2-Substituted adenosine derivatives: affinity and efficacy at four subtypes of human adenosine receptors

Abstract

The affinity and efficacy at four subtypes (A₁, A_2A, A_2B and A₃) of human adenosine receptors (ARs) of a wide range of 2-substituted adenosine derivatives were evaluated using radioligand binding assays and a cyclic AMP functional assay in intact CHO cells stably expressing these receptors. Similar to previous studies of the N⁶-position, several 2-substituents were found to be critical structural determinants for the A₃AR activation. The following adenosine 2-ethers were moderately potent partial agonists (K_i, nM): benzyl (117), 3-chlorobenzyl (72), 2-(3-chlorophenyl)ethyl (41), and 2-(2-naphthyl)ethyl (130). The following adenosine 2-ethers were A₃AR antagonists: 2,2-diphenylethyl, 2-(2-norbornan)ethyl, R- and S-2-phenylbutyl, and 2-(2-chlorophenyl)ethyl. 2-(S-2-Phenylbutyloxy)a-denosine as an A₃AR antagonist right-shifted the concentration–response curve for the inhibition by NECA of cyclic AMP accumulation with a K_B value of 212 nM, which is similar to its binding affinity (K_i = 175 nM). These 2-substituted adenosine derivatives were generally less potent at the A₁AR in comparison to the A₃AR, but fully efficacious, with binding K_i values over 100 nM. The 2-phenylethyl moiety resulted in higher A₃AR affinity (K_i in nM) when linked to the 2-position of adenosine through an ether group (54), than when linked through an amine (310) or thioether (1960). 2-[2-(l-Naphthyl)ethyloxy]adenosine (K_i = 3.8 nM) was found to be the most potent and selective (>50-fold) A_2A agonist in this series. Mixed A_2A/A₃AR agonists have been identified. Interestingly, although most of these compounds were extremely weak at the A_2BAR, 2-[2-(2-naphthyl)ethyloxy]adenosine (EC₅₀ = 1.4 µM) and 2-[2-(2-thienyl)-ethyloxy]adenosine (EC₅₀ = 1.8 (M) were found to be relatively potent A_2B agonists, although less potent than NECA (EC₅₀ = 140 nM).

1. Introduction

Extracellular adenosine acts as a local modulator at four subtypes of receptors (A₁, A_2A, A_2B, and A₃), which are involved in numerous physiological and pathophysiological processes [1]. For example, adenosine attenuates the effects of ischemia in the heart and brain. Acting through the A_2A adenosine receptor (AR), it suppresses prolonged inflammation [2] and causes vasodilation and inhibits platelet aggregation, thus increasing the amount of oxygen available to an organ under stress. Adenosine agonists selective for the A₃AR are of interest as cerebroprotective [3], cardioprotective [4, 5], and anticancer [6] agents.

Recently, we have characterized structure–efficacy relationships for adenosine derivatives as agonists at the A₃AR. The intrinsic efficacy of adenosine derivatives in activation of the A₃AR is more variable than at other subtypes [7–10]. Specific groups placed at the N⁶-position and on the ribose moiety have reduced or completely abolished the ability to activate this receptor, while maintaining high binding affinity. Thus, it has been possible to design nucleoside-based antagonists [11], which in many cases are selective for both the human and rat A₃ARs. Such A₃AR antagonists include adenosine derivatives bearing: steric constraints of the ribose moiety or its 5′-amide modification [11], N⁶-groups such as 2,2-diphenylethyl and cyclopropyl [12], and the combination of substituted N⁶-benzyl groups and various small substituents at the adenine 2-position (such as chloro, cyano, and methoxycarbonyl) [11, 13]. A radically altered analogue in which the properly functionalized adenine moiety was shifted from the 1′-position to the 4′-carbon proved to be an A₃AR selective antagonist [9].

The aim of the present study is to expand knowledge of the structure activity relationships at the A₃AR and at other subtypes, both in relation to binding affinity and intrinsic efficacy, of adenosine derivatives modified in the 2-position. In this manner, it will be possible to design new, selective A₃AR agonists, partial agonists, and antagonists based on nucleoside structures. Derivatives of adenosine modified at the 2-position of the adenine ring have been studied in both binding and/or functional assays at the four AR subtypes. For this purpose, we have expressed the human ARs stably in Chinese hamster ovary (CHO) cells [14]. Most of these analogues are 2-ether substituted adenosine derivatives, which have been previously evaluated at the rat A₁ and A_2AARs but not at the four human subtypes in a systematic manner [15–20]. From previous studies [7, 13, 21] and in greatly expanded form in the present study, it is clear that the intrinsic efficacy of adenosine derivatives at the A₃AR is dependent on structural changes at both the N⁶-position and the 2-position. The intrinsic efficacy at the A_2AAR tended to be insensitive to the same structural changes. Additionally, here we have identified several substituents at the 2-position that contribute significantly to the A_2BAR activity.

2. Materials and methods

2.1. Materials

[¹²⁵I]N⁶-(4-Amino-3-iodobenzyl)adenosine-5′-N⁶-methyluronamide ([¹²⁵I]I-AB-MECA; 2000 Ci/mmol), [³H]R-PIA (R-N⁶-[phenylisopropyl]adenosine, 34 Ci/mmol), [³H]CG-S21680 (2-[p-(2-carboxyethyl)phenylethylamino]-5′-N⁶-ethylcarboxamido-adenosine, 47 Ci/mmol) and [³H]cyclic AMP (40 Ci/mmol) were from Amersham Pharmacia Biotech (Buckinghamshire, UK). NECA, CGS21680, CPA, and R-PIA were purchased from Sigma-RBI (St. Louis, MO). Most of the 2-substituted adenosine derivatives examined were the kind gift of Dr. Ray A. Olsson (University of South Florida) and Dr. JohnW.Daly (NIDDK). Compound 24 was prepared as described [37]. Other chemicals were from standard commercial sources and of analytical grade.

2.2. Chemical synthesis

General synthetic procedure for substitution at the 2-position of adenosine (for compounds 21 and 44).

Adenosine derivatives were synthesized using the general method of Ueeda et al. [15]. A solution of appropriate alcohol or thiol (0.35 mmol) in 5 ml of dry 1,2-dimethoxyethane was cooled to 0 °C in an ice bath. To this solution was added 2.5 M n-BuLi (0.13 ml, 0.32 mmol), the reaction mixture was stirred for 1 h at 0 °C. The protected nucleoside 2-chloro-2′,3′-O-(ethoxymethylidene) adenosine (25 mg, 0.07 mmol) was added in one portion. The reaction mixture was refluxed for 4 days, at which time HPLC showed that starting material had almost completely disappeared.

The solvent was removed in vacuo, and a solution of the residue in 10 ml water was extracted with ethyl acetate (4 × 20 ml). The combined extracts were dried over Na₂SO₄ and evaporated in vacuo. The residue was purified with preparative thin layer chromatography (PTLC, silica gel) with the mobile phase consisting of mixtures of methanol (3% for 21 or 5% for 44, by volume) and chloroform. Fractions containing products were concentrated and dissolved in methanol–water. Acetic acid or trifluoroacetic acid was added and the solution refluxed until HPLC showed that the nucleoside was completely deblocked. The solution was adjusted to pH 9 with NH₃–EtOH and was refluxed for 30 min. The solvent was removed, and the residue was purified with PTLC, with the mobile phase consisting of mixtures of methanol (20% for 21 or 10% for 44, by volume) and chloroform, to afford the corresponding product (overall yield 25–34% in two steps). Proton NMR and mass spectra were consistent with the assigned structures.

2.3. Cell culture and membrane preparation

The CHO cells stably expressing recombinant ARs were cultured in DMEM and F12 (1:1) supplemented with 10% fetal bovine serum, 100 units/ml penicillin, 100 µg/ml streptomycin, 2 µmol/ml glutamine and 800 µg/ml geneticin. After harvest and homogenization, cells were centrifuged at 500 × g for 10 min, and the pellet was re-suspended in 50 mM Tris–HCl buffer (pH 7.4) containing 10 mM MgCl₂, 1 mM EDTA. The suspension was homogenized with an electric homogenizer for 10 s, and was then re-centrifuged at 20,000 × g for 20 min at 4 °C. The resultant pellets were resuspended in buffer in the presence of 3 units/ml adenosine deaminase, and the suspension was stored at −80 °C until the binding experiments. The protein concentration was measured using the Bradford assay [22].

2.4. Binding assays

For A₃AR binding assays [23], each tube contained 100 ml of membrane suspension, 50 µl of [¹²⁵I]I-AB-MECA (final concentration 0.5 nM), and 50 µl of increasing concentrations of compounds in Tris–HCl buffer (50 mM, pH 7.4) containing 10 mM MgC₂. Nonspecific binding was determined using 10 (M NECA. The mixtures were incubated at 25 °C for 60 min. Binding reactions were terminated by filtration through Whatman GF/B filters under reduced pressure using a MT-24 cell harvester (Brandell, Gaithersburg, MD). Filters were washed three times with ice-cold buffer. Radioactivity was determined in a Beckman 5500B γ-counter. The binding of [³H]R-PIA to A₁ receptors and the binding of [³H]CGS21680 to A_2A receptors were as previously described [8].

2.5. Cyclic AMP accumulation assay

Intracellular cyclic AMP levels were measured with a competitive protein binding method [24]. CHO cells expressing four subtypes of recombinant ARs were harvested by trypsinization. After resupension in medium, cells were planted in 24-well plates in 0.5 ml medium. After 24 h, the medium was removed and cells were washed three times with 0.5 ml DMEM, containing 50 mM HEPES, pH 7.4. Cells were then treated with agonists and/or test compounds in the presence of rolipram (10 (M) and adenosine deaminase (3 units/ml). In the case of A₁ and A₃ARs, after 45 min forskolin (10 (M) was added to the medium, and incubation was continued an additional 15 min. The reaction was terminated by removing the medium, and cells were lysed upon the addition of 200 (L of 0.1 M ice-cold HCl. The cell lysate was resuspended and stored at −20 °C. For determination of cyclic AMP production, protein kinase A (PKA) was incubated with [³H]cyclic AMP (2 nM) in K₂HPO₄/EDTA buffer (K₂HPO₄, 150 mM; EDTA, 10 mM), 20 µL of the cell lysate, and 30 µL 0.1 M HCl or 50 µL of cyclic AMP solution (0–16 pmol/200 µL for standard curve). Bound radioactivity was separated by rapid filtration through Whatman GF/C filters and washed once with cold buffer. Bound radioactivity was measured by liquid scintillation spectrometry.

2.6. Statistical analysis

Binding and functional parameters were estimated using Prism software (GraphPAD). IC₅₀ values obtained from competition curves were converted to K_i values using the Cheng–Prusoff equation [25]. For an antagonist, Schild analysis was carried out as reported [26]. Data were expressed as mean ± standard error.

3. Results

3.1. Nucleoside structures examined

The classes of compounds examined included 2-alkyloxy ethers 1–18, 2-alkylaryl derivatives 19–46 including mainly ethers, however, an amino derivative 22 and a thioether 23 were included for comparison, two 5′,8-cyclo analogues 47 and 48, and four standard reference AR agonists 49–52. Although 47 and 48 were not substituted at the 2-position, they were included in this study, since conformational constraint in the ribose region has previously formed the basis of AR selectivity [27], and in these analogues the 2-position remains available for further derivatization.

Adenosine derivatives synthesized for this study were prepared by nucleophilic substitution by the appropriate alkoxy ion of the 2-chloro substituent in a ribose-protected derivative of 2-chloroadenosine, using methods similar to those reported previously [16].

3.2. Assays of binding and activation of ARs

Binding at the A₁, A_2A, and A₃ARs was carried out using standard agonist radioligands [8, 23] in membranes of transfected CHO cells [14]. The activation of A₁ and A₃ARs (G_i-coupled) by the 2-substituted adenosine derivatives (Table 1) was examined by measuring the inhibition of forskolin-stimulated cyclic AMP accumulation in intact CHO cells stably expressing these receptors. The activation of A_2A and A_2BARs (G_s-coupled) in stably transfected CHO cells was measured in the absence of forskolin. The efficacy of each of these adenosine derivatives was evaluated at a fixed concentration of 10 (M and expressed as a percentage of the effect of a reference (full) agonist. In some cases, full concentration–response curves were measured. For compounds that bound weakly at a given AR, no indication of intrinsic efficacy was readily obtainable, however for the many examples of high affinity binding the degree of activation at 10 (M served as an approximate measure of intrinsic efficacy.

Table 1.

Binding affinities of adenosine derivatives at human A₁, A_2A, and A₃ARs expressed in CHO cells (expressed as K_i value or percent displacement at 10 (M) and maximal agonist effects at 10 (M (% of full agonist) at the ARs

Compound	2-Substitution	Ki at A1AR (nMa) (% activation)	Ki at A2AAR (nMa) (% activation)	% activation at A2BARa (and % inhibition), unless noted	Ki at A3AR (nMa) (% activation)
2-Alkoxy analogues
1	CH3O	155 ± 32 (119)	970 ± 310 (78.6)	−2.7 ± 5.3 (−0.2)	156 ± 37 (75.2 ± 5.1)
2	CH3CH2O	2640 ± 540 (81.3)	360 ± 139 (92.4)	−0.8 ± 1.9 (−6.7)	568 ± 205 (99.1 ± 4.2)
3	(CH3)2CHO	16% (15.9)	927 ± 204 (87.5)	0.4 ± 2.9 (−3.2)	457 ± 154 (101 ± 5)
4	(CH3)2CHCH2O	4410 ± 1150 (44.6)	222 ± 89 (93.2)	−1.2 ± 8.8 (−3.4)	84.2 ± 8.4 (108 ± 10)
5	(CH3CH2)2CHCH2O	42% (1.0)	45.8 ± 22.1 (90.0)	−4.4 ± 4.2 (2.0)	336 ± 116 (4.9 ± 5.2)
6	Cyclohexyl-CH2O	3350 ± 390 (27.5)	342 ± 22 (92.8)	−0.9 ± 2.3 (−4.7)	143 ± 32 (27.1 ± 4.9)
7	(CH3)2CH(CH2)2O	3560 ± 1120 (29)	37.7 ± 4.9 (92.4)	−0.6 ± 1.8 (−8.4)	81.1 ± 9.0 (96.3 ± 3.0)
8	Cyclohexyl-(CH2)2O	36% (27.8)	579 ± 250 (102)	−4.6 ± 8.8 (0.4)	578 ± 182 (51.6 ± 3.4)
9		3590 ± 670 (0)	137 ± 31 (98.6)	−0.1 ± 2.6 (2.8)	149 ± 45 (−3.5 ± 9.0)
10		3350 ± 720 (25.9)	21.2 ± 12.7 (97.7)	−0.5 ± 4.4 (2.8)	341 ± 132 (−0.2 ± 5.9)
11	(CH3)2CH(CH2)3O	3700 ± 790 (18.3)	77.8 ± 20.9 (96.4)	−1.7 ± 2.8 (0.0)	105 ± 31 (49.6 ± 8.1)
12	Cyclohexyl-(CH2)3O	1730 ± 330 (33.5)	92.0 ± 52.5 (105)	−2.5 ± 4.6 (2.2)	83.3 ± 8.4 (21.2 ± 11.8)
13	Cyclohexyl-(CH2)4O	883 ± 99 (109)	291 ± 73 (98.5)	−0.27 ± 2.8 (−0.5)	105 ± 13 (12.7 ± 1.5)
14	CH3(CH2)4O	2430 ± 620 (48.7)	6.9 ± 1.3 (94.1)	−2.2 ± 2.6 (9.9)	222 ± 68 (92.8 ± 13.4)
15	CH3C≡C–(CH2)2O	583 ± 78 (49.8)	63.2 ± 50.3 (95.9)	1.4 ± 3.7 (−7.9)	90.2 ± 36.2 (73.4 ± 4.4)
16	CH3(CH2)5O	1830 ± 340 (11.8)	156 ± 89 (100)	12.6 ± 0.5 (−6.5)	124 ± 28 (43.7 ± 6.3)
17	(CH3)2C=CH(CH2)2CH(R-CH3)(CH2)2O	34% (23.6)	382 ± 100 (91.3)	0.3 ± 3.4 (−5.0)	431 ± 130 (−1.5 ± 1.5)
18	(CH3)2C=CH(CH2)2CH(S-CH3)(CH2)2O	4550 ± 950 (21.3)	78.3 ± 10.3 (90.2)	−0.4 ± 1.7 (−1.7)	74.4 ± 22.5 (31.5 ± 6.4)
2-Aryl- and arylalkyloxo (or amino)
19	Phenyl-O	5140 ± 1110 (57.6)	44 ± 12 (100)	0.7 ± 1.8 (−0.9)	364 ± 96 (32.2 ± 3.5)
20	Benzyl-O	642 ± 79 (11.5)	585 ± 155 (85.1)	−3.3 ± 0.6 (9.5)	117 ± 8 (16.9 ± 3.9)
21	3-Chlorobenzyl-O	27.4 ± 3.9 (46)	228 ± 66	7.4 ± 2.1	71.6 ± 24.6 (16.2 ± 3.8)
22	Phenyl-(CH2)2O	221 ± 57 (112)	9.3 ± 2.9 (99.6)	3490 ± 1490c (0.0)	54.2 ± 14.3 (70.7 ± 2.7)
23	Phenyl-(CH2)2NH	530 ± 88 (70.5)	62.0 ± 17.6 (105)	−1.7 ± 3.1 (−6.9)	310 ± 163 (72.0 ± 3.2)
24	Phenyl-(CH2)2S	3700 ± 770	590 ± 260	ND	1960 ± 310
25	2-Methylphenyl-(CH2)2O	396 ± 83 (74.6)	17.4 ± 7.4 (110)	−1.0 ± 2.7 (3.8)	214 ± 47 (7.3 ± 6.0)
26	3-Methylphenyl-(CH2)2O	295 ± 8 (106)	41.6 ± 22.0 (98.4)	15.3 ± 2.5 (0.6)	242 ± 55 (70.9 ± 3.9)
27	4-Methylphenyl-(CH2)2O	1250 ± 250 (61.8)	118 ± 95 (98.3)	12.1 ± 1.9 (−12.2)	470 ± 81 (95.5 ± 15.1)
28	2-Methyloxyphenyl-(CH2)2O	490 ± 114 (111)	274 ± 142 (94.2)	7.4 ± 1.3 (0.6)	940 ± 354 (3.6 ± 6.7)
29	3-Methyloxyphenyl-(CH2)2O	246 ± 34 (114)	32.1 ± 1.6 (103)	0.4 ± 4.1 (0.0)	231 ± 53 (85.1 ± 5.8)
30	4-Methyloxyphenyl-(CH2)2O	288 ± 22 (109)	64.3 ± 7.8 (97.6)	11.3 ± 1.8 (−4.4)	105 ± 20 (91.1 ± 9.2)
31	3,4-Dimethyloxyphenyl-(CH2)2O	469 ± 118 (101)	30.3 ± 2.8 (99.8)	−1.4 ± 2.1 (6.2)	863 ± 313 (66.5 ± 3.7)
32	2-Fluorophenyl-(CH2)2O	331 ± 22 (18.9)	58.1 ± 24.9 (99.1)	16.6 ± 2.5 (0.3)	77.8 ± 13.5 (44.5 ± 5.1)
33	4-Fluorophenyl-(CH2)2O	467 ± 100 (115)	56.8 ± 16.3 (97.9)	17.3 ± 3.1 (−7.2)	112 ± 16 (73.4 ± 5.2)
34	2-Chlorophenyl-(CH2)2O	366 ± 33 (31.8)	17.9 ± 6.1 (94.9)	1.0 ± 3.4 (6.9)	144 ± 22 (1.4 ± 2.7)
35	3-Chlorophenyl-(CH2)2O	372 ± 116 (85.3)	11.5 ± 5.3 (96.7)	28.0 ± 4.9 (2.9)	41.0 ± 7.8 (31.0 ± 7.0)
36	4-Chlorophenyl-(CH2)2O	331 ± 51 (113)	58.5 ± 8.0 (97.4)	17.5 ± 2.8 (5.7)	116 ± 23 (69.0 ± 6.3)
37	R-2-Phenylbutyl-O	28% (24.5)	503 ± 98 (97.7)	−0.2 ± 6.0 (4.6)	201 ± 61 (−1.6 ± 3.8)
38	S-2-Phenylbutyl-O	4780 ± 990 (16.0)	26.9 ± 6.9 (96.0)	0.4 ± 5.6 (−5.5)	175 ± 31 (−3.9 ± 3.9)
39	2-(1-Naphthyl)ethyl-O	220 ± 18 (101)	3.8 ± 1.4 (102)	−2.1 ± 5.1 (−3.7)	205 ± 19 (12.8 ± 5.9)
40	2-(2-Naphthyl)ethyl-O	141 ± 51 (102)	16.1 ± 7.0 (105)	1440 ± 70c	130 ± 8 (45.1 ± 8.5)
41	2,2-Diphenylethyl-O	39% (10.4)	310 ± 119 (97.5)	−1.8 ± 5.6 (0.8)	53.6 ± 10.4 (−0.0 ± 0.6)
42	2-(2-Thienyl)ethyl-O	174 ± 20 (112)	10.9 ± 4.8 (105)	1780 ± 260c	93.3 ± 16.8 (79.7 ± 4.5)
43	2-(3-Thienyl)ethyl-O	280 ± 72 (117)	13.3 ± 4.1 (106)	8.9 ± 5.5 (−2.0)	101 ± 34 (61.6 ± 15.1)
44	trans-2-Phenylcyclopropyl-O	367 ± 27 (19.1)	2050 ± 700	1.2 ± 2.3	292 ± 46 (0.4 ± 1.6)
45	4-Phenylbutyl-O	1100 ± 230 (24.6)	243 ± 166 (103)	−0.4 ± 2.3 (−7.8)	251 ± 80 (19.3 ± 6.2)
46	5-Phenylpentyl-O	700 ± 126 (104)	249 ± 54 (101)	6.2 ± 3.6 (5.4)	429 ± 159 (9.8 ± 8.8)
5′,8-Cyclo analogues	Name
47	Cyclo-CPAb	418 ± 62 (117)	12% (11.7)	−0.6 (−2.9)	25% (11.7 ± 0.3)
48	Cyclo-R-PIAb	49% (26.3)	18% (27.6)	4.5 (−9.6)	27% (1.7 ± 1.7)
Reference compounds	Name
49	CPA	1.8 ± 0.5 (112)	820 ± 216 (98)	8.6 ± 2.9 (0.0)	72 ± 12 (99 ± 6)
50	R-PIA	2.0 ± 0.3 (101)	884 ± 188 (101)	1680 ± 498c	8.7 ± 0.9 (102 ± 6)
51	CGS21680	1570 ± 460 (99)	8.8 ± 1.6 (100)	4.7 ± 1.8 (0.0)	114 ± 16 (98 ± 5)
52	NECA	6.8 ± 2.4 (100)	2.2 ± 0.6 (100)	140 ± 19c (0.0)	16.0 ± 5.4 (100)

At the A_2BAR, EC₅₀ values or the percent stimulation at 1 (M (and in parentheses percent inhibition at 1 (M of the effects of 100 nM NECA) are shown.

^aAll experiments were performed using adherent CHO cells stably transfected with cDNA encoding a human adenosine receptor. Percent activation of the human A₁ A_2A and A₃AR was determined at 10 µM. Binding at A₁, A_2A and A₃ARs was carried out as described in Section 2. The A₁ and A_2AAR activation results were expressed as the mean values from two separate experiments, while the A₃AR activation results were from three separate experiments. The K_i and EC₅₀ values from the present study are means ± S.E.M., N = 3–5.

^bStructures shown in Fig. 4; ND: not determined.

^cEC₅₀ (nM) for activation of the A_2BAR.

3.3. Affinity and potency at the A₃AR

Various 2-substituents were found to be critical structural determinants for activation of the A₃AR. Less than maximal efficacy or lack of efficacy was frequently observed for the 2-modified adenosine derivatives at the A₃AR. The following adenosine 2-ethers were moderately potent partial agonists (K_i): benzyl 20 (117 nM), 2-(3-chlorophenyl)ethyl 35 (41 nM), and 2-(2-naphthyl)ethyl 40 (130 nM). Fig. 1 shows the concentration–response curves for inhibition of cyclic AMP accumulation by a variety of 2-position ethers, indicating the variation in efficacy.

Fig. 1.

Inhibition of forskolin-stimulated cyclic AMP production in CHO cells stably transfected with the human A₃AR, induced by various agonists. All experiments were performed in the presence of 10 µM rolipram and 3 units/ml adenosine deaminase. Forskolin (10 (M) was used to stimulate cyclic AMP levels. The data shown were from one experiment performed in duplicate and are typical of three independent experiments giving similar results. EC₅₀ values were (nM): NECA, 20.0 ± 4.5; 2-isobutyloxy derivative 4, 150 ± 32; 2-S-(3,7-dimethyl)oct-6-enyloxy derivative 18, 187 ± 66; 2-benzyloxy derivative 20, 160 ± 40; 2-S-(2-phenyl)butyloxy derivative 38, not applicable.

The following adenosine 2-ethers were A₃AR antagonists: 2,2-diphenylethyl 41, R-2-phenylbutyl 37, S-2-phenylbutyl 38, 2-(2-chlorophenyl)ethyl 34 and 2-nor-zbornylethyl 9. 2-(S-2-Phenylbutyloxy)adenosine 38 right-shifted the concentration–response curve for the inhibition by NECA of cyclic AMP accumulation with a K_B value of 212 nM (Fig. 2) calculated by Schild analysis [26], which is similar to its A₃AR affinity (K_i = 175 nM) determined in the binding assay.

Fig. 2.

Antagonism by compound 38 of the inhibition of cyclic AMP production elicited by NEC in CHO cells stably transfected with the human A₃AR (A) and Schild analysis of the data (B). The experiment was performed in the presence of 10 µM rolipram and 3 units/ml adenosine deaminase. Forskolin (10 (M) was used to stimulate cyclic AMP levels. The level of cAMP corresponding to 100% was 220 ± 30 pmol ml⁻¹. The K_B value for compound 38 was calculated to be 212 nM.

Among small 2-alkyloxy groups, efficacy at the A₃AR remained nearly full with increasing size until the branched hexyl derivative 5. Branching of the O-hexyl group more distally in 11 did not alter affinity, but increased efficacy. The straight-chain pentyl 14 and hexyl 16 ether derivatives were full and partial agonists, respectively. In the series of O-alkylcyclohexyl derivatives of varying chain length 6, 8, 12, and 13, all were partial agonists at the A₃AR with similar affinity (100–600 nM), but with efficacy diminishing as chain length increased beyond two methylenes. A diastereomeric pair of branched decenyl ethers 17 and 18 differed greatly in both affinity (see below) and efficacy at the A₃AR.

The series of phenyl 19 and phenylalkyl (20, 22, 45, and 46) ethers showed high A₃AR efficacy only at intermediate length, i.e. the 2-phenylethyl ether 22, which also demonstrated the highest A₃AR affinity among the five analogues with a K_i value of 54 nM. Nevertheless, 22 was not selective for the A₃AR. The benzyl ether 20 was modestly potent as a partial agonist at the A₃AR and slightly selective (5–6-fold) in comparison to A₁ and A_2AARs. The 3-chlorobenzyl ether 21 was approximately equipotent to 20 at the A₃AR with a K_i value of 71.6 nM, but 23-fold more potent in A₁AR binding. The 2-phenylethyl amino derivative 23 was less potent than corresponding O-ether 22 at all AR subtypes, with K_i values at the A₃AR of 310 and 54 nM, respectively. The corresponding 2-phenylethylthio derivative 24 was less potent at three AR subtypes. Thus, the 2-O-ether linkage was selected for further modification.

Numerous modifications were included on the 2-phenylethyl moiety of ether 22, including phenyl substitution with methyl 25–27, methyloxy 28–31, and halo 32–36. In a number of these cases (25, 28, and 34), the 2-substitution of the phenyl ring resulted in the greatest reduction of A₃AR efficacy. Methoxy substitution at both 3- and 4-positions of the phenyl ring in 31 reduced A₃AR affinity in comparison of substitution of either position alone (i.e. 29, 30). Substitution of the phenylethyl ring with groups having electron withdrawing character (i.e. F) versus donating (i.e. methoxy or methyl) had little effect on A₃AR affinity, however the efficacy of 3- and 4-substituted halo derivatives tended to be less than the corresponding methoxy or methyl derivatives. Also, the phenyl ring was substituted with other ring systems, as in 39, 40, and 42–44, all of which had reduced efficacy at the A₃AR. Substitution at the 2-position of the alkyl (ethyl) group of 22 with ethyl (the diastereomers 38 and 38) or with phenyl 41 completely abolished efficacy at the A₃AR.

3.4. Affinity and potency at the A₁AR

These 2-substituted adenosine derivatives were generally less potent at the A₁AR than at the A₃AR with binding K_i values over 100 nM. Among the most potent derivatives at the (K_i < 200 nM) were 1, 40, 42 and 44. Among the two 5′,8-cyclo derivatives 47 and 48, only moderate affinity was observed for the N⁶-cyclopentyl analogue 47 at the A₁AR. Compound 47 was a somewhat selective A₁AR agonist.

Those 2-modified compounds with high to moderate affinity at the A₁AR tended to be fully efficacious at that subtype. However, the alkynyl ether 15, 2-phenylethyl amine 23, 2-methoxyphenylethyl ether 25, 2-fluorophenylethyl ether 32, and 2-chlorophenylethyl ether 34 were not fully efficacious at the A₁AR.

3.5. Affinity and potency at the A_2AAR

Consistent with previous studies [15, 16], the 2-substituted ether derivatives of adenosine were generally fully efficacious at the A_2AAR. A number of substitutions at the 2-position, which were previously found to contribute to the affinity for the rat A_2AAR [20, 28, 29], were also demonstrated to be important for the affinity and selectivity at the human A_2AAR homologue. A single substitution at the 2-position might contribute significantly to both A_2AAR affinity and selectivity. For example, 2-[2-(l-naphthyl) ethyloxy]adenosine 39 was found to be the most potent (K_i = 3.8 nM) and selective (>50-fold) A_2A agonist in this series. Thus, 39 was roughly as potent as an A_2AAR agonist as the nonselective agonist NECA 52. The 2-naphthyl isomer 40 was also potent at the A_2AAR.

3.6. Potency at the A_2BAR

All of nucleosides were tested at a fixed concentration for the ability to stimulate or inhibit adenylate cyclase mediated by the A_2BAR, and activation curves were determined for only a few compounds (Fig. 3). Since an agonist radioligand for the A_2BAR is not yet available, the binding affinities of these nucleosides at this subtype have not been determined. Interestingly, although most of these compounds were extremely weak at the A_2BAR, 2-[2-(2-naphthyl)ethyloxy]adenosine 40 (EC₅₀ = 1.44 (M) and 2-[2-(2-thienyl)ethyloxy]adenosine 42 (EC₅₀ = 1.78 (M) were found to be moderately potent A_2BAR agonists, although less potent than NECA (EC₅₀ = 140 nM). Another 2-ether having substantial potency at the A_2BAR was the 2-phenylethyl ether 22, which displayed an EC₅₀ value of 3.49 (M. At the A_2AAR 39 was a more potent agonist than 40, while the opposite order was seen at the A_2BAR.

Fig. 3.

Stimulation of cyclic AMP production elicited by NECA and compounds 39 and 40 in CHO cells stably transfected with the human A_2BAR. The experiment was performed in the presence of 10 (M rolipram and 3 units/ml adenosine deaminase. The data shown were from one experiment performed in duplicate and are typical of three independent experiments giving similar results. The EC₅₀ values were listed in Table 1.

4. Discussion

Although in the present study we have not identified highly selective ligands for a given AR, we have found that substitution at the 2-position greatly modulates the pharmacological characteristics at the A₃AR. Thus, the aim of the study has been partially achieved, and in future studies, specific 2-ether groups identified here may be studied in combination with other modifications of adenosine known to provide subtype selectivity. Until present, the modifications of the 2-position to be included in A₃AR selective agonists have been somewhat limited (e.g. chloro, methylthio, and alkynyl groups). The SAR of alkynyl groups at the 2-position based on affinity at several subtypes of ARs has been explored [21, 29].

Several of the nucleosides studied here were found to have K_i values at the A₃AR of approximately 50 nM and compared favorably in selectivity to other analogues of adenosine in which only single sites have been modified [11–13, 21]. For example, compound 41 was >100-fold selective for the A₃AR in comparison to both the A₁ and A_2BARs. Alkyl ethers 4 and 7 were moderately selective full agonists of the A₃AR in comparison to the A₁AR. Compounds 13 and 20 were slightly selective for the A₃AR in comparison to the A₁AR and other subtypes.

Similar to previous studies of the N⁶-position [8] and the ribose moiety [9, 30, 31], structural determinants at the 2-position of adenosine have been found to be critical for A₃AR recognition and activation. Fig. 4 shows the structures of selected analogues having A₃AR selectivity in comparison to at least two other AR subtypes and/or reduced efficacy at the A₃AR and the 5′,8-cyclo analogues (47 and 48). A striking dependence of the A₃AR efficacy on relatively minor structural changes of the 2-ether group, for example, substitution of 2-O-phenylalkyl ethers, is illustrated. The benzyl ether 20 is a low efficacy partial agonist. Homologation to give 22 restored most of the ability to activate the A₃AR, however simple substitution of this ring in 34 and 35 or additional aryl substitution in 37, 39, and 41 greatly reduced the intrinsic efficacy. The alkyl ether 7 is structurally related to 18, however, the efficacy at the A₃AR was greatly reduced upon chain elaboration.

Fig. 4.

Representative adenosine derivatives examined in this study. Compounds having A₃AR selectivity in comparison to at least two other AR subtypes (4, 7, 35, 37, and 41) and/or reduced efficacy (2-O-benzyl ether 20 and others) at the A₃AR are included. A striking dependence of the efficacy on the position of substitution of 2-O-phenylethyl ethers (22 in comparison to 34, 35, 37, 37, and 41) is illustrated. The full agonism of alkyl ethers (4 and 7) is in contrast to the branched decenyl ether 18, which is a partial agonist. 5′,8-Cyclo analogues (47 and 48) of known adenosine agonists CPA and R-PIA are also shown. The percent of maximal activation of the human A₃AR at 10 (M is indicated, where applicable.

In binding experiments, typically these 2-substituted nucleosides displaced AR radioligand binding with the following order of potency: A_2A > A₃ > A₁. The benzyl ether 20 displayed selectivity (5–6-fold) for the A₃AR. Mixed A_2A/A₃AR agonists have been identified. For example, 7 was a mixed agonist with full A₃AR intrinsic efficacy. Compounds 11, 12, and 18 were mixed agonists, but with partial intrinsic efficacy at the human A₃AR. Finally, the 2-norbornanethyl ether 9, the branched decenyl ether 17, and the 2-phenylbutyl ethers 37 and 38 were full agonists at the A_2AAR and antagonists at the human A₃AR, with affinities roughly matched at the two subtypes and having selectivity over the A₁ and A_2B ARs. Nucleosides with combined action in activating the A_2AAR and antagonizing the A₃AR may be of use in treating inflammation. Several arylethyloxy substituents at the 2-position were found to contribute significantly to the A_2BAR activity, for which there are not yet any selective agonists although the SAR has been explored [32, 33].

In the absence of a physically-determined structure for the A₃AR, the use of molecular modeling [12, 13, 34] will be necessary to understand the structural basis for the loss of efficacy associated with certain 2-ether groups, such as benzyl, 2,2-diphenylethyl, and S-2-phenylbutyl. There were parallels between the effects on A₃AR binding and activation of the same substitution at either the 2-position (in the form of an ether) or at the N⁶-position (in the form of a secondary arylamine). For example, 2-(2,2-diphenylethyloxy)-adenosine and N⁶-(2,2-diphenylethyl)adenosine were both antagonists of the A₃AR [12].

Also, in the series of 2-phenylalkylethers, the benzyl ether had more favorable affinity particularly at the A₃AR in comparison to other subtypes and was a partial agonist. The same was observed for N⁶-benzyladenosine derivatives [11]. This raises the possible explanation, already invoked by Olsson and coworkers [16] prior to any knowledge of the receptor binding site, that these two substitutions might be overlayed in their receptor-bound positions [13, 34]. However, the trans-2-phenylcyclopropyl group at the N⁶-position resulted in high A₃AR affinity as an agonist, while at the 2-ether position (44), only moderate affinity was observed with no activation of the A₃AR.

Thus, we have demonstrated that affinity at the A₃AR may be enhanced by modifying the nucleoside in a systematic fashion at a position that was not previously explored for ether groups. The 2-ether modification of adenosine, previously known as a means of enhancing potency at the A_2AAR, has now been shown to variably enhance potency at the A₃AR. Typically, adenosine 2-benzyl and 2-phenylethyl ethers showed favorable binding affinity at the A₃AR and, depending on substitution, ranged from agonists to partial agonists to antagonists at this receptor. The 2-substituted adenosine derivatives examined in this study were generally less potent at the A₁AR in comparison to the A₃AR, but fully efficacious. However, a single substitution at the 2-position could also lead to a potent and selective A₁AR agonist. For example, compound 21 is somewhat selective for the A₁AR. Mixed A₂A/A₃AR agonists have been identified. Additionally, we have identified several substituents at the 2-position that contribute significantly to the A_2BAR activity.

In conclusion, a number of novel structural determinants for the A₃AR activation have been identified. Given the interest in A₃AR agonists as antiischemic agents [35] and antagonists as antiglaucoma agents [36] and for other therapeutic applications, there is need for additional selective ligands to interact with the receptor. Presently, only one A₃AR agonist (IB-MECA) is in clinical trials [6], and the drug-like properties of most of the analogues in this study are unexplored. Selective agonists and antagonists, especially those whose selectivity extend across species, are needed both as receptor probes for research and as clinical candidates. In some cases, nucleoside AR ligands of mixed selectivity (e.g. mixed A₁/A₃AR agonists for cardioprotection) would be desirable for a particular clinical application. Structural insights gained in the present study may now be extended to multiple substitutions of adenosine.

Abbreviations

AR

adenosine receptor

CHO

Chinese hamster ovary

CPA

N⁶-cyclopentyladenosine

DMEM

Dulbeccos modified Eagles medium

I-AB-MECA

N⁶-(4-amino-3-iodobenzyl)-5′-N-methylcarboxa-midoadenosine

NECA

5′-N-ethylcarboxamidoadenosine

PIA

N⁶-phenylisopropyladenosine

PTLC

preparative thin layer chromatography