2-Aryladenine Derivatives as a Potent Scaffold for Adenosine Receptor Antagonists: The 6-Morpholino Derivatives

A set of 2-aryl-9-H or methyl-6-morpholinopurine derivatives were synthesized and assayed through radioligand binding tests at human A1, A2A, A2B, and A3 adenosine receptor subtypes. Eleven purines showed potent antagonism at A1, A3, dual A1/A2A, A1/A2B, or A1/A3 adenosine receptors. Additionally, three compounds showed high affinity without selectivity for any specific adenosine receptor. The structure-activity relationships were made for this group of new compounds. The 9-methylpurine derivatives were generally less potent but more selective, and the 9H-purine derivatives were more potent but less selective. These compounds can be an important source of new biochemical tools and/or pharmacological drugs.


Introduction
Adenosine, a purine nucleoside composed of an adenine linked to a ribose via a β-N9-glycosidic bond, modulates many physiological conditions related to neurological, immunologic, and cardiovascular systems [1,2].Adenosine receptors, namely A 1 , A 2A , A 2B , and A 3 , belong to the G protein-coupled receptor (GPCR) superfamily and are widely recognized as attractive targets for the design and development of new therapeutic agents against different clinical disorders [3][4][5].The activation of the A 1 and A 3 receptors inhibits adenylyl cyclase activity through the action of G i/o proteins and, as a result, intracellular cyclic AMP (cAMP) decreases, while the A 2A and A 2B receptors increase cAMP production through the action of G s proteins [6].So, it is of therapeutic importance to synthesize new drugs active on adenosine receptors A 1 , A 3 , and, additionally, dual A 1 /A 3 to study the synergistic effect between these two receptor subtypes as they preferentially couple to the same G proteins [2].Indeed, there are many patents for adenosine receptor ligands, and some of these ligands are already in clinical trials or waiting for FDA approval [4,5,7].For example, A 1 ligands are useful for glaucoma, heart failure, angina, seizures, ischemia, depression, obesity, asthma, renal protection, edema associated with congestive heart failure, atrial arrhythmias, type II diabetes, neuropathic pain, neuroprotection, atrial fibrillation, failure, atrial arrhythmias, type II diabetes, neuropathic pain, neuroprotection, a brillation, tachycardia, cardioprotection, and sleep regulation [1,[8][9][10][11][12].In addition ands are useful for liver regeneration, hepatitis, psoriasis, rheumatoid arthritis, infl tion, dry eye syndrome, fibrotic diseases, neurodegeneration, ischaemia, asthma, obstructive pulmonary disease, glaucoma, and cancer [13][14][15][16][17][18][19][20], with A3 antagonist P (Palobiofarma SL), that has reached clinical trials for glaucoma, ulcerative colitis, sinophilic esophagitis [21,22].Finally, dual A1/A3 ligands can be useful as potentia peutics for treating glaucoma, kidney failure, pulmonary diseases, and Alzheim ease [23,24].Despite the fact that several adenosine receptor ligands are already in trials, achieving selectivity is still a challenge.Additionally, a number of side effec reported for A1 antagonists in the literature, such as an increase in the frequenc zures, strokes [25,26], dizziness, nausea, transient hypertension, and transient h sion [27] in treated patients, which justifies the search for new high-affinity and s drugs.
In our previous work, we synthesized and identified a new scaffold for ad receptor antagonists based on the adenine nucleus (Figure 1) [28].Highly potent lective compounds were identified as having a piperidinyl group in C-6, a proton and several selected aryl groups in C-2 of the purine nucleus.Indeed, the structur ity relationship (SAR) study indicated that an aryl unit at C-2 is crucial for activit proton at N-9 increased potency.Furthermore, a piperidinyl group in C-6 instead N-methylpiperazinyl group led to purines with higher or similar affinity for the ad receptors A1 and A3 but with higher selectivity.As an attempt to increase the affinity and selectivity of the compounds obta our previous work and to establish the importance of group X (Figure 1), herein, thesized new 6-morpholino purine derivatives substituted with a proton or a group in N-9 and a selection of aryl substituents in the C-2 position of the purine n The results will allow us to complete the SAR study for the adenine-based scaffo tified in our previous publication.

Chemistry
The target compounds 3a-x were synthesized through the synthetic appro scribed in Scheme 1.The starting reagents 1 [29] and 2 [28,30] were obtained follow procedures described in previous work.Briefly, the commercially available diamin onitrile was treated with triethyl orthoformate, under reflux.The solid obtained acted with the convenient amine followed by treatment with a base to generate imi 1.The compounds 2 were obtained in excellent yield through reaction of 1 with th of morpholine, in acetonitrile, at room temperature.
The products 3 were obtained through reaction of compounds 2 with differe hydes 4 following a previously described methodology [28,30].The experimenta tions used in each reaction mainly depended on the aldehyde.Generally, when oles 1 were reacted with non-phenolic aldehydes 4, the reactions were performed As an attempt to increase the affinity and selectivity of the compounds obtained in our previous work and to establish the importance of group X (Figure 1), herein, we synthesized new 6-morpholino purine derivatives substituted with a proton or a methyl group in N-9 and a selection of aryl substituents in the C-2 position of the purine nucleus.The results will allow us to complete the SAR study for the adenine-based scaffold identified in our previous publication.

Chemistry
The target compounds 3a-x were synthesized through the synthetic approach described in Scheme 1.The starting reagents 1 [29] and 2 [28,30] were obtained following the procedures described in previous work.Briefly, the commercially available diaminomaleonitrile was treated with triethyl orthoformate, under reflux.The solid obtained was reacted with the convenient amine followed by treatment with a base to generate imidazoles 1.The compounds 2 were obtained in excellent yield through reaction of 1 with the excess of morpholine, in acetonitrile, at room temperature.
celerate the reactions, most of them started at moderate temperature, but, when the TLC showed the absence of imidazoles 2, the temperature was increased and the reactions were monitored by TLC until the spots assigned to the intermediates were absent.In these cases, the reactions led to black-greyish solids that required purification.The purification was achieved through filtration of a dichloromethane solution of the compounds through a silica gel column.Scheme 1. Synthetic approach to 2-aryl-adenine derivatives 3a-x.
To obtain the phenolic derivatives, the reactions were performed in an acidic medium until complete consumption of the reagents (evidenced by TLC) and then continued in a basic medium at a temperature between 40 and 80 °C.In these reactions, degradation of the reaction mixture also occurred when higher temperatures were used.The pure products were obtained following the purification procedure described above for the other derivatives.The new compounds were fully characterized by usual techniques ( 1 H and 13 C NMR spectra are presented in the supporting information).

Pharmacology
All compounds synthesized in this work have a morpholine group as a substituent at C-6, an H atom or a methyl group at N-9, and different substituted aryl groups at C-2 of the purine nucleus.The compounds (3a-x) were assessed through radioligand binding Scheme 1. Synthetic approach to 2-aryl-adenine derivatives 3a-x.
The products 3 were obtained through reaction of compounds 2 with different aldehydes 4 following a previously described methodology [28,30].The experimental conditions used in each reaction mainly depended on the aldehyde.Generally, when imidazoles 1 were reacted with non-phenolic aldehydes 4, the reactions were performed in basic medium.Considering that compounds 2 are temperature-sensitive, the reactions always started at low to moderate temperature.When the reactions were carried out under those conditions (Method A), they were slower, but the products 3 were precipitated pure from solution and isolated through simple filtration from the reaction medium.In order to accelerate the reactions, most of them started at moderate temperature, but, when the TLC showed the absence of imidazoles 2, the temperature was increased and the reactions were monitored by TLC until the spots assigned to the intermediates were absent.In these cases, the reactions led to black-greyish solids that required purification.The purification was achieved through filtration of a dichloromethane solution of the compounds through a silica gel column.
To obtain the phenolic derivatives, the reactions were performed in an acidic medium until complete consumption of the reagents (evidenced by TLC) and then continued in a basic medium at a temperature between 40 and 80 • C. In these reactions, degradation of the reaction mixture also occurred when higher temperatures were used.The pure products were obtained following the purification procedure described above for the other derivatives.The new compounds were fully characterized by usual techniques ( 1 H and 13 C NMR spectra are presented in the supporting information).

Pharmacology
All compounds synthesized in this work have a morpholine group as a substituent at C-6, an H atom or a methyl group at N-9, and different substituted aryl groups at C-2 of the purine nucleus.The compounds (3a-x) were assessed through radioligand binding assays at all human adenosine subtype receptors (A 1 , A 2A , A 2B , and A 3 ) expressed in mammalian cell lines.The percentage of inhibition (% inhib ) of radioligand binding was determined for all compounds at 10 µM concentration.Those compounds showing a % inhib higher than 80% were tested at different concentrations to determine their affinities (calculated as pK i ) at the diverse receptors.Table 1 shows the binding affinities of all the synthesized compounds 3 at adenosine receptors A 1 , A 2A , A 2B , and A 3 .Analysis of the data presented in the Table 1 shows that at 10 µM, several compounds presented a percentage inhibition higher than 80%.The pK i was determined, being in the range of 5.28 ≤ pK i ≤ 8.23 (3l, 3x).Both series, R = H and R = Me, gave ligands with high affinity to the four receptors subtypes; however, higher affinities are observed for derivatives with R = H (3d vs. 3h, 3c vs. 3n, 3e vs. 3p, 3k vs. 3u, 3l vs. 3v, and 3m vs. 3x).The higher potency of derivatives with a proton at N-9 agrees with the results published in our previous work [28], but the increase in potency of N-9-methyl derivatives having a morpholine unit at C-6 may indicate that the morpholine ring is interacting with the target through the oxygen atom.When we look for selectivity, we find selective ligands either having a methyl group at N-9 (3d, 3e, 3l and 3m) or a proton (3p, 3r, 3s, 3t and 3v).Additionally, the ligands with a methyl group at N-9 show high potency mainly for the A 1 receptor (3c, 3d, 3e, 3i, 3j, 3k, and 3m), and ligands with a proton at N-9 presented high potency for both the A 1 and A 3 receptors (3n-3x).To evaluate the influence of the R 1 group on potency and selectivity against the four receptors, we analyzed first the R 1 = chloro-derivatives, compounds 3a, 3d, 3f, 3h, 3o, 3q, 3r, 3s, and 3t.
In this set, the most potent and selective compound for the A 1 receptor was 3d, with R 1 = 3-ClC 6 H 4 [pK i (A 1 ) = 6.80 ± 0.07].Compounds 3r (R 1 = 3,4-Cl 2 C 6 H 3 ) and 3s (R 1 = 3-CF 3 -4-ClC 6 H 3 ) were highly potent and selective for the A 3 receptor.Compound 3q (R 1 = 4-ClC 6 H 4 ) showed dual selectivity for A 1 /A 3 receptors.These results seem to indicate that the presence of an electron-withdrawing chlorine atom at the meta position together with a hydrophobic methyl group at N-9 interacts efficiently with the A 1 receptor (3d).However, when the hydrophilic hydrogen was at N-9, a promiscuous ligand resulted (3o).The presence of a chlorine atom at the para position and a proton at N-9 interacts efficiently with receptors A 1 and A 3 (3q).Compound 3q is a potential dual ligand for A 1 /A 3 receptors with therapeutic interest, as A 1 and A 3 receptors, principally coupled to G i/o proteins, can simultaneously inhibit the adenylyl cyclase activity [6,31,32].
In addition, selectivity was achieved when a second chlorine or a CF 3 group was simultaneously at the meta position (3r, 3s), suggesting volume constraints at the A 1 , A 2A , and A 2B pockets interacting with the ligands' meta position.This hypothesis is also supported by the lower pK i observed for compound 3s (R 1 = 3-CF 3 -4-Cl-C 6 H 3 ; pK i = 6.02 ± 0.04) when compared to the pK i of compound 3r (R 1 = 3,4-Cl 2 C 6 H 3 ; pK i = 7.15 ± 0.08).
The sub-set of derivatives with hydroxyl groups at ortho or meta positions of the R 1 group (3c, 3e, 3g, 3i, 3j, 3k, 3n, 3p, and 3u) will be analyzed.Generally, a hydroxyl group at ortho or meta positions of R 1 led to very potent but not selective ligands (3c, 3i, 3j, 3k, 3n, and 3u), except compounds 3e and 3p, which have a hydroxyl group at the meta position of R 1 and showed selectivity for the A 1 receptor.Derivative 3d (R 1 = 3-ClC 6 H 4 ) also showed selectivity for the A 1 receptor.These results suggest that an electron-withdrawing group produced through the inductive effect at the meta position of R 1 interacts efficiently with a polar moiety at the A 1 receptor.Comparing the affinity values between 3b (R 1 = 2-MeOC 6 H 4 ) and 3c (R 1 = 2-HOC 6 H 4 ), the increase of affinity of 3c suggests that a hydrogen-bond donor (HBD) group at the ortho position of R 1 is important for the interaction with the adenosine receptors A 1 , A 2A , and A 2B .However, the possibility of a hydrogen-bond-acceptor (HBA) interaction of the ligand with the target should also be considered.Hydroxyl and methoxyl groups may interact that way with the targets; however, if the receptor pockets have volume constraints, this will explain the lack of activity of derivative 3b and support the hypothesis of HBD interaction.Furthermore, compound 3c has almost the same submicromolar affinity for the A 1 and A 2A receptors.This result indicates that 3c is a dual ligand for A 1 /A 2A receptors.Dual affinity for the A 1 /A 2A receptors was also observed for ligand 2k.These ligands could have medicinal interest, as compounds with dual affinity at A 1 /A 2A receptors have shown therapeutic efficacy in animal models of Parkinson's disease [33,34].On the other hand, compound 3u showed a dual affinity for receptors A 1 /A 2B .A comparison of affinities registered for 3c, 3d, and 3e with 3n, 3o, and 3p suggests more hydrophilic or small pockets at receptors to accommodate the group at N-9 of the ligands.Additionally, compound 3e (R 1 = 3-HOC 6 H 4 ) is a very potent and selective ligand for the A 1 receptor with pK i = 6.16 ± 0.21, while compound 3g (R 1 = 2,3-(HO) 2 C 6 H 3 ) has a low affinity for all of the receptors.This may indicate that an intramolecular hydrogen bond between the hydroxyl groups of the ligand precludes the interaction with the target, leading to low activity.
Finally, the fluoro derivatives (3l, 3m, 3v, and 3x) presented high potency and selectivity for receptors A 1 and A 3 , except compound 3x, which showed no selectivity.Compounds 3l and 3v with R 1 = 2-F-5-MeOC 6 H 3 were selective to the A 3 receptor, with pK i = 5.28 ± 0.11 and pK i = 7.83 ± 0.16, respectively.Compound 3v also showed a high affinity for receptor A 2B (pK i = 6.40 ± 0.16).These results suggest volume restrictions in the pockets of receptors A 2B and A 3 to lodge the methyl group at N-9 of the ligand, with the available space in receptor A 3 bigger than in receptor A 2B .Compound 3m showed high affinity and selectivity for the A 1 receptor, while compound 3x showed higher potency for all receptors but no selectivity.These results also support the hypothesis that there is limited space in the receptors' pockets to accommodate bulky groups at N-9.The difference in the selectivity of compounds 3l and 3m may indicate that receptor A 3 has a hydrophobic pocket to accommodate the substituent MeO present at the meta position of the ligand, while receptor A 1 does not.This hypothesis is also supported by the results obtained for ligands 3v and 3x.The most potent compounds for A1 (3x; pKi = 8.23 ± 0.06; Figure 3a) and A3 (3v; pKi = 7.83 ± 0.16) receptors were selected and tested in intracellular cAMP tests to study their functional activity.The agonist/antagonist behavior of the two compounds selected was determined by assessing their effect on the modulation of cAMP levels by the receptor agonist NECA. Figure 3b shows the result of a representative test for the antagonist potency of compound 3x at A1 receptor.The results confirmed that the compounds 3v and 3x are antagonists of the A3 and A1 receptors, and their antagonist potencies, expressed as pKB values, were 8.24 ± 0.11 and 8.25 ± 0.16, respectively (Table 2).Based on their structural similarity, we can generalize that all compounds of this work are antagonists of adenosine receptors.The most potent compounds for A 1 (3x; pK i = 8.23 ± 0.06; Figure 3a) and A 3 (3v; pK i = 7.83 ± 0.16) receptors were selected and tested in intracellular cAMP tests to study their functional activity.The agonist/antagonist behavior of the two compounds selected was determined by assessing their effect on the modulation of cAMP levels by the receptor agonist NECA. Figure 3b shows the result of a representative test for the antagonist potency of compound 3x at A 1 receptor.The results confirmed that the compounds 3v and 3x are antagonists of the A 3 and A 1 receptors, and their antagonist potencies, expressed as pK B values, were 8.24 ± 0.11 and 8.25 ± 0.16, respectively (Table 2).Based on their structural similarity, we can generalize that all compounds of this work are antagonists of adenosine receptors.
Molecules 2024, 29, x FOR PEER REVIEW 6 of 18 receptor A1 does not.This hypothesis is also supported by the results obtained for ligands 3v and 3x. Figure 2 represents a SAR model that summarizes the results.The most potent compounds for A1 (3x; pKi = 8.23 ± 0.06; Figure 3a) and A3 (3v; pKi = 7.83 ± 0.16) receptors were selected and tested in intracellular cAMP tests to study their functional activity.The agonist/antagonist behavior of the two compounds selected was determined by assessing their effect on the modulation of cAMP levels by the receptor agonist NECA. Figure 3b shows the result of a representative test for the antagonist potency of compound 3x at A1 receptor.The results confirmed that the compounds 3v and 3x are antagonists of the A3 and A1 receptors, and their antagonist potencies, expressed as pKB values, were 8.24 ± 0.11 and 8.25 ± 0.16, respectively (Table 2).Based on their structural similarity, we can generalize that all compounds of this work are antagonists of adenosine receptors.In this work, we synthesized highly potent and selective hits based on the scaffold for adenosine receptor antagonists identified in our previous publication [28].These results indicate that the purines substituted with a morpholine group at C-6, a hydrogen atom or a methyl group at N-9, and a specific aryl group at C-2 of purine nucleus, gave high affinity and selective antagonists for adenosine A 1 and A 3 receptors, as well as dual antagonists for receptors A 1 /A 2A , A 1 /A 2B , and A 1 /A 3 .In addition, four compounds (3i, 3j, 3n, and 3o) showed high affinity, although they were not selective for any specific adenosine receptor subtype.

Docking Studies
Compound 3x was computationally docked to the orthosteric site of the A 1 receptor using a crystal structure representing an inactive conformation (PDB accession code: 5UEN) [35].The ligand was predicted to form hydrogen bonds with Asn254 and π-π stacking with Phe171, which are interactions characteristic of adenosine receptor antagonists (Figure 4) [36].The predicted binding mode of compound 3x in the A 1 receptor also explained observed SARs.For example, replacing the fluorine at the 5-position of the phenyl ring with a methoxy group led to a substantial loss of affinity at the A 1 receptor (pK i < 5, compound 3v).Consistent with this observation, increasing the size of this substituent in our model of the complex led to clashes with Asn184 and His251.In this work, we synthesized highly potent and selective hits based on the scaffold for adenosine receptor antagonists identified in our previous publication [28].These results indicate that the purines substituted with a morpholine group at C-6, a hydrogen atom or a methyl group at N-9, and a specific aryl group at C-2 of purine nucleus, gave high affinity and selective antagonists for adenosine A1 and A3 receptors, as well as dual antagonists for receptors A1/A2A, A1/A2B, and A1/A3.In addition, four compounds (3i, 3j, 3n, and 3o) showed high affinity, although they were not selective for any specific adenosine receptor subtype.

Docking Studies
Compound 3x was computationally docked to the orthosteric site of the A1 receptor using a crystal structure representing an inactive conformation (PDB accession code: 5UEN) [35].The ligand was predicted to form hydrogen bonds with Asn254 and π-π stacking with Phe171, which are interactions characteristic of adenosine receptor antagonists (Figure 4) [36].The predicted binding mode of compound 3x in the A1 receptor also explained observed SARs.For example, replacing the fluorine at the 5-position of the phenyl ring with a methoxy group led to a substantial loss of affinity at the A1 receptor (pKi < 5, compound 3v).Consistent with this observation, increasing the size of this substituent in our model of the complex led to clashes with Asn184 and His251.

Chemistry
The imidazoles 1 used in this work were synthesized according to previously described procedures [29], and compounds 2 were synthesized according to the procedure described in [28,30].Solvents and other commercially available chemicals were used as shipped.The melting points (m.p.) were determined with a Gallenkamp melting point apparatus, and they are uncorrected.The reactions were followed by thin layer chromatography (TLC) using Silica Gel 60 F 254 (Merck, Darmstadt, Germany) plates with detection by UV light.The NMR spectra were obtained on a Varian Unity Plus ( 1 H: 300 MHz, 13 C: 75 MHz) or on a Bruker Avance III NMR spectrometer ( 1 H: 400 MHz, 13 C: 100 MHz) including the 1 H and 13 C bidimensional correlation spectra (HMQC and HMBC).For solutions, [D 6 ]-DMSO residual [D 6 ]-DMSO (δ H = 2.49 ppm) or [D 6 ]-DMSO (δ C = 39.5 ppm) was used as the internal standard at 298 K.Chemical shifts (δ) were reported in parts per million (ppm) and the coupling constants, J, were presented in hertz (Hz).The purity of all tested compounds was higher than 95% according to elemental analysis, which was reported to be within 0.4% of the calculated values.The IR spectra were recorded with a FT-IR Bomem MB 104 using nujol mulls and NaCl cells.Elemental analyses were performed with a LECO CHNS-932 instrument.

General Procedure for the Synthesis of 3a-x
Method A: To a suspension of imidazole 2 in ethanol, an ethanol/acetonitrile mixture, or DMSO, the aldehyde 4 (1.1-1.5 equivalents) was added, followed by triethylamine (10 equivalents).The reaction was kept at room to moderate temperature and followed by TLC.Otherwise, when the TLC indicated the absence of the starting reagent but the presence of intermediates, the reaction continued at a higher temperature.When the TLC showed the absence of intermediate spots, the solvent was removed in the rotary evaporator, and an off-white solid was isolated after the addition of a small amount of ethanol to the residue.When the residue showed a dark color, it was dissolved in DCM.The solution was filtered through a column of silica gel (0.5 cm high), and the column was washed with an extra 30 mL of DCM.The resulting solution was concentrated until dry, diethyl ether was added to the oil, and an off-white solid was precipitated.The solid was filtered and washed with cold diethyl ether.
Method B: To a suspension of imidazole 2 in ethanol, the aldehyde 4 (1.1-1.5 equivalents) was added, followed by trifluoroacetic acid (1.3-2.0 equivalents).The reaction was kept at 22 • C under a magnetic stirrer until the TLC showed consumption of starting reagent 2. The solvent was eliminated in the rotary evaporator, dimethyl sulfoxide and/or ethanol was added, followed by triethylamine (10 equivalents), and the reaction continued at 40-80 • C. When the TLC indicated the end of the reaction, the solvents were removed in the rotary evaporator, and the product was precipitated through the addition of water.The isolated solid was dissolved in a mixture of DCM/THF/EtOH and purified through dry flash chromatography using DCM as the elution solvent.

Radioligand Binding Assays
The % inhib of specific radioligand binding at the receptors by the compounds was tested at the concentration of 10 µM at all adenosine receptors following the conditions described below.Competition binding curves at several receptors were made by testing six different concentrations (ranging from 10 nM to 100 µM) for all of the compounds showing an % inhib ≥ 80%.

Human A 1 Receptor
Competition binding experiments of adenosine A 1 receptor were made with membranes from CHO-A 1 cells (Euroscreen, Brussels, Belgium).Membranes were defrosted and suspended in an incubation buffer with 10 mM of MgCl 2 , 100 mM of NaCl, 20 mM of Hepes, 2 UI/mL of adenosine deaminase, and pH = 7.4.Each well of a GF/C multiscreen plate (Millipore, Madrid, Spain), prepared in duplicate, contained 15 µg of protein, 2 nM of [ 3 H]DPCPX, and the tested compound.Non-specific binding was calculated with 10 µM of (R)-PIA.The reaction mixture was incubated at room temperature (25 • C) for 60 min, after which samples were filtered and measured in a microplate beta scintillation counter (Microbeta Trilux, Perkin Elmer, Madrid, Spain).

Human A 2A Receptor
Competition binding experiments of adenosine A 2A receptor were made in membranes from HeLa-A 2A cells, which were provided by Dr. Mengod (Departament de Neuroquimica, Institut d'Investi-gacions Biomediques de Barcelona, CSIC-IDIBAPS, Barcelona, Spain).Membranes were defrosted and suspended in an incubation buffer with 1 mM of EDTA, 50 mM of Tris-HCl, 10 mM of MgCl 2 , and 2 UI/mL of adenosine deaminase, with pH = 7.4.Each well of a GF/C multiscreen plate (Millipore, Madrid, Spain), prepared in duplicate, contained 3 nM of [ 3 H]ZM241385, 10 µg of protein, and the tested compound.Nonspecific binding was calculated with 50 µM of NECA.The reaction mixture was incubated at room temperature (25 • C) for 30 min, after which samples were filtered and measured in a microplate beta scintillation counter (Microbeta Trilux, Perkin Elmer, Madrid, Spain).

Human A 2B Receptor
Competition binding experiments of adenosine A 2B receptor were made in membranes from HEK-293-A 2B cells (Euroscreen, Brussels, Belgium).Membranes were defrosted and suspended in incubation buffer with 1 mM of EDTA, 10 mM of MgCl 2 , 0.1 mM of benzamidine, 50 mM of Tris-HCl, 10 µg/mL of bacitracine, and 2 UI/mL of adenosine deaminase, with pH = 6.5.Each reaction well, prepared in duplicate, contained 35 nM of [ 3 H]DPCPX, 18 µg of protein, and the tested compound.Non-specific binding was calculated with 400 µM of NECA.The reaction mixture was incubated at room temperature (25 • C) for 30 min, after which samples were filtered through a multiscreen GF/C microplate and measured in a microplate beta scintillation counter (Microbeta Trilux, Perkin Elmer, Madrid, Spain).

Human A 3 Receptor
Competition binding experiments of adenosine A 3 receptor were made in membranes from HeLa-A 3 cells, which were provided by Dr. Mengod (Departament de Neuroquimica, Institut d'Investi-gacions Biomediques de Barcelona, CSIC-IDIBAPS, Barcelona, Spain).Membranes were defrosted and suspended in incubation buffer with 1 mM of EDTA, 5 mM of MgCl 2 , 50 mM of Tris-HCl, and 2 UI/mL of adenosine deaminase, with pH = 7.4.Each reaction well of a GF/B multiscreen plate (Millipore, Madrid, Spain), prepared in triplicate, contained 90 µg of protein, 30 nM of [ 3 H]NECA, and the tested compound.Non-specific binding was determined with 100 µM of (R)-PIA.The reaction mixture was incubated at room temperature (25 • C) for 180 min, after which samples were filtered and measured in a microplate beta scintillation counter (Microbeta Trilux, Perkin Elmer, Madrid, Spain).

Human A 1 Receptor
The agonist/antagonist behavior of tested compounds at A 1 receptor was determined in CHO-A 1 cells by measuring reversion induced by the tested compound of 10 µM of NECA-mediated inhibition of forkolin-stimulated cAMP production.Cells grown in 96-well plates with growth medium containing dialyzed fetal bovine serum were washed 2 times with F-12 nutrient mixture medium containing 25 mM of HEPES and 20 µM of the phosphodiesterase inhibitor rolipram, with pH = 7.4.After this time, the tested compounds were preincubated for 15 min in test medium at different concentrations (ranging from 0.1 nM to 10 µM).After this time, 10 µM of NECA and 3 µM of forskolin were added to each well.The incubation was continued for 15 min, and the reaction was stopped with a lysis buffer from the cAMP enzyme immunoassay kit (GE Healthcare, Chicago, IL, USA).Cell lysates were put in a plate with anti-IgG and anti-cAMP in the presence of cAMPperoxidase during 60 min.Then, the wells were washed 4 times with the wash buffer from the kit, and TMB was added to each well and incubated for 60 min.The peroxidase reaction was stopped with 1 M of sulphuric acid, and the cAMP amount was determined from the optical density at 450 nm (Tecan M100 reader).

Human A 3 Receptor
The agonist/antagonist behavior of tested compounds at A 3 receptor was determined in CHO-A3 cells, which were in-house generated at BioFarma Research Group (Department of Pharma-cology, University of Santiago de Compostela, Spain), by measuring reversion induced by the tested compound of 10 µM of NECA inhibition of forskolin-stimulated cAMP production.Cells grown in 96-well plates with growth medium containing dialyzed fetal bovine serum were washed 2 times with DMEM F-12 nutrient medium containing 25 mM of HEPES, 30 µM of rolipram, with pH = 7.4.After this time, the tested compounds were preincubated for 15 min in test medium at different concentrations (ranging from 0.1 nM to 10 µM).Finally, 10 µM of NECA and 10 µM of forskolin were added to each well.The incubation lasted 15 min, and cAMP was determined through an enzyme immunoassay (Perkin Elmer) for Human A 1 receptor.

Data Analysis
Concentration response curves were fitted with Prism 10.1.2(Graph Pad, San Diego, CA, USA).Radioligand binding pK i values were extrapolated from the equation where IC 50 is the concentration that displaced specific radioligand binding at 50%, extrapolated from non-linear fitting, [L] is the concentration of radioligand employed, and K D is the dissociation constant of the radioligand.In cAMP studies, K B values were derived by employing the derivation of the Cheng-Prusoff equation, reported by Leffand Dougall in 1993 [37].K B = IC 50 /(2 + (A/(EC 50 ) n ) 1/n − 1); where A is the concentration of the agonist employed to stimulate the receptor, EC 50 is the potency of the agonist for stimulating the receptor, and n is the Hill slope of the antagonist's concentration-response curve.

Docking Studies
The molecular docking calculations were carried out using the Molecular Operating Environment (MOE) software (version 2022.02,https://www.chemcomp.com,accessed on 25 April 2024) and the crystal structure of the A 1 receptor in complex with the ligand DU1 (PDB accession code: 5UEN) [35].The receptor was prepared by removing nonprotein atoms and adding missing side chains.Engineered mutations were reverted back to the wild-type sequence based on the side chain rotamers determined through MOE.The protonation states of ionizable residues were set to the most probable at physiological pH.The binding site was defined based on the co-crystallized ligand, and default parameters in MOE were used in the docking calculations.

Conclusions
In summary, several 6-morpholino-purine derivatives with different substituents in C-2 and a methyl or a hydrogen at N-9 of the purine ring were synthesized and pharmacologically tested at human A 1 , A 2A , A 2B , and A 3 adenosine receptor subtypes.Four of the twenty-three synthesized purines showed high affinity and selectivity for A 1 receptors, and five showed high affinity and selectivity for A 3 receptors.Three compounds (3k, 3q, and 3u) showed a dual affinity for A 1 /A 2A , A 1 /A 3 , and A 1 /A 2B receptors, respectively.In addition, three compounds showed high affinity without selectivity for any specific adenosine receptor subtype (3n, 3o, and 3x).SAR studies showed that the substituent groups in C-2, C-6, and N-9 positions affect the potency and selectivity of the compounds for adenosine receptors.To generate highly potent antagonist ligands for adenosine receptors, a morpholine group and a hydrogen atom must be present in C-6 and N-9, respectively.Indeed, 9H-purines bind some 5-358 times more strongly than their corresponding 9-CH 3purines for the adenosine receptors.The substituted aryl group in C-2 is very important for governing the selectivity towards the adenosine receptor subtypes.Additionally, our results suggested that the A 3 receptor subtype presents a larger hydrophobic pocket than that of the A 1 receptor subtype, as A 3 accommodates bigger substituent groups Y present at the meta position of substituent R 1 .In conclusion, several purine-based compounds are described herein, some with high affinity and antagonist selectivity for A 1 , A 3 , dual A 1 /A 2A , A 1 /A 2B , and A 1 /A 3 receptors.

Figure 1 .
Figure 1.The adenosine receptor antagonists scaffold identified in our previous work.

Figure 1 .
Figure 1.The adenosine receptor antagonists scaffold identified in our previous work.
Figure 2 represents a SAR model that summarizes the results.Molecules 2024, 29, x FOR PEER REVIEW 6 of 18receptor A1 does not.This hypothesis is also supported by the results obtained for ligands 3v and 3x.Figure2represents a SAR model that summarizes the results.

Figure 3 .Figure 2 .
Figure 3. (a) Competition binding curve for compound 3x at A1 receptors.(b) Concentration-response curve of compound 3x in the presence of 10 µM of NECA at human A1 receptor expressed in CHO cells.Points represent the mean ± standard deviation (vertical bars) of duplicate measurements.

Figure 3 .Figure 3 .
Figure 3. (a) Competition binding curve for compound 3x at A1 receptors.(b) Concentration-response curve of compound 3x in the presence of 10 µM of NECA at human A1 receptor expressed in CHO cells.Points represent the mean ± standard deviation (vertical bars) of duplicate measurements.

Figure 4 .
Figure 4. Predicted binding mode of compound 3x in the orthosteric site of the A1 receptor.The receptor is depicted as a grey cartoon, with the ligand and key side chains represented as sticks.Hydrogen bonds are shown as black dashed lines.

Figure 4 .
Figure 4. Predicted binding mode of compound 3x in the orthosteric site of the A 1 receptor.The receptor is depicted as a grey cartoon, with the ligand and key side chains represented as sticks.Hydrogen bonds are shown as black dashed lines.

Table 1 .
Binding affinities of compounds 3 at all human adenosine receptors expressed as % inhib at 10 µM or pK i .Values are reported as mean ± SEM of three experiments with duplicate measurements.

Table 2 .
Antagonist potency (pK B ) of the most potent compounds at human A 1 and A 3 adenosine receptors in cAMP tests.Values are reported as mean ± SEM of three experiments with duplicate measurements.

Table 2 .
Antagonist potency (pKB) of the most potent compounds at human A1 and A3 adenosine receptors in cAMP tests.Values are reported as mean ± SEM of three experiments with duplicate measurements.