Synthesis, biological activity and molecular modelling studies of tricyclic alkylimidazo-, pyrimido- and diazepinopurinediones.

Syntheses and biological activities of imidazo-, pyrimido- and diazepino[2,1-f]purinediones containing N-alkyl substituents (with straight, branched or unsaturated chains) are described. Tricyclic derivatives were synthesized by the cyclization of 8-bromo-substituted 7-(2-bromoethyl)-, 7-(3-chloropropyl)- or 7-(4-bromobutyl)-theophylline with primary amines under various conditions. Compound 22 with an ethenyl substituent was synthesized by dehydrohalogenation of 9-(2-bromoethyl)-1,3-dimethyltetrahydropyrimido[2,1-f]purinedione. The obtained derivatives (5-35) were initially evaluated for their affinity at rat A1 and A2A adenosine receptors (AR), showing moderate affinity for both adenosine receptor subtypes. The best ligands were diazepinopurinedione 28 (K i = 0.28 µM) with fivefold A2A selectivity and the non-selective A1/A2A AR ligand pyrimidopurinedione 35 (K i A1 = 0.28 µM and K i A2A = 0.30 µM). The compounds were also evaluated for their affinity at human A1, A2A, A2B and A3 ARs. All of the obtained compounds were docked to the A2A AR X-ray structure in complex with the xanthine-based, potent adenosine receptor antagonist-XAC. The likely interactions of imidazo-, pyrimido- and diazepino[2,1-f]purinediones with the residues forming the A2A binding pocket were discussed. Furthermore, the new compounds were tested in vivo as anticonvulsants in maximal electroshock, subcutaneous pentylenetetrazole (ScMet) and TOX tests in mice (i.p.). Pyrimidopurinediones showed anticonvulsant activity mainly in the ScMet test. The best derivative was compound 11, showing 100 % protection at a dose of 100 mg/kg without symptoms of neurotoxicity. Compounds 6, 7, 8 and 14 with short substituents showed neurotoxicity and caused death. In rat tests (p.o.), 9 was characterized by a high protection index (>13.3). AR affinity did not apparently correlate with the antiepileptic potency of the compounds.

Benzothiazinones: a novel class of adenosine receptor antagonists structurally unrelated to xanthine and adenine derivatives.

2-(Acyl)amino-4H-3,1-benzothiazin-4-ones and related thienothiazinones were identified as structurally novel antagonists at adenosine receptors (ARs). 6-Methyl-2-benzoylamino-4H-3,1-benzothiazin-4-one (10d) was found to be a balanced AR antagonist with affinity for all human (h) subtypes (K(i) hA(1) 65.6 nM; hA(2A) 120 nM; hA(2B) 360 nM; hA(3) 30.4 nM), while in rat (r), 10d was a highly potent A(1)-selective antagonist (rA(1) 7.7 nM; rA(2A) 546 nM; rA(2B) 679 nM, rA(3) >10000 nM). 2-(4-Methylbenzoylamino)-4H-3,1-benzothiazin-4-one (10g) was found to be a potent antagonist at human A(2A) (68.8 nM) and A(3) ARs (23.0 nM) with high selectivity versus the other human AR subtypes. In contrast to A(1) and A(3) ARs, A(2A) and A(2B) ARs tolerated bulky 2-acyl substituents. tert-Butyl (4-oxo-4H-3,1-benzothiazin-2-ylcarbamoyl)benzylcarbamate (15g, K(i) hA(2B) 186 nM; hA(2A) 603 nM) and 4-(4-benzylpiperazine-1-carbonyl)-N-(4-oxo-4H-3,1-benzothiazin-2-yl)benzamide (15k, hA(2A) 69.5 nM; hA(2B) 178 nM) were highly selective versus the other AR subtypes. 2-Acylamino-3,1-benzothiazin-4-ones represent novel scaffolds suitable for the development of potent and selective AR antagonists for each of the four receptor subtypes.

Synthesis and biological activity of tricyclic cycloalkylimidazo-, pyrimido- and diazepinopurinediones.

Syntheses and physicochemical properties of N-cycloalkyl-substituted imidazo-, pyrimido- and 1,3-diazepino[2,1-f]purinediones are described. These derivatives were synthesized by cyclization of 7-halogenoalkyl-8-bromo-1,3-dimethylxanthine derivatives with aminocycloalkanes. The obtained compounds (1-33) were evaluated for their affinity to rat adenosine A(1) and A(2A) receptors. Selected compounds were additionally investigated for affinity to the human A(1), A(2A), A(2B) and A(3) receptor subtypes. The results of the radioligand binding assays at adenosine A(1) and A(2A) receptors showed that most of the compounds exhibited adenosine A(2A) receptor affinity at micromolar or submicromolar concentrations; an annelated pyrimidine ring was beneficial for A(2A) affinity. The most potent A(2A) ligands of the present series were compounds 6 (K(i) 0.33 µM rat A(2A), 0.31 µM human A(2A)), 8 (K(i) 0.98 µM rat A(2A), 0.42 µM human A(2A)) and 15 (K(i) 0.24 µM rat A(2A), 0.61 µM human A(2A)) with the latter one showing high A(2A) selectivity. In NaCl shift assay, 15 was shown to be an antagonist at A(2A) receptors. This result was confirmed for the best compounds 6, 8, 15 in cAMP accumulation studies. A 3D-QSAR equation with a good predicting power (q(2) = 0.88) for A(2A) AR affinity was obtained. The compounds were evaluated in vivo as anticonvulsants in MES and ScMet tests and examined for neurotoxicity in mice (i.p.). Most of them showed anticonvulsant activity in chemically induced seizures; among them the diazepinopurinediones were the best (e.g. 31) showing protection in both tests on short time symptoms, without signs of neurotoxicity. Five compounds, 8, 17, 20, 29, and 31, exhibited anticonvulsant activity after peroral application in rats. Structure-activity relationships are discussed including the analysis of lipophilic and spatial properties. The new compounds, which contain a basic nitrogen atom and can therefore be protonated, may be good starting points for obtaining A(2A) antagonists with good water-solubility.