A2A Adenosine Receptor Antagonists: Are Triazolotriazine and Purine Scaffolds Interchangeable?

The A2A adenosine receptor (A2AAR) is one of the four subtypes activated by nucleoside adenosine, and the molecules able to selectively counteract its action are attractive tools for neurodegenerative disorders. In order to find novel A2AAR ligands, two series of compounds based on purine and triazolotriazine scaffolds were synthesized and tested at ARs. Compound 13 was also tested in an in vitro model of neuroinflammation. Some compounds were found to possess high affinity for A2AAR, and it was observed that compound 13 exerted anti-inflammatory properties in microglial cells. Molecular modeling studies results were in good agreement with the binding affinity data and underlined that triazolotriazine and purine scaffolds are interchangeable only when 5- and 2-positions of the triazolotriazine moiety (corresponding to the purine 2- and 8-positions) are substituted.


Introduction
The A 2A adenosine receptor (A 2A AR) subtype is one of the four G protein-coupled receptors (GPCRs,) named A 1 , A 2A , A 2B , and A 3 , which are ARs activated by the natural nucleoside adenosine (Ado) [1,2]. In contrast to the A 1 and A 3 receptor subtypes, which are coupled to G inhibitory proteins, the A 2A AR and A 2B AR are coupled to G stimulatory proteins and their activation produces an increase in the second messenger cAMP [3]. It is well-known that Ado is among the key modulators in the manifestation of the periphery and central nervous system (CNS) inflammation. This activity of Ado is mainly due to the activation of the A 2A AR among the AR subtypes. In the periphery, Ado reduces inflammation through the decreased recruitment of leucocyte [4][5][6]. Regarding the CNS, several studies have demonstrated that A 2A AR antagonists play a key role in counteracting neuroinflammation [7][8][9], a common event in neurodegenerative diseases, such Examples of other non-xanthine A 2A AR antagonists are adenine derivatives, which have been the subject of our research for a long time. In fact, we synthesized a number of such molecules substituted at the 2-, 9-, and/or 8-positions, which were found to be-have as A 2A AR antagonists endowed with various degrees of affinity and selectivity for ARs [29]. Among them, the 9-ethyl-8-ethoxyadenine (ANR 94, 5, Figure 1) showed good affinity and selectivity for the human A 2A AR subtype (Ki A 1 = 2400 nM, Ki A 2A = 46 nM, Ki A 2B > 30,000 nM, and Ki A 3 = 21,000 nM) and demonstrated very high efficacy in in vivo models of Parkinson's disease. In fact, it reversed the catalepsy induced by haloperidol and induced contralateral turning behavior in rats sensitized to L-DOPA. Moreover, this compound showed good efficacy in lowering the intensity of the tremor evoked by Parkinson's induced by tacrine and in reducing the motor deficit in a neuronal injury model induced by 6-hydroxydopamine [30]. Despite these promising results, ANR 94 does not show a remarkable affinity for A 2A ARs (Ki = 46 nM) with respect to other A 2A AR antagonists like as ZM241385, which is considered the reference compound for studying the A 2A AR.
Starting from these observations, and in order to verify whether the triazolotriazine scaffold of the ZM241385 can be replaced by the purine ring to give a high affinity A 2A AR ligand, we synthesized an adenine derivative bearing the same substituents present in the 2-and 5-positions of the ZM241385 structure. Hence, an adenine analogue substituted at the 2-position with a (4-hydroxy)-2-phenylethylamino group and bearing a furyl ring at the 8-position, was designed and synthesized together with its 4-methoxy analogue. At the 9-position of these molecules, we introduced an ethyl group that favors the interaction of adenine derivatives with A 2A AR [31] (Figure 2a). We also replaced the furyl ring present in the 8-position of ZM241385 with an ethoxy group (Figure 2b). Furthermore, aimed at improving the A 2A AR binding affinity of ANR 94, we designed and synthesized its triazolotriazine analogue, which lacks, for structural requirement reasons, the ethyl group present at the 9-position of the ANR 94 purine ring ( Figure 2b).
All the newly synthesized compounds were evaluated in binding and functional studies in order to assess their affinity/activity at ARs and the compound endowed with the most promising profile of A 2A AR affinity and selectivity was also evaluated in a microglial model of neuroinflammation.

Chemical Synthesis
The 9-ethyl-8-furyl-2-(4-hydroxyphenyl)-2-ethylaminoadenine (13) and its 4-methoxyphenyl analogue 12 were synthesized starting from the 2-chloro-9-ethyladenine (6), obtained from the commercially available 2,6-dichloropurine, in two steps [32]. Compound 6 was reacted with the p-methoxyphenethylamine, at 130 • C for 24 h, to give the disubstituted adenine derivative 7, which was brominated using N-bromosuccinimide (NBS) in dimethylformamide (DMF) for 30 min at r. t. to obtain the 8-bromo derivative 8 (Scheme 1). The latter was treated with hydrobromic acid (48% solution in water), at r. t. for 2 h, in order to convert the phenylmethoxy group to a phenol. In this reaction, compounds 10 and 11, deriving from the substitution of the 8-bromine atom and partial hydrolysis of the methyl ether, were obtained together with the desired 8-bromo-2-(4-hydroxy-phenyl)-2-ethylaminoadenine (9). Compounds 8 and 9 were, in turn, reacted with (2-tributylstannyl)furane, in tetrahydrofurane (THF) at r. t. for 16 h, to obtain the desired trisubstituted adenine derivatives 12 and 13.  3,5]-triazine derivatives 20 and 21, a targeted approach, aimed at forming first the triazole nucleus and then the triazine ring, was used (Scheme 2). A similar approach was also utilized for the synthesis of ZM241385 [33], but here we chose the commercially available dimethyl Ncyanodithioiminocarbonate (14) as the starting material which presents two very important advantages: the low price and the fact that it can be used both for the formation of the triazole and the triazine ring. In fact, 14 was reacted with aqueous hydrazine in toluene at 0 • C to afford the 5-(methylthio)-1H-1,2,4-triazol-3-amine (15 The formation of the triazole ring was followed by the oxidation of the 15-thiol group, using meta-chloroperbenzoic acid (m-CPBA) in dichloromethane at r. t., to give the sulfone derivatives 16. As already mentioned, the triazine ring was obtained by treating 16 with the starting reagent 14 in a fusion reaction performed in a steel bomb at 170 • C for 4 h. The reaction led to the formation of a main product which was isolated after chromatography on a silica gel column. Since this reaction could lead to the formation of two or more isomers of 17 [34], the structure of the obtained compound was unequivocally attributed through a further crystallographic X-ray analysis (see next section).
The reaction of 17 with sodium ethoxyde, obtained by treating ethanol with sodium hydride for 24 h, furnished the 5-ethoxy derivative 18 as the main product together with a small percentage of the disubstituted derivative 19.
The ZM241385 analogue 20 was obtained in a two-step reaction. First, the oxidation of the thiomethyl group of 18 with m-CPBA to the sulfone, which is a better leaving group, followed after 2 h when the starting material was totally consumed by the addition of tyramine. Then, the reaction mixture was left at 40 • C for 12 h.
Finally, hydrogenolisis of 18, using triethylsilane as a proton donor and Pd/C 10% as the catalyst, allowed the removal of the thiomethyl goup to obtain the desired triazolotriazine 21.
As already mentioned in the introduction, ZM241385 (4) possesses a high affinity for the A 2A AR (Ki = 1.6 nM), good selectivity versus the A 1 and A 3 AR subtypes by 484-and 464-fold, respectively, and lower selectivity towards the A 2B AR by just 47 times.
Its purine analogue maintained high affinity at the A 2A AR but showed an inverse profile of selectivity that decreased towards the A 1 and A 3 subtypes and increased versus the A 2B AR (13; Ki A 2A = 1.8 nM; selectivity A 1 /A 2A = 28, A 3 /A 2A = 7, and A 2B /A 2A = 796). Supposing that the triazoletriazine scaffold of ZM241385 and the purine core of 13 are positioned with the same orientation in the binding pocket of the A 2A AR, the presence of the ethyl group in 9-position of 13 seems to favor the interaction of this molecule with the A 1 and A 3 subtypes and disadvantage the one with the A 2B AR, while it does not seem to influence the interaction with the A 2A AR. These findings indicate that, at least with regards to A 2A AR affinity, the triazoletriazine and purine scaffolds are interchangeable. When the phenolic hydroxyl group of 13 is replaced by a methoxy substituent, a further increase in affinity at the A 1 AR is observed, leading to an A 1 /A 2A unselective ligand (12; Ki A 2A = 1.0 nM and Ki A 1 = 2.6 nM). As expected, the presence of an 8-bromine atom, instead of a furyl ring, led to a decrease in affinity at A 1 , A 2A , and A 3 ARs; in fact, 9 and 8 showed higher Ki values with respect to 12 and 13 while remaining A 2A selective ligands These findings are in agreement with our previous results which demonstrated that the presence of an 8-bromine atom and, even more, a 8-furyl ring, favor the interaction of 2-9-disubstituted adenine derivatives at ARs [29]. Furthermore, the presence of a hydroxyl group at the 8-position of these molecules is detrimental for the affinity, especially at the A 2A and A 3 subtypes (10; Ki  Unfortunately, its triazolotriazine analogue 20 showed a decrease in A 2A AR affinity by 44fold (Ki = 2029 nM) and a decrease at the A 1 subtype but not at the A 3 AR (Ki = 1019 nM). In this case, the replacement of the purine ring with the triazolotriazine scaffold had a negative impact on the A 2A AR affinity. Among the 2,5-disubstituted triazoletriazines reported here, only compounds 18 and 21, the latter being the 8-ethoxy analogue of ZM241385, showed a sub-microM affinity at the A 2A AR with Ki values of 583 nM and 178 nM, respectively.
The obtained results did not allow to give a univocal answer to the question of the interchangeability between the two scaffolds that we have tried to explain with molecular modeling studies.

Study of the Anti-Inflammatory Activity of 13
In order to investigate whether the promising affinity profile of 13 could be associated with potential protective activity against PD, we studied its anti-inflammatory properties in a microglial model of neuroinflammation. Increasing evidence indicates that neuroinflammation mediated by microglia plays a major role in PD [40][41][42]. Although it is still unclear whether the neuroimmune response may or not represent a primary cause for neuronal loss, there is wide consensus that it participates in disease progression [43]. ARs expressed in microglial cells modulate inflammatory response in PD and preliminary data suggest that A 2A AR antagonists may have therapeutic efficacy [8,44]. Activated microglia can assume two different polarization states: a pro-inflammatory "M1" phenotype and an anti-inflammatory "M2" phenotype. Interestingly, in PD patients, an increase in M1-polarized microglia is observed [45]. M1-phenotype microglia is characterized by the increased production of inflammatory cytokines, such as interleukin-1β (IL-1β) and IL-6, that leads to tissue damage. On the other hand, the M2-phenotype is characterized by an upregulation of anti-inflammatory mediators, such as IL-10. In activated microglia, these two phenotypes can coexist, revealing the complexity of microglia function and the dynamic changes of the environment in vivo [46]. Different studies associate the release of pro-inflammatory mediators with the activation of the NOD-like receptor family, the pyrin domain containing-3 protein (NLRP3) inflammasome in BV-2 cells [47]. NLRP3, a multi-protein complex, modulates the maturation and secretion of pro-inflammatory cytokines, including IL-1β. Of note, the inhibition of the NLRP3 inflammasome activation could protect dopaminergic neurons [48,49]. On these bases, the identification of new compounds able to inhibit the activation of NLRP3 Inflammasome are a promising therapeutic intervention against neurodegeneration in PD.
The microglial BV-2 cell line exposed to bacterial endotoxin lipopolysaccharide (LPS) was chosen as the in vitro model of neuroinflammation. LPS, a pro-inflammatory mediator, is widely used in in vitro studies to activate microglia cells and induce proinflammatory transduction pathways [50][51][52]. In addition, BV-2 cells are a valid alternative to primary microglia as 90% of the genes regulated by LPS in BV-2 cells are also modulated in primary microglia and both display similar reaction patterns [53].
First of all, we verified the potential cytotoxicity of 13 in our cell model system. BV-2 cells were treated with increasing concentrations of the compound for 24 h and cell viability was evaluated by MTT assay ( Figure 4A). Of note, the tested A 2A AR antagonist did not show any cytotoxicity up to 5 µM, demonstrating its high safety profile.
To determine the effects of 13 on LPS-induced inflammatory mediators, we initially evaluated the production of NO. BV-2 cells were treated with different concentrations of 13, then, were exposed to 100 ng/mL LPS for 24 h, and the release of NO in the culture medium was measured by Griess assay ( Figure 4B). LPS strongly and significantly increased the release of NO in the culture medium with respect to the control cells (CTRL). The treatments with 13 at the concentrations 1 and 10 nM significantly reduced the NO level with respect to LPS treated cells. Interestingly, 10 nM treatment was the most effective in counteracting NO release as 1 nM led to significantly higher NO production than 10 nM. On theses bases, the concentration 10 nM was chosen for the following experiments.
To better characterize the anti-inflammatory activity of 13, we evaluated the expression of the main pro-inflammatory cytokines and enzymes up-regulated in activated microglial cells, such as IL-1β, IL-6, inducible NO synthase (iNOS), and cyclooxygenase 2 (COX2). ANR 94 was chosen as the reference compound as we already demonstrated its ability to reduce microglia activation in a murine model of PD [8]. BV-2 cells were treated with 13 or ANR 94 (10 nM) and activated with 100 ng/mL LPS for 24 h. Total RNA was isolated, and proinflammatory cytokine and enzyme expressions were measured using RT-PCR ( Figure 5). As expected, LPS significantly increased the expression of IL-1β, IL-6, COX-2, and iNOS with respect to the control cells. 13 significantly reduced the expression of all the pro-inflammatory mediators tested with respect to cells exposed to LPS. On the other hand, ANR 94 was only able to significantly reduce the expression of IL-1β. In particular, the treatment with 13 led to a significant down-regulation of IL-6, iNOS, and COX-2 with respect to ANR 94, demonstrating it possesses higher efficacy than ANR 94 in reducing the expression level of these pro-inflammatory mediators. Cells were treated with 10 nM 13 or 10 nM ANR 94 for 24 h, exposed to 100 ng/mL LPS for 24 h and cytokines and enzymes mRNA levels were measured by RT-PCR. Data are expressed as relative abundance compared to untreated cells. Each bar represents the mean ± SEM of three independent experiments. Data were analyzed with one-way ANOVA followed by Bonferroni's test. * p < 0.05 vs. CTRL; • p < 0.05 vs. LPS; § p < 0.05 vs. ANR 94 + LPS.
To verify if the ability of 13 to reduce the previous inflammatory mediators is also associated with a reduction in the inflammasome, we measured the expression of NLRP3 inflammasome ( Figure 6). The data demonstrated that, after LPS stimulation, the expression of NLRP3 mRNA, which is the rate-limiting step for inflammasome activation [54], was significantly increased. Moreover, the increased mRNA of NLRP3 was significantly reduced by 13 pre-treatment.
As IL-10 was reported to promote microglial M2 polarization [55], we then evaluated the expression of IL-10 in BV-2 cells treated with 13 before LPS activation ( Figure 6). As expected, LPS significantly reduced the expression of this anti-inflammatory cytokine with respect to the control cells, while treatment with 13, in agreement with the previous results, maintained IL-10 expression to a level comparable to the control cells, suggesting that 13 promotes the protective M2 phenotype. Figure 6. Expression of NLRP3 and IL-10 in BV-2 cells treated with 13. Cells were treated with 10 nM 13, exposed to 100 ng/mL LPS for 24 h and NLRP3 and IL-10 mRNA levels were measured by RT-PCR. Data are expressed as relative abundance compared to untreated cells. Each bar represents the mean ± SEM of three independent experiments. Data were analyzed with one-way ANOVA followed by Bonferroni's test. * p < 0.05 vs. CTRL; • p < 0.05 vs. LPS.
In conclusion, 13 could represent a promising A 2A AR antagonist in the fight against PD due to its ability to reduce pro-inflammatory mediators and to increase anti-inflammatory cytokines in microglia cells. Future studies are needed to increase the knowledge of its anti-inflammatory molecular mechanisms and to verify its efficacy in animal models of PD before suggesting this interesting compound for clinical applications.

Molecular Modeling Studies
Molecular modeling studies, consisting of docking experiments at the A 2A AR, A 1 AR, and A 3 AR 3D structures, were performed to analyze the binding data of the developed compounds. We employed the crystal structures of the human A 2A AR and A 1 AR in combination with ZM241385 and the antagonist PSB36, respectively, downloaded from the Protein Data Bank webpage (http://www.rcsb.org accessed on 7 January 2022; pdb code: 5NM4; 1.7-Å resolution [26], and pdb code: 5N2S; 3.6-Å resolution [56], respectively). A homology model of the human A 3 AR was built using the above cited X-ray structure of the antagonist-bound A 1 AR as a template (pdb code: 5N2S). Docking analyses were performed by using CCDC Gold [57] and then analyzed within the Molecular Operating Environment (MOE, version 2020.09) suite [58].
The docking analysis with the compound ZM241385 at the A 2A AR was useful for checking the reliability of the docking protocols, by comparing its obtained top score docking conformation with the co-crystallized arrangement of the same molecule within the receptor structure. Results were in good agreement with the experimental data (RMSD = 1.31 Å).
The docking conformations generally observed for the new adenine derivatives at the A 2A AR are similar to those obtained in our previous studies and are showed in Figure 7A, where the purine derivative 13 is perfectly overlapped to the co-crystallized ligand ZM241385. The bicyclic purine core is positioned between the side chains of Phe168 (EL2) and Leu249 6.51 and gives non-polar interactions with these residues. The N 6 -amine group makes H-bond interactions with Asn253 6.55 and Glu169 (EL2), while the 8-substituent is located in the depth of the binding cavity. The 2-substituent is positioned at the entrance of the binding site and points toward the extracellular environment. Purine derivatives lacking the 2-substituent (i.e., ANR 94 (5), Figure 7B) or presenting a small moiety in the same position, are observed to give an analogue binding mode (respect to 13 and ZM241385), with interactions given by the purine core and the exocyclic amine group in the 6-position, and further interactions made by the other substituents. Hence, the lack of the 2-substituent or its substitution with a small group does not seems critical for the binding mode of the 9-ethyladenines. In fact, both 13 and 2-unsubstituted ANR 94 possess nM affinity at the A 2A AR. In the case of compound 21, based on a triazoletriazine scaffold and lacking the exocyclic ethyl chain (compared to ANR 94) and the (4-hydroxy)-2-phenylethylamino group (compared to ZM241385), we observed two potential binding modes ( Figure 7C,D).
One of these binding modes presents the triazoletriazine scaffold of 21 as almost superimposable to the one of ZM241385, while the other docking conformation makes the triazoletriazine moiety oppositely oriented (see the docking conformation colored in cyan, Figure 7C,D, respectively). We may assume that the presence of a 9-ethyl group in the adenine derivatives makes their binding mode stably anchored within the binding cavity. On the contrary, the triazolotriazine derivatives lacking the long (4-hydroxy)-2phenylethylamino group are less stably inserted within the binding cavity, giving more potential binding modes. This feature appears to be slightly beneficial in the case of the A 3 AR, for which the affinity of 21 is higher than ANR 94.
Compounds ZM241385 and 13 make the same interactions with the A 2A AR, apart from the ligand region corresponding to the 9-position of the purine core. Considering the purine derivative, its 9-ethyl substituent is located in the depth of the cavity and makes hydrophobic interactions with the residues in proximity, such as Ala63 2.61 , Ile66 2.64 , Val84 3.32 , Leu85 3.33 , Thr88 3.36 , Trp246 6.48 , Ile274 7.39 , Ser277 7.42 , and His278 7.43 (see Figure 8). Binding affinity data of ZM241385 and 13 show that at the A 2A AR they are endowed with comparable affinity, while 13 presents higher affinity (with respect to ZM241385) at the A 1 AR and A 3 AR. Given that the unique structural difference between these two molecules is the presence of a 9-ethyl group in the purine derivative and a nitrogen atom in the corresponding position of ZM241385, this feature appears responsible for the different affinity of the above cited molecules at these two receptors. We analyzed and compared the ability of these two compounds to interact with the A 2A AR, A 1 AR, and A 3 AR with the aid of the IF-E 6.0 tool (see Experimental section for details), which calculates the atomic and residue interaction forces and the per-residue interaction energies (expressed as kcal × mol −1 ). This tool was previously used for analogue analyses at hARs obtaining useful interpretation of the compound activity [59][60][61]. We focused on the above residues located in proximity of the 9-ethyl group of 13. The results are displayed in Figure 8.
The comparison of the interaction energies calculated for ZM241385 and 13 at the A 2A AR shows that the two compounds make analogue interactions with a sum of −5.53 and −5.63 kcal mol −1 , respectively. This is in agreement with the similar binding affinity data of the two molecules at this AR subtype. At the A 1 AR and the A 3 AR, the two molecules show a different pharmacological behavior, with significantly higher affinity for both receptors shown by 13 compared to ZM241385. The interaction energies calculated at these receptors are again in agreement with the biological data, since at the A 1 AR and the A 3 AR the sum of the interaction energies for 13 is −5.84 and −6.76 kcal mol −1 , respectively, while at the same receptors the sum of the interaction energies for ZM241385 is −4.82 and −4.88 kcal mol −1 , respectively. Hence, the presence of an ethyl group in the 9-position of adenine appears beneficial to achieve affinity at the analyzed AR subtypes, while its absence appears to provide some selectivity for the A 2A AR.

Crystallographic Study
The crystallographic data for complex 17 were collected on a Bruker APEX II singlecrystal diffractometer working with monochromatic Mo-Kα radiation and equipped with an area detector. The structure was solved by direct methods and refined against F 2 with SHELXL-2014/7 with anisotropic thermal parameters for all non-hydrogen atoms [62,63]. Idealized geometries were assigned to the hydrogen atoms. Crystallographic data were deposited with the Cambridge Crystallographic Data Centre as a supplementary publication (reference code CCDC 2133489). Copies of the data can be obtained free of charge upon application to the CCDC, 12 Union Road, Cambridge CB2 1EZ, UK (fax, (+44) 1223 336033; e-mail, deposit@ccdc.cam.ac.uk).
Preparation of membranes: Cell membranes for binding assays were prepared mechanically detaching the cells from the petri and suspending them in a cold hypotonic Concentration of nitrites released in the medium was determined from the standard curve generated with known concentrations of sodium nitrite. Results are expressed as mean concentration of nitrites (µM) ± SEM, from three different samples.
Real-Time Polymerase Chain Reaction (PCR): At the end of the experiments, total RNA was extracted by BV-2 cells using RNeasy Mini Kit (QIAGEN GmbH, Hilden, Germany). RNA concentraion and quality were measured on a NanoVue Spectrophotometer (GE Healthcare, Milano, Italy). iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA, USA) was used to synthesize cDNA starting from the extracted RNA, following the supplier's instructions. PCR was performed by adding 2.5 µL (12.5 ng) of cDNA, 5 µL SsoAdvanced Universal SYBR Green Supermix (Bio-Rad), and 0.5 µL (500 nM) of each primer (Table 2) to a PCR tube. cDNA amplification was started at 95 • C for 30 s to activate the polymerase, followed by 40 cycles of 5 s at 95 • C and 30 s at 60 • C. Normalized expression levels were calculated relative to the control cells according to the 2 −∆∆CT method. Table 2. Reports the primers used (Sigma-Aldrich-Merck). GAPDH was used as the reference gene.

Molecular Modeling
Receptor refinement and energy minimization tasks were carried out using MOE. Docking experiments were performed with CCDC Gold [57].
A 2A AR and A 1 AR crystal structures refinement: The crystal structures of the human A 2A AR and A 1 AR in complex with ZM241385 and the antagonist PSB36, respectively, were downloaded by the Protein Data Bank webpage (http://www.rcsb.org accessed on 7 January 2022; pdb code: 5NM4; 1.7-Å resolution [26], and pdb code: 5N2S; 3.6-Å resolution [56], respectively). The two structures were checked into MOE by restoring missing loops and the wild-type receptor sequences, and by adding all the hydrogen atoms. The protein structures were then energetically minimized with MOE using the AMBER99 force field until the RMS gradient of the potential energy was less than 0.05 kJ mol −1 Å −1 .
Homology modeling of the human A 3 AR structure: A homology model of the human A 3 AR was built using the above cited X-ray structure of the antagonist-bound A 1 AR as a template (pdb code: 5N2S). A multiple alignment of the AR primary sequences was built within MOE as the preliminary step. The Homology Modeling tool of MOE was employed for this task. The obtained A 3 AR model was then energetically minimized with the same protocol of the other receptors (see above).
Molecular docking analysis: Docking analyses were performed by using CCDC Gold [57], which was set with default efficiency settings by selecting ChemScore as the scoring function and generating 50 poses for each ligand.
Post Docking analysis. Residue interaction analysis: The interactions between the ligands and the AR receptors binding sites were analyzed by using the IF-E 6.0 tool retrievable at the SVL exchange service (Chemical Computing Group, Inc. SVL exchange: http://svl.chemcomp.com accessed on 7 January 2022). The program calculates and displays the atomic and residue interaction forces as 3D vectors. It also calculates the per-residue interaction energies, where negative and positive energy values are associated to favorable and unfavorable interactions, respectively. For each AR structure, a shell of residues contained within a 10 Å distance from the ligand were considered for this analysis.

Conclusions
On the base of ZM241385 and ANR 94 structures, two series of new compounds endowed with the triazolotriazine, and purine scaffolds were designed and synthesized. Radioligand binding and functional assays at human recombinant ARs showed that the 9-ethyl-8-furyl-2-(4-hydroxyphenyl)-2-ethylaminoadenine (13) and its 4-methoxyphenyl analogue 12 showed high affinity (Ki in the low nM range) and a different degree of selectivity at the A 2A AR. Furthermore, 13 was found to exert anti-inflammatory properties in a microglial model of neuroinflammation, since it was able to reduce pro-inflammatory mediators and to increase anti-inflammatory cytokines in microglia cells.
Comparing the ARs binding mode of selected compounds belonging to the two series, molecular modeling studies support the hypothesis that the triazolotriazine and the purine scaffolds are interchangeable only when the 2 and 5 positions of the triazolotriazine moiety (corresponding to the purine 2-and 8-positions) are substituted. Taken together, the results of our experiments reveal that the novel A 2A AR antagonist 13 is a potential tool for studying neuroinflammation and neuroprotective processes and could be a useful anti-PD agent suitable for further investigations.