Derivatives of the Triazoloquinazoline Adenosine Antagonist (CGS15943) Are Selective for the Human A₃ Receptor Subtype

Abstract

The adenosine antagonist 9-chloro-2-(2-furanyl)[1,2,4]triazolo[1,5-c]quinazolin-5-amine (CGS15943) binds to human A₃ receptors with high affinity (K_i = 14 nM), while it lacks affinity at rat A₃ receptors. Acylated derivatives of the 5-amino group and other modifications were prepared in an effort to provide A₃ subtype selectivity. Affinity was determined in radioligand binding assays at rat brain A₁ and A_2A receptors using [³H]-(R)-PIA ([³H]-(R)-N⁶-(phenylisopropyl)-adenosine) and [³H]CGS 21680 ([³H]-2-[[4-(2-carboxy ethyl)phenyl]ethylamino]-5′-(N-ethylcarbamoyl)adenosine), respectively. Affinity was determined at cloned human A₃ receptors using [¹²⁵I]AB-MECA (N⁶-(4-amino-3-iodobenzyl)-5′-(N-methylcarbamoyl)adenosine). A series of straight chain alkyl amides demonstrated that the optimal chain length occurs with the 5-N-propionyl derivative, 3, which had a K_i value of 7.7 nM at human A₃ receptors, and was 40- and 14-fold selective vs rat A₁ and A_2A receptors, respectively. The 5-N-benzoyl derivative, 10, displayed K_i values of 680 and 273 nM at rat A₁ and A_2A receptors, respectively, and 3.0 nM at human A₃ receptors. A 5-N-phenylacetyl derivative, 12, was 470-fold selective for human A₃ vs rat A₁ receptors with a K_i value of 0.65 nM. A conjugate of Boc-γ-aminobutyric acid was also prepared but was nonselective. Conversion of the 5-amino group of CGS15943 to an oxo function resulted in lower affinity but 15-fold selectivity for human A₃ receptors.

Introduction

Four subtypes of adenosine receptors (A₁, A_2A, A_2B, and A₃) have been cloned and studied pharmacologically.¹ Since the recent discovery of the adenosine A₃ receptor and the cloning of its species homologues,² much effort has been made in order to characterize the receptor biochemically and pharmacologically. Principally, the development of selective agonist ligands³ has aided in this effort. The A₃ receptor mediates processes of inflammation,² hypotension,⁴ and mast cell degranulation.⁵ This receptor apparently also has a role in the central nervous system. The A₃ selective agonist IB-MECA induces behavioral depression⁶ and upon chronic administration protects against cerebral ischemia.⁷ A₃ selective agonists at high concentrations were also found to induce apoptosis in HL-60 human leukemia cells.⁸ These and other findings have made the A₃ receptor a promising therapeutic target.³ Selective antagonists for the A₃ receptor are sought as potential antiinflammatory or possibly antiischemic agents in the brain.^9–11

Thousands of analogues of the classical adenosine antagonists, the xanthines, have been prepared, resulting in high degrees of selectivity for A₁ or A_2A receptor subtypes.^12–14 However, the xanthines have not provided fruitful leads for A₃ receptor antagonists, since they are generally much weaker in binding at the A₃ subtype than at A₁ or A_2A subtypes.^9,10 Thus, the search for A₃ receptor antagonists has been directed toward the screening of libraries to identify structurally novel leads.^11,15–18 Diverse classes of heterocycles, non-xanthine adenosine antagonists, have already been found to bind to A₁ and A_2A receptors and more recently to A₃ adenosine receptors. We recently reported that two chemical classes, the flavonoids¹⁶ and dihydropyridines,¹⁸ are amenable to modification, resulting in ligands selective for the recently discovered A₃ receptor. The flavonoid I (MRS 1067)¹⁷ and the dihydropyridine derivative II (MRS 1097)¹⁸ (Figure 1) are selective human A₃ receptor antagonists.

Figure 1.

Structures and affinities (K_i value in μM) of novel A₃ adenosine receptor antagonists the flavonoid I (MRS 1067),¹⁷ the dihydopyridine II (MRS 1097),¹⁸ and the nonselective antagonist triazoloquinazoline 1 (CGS15943).^19,28,29

The non-xanthine adenosine antagonist 1 (CGS15943, 9-chloro-2-(2-furanyl)[1,2,4]triazolo[1,5-c]quinazolin-5-amine)^19,20 is very potent and slightly selective at human brain A_2A receptors with a K_i value of 1.7 nM.^21,29 It had been under development as an antiasthmatic agent, until unanticipated skin sensitivity was observed.¹² This antagonist was reported to be inactive in binding at rat A₃ receptors.¹⁰ Nevertheless, in light of the considerable species variability in affinity of non-adenosine derivatives at A₃ receptors, the affinity of 1 (CGS15943) was determined at cloned human A₃ receptors and was found to be relatively high. In this study we have prepared a series of derivatives of 1 and found that N-acylation greatly enhances affinity and selectivity for human A₃ receptors.

Results

Synthesis.

The structures of the triazoloquinazoline derivatives tested for affinity in radioligand binding assays at adenosine receptors are shown in Table 1. Several of the derivatives, e.g. the bromo derivative, 13, and the 5-oxo derivative, 15, had been reported previously by Francis et al.¹⁹ to bind to A₁ and A_2A receptors. The 5-oxo derivative¹⁹ was prepared in the present study by an alternate route using an acid treatment of commercially available 1 (Figure 2). Various N-acyl derivatives were synthesized from 1 using standard acylation methodologies (Figure 2). Compound 13^19,20 was prepared by first protecting the amino group as the tert-butyloxycarbonyl derivative, followed by bromination using N-bromosuccinimide and acidic deprotection. The yields and chemical characterization of these compounds are reported in Table 2.

Table 1.

Affinities of Triazoloquinazoline Derivatives in Radioligand Binding Assays at A₁, A_2A, and A₃ Receptors

				Ki (nM), IC50 (nM), or % inhibition
compd	R1	R2	R3	rA1a	rA2Ab	hA3c	rA1/hA3	rA2A/hA3
1 (CGS15943)	H	H		21 ± 3d	3.3 ± 1.7d	13.8 ± 2.4
2	COCH3	H		52.2 ± 2.6	36.3 ± 3.2	13.9 ± 2.5	3.8	2.6
3 (MRS1186)	COCH2CH3	H		283 ± 42	106 ± 30	7.66 ± 3.03	40	14
4	CO(CH2)2CH3	H		32.5 ± 9.4	37.6 ± 5.6	14.6 ± 2.8	2.2	2.6
5	CO(CH2)3CH3	H		28.9 ± 3.7	48.8 ± 10.5	21.5 ± 6.2	1.3	2.3
6	COC(CH3)3	H		205 ± 20	88.8 ± 20.5	244 ± 6	0.84	0.36
7	COOC(CH3)3	H		190 ± 16	92.0 ± 8.0	82.5 ± 23.3	2.3	1.1
8	CO(CH2)3NH-Boc	H		91.3 ± 13.2	135 ± 38	32.9 ± 1.7	2.8	4.1
9	CO(CH2)3NH2	H		11.3 ± 2.4	14.4	80.8 ± 7.4	0.14	0.18
10 (MRS1177)	COPh	H		680 ± 82	273 ± 42	3.03 ± 1.73	220	90
11	CO(3-I-Ph)	H		200	310 ± 151	23.8 ± 5.2	8.4	13
12 (MRS1220)	COCH2Ph	H		305 ± 51	52.0 ± 8.8	0.65 ± 0.25	470	80
13	H	Br		1570d	531d	64.0 ± 13.1
14	COPh	Br		<10%e	<10%e	856 ± 156
15			H	3950 ± 250	4380	260 ± 87	15	17
16			(CH2)2CH3	1850 ± 420	6960	1813 ± 720	1.0	3.8

^aDisplacement of specific [³H]-(R)-PIA binding in rat brain membranes, expressed as K_i ± SEM in nM (n = 3–5).

^bDisplacement of specific [³H]CGS 21680 binding in rat striatal membranes, expressed as K_i ± SEM in nM (n = 3–6).

^cDisplacement of specific [¹²⁵I]AB-MECA binding at human A₃ receptors expressed in HEK-293 cells, in membranes, expressed as K_i ±SEM in nM (n = 3–4).

^dIC₅₀ values from Francis et al.¹⁹

^ePercent displacement of specific binding at concentration 1 μM.

Figure 2.

Synthesis of N-acylated and various other derivatives of 1 (CGS15943).

Table 2.

Yields and Chemical Characterization of Triazoloquinazoline Derivatives

compd no.	methoda	% yield	mp (°C)	MS	formula	analysis
2	A	96	245	CI: 328	C15H10cN5O2C1·0.16CH2C12	C,H,N
3	A	84	224	CI: 342	C16H12N5O2Cl	C,H,N
4	A	96	232	CI: 356	C17H14N5O2Cl·0.2MeOH	C,H,N
5	A	73	220	FAB: 370	C18H16N5O2Cl·0.1EtOAc	C,H,N
6	B	70	200	CI: 370	C18H16N5O2Cl	C,H,N
7	A	98	285	CI: 386	C18H16N5O3Cl	C,H,N
8	C	49	209	CI: 471	C22H23N6O4Cl·0.11CH2Cl2	C,H,N
9		83	252	FAB: 371	C17H15N6O2Cl·C2H1O2F3·0.15CH2Cl2	C,H,N
10	A	73	239	CI: 390	C20H12N5O2Cl·0.6CH2Cl2	C,H,N
11	C	50	243	CI: 515	C20cH11N5O2ClI·0.75MeOH	C,H,N
12	B	35	233	EI: 403	C21H14N5O2Cl·0.38EtOAc	C,H,N
14		28	266	FAB: 469	C20H11N5O2BrCl	b
16		50	178	EI: 328	C16H13N502C1	b

^aA, symmetric anhydride method; B, acid chloride method; C, carbodiimide method.

^bHigh-resolution mass in EI or FAB⁺ mode (m/z) determined to be within acceptable limits. 14: calcd, 467.9863; found, 467.9871. 16: calcd, 328.0727; found, 328.0726.

Binding at Adenosine Receptors.

K_i values were determined in radioligand binding assays at rat cortical A₁ receptors vs [³H]-(R)-PIA ([³H]-(R)-N⁶-(phenylisopropyl)adenosine) or rat striatal A_2A receptors vs [³H]CGS 21680 ([³H]-2-[[4-(2-carboxyethyl)phenyl]ethylamino]-5′-(N-ethylcarbamoyl)adenosine.^22,23 Affinity at cloned human A₃ receptors expressed in HEK-293 cells²⁴ was determined using [¹²⁵I]AB-MECA (N⁶-(4-amino-3-[¹²⁵I]iodobenzyl)-5′-(N-methyl carbamoyl)adenosine).²⁵ In this study, affinities at adenosine receptors from different species are compared. The species differences in ligand affinity for triazoloquinazolines at A₁ and A_2A receptors are minor, whereas those at A₃ receptors are significant (Figure 1).^2,26 Thus, our ratios of affinity are indicative of selectivities in human, but not rat tissue.

The primary amine and 2-furanyl moieties in 1 have been considered most important for binding to rat A₁ and rat A_2A receptors.¹⁹ In an effort to investigate the effects on binding at human A₃ receptors, various structural modifications of these groups were made. The structure–activity relationships (SAR) for binding at adenosine receptors indicated that a variety of N-acyl groups are tolerated at the 5-amino position (Table 1). As with the parent compound 1,¹⁹ the affinity of the N-acylated derivatives (2–12) at A_2A receptors tended to equal or surpass the affinity at A₁ receptors. The substituent R₁ was varied from acetyl to valeryl in a homologous series (2–5). Also several acyl congeners containing either bulky tert-butyl or derivatized aromatic moieties were synthesized to investigate steric and electrostatic effects on the competitive binding at each of the adenosine receptors.

Among the straight acyl chain derivatives, the N-propionyl derivative, 3, had the highest binding affinity (K_i = 7.7 nM) and selectivity (40-fold vs rat A₁, 14-fold vs rat A_2A) at human A₃ receptors. Within this series, the affinity at A₃ receptors varied only 3-fold with varying chain lengths, while at A_2A receptors a greater variation was observed. The affinities of compounds 6 and 7 indicated that bulky groups at position R₁ were less well tolerated than straight alkyl chains in human A₃ receptor binding. The pivaloyl derivative, 6, was 32-fold less potent at A₃ receptors than the propionyl derivative, 3, from which it differed only in the presence of two methyl groups at a branched carbon. The change from an amide, 6, to the corresponding urethane, 7, had no effect on affinity at A₁ and A_2A receptors, while the affinity at A₃ receptors was enhanced ~3-fold. The competitive binding curves of compounds 3, 6, and 7 at human A₃ receptors are shown in Figure 3A.

Figure 3.

Representative competition curves comparing inhibition of binding at adenosine receptors. (A) Inhibition by compounds 3 (circles), 6 (squares), and 7 (triangles) of [¹²⁵I]AB-MECA binding at human A₃ receptors, expressed in HEK-293 cells. (B) Inhibition by compound 12 of [¹²⁵I]AB-MECA binding at human A₃ receptors (filled circles, transfected HEK-293 cell membranes, n_H = 0.75), [³H]-(R)-PIA binding at rat brain A₁ (filled diamonds, nH) = 0.89), and [³H]CGS 21680 binding at A_2A (filled squares, n_H = 0.84) receptors (in brain membranes).

A Boc-γ-amino butyric acid conjugate, 8, of 1 was slightly selective for human A₃ receptors. The removal of the Boc group resulting in an acyl chain containing the primary amine functionality, 9, which is positively charged at pH 7.4, reduced potency at human A₃ receptors, while increasing potency at rat A₁ and rat A_2A receptors.

Among aromatic acyl derivatives, the N-benzoyl derivative, 10, showed higher binding affinity (K_i = 3.0 nM) at the human A₃ receptor than 1, while affinities at rat A₁ and rat A_2A receptors were significantly decreased. At rat A₃ receptors,⁹ compound 10 was much less potent (data not shown), with only 23% of radioligand binding displaced at 3 μM. Substitution of the phenyl ring of the N-benzoyl group with a 3-iodo substituent, 11, caused a marked reduction in affinity only at human A₃ receptors. However, the less sterically hindered N-phenylacetyl compound, 12, showed yet higher affinity (K_i = 0.65 nM) and selectivity (470-fold vs rat A₁ receptors) for human A₃ receptors than 10. The competitive binding curves of compound 12 at rat A₁ and A_2A receptors and at human A₃ receptors are shown in Figure 3B.

Two 5-bromofuranyl derivatives were prepared. Compound 13, which was used as the synthetic precursor in the preparation of tritiated 1²⁰ and reported to be much less potent at rat A₁ and rat A_2A receptors,¹⁹ showed appreciable potencies at human A₃ receptors. Compound 14 was prepared with an expectation of improving selectivity and affinity with respect to compounds 10 and 13. Unfortunately, this combination of functionalities displayed a dramatic loss of binding activity at all three adenosine receptor subtypes.

Finally, two 5-oxo compounds were also prepared and tested. Compound 15 was previously reported to be much less potent than 1 at A₁ and A_2A receptors,¹⁹ and we found it to be 15–17-fold A₃ receptor selective. However, compound 16, the N⁶-propyl derivative of 15 displaced radioligand binding with lower affinity at the A₃ adenosine receptor subtype.

Discussion

Until the present there have been few reports of leads for the development of selective antagonists for the A₃ receptor,^17,18 especially having high potency. We have screened chemical libraries¹¹ and known adenosine receptor ligands¹⁰ for potential A₃ receptor antagonists. Initially we reported that the triazoloquinazoline 1, which was reported as the first potent, non-xanthine, but nonselective adenosine A₁ and A_2A receptor antagonist, did not have appreciable binding activity at rat A₃ receptors.¹⁰ Nevertheless, in this study, it proved to be highly potent at human A₃ receptors (K_i = 13.8 nM), suggesting that it may serve as a lead compound in this species. This finding encouraged us to investigate structure–activity relationships of 1 derivatives for the development of highly potent and selective antagonists at human but not rat A₃ receptors. There are dramatic species differences^2,26 in the affinity of antagonists binding at A₃ receptors. The overall identity of protein sequences between the A₃ adenosine receptors of rat and human was reported as 72%,^1,2 which is relatively low, consistent with the high variability in antagonist affinity.

The present findings indicated that certain less polar derivatives of 1 displayed increased affinity at the human A₃ receptor, and this enhanced binding affinity could be distinguished from effects at rat A₁ and rat A_2A receptors. Compound 13 and 15, which were previously reported to have poor binding affinities at rat A₁ and A_2A receptors, showed binding affinities to human A₃ receptors in the sub-micromolar range; thus some degree of selectivity was present. The most significant findings in the present study are that compounds 10 and 12, containing the N-benzoyl and N-phenylacetyl substituents, respectively, are highly potent and selective for human A₃ receptors. A competitive binding assay of 10 showed high affinity at human A₃ receptors (K_i = 3.0 nM), with 220-fold selectivity vs rat A₁ receptors and 90-fold selectivity vs rat A_2A receptors. Compound 12 displayed a K_i value of 0.65 nM, with 470-fold selectivity vs rat A₁ receptor and 80-fold selectivity vs rat A_2A receptors. The displacement of radioligand binding at rat A₃ receptors by 10 and 12 was only 23% (3 μM) and 12% (1 μM), respectively, showing that there are marked differences of affinities between human A₃ and rat A₃ receptors.

Compound 14, the benzoyl derivative of compound 13, however, had greatly diminished affinity at adenosine receptors; thus the A₃ receptor selectivity enhancing effects induced by N-acylation and by 5-bromination are not additive.

N-Acylations with various alkanecarboxylic acids resulted in less selective ligands than the aryl carboxylic acid derivatives. A structure–activity relationship analysis of various length of acyl chain substituents (2–7) resulted in the interesting finding that compound 3 displayed the optimum acyl chain length for A₃ receptor binding.

There appears to be a hydrophobic pocket in the receptor in the vicinity of the N⁵-amino group, since hydrophobic N-acyl and N-aryl substituents could enhance potency. In contrast, removal of the hydrophobic Boc group from 8 resulted in the positively charged congener 9 that displayed a lower affinity at A₃ receptors. According to molecular modeling,¹⁵ in the receptor bound states the N⁶ region of adenosine corresponds to the position of N⁵ of 1. In the present study, the introduction of an aryl ring in that region enhanced selectivity, as was found in the case of N⁶-benzylad-enosine derivatives.³ Compound 11 was synthesized to test further the hypothesis of overlap with the N⁶ region of adenosine, since the N⁶-3-iodobenzyl substituent, vs unsubstituted N⁶-3-benzyl, has been shown to enhance A₃ receptor affinity and selectivity. Nevertheless, in the present series, the iodo substituent offered no advantage. N⁵ is also proposed to correspond to the N3-position of xanthines,¹⁵ at which large hydrophobic substituents are tolerated,¹⁰ consistent with our findings.

The selective ligands introduced here should be useful as antagonists in characterizing and probing the physiological role of human A₃ receptors. They may also provide affinity probes such as radioligands for human A₃ receptors. These selective agents must now be studied in functional assays.

Experimental Section

Materials.

Compound 1, (R)-PIA, and 2-chloroadenosine were purchased from Research Biochemicals International (Natick, MA). All acylating agents were obtained from Aldrich (St. Louis, MO).

Synthesis.

Proton nuclear magnetic resonance spectroscopy was performed on a Varian GEMINI-300 spectrometer and spectra were taken in DMSO-d₆ or CDCl₃. Unless noted, chemical shifts are expressed as ppm downfield from tetramethylsilane. Chemical-ionization (CI) mass spectrometry was performed with a Finnigan 4600 mass spectrometer and electron-impact (EI) mass spectrometry with a VG7070F mass spectrometer at 6 kV. Elemental analysis was performed by Atlantic Microlab Inc. (Norcross, GA) or Galbraith Laboratories, Inc. (Knoxville, TN). All melting points were determined with a Unimelt capillary melting point apparatus (Arthur H. Thomas Co., PA) and were uncorrected. All triazoloquinone derivatives showed one spot on TLC (MK6F silica, 0.25 mm, glass backed, Whatman Inc., Clifton, NJ). Where needed evaluation of purity was done on a Hewlett-Packard 1090 HPLC system using OD-5–60 C18 analytical column (150 mm × 4.6 mm, Separation Methods Technologies, Inc., Newark, DE) in two different linear gradient solvent systems. One solvent system (A) was 0.1 M TEAA/CH₃CN, 30:70 to 10:90, in 20 min with flow rate 1 mL/min. The other (B) was H₂O/MeOH, 40:60 to 10:90, in 20 min with flow rate 1 mL/min. Peaks were detected by UV absorption using a diode array detector.

General Procedure for Preparation of 5-N-Acyl Derivatives of 1.

Method A (Symmetrical Anhydride).

To a stirred solution of 1 (10 mg, 0.035 mmol), anhydride (0.105 mmol) and (dimethylamino)pyridine (0.5 mg, 0.004 mmol) in 1.5 mL of anhydrous DMF was added triethylamine (73 μL, 0.525 mmol) at room temperature. The mixture was stirred for 48 h and then evaporated to dryness under reduced pressure. The residue was purified by preparative silica gel TLC (CH₂Cl₂/MeOH, 50:1 ~ 75:1) to afford the desired compounds (2–5, 7, and 10).

Method B (Acid Chloride).

To a stirred solution of 1 (10 mg, 0.035 mmol) and anhydrous pyridine (40 μL, 0.5 mmol) in 1.5 mL of anhydrous CH₂Cl₂ was added acyl chloride (0.105 mmol) at 0 °C. The mixture was stirred at room temperature for 24–48 h and then treated with same procedure as method A for purification of the desired compounds (6 and 12).

Method C (Carbodiimide).

A solution of 1 (10 mg, 0.035 mmol), the carboxylic acid component (0.210 mmol), 1-(3-(dimethylamino)propyl)-3-ethylcarbodiimide (41 mg, 0.210 mmol), 1-hydroxybenzotriazole (28 mg, 0.210 mmol), 4-(dimethylamino)pyridine (0.5 mg, 0.004 mmol), and triethylamine (74 μL, 0.530 mmol) in 2 mL of anhydrous DMF/CH₂Cl₂ (1:1 v/v) was stirred at room temperature for 48 h. The mixture was treated with same procedure as method A for purification of desired compounds (8 and 11).

5-(Acetylamino)-9-chloro-2-(2-furanyl)[1,2,4]triazolo-[1,5-c]quinazoline (2):

¹H NMR (CDCl₃) δ 2.78 (3H, s, CH₃CO), 6.63–6.65 (1H, m, H-4′), 7.30 (1H, d, J = 2.9, H-3′),7.68 (1H, broad s, H-5′), 7.73 (1H, dd, J = 1.9, 8.8, H-8), 7.86 (1H, d, J = 8.8, H-7), 8.48 (1H, d, J = 2.9, H-10), 8.99 (NH, broad s).

9-Chloro-2-(2-furanyl)-5-(n-propionylamino)[1,2,4]-triazolo[1,5-c]quinazoline (3):

¹H NMR (CDCl₃) δ 1.34 (3H, t, J = 7.8, CH₃CH₂CO), 3.10 (2H, q, J = 7.8, CH₃CH₂CO), 6.63–6.65 (1H, m, H-4′), 7.31 (1H, d, J = 3.9, H-3′), 7.68 (1H, d, J = 1.9, H-5′), 7.73 (1H, dd, J = 2.0, 8.8, H-8), 7.88 (1H, d, J = 8.8, H-7), 8.48 (1H, d, J = 2.0, H-10), 9.01 (NH, broad s).

5-(n-Butyrylamino)-9-chloro-2-(2-furanyl)[1,2,4]triazolo-[1,5-c]quinazoline (4):

¹H NMR (CDCl₃) δ 1.10 (3H, t, J = 7.4, 7.3, CH₃CH₂CH₂CO), 1.84–1.91 (2H, m, CH₃CH₂CH₂CO), 3.03 (2H, t, J = 7.4, 7.3, CH₃CH₂CH₂CO), 6.63–6.65 (1H, m, H-4′), 7.31 (1H, d, J = 3.4, H-3′), 7.68 (1H, d, J = 1.8, H-5′), 7.73 (1H, dd, J = 2.3, 8.8, H-8), 7.88 (1H, d, J = 8.8, H-7), 8.48 (1H, d, J = 2.3, H-10), 8.97 (NH, broad s).

9-Chloro-2-(2-furanyl)-5-(n-pentanoylamino)[1,2,4]-triazolo[1,5-c]quinazoline (5):

¹H NMR (CDCl₃) δ 1.01 (3H, t, J = 6.8, 7.8, CH₃CH₂CH₂CH₂CO), 1.48–1.55 (2H, m, CH₃CH₂CH₂CH₂CO), 1.77–1.85 (2H, m, CH₃CH₂CH₂CH₂CO), 3.05 (2H, t, J = 7.82, 6.8, CH₃CH₂CH₂CH₂CO), 6.63–6.65 (1H, m, H-4′), 7.31 (1H, d, J = 2.9, H-3′), 7.68 (1H, broad s, H-5′), 7.73 (1H, dd, J = 2.0, 8.8, H-8), 7.88 (1H, d, J = 8.8, 7-H), 8.48 (1H, d, J = 2.9, 10-H), 8.98 (NH, broad s).

9-Chloro-2-(2-furanyl)-5-[(trimethylacetyl)amino][1,2,4]-triazolo[1,5-c]quinazoline (6):

¹H NMR (CDCl₃) δ 1.47 (9H, s, (CH₃)₃CCO), 6.63–6.65 (1H, m, H-4′), 7.31 (1H, d, J = 3.4, H-3′), 7.68–7.69 (1H, m, H-5′), 7.74 (1H, dd, J = 2.5, 8.9, H-8), 8.01 (1H, d, J = 8.9, H-7), 8.49 (1H, d, J = 2.5, H-10), 9.39 (NH, broad s).

5-[(tert-Butoxycarbonyl)amino]-9-chloro-2-(2-furanyl)-[1,2,4]triazolo[1,5-c]quinazoline (7):

¹H NMR (CDCl₃) δ 1.63 (9H, s, (CH₃)₃COCO), 6.65–6.66 (1H, m, H-4′), 7.35 (1H, d, J = 3.5, H-3′), 7.69 (1H, broad s, H-5′), 7.74 (1H, dd, J = 2.6, 8.9, H-8), 8.01 (1H, d, J = 8.9, H-7), 8.49 (1H, d, J = 2.6, H-10), 8.56 (NH, broad s).

5-[[4-[(tert-Butoxycarbonyl)amino]butyryl]amino]-9-chloro-2-(2-furanyl)[1,2,4]triazolo[1,5-c]quinazoline (8):

¹H NMR (CDCl₃) δ 1.44 (9H, s, (CH₃)₃COCONH), 2.03 (2H, pen, J = 6.8, NHCH₂CH₂CH₂CO), 3.13 (2H, t, J = 6.8, 7.3, NHCH₂CH₂CH₂CO), 3.32 (2H, q, J = 6.4, NHCH₂CH₂CH₂-CO), 4.81 (NH, broad s), 6.63–6.65 (1H, m, H-4′), 7.27–7.32 (1H, m, H-3′), 7.67–7.68 (1H, m, H-5′), 7.71–7.75 (1H, m, H-8), 7.91 (1H, d, J = 8.8, H-7), 8.48 (1H, d, J = 2.0, H-10), 9.15 (NH, broad s).

5-[(4-aminobutyryl)amino]-9-chloro-2-(2-furanyl)[1,2,4]-triazolo[1,5-c]quinazoline (9).

A solution of 8 (3 mg, 6.4 μmol) in 1 mL of 50% TFA in CH₂Cl₂ was left to stand for 10 min and evaporated under reduced pressure. Crystallization of the residue with MeOH/CH₂Cl₂ gave 3.2 mg of 9 (83% as TFA salt) as a white solid: ¹H NMR (DMSO-d₆) δ 1.87–1.97 (2H, m, H₂NCH₂CH₂CH₂CO), 2.82 (2H, t, J = 7.32, H₂NCH₂-CH₂CH₂CO), 2.88–2.94 (2H, m, H₂NCH₂CH₂CH₂CO), 6.78–6.80 (1H, m, H-4′), 7.34 (1H, d, J = 3.05, H-3′), 7.77 (NH₂, broad s), 7.93–8.02 (3H, m), 8.40 (1H, d, J = 1.2, H-10), 11.17 (NH, s).

5-(Benzoylamino)-9-chloro-2-(2-furanyl)[1,2,4]triazolo-[1,5-c]quinazoline (10):

¹H NMR (CDCl₃) δ 6.64–6.66 (1H, m, H-4′), 7.35 (1H, d, J = 2.9, H-3′), 7.61–7.68 (3H, m), 7.68–7.70 (1H, m, H-5), 7.77 (1H, dd, J = 2.0, 8.8, H-8), 8.04–8.10 (3H, m), 8.52 (1H, d, J = 2.0), 9.75 (NH, broad s).

9-Chloro-2-(2-furanyl)-5-[(3-iodobenzoyl)amino][1,2,4]-triazolo[1,5-c]quinazoline (11):

¹H NMR (CDCl₃) δ 6.64–6.66 (1H, m, H-4′), 7.30–7.36 (3H, m), 7.70 (1H, broad s, H-5′), 7.77 (1H, dd, J = 2.0, 10.7, H-8), 7.92–8.10 (2H, broad m), 8.39 (1H, m), 8.52 (1H, d, J = 2.0, hH-10), 9.62 (NH, broad s).

9-Chloro-2-(2-furanyl)-5-[(phenylacetyl)amino][1,2,4]-triazolo[1,5-c]quinazoline (12):

¹H NMR (CDCl₃) δ 4.38 (2H, s, CH₂CO), 6.62–6.65 (1H, m, H-4′), 7.24–7.26 (1H, m, H-3′), 7.35–7.44 (5H, m), 7.67–7.68 (1H, m, H-5′), 7.74 (1H, dd, J = 2.2, 9.0, H-8), 7.93 (1H, d, J = 8.8, H-7), 8.49 (1H, m, H-10), 9.10 (NH, broad s).

5-Amino-2-[2-(5-bromofuranyl)]-9-chloro[1,2,4]triazolo-[1,5-c]quinazoline (13):

A solution of 7 (0.01 g, 0.026 mmol) and N-bromosuccinimide (0.005 g, 0.028 mmol) in 2 mL of AcOH/CHCl₃ (1:1) was stirred for 1 h at room temperature. The mixture was poured into 10 mL of saturated NaHCO₃ solution, and the product was extracted with 10 mL of CHCl₃ three times. The combined CHCl₃ solution was washed with brine, dried over anhydrous Na₂SO₄, and evaporated to dryness under reduced pressure. The residue was purified by preparative silica gel TLC (CHCl₃/MeOH, 80:1) to afford 2-[2-(5-bromofuranyl)]-5-[(tert-butoxycarbonyl)amino]-9-chloro[1,2,4]-triazolo[1,5-c]quinazoline (0.012 g, 99%) as a white solid: MS (CI, NH₃) 466 (M⁺ + 1); ¹ H-NMR (CDCl₃) δ 1.63 (9H, s, (CH₃)₃-OCO), 6.58 (1H, d, J = 3.6, H-4′), 7.29 (1H, d, J = 3.6, H-3′), 7.73 (1H, dd, J = 2.3, 8.8, H-8), 7.98 (1H, d, J = 9.0, H-7), 8.46 (1H, d, J = 2.3, H-10), 8.53 (NH, broad s). To a solution of this compound in 2 mL of CH₂Cl₂ was added TFA (0.05 mL, 0.67 mmol), and the mixture was stirred for 2 h at room temperature. The reaction mixture was treated with same workup procedure above. A preparative silica gel TLC (n-hexane/CHCl₃/MeOH, 1:1:0.1) of the crude product gave 13 (4.5 mg, 48%) as a white solid: MS (CI, NH₃) 366 (M⁺ + 1), 383 (M⁺ + 18); ¹H-NMR (CDCl₃) δ 5.94 (NH₂, broad s), 6.56 (1H, d, J = 3.8, H-4′), 7.25 (1H, d, J = 3.8, H-3′), 7.63–7.65 (2H, m, H-8 + H-7), 8.41 (1H, d, J = 2.1, H-10).

5-(Benzoylamino)-2-[2-(5-bromofuranyl)]-9-chloro[1,2,4]-triazolo[1,5-c]quinazoline (14).

The furanyl group in compound 10 was brominated by the same method as above: ¹H-NMR (CDCl₃) δ 6.64–6.66 (1H, m, H-4′), 7.35 (1H, d, J = 2.9, H-3′), 7.61–7.68 (2H, m), 7.68–7.70 (1H, m), 7.77 (1H, dd, J = 2.0, 8.8, H-8), 8.04–8.10 (3H, m), 8.52 (1H, d, J = 2.0, H-10), 9.75 (NH, broad s); HPLC retention time: 4.6 min (>95% purity) using solvent system A, 12.5 min (>95% purity) using solvent system B.

9-Chloro-2-(2-furanyl)[1,2,4]triazolo[1,5-c]quinazolin-5(6H)-one (15).

A solution of 1 (0.075 g, 0.263 mmol) in 8.0 mL of AcOH and 2.0 mL of H₂O in a sealed tube was heated for 72 h at 100 °C. The solution was coevaporated with toluene under reduced pressure, and the residue was purified by preparative silica gel TLC (CHCl₃/MeOH, 15:1) to afford 15 (0.065 g, 86%) as a white solid: mp >310 °C; MS (CI NH₃) 287 (M⁺ + 1), 304 (M⁺ + 18); ¹H-NMR (DMSO-d₆) δ 6.73–6.74 (1H, m, H-4′), 7.26 (1H, d, J = 3.5, H-3′), 7.46 (1H, d, J = 8.8, H-7), 7.77 (1H, dd, J = 2.5, 8.9, H-8), 7.97 (1H, s, H-5′), 8.16 (1H, d, J = 2.5, H-10), 12.47 (NH, s).

9-Chloro-2-(2-furanyl) 6-n-propyl[1,2,4]triazolo[1,5-c]-quinazolin-5-one (16).

To a suspension of 15 (0.021 g, 0.073 mmol) in 2 mL of anhydrous THF was added a suspension of NaH (6 mg, 60% in mineral oil, prewashed with n-hexane, 0.15 mmol) in 2 mL of anhydrous THF followed by HMPA (0.21 mL, 12 mmol) under N₂ atmosphere at room temperature. The mixture was stirred vigorously for 30 min (H₂ gas evolved). 1-Bromopropane (28 μL, 0.3 mmol) was added, and the reaction mixture was refluxed for 6 h. After cooling, the precipitate was removed by filtration through a small volume of silica gel bed and the filtrate was evaporated. The residue was purified by preparative silica gel TLC (n-hexane/EtOAc, 2:1) to afford 16 (0.012 g, 50%) as a white solid: 1H-NMR (CDCl₃) δ 1.11 (3H, t, J = 7.5, 7.5, CH₃CH₂CH₂N), 1.85–1.93 (2H, m, CH₃CH₂-CH₂N), 4.35 (2H, t, J = 8.0, 7.5, CH₃CH₂CH₂N), 6.60–6.61 (1H, m, H-4′), 7.32 (1H, d, J = 3.4, H-3′), 7.37 (1H, d, J = 9.3, H-7), 7.66 (1H, broad s, H-5′), 7.68 (1H, dd, J = 2.39, 9.0, H-8), 8.52 (1H, d, J = 2.5, H-10); HPLC retention time 3.5 min (>95% purity) using solvent system A, 9.1 min (>95% purity) using solvent system B.

Pharmacology: Radioligand Binding Studies.

Binding of [³H]-(R)-PIA to A₁ receptors from rat cerebral cortex membranes and of [³H]CGS 21680 to A_2A receptors from rat striatal membranes was performed as described previously.^6,8 Adenosine deaminase (3 units/mL) was present during the preparation of the brain membranes, in a preincubation of 30 min at 30 °C and during the incubation with the radioligands.

Binding of [¹²⁵I]AB-MECA in membranes prepared from HEK-293 cells stably expressing the human A₃ receptor (Receptor Biology, Inc., Baltimore, MD), or from CHO cells stably expressing the rat A₃ receptor, was as described.^17,25 The assay medium consisted of a buffer containing 50 mM Tris, 10 mM Mg²⁺, and 1 mM EDTA, at pH 8.0. The glass incubation tubes contained 100 μL of the membrane suspension (0.3 mg of protein/mL, stored at −80 °C in the same buffer), 50 μL of [¹²⁵I]AB-MECA (final concentration 0.3 nM), and 50 μL of a solution of the proposed antagonist. Nonspecific binding was determined in the presence of 200 μM NECA.

All nonradioactive compounds were initially dissolved in DMSO and diluted with buffer to the final concentration, where the amount of DMSO never exceeded 2%.

Incubations were terminated by rapid filtration over Whatman GF/B filters, using a Brandell cell harvester (Brandell, Gaithersburg, MD). The tubes were rinsed three times with 3 mL buffer each.

At least five different concentrations of competitor, spanning 3 orders of magnitude adjusted appropriately for the IC₅₀ of each compound, were used. IC₅₀ values, calculated with the nonlinear regression method implemented in the InPlot program (Graph-PAD, San Diego, CA), were converted to apparent K_i values using the Cheng-Prusoff equation²⁷ and K_d values of 1.0 nM, 14 nM for [³H]-(R)-PIA and [³H]CGS 21680, and 0.59 nM for binding of [¹²⁵I]AB-MECA at human A₃ receptors, respectively. Most Hill coefficients of tested compounds were in the range 0.8–1.1.

Abbreviations:

AcOH

acetic acid

Boc

tert-butoxycarbonyl

CGS 21680

2-[[4-(2-carboxyethyl)phenyl]ethylamino]-5′-N-(ethylcarbamoyl)adenosine

CGS15943

9-chloro-2-(2-furanyl)[1,2,4]triazolo[1,5-c]quinazolin-5-amine

CHO cells

Chinese hamster ovary cells

CI

chemical ionization

DMF

N,N-dimethylformamide

DMSO

dimethyl sulfoxide

EDTA

ethylenediaminetetraacatate

HEK cells

human embryonic kidney cells

HMPA

hexamethylphosphotriamide

[¹²⁵I]AB-MECA

[¹²⁵I]-N⁶-(4-amino-3-iodobenzyl)-5′-(N-methylcarbamoyl)adenosine

K _i

equilibrium inhibition constant

MS

mass spectrum

NECA

(N-ethylcarbamoyl)adenosine

(R)-PIA

(R)-N⁶-(phenylisopropyl)adenosine

SAR

structure-activity relationship

TFA

trifluoroacetic acid

THF

tetrahydrofuran

TLC

thin layer chromatography

Tris

tris(hydroxymethyl)amino-methane