3-Unsubstituted xanthine derivatives bearing a cyclopentyl or a phenyl residue in the 8-position were synthesized and developed as A2B adenosine receptor antagonists. Compounds bearing polar substituents were prepared to obtain water-soluble derivatives. 1-Alkyl-8-phenylxanthine derivatives were found to exhibit high affinity for A2B adenosine receptors (ARs). 1,8-disubstituted xanthine derivatives were equipotent to or more potent than 1,3,8-trisubstituted xanthines at A2B ARs, but generally less potent at A1 and A2A, and much less potent at A3 ARs. Thus, the new compounds exhibited increased A2B selectivity versus all other AR subtypes. 9-Deazaxanthines (pyrrolo[2,3-d]pyrimidindiones) appeared to be less potent at A2B ARs than the corresponding xanthine derivatives. 1-Propyl-8-p-sulfophenylxanthine (17) was the most selective compound of the present series, exhibiting a K(i) value of 53 nM at human A2B ARs and showing greater than 180-fold selectivity versus human A1 ARs. Compound 17 was also highly selective versus rat A1 ARs (41-fold) and versus the other human AR subtypes (A2A > 400-fold and A3 > 180-fold). The compound is highly water-soluble due to its sulfonate function. 1-Butyl-8-p-carboxyphenylxanthine (10), another polar analogue bearing a carboxylate function, exhibited a K(i) value of 24 nM for A2B ARs, 49-fold selectivity versus human and 20-fold selectivity versus rat A1 ARs, and greater than 150-fold selectivity versus human A2A and A3 ARs. 8-[4-(2-Hydroxyethylamino)-2-oxoethoxy)phenyl]-1-propylxanthine (29) and 1-butyl-8-[4-(4-benzyl)piperazino-2-oxoethoxy)phenyl]xanthine (35) were among the most potent A2B antagonists showing K(i) values at A2B ARs of 1 nM, 57-fold (29) and 94-fold (35) selectivity versus human A1, ca. 30-fold selectivity versus rat A1, and greater than 400-fold selectivity versus human A2A and A3 ARs. The new potent, selective, water-soluble A2B antagonists may be useful research tools for investigating A2B receptor function.

Water-soluble prodrugs of potent, A(2A)-selective adenosine receptor (AR) antagonists were prepared. 8-(m-Bromostyryl)-3, 7-dimethyl-1-propargylxanthine (BS-DMPX, 11) and the analogous 8-(m-methoxystyryl)xanthine derivative (MS-DMPX, 5b) were used as starting points. It was found that polar functional groups suitable for the attachment of a prodrug moiety were tolerated on the styryl ring and even better on the 3-substituent. 8-(m-Hydroxystyryl)-DMPX (7) and 3-(3-hydroxypropyl)-8-(m-methoxystyryl)-1-propargylxanthine (5e, MSX-2) were the most potent and A(2A)-selective compounds and were selected for prodrug formation. For the preparation of 5e a new ring-closure method was applied. Treatment of 6-amino-1-(3-hydroxypropyl)-5-(m-methoxycinnamoylamino)-3-propa rgylur acil with hexamethyldisilazane at high temperature resulted in higher yields of the target xanthine than the standard ring-closure procedure using sodium hydroxide. Phosphate prodrugs were prepared by classical phosphorylation using phosphorus oxychloride and alternatively by using a phosphoramidite method. Phosphates of the aliphatic alcohol 5e could be obtained by both methods in similar yields. The phenolic compound 7, however, could be phosphorylated only by using the phosphoramidite method. The disodium salts of the phosphate prodrugs exhibited high water solubility (8-(m-methoxystyryl)-7-methyl-3-[3-O-phosphatylpropyl]-1- propargylxan thine disodium salt, 9b: 17 mM, 9 mg/mL). Prodrug 9b was found to be stable in aqueous solution (pH 7) but readily cleaved by phosphatases to liberate 5e (MSX-2). Compound 5e showed high affinity for rat A(2A) AR (K(i) = 8 nM), human recombinant A(2A) AR (K(i) = 5 nM), and human native A(2A) AR (K(i) = 15 nM) and was highly selective versus rat A(1) AR (110-fold), human recombinant A(2A) AR (500-fold), human A(2B) AR (>2000-fold), and human A(3) AR (>2000-fold).