A3 adenosine receptor agonists containing dopamine moieties for enhanced interspecies affinity.

Following our study of 4'-truncated (N)-methanocarba-adenosine derivatives that displayed unusually high mouse (m) A3AR affinity, we incorporated dopamine-related N6 substituents in the full agonist 5'-methylamide series. N6-(2-(4-Hydroxy-3-methoxy-phenyl)ethyl) derivative MRS7618 11 displayed Ki (nM) 0.563 at hA3AR (∼20,000-fold selective) and 1.54 at mA3AR. 2-Alkyl ethers maintained A3 affinity, but with less selectivity than 2-alkynes. Parallel functional assays of G protein-dependent and ß-arrestin 2 (ßarr2)-dependent pathways indicate these are full agonists but not biased. Through use of computational modeling, we hypothesized that phenyl OH/OMe groups interact with polar residues, particularly Gln261, on the mA3AR extracellular loops as the basis for the affinity enhancement. Although the pharmacokinetics indicated facile clearance of parent O-methyl catechol nucleosides 21 and 31, prolonged mA3AR activation in vivo was observed in a hypothermia model, suggested potential formation of active metabolites through demethylation. Selected analogues induced mouse hypothermia following i.p. injection, indicative of peripheral A3AR agonism in vivo.

Truncated (N)-Methanocarba Nucleosides as Partial Agonists at Mouse and Human A3 Adenosine Receptors: Affinity Enhancement by N6-(2-Phenylethyl) Substitution.

Dopamine-derived N6-substituents, compared to N6-(2-phenylethyl), in truncated (N)-methanocarba (bicyclo[3.1.0]hexyl) adenosines favored high A3 adenosine receptor (AR) affinity/selectivity, e.g., C2-phenylethynyl analogue 15 (MRS7591, Ki = 10.9/17.8 nM, at human/mouse A3AR). 15 was a partial agonist in vitro (hA3AR, cAMP inhibition, 31% Emax; mA3AR, [35S]GTP-γ-S binding, 16% Emax) and in vivo and also antagonized hA3AR in vitro. Distal H-bonding substitutions of the N6-(2-phenylethyl) moiety particularly enhanced mA3AR affinity by polar interactions with the extracellular loops, predicted using docking and molecular dynamics simulation with newly constructed mA3AR and hA3AR homology models. These hybrid models were based on an inactive antagonist-bound hA1AR structure for the upper part of TM2 and an agonist-bound hA2AAR structure for the remaining TM portions. These species-independent A3AR-selective nucleosides are low efficacy partial agonists and novel, nuanced modulators of the A3AR, a drug target of growing interest.

Design and in Vivo Characterization of A1 Adenosine Receptor Agonists in the Native Ribose and Conformationally Constrained (N)-Methanocarba Series.

(N)-Methanocarba ([3.1.0]bicyclohexyl) adenosines and corresponding ribosides were synthesized to identify novel A1 adenosine receptor (A1AR) agonists for CNS or peripheral applications. Human and mouse AR binding was determined to assess the constrained ring system's A1AR compatibility. N6-Dicyclobutylmethyl ribose agonist (9, MRS7469, >2000-fold selective for A1AR) and known truncated N6-dicyclopropylmethyl methanocarba 7 (MRS5474) were drug-like. The pure diastereoisomer of known riboside 4 displayed high hA1AR selectivity. Methanocarba modification reduced A1AR selectivity of N6-dicyclopropylmethyl and endo-norbornyladenosines but increased ribavirin selectivity. Most analogues tested (ip) were inactive or weak in inducing mouse hypothermia, despite mA1AR full agonism and variable mA3AR efficacy, but strong hypothermia by 9 depended on A1AR, which reflects CNS activity (determined using A1AR or A3AR null mice). Conserved hA1AR interactions were preserved in modeling of 9 and methanocarba equivalent 24 (∼400-fold A1AR-selective). Thus, we identified, and characterized in vivo, ribose and methanocarba nucleosides, including with A1AR-enhancing N6-dicyclobutylmethyl-adenine and 1,2,4-triazole-3-carboxamide (40, MRS7451) nucleobases.

Structure activity relationship of 2-arylalkynyl-adenine derivatives as human A3 adenosine receptor antagonists.

Recognition of nucleosides at adenosine receptors (ARs) is supported by multiple X-ray structures, but the structure of an adenine complex is unknown. We examined the selectivity of predicted A1AR and A3AR adenine antagonists that incorporated known agonist affinity-enhancing N 6 and C2 substituents. Adenines with A1AR-favoring N 6-alkyl, cycloalkyl and arylalkyl substitutions combined with an A3AR-favoring 2-((5-chlorothiophen-2-yl)ethynyl) group were human (h) A3AR-selective, e.g. MRS7497 17 (∼1000-fold over A1AR). In addition, binding selectivity over hA2AAR and hA2BAR and functional A3AR antagonism were demonstrated. 17 was subjected to computational docking and molecular dynamics simulation in a hA3AR homology model to predict interactions. The SAR of nucleoside AR agonists was not recapitulated in adenine AR antagonists, and modeling suggested an alternative, inverted binding mode with the key N2506.55 H-bonding to the adenine N 3 and N 9, instead of N 6 and N 7 as in adenosine agonists.