Synthesis of 2-Chloro-N⁶-Substituted-4′-thioadenosine-5′-N,N-dialkyluronamides as Potent and Selective A₃ Adenosine Receptor Antagonists

Abstract

The highly selective A₃ receptor agonist, 4′-thio-Cl-IB-MECA was successfully converted into selective A₃ receptor antagonists by appending a second N-alkyl group on the 5′-uronamide position. This result indicates that the hydrogen bonding ability of the 5′-uronamide is essential for the conformational change required for the receptor activation. Among compounds tested, a N⁶-(3-bromobenzyl) derivative with 5′-dimethyluronamide exhibited the highest binding affinity (K_i = 9.32 nM) at the human A₃ AR with very-low binding affinities to other AR subtypes.

INTRODUCTION

Adenosine regulates cell signaling through binding to four subtypes (A₁, A_2A, A_2B, and A₃) of adenosine receptors (ARs).¹ Among these, A₃ AR has been the most recently identified and good therapeutic target for the treatment of cancer, ischemia, asthma, and inflammation.² Modification on N⁶ position and/or 4′-hydroxymethyl group of adenosine afforded potent and selective A₃ AR agonists³ because of the structural similarity to adenosine, but not AR antagonists. Most of A₃ AR antagonists belong to nonpurine heterocyclic compounds. However, because these derivatives showed poor binding affinity at the rat A₃ AR, these are not suitable for establishing efficacy in animal model.⁴

Compounds with a nucleoside skeleton were found to be species-independent A₃ AR antagonists,⁵ but with very low binding affinity to the A₃ AR. Thus, it is of great interest to synthesize nucleoside analogues with high binding affinity to the A₃ AR, showing equal binding affinity at the rat as well as human A₃ ARs. Recently, we discovered the 4′-thio analogue, thio-Cl-IB-MECA⁶ as potent and selective human A₃ AR agonist (K_i for human A₃ AR = 0.38 nM) (Fig. 1). Since it is believed that the hydrogen of the methylamide is essential for the hydrogen bonding and thus for the induced-fit required for the receptor activation, it was hypothesized that removal of the hydrogen bonding ability results in the conversion of the agonistic activity into the antagonistic activity. Thus, 5′-N,N-dialkyluronamide derivatives of thio-Cl-IB-MECA were designed and synthesized as human A₃ AR antagonists.⁷

Fig. 1.

Rationale for the design of A₃ AR antagonists

RESULTS AND DISCUSSION

Synthesis of the desired nucleosides started from 2,6-dichloropurine derivative 3⁶, which was prepared from D-gulonic γ-lactone (Scheme 1).

Scheme 1.

Rationale for the design of A₃ AR antagonists

Treatment of 3 with various alkyl or arylamines afforded N⁶-substituted derivative 4. The isopropylidene group of 4 was changed to the TBS group because its removal at the final stage resulted in deglycosylation, and then 5′-benzoyl group was deprotected to give 5. Oxidation of the primary alcohol of 5 with PDC in DMF yielded the acid 6. Conversion of the acid 6 into the dialkylamide derivatives 7 was achieved by treating with various amines in the presence of EDC and HOBt.⁷

Radioligand binding assays were carried out in Chinese hamster ovary (CHO) cells stably expressing a human AR subtype. 5′-N,N-Dimethylamide derivatives showed better binding affinity than larger 5′-N,N-dialkyl or 5′-N,N-cycloalkylamide derivatives, irrespective of N⁶-substituents. With dimethylamide at the 5′-position (R₂ = R₃ = Me) fixed, the binding affinity of various N⁶-substituted analogues was examined. The N⁶-(3-halobenzyl)amino series generally exhibited higher binding affinity at the human A₃AR than the N⁶-dialkylamino or N⁶-cycloalkylamino series. Within the N⁶-(3-halobenzyl)amino series (R₁ = 3-halobenzyl, R₂ = R₃ = Me), the binding affinity at the human A₃AR was in the following order: 3-bromobenzyl (K_i = 9.32 nM) > 3-iodobenzyl (K_i = 15.5 nM) > 3-chlorobenzyl (K_i = 21.3 nM) > 3-fluorobenzyl (K_i = 121 nM). All synthesized compounds exhibited almost no binding affinities to other AR subtypes.

The functional assay indicated that compound 7 (R₁ = 3-iodobenzyl, R₂ = R₃ = Me) indicated that this is a pure A₃AR antagonist. The binding affinity of compound 7 was also measured at the rat A₃AR to determine species differences, in which this compound showed moderate affinity (K_i = 321 ±74 nM) at the rat A₃AR.⁷

CONCLUSION

Based on the hypothesis that the hydrogen bond-donating ability of the 5′-uronamide is responsible for the conformational change needed for receptor activation, we carried out the structure-activity relationship of 2-chloro-N⁶-substituted-4′-thioadenosine-5′-N,N-dialkyluronamides as pure A₃ AR antagonists. From this study, it was found that the hydrogen bond-donating ability of the 5′-uronamide was essential for the pure A₃AR antagonism. All synthesized A₃AR antagonists could be evaluated in models of a number of disorders related to the A₃AR, such as glaucoma, inflammation, and asthma.