Gen. Pharmac. Vol. 23, No. 3, pp. 303-307, 1992 Printed in Great Britain. All rights reserved
0306-3623/92 $5.00 + 0.00 Copyright © 1992 Pergamon Press Ltd
MINIREVIEW DO ADENOSINE
A3
RECEPTORS EXIST?
IAN KENNEDY,l MARK GURDENt and PETERSTRONG2. The Departments of ~Peripheral and 2Gastrointestinal Pharmacology, Glaxo Group Research Ltd, Park Road, Ware, Herts SG12 0DP, England (Received 17 October 1991)
Abstract--1. It has been suggested that adenosine A~ receptors may be sub-divided into A~ and A 3 types, based on the relative potencies of 5'-N-ethylcarboxamidoadenosine (NECA) and selected N6-substituted adenosine analogues. At A~ receptors (rat adipocytes) N6-phenylisopropyladenosine(PIA) was reported to be approx. 100-foldmore potent than NECA, whereas the compounds were equipotent at A3receptors (those in cardiac and neuronal preparations). 2. The study reported here has systematically evaluated this proposal, the rank orders of potency of NECA, R- and S-PIA, N6-cyclopentyladenosine (CPA) and N6-cyclohexyladenosine(CHA) being determined in rat adipocytes, guinea-pig ileum and rat and guinea-pig atria. 3. R-PIA, CHA and CPA were found to have consistent potencies relative to NECA across all 6 tissues, including rat adipocytes. The rank order was CPA ~>CHA, R-PIA >/NECA > S-PIA. 4. We conclude that the relative potencies of these agonists do not support the concept that adenosine A~ receptors in rat adipocytes differ from those in neuronal and cardiac tissues.
INTRODUCTION Adenosine receptors were first classified into two types by Van Calker et aL (1979), based on the potency of adenosine and some of its derivatives as stimulators or inhibitors of cyclic AMP formation in cultured brain cells. The receptors mediating inhibition were termed Aj and those mediating stimulation A 2. Although other nomenclature has been used, the A~/A2 system for adenosine receptors is now generally accepted (Stone, 1985). A key study which contributed to the general acceptance of the two receptor classification was that of Londos et al. (1980). These workers, who termed AI and A2 receptors Ri and Ra respectively, examined the effects of adenosine and two synthetic derivatives, 5'-Nethylcarboxamidoadenosine (NECA) and N 6phenylisopropyladenosine (PIA), on the adenylate cyclase activity and other responses of a range of isolated cell types. The AI receptor-containing cell type used was the rat adipocyte. Londos et al. (1980) reported that PIA was approx. 100 times more potent than NECA as an inhibitor of adenylate cyclase and lipolysis in the adipocyte. In contrast, on cells containing A2 receptors (hepatocytes and Leydig cells), NECA was found to be more potent than PIA. The concept of two receptor sub-types was also supported by the work of Smellie et aL (1979), who reported that the R- and S-isomers of PIA could differentiate between A~ and A 2 receptors. At Al receptors R-PIA was reported to be 100-times more potent than S-PIA whereas at A2 receptors R-PIA was 10-times less potent than S-PIA. *To whom all correspondence should be addressed.
Subsequently, however, a number of workers have found that for many tissues containing receptors classified as A~, on the basis of the relative potencies of R- and S-PIA, the potency of R-PIA was little different from that of NECA. Examples include pre-synaptic receptors mediating inhibition of neurotransmitter release (Paton and Webster, 1984) and those on cardiac tissue mediating negative inotropic and chronotropic effects (Collis, 1983; Kurahasai and Paton, 1986). Indeed, in one such study Hughes and Stone (1983) concluded that the receptors mediating negative inotropic activity in rat atria could not be classifed as either A~ or A 2. Furthermore it is now clear that not all effects following activation of A~ receptors are a consequence of inhibition of adenylate cyclase (Cooper and Caldwell, 1990). Findings such as these led Ribeiro and Sebastifio (t986), in a review of the literature in this area, to propose the existence of a third type of adenosine receptor, termed A 3. This receptor was postulated to be present in excitable tissues, such as nerves and cardiac muscle, and, unlike the A I receptor, it was envisaged not to be coupled to adenylate cyclase. Activation of A3 receptors was suggested to decrease the entry of calcium ions into cells and/or to block the effects of intracellular calcium. The A~ and A3 receptors were defined pharmacologically in terms of the rank orders of agonist potency of: NECA, Rand S-PIA, N6-cyclohexyladenosine (CHA) and 2chloroadenosine (2-CA). At A~ receptors the rank order is R-PIA, C H A > 2 - C A > S-PIA, NECA whilst at m3 receptors R-PIA, CHA, NECA > 2-CA, with S-PIA usually less potent than 2-CA (Ribeiro and Sebasti~o, 1986). Since we are not aware of any studies which have systematically addressed this hypothesis we have compared the potencies of a
303
304
IAN KENNEDY et al.
range of adenosine analogues on rat adipocytes (A l receptors), guinea-pig ileum a n d left a n d right atria of b o t h guinea-pig a n d rat, tissues suggested to contain A 3 receptors. The c o m p o u n d s tested included N E C A , R- a n d S-PIA, C H A and a n o t h e r N6-substituted derivative o f adenosine, N6-cyclopentyladenosine (CPA). We chose not to study 2-CA, however, because of the evidence which suggests t h a t it m a y be subject to uptake (Brown a n d Collis, 1982; Jarvis et al., 1985) a n d for this reason is p r o b a b l y unsuitable for receptor classification (Kenakin, 1984). A preliminary account of some of this work has been published elsewhere ( G u r d e n et al., 1989). MATERIALS AND M E T H O D S
Tissues were obtained from male Dunkin Hartley guineapigs (250--300 g) and from male AH/A rats (Glaxo Group Research 200-300 g; 100-120 g for adipocyte experiments). NECA and cyclopentyladenosine were synthesised in the Medicinal Chemistry Department, Glaxo Group Research. R-PIA, S-PIA and CHA were obtained from Boehringer Mannheim (Diagnostics and Biochemicals) Ltd (Sussex, England). Bovine serum albumin (fraction V, microbiological grade) was obtained from Armour Pharmaceutical Co. (Eastbourne, England). All tension changes were recorded isometrically by means of either Dynamometer UFI or Statham UC3 force transducers. Ileum and cardiac preparations
A modified Krebs solution of the following composition was used: (raM) NaC1, 118; NaH2PO4, 25; KC1, 4.7; MgSO4, 0.6; KH2PO4, 1.2; D-glucose, 11.2; CaC12, 1.3. However, for experiments utilising guinea-pig ileum longitudinal muscle, the concentrations of MgSO 4 and CaC12 were increased to 1.2 and 2.6mM respectively. Solutions were gassed with 95% 02/5% CO 2 and maintained at 37°C (ileum) or 32°C (atria). Ileum. Strips of guinea-pig ileum longitudinal muscle, 15-20 mm long, with the myenteric plexus intact, were prepared from the proximal ileum by gently rubbing the tissue with a cotton bud. The preparations were then suspended under a resting tension of I g and equilibrated for 45 rain, before being field stimulated (platinum electrodes, square wave pulses at 0-2 Hz, 1 msec pulse width at supramaximal voltage). Under these conditions, preliminary studies showed that the twitch response was blocked by tetrodotoxin (3 × 10-TM), confirming the neuronal nature of the response. Preparations were stimulated for periods of 4min, at 10 min intervals, during which the tissues were washed. When constant twitch responses were obtained, the inhibitory effects of adenosine analogues were studied. Responses were expressed as percentage inhibitions of the original twitch height. Atria. Left atrial preparations were placed on punctate platinum electrodes (resting tension 1-1.5g) and paced continuously at a frequency of 1 Hz (rat) or 3 Hz (guineapig), using 1 msec pulse width at twice the threshold voltage. After equilibration (45 min), during which time constant responses to electrical stimulation were achieved, a cumulative concentration-effect curve to orciprenaline (1 × 10-8-3 × 10-SM) was constructed. The tissues were washed, and a sub-maximum concentration of orciprenaline added. Tissues were washed and re-exposed to the selected concentration of orciprenaline until constant responses were obtained. The inhibitory effect of adenosine derivatives on the total force of contraction was then studied in preparations stimulated with orciprenaline, and responses were expressed as percentage inhibitions of this force of contraction. Right atrial preparations were allowed to beat spontaneously (the contractile frequency being derived
electronically using a rate meter). When a stable rate had been achieved after equilibration (45rain), the negative chronotropic effects of adenosine derivatives were assessed, and all responses were expressed as percentage decreases in the original rate. Rat adipocytes
Rat adipocytes were obtained from epididymal fat pads using the methods described by Stratton et al. (1985). Adenosine deaminase (1 unit/ml) and bovine serum albumin (4%) were included in all incubations. Adenosine derivatives in a range of concentrations were incubated with adipocytes (at 37°C) for 15min before lipolysis was activated by addition of noradrenaline (10-TM). One hour later the experiment was terminated by addition of perchloric acid. Lipolytic activity was then assessed by measuring the glycerol concentration in neutralised supernatants. Effects were estimated by calculating the % of release stimulated by noradrenaline in the presence and absence of test compounds. Analysis o f results
The responses obtained in each preparation were measured and a log concentration-effect curve constructed. From this the ECs0, the concentration of agonist producing 50% of its own maximum response, was estimated for each compound. In the guinea-pig ileum and all 4 atrial preparations the effect of NECA was rapidly reversed by washing. Consequently ECs0s for compounds were determined from a second curve constructed some 30~i0 min after a NECA concentration-effect curve had been established. An equipotent molar concentration ratio (EMCR) was then calculated by dividing the ECs0 of the test agonist by the ECs0 of NECA obtained in the same tissue, or batch of cells. Geometric mean EMCRs obtained with the agonists on each preparation were then subjected to a one-way analysis of variance (ANOVA). Where statistical differences were detected the data were further analysed using the Duncan's multiple comparisons tests. Differences were taken to be statistically significant where P < 0.05. RESULTS The reference agonist, N E C A , produced a concentration-related inhibition of lipolysis in the rat adipocyte, inhibition of the neurogenic twitch response in the guinea-pig ileum and a negative inotropic a n d c h r o n o t r o p i c response in left a n d right atrial p r e p a r a t i o n s respectively. All of the o t h e r adenosine derivatives produced responses qualitatively similar to that of N E C A . M e a n E M C R values, obtained on the above p r e p a r a t i o n s for each of the agonists are s h o w n in Table 1. M e a n agonist concentration-effect curves are shown in Figs 1 and 2. The profile o f agonist activity obtained on each of the 6 p r e p a r a t i o n s studied was very similar. C P A was the most p o t e n t c o m p o u n d tested o n all preparations, being between 2- and 5-times more potent t h a n N E C A ( P < 0.05). In contrast, S - P I A was 18-36 times weaker t h a n N E C A , a n d was significantly weaker t h a n any of the o t h e r c o m p o u n d s studied ( P < 0.05). R - P I A was equipotent with N E C A on the guinea-pig ileum and rat right atrium. This c o m p o u n d was slightly more p o t e n t t h a n N E C A ( P < 0.05) on the rat adipocyte and rat left atrium, but nearly 2-times less p o t e n t t h a n N E C A ( P < 0.05) on the guinea-pig atrial preparations. C H A was equipotent with R - P I A on all of the p r e p a r a t i o n s tested.
D o a d e n o s i n e A 3 receptors exist?
305
Table 1. Mean equipotent molar concentration ratios for the adenosine derivatives evaluated EMCR (NECA = 1) Preparation
NECA EC~ (nM)
R-PIA
CPA
CHA
S-PIA
8.20 (5.5-12.2) n=13 52 (44-61) n=20 46 (28-76) n =22 30 (24-37) n =22 93 (72-119) n =20 148 (115-189) n=23
0.43 (0.20-0.92) n=5 1.15 (0.81-1.64) n=7 0.50 (0.31-0.82) n =6 1.74 (0.07-2.81) n =6 0.70 (0.31-1.57) n =6 1.66 (0.99-2.78) n=6
0.21 (0.13~).37) n=4 0.37 (0.3(~).50) n=8 0.24 (0.14-0.40) n =6 0.57 (0.52~).62) n =6 0.21 (0.17-0.26) n=6 0.45 (0.23~0,88) n=6
0.28
39 (16-90) n=5 56 (41-75) n=15 18 (I 1-28) n =7 32 (24-43) n =7 25 (12 52) n =7 49 (22-107) n=5
Rat adipocyte Guinea-pig ileum Rat left atrium Guinea-pig left atrium Rat right atrium Guinea-pig right atrium
n=2 1.33 (1.14-1.55) n=5 0.52 (0.36-0.76) n =6 1.21 (0.78-1.88) n =6 0.77 (0.44-1.36) n =6 2,45 (1.54.1) n=7
EMCR, equipment molar concentration ratio. Figures in parentheses are 95% confidence limits.
~ 100 " ~ 80
3 .~
60
~~1,~ ~
~1oo 8o
4o
.E
6o 4o
ZO
~o
Qo
H
1 ~ 10
10 9
10- 8
10 7
10- 6
I~ 5
~
0 i
~ 100 ~ 80
10-9
10~B
10-7
10-6
10 -5
10-9
10-8
10-9
10-8 10-? 10-6 4o-5 Concentration (M)
10-4
100
~
~
/
i
~ 8o L. 0
-~ 60
~- 60 .E ~ 4o o
.1~ 2o H
b
20
0
10- 9
t0.8
40- 7
10- 6
10- 5
I
t 0- 4
I
10-7
I
I0-6
I
10-5
!
10-4
lOO 100 ~
BO o
'- 60 .E
._.
81 40 Q
6O
~ 4o
o
~
8O
20
~ 2o
o
nC9
(M]
to-e 10-7 ao-s Concentration
10-5
a
lo.4
Fig. 1. M e a n c o n c e n t r a t i o n - e f f e c t curves for N E C A ( 0 ) , R - P I A (11), C P A ( A ) , C H A ((3) a n d S - P I A ( r l ) for i n h i b i t i o n of lipolysis in rat adipocytes, o f electrically-induced twitch in g u i n e a - p i g ileum a n d o f s p o n t a n e o u s rate in rat right atria. Results are s h o w n as the m e a n + S E M for a m i n i m u m of 4 observations, except for C H A in rat adipocytes where n = 2.
o'
10-4
Fig. 2. M e a n c o n c e n t r a t i o n - e f f e c t curves for N E C A ( 0 ) , R - P I A (mR), C P A ( A ) , C H A ((3) a n d S - P I A (r-I) for i n h i b i t i o n of contractile force in rat a n d guinea-pig left a t r i a and of s p o n t a n e o u s rate in guinea-pig right atria. Results are s h o w n as the m e a n + S E M for a m i n i m u m of 5 observations.
306
IAN KENNEDYet al. DISCUSSION
The agonist potency ratios obtained in this study on guinea-pig ileum and guinea-pig and rat atrial preparations are in good agreement with published data (Collis, 1983; Ribeiro and Sebasti~o, 1986). R-PIA and CHA were essentially equipotent with NECA, whilst CPA was, at most, up to 5-fold more potent than NECA. The ratio of potencies of R- and S-PIA ranged from 18 on guinea-pig left atrium to 43 on guinea-pig ileum, which is broadly consistent with that anticipated for Aj receptors (Smellie et al., 1979). Results obtained on rat adipocytes with NECA and R-PIA were consistent with those obtained on the ileum and atrial preparations. However, these data differ markedly from those of Londos et al. (1980), in that R-PIA was only twice as potent as NECA, rather than 100-times more potent. It is not clear whether Londos et al. (1980) used the R-isomer of PIA or the racemate, although it is difficult to see how this alone could account for such a profound difference. Similarly, there are no obvious methodological differences which might account for this discrepancy. Ukena et al. (1984) also found R-PIA to be only twice as potent as NECA in inhibiting lipolysis, and 5-times more potent than NECA in inhibiting adenylate cyclase, in rat adipocytes. Comparable data have also been reported by Fredholm and Lindgren (1984), R-PIA being 2-3-fold more potent than NECA as an inhibitor of both lipolysis and adenylate cyclase activity in rat adipocytes. Clearly the results of these two studies are in agreement with our findings, but differ from those of Londos et al. (1980). Ribeiro and Sebasti~.o (1986) drew a distinction between adenosine receptors coupled to adenylate cyclase, such as those in the adipocyte, and receptors in excitable tissues, for example cardiac preparations, which are probably not coupled to this enzyme. In fact there are few quantitative studies of the pharmacological characteristics of adenosine receptor-mediated inhibition of adenylate cyclase. Apart from those cited above in rat adipocytes, Cooper et al. (1980) found PIA (isomer not specified) and CHA to be 10-times more potent than NECA as inhibitors of adenylate cyclase in GH 3 cells. In fact, we are aware of only one study in which chick embryo retinal cells were used, other than that of Londos et al. (1980), in which PIA was found to be 100-times more potent than NECA as an inhibitor of adenylate cyclase (Paes de Carvalho and de Mello, 1985). In summary, the results of the present study, combined with the majority of published data, do not support the view that the agonist potencies of NECA relative to N6-substituted compounds such as PIA, CHA and CPA can be used to distinguish between "AI" and "A3" receptors (Ribeiro and Sebasti~.o, 1986). Rather we suggest that the A~ receptor should be defined by the rank order of potency CPA/> CHA, R-PIA i> NECA > S-PIA, with CPA being not more than 10-times more potent than NECA. We note that the earlier ligand binding studies of Bruns et al. (1986) demonstrated a rank order of affinity for these compounds at A~ receptors which was the same as the rank order of potency reported here. However, none of these results preclude the possibility that there are sub-types of the AI receptor.
Indeed, Fredholm and Dunwiddie (1989) recently suggested that AI receptors, by interacting with different G-proteins, may generate three different, intracellular signals: inhibition of adenylate cyclase, stimulation of K + efflux and inhibition of Ca ++ influx. The authors' favoured explanation of this involved the same A~ receptor interacting with different G-proteins. They did not, however, exclude the possibility that a different receptor sub-type might interact with each G-protein. SUMMARY
In 1986 Ribeiro and Sebastiao suggested that the adenosine receptors in rat isolated adipocytes were different from those in cardiac and neuronal preparations ("excitable tissues") on the basis of the agonist potencies of selected adenosine analogues. At adenosine A~ receptors, N6-substituted compounds, such as N6-phenylisopropyladenosine (PIA), were postulated to be approx. 100-fold more potent than 5'-N-ethylcarboxamidoadenosine (NECA), whilst at adenosine A 3 receptors N6-substituted compounds were of comparable potency to NECA. They suggested that the adipocyte receptors should be termed "A~", whilst the others were "A3". The purpose of this paper is to report a systematic evaluation of this proposal. The rank orders of agonist potency of NECA, Rand S-PIA, N6-cyclopentyladenosine (CPA) and N 6cyclohexyladenosine (CHA) have been determined for inhibition of lipolysis in rat isolated adipocytes, inhibition of neuronally-mediated twitch in guineapig ileum, inhibition of catecholamine-induced contraction in rat and guinea-pig left atria, and inhibition of spontaneous contractile rate in rat and guinea-pig right atria. The rank order of agonist potency for all of these preparations was CPA/>CHA, RPIA/> NECA > S-PIA. These data, therefore, do not support the hypothesis that this range of compounds can differentiate between the adenosine A~ receptors present in the range of tissues studied here. REFERENCES
Brown C. M. and Collis M. G. (1982) Evidence for an A2/Ra adenosine receptor in the guinea-pig trachea. Br. J. Pharmac. 76, 381 387. Bruns R. F., Lu G. H. and Pugsley T. A. (1986) Characterisation of the A2 adenosine receptor labelled by [3H]NECA in rat striatal membranes. Molec. Pharmac. 29, 331-346. Collis M. G. (1983) Evidence for an Ai-adenosine receptor in the guinea-pig atrium. Br. J. Pharmac. 78, 207-212. Cooper D. M. F. and Caldwell K. K. (1990) Signal transduction mechanisms for adenosine. In Adenosine and Adenosine Receptors (Edited by Williams M.), pp. 105-141. Humana Press, Clifton. Cooper D. M. F., Londos C. and Rodbell M. C. (1980) Adenosine receptor-mediated inhibition of rat cerebral adenylate cyclase by a GTP-dependent process. Molec. Pharmac. 18, 598~501. Fredholm B. B. and Dunwiddie T. V. (1989) How does adenosine inhibit transmitter release? Trends Pharmac. Sci. 9, 130-134. Fredholm B. B. and Lindgren E. (1984) The effects of alkylxanthines and other phosphodiesterase inhibitors on adenosine-receptor mediated decreases in lipolysis and cyclic AMP accumulation in rat fat cells. Acta pharmac. tox. 54, 64-71.
Do adenosine A 3 receptors exist? Gurden M. F., Strong P. and Kennedy I. (1989) Does the adenosine A 3 receptor exist? In Adenosine Receptors in the Nervous System (Edited by Ribeiro J. A.), p. 199. Taylor & Francis, London. Hughes P. R. and Stone T. W. (1983) Inhibition by purines of the inotropic action of isoprenaline in rat atria. Br. J. Pharmac. 80, 149-153. Jarvis S. M., Martin B. W. and Ng A. S. (1985) 2Chloroadenosine, a permeant for the nucleoside transporter. Biochem. Pharmac. 34, 3237-3241. Kenakin T. P. (1984) The classification of drugs and drug receptors in isolated tissues. Pharmac. Rev. 36, 165 222. Kurahasi K. and Paton D. M. (1986) Negative chronotropic action of adenosine in rat atria: evidence for action at A~ receptors. Nucleosides Nucleotides 5, 493 501. Londos C., Cooper D. M. F. and Wolff J. (1980) Subclasses of external adenosine receptors. Proc. natn Acad. Sci. U.S.A. 77, 2551-2554. Paes de Carvalho R. and de Mello F. G. (1985) Expression of A~ adenosine receptors modulating dopamine-dependent cyclic AMP accumulation in the chick embryo retina. J. Neurochem. 44, 845-851. Paton D. M. and Webster D. R. (1984) On the classification of adenosine and purinergic receptors in rat atria and in peripheral adrenergic and cholinergic nerves. In Neurones and Extraneuronal Events in Autonomic Pharmacology
307
(Edited by Flemming W. W., Langer S. Z., Graefe K-H. and Weiner N.), pp. 193-204. Raven Press, New York. Ribeiro J. A. and Sebastifio A. M. (1986) Adenosine receptors and calcium: basis for proposing a third (A3) adenosine receptor. Prog. Neurobiol 26, 179-209. Smellie F. W., Daly J. W., Dunwiddie T. V. and Hoffer B. J. (1979) The dextro and levorotatory isomers of Nphenylisopropyl adenosine: stereospecific effects on cyclic AMP-formation and evoked synaptic responses in brain slices. Life Sci. 25, 1739-1748. Stone T. W. (1985) Summary of a symposium discussion on purine receptor nomenclature. In Purines: Pharmacology and Physiological Roles (Edited by Stone T. W.), pp. 1-4 Macmillan, London. Stratton G. D., Myles D. D., Strong P. and Skidmore I. F. (1985) The development of tolerance to antilipolytic agents by isolated rat adipocytes. Biochem. Pharmac. 34, 275-279. Ukena D., B6hme E. and Schwabe V. (1984) Effects of several 5-carboxamide derivatives of adenosine on adenosine receptors of human platelets and fat cells. Naunyn-Schmiedebergs Arch. Pharmac. 327, 3~42. Van Calker D., Muller M. and Hamprecht B, (1979) Adenosine regulates via two different types of receptors the accumulation of cyclic AMP in cultured brain cells. J. Neurochem. 33, 999-1005.
0306-3623/92 $5.00 + 0.00 Copyright © 1992 Pergamon Press Ltd
MINIREVIEW DO ADENOSINE
A3
RECEPTORS EXIST?
IAN KENNEDY,l MARK GURDENt and PETERSTRONG2. The Departments of ~Peripheral and 2Gastrointestinal Pharmacology, Glaxo Group Research Ltd, Park Road, Ware, Herts SG12 0DP, England (Received 17 October 1991)
Abstract--1. It has been suggested that adenosine A~ receptors may be sub-divided into A~ and A 3 types, based on the relative potencies of 5'-N-ethylcarboxamidoadenosine (NECA) and selected N6-substituted adenosine analogues. At A~ receptors (rat adipocytes) N6-phenylisopropyladenosine(PIA) was reported to be approx. 100-foldmore potent than NECA, whereas the compounds were equipotent at A3receptors (those in cardiac and neuronal preparations). 2. The study reported here has systematically evaluated this proposal, the rank orders of potency of NECA, R- and S-PIA, N6-cyclopentyladenosine (CPA) and N6-cyclohexyladenosine(CHA) being determined in rat adipocytes, guinea-pig ileum and rat and guinea-pig atria. 3. R-PIA, CHA and CPA were found to have consistent potencies relative to NECA across all 6 tissues, including rat adipocytes. The rank order was CPA ~>CHA, R-PIA >/NECA > S-PIA. 4. We conclude that the relative potencies of these agonists do not support the concept that adenosine A~ receptors in rat adipocytes differ from those in neuronal and cardiac tissues.
INTRODUCTION Adenosine receptors were first classified into two types by Van Calker et aL (1979), based on the potency of adenosine and some of its derivatives as stimulators or inhibitors of cyclic AMP formation in cultured brain cells. The receptors mediating inhibition were termed Aj and those mediating stimulation A 2. Although other nomenclature has been used, the A~/A2 system for adenosine receptors is now generally accepted (Stone, 1985). A key study which contributed to the general acceptance of the two receptor classification was that of Londos et al. (1980). These workers, who termed AI and A2 receptors Ri and Ra respectively, examined the effects of adenosine and two synthetic derivatives, 5'-Nethylcarboxamidoadenosine (NECA) and N 6phenylisopropyladenosine (PIA), on the adenylate cyclase activity and other responses of a range of isolated cell types. The AI receptor-containing cell type used was the rat adipocyte. Londos et al. (1980) reported that PIA was approx. 100 times more potent than NECA as an inhibitor of adenylate cyclase and lipolysis in the adipocyte. In contrast, on cells containing A2 receptors (hepatocytes and Leydig cells), NECA was found to be more potent than PIA. The concept of two receptor sub-types was also supported by the work of Smellie et aL (1979), who reported that the R- and S-isomers of PIA could differentiate between A~ and A 2 receptors. At Al receptors R-PIA was reported to be 100-times more potent than S-PIA whereas at A2 receptors R-PIA was 10-times less potent than S-PIA. *To whom all correspondence should be addressed.
Subsequently, however, a number of workers have found that for many tissues containing receptors classified as A~, on the basis of the relative potencies of R- and S-PIA, the potency of R-PIA was little different from that of NECA. Examples include pre-synaptic receptors mediating inhibition of neurotransmitter release (Paton and Webster, 1984) and those on cardiac tissue mediating negative inotropic and chronotropic effects (Collis, 1983; Kurahasai and Paton, 1986). Indeed, in one such study Hughes and Stone (1983) concluded that the receptors mediating negative inotropic activity in rat atria could not be classifed as either A~ or A 2. Furthermore it is now clear that not all effects following activation of A~ receptors are a consequence of inhibition of adenylate cyclase (Cooper and Caldwell, 1990). Findings such as these led Ribeiro and Sebastifio (t986), in a review of the literature in this area, to propose the existence of a third type of adenosine receptor, termed A 3. This receptor was postulated to be present in excitable tissues, such as nerves and cardiac muscle, and, unlike the A I receptor, it was envisaged not to be coupled to adenylate cyclase. Activation of A3 receptors was suggested to decrease the entry of calcium ions into cells and/or to block the effects of intracellular calcium. The A~ and A3 receptors were defined pharmacologically in terms of the rank orders of agonist potency of: NECA, Rand S-PIA, N6-cyclohexyladenosine (CHA) and 2chloroadenosine (2-CA). At A~ receptors the rank order is R-PIA, C H A > 2 - C A > S-PIA, NECA whilst at m3 receptors R-PIA, CHA, NECA > 2-CA, with S-PIA usually less potent than 2-CA (Ribeiro and Sebasti~o, 1986). Since we are not aware of any studies which have systematically addressed this hypothesis we have compared the potencies of a
303
304
IAN KENNEDY et al.
range of adenosine analogues on rat adipocytes (A l receptors), guinea-pig ileum a n d left a n d right atria of b o t h guinea-pig a n d rat, tissues suggested to contain A 3 receptors. The c o m p o u n d s tested included N E C A , R- a n d S-PIA, C H A and a n o t h e r N6-substituted derivative o f adenosine, N6-cyclopentyladenosine (CPA). We chose not to study 2-CA, however, because of the evidence which suggests t h a t it m a y be subject to uptake (Brown a n d Collis, 1982; Jarvis et al., 1985) a n d for this reason is p r o b a b l y unsuitable for receptor classification (Kenakin, 1984). A preliminary account of some of this work has been published elsewhere ( G u r d e n et al., 1989). MATERIALS AND M E T H O D S
Tissues were obtained from male Dunkin Hartley guineapigs (250--300 g) and from male AH/A rats (Glaxo Group Research 200-300 g; 100-120 g for adipocyte experiments). NECA and cyclopentyladenosine were synthesised in the Medicinal Chemistry Department, Glaxo Group Research. R-PIA, S-PIA and CHA were obtained from Boehringer Mannheim (Diagnostics and Biochemicals) Ltd (Sussex, England). Bovine serum albumin (fraction V, microbiological grade) was obtained from Armour Pharmaceutical Co. (Eastbourne, England). All tension changes were recorded isometrically by means of either Dynamometer UFI or Statham UC3 force transducers. Ileum and cardiac preparations
A modified Krebs solution of the following composition was used: (raM) NaC1, 118; NaH2PO4, 25; KC1, 4.7; MgSO4, 0.6; KH2PO4, 1.2; D-glucose, 11.2; CaC12, 1.3. However, for experiments utilising guinea-pig ileum longitudinal muscle, the concentrations of MgSO 4 and CaC12 were increased to 1.2 and 2.6mM respectively. Solutions were gassed with 95% 02/5% CO 2 and maintained at 37°C (ileum) or 32°C (atria). Ileum. Strips of guinea-pig ileum longitudinal muscle, 15-20 mm long, with the myenteric plexus intact, were prepared from the proximal ileum by gently rubbing the tissue with a cotton bud. The preparations were then suspended under a resting tension of I g and equilibrated for 45 rain, before being field stimulated (platinum electrodes, square wave pulses at 0-2 Hz, 1 msec pulse width at supramaximal voltage). Under these conditions, preliminary studies showed that the twitch response was blocked by tetrodotoxin (3 × 10-TM), confirming the neuronal nature of the response. Preparations were stimulated for periods of 4min, at 10 min intervals, during which the tissues were washed. When constant twitch responses were obtained, the inhibitory effects of adenosine analogues were studied. Responses were expressed as percentage inhibitions of the original twitch height. Atria. Left atrial preparations were placed on punctate platinum electrodes (resting tension 1-1.5g) and paced continuously at a frequency of 1 Hz (rat) or 3 Hz (guineapig), using 1 msec pulse width at twice the threshold voltage. After equilibration (45 min), during which time constant responses to electrical stimulation were achieved, a cumulative concentration-effect curve to orciprenaline (1 × 10-8-3 × 10-SM) was constructed. The tissues were washed, and a sub-maximum concentration of orciprenaline added. Tissues were washed and re-exposed to the selected concentration of orciprenaline until constant responses were obtained. The inhibitory effect of adenosine derivatives on the total force of contraction was then studied in preparations stimulated with orciprenaline, and responses were expressed as percentage inhibitions of this force of contraction. Right atrial preparations were allowed to beat spontaneously (the contractile frequency being derived
electronically using a rate meter). When a stable rate had been achieved after equilibration (45rain), the negative chronotropic effects of adenosine derivatives were assessed, and all responses were expressed as percentage decreases in the original rate. Rat adipocytes
Rat adipocytes were obtained from epididymal fat pads using the methods described by Stratton et al. (1985). Adenosine deaminase (1 unit/ml) and bovine serum albumin (4%) were included in all incubations. Adenosine derivatives in a range of concentrations were incubated with adipocytes (at 37°C) for 15min before lipolysis was activated by addition of noradrenaline (10-TM). One hour later the experiment was terminated by addition of perchloric acid. Lipolytic activity was then assessed by measuring the glycerol concentration in neutralised supernatants. Effects were estimated by calculating the % of release stimulated by noradrenaline in the presence and absence of test compounds. Analysis o f results
The responses obtained in each preparation were measured and a log concentration-effect curve constructed. From this the ECs0, the concentration of agonist producing 50% of its own maximum response, was estimated for each compound. In the guinea-pig ileum and all 4 atrial preparations the effect of NECA was rapidly reversed by washing. Consequently ECs0s for compounds were determined from a second curve constructed some 30~i0 min after a NECA concentration-effect curve had been established. An equipotent molar concentration ratio (EMCR) was then calculated by dividing the ECs0 of the test agonist by the ECs0 of NECA obtained in the same tissue, or batch of cells. Geometric mean EMCRs obtained with the agonists on each preparation were then subjected to a one-way analysis of variance (ANOVA). Where statistical differences were detected the data were further analysed using the Duncan's multiple comparisons tests. Differences were taken to be statistically significant where P < 0.05. RESULTS The reference agonist, N E C A , produced a concentration-related inhibition of lipolysis in the rat adipocyte, inhibition of the neurogenic twitch response in the guinea-pig ileum and a negative inotropic a n d c h r o n o t r o p i c response in left a n d right atrial p r e p a r a t i o n s respectively. All of the o t h e r adenosine derivatives produced responses qualitatively similar to that of N E C A . M e a n E M C R values, obtained on the above p r e p a r a t i o n s for each of the agonists are s h o w n in Table 1. M e a n agonist concentration-effect curves are shown in Figs 1 and 2. The profile o f agonist activity obtained on each of the 6 p r e p a r a t i o n s studied was very similar. C P A was the most p o t e n t c o m p o u n d tested o n all preparations, being between 2- and 5-times more potent t h a n N E C A ( P < 0.05). In contrast, S - P I A was 18-36 times weaker t h a n N E C A , a n d was significantly weaker t h a n any of the o t h e r c o m p o u n d s studied ( P < 0.05). R - P I A was equipotent with N E C A on the guinea-pig ileum and rat right atrium. This c o m p o u n d was slightly more p o t e n t t h a n N E C A ( P < 0.05) on the rat adipocyte and rat left atrium, but nearly 2-times less p o t e n t t h a n N E C A ( P < 0.05) on the guinea-pig atrial preparations. C H A was equipotent with R - P I A on all of the p r e p a r a t i o n s tested.
D o a d e n o s i n e A 3 receptors exist?
305
Table 1. Mean equipotent molar concentration ratios for the adenosine derivatives evaluated EMCR (NECA = 1) Preparation
NECA EC~ (nM)
R-PIA
CPA
CHA
S-PIA
8.20 (5.5-12.2) n=13 52 (44-61) n=20 46 (28-76) n =22 30 (24-37) n =22 93 (72-119) n =20 148 (115-189) n=23
0.43 (0.20-0.92) n=5 1.15 (0.81-1.64) n=7 0.50 (0.31-0.82) n =6 1.74 (0.07-2.81) n =6 0.70 (0.31-1.57) n =6 1.66 (0.99-2.78) n=6
0.21 (0.13~).37) n=4 0.37 (0.3(~).50) n=8 0.24 (0.14-0.40) n =6 0.57 (0.52~).62) n =6 0.21 (0.17-0.26) n=6 0.45 (0.23~0,88) n=6
0.28
39 (16-90) n=5 56 (41-75) n=15 18 (I 1-28) n =7 32 (24-43) n =7 25 (12 52) n =7 49 (22-107) n=5
Rat adipocyte Guinea-pig ileum Rat left atrium Guinea-pig left atrium Rat right atrium Guinea-pig right atrium
n=2 1.33 (1.14-1.55) n=5 0.52 (0.36-0.76) n =6 1.21 (0.78-1.88) n =6 0.77 (0.44-1.36) n =6 2,45 (1.54.1) n=7
EMCR, equipment molar concentration ratio. Figures in parentheses are 95% confidence limits.
~ 100 " ~ 80
3 .~
60
~~1,~ ~
~1oo 8o
4o
.E
6o 4o
ZO
~o
Qo
H
1 ~ 10
10 9
10- 8
10 7
10- 6
I~ 5
~
0 i
~ 100 ~ 80
10-9
10~B
10-7
10-6
10 -5
10-9
10-8
10-9
10-8 10-? 10-6 4o-5 Concentration (M)
10-4
100
~
~
/
i
~ 8o L. 0
-~ 60
~- 60 .E ~ 4o o
.1~ 2o H
b
20
0
10- 9
t0.8
40- 7
10- 6
10- 5
I
t 0- 4
I
10-7
I
I0-6
I
10-5
!
10-4
lOO 100 ~
BO o
'- 60 .E
._.
81 40 Q
6O
~ 4o
o
~
8O
20
~ 2o
o
nC9
(M]
to-e 10-7 ao-s Concentration
10-5
a
lo.4
Fig. 1. M e a n c o n c e n t r a t i o n - e f f e c t curves for N E C A ( 0 ) , R - P I A (11), C P A ( A ) , C H A ((3) a n d S - P I A ( r l ) for i n h i b i t i o n of lipolysis in rat adipocytes, o f electrically-induced twitch in g u i n e a - p i g ileum a n d o f s p o n t a n e o u s rate in rat right atria. Results are s h o w n as the m e a n + S E M for a m i n i m u m of 4 observations, except for C H A in rat adipocytes where n = 2.
o'
10-4
Fig. 2. M e a n c o n c e n t r a t i o n - e f f e c t curves for N E C A ( 0 ) , R - P I A (mR), C P A ( A ) , C H A ((3) a n d S - P I A (r-I) for i n h i b i t i o n of contractile force in rat a n d guinea-pig left a t r i a and of s p o n t a n e o u s rate in guinea-pig right atria. Results are s h o w n as the m e a n + S E M for a m i n i m u m of 5 observations.
306
IAN KENNEDYet al. DISCUSSION
The agonist potency ratios obtained in this study on guinea-pig ileum and guinea-pig and rat atrial preparations are in good agreement with published data (Collis, 1983; Ribeiro and Sebasti~o, 1986). R-PIA and CHA were essentially equipotent with NECA, whilst CPA was, at most, up to 5-fold more potent than NECA. The ratio of potencies of R- and S-PIA ranged from 18 on guinea-pig left atrium to 43 on guinea-pig ileum, which is broadly consistent with that anticipated for Aj receptors (Smellie et al., 1979). Results obtained on rat adipocytes with NECA and R-PIA were consistent with those obtained on the ileum and atrial preparations. However, these data differ markedly from those of Londos et al. (1980), in that R-PIA was only twice as potent as NECA, rather than 100-times more potent. It is not clear whether Londos et al. (1980) used the R-isomer of PIA or the racemate, although it is difficult to see how this alone could account for such a profound difference. Similarly, there are no obvious methodological differences which might account for this discrepancy. Ukena et al. (1984) also found R-PIA to be only twice as potent as NECA in inhibiting lipolysis, and 5-times more potent than NECA in inhibiting adenylate cyclase, in rat adipocytes. Comparable data have also been reported by Fredholm and Lindgren (1984), R-PIA being 2-3-fold more potent than NECA as an inhibitor of both lipolysis and adenylate cyclase activity in rat adipocytes. Clearly the results of these two studies are in agreement with our findings, but differ from those of Londos et al. (1980). Ribeiro and Sebasti~.o (1986) drew a distinction between adenosine receptors coupled to adenylate cyclase, such as those in the adipocyte, and receptors in excitable tissues, for example cardiac preparations, which are probably not coupled to this enzyme. In fact there are few quantitative studies of the pharmacological characteristics of adenosine receptor-mediated inhibition of adenylate cyclase. Apart from those cited above in rat adipocytes, Cooper et al. (1980) found PIA (isomer not specified) and CHA to be 10-times more potent than NECA as inhibitors of adenylate cyclase in GH 3 cells. In fact, we are aware of only one study in which chick embryo retinal cells were used, other than that of Londos et al. (1980), in which PIA was found to be 100-times more potent than NECA as an inhibitor of adenylate cyclase (Paes de Carvalho and de Mello, 1985). In summary, the results of the present study, combined with the majority of published data, do not support the view that the agonist potencies of NECA relative to N6-substituted compounds such as PIA, CHA and CPA can be used to distinguish between "AI" and "A3" receptors (Ribeiro and Sebasti~.o, 1986). Rather we suggest that the A~ receptor should be defined by the rank order of potency CPA/> CHA, R-PIA i> NECA > S-PIA, with CPA being not more than 10-times more potent than NECA. We note that the earlier ligand binding studies of Bruns et al. (1986) demonstrated a rank order of affinity for these compounds at A~ receptors which was the same as the rank order of potency reported here. However, none of these results preclude the possibility that there are sub-types of the AI receptor.
Indeed, Fredholm and Dunwiddie (1989) recently suggested that AI receptors, by interacting with different G-proteins, may generate three different, intracellular signals: inhibition of adenylate cyclase, stimulation of K + efflux and inhibition of Ca ++ influx. The authors' favoured explanation of this involved the same A~ receptor interacting with different G-proteins. They did not, however, exclude the possibility that a different receptor sub-type might interact with each G-protein. SUMMARY
In 1986 Ribeiro and Sebastiao suggested that the adenosine receptors in rat isolated adipocytes were different from those in cardiac and neuronal preparations ("excitable tissues") on the basis of the agonist potencies of selected adenosine analogues. At adenosine A~ receptors, N6-substituted compounds, such as N6-phenylisopropyladenosine (PIA), were postulated to be approx. 100-fold more potent than 5'-N-ethylcarboxamidoadenosine (NECA), whilst at adenosine A 3 receptors N6-substituted compounds were of comparable potency to NECA. They suggested that the adipocyte receptors should be termed "A~", whilst the others were "A3". The purpose of this paper is to report a systematic evaluation of this proposal. The rank orders of agonist potency of NECA, Rand S-PIA, N6-cyclopentyladenosine (CPA) and N 6cyclohexyladenosine (CHA) have been determined for inhibition of lipolysis in rat isolated adipocytes, inhibition of neuronally-mediated twitch in guineapig ileum, inhibition of catecholamine-induced contraction in rat and guinea-pig left atria, and inhibition of spontaneous contractile rate in rat and guinea-pig right atria. The rank order of agonist potency for all of these preparations was CPA/>CHA, RPIA/> NECA > S-PIA. These data, therefore, do not support the hypothesis that this range of compounds can differentiate between the adenosine A~ receptors present in the range of tissues studied here. REFERENCES
Brown C. M. and Collis M. G. (1982) Evidence for an A2/Ra adenosine receptor in the guinea-pig trachea. Br. J. Pharmac. 76, 381 387. Bruns R. F., Lu G. H. and Pugsley T. A. (1986) Characterisation of the A2 adenosine receptor labelled by [3H]NECA in rat striatal membranes. Molec. Pharmac. 29, 331-346. Collis M. G. (1983) Evidence for an Ai-adenosine receptor in the guinea-pig atrium. Br. J. Pharmac. 78, 207-212. Cooper D. M. F. and Caldwell K. K. (1990) Signal transduction mechanisms for adenosine. In Adenosine and Adenosine Receptors (Edited by Williams M.), pp. 105-141. Humana Press, Clifton. Cooper D. M. F., Londos C. and Rodbell M. C. (1980) Adenosine receptor-mediated inhibition of rat cerebral adenylate cyclase by a GTP-dependent process. Molec. Pharmac. 18, 598~501. Fredholm B. B. and Dunwiddie T. V. (1989) How does adenosine inhibit transmitter release? Trends Pharmac. Sci. 9, 130-134. Fredholm B. B. and Lindgren E. (1984) The effects of alkylxanthines and other phosphodiesterase inhibitors on adenosine-receptor mediated decreases in lipolysis and cyclic AMP accumulation in rat fat cells. Acta pharmac. tox. 54, 64-71.
Do adenosine A 3 receptors exist? Gurden M. F., Strong P. and Kennedy I. (1989) Does the adenosine A 3 receptor exist? In Adenosine Receptors in the Nervous System (Edited by Ribeiro J. A.), p. 199. Taylor & Francis, London. Hughes P. R. and Stone T. W. (1983) Inhibition by purines of the inotropic action of isoprenaline in rat atria. Br. J. Pharmac. 80, 149-153. Jarvis S. M., Martin B. W. and Ng A. S. (1985) 2Chloroadenosine, a permeant for the nucleoside transporter. Biochem. Pharmac. 34, 3237-3241. Kenakin T. P. (1984) The classification of drugs and drug receptors in isolated tissues. Pharmac. Rev. 36, 165 222. Kurahasi K. and Paton D. M. (1986) Negative chronotropic action of adenosine in rat atria: evidence for action at A~ receptors. Nucleosides Nucleotides 5, 493 501. Londos C., Cooper D. M. F. and Wolff J. (1980) Subclasses of external adenosine receptors. Proc. natn Acad. Sci. U.S.A. 77, 2551-2554. Paes de Carvalho R. and de Mello F. G. (1985) Expression of A~ adenosine receptors modulating dopamine-dependent cyclic AMP accumulation in the chick embryo retina. J. Neurochem. 44, 845-851. Paton D. M. and Webster D. R. (1984) On the classification of adenosine and purinergic receptors in rat atria and in peripheral adrenergic and cholinergic nerves. In Neurones and Extraneuronal Events in Autonomic Pharmacology
307
(Edited by Flemming W. W., Langer S. Z., Graefe K-H. and Weiner N.), pp. 193-204. Raven Press, New York. Ribeiro J. A. and Sebastifio A. M. (1986) Adenosine receptors and calcium: basis for proposing a third (A3) adenosine receptor. Prog. Neurobiol 26, 179-209. Smellie F. W., Daly J. W., Dunwiddie T. V. and Hoffer B. J. (1979) The dextro and levorotatory isomers of Nphenylisopropyl adenosine: stereospecific effects on cyclic AMP-formation and evoked synaptic responses in brain slices. Life Sci. 25, 1739-1748. Stone T. W. (1985) Summary of a symposium discussion on purine receptor nomenclature. In Purines: Pharmacology and Physiological Roles (Edited by Stone T. W.), pp. 1-4 Macmillan, London. Stratton G. D., Myles D. D., Strong P. and Skidmore I. F. (1985) The development of tolerance to antilipolytic agents by isolated rat adipocytes. Biochem. Pharmac. 34, 275-279. Ukena D., B6hme E. and Schwabe V. (1984) Effects of several 5-carboxamide derivatives of adenosine on adenosine receptors of human platelets and fat cells. Naunyn-Schmiedebergs Arch. Pharmac. 327, 3~42. Van Calker D., Muller M. and Hamprecht B, (1979) Adenosine regulates via two different types of receptors the accumulation of cyclic AMP in cultured brain cells. J. Neurochem. 33, 999-1005.
