Navigating in chromone chemical space: discovery of novel and distinct A₃ adenosine receptor ligands

Abstract

One of the major hurdles in the development of safe and effective drugs targeting G-protein coupled receptors (GPCRs) is finding ligands that are highly selective for a specific receptor subtype. The search for novel compounds with therapeutic value by targeting the A₃ adenosine receptor (A₃AR) is still in its early stages. The increasing knowledge about the biological, physiological and pathological role of the A₃AR subtype was accompanied by the design and development of the A₃AR ligands, but the particular role of A₃AR agonists and antagonists is still an open issue. Among the large variety of chemical classes screened towards ARs flavonoids have been indicated as remarkable A₃AR antagonists. However, the search of A₃AR ligands based on this framework seems to be discontinued. In this context, our research group focused its investigation into the discovery and development of novel, potent and selective AR ligands based on the chemical core of flavonoids, the chromone scaffold. The ongoing research has shown that chromone-2-phenylcarboxamide derivatives display a remarkable preference for hA₃AR. In this work we report stimulating results, supported by A_2A/A₃ molecular docking simulations and structure–affinity-relationship (SAR) studies by which N-(4,5-methylthiazol-2-yl)-4-oxo-4H-chromene-2-carboxamide (compound 31) emerged as the most potent and selective compound, displaying an hA₃ K_i of 167 nM and a selectivity ratio of 590 vs. the hA₁ and 480 vs. the hA_2AAR subtypes. The chromone-based ligand was obtained by a simple synthetic approach and will enter in a lead optimization program to enhance its potency and drug-like properties.

---

Introduction

Adenosine is a biomolecule that plays a well-known biological role as a nucleic acid building block and as a component of the cellular ATP energy system. Outside of the cell, adenosine can act as a signalling molecule by interacting with four distinct adenosine receptors (ARs) — designated as A₁, A_2A, A_2B and A₃.^1,2 The ARs are G protein-coupled receptors (GPCRs) that modulate adenylate cyclase activity. A₁ and A₃ receptor subtypes couple preferentially to G_i protein and thereby inhibit adenylyl cyclase leading to a reduced production of cyclic AMP (cAMP) while A_2A and A_2B subtypes stimulate the production of cAMP by coupling to G_s proteins.³ The extensive research dedicated to the physiological and pathological role of extracellular adenosine allowed the connection between the cellular signals promoted by ARs and several types of pathologies, namely neurological, cardiovascular, inflammatory diseases and cancer.^2,4 Nowadays, all four adenosine receptors are recognized as valid targets for drug discovery and development programs.⁵

Adenosine plays a crucial role in the cell progression pathway, either during apoptosis or during cytostatic state. It has been postulated that high levels of extracellular adenosine can have a profound impact on the growth of tumour masses.⁶ Several lines of evidence show that ARs are potential drug targets for cancer treatment, but it seems that the A_2A and A₃ subtypes are the most promising candidates.⁷ In fact, human A₃AR are overexpressed in tumor cells including leukemia, lymphoma, astrocytoma and melanoma tumor cells, and are in turn considered as a target for cancer therapy.⁸ The receptor is also overexpressed in inflammatory cells being a putative target for the treatment of inflammatory disorders, such as rheumatoid arthritis.^8,9 Moreover, A₃AR mediated neuroprotective effects have been demonstrated although further studies are still needed.^8,9

The increasing knowledge about the physiological and pathological role of A₃AR subtype was escorted by the design and development of A₃AR ligands. From these research projects, many chemical structures successfully emerged as potent A₃AR agonists (Fig. 1)^8,10–12 and antagonists (Fig. 2).^8,10 Paradoxically, both agonists and antagonists showed encouraging results namely in anticancer therapy.^13–19

Fig. 1 Examples of A₃AR agonists based on adenosine and xanthine scaffolds.

Fig. 2 Examples A₃AR antagonists based on triazoloquinazoline, dihydropyridine, pyrimidine, quinazoline, xanthine and adenosine scaffolds.

However, despite the intensive efforts in medicinal chemistry to design and synthesize selective A₃AR agonists and antagonists (Fig. 2 and 3) from a diversity of chemical classes few of them have reached the clinical phase.⁸ Central drawbacks found for the lack of success include side effects, which are related to the ubiquity of the receptors, low absorption, short half-life and toxicity.⁸

Fig. 3 A₃AR antagonists based on flavonoid scaffold.

Although benzopyran is considered a privileged structure, few studies were addressed towards its application in the discovery of new AR ligands. In this context, coumarin (2H-chromen-2-one)^20,21 and flavonoid (2-phenyl-4H-chromen-4-one)^22–24 cores were reported as motivating scaffolds. The chemical optimization of A₃AR ligands based on a flavonoid scaffold was reported by Karton et al.²³ in a study that resulted in two selective A₃AR ligands, MRS 1088 and MRS 1067 with K_i values of 741 nM and 561 nM, respectively (Fig. 3). Yet, albeit interesting data has been acquired the compounds' low aqueous solubility, among other explanations, may be one of the major problems that hampered progress of the drug discovery project. To the best of our knowledge, the search for new and more potent A₃AR ligands based on the flavonoid core was discontinued.

Chromones are a group of naturally occurring compounds that are ubiquitous in nature.²⁵ Chemically they are oxygen-containing heterocyclic compounds with a benzoannelated γ-pyrone ring (benzo-γ-pyrone, [(4H)-1-benzopyran-4-one]) having a great potential for chemical decoration allied with a broad spectrum of pharmacological activities.²⁵ These features, along with the structural resemblance with flavonoids, inspired our group to initiate a medicinal chemistry program based on the chromone core in order to study its importance in the development of novel AR ligands.^26,27 Preceding research performed by our group has shown that chromone-2-phenylcarboxamide is a valid scaffold for the design of novel A₃AR ligands. Therefore, the present study was accomplished to acquire new data to validate the significance of the chromone-2-phenylcarboxamide framework. The study was guided by structure–affinity-relationships (SAR) performed with a new library of chromone compounds and complemented by molecular modeling studies.

Results and discussion

Chemistry

The novel chromone phenylcarboxamides were attained following synthetic strategies (Scheme 1) that encompassed a one-pot condensation between a chromone carboxylic acid and aryl or heteroaryl amine. The desired compounds were obtained in moderate to high yields. Strategy 1-A involved the activation of chromone-2-carboxylic (1) or chromone-3-carboxylic (2) acids by (benzotriazol-1-yloxy)tripyrrolidinophosphonium hexafluorophosphate (PyBOP) in the presence of N,N-diisopropylethylamine (DIPEA). After the subsequent addition of the suitable amine compounds 5–12, 15–18, 21, 22, 26–31 and 35 (Tables 1 and 2) were attained. In addition some chromone-2-carboxamides (19, 20, 23, 25, 32, Table 1) were synthesized taking advantage of a microwave-assisted chemistry strategy 1-B. Briefly, the chromone-2-carboxylic acid (1) was activated by reaction with phosphorus oxychloride (POCl₃) with formation of the corresponding acyl chloride, followed by a microwave-assisted amidation reaction that occurs between the obtained acyl chloride and the corresponding amine.^28,29 Additionally, chromone-3-carboxamides 33 and 34 (Table 2) were synthesized by a slightly modified strategy 1-B due to the lability of the starting material and the inherent formation of enamine by-products.³⁰ The reaction of chromone-3-carboxylic acid (2) with POCl₃ led to the formation of the corresponding acyl chloride and the amidation reaction took place at room temperature, without the employment of microwave heating (Scheme 1-C).

Scheme 1 Synthesis of chromone carboxamides. A – PyBOP, DIPEA, DMF, CH₂Cl₂, 0 °C to rt.; B – POCl₃, DMF, 120 °C, MW; C – POCl₃, DMF, rt.

Table 1 Affinity (K_i, nM; 95% confidence intervals in parentheses) of chromones 3–32 in radioligand binding assays at human A₁, A_2A, and A₃ adenosine receptor subtypes

Compound		A1	A2A	A3	Selectivity
	R	Ki (nM)	Ki (nM)	Ki (nM)	A1/A3	A2A/A3
3^26		>100 000	>100 000	14 200 (11 800–17 100)	>7.0	>7.0
4^26		>100 000	>100 000	>100 000	—	—
5		>100 000	>100 000	>100 000	—	—
6		>100 000	>100 000	>100 000	—	—
7		>100 000	>100 000	>100 000	—	—
8		>100 000	>100 000	>100 000	—	—
9		>100 000	>100 000	>100 000	—	—
10		>100 000	>100 000	>100 000	—	—
11		>100 000	>100 000	>100 000	—	—
12		>100 000	>100 000	>100 000	—	—
13^26		>100 000	>100 000	15 800 (12 200–20 400)	>6.3	>6.3
14^26		>100 000	>100 000	9580 (7600–12 100)	>10	>10
15		>100 000	>100 000	>30 000	—	—
16		>100 000	>100 000	>30 000	—	—
17		>100 000	>100 000	21 800 (17 300–27 500)	>4.5	>4.5
18		>100 000	>100 000	>100 000	—	—
19		>100 000	>100 000	>30 000	—	—
20		>100 000	>30 000	>30 000	—	—
21		>100 000	>100 000	>100 000	—	—
22		>100 000	>100 000	>30 000	—	—
23		>100 000	>100 000	20 000 (16 500–24 300)	>5.0	>5.0
24^26		>100 000	>100 000	27 900 (18 300–42 700)	>3.5	>3.5
25		>100 000	>100 000	>100 000	—	—
26		>100 000	>100 000	>100 000	—	—
27		10 200 (7620–13 700)	18 900 (13 300–26 900)	10 700 (8610–13 300)	0.96	1.8
28		12 000 (8450–17 000)	18 500 (11 700–29 100)	>100 000	<0.12	<0.18
29		17 300 (11 400–26 300)	15 100 (9160–24 900)	11 900 (10 900–13 100)	1.4	1.3
30		8180 (5600–11 900)	10 400 (8040–13 300)	1310 (982–1740)	6.3	7.9
31		>100 000	>80 000	167 (151–183)	>590	>480
32		>100 000	>100 000	4780 (2970–7700)	>21	>21

---

Table 2 Affinity (K_i, nM) of chromones 33–35 in radioligand binding assays at human A₁, A_2A, and A₃ adenosine receptor subtypes

Compound		A1	A2A	A3	Selectivity
	R	Ki (nM)	Ki (nM)	Ki (nM)	A1/A3	A2A/A3
33		>30 000	>30 000	32 400 (25 100–41 800)	>0.9	>0.9
34		>10 000	>10 000	>10 000	—	—
35		>10 000	>10 000	10 000	—	—

---

Pharmacology

The affinity of the new potential antagonists for the human adenosine receptor subtypes hA₁, hA_2A, hA₃ (expressed in Chinese hamster ovary (CHO) cells) was determined in radioligand competition experiments.^11,31–34 In this assay, the competition with the following agonist radioligands: (i) [³H]CCPA at hA₁ receptors, (ii) [³H]NECA at hA_2A and [³H]HEMADO at hA₃ receptors was measured. The data were expressed as K_i (dissociation constant). The receptor binding affinities of the synthesised compounds (3–35) are reported in Tables 1 and 2.

Structure–affinity relationship studies

Foregoing research of our group allowed acquiring evidence that validates chromone as a privileged structure for the design of adenosine receptor ligands and chromone-2-phenylcarboxamide (Scheme 2) as a new lead compound for the development of A₃AR ligands.^26,27 The structure–activity relationship (SAR) studies performed so far revealed the importance of the phenylcarboxamide moiety as a key feature for receptor binding. Moreover, the location of the amide function at C-2 or C-3 position of the benzopyran nucleus and the type of amide, preferentially of aromatic type (3, Scheme 2, Table 1), has been found to modulate the affinity/selectivity of the compounds.^26,27 It has also been postulated that the presence of electron withdrawing (Scheme 2, Table 1) or donating groups (Scheme 2, Table 1) on the aromatic exocyclic ring seems to modulate the affinity and selectivity of chromone-2-phenylcarboxamide towards ARs subtypes.^26,27 New derivatives with carefully chosen substituents were synthesised in order to acquire more information about the significance of chromone-2-phenylcarboxamide as a putative A₃AR ligand (Scheme 1, Table 1). Their influence on affinity and selectivity for human A₁, A_2A and A₃ ARs was evaluated in radioligand binding while the interaction with the A_2B AR was analyzed by adenylyl cyclase activity measurement following the procedures previously described by Klotz et al. with minor modifications.^11,31,32 For all the tested compounds no measurable activity for A_2B AR (K_i > 10 000 or 30 000 nM) was detected and therefore these values were omitted in Table 1.

Scheme 2 Preliminary structure–affinity relationships for chromone-2-carboxamide derivatives as A₃AR ligands (see also Table 1).²⁶

Previously, it was postulated that the presence of electron donors or withdrawing groups on the exocyclic aromatic system can modulate the affinity and selectivity of chromones.²⁶ Therefore, it was decided to synthesize a new chromone-2-phenylcarboxamide series that incorporates diverse types of substituents located at different positions of the exocyclic aromatic nucleus. These structural modifications, namely in para position, enabled the development of an extended SAR study. In accordance with the previous data, it was found that the introduction of different electron withdrawing substituents with dissimilar relative strengths from weak (–F, –Br, –I, compounds 5–7), to moderate (–COCH₃, –COOC₂H₅, compounds 8 and 9) and strong (–CN, compound 10) results in all cases in compounds with no measurable affinity for all AR subtypes. Furthermore, a lack of affinity is also observed for a Cl-substituent in para (compound 4), ortho (compound 11) and meta (compound 12) position of the exocyclic phenyl ring. Thus, the binding of chromone-2-phenylcarboxamides to AR is strongly influenced by electronic properties of the substituents and that any type of electron withdrawing groups located at any position of the exocyclic aromatic substituent is not tolerated.

The results obtained from the introduction of electron donating substituents in the exocyclic ring were not as straightforward as it was for the electron withdrawing groups. Preliminary data obtained in previous studies emphasizes the effect of electron donating groups located in para position of the exocyclic phenyl ring of the chromone-2-phenylcarboxamide (3) as they appear to modulate the affinity/selectivity for the A₃AR.²⁶ From these studies one can highlight the data showing the influence of a p-methyl (K_i = 15 800 nM, compound 13) and a p-methoxyl substituent (K_i = 9580 nM compound 14) on A₃AR affinity (Table 1).²⁶ Hence, guided by this information it was decided to study the influence on affinity of the lengthening of the saturated carbon side chain, from one (–CH₃, compound 13), two (–C₂H₅, compound 15) and four (–C₄H₉, compound 16) carbon atoms. A complete loss of affinity towards all ARs subtypes was observed. These results suggest that the affinity of the chromone-2-phenylcarboxamides derivatives to ARs may be circumscribed by steric hindrance factors. Then, to assess the influence of the position of the electron donating substituent on the aromatic exocyclic nucleus, several positional isomers were synthesized. A decrease or absence of affinity was observed for substituents located either in ortho (CH₃, compound 17; OCH₃, compound 19) or meta (CH₃, compound 18; OCH₃, compound 20) positions. Noteworthy data was obtained by the isosteric replacement of the methoxyl group (compound 14) by a thiomethyl (compound 21) as it results in a compound with no affinity to all ARs. The effect of the presence of two electron donating substituents (–CH₃ and/or –OCH₃) located in diverse positions (vicinal or meta position) of the exocyclic phenyl ring (compounds 22–26) on ARs affinity was also inspected. A decrease (2,3-diCH₃, compound 23 and 3,4-diOCH₃, compound 24) or no measurable affinity (3,4-diCH₃, compound 22) for the A₃AR subtype was observed. Interestingly, the introduction of a para methoxyl and an ortho methyl substituents (2-CH₃-4-OCH₃, compound 25) gave rise to an inactive compound. The same tendency has been previously observed with compound 24, with two vicinal methoxy groups,²⁶ and compound 26 that possesses a methylenedioxy substituent on the exocyclic phenyl ring.

Our results reported here corroborate the preliminary SAR data observed in a previous study.²⁶ In order to check the reliability of the SAR studies, and taking into account the importance of the aromaticity of the carboxamide moiety for the affinity towards hA₃AR, a series of heterocyclic carboxamide derivatives were synthesised and evaluated towards all AR subtypes (compounds 27–32). In general, the introduction of a heterocyclic ring, directly linked to the NH of the carboxamide, resulted in active chromone-2-carboxamides with moderate to good binding affinities for human A₁, A_2A and A₃ ARs. However, the replacement of the phenyl (compound 3, K_i = 14 200 nM) by an isosteric pyridine ring (compound 27, K_i = 10 700 nM) or thiazol (compound 29, K_i = 11 900 nM) gave rise to compounds with similar affinity towards A₃ARs but a noticeable loss of selectivity was observed. Moreover, the presence of a CH₂ spacer between the carboxamide and the exo-heterocyclic ring (compound 28) led to a complete loss of affinity towards hA₃AR, although hA₁ (K_i = 12 000 nM) and hA_2A receptor (K_i = 18 500 nM) binding affinities were detected. Inspiring results were obtained with the chromone-2-carboxamides bearing a substituted thiazolyl group (compounds 30–32). The introduction of a methyl substituent at position 5 of the heterocyclic ring (compound 30, K_i = 1310 nM) gave rise to compounds about six- and eightfold selective for hA₃ versus hA₁ (K_i = 8180 nM) and hA_2A (K_i = 10 400 nM) ARs, respectively. The introduction of an additional substituent at position 4 of the thiazole moiety (compounds 31 and 32) led to even more selective and potent chromone-2-carboxamide derivatives. In fact, compound 31 was found to be the most potent and selective compound of all the chromone-2-carboxamide series tested so far, displaying a K_i = 167 nM for A₃AR and selectivity ratios of 590 and 480 compared to A₁ and A_2A ARs, respectively. The loss in potency observed with compound 32 (hA₃ K_i = 4780 nM) can be explained with the introduction of an electron withdrawing group (–COOEt) in the mentioned position.

The compelling results associated to the chromone-2-(thiazol-2-yl)carboxamides prompted us to synthesize their positional C-3 isomers in order to corroborate the current SAR studies. The results for the binding affinity obtained for the chromone-3-(thiazol-2-yl)carboxamides (33–35) are summarized in Table 2.

In analogy to the C-2 isomers no measurable activity for A_2B AR (K_i > 100 000 or 10 000 nM) was detected for the C-3 carboxamides. From the analysed data it is possible to conclude that the position of the (thiazol-2-yl)carboxamide moiety at the chromone core is crucial for the observed affinity. In fact a decrease or a lack of affinity for all ARs was observed for compounds 33, 34 and 35 that are positional isomers of compounds 30, 31, and 32, respectively.

Molecular docking

Molecular docking simulations on hA_2A and hA₃ receptors were performed to study the binding modes of the most active compounds in the series. To carry out this step, the crystal structure 3EML (PDB code)³⁵ for the hA_2A and a homology model for the hA₃ ARs were used (see Experimental section for more details).²¹ The ligands were docked using Glide in the Schrodinger package.³⁶ For validation process the co-crystallized ligands ZM241385 and T4E in the structures 3EML³⁵ (including the role of water in the pocket) and 3UZC³⁷ were docked to the hA_2A protein. A RMSD (root mean square deviation) was obtained between the theoretical poses and the co-crystallized conformations of 0.69 and 1.92 for the ZM241385 and the triazine derivative T4E, respectively.

The docking calculations of chromone compounds showed a high degree of variability regarding possible binding modes in the pocket, in particular in the case of the hA₃ homology model. Selection of the described binding modes was carried out taking into account the energetic parameter E_model and the number of similar poses yielded by the simulations. In both hA_2A and hA₃ the active compounds with a thiazole ring (compounds 29 and 30) shared a similar binding mode inside the AR subtypes. For instance compound 30 oriented the thiazole ring towards the bottom of the cavity whereas the chromone ring is placed closer to the extracellular environment in both receptors. The described binding mode presents some similarity to the co-crystallized antagonist ZM241385 in the hA_2A as it orientates the furan ring towards the bottom of the pocket in the same area as the thiazole ring of compound 30 (see Fig. 4a). In the same manner as the co-crystallized ligand the compound established hydrogen bonds with the residue Asn253 of the hA_2A (see Fig. 4b). Moreover, the hydrophobic/hydrophilic surface generated by the protein was also plotted. The surface revealed favored hydrophobic and hydrophilic areas for ligand interaction that can help to understand the pose of the compounds inside the protein. In fact, for compounds 29 and 30 a good accommodation for the thiazole ring was attained (see Fig. 4c) whereas the introduction of a second 4-methyl substituent in the thiazole (compound 31) led to a loss of its fitting in the hydrophobic area. The loss of this hydrophobic interaction can cause the disruption of the binding mode being an important factor to explain the lack of affinity of compound 31 for the hA_2AAR (K_i > 80 000 nM) compared to the derivatives 29 and 30 (K_i = 15 100 nM and K_i = 10 400 nM, respectively). Binding mode of compound 30 was also inspected by docking studies with water molecules present in the pocket of the crystallized structure 3EML. Compound 30 binding mode was similar to as the one previously described. Thiazole ring pose is oriented towards the bottom of the cavity and established hydrogen bonds with residue Asn253 and water molecules (see Fig. 4d).

Fig. 4 (a) Comparison between the co-crystallized compound (green carbons) in the hA_2A (3EML) and the pose described by docking for compound 30 (grey carbons). (b) Binding mode extracted from docking for compound 30 inside the hA_2A (protein in cartoon style). Hydrogen bonds are represented in yellow. (c) Hydrophobic/hydrophilic surface generated inside the hA_2A (magenta color for hydrophobic and grey color for hydrophilic areas). (d) Binding mode from docking for compound 30 in the hA_2A with water molecules in the pocket.

Thiazole chromone derivatives were also inspected in docking to the hA₃ AR. The pose obtained for compound 31, the most active compound in the series, is very similar to that obtained in the binding mode studies already described for the hA_2A (see Fig. 5a). The analysis of the binding mode showed that the thiazole ring plays a key role in the anchoring of the amide moiety to the residue Asn250 through hydrogen bond interactions. Moreover, the oxygen of the carbonyl group in the chromone nucleus played an important role interacting with residue Gln167 of the second extracellular loop (EL2). In the case of the hA_2A, the residue Gln167 is not present as the protein contains a Leu in this position and so no hydrogen bond with the ligand is detected. This residue can be an operative feature in the ligand binding that can explain the enhanced affinity shown by the thiazole derivatives in the hA₃.

Fig. 5 (a) Pose for compound 31 in the hA₃ determined by molecular docking. Hydrogen bonds are represented in yellow dashed lines and the protein in ribbon style. (b) Hydrophobic/hydrophilic surface calculated in the hA₃ (hydrophobic regions in green and hydrophilic domains in maroon). (c) Binding mode calculated for compound 14 inside the hA₃. Hydrogen bonds with residues Asn250 and Gln167 are represented in yellow color.

On the other hand, hydrophobic/hydrophilic surfaces in the hA₃ pocket showed important differences compared to the hA_2A in the area occupied by the thiazole ring. In the case of the hA₃, the hydrophobic surface is larger thus allowing hydrophobic substitutions at position 4 of the thiazole (see Fig. 5b). This fact could explain the enhancement of hA₃ activity of compound 31 (K_i = 167 nM) with regard to derivatives 29 and 30 (K_i = 11 900 nM and K_i = 1310 nM respectively). Moreover, the distinct hydrophobic distribution detected in both proteins (hA_2A and hA₃) can elucidate the distinct receptor selectivity of compound 31.

The hA₃ also presents a favored hydrophobic surface close to the extracellular environment. As it was previously reported,³⁸ there are some differences in the extracellular loops of the hA_2A and hA₃ that could determine not only the way how ligands bind to the protein but also modulate the entrance of ligands to the interior of the binding cavity. The different nature of some extracellular residues in both proteins, such as the hydrophobic Val169 in the hA₃ and the ionic Glu169 in the hA_2A, could influence the access and accommodation of the ligands. Introduction of a second methyl in the thiazole fragment contributed to an increase of hA₃ activity as the receptor present a deeper hydrophobic surface.

Moreover, the contribution to the interaction energy of the different residues placed in a distance of 4 Å from the ligand in the hA_2A and hA₃ was calculated. The interaction energy (Fig. 6) was calculated as the sum of three scores: Coulomb, van der Waals and hydrogen bonding. Compound 30 inside the hA_2A interacts preferably with residues Phe168, Asn253, Leu249, Ile274 and Met177. Although compound 31 presents a similar interaction outline a lower interaction with residue His250 was noticed. The important role of some residues in ligand binding, such as Phe168, Asn253, Leu249 or Met177 was previously confirmed in mutagenesis studies that examined the effect of mutating residues in hA_2A activity.³⁸ On the other hand, and using the same approach, the most important residues in the interaction between compounds 30–31 and the hA₃ were determined: Phe168, Asn250, Leu246, Ile268 and Gln167. The higher interaction of compound 31 with the key residue Asn250 can also validate its superior hA₃ affinity. It is noteworthy that although hA₃ possesses high identity with the hA_2A there are also differences in some residues of the binding pocket which can condition the binding of the different ligands.

Fig. 6 Residue interaction energy scores between hA_2A and compounds 30, 31 (panel a) and hA₃ and compounds 30, 31 (panel b). Interaction energy is calculated as the sum of Coulomb, van der Waals and hydrogen bonding energies.

The position of the carboxamide in the chromone ring was also found to be important for A₃ adenosine receptor affinity. In fact, compounds with the carboxamide function at position 3, such as compound 34, showed a drastic loss of hA₃ activity. Docking calculations in hA₃ regarding compound 34 demonstrated a hypothetical binding mode with the chromone ring in a similar position as the one described for compound 31. However, compound 34 oriented the carboxamide substituent towards the extracellular environment. The different binding mode could be responsible for the loss of hA₃ activity.

In addition, molecular modeling studies were performed with the chromone-2-phenylcarboxamide with a methoxyl substituent located in para position of the exocyclic aromatic ring (14). It was concluded that this type of chromone presented a different binding mode in the hA₃AR (see Fig. 5c). Compound 14 oriented the chromone scaffold towards the bottom of the cavity and the para-methoxyphenyl moiety towards the extracellular region through the interaction of hydrogen bonds with residues Asn250 and Gln167. Similar binding modes for this type of chromone-2-phenylcarboxamide derivatives were previously reported by our research group.²⁶ So, aromatic heterocyclic systems are considered to be positive bioisosteric replacements to improve hA₃AR affinity of this type of ligands.

From the overall studies one can conclude that the hydrophobic area in the extracellular environment due to the presence of hydrophobic residues, such as Val169 and Leu264, can be considered a significant feature to be accomplished along subsequent compound 31 optimization steps. The change of the type of heterocycle is considered to be an important approach to obtain additional information, improve the affinity and validate the docking studies. Also the presence of other type of electron donating substituents, which must be oriented toward the extracellular area, must be explored.

Scoring function energies for each compound docked to the hA₂A and hA₃ are shown in Table 3. For AUROC calculation, compounds 27–30 were considered to be active in the hA_2A, and compounds 14, 30, 31, 32 and 35 were considered to be active in the hA₃ (K_i ≤ 10 000 nM). Calculations with and without water molecules in the hA_2A pocket were also performed. From the data one can conclude that molecular docking showed limited ability to discriminate active from non-active compounds, even though the results are variable depending on the scoring function (SP or XP) and the adenosine subtype. SP (standard precision) scoring function showed better discriminatory power than the XP (extra-precision) scoring. The area under the ROC curves (AUROC) was also calculated to evaluate the discriminatory ability in the series. SP scoring showed AUROCs of 0.77 and 0.64 in the hA_2A (including water or not in the pocket). The most hA_2A active compound (30) in the series is identified by the SP models in position 2 and 8, respectively. Docking in hA₃ showed more limitations to differentiate between active and non-active compounds. However, hA₃ docking with SP function showed some ability to capture the most active compounds in the series in top positions (positions 9, 10, 13, 15 and 16 for compounds 30, 32, 14, 31 and 35 respectively). An AUROC value of 0.66 was attained when considering as actives the compounds with K_i ≤ 10 000 nM. Docking studies showed some limitations from the energetic point of view in the hA_2A crystal structure and the hA₃ homology model.

Table 3 Scoring function energies for chromone compounds docked to the hA₂A and hA₃ adenosine receptors

	hA2A		hA2A		hA3
	SP (no water)	XP (no water)	SP (water)	XP (water)	SP (no water)	XP (no water)
a SP (standard precision scoring function).b XP (extra-precision scoring function, kcal mol^−1).
AUROC	0.64	0.47	0.77	0.66	0.66	0.36

---

Comp.	SP (no water)^a	XP (no water)^b	SP (water)^a	XP (water)^b	SP (no water)^a	XP (no water)
3	−7.80	−8.33	−7.27	−6.95	−7.78	−7.06
4	−6.61	−8.62	−7.56	−6.79	−8.48	−8.22
5	−7.60	−7.72	−7.75	−6.28	−7.93	−7.56
6	−6.81	−7.82	−7.09	−6.39	−8.30	−7.87
7	−7.16	−7.56	−6.76	−5.98	−7.28	−7.46
8	−7.80	−8.60	−6.10	−5.08	−9.01	−8.19
9	−6.32	−9.07	−6.77	−6.15	−8.68	−7.47
10	−7.53	−7.53	−6.51	−6.40	−8.62	−7.42
11	−8.37	−8.30	−5.46	−5.23	−7.54	−9.49
12	−7.23	−8.66	−7.68	−7.74	−7.93	−7.69
13	−7.69	−8.28	−7.49	−6.33	−7.65	−8.05
14	−7.83	−8.37	−5.13	−5.95	−8.35	−8.10
15	−7.58	−7.86	−6.74	−6.67	−8.30	−7.77
16	−5.94	−9.45	−5.47	−5.24	−8.66	−8.40
17	−6.84	−3.70	−7.12	−6.85	−7.97	−7.49
18	−7.43	−8.58	−7.76	−7.64	−7.82	−8.01
19	−5.82	−8.71	−7.62	−6.87	−8.03	−7.98
20	−7.72	−8.60	−6.14	−5.11	−8.44	−7.94
21	−7.64	−7.32	−6.38	−5.84	−7.86	−7.32
22	−8.31	−8.70	−8.19	−6.62	−8.40	−8.48
23	−6.89	−8.73	−7.18	−6.26	−8.29	−8.13
24	−6.76	−5.43	−5.94	−6.22	−8.62	−8.77
25	−7.11	−8.86	−5.62	−5.36	−8.71	−9.04
26	−7.13	−6.84	−6.40	−7.21	−8.32	−7.82
27	−7.57	−9.77	−6.34	−6.52	−8.57	−6.80
28	−7.54	−8.09	−7.88	−7.28	−8.29	−7.91
29	−7.46	−6.44	−7.94	−6.50	−8.11	−6.60
30	−7.72	−7.88	−8.31	−7.14	−8.45	−6.29
31	−7.50	−7.85	−8.38	−7.21	−8.32	−7.08
32	−6.83	−8.11	−6.10	−7.15	−8.44	−7.76
33	−7.73	−7.89	−7.65	−7.34	−7.55	−7.37
34	−6.57	−8.17	−8.13	−7.58	−7.14	−7.87
35	−7.06	−7.49	−6.25	−7.21	−8.31	−7.96

---

Conclusions

The search for novel compounds with therapeutic value by targeting A₃ ARs is still in its infancy and additional research in the field is still needed. The role of A₃AR agonists and antagonists is still an open issue and the development of new and superior agonists/antagonists that can reach clinical trials remains an emergent topic. Previous research performed by our group has shown that chromone-2-phenylcarboxamide is a valid scaffold for tracking novel ligands with affinity for A₃AR. Thus, the study presented here provides new insights into the SAR regarding the chromone-2-phenylcarboxamide framework. This input, supported by A_2A/A₃ molecular docking simulations and structure–affinity-relationship (SAR) studies, lead to the identification of the crucial role for A₃AR affinity of a heteroamide substituent located at C-2 of the pyrone ring. The most remarkable chromone-2-carboxamides with affinity towards A₃ARs were compounds 30, 31 and 32 with K_i values of 1310 nM, 167 nM and 4780 nM, respectively. Compound 31, displaying an A₃ K_i of 167 nM and a selectivity ratio of 590 vs. the A₁ and 480 vs. the A_2AAR subtypes, stood out as the best chromone derivative tested so far, showing an enhanced A₃AR affinity and selectivity compared to the flavonoids presently known. Moreover, compound 31 presents appealing drug-like properties (e.g. c log P 2.44 and molecular polar surface area of 72.20) and no violations of the Lipinski rule.

In addition, the present study supports the belief that chromone due to its synthetic accessibility and decoration capability, is a privileged structure for the design and development of libraries in the discovery of novel and innovative AR ligands. Finally, the receptor-driven molecular modeling studies have provided valuable information on the molecular interactions responsible for selective high affinity binding to the A₃ adenosine receptor of the most promising chromones. In this respect the data are instrumental for future optimization of potency and drug-like properties of chromone-2-(hetero)carboxamides.

Experimental section

Chemistry

Chromone-2-carboxylic acid, (benzotriazol-1-yloxy)tripyrrolidinophosphonium hexafluorophosphate (PyBOP), N,N-diisopropylethylamine (DIPEA), dimethylformamide (DMF), phosphoryl trichloride (POCl₃), arylamines and heteroamines were purchased from Sigma-Aldrich Química S.A. (Sintra, Portugal). All other reagents and solvents were pro analysis grade and were acquired from Merck (Lisbon, Portugal), Panreac (Lisbon, Portugal) and Sigma Aldrich (Sintra, Portugal), used without additional purification.

Thin-layer chromatography (TLC) was carried out on pre-coated silica gel 60 F254 (Merck) with layer thickness of 0.2 mm. The spots were visualized under UV detection (254 and 366 nm) and iodine vapour. Normal-phase column chromatography was performed using silica gel 60 0.2–0.5 or 0.040–0.063 mm (Merck).

Following the workup and after extraction, the organic phases were dried over Na₂SO₄. Solutions were decolorized with activated charcoal, when necessary. Solvents were evaporated in a Buchi Rotavapor.

The purity of the final products (>97% purity) was verified by high-performance liquid chromatography (HPLC) equipped with a UV detector. Chromatograms were obtained in an HPLC/DAD system, a Jasco instrument (pumps model 880-PU and solvent mixing model 880-30, Tokyo, Japan), equipped with a commercially prepacked RP-18 analytical column (250 mm × 4.6 mm, 5 μm, Macherey-Nagel, Duren, Germany), and UV detection (Jasco model 875-UV) at the maximum wavelength of 254 nm. The mobile phase consisted of a methanol/water or acetonitrile/water (gradient mode, room temperature) at a flow rate of 1 mL min⁻¹. The chromatographic data was processed in a Compaq computer, fitted with CSW 1.7 software (DataApex, Czech Republic).

Apparatus. The main paragraph text ¹H NMR and ¹³C NMR data were acquired, at room temperature, on a Brüker AMX 400 spectrometer operating at 400.15 and 101.0 MHz, respectively. Chemical shifts are expressed in δ (ppm) values relative to tetramethylsilane (TMS) as internal reference; coupling constants (J) are given in Hz. Electron impact mass spectra (EI-MS) were carried out on a VG AutoSpec instrument; the data are reported as m/z (% of relative intensity of the most important fragments).

Synthesis. The chromone derivatives were synthesised by three synthetic approaches previously developed by our group.^28,29

Method A. To a solution of the chromone-2-carboxylic acid or chromone-3-carboxylic (1.1 mmol) in DMF (2.5 mL) at 4 °C it was added the reagent N,N-diisopropylethylamine (1.1 mmol) and a solution of PyBOP (1 mmol) in CH₂Cl₂ (2.5 mL). The mixture was kept in an ice bath and stirred for half hour. After this period the amine with the desired (hetero)aromatic pattern was added and the mixture was allowed to warm up to room temperature. The reaction was kept under stirring for 4 hours. The working-up of the crude material enclosed a liquid–liquid extraction with CH₂Cl₂ followed by flash chromatography (CH₂Cl₂/MeOH or EtOAc/n-hexane) and final purification by recrystallization (EtOAc/n-hexane).
N-Phenyl-4-oxo-4H-chromene-2-carboxamide (3). The synthetic procedure and structural elucidation data was described elsewhere.³⁹
N-(4-Chlorophenyl)-4-oxo-4H-chromene-2-carboxamide (4). The synthetic procedure and structural elucidation data was described elsewhere.⁴⁰
N-(4-Fluorphenyl)-4-oxo-4H-chromene-2-carboxamide (5). Yield 75%. ¹H NMR (DMSO-D6): δ = 6.98 (1H, s, H(3)), 7.28 (2H, dd, J = 8.9, H(3′), H(5′)), 7.57 (1H, ddd, J = 8.0, 7.0, 1.0, H(6)), 7.81–7.85 (3H, m, H(8), H(2′), H(6′)), 7.94 (1H, ddd, J = 1.6, 7.0, 8.6, H(7)), 8.09 (1H, dd, J = 1.6, 8.0, H(5)), 10.81 (1H, s). ¹³C NMR (DMSO-D6): δ = 112.5 (C(3)), 116.9 (d, J = 22.5, C(3′), C(5′)), 120.4 (C(8)), 124.5 (d, J = 8.2, C(2′), C(6′)), 125.1 (C(4a)), 126.3 (C(6)), 127.5 (C(5)), 135.2 (C(7)), 136.5 (C(1′)), 156.5 (C(8a)), 157.0 (CONH), 159.1 (C(2)), 160.4 (d, J = 241.9, C(4′)) 178.7 (C(4)). EM/IE m/z: 284 (M + 1, 21), 283 (M⁺˙, 100), 282 (83), 254 (29), 145 (23), 89 (58).
N-(4-Bromophenyl)-4-oxo-4H-chromene-2-carboxamide (6). The synthetic procedure and structural elucidation data was described elsewhere.²⁸
N-(4-Iodophenyl)-4-oxo-4H-chromene-2-carboxamide (7). Yield 95% ¹H NMR (DMSO-D6): δ = 6.97 (1H, s, H (3)), 7.56 (1H, ddd, J = 1.0, 7.1, 8.0, H(6)), 7.65 (2H, d, J = 8.8, H(3′), H(5′)), 7.77 (2H, d, J = 8.8, H(2′), H(6′)), 7.84 (1H, d, J = 8.5, H(8)), 7.93 (1H, ddd, J = 1.6, 7.0, 8.6, H(7)), 8.08 (1H, dd, J = 1.4, 8.0, H(5)), 10.82 (1H, s, NH). ¹³C NMR (DMSO-D6): δ = 90.5 (C(4′)), 112.6 (C(3)), 120.4 (C(8)), 124.5 (C(2′), C(6′)), 124.6 (C(4a)), 126.3 (C(6)), 127.5 (C(5)), 136.5 (C(7)), 138.8 (C(1′)), 138.9 (C(3′), C(5′)), 156.5 (C(8a)), 156.8 (C(2)), 159.2 (NH), 178.6 (C(4)). MS/EI m/z: 392 (22), 391 (M⁺˙, 100), 390 (47), 362 (16), 145 (13), 89 (45), 86 (12), 84 (18), 70 (10).
N-(4-Acetylphenyl)-4-oxo-4H-chromene-2-carboxamide (8). Yield 75% ¹H NMR (CDCl₃): δ = 2.66 (3H, s, CH₃), 7.38 (1H, s, H(3)), 7.55 (1H, app t, J = 7.8, H(6)), 7.78 (1H, d, J = 8.6, H (8)), 7.85 (1H, ddd, J = 1.5, 7.8, 8.0, H(7)), 7.95 (2H, d, J = 8.7, H(3′), H(5′)), 8.06 (2H, d, J = 8.7, H(2′), H(6′)), 8.26 (1H, dd, J = 7.8, 1.5, H(5)), 9.86 (1H, s, CONH). ¹³C NMR (CDCl₃): δ = 26.4 (CH₃), 112.4 (C(3)), 118.4 (C(8)), 120.2 (C(3′), C(5′)), 124.1 (C(4a)), 125.9 (C(5)), 126.3 (C(6)), 129.7 (C(2′), C(6′)) 133.7 (C(1′)), 135.1 (C(7)), 141.3 (C(4′)), 154.9 (C(8a)), 155.5 (C(2)), 157.8 (CONH), 178.7 (C(4)), 197.8 (C̲OCH₃). MS/EI m/z: 307 (M⁺˙, 87), 292 (100), 146 (19), 121 (37), 89 (56).
N-(4-Ethoxycarbonylphenyl)-4-oxo-4H-chromene-2-carboxamide (9). Yield: 82%. ¹H NMR (CDCl₃): δ = 1.41 (3H, t, J = 7.1, CH₂CH̲₃), 4.39 (2H, d, J = 7.1, CH̲₂CH₃), 7.28 (1H, s, H(3)), 7.50 (1H, ddd, J = 1.0, 7.1, 8.0, H(6)), 7.61 (1H, d, J = 8.5, H(8)), 7.79 (1H, ddd, J = 1.6, 7.0, 8.6, H(7)), 7.82 (2H, d, J = 8.8, H(3′), H(5′), 8.11 (2H, d, J = 8.8, H(2′)), H(6′)), 8.25 (1H, dd, J = 1.4, 8.0, H(5)), 8.74 (1H, s, CONH). ¹³C NMR (CDCl₃): δ = 14.3 (CH₂C̲H₃), 61.1 (C̲H₂CH₃), 112.9 (C(3)), 118.0 (C(8)), 119.5 (C(2′), C(6′)), 123.9 (C(4a)), 124.3 (C(4′)), 126.3 (C(6)), 127.4 (C(5)), 130.9 (C(3′), C(5′)), 134.8 (C(7)), 140.3 (C(1′)), 154.1 (C(8a)), 155.1 (C(2)), 157.1 (CONH), 177.0 (C(4)). EM/IE m/z: 338 (37), 337 (94), 336 (M⁺˙, 100), 320 (24), 309 (13), 308 (53), 293 (18), 292 (81), 280 (14), 265 (17), 264 (18), 146 (17), 145 (18), 121 (29), 120 (12), 101 (12), 89 (63).
N-(4-Cyanophenyl)-4-oxo-4H-chromene-2-carboxamide (10). Yield: 71%. ¹H NMR (DMSO-D6): δ = 7.01 (1H, s, H(3)), 7.56 (1H, ddd, J = 1.2, 7.4, 7.8, H(6)), 7.82–7.87 (3H, m, H(3′), H(5′), H(8)), 7.93 (1H, ddd, J = 1.6, 7.6, 8.0, H(7)), 7.98 (2H, d, J = 8.8, H(2′), H(6′)), 8.08 (1H, dd, J = 1.2, 8.0, H(5)), 11.09 (1H, s, NH). ¹³C NMR (DMSO-D6): δ = 105.3 (C(4′)), 110.6 (C(3)), 110.7 (CN), 118.4 (C(2′), C(6′)), 121.0 (C(8)), 123.1 (C(4a)), 124.3 (C(6)), 125.5 (C(5)), 132.5 (C(3′), C(5′)), 134.4 (C(7)), 143.1 (C(1′)), 154.6 (C(8a)), 155.6 (C(2)), 158.2 (NH), 176.7 (C(4)). EM/IE m/z: 290 (M⁺˙, 91), 289 (100), 261 (25), 173 (26), 145 (39), 89 (88), 69 (20).
N-(2-Chlorophenyl)-4-oxo-4H-chromene-2-carboxamide (11). The synthetic procedure and structural elucidation data was described elsewhere.⁴¹
N-(3-Chlorophenyl)-4-oxo-4H-chromene-2-carboxamide (12). The synthetic procedure and structural elucidation data was described elsewhere.⁴¹
N-(p-Tolyl)-4-oxo-4H-chromene-2-carboxamide (13). The synthetic procedure and structural elucidation data was described elsewhere.²⁶
N-(4-Methoxyphenyl)-4-oxo-4H-chromene-2-carboxamide (14). The synthetic procedure and structural elucidation data was described elsewhere.²⁶
N-(4-Ethylphenyl)-4-oxo-4H-chromene-2-carboxamide (15). Yield: 67%. ¹H NMR (CDCl₃): δ = 1.24 (3H, t, J = 7.6, CH₃), 2.65 (2H, J = 7.6, CH₂), 7.23 (2H, d, J = 8.5, H(3′), H(5′)), 7.25 (1H, s, H(3)), 7.47 (1H, ddd, J = 1.0, 7.2, 8.0, H(6)), 7.57–7.65 (3H, m, H(8), H(2′), H(6′)), 7.76 (1H, ddd, J = 1.5, 7.2, 8.7, H(7)), 8.23 (1H, dd, J = 1.5, 8.0, H(5)), 8.55 (1H, s, NH). ¹³C NMR (CDCl₃): δ = 15.5 (CH₃), 28.4 (CH₂), 112.3 (C(3)), 118.1 (C(8)), 120.6 (C(2′), C(6′)), 124.4 (C(4a)), 126.1 (C(6)), 126.2 (C(5)), 128.6 (C(3′), C(5′)), 134.0 (C(4′)), 134.7 (C(7)), 142.0 (C(1′)), 154.7 (C(8a)), 155.2 (C(2)), 156.8 (CONH), 178.0 (C(4)). MS/EI m/z: 293 (M⁺˙, 100), 278 (20), 264 (20), 120 (31), 107 (23), 89 (99), 77 (25).
N-(4-Butylphenyl)-4-oxo-4H-chromene-2-carboxamide (16). Yield: 69%. ¹H NMR (CDCl₃): δ = 0.93 (3H, t, J = 7.3, CH₃), 1.31–1.41 (2H, m, CH₂), 1.59–1.62 (2H, m, CH₂), 2.61 (2H, t, J = 7.7, CH₂). 7.21 (2H, d, J = 8.5, H(3′), H(5′)), 7.26 (1H, s, H(3)), 7.48 (1H, ddd, J = 1.0, 7.2, 8.1, H(6)), 7.57–7.63 (3H, m, H(8), H(2′), H(6′)), 7.77 (1H, ddd, J = 1.6, 7.2, 8.5, H(7)), 8.24 (1H, dd, J = 1.6, 8.1, H(5)), 8.55 (1H, s, NH). ¹³C NMR (CDCl₃): δ = 13.9 (CH₃), 22.3 (CH₂), 33.5 (CH₂) 35.1 (CH₂), 112.5 (C(3)), 118.1 (C(8)), 120.5 (C(2′), C(6′)), 124.4 (C(4a)), 126.1 (C(6)), 126.2 (C(5)), 129.2 (C(3′), C(5′)), 133.9 (C(4′)), 140.6 (C(1′)), 134.4 (C(7)), 154.7 (C(8a)), 155.2 (C(2)), 156.8 (CONH), 178.1 (C(4)). MS/EI m/z: 321 (M⁺˙, 81), 278 (100), 107 (68), 89 (88).
N-(2-Methylphenyl)-4-oxo-4H-chromene-2-carboxamide (17). The synthetic procedure and structural elucidation data was described elsewhere.⁴¹
N-(3-Methylphenyl)-4-oxo-4H-chromene-2-carboxamide (18). The synthetic procedure and structural elucidation data was described elsewhere.⁴¹
N-(4-(Methylthio)phenyl)-4-oxo-4H-chromene-2-carboxamide (21). The synthetic procedure and structural elucidation data was described elsewhere.⁴²
N-(3,4-Dimethylphenyl)-4-oxo-4H-chromene-2-carboxamide (22). Yield: 75% ¹H NMR (DMSO-D6): δ = 2.21 (3H, s, CH₃), 2.24 (3H, s, CH₃), 6.95 (1H, s, H(3)), 7.16 (1H, d, J = 8.0, H(5′)), 7.52–7.54 (1H, m, H(2′)), 7.56–7.58 (2H, m, H(6′), H(6)), 7.84 (1H, d, J = 8.0, H(8)), 7.93 (1H, ddd, J = 1.2, 6.8, 8.0, H(7)), 8.08 (1H, dd, J = 1.2, 8.0, H(5)), 10.60 (1H, s, NH). ¹³C NMR (DMSO-D6): δ = 20.3 (CH₃), 21.0 (CH₃), 112.3 (C(3)), 119.9 (C(8)), 121.4 (C(2′)), 123.5 (C(6)), 125.1 (C(4a)), 126.3 (C(6′)), 127.5 (C(5)), 131.0 (C(3′)), 134.3 (C(4′), C(5′)), 136.4 (C(7)), 137.9 (C(1′)), 156.5 (C(8a)), 157.2 (C(2)), 158.8 (CONH), 178.7 (C(4)). MS/EI m/z: 293 (M⁺˙, 100), 292 (90), 276 (25), 264 (35), 173 (22), 120 (45), 89 (49), 77 (24).
N-(3,4-Dimethoxyphenyl)-4-oxo-4H-1-benzopyran-2-carboxamide (24). The synthetic procedure and structural elucidation data was described elsewhere.²⁶
N-(Benzodioxol-5-yl)-4-oxo-4H-chromene-2-carboxamide (26). Yield: 62%. ¹H NMR (DMSO-D6): δ = 6.05 (2H, s, CH₂), 6.95 (1H, s, H(3)), 6.97 (1H, d, J = 8.4, H(5′)), 7.24 (1H, dd, J = 2.4, 8.4, H(6′)), 7.43 (1H, d, J = 2.4, H(2′)), 7.57 (1H, ddd, J = 1.0, 7.6, 9.2, H(6)), 7.84 (1H, dd, J = 0.8, 8.4, H(8)), 7.93 (1H, app dt, J = 1.6, 7.8, H(7)), 8.08 (1H, dd, J = 1.6, 8.0, H(5)), 10.65 (1H, s, NH). ¹³C NMR (DMSO-D6): δ = 100.7 (CH₂), 102.4 (C(2′)), 107.5 (C(5′)), 110.4 (C(6′)), 113.8 (C(3)), 118.4 (C(8)), 123.1 (C(4a)), 124.4 (C(6)), 125.5 (C(5)), 131.1 (C(1′)), 134.5 (C(7)), 143.7 (C(4′)), 146.5 (C(3′)), 154.6 (C(8a)), 155.1 (C(2)), 156.8 (NH), 176.7 (C(4)). MS/EI m/z: 310 (M + 1, 25), 309 (M⁺˙, 100), 292 (17), 280 (18), 189 (30), 136 (65), 89 (33).
N-(Pyridin-2-yl)-4-oxo-4H-chromene-2-carboxamide (27). Yield: 31%. ¹H NMR (CDCl₃): δ = 7.19 (1H, ddd, J = 0.9, 5.0, 7.4, H(4′)), 7.30 (1H, s, H(3)), 7.50 (1H, ddd, J = 1.0, 7.2, 8.1, H(6)), 7.63 (1H, dd, J = 0.6, 8.5, H(6′)), 7.89–7.72 (2H, m, H(5′), H(8)), 8.25 (1H, dd, J = 1.4, 8.0, H(3′)), 8.45–8.32 (2H, m, H(5), H(7)), 9.31 (1H, s, NH). ¹³C NMR (CDCl₃): δ = 113.0 (C(3)), 114.8 (C(6′)), 118.2 (C(8)), 121.1 (C(4′)), 124.4 (C(4a)), 126.2 (C(6)), 126.3 (C(5)), 134.9 (C(7)), 139.0 (C(5′)), 148.1 (C(3′)), 150.1 (C(1′)), 154.0 (C(8a)), 155.2 (C(2)), 157.3 (CONH), 177.9 (C(4)). MS/EI m/z: 267 (22), 266 (M⁺˙, 99), 238 (58), 237 (66), 210 (83), 89 (100).
N-(Furan-2-ylmethyl)-4-oxo-4H-chromene-2-carboxamide (28). Yield: 35%. ¹H NMR (CDCl₃): δ = 4.68 (2H, d, J = 5.7, CH₂), 6.42–6.32 (2H, m, H(4′), H(5′)), 7.19 (1H, s, H(3)), 7.23 (1H, bs, NH), 7.42 (1H, dd, J = 0.9, 1.7, H(3′)), 7.48–7.43 (1H, m, H(6)), 7.52 (1H, dd, J = 0.5, 8.5, H(8)), 7.73 (1H, ddd, J = 1.7, 7.2, 8.7, H(7)), 8.22 (1H, dd, J = 1.4, 8.0, H(5)). ¹³C NMR (CDCl₃): δ = 36.9 (CH₂), 108.6 (C(5′)), 110.8 (C(4′)), 112.5 (C(3)), 118.2 (C(8)), 124.4 (C(4a)), 126.1 (C(6)), 126.2 (C(5)), 134.7 (C(7)), 142.8 (C(3′)), 150.1 (C(1′)), 154.6 (C(8a)), 156.3 (C(2)), 159.2 (CONH), 178.2 (C(4)). MS/EI m/z: 270 (22), 269 (M⁺˙, 100), 297 (11), 146 (21), 145 (16).
N-(Thiazol-2-yl)-4-oxo-4H-chromene-2-carboxamide (29). The synthetic procedure and structural elucidation data was described elsewhere.⁴³
N-(5-Methylthiazol-2-yl)-4-oxo-4H-chromene-2-carboxamide (30). The synthetic procedure and structural elucidation data was described elsewhere.⁴³
N-(4,5-Dimethylthiazol-2-yl)-4-oxo-4H-chromene-2-carboxamide (31). The synthetic procedure and structural elucidation data was described elsewhere.⁴³
Ethyl 4-methyl-2-(4-oxo-4H-chromene-3-carboxamido)thiazole-5-carboxylate (35). Yield: 15%. ¹H NMR (CDCl₃): δ = 1.37 (3H, t, J = 7.1, OCH₂CH̲₃), 2.69 (3H, s, CH₃), 4.33 (2H, q, J = 7.1, OCH̲₂CH₃), 7.57 (1H, m, H(6)), 7.61 (1H, d, J = 8.4, H(8)), 7.82 (1H, ddd, J = 1.6, 7.3, 8.6, H(7)), 8.37 (1H, dd, J = 1.6, 8.0, H(5)), 9.07 (1H, s, H(2)), 12.65 (1H, s, NH). ¹³C NMR (CDCl₃): δ = 14.4 (OCH₂C̲H₃), 17.3 (CH₃), 60.8 (OC̲H₂), 114.1 (C(3)), 116.6 (C(3′)), 118.5 (C(8)), 123.9 (C(4a)), 126.6 (C(6)), 126.9 (C(5)), 135.3 (C(7)), 156.1 (C(8a)), 157.1 (C(4′)), 158.4 (C(1′)), 161.0 (CONH), 162.8 (C̲OOCH₂), 163.3 (C(2)), 176.5 (C(4)). MS/EI m/z: 359 (M + 1, 64), 358 (M⁺˙, 98), 313 (37), 174 (73), 173 (93), 121 (100).

Method B. To a solution of chromone-2-carboxylic acid (1.1 mmol) in DMF (1.5 mL), POCl₃ (1 mmol) was added. The mixture was stirred at room temperature for 30 min, with the formation in situ of the corresponding acyl chloride. Then the aromatic amine or the heteroaromatic amine was added. The system was heated at 120 °C for 5 min in a microwave apparatus. Subsequently, the mixture was poured in a beaker and water was added. The formed solid was filtered and purified by recrystallization (CH₂Cl₂/n-hexane) or flash chromatography (AcOEt/n-hexane).
N-(2-Methoxyphenyl)-4-oxo-4H-chromene-2-carboxamide (19). The synthetic procedure and structural elucidation data was described elsewhere.²⁸
N-(3-Methoxyphenyl)-4-oxo-4H-chromene-2-carboxamide (20). The synthetic procedure and structural elucidation data was described elsewhere.²⁸
N-(2,3-Dimethylphenyl)-4-oxo-4H-chromene-2-carboxamide (23). Yield: 68%. ¹H NMR (CDCl₃): δ = 2.15 (3H, s, CH₃), 2.32 (3H, s, CH₃), 6.96 (1H, s, H(3)), 7.25–7.12 (3H, m, H(4′), H(5′), H(6′)), 7.58 (1H, ddd, J = 1.08, 7.18, 8.07, H(6)), 7.83 (1H, dd, J = 0.58, 8.41, H(8)), 7.94 (1H, ddd, J = 1.70, 7.11, 8.68, H(7)), 8.10 (1H, dd, J = 1.41, 7.99, H(5)), 10.60 (1H, s, NH). ¹³C NMR (CDCl₃): δ = 14.2 (CH₃), 20.0 (CH₃), 110.8 (C(3)), 118.9 (C(8)), 123.6 (C(4a)), 124.4 (C(6′)), 124.9 (C(6)), 125.4 (C(5′)), 126.0 (C(5)), 128.2 (C(4′)), 132.7 (C(2′)), 134.6 (C(1′)), 135.0 (C(7)), 137.0 (C(3′)), 155.1 (C(8a)), 155.7 (C(2)), 157.9 (CONH), 177.2 (C(4)). MS/EI m/z: 293 (M⁺˙, 40), 292 (100), 264 (13), 120 (21).
N-(4-Methoxy-2-methylphenyl)-4-oxo-4H-chromene-2-carboxamide (25). Yield: 64%. ¹H NMR (CDCl₃): δ = 2.35 (3H, s, CH₃), 3.79 (3H, s, OCH₃), 6.81–6.75 (2H, m, H(3′), H(5′)), 7.22 (1H, s, H(3)), 7.47 (1H, app t, J = 8.0, 8.0, H(6)), 7.56 (1H, d, J = 8.1, H(8)), 7.72 (1H, d, J = 8.4, H(6′)), 7.76 (1H, ddd, J = 1.7, 7.2, 8.7, H(7)), 8.23 (1H, dd, J = 8.0, 1.6, H(5)), 8.43 (1H, s, NH). ¹³C NMR (CDCl₃): δ = 18.0 (CH₃), 55.4 (OCH₃), 111.8 (C(3)), 112.4 (C(6′)), 116.2 (C(5′)), 118.0 (C(8)), 124.3 (C(4a)), 125.0 (C(6)), 126.1 (C(3′)), 126.1 (C(5)), 127.0 (C(1′)), 132.2 (C(2′)), 134.6 (C(7)), 154.8 (C(8a)), 155.1 (C(2)), 157.1 (C(4′)), 157.8 (CONH), 178.0 (C(4)). MS/EI m/z: 310 (23), 309 (M⁺˙, 100), 195 (26), 136 (72).
Ethyl 4-methyl-2-(4-oxo-4H-chromene-2-carboxamido)thiazole-5-carboxylate (32). The synthetic procedure and structural elucidation data was described elsewhere.⁴³

Method C. To a solution of chromone-3-carboxylic acid (500 mg, 2.6 mmol) in DMF (4 mL), POCl₃ (241 μL, 2.6 mmol) was added. The mixture was stirred at room temperature for 30 min, with the formation in situ of the corresponding acyl chloride. Then the heteroaromatic amine was added. After 1–5 hours, the mixture was diluted with dichloromethane (20 mL), washed with H₂O (2 × 10 mL) and with saturated NaHCO₃ solution (2 × 10 mL). The organic phase was dried with Na₂SO₄, filtered, and concentrated under reduced pressure. The residue was purified by flash chromatography or crystallization.
N-(5-Methylthiazol-2-yl)-4-oxo-4H-chromene-3-carboxamide (33). Yield: 15%. ¹H NMR (CDCl₃): δ = 2.43 (3H, d, J = 1.2, CH₃), 7.19–7.12 (1H, m, H(4′)), 7.54 (1H, ddd, J = 1.0, 7.2, 8.1, H(6)), 7.59 (1H, dd, J = 0.6, 8.5, H(8)), 7.80 (1H, ddd, J = 1.7, 7.2, 8.7, H(7)), 8.35 (1H, dd, J = 1.5, 8.0, H(5)), 9.05 (1H, s, H(2)), 12.43 (1H, s, NH). ¹³C NMR (CDCl₃): δ = 11.7 (CH₃), 114.6 (C(3)), 118.6 (C(8)), 124.1 (C(4a)), 126.7 (C(6)), 126.9 (C(5)), 128.4 (C(3′)), 135.2 (C(7)), 135.4 (C(4′)), 155.5 (C(1′)), 156.2 (C(8a)), 160.3 (CONH), 163.0 (C(2)), 176.7 (C(4)). MS/EI m/z: 286 (M⁺˙, 91), 174 (30), 173 (100), 121 (61).
N-(4,5-Dimethylthiazol-2-yl)-4-oxo-4H-chromene-3-carboxamide (34). Yield: 15%. ¹H NMR (CDCl₃): δ = 2.28 (1H, s, CH₃), 2.32 (1H, s, CH₃), 7.54 (1H, ddd, J = 1.0, 7.2, 8.1, H(6)), 7.59 (1H, dd, J = 0.6, 8.5, H(8)), 7.79 (1H, ddd, J = 1.7, 7.2, 8.7, H(7)), 8.35 (1H, dd, J = 1.6, 8.0, H(5)), 9.04 (1H, s, H(2)), 12.37 (1H, s, NH). ¹³C NMR (CDCl₃): δ = 11.0 (CH₃), 14.7 (CH₃), 114.7 (C(3)), 118.6 (C(8)), 120.9 (C(3′)), 124.1 (C(4a)), 126.7 (C(6)), 126.8 (C(5)), 135.2 (C(5)), 143.0 (C(4′)), 152.9 (C(1′)), 156.2 (C(8a)), 160.3 (CONH), 163.0 (C(2)), 176.7 (C(4)). MS/EI m/z: 302 (M + 2), 300 (M⁺˙, 100), 199 (45), 139 (73).

Pharmacology

CHO membrane preparation. All methods involving membrane preparation for radioligand binding experiments followed the procedures as described earlier.³² Membranes for radioligand binding were prepared from CHO cells stably transfected with the human adenosine receptor subtypes A₁, A_2A, and A₃ in a two-step procedure. In the first low-speed step (1000g for 10 min at 4 °C), the cell fragments and nuclei were removed. After that, the crude membrane fraction was sedimented from the supernatant at 100 000×g for 30 min. The membrane pellet was then resuspended in the specific buffer used for the respective binding experiments, frozen in liquid nitrogen, and stored at −80 °C. For the measurement of the adenylyl cyclase activity in A_2B receptor expressing CHO cells, only one step of centrifugation was used in which the homogenate was sedimented for 30 min at 54 000×g. The resulting crude membrane pellet was resuspended in 50 mM Tris/HCl, pH 7.4 and immediately used for the adenylyl cyclase assay.

Human cloned A₁, A_2A, A₃ adenosine receptor binding assay. All binding of [³H]CCPA to CHO cells transfected with the human recombinant A₁ adenosine receptor was performed as previously described.³² Competition experiments were performed for 3 h at 25 °C in 200 μL of buffer containing 1 nM [³H] CCPA, 0.2 U mL⁻¹ adenosine deaminase, 20 μg of membrane protein in 50 mM Tris/HCl, pH 7.4 and tested compound in different concentrations. Nonspecific binding was determined in the presence of 1 mM theophylline. Binding of [³H]NECA to CHO cells transfected with the human recombinant A_2A adenosine receptors was performed following the conditions as that described for the A₁ receptor binding.³² In the competition experiments, samples containing a protein concentration of 50 μg, 10 nM [³H]NECA and tested compound in different concentrations were incubated for 3 h at 25 °C. Nonspecific binding was determined in the presence of 100 μM R-PIA (R-N⁶-phenylisopropyladenosine). Binding of [³H]HEMADO to CHO cells transfected with the human recombinant A₃ adenosine receptors was carried out as previously described.³² The competition experiments were performed for 3 h at 25 °C in buffer solution containing 1 nM [³H]HEMADO, 20 μg membrane protein in 50 mM Tris–HCl, 1 mM ethylenediaminotetraacetate (EDTA), 10 mM MgCl₂, pH 8.25 and tested compound in different concentrations. Nonspecific binding was determined in the presence of 100 μM R-PIA.

Adenylyl cyclase activity. Due to the lack of a suitable radioligand for hA_2B receptor in binding assay, the potency of antagonists at A_2B receptor (expressed on CHO cells) was determined in adenylyl cyclase experiments instead. The procedure was carried out as described previously with minor modifications.^31,32 Membranes were incubated with about 150 000 cpm of [α-³²P]ATP for 20 min in the incubation mixture as described, without EGTA and NaCl. For agonists, the EC₅₀ values for the stimulation of adenylyl cyclase were calculated with the Hill equation. Hill coefficients in all experiments were near unity. IC₅₀ values for concentration-dependent inhibition of NECA-stimulated adenylyl cyclase caused by antagonists were calculated accordingly. Dissociation constants (K_i) for antagonist were then calculated from the Cheng and Prusoff equation.³³

Molecular modeling

A homology model of hA₃ was constructed using MOE software.⁴⁴ Detailed description of the homology models are provided in previous publications of our research group.²¹ The crystal structure of the hA_2A adenosine receptor (PDB: 3EML)³⁵ was used as a template in the modeling. Alignment between both proteins was carried out considering highly conserved residues of the transmembrane helices. The same alignment described by Katritch et al. was followed.⁴⁵ The quality of the geometry in the protein homology models was assessed using the Protein Geometry module in MOE that takes into account phi–psi dihedral plots, bond lengths, bond angles, dihedrals, side chain rotamers and non-bonded contacts. The presence of conserved disulfide bridges between adenosine receptors was checked, such as the disulfide bridge in hA₃ between Cys166 (EL2) and Cys83 (TM3). An optimization of the protein pocket using the Induce Fit Docking workflow was carried out.³⁶ This procedure included docking of high affinity ligands through Glide and protein active site optimization with Prime. Different hA₃ models were tested, in a similar way as Katritch et al.,⁴³ according to their potential to discriminate ligands from decoys and hA₃ high affinity compounds from other adenosine sub-types ligands. Areas under the ROC curves (AUROCs) were calculated showing results for some models greater than 0.80. The best hA₃ models were selected to run molecular docking calculations.

Molecular docking of adenosine receptors ligands (hA_2A and hA₃)

Molecular docking simulations were carried out using the Schrödinger package.³⁶ The database of chromone derivatives was pre-processed with the LigPrep module. The structures were optimized and different tautomers and protonation states (pH = 7.0 ± 2.0) were generated. Protein structures were also prepared with the Protein Preparation Wizard workflow.³⁶ This process includes different steps, i.e., assigned bond orders, added cap termini, optimized the hydrogen-bonding network and protonation states of some residues, among others. The molecular docking using the hA_2A crystallized structure (PDB code: 3EML)³⁵ and the homology model for the hA₃ was carried out. No water molecules in the pocket of hA₃ were included in the docking calculations. For the hA_2A, calculations without water and with water were performed. The compounds were docked with the Glide module. Selection of the described binding modes was carried out taking into account the energetic parameter E_model and number of similar poses yielded by the calculations.

Acknowledgements

The authors would like to thank Fundação para a Ciência e Tecnologia (FCT)-Pest/C-QUI/UI0081/2013- and “Angeles Alvariño, Plan Galego de Investigación, Innovación e Crecemento 2011–2015 (I2C)” for the financial support. Thanks are due to Fundação para a Ciência e Tecnologia (FCT), Programa Operacional Potencial Humano (POPH) and Quadro de Referência Estratégica Nacional (QREN) for the post-doctoral and doctoral grants: A. Gaspar (SFRH/BPD/93331/2013), F. Cagide (SFRH/BPD/74491/2010) and J. Reis (SFRH/BD/96033/2013).

Notes and references

1. J. F. Chen, H. K. Eltzschig and B. B. Fredholm, Nat. Rev. Drug Discovery, 2013, 12, 265–286.
2. B. B. Fredholm, A. P. IJzerman, K. A. Jacobson, J. Linden and C. E. Muller, Pharmacol. Rev., 2011, 63, 1–34.
3. S. Moro, Z. G. Gao, K. A. Jacobson and G. Spalluto, Med. Res. Rev., 2006, 26, 131–159.
4. P. G. Baraldi, D. Preti, P. A. Borea and K. Varani, J. Med. Chem., 2012, 55, 5676–5703.
5. C. E. Muller and K. A. Jacobson, Biochim. Biophys. Acta, 2011, 1808, 1290–1308.
6. J. Spychala, Pharmacol. Ther., 2000, 87, 161–173.
7. S. Gessi, S. Merighi, V. Sacchetto, C. Simioni and P. A. Borea, Biochim. Biophys. Acta, 2011, 1808, 1400–1412.
8. P. A. Borea, K. Varani, F. Vincenzi, P. G. Baraldi, M. A. Tabrizi, S. Merighi and S. Gessi, Pharmacol. Rev., 2015, 67, 74–102.
9. P. Fishman, S. Bar-Yehuda, B. T. Liang and K. A. Jacobson, Drug Discovery Today, 2012, 17, 359–366.
10. P. G. Baraldi, B. Cacciari, R. Romagnoli, S. Merighi, K. Varani, P. A. Borea and G. Spalluto, Med. Res. Rev., 2000, 20, 103–128.
11. K. N. Klotz, N. Falgner, S. Kachler, C. Lambertucci, S. Vittori, R. Volpini and G. Cristalli, Eur. J. Pharmacol., 2007, 556, 14–18.
12. R. Volpini, M. Buccioni, D. Dal Ben, C. Lambertucci, C. Lammi, G. Marucci, A. T. Ramadori, K. N. Klotz and G. Cristalli, J. Med. Chem., 2009, 52, 7897–7900.
13. S. Taliani, C. La Motta, L. Mugnaini, F. Simorini, S. Salerno, A. M. Marini, F. Da Settimo, S. Cosconati, B. Cosimelli, G. Greco, V. Limongelli, L. Marinelli, E. Novellino, O. Ciampi, S. Daniele, M. L. Trincavelli and C. Martini, J. Med. Chem., 2010, 53, 3954–3963.
14. S. Gessi, S. Merighi, K. Varani, E. Leung, S. Mac Lennan and P. A. Borea, Pharmacol. Ther., 2008, 117, 123–140.
15. P. Fishman, K. A. Jacobson, A. Ochaion, S. Cohen and S. Bar-Yehuda, Immunol., Endocr. Metab. Agents Med. Chem., 2007, 7, 298–303.
16. L. Madi, S. Bar-Yehuda, F. Barer, E. Ardon, A. Ochaion and P. Fishman, J. Biol. Chem., 2003, 278, 42121–42130.
17. P. G. Baraldi, M. A. Tabrizi, A. Bovero, B. Avitabile, D. Preti, F. Fruttarolo, R. Romagnoli, K. Varani and P. A. Borea, Eur. J. Med. Chem., 2003, 38, 367–382.
18. H. Kim, J. W. Kang, S. Lee, W. J. Choi, L. S. Jeong, Y. Yang, J. T. Hong and D. Y. Yoon, Anticancer Res., 2010, 30, 2823–2830.
19. P. Mlejnek, P. Dolezel and I. Frydrych, J. Physiol. Biochem., 2013, 69, 405–417.
20. M. J. Matos, A. Gaspar, S. Kachler, K. N. Klotz, F. Borges, L. Santana and E. Uriarte, J. Pharm. Pharmacol., 2013, 65, 30–34.
21. M. J. Matos, S. Vilar, S. Kachler, A. Fonseca, L. Santana, E. Uriarte, F. Borges, N. P. Tatonetti and K. N. Klotz, ChemMedChem, 2014, 9, 2245–2253.
22. X. Ji, N. Melman and K. A. Jacobson, J. Med. Chem., 1996, 39, 781–788.
23. Y. Karton, J. Jiang, X. Ji, N. Melman, M. E. Olah, G. L. Stiles and K. A. Jacobson, J. Med. Chem., 1996, 39, 2293–2301.
24. K. A. Jacobson, S. Moro, J. A. Manthey, P. L. West and X. D. Ji, Adv. Exp. Med. Biol., 2002, 505, 163–171.
25. A. Gaspar, M. J. Matos, J. Garrido, E. Uriarte and F. Borges, Chem. Rev., 2014, 114, 4960–4992.
26. A. Gaspar, J. Reis, S. Kachler, S. Paoletta, E. Uriarte, K. N. Klotz, S. Moro and F. Borges, Biochem. Pharmacol., 2012, 84, 21–29.
27. A. Gaspar, J. Reis, M. J. Matos, E. Uriarte and F. Borges, Eur. J. Med. Chem., 2012, 54, 914–918.
28. A. Gaspar, F. Cagide, E. Quezada, J. Reis, E. Uriarte and F. Borges, Magn. Reson. Chem., 2013, 51, 251–254.
29. J. Reis, A. Gaspar, F. Borges, L. R. Gomes and J. N. Low, J. Mol. Struct., 2014, 1056–1057, 31–37.
30. F. Cagide, T. Silva, J. Reis, A. Gaspar, F. Borges, L. R. Gomes and J. N. Low, Chem. Commun., 2015, 51, 2832–2835.
31. K. N. Klotz, G. Cristalli, M. Grifantini, S. Vittori and M. J. Lohse, J. Biol. Chem., 1985, 260, 14659–14664.
32. K. N. Klotz, J. Hessling, J. Hegler, C. Owman, B. Kull, B. B. Fredholm and M. J. Lohse, Naunyn Schmiedebergs Arch. Pharmacol., 1998, 357, 1–9.
33. Y. Cheng and W. H. Prusoff, Biochem. Pharmacol., 1973, 22, 3099–3108.
34. A. Delean, A. A. Hancock and R. J. Lefkowitz, Mol. Pharm., 1982, 21, 5–16.
35. V. P. Jaakola, M. T. Griffith, M. A. Hanson, V. Cherezov, E. Y. Chien, J. R. Lane, A. P. Ijzerman and R. C. Stevens, Science, 2008, 322, 1211–1217.
36. Schrödinger suite 2014-3, Schrödinger, LLC, New York, USA, 2014, Available at http://www.schrodinger.com/accessed: Nov 2014.
37. M. Congreve, S. P. Andrews, A. S. Dore, K. Hollenstein, E. Hurrell, C. J. Langmead, J. S. Mason, I. W. Ng, B. Tehan, A. Zhukov, M. Weir and F. H. Marshall, J. Med. Chem., 2012, 55, 1898–1903.
38. V. P. Jaakola, J. R. Lane, J. Y. Lin, V. Katritch, A. P. Ijzerman and R. C. Stevens, J. Biol. Chem., 2010, 285, 13032–13044.
39. M. F. Martins Borges, A. M. Neves Gaspar, J. M. Pinto de Jesus Garrido and N. J. Da Silva Pereira Milhazes, PT 103665, WO 2008/104925 A1, 2008.
40. A. Gaspar, T. Silva, M. Yáñez, D. Vina, F. Oralo, F. Ortuso, E. Uriarte, S. Alcaro and F. Borges, J. Med. Chem., 2011, 54, 5165–5173.
41. F. Cagide, J. Reis, A. Gaspar and F. Borges, Tetrahedron Lett., 2011, 52, 6446–6449.
42. A. Gaspar, J. Reis, A. Fonseca, N. Milhazes, D. Viña, E. Uriarte and F. Borges, Bioorg. Med. Chem. Lett., 2011, 21, 707–709.
43. F. Cagide, F. Borges, L. R. Gomes and J. N. Low, J. Mol. Struct., 2015, 1089, 206–215.
44. MOE, version 2011.10, Chemical Computing Group, Inc., Available at: http://www.chemcomp.com accessed: Jan 2012.
45. V. Katritch, I. Kufareva and R. Abagyan, Neuropharmacology, 2011, 60, 108–115.

---