Synthesis of adenosine analogues with indole moiety as human adenosine A₃ receptor ligands

Abstract

Adenosine is an endogenous modulator exerting its functions through the activation of four adenosine receptor (AR) subtypes, termed A₁, A_2A, A_2B and A₃, which belong to the G-protein-coupled receptor superfamily. The human A₃AR (hA₃AR) subtype is implicated in several cytoprotective functions. Therefore, hA₃AR modulators, and in particular agonists, are sought for their potential application as anti-inflammatory, anti-cancer and cardioprotective agents. Here, we prepared novel adenosine derivatives with indole moiety as hA₃AR ligands. According to the biological assay, we found that 2-substituents 11 were critical structural determinants for A₃AR ligands (K_i = 111 nM). The observed structure–affinity relationships of this class of ligands were also exhaustively rationalized using the molecular modelling approach. This allows the investigation on the binding mode of the potential compound in the ligand-binding pocket of the human A₃ receptor. The results demonstrated that 11 can interact with the ASN250, GLN167, PHE168 and VAL178 through hydrogen bonding, which are shown to be important for ligand–receptor interaction.

1. Introduction

Adenosine is an endogenous purine nucleoside that modulates many physiological processes, which is composed of a molecule of adenine attached to a ribose sugar molecule (ribofuranose) moiety via a β-N₉-glycosidic bond [1–3]. Adenosine is widely found in nature and plays an important role in biochemical processes, such as energy transfer as adenosine triphosphate (ATP) and adenosine diphosphate (ADP), as well as in signal transduction as cyclic adenosine monophosphate (cAMP). It is also a neuromodulator, believed to play a role in promoting sleep and suppressing arousal. Adenosine also plays a role in the regulation of blood flow to various organs through vasodilation [4–6]. Cellular signalling by adenosine occurs through four known adenosine receptor (AR) subtypes (A₁, A_2A, A_2B and A₃) [7]. All AR subtypes are G-protein-coupled receptors. The four receptor subtypes are further classified based on their ability to either stimulate or inhibit adenylate cyclase activity. They have long been considered to be promising therapeutic targets in a wide range of conditions, ranging from cerebral diseases to cancer, including inflammatory and immunological disorders [8,9]. Adenosine contributes in a significant manner to the maintenance of tissue integrity by modulating the immune system. Encouraging results have emerged with AR ligands for the management of several physiological conditions in preclinical and clinical settings [7,10].

Among the four AR subtypes, the A₃AR, probably the most studied AR subtype, is also ubiquitously expressed [11]. The distribution of A₃AR is species-dependent, and in humans, this subtype is expressed in the lungs, liver, heart, kidneys and brain [12–14]. The widespread distribution in different cells and tissues of the A₃AR suggests a potential involvement in various pathologies and the possible use as a selective pharmacological target [15]. This subtype of ARs is involved in a variety of important pathophysiological processes, ranging from modulation of cerebral and cardiac ischaemic damage to regulation of immunosuppression and inflammation [16].

The increasing knowledge about A₃ARs, in particular regarding the molecular biology of this subtype, has provided important evidence to consider this receptor as a novel therapeutic target. In addition, it enables rational design and the development of potent and selective A₃AR ligands as promising therapeutic options for a variety of diseases [17,18]. Therefore, small molecule modulators targeting the A₃AR have been sought for their potential application as anti-inflammatory, anti-cancer and cardioprotective agents [19–21].

Indoles probably represent one of the most important structural classes in drug discovery [22]. The indole substructure is a basic element for a number of biologically active natural and synthetic products. Indoles are found in a wide range of therapeutically important drugs [23,24]. Over the years, a considerable amount of effort has been made to find indoles with various biological activities and select certain agents for leads in drug discovery research. On account of their potent biological activities, indoles have continued to attract the interest of chemists and biologists alike in drug discovery.

In addition, the scientific community is making intensive efforts to design AR ligands endowed with greater selectivity or to develop innovative compounds acting as receptor modulators. To further explore the importance of indole framework as the basis for AR ligands, we investigate the receptor subdomain that binds the purine moiety by the study of 2-O-alkyl derivatives of 2-chloroadenosine. The present paper reports on the synthesis and binding studies of these compounds as well as of adenosine. We describe the lengths of alkyl group in terms of potency at the A₃ receptor as well as receptor subtype selectivity. Here, we use the conversion of the 2-alkynyl group into flexible 2-O-alkyl between purine and indole moiety.

2. Results and discussion

2.1. Synthesis

In an effort to discover new nucleoside analogues as potential A₃AR ligands, we pursued the 2-oxypurine nucleoside using indole alkyl iodides as shown in scheme 1. The synthesis of the 5′-CH₂OH analogues started from 2-amino-6-chloropurine riboside 1, which was converted into 6-chloro-2-hydroxy-9-(2,3,5-tri-O-acetyl-β-d-ribofuranosyl)purine 2, as reported [25]. The reaction of the hydroxyl group at the 2-position of 2 with various indole iodides was conducted in the presence of caesium carbonate to affect compounds 3–8. Simultaneous removal of the acetyl group and amination at the 6-position of 3–8 using ammonium hydroxide solution yielded compounds 9–14.

Scheme 1.

Synthesis of compounds 9–14.

For synthesizing the key compounds, we should synthesize intermediates for the 2-ether component depicted in schemes 2–4. Scheme 2 illustrates the synthesis of two carbons as linker indole iodides. The indole 2-oxoacetate 16 was accessed by the treatment of 6-bromoindole 15 to oxalyl chloride, followed by a reaction with ethanol. Alkylation of indole nitrogen by CH₃I and BnBr provided N-alkyl 17a and 17b, respectively. The following exposure to BH₃-SMe₂ provided corresponding alcohols 18a and 18b. The alcohols were transformed to corresponding iodides 19a–b by iodine, triphenylphosphine and imidazole.

Scheme 2.

Synthesis of intermediates 19a–b.

Scheme 4.

Synthesis of intermediates 32a–b.

The three carbons used as the linker of indole iodides shown in scheme 3 were obtained from the commercially available 3-indolepropionic acid 20. Propionate 21 was prepared by the conversion of carboxylic acid into methyl ester by treatment with CH₃I and KHCO₃. Subsequent N-alkylation of the indole nitrogen provided 22a and 22b followed by reduction of carboxylic methyl ester to produce the corresponding alcohols 23a and 23b. Finally, iodides 25a and 25b were derived from 23a–b via tosylate alcohol, which underwent substitution with iodine.

Scheme 3.

Synthesis of intermediates 25a–b.

The synthetic strategy for the preparation of substituted indole derivatives is described in scheme 4. The commercially available 4-bromoindole-3-carboxaldehyde 26 was reacted with CH₃I and BnBr to give rise to N-alkylation 27a–b, respectively. The Wittig reaction of aldehyde provided trans-indole acrylate 28a–b. The following treatment with NaBH₄ and BiCl₃ yielded propionates 29a–b. LiAlH₄ reduction of the carboxylic ester produced intermediates 30a–b followed by treatment with tosyl chloride to afford the corresponding 31a–b, and the tosylate moiety was displaced with good yield with iodine to incorporate the iodo functionality. Detailed synthetic procedures including the yield of reactions and characteristic data can be found in the electronic supplementary material.

2.2. Binding affinity studies

The aim of the present study is to expand knowledge of the structure–activity relationships at the A₃AR and at other subtypes, both in relation to binding affinity and intrinsic efficacy, of adenosine derivatives modified in the 2-position. A screening campaign to discover new scaffolds for A₃AR inhibition yielded the moderate-potency lead 11 (table 1). Among 2-substituted derivatives, 2-ethers were more potent than the corresponding amines or thioethers [26]. The effect of bromine substitution of the phenyl ring was evaluated. This series of 5′-brom analogue showed a tendency towards increased K_i values with A₁ and A₃ARs, depending on the bulkiness of the bromine atom. Of these analogues, compound 10 was equipotent to 13 with A₁AR. However, its selectivity to binding A₃AR was improved. Unexpectedly, 14 was threefold less potent than 13 in binding to A₃AR and was tolerated by A_2AAR. Although the 5′-bromo derivative 13 was somewhat equipotent to 11 with A₃AR, its selectivity was reduced.

Table 1.

Potency of 2-alkoxyadenosine derivatives to bind human A₁, A_2A and A₃ARs expressed in CHO cells.^a

		Ki (nM ± SEM) or % inhibition at 10 µM
entry	R	hA1ARb	hA2AARb	hA3ARb
9		(47 ± 5%)	2770 ± 500	679 ± 149
10		217 ± 57	380 ± 81	532 ± 144
11		300 ± 70	880 ± 200	111 ± 30
12		(34 ± 2%)	(56 ± 5%)	731 ± 209
13		230 ± 26	262 ± 192	177 ± 43
14		152 ± 49	371 ± 79	491 ± 143
	adenosine	—	—	290

^aAll experiments were performed on CHO cells stably expressing one of three subtypes of human ARs. The binding affinities for A₁, A_2A and A₃ARs were expressed as K_i values and were determined using agonist radioligands ([³H]CCPA), ([³H]CGS21680) and [¹²⁵I]I-AB-MECA, respectively. Values in parentheses are for weak binding, corresponding to an IC₅₀ ≥ 10 µM. Data are expressed as mean ± s.e.

^bK_i in binding, unless noted.

Interestingly, the 3-indolyl analogue 11 was sevenfold more potent than 12 with A₃AR. The increased size or steric hindrance of the N-substituent markedly decreases A₃AR potency. Nevertheless, 12 was less potent than 13 in binding to A_2AAR. The effect of the space of the alkyl chain between the 2-ethers and the indole moiety was tested. Elongation increased the affinity for A₃AR. Compound 11 showed a fivefold increased potency with A₃AR, while it was somewhat tolerated by A_2AAR. The affinity of 9 to all three AR subtypes showed low potency compared with that of 13, similar to the results with 10.

However, the corresponding 2-indolyl derivative 13 was more potent than compound 11 in affinity for A₁ and A_2ARs. The bulkiness of the N-substituent may be related to the increased affinity. Compound 14 was more potent than 12 in binding to all three ARs. Meanwhile, compound 11 displayed a fivefold potency enhancement over 9 with A₃AR. The N-Bn derivative 10 with potency close to that of 9 was invariant in affinity for A₁ and A_2ARs.

2.3. Molecular docking analysis

Driven by docking of several derivatives with hA₃R, we performed the molecular modelling studies [27] to explore the binding modes of all six aforementioned compounds as shown in table 2. Among these compounds, the compounds cpd10, 12 and 14 with the N-Bn substituent failed to dock into the ligand-binding site of the protein. The current docking results are consistent with the biological studies mentioned above and serve as an explanation for the sharply reduced affinity. On the contrary, all other three compounds (cpd9, 11 and 13) with N-methyl substituent were found to dock into the binding pocket of the protein of interest and further interact with the amino acids GLN167, PHE168, ASN250, etc., through hydrogen bonding or hydrophobic interactions (figures 1 and 2). Detailed information for docking experiment can be found in the electronic supplementary material. In particular, the common interactions for these three compounds were the hydrogen bonding with the amino acid ASN250 of the protein, which is shown to be closely relevant to antagonist interaction [28,29]. Furthermore, the docking results showed that the compound cpd11 with the highest K_i value has the highest predicted affinity (−9.08 kcal mol⁻¹, table 2).

Table 2.

The predicted free energy and affinity compared with experimental values.

entry	experimental Ki (hA3AR, nM)	predicted free energy (kcal mol−1)	predicted affinity (nM)
11	111 ± 30	−9.08	222.18
13	177 ± 43	−8.73	398.82
9	679 ± 149	−8.38	719.10

Figure 1.

The two-dimensional image of ligand–hA₃AR interactions with the hydrogen bonds and hydrophobic interactions (a–c); (d) the cpd9 is coloured green, the cpd11 is coloured magenta, the cpd13 is coloured cyan, respectively. The surrounding residues interacting with these ligands are also shown and labelled. The meaning of the items on the two-dimensional plot above is listed in the bottom panel.

Figure 2.

The most active compound cpd11 (magenta colour) in the active site of human adenosine A₃ receptor showing hydrogen bonding interactions.

Figure 2 depicts the binding mode of the compound cpd11 in the ligand-binding pocket of the protein in details. Cpd11 can be well docked into the binding site of interest. In the most potent molecule cpd11, the adenosine core contributed strongly to the binding affinity, which also demonstrates the rationality of the core as a key scaffold for the further chemical modifications. The hydroxyl oxygen of furan ring and the 6-postion N atom of adenyl group interact with the ASN250, GLN167, PHE168 and VAL178 through hydrogen bonding. On the other hand, the high affinity also requires the presence of the 2-O-alkyl-substituted groups for producing the strong hydrophobic interactions.

From these results, we have illustrated the interactions of our newly synthesized 2-O-alkyl-substituted adenosine analogues with a ligand-binding site of hA₃AR from the molecular modelling point of view. The results also showed the different roles of the substitutions which are strongly linked to the increased or decreased affinity. The exploration for these interactions can provide us the guidance for the future chemical modifications.

3. Conclusion

In this work, we designed and synthesized a series of 2-O-alkyl-substituted adenosine analogues with indole moiety. The 2-substituents 11 was the most potent among the series, and it was confirmed to be a modulator in a functional assay measuring its capacity to bind receptors in CHO cells expressing the hA₃A receptor. We found that 2-substituents 11 were critical structural determinant for A₃AR ligands (K_i = 111 nM). The promising compound can be considered a valuable seed for the design and development of new and even more selective and potent compounds. The molecular modelling studies have also been performed to investigate the binding mode of the potential compound in the ligand-binding pocket of human A₃ receptor. Here, this study provides useful foundations for the attainment of a detailed pharmacological and physiological characterization of the adenosine A₃ receptor.

Data accessibility

The detailed experimental synthetic procedures and spectra of the final compounds are provided in the electronic supplementary material.

Authors' contributions

Y.X. carried out the synthetic work and performed the NMR experiments. X.Z. carried out the biological screening. R.H. helped analyse and interpret the data. J.W. conceived the study, designed it and drafted the manuscript. All authors gave final approval for publication.

Competing interests

The authors declare that they have no competing interests.

Funding

This research was funded by the National Research Foundation of China (grant nos. 21442004, 21502008 and 91430217) and the Ministry of Science and Technology (grant no. 2016YFA0203200).