Structure–Activity Relationships at Human and Rat A_2B Adenosine Receptors of Xanthine Derivatives Substituted at the 1-, 3-, 7-, and 8-Positions

Abstract

In the search for improved selective antagonist ligands of the A_2B adenosine receptor, which have the potential as antiasthmatic or antidiabetic drugs, we have synthesized and screened a variety of alkylxanthine derivatives substituted at the 1-, 3-, 7-, and 8-positions. Competition for ¹²⁵I-ABOPX (¹²⁵I-3-(4-amino-3-iodobenzyl)-8-(phenyl-4-oxyacetate)-1-propylxanthine) binding in membranes of stably transfected HEK-293 cells revealed uniformly higher affinity (<10-fold) of these xanthines for human than for rat A_2B adenosine receptors. Binding to rat brain membranes expressing A₁ and A_2A adenosine receptors revealed greater A_2B selectivity over A_2A than A₁ receptors. Substitution at the 1-position with 2-phenylethyl (or alkyl/olefinic groups) and at N-3 with hydrogen or methyl favored A_2B selectivity. Relative to enprofylline 2b, pentoxifylline 35 was equipotent and 1-propylxanthine 3 was >13-fold more potent at human A_2B receptors. Most N-7 substituents did not enhance affinity over hydrogen, except for 7-(2-chloroethyl), which enhanced the affinity of theophylline by 6.5-fold to 800 nM. The A_2B receptor affinity-enhancing effects of 7-(2-chloroethyl) vs 7-methyl were comparable to the known enhancement produced by an 8-aryl substitution. Among 8-phenyl analogues, a larger alkyl group at the 1-position than at the 3-position favored affinity at the human A_2B receptor, as indicated by 1-allyl-3-methyl-8-phenylxanthine, with a K_i value of 37 nM. Substitution on the 8-phenyl ring indicated that an electron-rich ring was preferred for A_2B receptor binding. In conclusion, new leads for the design of xanthines substituted in the 1-, 3-, 7-, and 8-positions as A_2B receptor-selective antagonists have been identified.

Introduction

Four extracellular G protein-coupled receptors for adenosine have been identified as follows: A₁, A_2A, A_2B, and A ₃.¹ A_2B receptors, which are coupled to stimulation of adenylyl cyclase^2,3 and also lead to a rise in intracellular calcium,⁴ are involved in the control of vascular tone,⁵ hepatic glucose balance,⁶ cell growth and gene expression,⁷ mast cell degranulation,⁸ and intestinal water secretion.⁹ Activation of A_2B receptors in human retinal endothelial cells may lead to neovascularization by a mechanism involving increased vascular endothelial growth factor expression.⁷ Selective xanthine antagonists of the A_2B receptor have recently been reported.^10,11 Such antagonists are potentially useful in the treatment of asthma,^12,13 intestinal disorders,⁹ and diabetes (through improved glucose uptake in skeletal muscle and suppression of glucose production in hepatocytes).⁶

On the basis of adenosine receptor binding assays, we have identified several new xanthines with improved potency and selectivity for human A_2B receptors.^10,14 A P-cyanoanilide derivative in this series, N-(4-cyanophenyl)-2-[4-(2,3,6,7-tetrahydro-2,6-dioxo-1,3-dipropyl-1H-purin-8-yl)phenoxy]acetamide) 1, was shown to be 204-, 255-, and 289-fold selective for human A_2B receptors vs human A₁, A_2A, and A₃ receptors, although less selective vs rat A₁ and A_2A receptors. Compound 1 (100 nM) was shown to completely inhibit calcium mobilization stimulated by 1 μM NECA in HEK-293 (human embryonic kidney) cells expressing human A_2B receptors.¹⁰

Nonselective radioligands have been used to characterize binding to recombinant human A_2B receptors overexpressed in cell lines such as HEK-293. These include ¹²⁵I-ABOPX (¹²⁵I-3-(4-amino-3-iodobenzyl)-8-(phenyl-4-oxyacetate)-1-propylxanthine),¹⁵ [³H]8-cyclopentyl-1,3-dipropylxanthine,¹³ and [³H]4-(2-[7-amino-2-{furyl}{1,2,4}triazolo{2,3-a}{1,3,5}triazin-5-ylamino-ethyl)phenol.¹⁶ [³H]N-(4-Cyanophenyl)-2-[4-(2,3,6,7-tetrahydro-2,6-dioxo-1,3-dipropyl-1H-purin-8-yl)phenoxy]-acetamide) 1 was recently introduced as a new selective radioligand that bound to human A_2B receptors with a k_D of 1 nM.¹⁷

In the present study, in an effort to better understand the structure–activity relationships (SAR) of antagonist ligands of the A_2B adenosine receptor, we have screened a variety of alkylxanthine derivatives substituted at the 1-, 3-, 7-, and 8-positions and synthesized some derivatives that were designed by analysis of the biological results obtained from screening. In this study, previously uncharacterized patterns in the SAR of rat and human A_2B receptors have been identified.

Results

Chemistry.

Sixty-five xanthine derivatives (Table 1) were examined as antagonists of binding at the human A_2B adenosine receptor. Many of these ligands had been investigated in earlier studies of A₁ and A_2A receptors.^{14,18–25,37} Other compounds were synthesized, based on SAR indications that arose during the course of the present study.

Table 1.

Affinities of Xanthine Derivatives in a Radioligand Binding Assay at Human A_2B Receptors

compd	R1	R3	R7	R8	Kia (nM)	compd	R1	R3	R7	R8	Kia (nM)
2a (theophylline)	Me	Me	H	H	9070 ± 1490	31	CH2OCH3	Me	Me	H	14 000 ± 2400
2b (enprofylline)	H	Pr	H	H	4730 ± 270	32	CH2CN	Me	Me	H	21 600 ± 5200
3	Pr	H	H	H	360 ± 70	33	(CH2)2OC-OCH3	Me	Me	H	8710 ± 1320
4	propargyl	H	H	H	552 ± 76	34	CH2COCH3	Me	Me	H	20 800 ± 4800
5	allyl	H	H	H	461 ± 18 (4)	35 (pentoxifylline)	(CH2)4CO-CH3	Me	Me	H	5180 ± 2010
6	n-Bu	H	H	H	421 ± 38 (4)	36	Me	Propargyl	Me	H	9480 ± 1340
7	2-phenylethyl	H	H	H	408 ± 67 (4)	37	Pr	H	H	cyclopentyl	34 ± 10
8	cyclopentyl	H	H	H	965 ± 64 (4)	38 (CPX)	Pr	Pr	H	cyclopentyl	63.8 ± 8.3
9	Et	Et	H	H	1770 ± 260	39 (8-PT)	Me	Me	H	C6H5	415 ± 219
10	n-Pr	n-Pr	H	H	1110 ± 330 (4)	40 (SPT)	Me	Me	H	C6H4-4-S-O3H	1330 ± 220
11	allyl	allyl	H	H	1330 ± 240	41	Me	Me	H	C6H4-4-OH	60.7 ± 3.1
12	n-hexyl	n-hexyl	H	H	7580 ± 1900	42	Me	Me	II	C6H4-4-OCH3	394 ± 38
13	Et	Me	H	H	1620 ± 440	43	Me	Me	H	C6H4-4-NO2	6940 ± 1960
14a	propargyl	Me	H	H	511 ± 52 (4)	44	Me	Me	H	C6H4-4-N-(CH3)2	289 ± 11
14b	benzyl	Me	H	H	10 200 ± 2000	45	Me	Me	II	C6H4-2-CO2H	11 100 ± 4700
14c	2-phenylethyl	Me	H	H	646 ± 15	46	Me	Me	H	C6H4-3-CO2H	5530 ± 640 (4)
14d	3-phenylpropyl	Me	H	H	2330 ± 250	47	Me	Me	H	C6H3-2,4-(NO2)2	15 400 ± 6600
15	Me	i-Pr	H	H	4890 ± 870	48 (DPX)	Et	Et	H	C6H5	62.0 ± 11.4
16	n-Bu	Et	H	H	1720 ± 160	49	Pr	Pr	H	C6H5	18.9 ± 3.2 (4)
17	i-amyl	i-Bu	H	H	6240 ± 1340	50	Pr	Pr	H	pyrazine	79.6 ± 11.3
18 (caffeine)	Me	Me	Me	H	10 400 ± 1800	51	allyl	Me	H	C6H5	37 ± 6
19a	Me	Me	2-Cl-ethyl	H	1390 ± 200	52	i-amyl	i-Bu	H	C6H5	9890 ± 1440
19b	Me	Me	3-Cl-propyl	H	5620 ± 980	53	Pr	Pr	H	C6H3-2,6-F2	310 ± 42
19c	Me	Me	4-Cl-butyl	H	15 000 ± 1100	54	Pr	Pr	H	C6H3-2,3-(OH)2	125 ± 10
20	Me	Me	propargyl	H	9460 ± 1880	55	Pr	Pr	H	C6H3-2,4-(OH)2	20.7 ± 2.0
21	Me	Me	allyl	H	9490 ± 1220	56	Pr	Pr	H	C6H3-2,5-(OH)2	138 ± 6.9
22	Me	Me	benzyl	H	19 400 ± 4140	57	Pr	Pr	H	C6H3-2-OH-4-OCH3	46.5 ± 0.7
23	Me	Me	CH2COOH	H	27 600 ± 11 400	58	Pr	Pr	H	C6H3-3,5-(OH)2-4-OCH3	59.5 ± 7.0
24	Me	Me	CH2CN	H	14 500 ± 300	59	Pr	Pr	H	C6H4-4-NHC-OCH3	36.7 ± 9.5 (4)
25	Me	Me	(CH2)2-NH2	H	23 500 ± 6700	60	Pr	Pr	H	C6H3-2-OMe-4-Cl	6530 ± 570 (4)
26	Me	Me	(CH2)2OH	H	17 300 ± 2900	61 (XCC)	Pr	Pr	H	C6H4-4-OCH2-CO2H	40 ± 4
27	Me	Me	(CH2)2OCO-CH3	H	23 200 ± 8300	62	Pr	Pr	H	C6H3-2-OH-4-OCH2-CO2H	188 ± 22
28	allyl	allyl	Me	H	3390 ± 760	63	allyl	Me	Me	C6H5	11 400 ± 1000
29	propargyl	Me	Me	H	4130 ± 1040	64	allyl	Me	2-Cl-ethyl	C6H5	311 ± 22
30	CH2COOH	Me	Me	H	23 500 ± 7600	65	Pr	Pr	2-Cl-ethyl	C6H4	691 ± 72

^aDisplacement of specific [¹²⁵I]ABOPX binding at human A_2B receptors expressed in HEK-293 cells, in membranes, expressed as K_i ± SEM. n = 3, unless noted. Cases in which n = 4 are noted in parentheses.

Compounds 14b–d, which were substituted with various phenylalkyl groups in the 1-position, were synthesized in four steps (Scheme 1). Alkylation of 6-amino-1-methyluracil was performed with 10% aqueous NaOH and the appropriate alkyl halide. Compounds 66b–d were converted to xanthines 14b–d according to standard procedures as reported.^10,25 The synthesis of several 8-H and 8-aryl derivatives alkylated at the 7-position, 19b, 19c, 64, and 65, was performed using NaH and the alkyl halide in dimethylformamide (DMF) (Schemes 2 and 3). The yields and characterization of the xanthine derivatives prepared are listed in Table 2.

Scheme 1.

Synthesis of Xanthine Derivatives 14b–d^a

^a Reagents: (a) 10% NaOH, EtOH, R–Br; (b) NaNO₂, 6 N HCl, AcOH; (c) Na₂S₂O₄; (d) HC(OMe)₃.

Scheme 2.

Synthesis of Xanthine Derivatives 19b,c^a

^a Reagents: (a) NaH, Br(CH₂)_nCl, DMF.

Scheme 3.

Synthesis of Xanthine Derivatives 64 and 65^a

^a Reagents: (a) NaH, Cl₂(CH₂)₂, DMF.

Table 2.

Synthetic Yields and Characterization of Xanthine Derivatives

compd	formula	yield (%)	analysis
14b	C13H12N4O2	44	C, H, N
14c	C14H14N4O2·0.2H2O	32	C, H, N
14d	C15H16N4O2	60	C, H, N
19b	C10H13ClN4O2	88	C, H, N
19c	C11H15ClN4O2	95	C, H, N
64	C17H18C1N4O2	66	HRMSa
65	C19H24ClN4O2	32	HRMSa

^aHigh-resolution mass in positive ion FAB mode determined to be within acceptable limits. Compound 64: calcd, 345.1118; found, 345.1110. Compound 65: calcd, 375.1588; found, 375.1573. HPLC demonstrated >95% purity, retention times (mobile phase, min): 64, (A) 22.69; (B) 23.63; 65, (A) 23.58; (B) 30.18.

Biology.

Binding experiments were carried out using ¹²⁵I-ABOPX¹⁵ for binding to recombinant human or rat A_2B receptors overexpressed in HEK-293 cells (Tables 1 and 3). The xanthines examined included those substituted at the 1-position alone (3–8), the 3-position alone (2b), the 1- and 3-positions (2a and 9–17), the 1-, 3-, and 7-positions (18–36), the 1- and 8-positions (37), the 1-, 3-, and 8-positions (38–62), and the 1-, 3-, 7-, and 8-positions (63–65). At the 8-position, cycloalkyl or aryl substituents were included. For selected derivatives, binding to membranes expressing rat A₁ and A_2A adenosine receptors as an indication of rat A_2B selectivity was also carried out (Table 3).

Table 3.

Affinities of Xanthine Derivatives in Radioligand Binding Assays at Rat A₁ and A_2A Receptors and Selected Compounds at Other Subtypes^a–d

	Ki (nM)			Ki ratio
compd	rA1a	rA2Ab	rA2Bc	rA1/rA2B	rA2A/rA2B
1	16.8 ± 3.6f	612 ± 287f	12.8 ± 1.1	1.3	48
2a	6920 ± 160f	6700 ± 320f	15 100 ± 2700	0.46	0.44
3	3310 ± 580	< 10% (10−5)	1880 ± 190	1.8	>10
4	12 700 ± 5600	<10% (10−5)
5e	24 300 ± 4300	<10% (10−5)
6	8890 ± 1090	<10% (10−5)
7	37 600 ± 8900	<10% (10−5)
8	5570 ± 760	<10% (10−4)
13	24 700 ± 8100	9780 ± 220
14ae	5830 ± 250	33 600 ± 7700	2150 ± 260	2.7	16
14b	1660 ± 620	3190 ± 1970	10 200 ± 1500	0.16	0.31
14ce	27 600 ± 3900	<10% (10−4)	5530 ± 610	5.0	>20
14d	24 400 ± 4800	<10% (10−4)	15 300 ± 2000	1.6	>10
16	2170 ± 940	<10% (10−4)
19a	15 200 ± 1100	4230 ± 1580	5390 ± 780	2.8	0.78
19b	23 600 ± 3900	<10% (10−4)	14 200 ± 100	1.7	>10
19c	17 700 ± 3700	33 300 ± 4300	18 700 ± 500	0.95	1.8
37	7.68 ± 1.46
38	1.0	500	186 ± 5	0.005	3
51e	302 ± 49	1920 ± 400	174 ± 35	1.7	11
54	39.0 ± 5.2	<10% (10−5)
55	220 ± 11	107 ± 10
56	7.46 ± 0.58	17 300 ± 6500
57	97.7 ± 2.7	5500 ± 760
58	30.3 ± 3.5	1340 ± 680
59	39.2 ± 10.8	<10% (10−5)
62	757 ± 109	2720 ± 470
63	< 10% (10−5)	23 500 ± 9800
64e	35 ± 4% (10−5)	3570 ± 1360	1700 ± 90	>6	2.1
65e	1660 ± 130	16 700 ± 4300	3360 ± 530	0.49	5.0

^aDisplacement of specific [³H]R-PIA binding in rat brain membranes in HEK-293 cells, expressed as K_i ± SEM or percent displacement at the indicated concentration (n = 3–5).

^bDisplacement of specific [³H]CGS 21680 binding in rat striatal membranes, expressed as K_i ± SEM or percent displacement at the indicated concentration (n = 3–5).

^cRat A_2B receptor expressed in HEK cells, using [¹²⁵I]I-ABOPX as radioligand.

^dHuman A₃ receptor, K_i ± SEM (μM), using [¹²⁵I]I-AB-MECA as radioligand: 3, 2.37 ± 0.08; 4, 3.08 ± 0.43; 5, 16.1 ± 2.9; 6, 4.61 ± 1.26; 7, 7.51 ± 1.64; 8, 0.639 ± 0.158; 14a, 10.9; 14b, 4.44 ± 0.89; 14c, 6.88 ± 0.47; 14d, 8.65 ± 2.73.

^e5, MRS 1973; 7, MRS 1975; 14a, MRS 1980; 14c, MRS 3005; 51, MRS 1916; 64, MRS 3000; 65, MRS 1995.

^fValues from refs 10 and 14.

Enprofylline (3-propylxanthine 2b) was previously identified as being roughly 1 order of magnitude selective for the human A_2B vs human A₁ and A_2A receptors.¹³ Surprisingly, the A_2B affinity of the enprofylline isomer (1-propylxanthine 3) was 13-fold greater than enprofylline itself. Another monosubstitution was examined at N-1: propargyl, allyl, and n-butyl analogues 4–6 displayed similar affinity with K_i values (equilibrium inhibition constant) of approximately 0.5 μM, and even the more bulky substitutents, 2-phenylethyl 7 and cyclopentyl 8, were relatively potent with K_i values of 0.4 and 1.0 μM, respectively.

Symmetrical 1,3-substitution of xanthine with methyl 2a, ethyl 9, n-Pr 10, or allyl 11 resulted in K_i values of roughly 1 μM at A_2B receptors, while the 1,3-di-n-hexyl derivative 12 was less potent. Among asymmetrically substituted xanthines, the 1-ethyl-3-methyl derivative 13 was equipotent to the 1,3-diethyl analogue 9 and elongation at the 1-position, as in 16, did not have a significant effect. The 1-propargyl-3-methyl derivative 14a was equipotent to the 1-propargyl derivative 4. Other 1-position substitutions were included (14b–d) and studied in binding (Table 3). Branched alkyl chains, in 15 and 17, reduced affinity of A_2B binding. A comparison of 6 and 16 indicated that addition of a 3-ethyl group reduced affinity 4-fold.

Most N-7 substituents did not enhance affinity of alkylxanthines over hydrogen at this position. For example, theophylline 2a and caffeine 18 were equipotent with K_i values of approximately 10 μM for the human A_2B receptor. Most other 7-position substitutions of theophylline examined, i.e., 20–27, including charged and heteroatom substitution, produced K_i values of 10–20 μM. The notable exception to this pattern was 2-chloroethyltheophylline 19a (K_i value of 1.4 μM), which displayed a 7.5-fold greater affinity than caffeine. Homologues of 19a were also prepared and studied in binding (Table 3). As illustrated in Figure 1, the potency order of 7-substituted compounds was 2-Cl-ethyl 19a > 3-Cl-propyl 19b > 4-Cl-butyl 19c. A 7-methyl group slightly reduced the affinity of 1,3-diallylxanthine (28 vs 11). Among caffeine analogues in which the 1-substituent was modified, including charged and heteroatom substitution in 29-35, some variation in affinity was observed (K_i values of 4–21 μM), with the 1-propargyl analogue 29 as the most potent. The relatively high affinity of pentoxifylline 35 (K_i) 5.2 μM) is notable since this compound is used therapeutically to treat intermittent claudication. An isomer 36 was less potent than 29.

Figure 1.

Competition by 7-substituted xanthines for ¹²⁵I-ABOPX binding to recombinant human A_2B receptors. The indicated antagonists (see Table 1) were incubated to equilibrium (2 h) in 100 μL with radioligand (0.7 nM ¹²⁵I-ABOPX) and human HEK-A_2B cell membranes (25 μg of membrane protein). Each point is the mean ± SEM of triplicate determinations. Where not shown, the SEM is smaller than the symbol. K_i values derived from triplicate experiments are summarized in Table 1.

For the 8-cyclopentyl analogues 37 and 38, the absence of a 3-alkyl group slightly increased affinity. An 8-phenyl substituent, well-explored as a means of enhancing affinity at A₁ and A_2A receptors,^19–21 generally enhanced affinity at the human A_2B receptor by roughly 20-fold. 8-Phenyltheophylline 39 (K_i value of 415 nM) displayed a 22-fold greater human A_2B receptor affinity than theophylline 2a. Substitution of the 8-phenyl ring indicated that an electron-rich ring was preferred in receptor binding. The rank order of potency of substitution at the 8-aryl ring in theophylline analogues 39–47 was 4-hydroxy > 4-dimethylamino > 4-methoxy > 4-sulfo > 3-carboxy and 4-nitro > 2-carboxy and 2,4-dinitro.

Homologation of 1,3-dialkyl groups in 8-phenylxanthines enhanced affinity, by 22-fold in the case of 1,3-dipropyl analogue 49 in comparison to 39. As with 8-H xanthines, a 1-allyl substitutent greatly increased affinity at the A_2B receptor in the 8-aryl series. Thus, 1-allyl-3-methyl-8-phenylxanthine 51 displayed a K_i value of 37 nM. Branched substitution, in 52 as before, greatly decreased affinity. Phenyl ring-substituted analogues of 49 were also tested: The electron-withdrawing 2,6-difluoro and 4-chloro groups on the 8-aromatic ring, in 53 and 60, reduced affinity. Various di- and trisubstituted phenyl rings, containing electron-donating groups, displayed considerable affinity but not exceeding that of 49. The functionalized congener XCC (xanthine carboxylic congener, 8-[4-[(carboxymethyl)oxy]phenyl]-1,3-dipropylxanthine, 61), which served as the basis of A_2B receptor-selective amide derivatives such as 1,¹⁰ was nearly as potent as 49.

The 7-(2-chloroethyl) group of 65 decreased the affinity of compound 49 by 36-fold to a K_i value of 690 nM. However, in comparison to the 7-methyl modification of 8-phenylxanthines, this group increased affinity. Compound 64, the 7-(2-chloroethyl) derivative of the 8-phenylxanthine derivative 63, displayed a 37-fold greater affinity at human A_2B receptors.

Selected compounds were tested in binding experiments at rat A_2B receptors (Table 3). A close correlation between relative potencies at human and rat A_2B receptors was found (Figure 2). Affinities were usually higher by 2–9-fold at the human vs the rat A_2B receptor, with more pronounced differences observed for the more potent compounds. Compound 14c, a 1-phenylethyl derivative, showed slight A_2B receptor selectivity in the rat. The reference compound 1 was still highly potent at the rat A_2B receptor, with a K_i value of 12.8 nM, although not selective for A_2B vs A₁ receptors in the rat. Several other compounds were selective for A_2B vs A_2A but not A₁ receptors in the rat: 3, 14a, 14d, 19b, and 51. Selected compounds were also tested at human A₃ receptors (Table 3, footnote e), and most were found to be approximately an order of magnitude more selective for human A_2B vs A₃ receptors, except compound 8, which was equipotent at the two subtypes.

Figure 2.

Correlation in the affinities of various xanthines for binding to human and rat A_2B adenosine receptors. The pK_i values of compounds that were evaluated for binding to rat and human receptors (Table 2) are plotted and fit by linear regression.

Discussion

Theophylline 2a is widely used as an antiasthmatic drug, although its mechanism of action is uncertain. The related xanthine, enprofylline^13,14 2b, which is also therapeutically efficacious in the treatment of asthma, was earlier thought to act through a nonadenosine receptor-mediated mechanism due to its low affinity at A₁ and A_2A receptors. However, the discovery that enprofylline has greater than anticipated affinity and slight selectivity at the A_2B subtype¹³ supports the hypothesis that A_2B receptor antagonism may contribute to antiasthmatic activity of xanthines.^12,26,27 This hypothesis was strengthened by functional effects of A_2B receptor activation observed in mast cells of dog and human.^8,12,28 Thus, potent and/or selective A_2B receptor antagonists may provide new therapeutic agents.

The SAR of xanthines in binding to A_2B adenosine receptors was investigated in the present study. Substitution at the 1-position by a group larger than at the 3-position (H or Me) favored affinity at the A_2B receptor. The affinity of xanthines in binding to the human A_2B receptor was enhanced by substitution with propyl, butyl, allyl, propargyl, or 2-phenylethyl groups at N-1, while hydrogen or methyl at N-3 favored affinity at this subtype.

A 7-(2-chloroethyl) substituent strikingly favored A_2B receptor selectivity. Most N-7 substituents did not enhance affinity over hydrogen, except for 2-chloroethyl, which enhanced the human A_2B receptor affinity of theophylline by 6.5-fold. Thus, 7-(2-chloroethyl)theophylline displayed a K_i value of 800 nM. Previously, 7-(2-chloroethyl) theophylline was noted to be slightly selective for adenosine A_2A vs A₁ receptors.²⁹ At the A_2B receptor, the effects of the 7-(2-chloroethyl) group were compatible with the enhancement produced by 8-aryl substitution. For example, the affinity of an 8-aryl-7-chloroethyl analogue 64 was intermediate between those of the corresponding 7-H and 7-methyl analogues. Compound 64 was roughly 35-fold more potent at the human A_2B receptor than the 7-methyl analogue 63 and 8-fold less potent than the 7-H analogue, 51.

An 8-phenyl substituent, well-explored as a means of enhancing affinity at A₁ and A_2A receptors,^19–24 enhanced affinity at A_2B receptors by roughly 20-fold. The detail in the present study concerning 8-aryl ring substitution and compatibility with other xanthine substituents greatly extended previous findings.^10,14 As for simpler xanthines, even in the 8-phenyl series, a larger alkyl group at the 1-position than at the 3-position favored affinity at the A_2B receptor. This was illustrated by 1-allyl-3-methyl-8-phenylxanthine, 51, which displayed a K_i value of 37 nM at A_2B receptors. Substitution of the 8-phenyl ring indicated that an electron-rich ring (e.g., 4-hydroxy substitution, 41) was preferred in receptor binding. Consistent with this finding, nitro and other electron-withdrawing substituents greatly reduced affinity.

The initial screen of A_2B receptor affinity in the present study utilized the human subtype, and selected compounds were also evaluated at rat A_2B receptors. We found uniformly higher affinity of substituted xanthines for human than for rat A_2B adenosine receptors, but these differences were consistently less than 10-fold. Compounds having various degrees of selectivity for rat A_2B receptors were 3–7 (1-substitituted xanthines), 14a–d (1-substitituted theophyllines), 51 (1-allyl-3-methyl-8-phenylxanthine), and 64 (1-allyl-7-chloroethyl-3-methyl-8-phenylxanthine). The 1-phenylethyl group appeared to enhance A_2B receptor selectivity in the rat.

Selectivity for the human A_2B receptor has been easier to achieve than selectivity for the rat homologue, mainly due to the large species differences at A₁ and A_2A receptors.¹⁰ To further define the selectivity of A_2B receptor antagonists, additional characterization at receptor homologues of various species will be needed.

In conclusion, we have identified new leads for the design of xanthine derivatives as selective antagonists of the A_2B receptor, including asymmetric substitution at the 1- and 3-positions, 7-haloalkyl groups, and 8-aryl ring substitution patterns. Furthermore, the affinities at rat and human A_2B receptors have been compared. Optimal combinations of substituents on the 1-, 3-, 7-, and 8-positions of xanthines may lead to more potent and selective antagonists. Such selective compounds will aid in the elucidation of the physiological role of this receptor and possibly lead to therapeutically useful agents for treating asthma, diabetes, and other diseases.^6,10,30,31

Experimental Section

Synthetic Methods.

¹H nuclear magnetic resonance (NMR) spectra were obtained with a Varian Gemini-300 spectrometer using CDCl₃ or DMSO-d₆ as solvent. The chemical shifts are expressed as parts per million (ppm) downfield from tetramethylsilane or as relative ppm from CDCl₃ (7.27 ppm) and dimethyl sulfoxide (DMSO) (2.5 ppm). FAB (fast atom bombardment) mass spectrometry was performed with a JEOL SX102 spectrometer using 6-kV Xe atoms. CI-NH₃ (chemical ionization) mass spectra were carried out with a Finnigan 4600 mass spectrometer. Elemental analysis (±0.4% acceptable) was performed by Atlantic Microlab Inc. (Norcross, GA). All melting points were determined with a Unimelt capillary melting point apparatus (Arthur H. Thomas Co., PA) and were uncorrected. All xanthine derivatives were homogeneous as judged using thin-layer chromatography (TLC) (MK6F silica, 0.25 mm, glass backed; Whatman, Clifton, NJ). Where needed, determinations of purity were performed with a Hewlett-Packard 1090 highperformance liquid chromatography (HPLC) system using an SMT OD-5–60 C18 analytical column (250 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 triethylammonium acetate buffer:CH₃CN in ratios of 80:20 to 20:80 in 30 min with a flow rate of 1 mL/min. The other (B) was H₂O:CH₃OH in ratios of 60:40 to 10:90 in 30 min, 10:90 after 30 min, with a flow rate of 1 mL/min. Peaks were detected by UV absorption (250 nm) using a diode array detector.

6-Amino-1-methyl-3-benzyl-1H-pyrimidine-2,4-dione (66b).

To a suspension of 6-amino-1-methyluracil (706 mg, 5 mmol) in methanol (15 mL) was added 10% aqueous NaOH solution (2 mL, 5 mmol) and benzyl bromide (0.65 mL, 5.5 mmol). The reaction mixture was refluxed for 3 h and quenched by the addition of saturated NH₄Cl solution. The solvent was evaporated under reduced pressure, and the residue was diluted with chloroform (50 mL). The organic phase was washed with water and brine and dried over sodium sulfate, filtered, and evaporated under reduced pressure. The residue was purified by flash column chromatography (silica, chloroform:methanol = 10:1) to give 97 mg of 66b (33%, based on recovered starting material) as a white foaming solid. ¹H NMR (CDCl₃): δ 3.33 (s, 3H), 5.00 (s, 1H), 5.08 (s, 2H), 7.23–7.42 (m, 5H). MS (positive ion FAB): 232 [M + H]⁺.

6-Amino-1-methyl-5-nitroso-3-benzyl-1H-pyrimidine-2,4-dione (67b).

Compound 66b (53 mg, 0.23 mmol) was dissolved in a mixed solvent of water (1 mL), acetic acid (0.2 mL), and 6 N HCl (0.05 mL) and treated with NaNO₂ (24 mg, 0.35 mmol) in small portions over a period of 10 min. The mixture was filtered, and the filter cake was washed with water to afford 67b (30 mg, 50%) as a violet solid. ¹H NMR(DMSO-d₆): δ 3.25 (s, 3H), 5.10 (s, 2H), 7.23–7.40 (m, 5H). MS (positive ion FAB): 261 [M + H]⁺.

1-Benzyl-3-methylxanthine (14b).

Sodium hydrosulfite (120 mg, 0.69 mmol) was added in small portions to a suspension of 67b (30 mg, 0.12 mmol) in ethyl acetate (4 mL) and water (2 mL), until the violet color disappeared. The organic layer was separated, and the aqueous layer was extracted with ethyl acetate (2 × 10 mL). The combined organic layer was washed with brine and dried over sodium sulfate, filtered, and evaporated under reduced pressure to dryness. The crude product was suspended in trimethylorthoformate (2 mL) and refluxed for 2 h. Trimethylorthoformate was removed by a nitrogen stream, and the residue was purified by preparative TLC (chloroform:methanol = 10:1). The solid was recrystallized from ethyl acetate/hexane to furnish xanthine 14b (13 mg, 44%) as a white solid having a mp 224–226 °C. ¹H NMR (CDCl₃): δ 3.65 (s, 3H), 5.28 (s, 2H), 7.27–7.35 (m, 3H) 7.49 (d, 2H, J = 6.6 Hz), 7.73 (s, 1H), 11.93 (bs, 1H). MS (positive ion FAB): 257 [M + H]⁺.

6-Amino-1-methyl-3-phenethyl-1H-pyrimidine-2,4-di-one (66c).

Compound 66c was prepared by the procedure described for 66b, except using ethanol as solvent, instead of methanol, and obtained in 21% yield. ¹H NMR (CDCl₃): δ 2.91 (m, 2H), 3.42 (s, 3H), 4.12 (m, 2H), 4.56 (bs, 2H), 5.00 (s, 1H),7.20–7.31 (m, 5H). MS (positive ion FAB): 246 [M + H]⁺.

6-Amino-1-methyl-5-nitroso-3-phenethyl-1H-pyrimidine-2,4-dione (67c).

Compound 67c was prepared by the procedure described for 67b in 51% yield. ¹H NMR (DMSO-d₆): δ 2.88 (t, 2H, J = 7.8 Hz), 3.26 (s, 3H), 4.10 (t, 2H, J = 8.0 Hz),7.20–7.38 (m, 5H). MS (positive ion FAB): 275 [M + H]⁺.

3-Methyl-1-(2-phenylethyl)xanthine (14c).

Compound 14c was prepared by the procedure described for 14b in 32% yield; mp 240–243 °C. ¹H NMR (CDCl₃): δ 2.97 (m, 2H), 3.66 (s, 3H), 4.29 (m, 2H), 7.22–7.29 (m, 2H) 7.32 (d, 3H, J =4.4 Hz), 7.78 (s, 1H), 11.45 (bs, 1H). MS (positive ion FAB): 271 [M + H]⁺.

6-Amino-1-methyl-3-(3-phenyl-propyl)-1H-pyrimidine-2,4-dione (66d).

Compound 66d was prepared by the procedure described for 66c in 45% yield. ¹H NMR (CDCl₃): δ 1.97 (m, 2H), 2.68 (t, 2H, J = 8.0 Hz), 3.39 (s, 3H), 3.97 (t, 2H, J = 7.4 Hz), 4.45 (bs, 2H), 4.97 (s, 1H), 7.10–7.28 (m, 5H). MS (positive ion FAB): 260 [M + H]⁺.

6-Amino-1-methyl-5-nitroso-3-(3-phenyl-propyl)-1H-pyrimidine-2,4-dione (67d).

Compound 67d was prepared by the procedure described for 67b in 78% yield. ¹H NMR(DMSO-d₆): δ 1.91 (m, 2H), 2.66 (t, 2H, J = 7.7 Hz), 3.22 (s, 3H), 3.95 (t, 2H, J = 7.3 Hz), 7.11–7.30 (m, 5H). MS (CI/NH₃): 289 [M + H]⁺

3-Methyl-1-(3-phenylpropyl)xanthine (14d).

Compound 14d was prepared by the procedure described for 14b in 60% yield; mp 176–180 °C. ¹H NMR (CDCl₃): δ 2.06 (quintet, 2H, J = 7.4 Hz), 2.75 (t, 2H, J = 7.4 Hz), 3.63 (s, 3H), 4.14 (t, 2H, J = 7.4 Hz), 7.12–7.30 (m, 5H), 7.74 (s, 1H), 12.28 (bs, 1H). MS (positive ion FAB): 285 [M + H]⁺.

7-(3-Chloropropyl)theophylline (19b).

Sodium hydride (24 mg, 60% in mineral oil, 0.61 mmol) was added to a solution of theophylline (100 mg, 0.56 mmol) in DMF (2 mL), and the reaction mixture was stirred at room temperature for 10 min. 1-Bromo-3-chloropropane (0.07 mL, 0.70 mmol) was then added dropwise. The reaction mixture was stirred at room temperature for 24 h, and the solvent was removed by a nitrogen stream. The residue was purified using preparative TLC (ethyl acetate:hexane:methanol = 10:10:1) to give 19b (126 mg, 88%) as a white solid. The solid was recrystallized from ethyl acetate/n-hexane; mp 1223–124 °C. ¹H NMR (CDCl₃): δ 2.39 (m, 2H), 3.42 (s, 3H), 3.50 (t, 2H, J = 5.9 Hz), 3.61 (s, 3H), 4.49 (t, 2H, J = 6.5 Hz), 7.62 (s, 1H). MS (positive ion FAB): 257 [M + H]⁺.

7-(4-Chlorobutyl)theophylline (19c).

Compound 19c was prepared by the procedure described for 19b in 95% yield; mp 115–117 °C. ¹H NMR (CDCl₃): δ 1.81 (m, 2H), 2.07 (m, 2H), 3.42 (s, 3H), 3.57 (t, 2H, J = 6.3 Hz), 3.60 (s, 3H), 4.35 (t, 2H, J = 7.1 Hz), 7.56 (s, 1H). MS (positive ion FAB): 271 [M + H]⁺.

1-Allyl-7-chloroethyl-3-methyl-8-phenylxanthine (64).

Sodium hydride (10 mg, 60% in mineral oil, 0.25 mmol) was added dropwise to a solution of 1-allyl-3-methyl-8-phenylxan-thine 51³⁶ (6 mg, 0.02 mmol) in DMF (1 mL), and the reaction mixture was stirred at room temperature for 10 min. 1,2-Dichloroethane (0.05 mL, 0.6 mmol) was added, and the reaction mixture was stirred at 70 °C for 36 h. The solvent was removed by a nitrogen stream. The residue was purified using preparative TLC (dichloromethane:2-propanol = 100:1) to give 64 (4.8 mg, 70%) as a white solid; mp 130–133 °C. ¹H NMR (CDCl₃): ä 3.66 (s, 3H), 3.94 (t, 2H, J = 6.0 Hz) 4.66 (t, 2H, J = 6.3 Hz) 4.68 (d, 2H, J = 5.5 Hz) 5.22 (dd, 1H, J = 10.2, 1.1 Hz) 5.31 (dd, 1H, J = 17.3, 1.1 Hz) 5.96 (m, 1H) 7.54–7.56 (m, 3H) 7.64–7.68 (m, 2H). High-resolution MS (positive ion FAB) calcd for C₁₇H₁₈ClN₄O₂ [M + H]⁺, 345.1118; found, 345.1110. HPLC indicated 99% purity (retention times (min): A, 22.7; B, 23.6).

7-Chloroethyl-1,3-dipropyl-8-phenylxanthine (65).

To a solution of 1,3-dipropyl-8-phenylxanthine 49²⁵ (13 mg, 0.042 mmol) in DMF (2 mL) was added NaH (18 mg, 60% in mineral oil, 0.45 mmol), and the reaction mixture was stirred for 10 min at room temperature. 1,2-Dichloroethane (0.10 mL, 1.27 mmol) was added dropwise, and the reaction mixture was stirred at 70 °C for 48 h. The solvent was removed by a nitrogen stream. The residue was purified using preparative TLC (dichloromethane:2-propanol = 100:1) to give 65 (5.0 mg, 32%) as a white solid; mp 128–132 °C. ¹H NMR (CDCl₃): δ 1.09 (t, 3H, J = 7.4 Hz), 1.10 (t, 3H, J = 7.7 Hz), 1.82 (m, 2H), 1.95 (m, 2H), 4.06 (t, 2H, J = 6.3 Hz), 4.11 (dd, 2H, J = 7.4, 7.7 Hz), 4.23 (dd, 2H, J = 7.7, 7.4 Hz), 4.74 (t, 2H, J = 6.2 Hz), 7.64–7.66 (m, 3H), 7.74–7.77 (m, 2H). High-resolution MS (positive ion FAB) calcd for C₁₉H₂₄ClN₄O₂ [M + H]⁺, 375.1588; found, 375.1573. HPLC indicated 96% purity (retention times (min): A, 23.6; B, 30.2).

Pharmacological Methods. Cloning of A_2B Receptor.

The rat A_2B receptor was cloned by Dr. Eric Yuan-Ji Day (University of Virginia) by the following method: Sprague–Dawley rats were euthanized with an overdose of pentobarbital. The urinary bladders were removed and stored in RNALater (Qiagen). mRNA was extracted from rat bladder tissue using a Qiagen mRNA extraction kit. The rat A_2B receptor was amplified by RT-polymerase chain reaction (PCR) with the forward primer GGCCATGCAGCTAGAGACGCAGGAC and reverse primer TAGGTCACAAGCTCAGACTGA and then cloned into TOPO2.1 and confirmed by sequencing. The confirmed PCR product was then subcloned in a pDouble-Trouble vector using the restriction enzyme sites for HindIII and ECoRV.

Stable Transfection.

HEK-293 cells were grown in DMEM/F12 with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin to 50% confluence in six well dishes. Cells were transfected with 2 μg of plasmid DNA and 5 μL of lipofectamine in accordance with the manufacturer’s instructions (GibcoBRL). Transfected cells were grown for 48 h prior to selection by the addition of 1 mg/mL G418 in the medium. Resistant colonies were isolated and screened using radioligand binding with ¹²⁵I-ABOPX. Clonal lines with high specific binding were expanded and maintained in 0.5 mg/mL G418. Several clones with expression levels of approximately 20 000 fmol/mg were preserved. The average K_D for ¹²⁵I-ABOPX for the rat A_2B receptor was determined to be 40.9 nM (SEM = 4.4, n = 7).

Binding Assays.

Membranes from HEK-293 cells stably expressing the human or rat A_2B receptor were used for competition binding assays with ¹²⁵I-ABOPX (2200 Ci/mmol).^10,32 Radioligand binding experiments were performed in triplicate with 20–25 μg of membrane protein in a total volume of 0.1 mL of HE buffer (10 mM HEPES and 1 mM EDTA, pH 7.4) supplemented with 1 U/mL adenosine deaminase and 5 mM MgCl₂. Nonspecific binding was measured in the presence of 100 μM NECA. Xanthine derivatives for competition assays were diluted in HE buffer with 10% DMSO. The incubation time was 3 h at 21 °C. Competition experiments were carried out using between 0.5 and 1.0 nM ¹²⁵I-ABOPX. Membranes were filtered on Whatman GF/C filters using a Brandel cell harvester (Gaithersburg, MD) and washed three times during 15–20 s with ice-cold buffer (10 mM Tris, 1 mM MgCl₂, pH 7.4). K_i values for different compounds were derived from IC₅₀ values as described, assuming a K_D value for ¹²⁵I-ABOPX of 36 nM at the human A_2B receptor. Data from replicate experiments were tabulated as means ± SEM. Nonspecific binding, measured in the presence of 100 μM NECA (RBI-Sigma, St. Louis, MO), was 25% of total binding. All nonradioactive compounds were initially dissolved in DMSO and diluted with buffer to the final concentration, with the amount of DMSO in the final assay tubes consistently ≤0.5%.

For competition experiments, at least six 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 Prism program (GraphPAD, San Diego, CA), were converted to apparent K_i values.³³ Equilibrium binding competition experiments at rat A₁, rat A_2A, and human A₃ adenosine receptors were carried out as previously reported.^34–36