A SELECTIVE AGONIST AFFINITY LABEL FOR A₃ ADENOSINE RECEPTORS

Abstract

A newly synthesized, chemically reactive adenosine derivative, N⁶-(3-isothiocyanatobenzyl)adenosine-5'-N-methyluronamide, was found to bind selectively to A₃ receptors. K_i values for this isothiocyanate derivative in competition binding at rat brain A₁, A_2a, and A₃ receptors were 145, 272 and 10.0 nM, respectively. A preincubation with this derivative resulted in irreversible inhibition of radioligand binding at rat A₃ receptors in membranes of transfected CHO cells or RBL-2H3 mast cells, but not at rat A₁ or A_2a receptors. The loss of binding sites for 0.1 nM [¹²⁵I]N⁶-(4-aminobenzyl)adenosine-5'-N-methyluronamide, a high affinity A₃ receptor radioligand, in transfected CHO cell membranes was concentration-dependent with an IC₅₀ of 50 nM. No change was observed in the K_d value of the remaining A₃ receptor sites. The inhibition was also insensitive to theophylline (1 mM), consistent with the pharmacology of rat A₃ receptors. Structurally similar adenosine analogues lacking the chemically reactive isothiocyanate group failed to irreversibly inhibit A₃-binding.

Adenosine agonists such as NECA (5'-N-ethylcarboxamidoadenostne, 1, Figure 1) are known to facilitate the release of inflammatory mediators from mast cells (1,2). The action occurs via activation of phospholipase C and is accompanied by a rise in intracellular calcium concentration. Recently, a molecular explanation was offered for this unique action of adenosine, i.e. mast cells express the novel A₃ adenosine receptor (3), which was discovered through cloning from a rat brain cDNA library (4). Xanthines that act as antagonists at A₁ and A₂ adenosine receptors do not antagonize the response (4,5). For example, the xanthine amine congener, which has nanomolar affinity at rat brain A₁ receptors, binds to rat brain A₃ receptors with a K_i value of 29 μM (6) and does not antagonize functional effects (inhibition of adenylate cyclase) of activation of rat A₃ receptors (4).

Figure 1.

Structures and synthesis of adenosine derivatives having high affinity and/or selectivity as A₃ receptor agonists. NECA, 1, has K_i values of 6.3, 10.3, and 113 nM at rat A₁-, A_2a-, and A₃-receptors, respectively (5). The 3-amino derivative (13), 3, was converted to the electrophilic isothiocyanate analogue, 4, using thiophosgene. The 3-amino derivative, 3, is also selective for A₃-receptore (36-fold and 28-fold vs. A₁ and A_2a receptors, respectively), but is not chemically reactive in receptor binding.

A₃ adenosine receptors have been proposed as a therapuetic target in inflammatory diseases (7–9) and in diseases of the central nervous system (10,11). Adenosine is known to be a bronchoconstrictor In asthmatic patients (8,9), although the mechanism of this effects is currently unclear, since until recently selective probes have not been available. A selective A₃ antagonist might be useful as an anti-inflammatory drug (7). Chronic activation of A₃ receptors has also been proposed in treating cerebral ischemia (11).

In this study we have designed a chemically-reactive receptor probe, an adenosine derivative bearing an isothiocyanate group, 4 (Figure 1), and have shown that it is selective for A₃ receptors in models of reversible and irreversible binding in membranes from CHO cells stably transfected with rat brain A₃ receptors and in RBL-2H3 cell membranes.

MATERIALS AND METHODS

Synthesis

[³H]CGS 21680 was obtained from Dupont NEN (Boston, MA) and [¹²⁵I]AB-MECA was prepared as described (12). Compound 3 was prepared as described (13).

N⁶-(3-Isothiocyanatobenzyl)adenosine-5'-N-methyluronamide (4)

N⁶-(3-Aminobenzyl)adenosine-5'-N-methyluronamide (3, 14.1 mg, 35.3 μmol) was partially dissolved in dry chloroform (6 mL), and saturated sodium bicarbonate solution (2 mL) was added. Thiophosgene (16 μL, 210 μmol) was added at once with vigorous stirring. After 5 min. the reaction was complete (silica TLC, Rf 0.79, chloroform:methanol:ammonium hydroxide 80:20:1), and more chloroform (20 ml) and more water (5 mL) were added to break the emulsion. The phases were separated, and the aqueous phase was washed with chloroform (3 × 25 mL). The organic phases were combined, dried (MgSO₄) and the solvent was removed under vacuum to obtain 8 mg (51 % yield) of the pure compound. Mp: 170–174 °C decomp.

MS EI: 441 (M+), 325, 311, 282. IR (NaBr) 2051.3, 2103.5 cm⁻¹

¹H NMR DMSO-d₆ 2.70 (d, J = 4.6 Hz, 3H, NHCH₃), 4.14 (dd, J = 4.5 Hz J = 1.1 Hz, 1H, H-3'), 4.31 (d, J = 1.2, 1H, H-4'), 4.59 (dd, J = 7.3 Hz J = 4.7 Hz, 1H, H-2'), 4.70 (br s, 2H, N⁶-CH₂Ph), 5.97 (d, J = 7.6 Hz, 1H, H-1'), 7.24–7.31 (m, 1H). 7.31–7.42 (m, 3H), 8.29 (s, 1H, H-2). 8.44 (s, 1H, H-8), 8.58 (br s, 1H, N⁶H-CH₂Ph), 8.80–8.87 (m, 1H, NHCH₃).

Biochemical assays

Preparation of membranes for binding assays

CHO cells stably transfected with cDNA for rat A₃ receptors and RBL-2H3 cells were grown by methods described previously (5,12). Membranes from CHO cells, from RBL-2H3 cells were homogenized in ice cold 50 mM Tris HCI / 10 mM MgCl₂ /1 mM EDTA (pH 8.25, “50/10/1”) using a Polytron (Kinematica Gmbh, Luzern, Switzerland) at a setting of 6 for 10 sec. Rat forebrain (for A₁ binding) or striatum (for A_2a binding) were homogenized in ice cold 50 mM Tris HCI (pH 7.4). The membrane suspension was then centrifuged at 37,000 × g for 10 min at 4°C. The pellet was resuspended (20 mg tissue/ml) in the above buffer solution, preincubated at 30 °C for 30 min with 3 IU/ml of adenosine deaminase, and the membranes were again homogenized and centrifuged. Finally the pellet was suspended in buffer (100 mg wet weight per ml) and stored at −70°C. Protein was determined using the BCA protein assay reagents (Pierce Chemical Co., Rockford, IL).

Radioligand binding

For all binding experiments, adenosine deaminase was present (3 **lU/ml) during the incubation with radioligand. [³H]CGS 21680 binding to striatal A_2a-receptors in rat brain was carried out as described (14) using 20 μM 2-chloroadenosine to determine non-specefic binding. The binding of [³H]R-PIA to rat cortical A₁-receptors was carried out by the method described (5).

The binding of [¹²⁵I]AB-MECA to A3-receptors membranes from transfected CHO cells and RBL-2H3 cells was carried out as described (12). Unlike our previous study of A₃ receptors in brain membranes (6), a xanthine (A₁/A_2a antagonist) was not added to prevent non-selective binding of the radioligand, since the background level of A₁/A_2a receptors was minimal. A₁ receptor binding of 0.41±0.38 and 3.3±0.5 fmol/mg protein was found in CHO and RBL cell membranes, respectively, using 1.0 nM [³H]PLA. Similarly, A_2a receptor binding of 2.8±2.0 and 26±3 fmol/mg protein was detected using 5.0 nM [³H]CGS 21680.

For competition studies, IC₅₀ values were determined using the Inplot computer program (Graphpad, San Diego, CA) and converted to apparent K_i values using K_D values and the Cheng-Prusoff equation (15). K_D values for [¹²⁵I]AB-MECA binding to A₃ receptors in transfected CHO cells and in RBL-2H3 cells were 1.48 and 3.61 nM, respectively (12). K_D values in rat brain for [³H]PIA and [³H]CGS 21680 binding were 1.0 and 15 nM, respectively (5,14), at A₁ and A_2a receptors. Concentrations of [¹²⁵I]AB-MECA, [³H]PIA, and [³H]CGS 21680 used in competition experiments were 0.1, 1.0, and 5.0 nM, respectively.

Preincubation of membranes with inhibitor

Membranes were incubated with adenosine derivatives in pH 8.25 50/10/1 buffer containing adenosine deaminase, for 1 h at 25 °C, subjected to three washing cycles, consisting of centrifugation at 37,000 × g and resuspended of the pellet in 50/10/1 buffer prior to radioligand binding. For theophylline exposure experiments, membranes were preincubated with theophylline at 25°C for 30 min, and then 4 was added immediately for an additional incubation at 25°C for 60 min. At the end of this sequence, the membranes were washed by repeated centrifugation and resuspension and subjected to binding with the appropriate radioligand. At the final step, prior to radioligand binding, the membranes were homogenized manually using a glass tissue grinder.

RESULTS AND DISCUSSION

The structure activity relationships of N⁶ and 5'-substituted adenosine derivatives at rat A₃ adenosine receptors have been elucidated (5,13). The combination of the 5'-N-methyluronamide modification and a 3-substituted N⁶-benzyl substituent was found to generally result in moderate selectivity as agonists for A₃ receptors. One such derivative, IB-MECA, 2, which was 50-fold A₃-selective in binding assays, was found to selectively activate A₃ receptors also in vivo (10). Thus, the 3-position of the benzyl ring of this analogue was chosen as the site for incorporation of a chemically reactive functional group. The etectrophilic group selected for this study was the isothiocyanate group, which was shown previously to serve as a cross-linking point for covalent reaction of ligands selective for A₁ and A₂ subtypes of adenosine receptors (16–18).

An N⁶-(3-aminobenzyl) derivative of adenosine-5'-N-methytearboxamide (13), 3, was converted to the corresponding isothiocyanate derivative, 4, by treatment with thiophosgene (Figure 1). This conversion was carried out without protection ot the 2'-and 3'-hydroxyl groups.

Compound 4 was then characterized in adenosine receptor binding assays. Competition of binding of [¹²⁵I]AB-MECA, a high affinity A₃-agonist (12), in membranes from CHO cells stably transfected with rat brain A₃-receptors was measured under “reversible” conditions. The apparent K_i value of 4 at A₃ receptors was found to be 10.0±2.3 nM (n = 3), comparable to the affinity of the selective agonist 2 and the precursor 3 (13). At A₁-receptors in rat cortex, a K_i value of 145±41 nM (n = 3) for inhibition of [³H]PIA was determined. Thus, the selectivity ratio of 4 for A₃- versus A₁-receptors in the rat, based on K_i values, was 14-fold. At A_2a-receptors in rat striatum, the K_i value in displacement of the A_2a-selective agonist [³H]CGS2l680 was 272±93 nM (n = 3), indicating a 27-fold selectivity ratio in binding of 4 at A₃- vs. A_2a-receptors in the rat brain. The Hill coefficients for displacement of binding at rat brain A₁, A_2a-, or A₃-receptors were approximately equal to 1.

Competition by 4 of binding of [¹²⁵I]AB-MECA, in RBL-2H3 cell membranes was also measured under “reversible” conditions. An apparent K_i value of 23.0±7.8 nM (n = 3) was determined. Non-specific binding in control and treated RBL-2H3 cell membranes was nearly identical and amounted to 15–20% of total binding at 0.1 nM [¹²⁵I]AB-MECA.

At higher concentrations, compound 4 was found to significantly inhibit radioligand binding at A₃-receptors in an irreversible manner, in both transfected CHO cells and RBL-2H3 cells (Figure 2). Preincubation of transfected CHO cell membranes with 4 caused a dose-dependent, irreversible antagonism of the specific binding of 0.1 nM [¹²⁵I]AB-MECA, with an IC₅₀ value of approximately 50 nM. This IC₅₀ value was 4-times greater than the IC₅₀ value in competitive displacement of [¹²⁵I]AB-MECA in the same membranes. The IC₅₀ value for irreversible inhibition of A₃ receptors in RBL-2H3 cell membranes was approximately 400 nM. The irreversible nature of inhibition by the isothiocyanate derivative was demonstrated by the failure of repeated washing to regenerate the A₃-receptor binding site. At A₁-adertosine receptors in rat cortical membranes, 4 was not effective as an irreversible inhibitor (Figure 2), even at a concentration 70× the K_i value. Following preincubation with 4 at 1 μM the percentage of [³H]PIA binding sites was found to be 104±9% (n = 4) of specific control binding (membranes incubated in parallel) versus 12% (n = 2) of specific control binding [¹²⁵I]AB-MECA binding in the CHO cells.

Figure 2.

Dose-dependent Inhibition by 4 of radioligand binding at A₃-adenosine receptors in: A₃ receptors in stably transfected CHO cell membranes (filled circles, n = 2–4) and A₃ receptors in RBL-2H3 cell membranes (triangles, n = 3), and A₁ receptors in rat brain cell membranes (open circles, n = 3–4). The preincubation with 4 or control was carried out for 1 h at 25°C, and the subsequent binding assay involved a 60 min at 37°C incubation followed by rapid filtration. Concentrations of [¹²⁵I]AB-MECA and [³H]PIA used in competition experiments were 0.1, 1.0, and 5.0 nM, respectively. Each data point represents the mean±S.E.M of three separate determinations, each done in duplicate.

As a control experiment for demonstrating that irreversible binding depends on the presence of the isothiocyanate group, CHO cell membranes were pre-incubated with compounds 2 and 3, which are not chemically reactive due to the absence of an electrophilic group. The K_i values for these compounds at rat A₃ receptors were found to be 1.1 and 16 nM, respectively (13). The [¹²⁵I]AB-MECA binding sites in these membranes were nearly completely restored following the same washing procedure that was ineffective in regenerating the binding following incubation with 4. Thus following a preincubation with compound 2(10 nM) or 3 (100 nM), 98 or 90% (n = 2), respectively, of the specific A₃ binding was regenerated.

Binding of [¹²⁵I]AB-MECA to A₃ receptors in CHO cell membranes exposured to 4 was saturable (Figure 3) and of the same affinity as in control membranes, but the B_max was diminished. A preincubation with 100 nM 4 followed by washing resulted in a K_d value at the remaining sites of 1.63±0.22 nM, compared to 1.82±0.45 nM (n = 4) in the control membranes. The B_max values in treated and control membranes were 1.36±0.05 and 2.30±0.12 pmol/mg protein, respectively, representing a loss in the density of binding sites of 41%. The loss of binding sites determined by Scatchard analysis was also concentration dependent. Following a preincubation with 1000 nM 4, K_d and B_max values of 2.32 nM and 0.83 pmol/mg were determined.

Figure 3.

Saturation curve (A) and Scatchard transformation (B) for the binding of [¹²⁵I]AB-MECA to A₃-adenosine receptors in CHO cell membranes following a preincubation with compound 4 at 25°C for 1 hr. Specific binding for control membranes (open circles), and following exposure to compound 4 at 0.1 μM (filled circles) and 1.0 μM (squares) are shown. After three cycles of washing, the membranes were incubated with radioligand at 37°C for 60 min. The volume of incubation for radioligand binding (20–40 μg protein/tube) was 0.2 ml.

The degree of ireversible inhibition of binding of [¹²⁵I]AB-MECA at rat brain A₃-receptors by 4 (at either 0.1 or 1.0 μM) was not affected by co-incubation with the non-selective A₁ and A₂ adenosine receptor antagonist theophylline (1 mM). This is consistent with the inability of theophylline to block rat A₃ receptor-mediated inhibition of adenylate cyclase (5).

Although most chemical affinity labels for receptors are antagonist-derived, a few irreversibly binding agonists have been reported. An A_2a-selective isothiocyanate derivative, p-DITC-APEC, effected the prolonged and irreversible activation of A_2a adenosine receptors in the guinea pig coronary artery (17). An A₁-selective isothiocyanate derivative, m-DITC-ADAC, was found to be an irreversible, full agonist in prolonging the S-H interval in guinea pig isolated hearts (18). An agonist-derived chemical affinity labels for adrenergic receptors has proven useful in pharmacological studies of cross-talk with adenosine receptors (19). Agonist affinity labels are also potentially of use in studying spare receptors, receptor turnover, affinity states, peptide mapping of the binding site, and receptor activation and desensitization (20).

Indeed we have established that the isothiocyanate derivative, 4, binds irreversibly and selectively to A₃ receptors in transfected CHO cell membranes and in membranes from rat mast cells. Thus, we have introduced a new tool for the study of A₃ receptors, their molecular structure, regulation, and physiological role in the inflammatory (1–3), pulmonary (8,9), cardiovascular (21), and central nervous systems (10,11). It will be necessary to examine the functional effects (3–5) of this isothiocyanate derivative as a function of time in order to establish whether it acts as a full agonist in functional assays and whether this action is maintained or rapid desensitization occurs.

Abbreviations used

[¹²⁵I]AB-MECA

N⁶-(4-amino-3-iodobenzyl)adenosine-5'-N-methyl-uronamide

CGS 21680

2-(carboxyethylphenylethylamino)adenosine-5'-carboxamide

CHO

Chinese hamster ovary

IB-MECA

N⁶-(3-iodobenzyl)adenosine-5'-N-methyluronamide

PIA

fl-N⁶phenylisopropyl-adenosine

Tris

tris(hydroxy-methyl)aminomethane.