Synthesis, Biological Properties, and Molecular Modeling Investigation of the First Potent, Selective, and Water-Soluble Human A₃ Adenosine Receptor Antagonist

Abstract

A new, highly potent, selective, and water-soluble antagonist of the hA₃ adenosine receptor was synthesized and tested in binding and functional assays. Compound 4 (5-[[(4-pyridyl)amino]carbonyl]amino-8-methyl-2-(2-furyl)-pyrazolo-[4,3-e]1,2,4-triazolo[1,5-c]pyrimidine hydrochloride) displayed high water solubility (15 mM) and the highest affinity (K_i = 0.01 nM) and selectivity for the hA₃ versus A₁, A_2A, and A_2B receptors (>10000-fold) ever reported. A Schild analysis of the antagonism by 4 of agonist-induced inhibition of cAMP production in CHO cells expressing the hA₃ receptor indicated a K_B value of 0.20 nM.

Introduction.

It is well-known that adenosine regulates many physiological processes through the interaction with four known receptor subtypes classified as A₁, A_2A, A_2B, and A₃.^1,2 In particular, the A₃ adenosine receptor subtype, which is distributed in different organs (lung, liver, kidney, heart, and, with a lower density, the brain),³ exerts its action through the modulation of two second messenger systems: inhibition of adenylate cyclase⁴ and stimulation of phospholipases C⁵ and D.⁶ The potential therapeutic applications of activating or antagonizing this receptor subtype have been investigated in recent years. In particular, antagonists for the A₃ receptor promise to be useful for the treatment of inflammation⁷ and in the regulation of cell growth.^8,9 Consequently, much effort has been directed toward searching for potent and selective human A₃ adenosine antagonists.¹⁰ Recently, Baraldi and co-workers reported a large series of pyrazolotriazolopyrimidines, bearing substituted phenylcarbamoyl residues at the amino group at the 5-position, as highly potent and selective antagonists of the human A₃ adenosine receptor.^11–13 In particular, 1 (5-[[(phenyl)-amino]carbonyl]amino-8-methyl-2-(2-furyl)-pyrazolo[4,3-e] 1,2,4-triazolo[1,5-c]pyrimidine (Chart 1) showed the most favorable binding affinity and selectivity for the human A₃ adenosine receptor ever reported.¹³

Chart 1.

Structures, Biological Characterization, and Water Affinities of Reference Compounds

Unfortunately, a major problem within this class of compounds is the typical low water solubility, which has limited their use as pharmacological and diagnostic tools. The hydrophobicity of compound 1 is indicated by its high R_m value of 4.06. Previously an attempt for obtaining a water-soluble derivative has been made through the introduction of a sulfonic acid group at the para position of the phenyl ring (Chart 1) affording compound 2 (4-[3-(2-furan-2-yl-8-methyl-8H-pyrazolo-[4,3-e][1,2,4]triazolo[1,5-c]pyrimidin-5-yl)ureido]benzenesulfonic acid). As expected, a completely water-soluble derivative (R_m = 1.66, water solubility greater than 20 mM) was obtained; however, a significant loss of affinity (156-fold) and selectivity was observed.¹³ A hypothesis for the dramatic loss of affinity was provided through molecular modeling studies, which indicated that steric control seemed to be taking place around the para position of the phenyl ring in the putative A₃ receptor binding site.¹³ Taking into account these observations and with the aim of obtaining derivatives with high affinity, selectivity, and water solubility, we synthesized a new derivative (Chart 2) by bioisosteric replacement of the phenyl ring with the 4-pyridyl moiety (3), thus providing higher water solubility while avoiding the steric hindrance of a substituent at the para position, which seemed to be responsible for the reduction of hA₃ adenosine receptor affinity.

Chart 2.

Rational Design of Water-Soluble hA₃ Adenosine Receptor Antagonist

In fact, the introduction of a basic nitrogen was intended to further improve water solubility when protonated, i.e., in the form of the corresponding HCl salt (4).

Results and Discussion.

The synthesis of the desired compound has been performed following the general strategy depicted in Scheme 1. The major problem in the synthesis of derivative 3 involved the preparation of a 4-pyridyl isocyanate (5), which could not be synthesized by the usual method, i.e., reaction of the corresponding amine with phosgene or a phosgene equivalent, because of the reactivity and instability of pyridyl isocyanates.¹⁴

Scheme 1^a.

^a (i) NaNO₂, aqueous HCl, 0 °C, 1 h; (ii) benzene, reflux, 2 h; (iii) THF reflux overnight; (iv) HCl/MeOH, 0 °C, 30 min.

For this reason, 5 was prepared as reported in the literature starting from the commercially available isonicotinoyl hydrazide (6), which after reaction with sodium nitrite under acid conditions afforded the corresponding acyl azide 7.¹⁵ The latter was in turn converted into 5 upon Curtius rearrangement induced by heating 7 at reflux in dry benzene for 2 h.¹⁶ The crude isocyanate was refluxed overnight in dry THF with the precursor containing the tricyclic system bearing a methyl group at the N8 position (8). The desired product 3 was purified using flash chromatography and eluted with a MeOH/EtOAc gradient of 0–30%. The corresponding hydrochloride 4 was obtained by treatment of 3 for 30 min at 0 °C with methanol saturated with HCl gas. The hydrophobicity of the newly synthesized substances was measured in reverse-phase TLC experiments and reported as R_m values¹⁷ (R_m = log(1/R_f − 1) as shown in Table 1. As expected, both derivatives showed increased hydrophilicity with respect to reference 1, but most importantly, the hydrochloride salt 4 freely dissolved in water to a maximum concentration of 15 mM. Table 1 also compares the receptor binding affinities of 3 and 4 determined at human A₁,¹⁸ A_2A,¹⁹ A_2B,²⁰ and A₃²⁰ receptors expressed in CHO (A₁, A_2A, A₃) and HEK-293 (A_2B) cells.

Table 1.

Binding Affinity at hA₁, hA_2A, hA_2B, and hA₃ Adenosine Receptors and Water Solubility of Synthesized Compounds

compd	Rm(0)a	hA1 Ki (nM)b	hA2A Ki (nM)c	hA2B Ki (nM)d	hA3 Ki (nM)e	hA1/hA3	hA2A/hA3	hA2B/hA3
3	3.06 ± 0.06	250 ± 23	60 ± 10	200 ± 16	0.04 ± 0.009	6250	1500	5000
4	2.29 ± 0.05	350 ± 22	100 ± 12	250 ± 24	0.01 ± 0.005	35000	10000	25000

^aThe R_m values of 3 and 4 were measured with a mobile phase of different concentrations of MeOH/H₂O. R_m values are reported as theoretical at 0% organic solvent in the mobile phase (R_m(0)).

^bDisplacement of specific [³H]DPCPX²⁶ binding at human A_2A receptors expressed in CHO cells (n = 3–6).

^cDisplacement of specific [³H]ZM 241385²⁶ binding at human A_2A receptors expressed in HEK-293 cells.

^dDisplacement of specific [³H]DPCPX binding at human A_2B receptors expressed in HEK-293²⁶ cells (n = 3–6).

^eDisplacement of specific [³H]MRE3008-F2 0²⁶ binding at human A₃ receptors expressed in HEK-293 cells. Data are expressed as K_i ± SEM (n = 3–6). The affinity of 4 at the rat adenosine receptor in RBL-2H3 cells was also determined using previously reported methods.²¹ K_i values (n = 3) were 226 ± 50 nM (A₁), 97.6 ± 26.2 nM (A_2A), and >1 μM (A₃).

Surprisingly, both substances showed very high affinity at the human A₃ adenosine receptor subtype, with K_i values in the picomolar range (10–40 pM) and with high levels of selectivity. In particular, the hydrochloride salt 4 showed increased affinity and selectivity with respect to the reference compound bearing the phenylcarbamoyl moiety 1. The K_i value of 4 at the hA₃ receptor was 0.01 nM, thus indicating high selectivity versus other subtypes: hA₁/hA₃ = 35 000, hA_2A/hA₃ = 10 000, hA_2B/hA₃ = 25 000. These values were more favorable than the selectivity ratios for compound 1, for which the K_i value at the hA₃ was reported to be 0.16 nM: hA₁/hA₃ = 3700, hA_2A/hA₃ = 2400, hA_2B/hA₃ = 1400. Moreover, this class of compounds, as previously demonstrated, proved to be inactive in a rat model with K_i values at the rA₃ typically greater than 1 μM.¹¹ 4 at 1 μM displaced only 35% of the specific binding of [¹²⁵I]I-AB-MECA at A₃ receptors in membranes of rat basophilic RBL-2H3 cells.^21b Concerning the affinity differences observed between the salt and neutral species, we speculate that a difference in their relative dissociation/association rates during the biological assay could be a reasonable explanation. These results not only represent the first example of a highly potent, selective, and water-soluble human A₃ adenosine antagonist but strongly suggest an involvement of the pyridine nitrogen in the receptor recognition. Using a homology modeling approach based on the crystal structure of the rhodopsin²² as a template, we have built an improved model of the transmembrane helical domains (TMs) of the human A₃ receptor,²³ which can be considered a further refinement of the hypothetical binding site for A₃ receptor antagonists already proposed.^13,22,24 As shown in Figure 1, after the Monte Carlo/annealing sampling, we propose that the hypothetical binding site of 4 is surrounded by TMs 3, 5, 6, and 7, with the furan ring pointing toward the extracellular environment. Similar to conformational results already described for other pyrazolotriazolopyrimidines,^12,13,22 the lowest energy conformation of 4 featured the carbamoyl moiety in the 5-position surrounded by four polar amino acids: Ser242 (TM6), Ser271 (TM7), His274, and Ser275 (TM7).

Figure 1.

Side view of the human A₃-4 complex model. The side chains of the important residues in proximity to the docked 4 molecule are highlighted and labeled.

Accordingly, this region seemed to be critical for the recognition of this class of antagonists. Moreover, additional strong electrostatic interactions appeared to occur between the positively charged pyridinium moiety of 4 and the carbonyl oxygen atoms of Asn274 (N⁺H····OC distance = 2.5 Å) and Asn278 (N⁺H····OC distance = 3.1 Å), both located on TM7. These electrostatic interactions might be responsible for the increase of the affinity in the protonated form, i.e., the hydrochloride derivative 4. Interestingly, these two asparagine residues (Asn274 and Asn278) are largely conserved among a number of GPCRs. It should be noted, however, that this human A₃ receptor model, based on the TM region of the receptor, is not able to clearly explain the corresponding increase of selectivity for this subtype even taking into account which of the important TM residues are different in the other receptor subtpyes. However, as already reported by Jacobson and co-workers, multiple regions of the adenosine receptors, including a segment of the second extracellular loop, are involved in ligand recognition.²⁵ Other investigations are in progress in our lab to better describe the role of the extracellular domain on the ligand recognition process. For confirmation of the high potency of this compound in a functional assay, the inhibition of cAMP₂₆ generation by Cl-IB-MECA in membranes of CHO cells stably transfected with the human A₃ receptor was evaluated.

Consistent with its binding affinity, 4 showed an IC₅₀ value of 0.7 ± 0.06 nM, compared to an IC₅₀ of 2.10 ± 0.21 nM for 1. A Schild analysis of the antagonism of the effects of 4 on Cl-IB-MECA²⁶-induced inhibition of forskolin-stimulated cAMP was carried out (Figure 2).¹⁹ A K_B value of 0.20 ± 0.03 nM was calculated, thus demonstrating 4 to be the most potent antagonist of the human A₃ receptor ever reported.

Figure 2.

Effects of 4 on the inhibition of cyclic AMP production induced by the agonist Cl-IB-MECA in human A₃ adenosine receptor-expressing CHO cells (A) and Schild analysis (B) of the data. The procedures used are described in Supporting Information. The data shown in panel A were derived from one experiment performed in duplicate and are typical of three independent experiments giving similar results. The K_B value for antagonism by 4 was calculated from three independent experiments.

Conclusions.

The present study revealed a novel, potent, selective, and most importantly, water-soluble hA₃ adenosine receptor antagonist. This derivative featured a basic 4-pyridylcarbamoyl moiety at the N5 position of the pyrazolotriazolopyrimidine nucleus, and the corresponding hydrochloride salt 4 displayed a K_i value of 0.01 nM at the hA₃ and selectivities versus the other adenosine receptor subtypes ranging from 10 000 to 35 000. This increase of affinity compared to neutral arylcarbamate derivatives could be attributed in receptor modeling to strong electrostatic interactions between the pyridinium moiety of 4 and the side chain carbonyl oxygen atoms of Asn274 and Asn278, both located on TM7. In view of the potency, selectivity, and water solubility, 4 could be an ideal candidate for pharmacological and clinical investigation of the hA₃ adenosine receptor subtype.