenosine and [‘Hlnitrobenzylthioinosine binding sties in rat striatum and superior colliculus. Brain Res. 363: 404-408.

Geiger, J. D., and J. I. Nagy (1984) Heterogeneous distribution of adenosine transport sites labelled by [3H]nitrobenzylthioinosine in rat brain: An autoradiographic and membrane binding study. Brain Res. Bull. 13: 657-666.

Geiger, J. D., F. S. La Bella, and J. I. Nagy (1985) Characterization of nitrobenzylthioinosine binding to nucleoside transport sites selective for adenosine in rat brain. J. Neurosci. 5: 735-740.

Goodman, R. R., and S. H. Snyder (1982) Autoradiographic localization of adenosine ieceptors in rat brain using PHlcvclohexvladenosine. J. Neurosci. 2: 1230-1241.

Goodman, R. R., M. J. Cooper, M. Gavish, and S. H. Snyder (1982) Guanine nucleotide and cation regulation of the binding of [3H]cyclohexyladenosine and [)H]diethylphenylxanthine to adenosine A, receptors in brain membranes. Mol. Pharmacol. 21: 329-335.

Goodman, R. R., M. J. Kuhar, L. Hester, and S. H. Snyder (1983) Adenosine receptors: Autoradiographic evidence for their location on axon terminals of excitatory neurons. Science 220: 967-969.

Green, R. D. (1984) Reciprocal modulation of agonist and antagonist binding to inhibitory adenosine receptors by 5’-guanylimidodiphosphate and monovalent cations. J. Neurosci. 4: 2472-2476.

Harper, J. F., and G. Brooker (1975) Femtomole sensitive radioimmunoassay for cyclic AMP and cyclic GMP after 2’0 acetylation by acetic anhydride in aqueous solution. J. Cyclic Nucleotide Res. I: 207-2 18.

Jakobs, K. H., K. Aktories, M. Minuth, and G. Schultz (1985) Inhi- bition of adenylate cyclase. Adv. Cyclic Nucleotide Prot. Phos. Res. 19: 137-150.

Li, Y.-O., and B. B. Fredholm (1985) Adenosine analogues stimulate cyclic AMP formation in rabbit cerebral microvessels via adenosine A,-receptors. Acta Physiol. Stand. 124: 253-259.

Londos, C., D. M. F. Cooper, and J. Wolff (1980) Subclasses ofexternal adenosine receptors. Proc. Natl. Acad. Sci. USA 77: 2551-2554.

Lowrv. 0. H.. N. J. Rosebrouah. A. L. Farr. and R. J. Randall (1951) Protein measurement with Fhe Folin phenol reagent. J. Biol. Chem: 193: 265-275.

Marangos, P. J. (1984) Differentiating adenosine receptors and adenosine uptake sites in brain. J. Receptor Res. 4: 23 l-244.

Marani, E. (1977) The subcellular distribution of 5’-nucleotidase activity in mouse cerebellum. Exp. Neurol. 57: 1042-1048.

Nagy, J. I., L. A. La Bella, and M. Buss (1984) Immunohistochem- mistry of adenosine deaminase: Implications for adenosine neurotransmission. Science 224: 166-l 68.

2528 Blazynski * A, Receptor Inhibition of Adenylate Cyclase

Paes de Carvalho, R., and F. G. de Mello (1982) Adenosine-elicited accumulation of adenosine 3’,5’-cyclic monophosphate in the chick embryo retina. J. Neurochem. 38: 493-500.

Paes de Carvalho, R., and F. G. de Mello (1985) Expression of A, adenosine receptors modulating dopamine-dependent cyclic AMP accumulation in the chick embryo retina. J. Neurochem. 44: 845-85 1.

Perez, M. T. R., B. E. Ehinger, K. Lindstrom, and B. B. Fredholm (1986) Release of endogenous and radioactive purines from the rabbit retina. Brain Res. 398: 106-l 12.

Phillis, J. W., and R. A. Barraco (1985) Adenosine, adenylate cyclase, and transmitter release. Adv. Cyclic Nucleotide Prot. Phos. Res. 19: 243-257.

Phillis, J. W., and P. H. Wu (1981) The role of adenosine and its nucleotides in central synaptic transmission. Prog. Neurobiol. 16: 187-239.

Pull, I., and H. McIlwain (1972) Adenine derivatives as neurohumoral agents in the brain. The quantities liberated on excitation of super- fused cerebral tissues. Biochem. J. 130: 975-98 1.

Ribeiro, J. A, and A. M. Sebastiao (1986) Adenosine receptors and calcium: Basis for proposing a third (A,) adenosine receptor. Prog. Neurobiol. 26: 179-209.

Schubert, P., W. Komp, and G. W. Kreutzberg (1979) Correlation of 5’-nucleotidase activity and selective transneuronal transfer of adenosine in the hippocampus. Brain Res. 168: 419-424.

Seamon, K. B., and J. W. Daly (198 1) Forskolin: A unique diterpene activator of cyclic AMP-generating systems. J. Cyclic Nucleotide Res. 7: 20 l-224.

Shimizu, H., S. Tanaka, and T. Kodama (1972) Adenosine kinase of mammalian brain: Partial purification and its role for the uptake of adenosine. J. Neurochem. 19: 687-698.

Snyder, S. H. (1986) Brain receptors: The emergence of a new phar- macology. TINS 9: 455-459.

Trussell, L. O., and M. B. Jackson (1985) Adenosine-activated potassium conductance in cultured striatal neurons. Proc. Natl. Acad. Sci. USA 82: 4857-4861.

Wojcik, W. J., and N. H. Neff (1982) Characteristics of the adenosine inhibited adenylate cyclase (A, receptor) in rat and mouse cerebellum. Sot. Neurosci. Abstr. 8: 157.

Yeung, S. M. H., and R. D. Green (1983) Agonist and antagonist affinities for inhibitory adenosine receptors are reciprocally affected by 5’-guanylimidodiphosphate or N-ethylmaleimide. J. Biol. Chem. 258: 2334-2339.

Zarbin, M. A., K. Braas, A. Newby, W. R. Green, and S. H. Snyder (1986) Localization of adenosine and adenosine receptors in the retina. Invest. Ophthalmol. Vis. Sci. (Suppl.) 27: 185.

Documents

Adenosine A, Receptor-Mediated Inhibition of Adenylate ... · cellular adenosine receptors. There have been relatively few reports of adenosine effects on visual biochemistry or physiology