Acyl-Hydrazide Derivatives of a Xanthine Carboxylic Congener (XCC) as Selective Antagonists at Human A2B Adenosine Receptors

The structure–activity relationships (SAR) of 8-phenyl-1,3-dipropylxanthine derivatives in binding to recombinant human A2B adenosine receptors were explored, in order to identify selective antagonists. Based on the finding of receptor selectivity in MRS 1204, containing an N-hydroxysuccinimide ester attached through the p-position of the 8-phenyl substituent [Jacobson et al. (1999): Drug Dev. Res., 47:45–53], a hydrazide and its more stable imide derivatives were synthesized. The hydrazide of XCC (8-[4-[[[carboxy]methyl]oxy]phenyl]-1,3-dipropylxanthine) was acylated with a variety of mono- and dicarboxylic acids. Ki values were determined in the adenosine receptor binding assays. At recombinant human A2B receptors expressed in membranes of HEK-293 cells, antagonist radioligands used were the xanthine 125I-ABOPX (125I-3-(4-amino-3-iodobenzyl)-8-oxyacetate-1-propyl-xanthine) and the nonxanthine antagonist [3H]ZM 241385 ([3H]4-(2-[7-amino-2-{furyl}{1,2,4}triazolo{2,3-a}{1,3,5}triazin-5-ylamino-ethyl)phenol). The initial screening utilized rat A1/A2A receptors and human A3 receptors, and selected compounds were examined at the human A1/A2A subtypes. A 1,2-dimethylmaleimide derivative, 14 (MRS 1595), bound to human A2B receptors with a Ki of 19 nM and proved to be selective vs. human A1/A2A/A3 receptors by 160-, 100-, and 35-fold, respectively. Enprofylline (3-propylxanthine) is slightly selective for A2B receptors, suggesting removal of the 1-propyl group; however, combination of the 1-H-3-Pr and 8-phenyl substituents eliminated the selectivity. Other potent and moderately selective A2B antagonists were a tetrahydrophthaloyl derivative 18b (MRS 1614, Ki value 10 nM) and amino acid conjugates of the XCC-hydrazide, i.e., the glutarimide 24b (MRS 1626, Ki value 13 nM), and protected dipeptide 27 (MRS 1615, Ki value 11 nM). Drug Dev. Res. 47:178–188, 1999.


INTRODUCTION
Adenosine receptors [Linden and Jacobson, 1998] constitute four members of the G proteincoupled receptor superfamily, have structure-function homology to the biogenic amine receptors [Jiang et al., 1997], and are widely distributed in the body. Adenosine is a local modulator in the cardiovascular, renal, and immune systems and in the central nervous system. The A 2B adenosine receptor [Daly et al., 1983; see review by Feoktistov and Biaggioni, 1997] is involved in the control of cell growth and gene expression [Neary et al., 1996], vasodilation [Martin et al., 1993], and fluid secretion from intestinal epithelia [Strohmeier et al., 1995].
A selective A 2B receptor antagonist may have potential use as an antiasthmatic agent [Feoktistov and Biaggioni, 1997]. A possible role for A 2B ARs in asthma is consistent with the therapeutic efficacy of enprofylline, 1, and theophylline, 2, in treating asthma. In radioligand binding assays, both of these xanthines were confirmed to be effective, although not very potent, antagonists of human A 2B ARs in the therapeutic dose range . Furthermore, enprofylline, with a K i value of 7 μM, even appears to be somewhat selective for human A 2B ARs [Robeva et al., 1996b]. A 2B ARs are expressed in some mast cells, such as canine BR mastocytoma cells, in which they appear to be responsible for triggering acute Ca 2+ mobilization and degranulation [Auchampach et al., 1997]. A 2B ARs also participate in a delayed IL8 release from human HMC-1 mast cells [Feoktistov et al., 1999]. The A 3 AR may also play a role in asthma, since it mediates the degranulation of rat RBL mast-like cells [Ramkumar et al., 1993] and is present in high density in human blood eosinophils [Kohno et al., 1996].
Although adenosine receptor subtype-selective probes are available for the A 1 , A 2A , and A 3 adenosine receptors ], very few selective antagonists and agonists are known for the A 2B receptor, in part because the absence of radioligand binding assays has precluded a detailed investigation of the SAR at this subtype. MRS 1224, 7b, a derivative of the triazoloquinazoline, CGS15943, 7a, was highly potent at the A 2B receptor [Kim et al., 1998]. Although selective for the A 2A receptor, the triazolotriazine ZM 241385 was also shown to be a potent antagonist at the A 2B receptor and useful as a radioligand in cells expressing the recombinant A 2B receptor [Ji and . Alloxazine, 6, [Brackett and Daly, 1994] has been reported to be approximately one order of magnitude selective as antagonists at the A 2B receptor vs. other subtypes. Among xanthines, an 8-phenyl group is associated with increased affinity at A 2B receptors. The 8-phenyl analog, 3, of theophylline, 2, displayed a 22-fold enhancement of affinity at A 2B receptors . A lead for achieving moderate selectivity (at least 20-fold vs. A 1 , A 2A , and A 3 adenosine receptors) have been found in the category of complex 8-phenylxanthine derivatives. methyl]oxy]phenyl]-1,3-dipropylxanthine (XCC), 4a, and its ethyl ester, 4b, displayed high affinity for the A 2B receptor. Moreover, MRS 1204 (Nhydroxysuccinimide ester of XCC), 4d, displayed moderate selectivity (at approximately 20-fold for human A 2B receptors  vs. A 1 , A 2A , and A 3 adenosine receptors).
As an approach to finding selective antagonists for the A 2B receptor, we synthesized novel 8-phenyl-1,3-dialkylxanthines related structurally to 4d, in most of which the active ester bond has been replaced by a more stable acyl-hydrazide bond, and screened them for receptor affinity and selectivity in binding to the recombinant human A 2B receptor and other adenosine receptor subtypes. In order to identify potent adenosine receptor subtypeselective antagonists, in this study we utilized radioligand binding assays based on the use of membranes derived from HEK-293 cells that overexpress recombinant human A 2B ARs.

Synthesis
Proton nuclear magnetic resonance spectroscopy was performed on a Varian GEMINI-300 spectrometer and spectra were taken in DMSO-d 6 or CDCl 3 . Unless noted, chemical shifts are expressed as ppm downfield from tetramethylsilane, or relative ppm from DMSO (2.5 ppm). Chemical-ionization (CI) mass spectrometry was performed with a Finnigan 4600 mass spectrometer, and Electron-impact (EI) mass spectrometry with a VG7070F mass spectrometer at 6 kV. FAB (fast atom bombardment) mass spectrometry was performed with a JEOL SX102 spectrometer using 6-kV Xe atoms. All xanthine derivatives tested in binding assays were shown to be homogeneous by TLC (MK6F silica, 0.25 mm, glass-backed; Whatman Inc., Clifton, NJ). NMR and mass spectra were shown to be consistent with the assigned structure.

8-[4-[(Carboxymethyl)oxy]phenyl]-1H-3-(n-propyl)xanthine methyl ester (36).:
To a suspension of 3.2 g of 32 [Papesch and Schroeder, 1951] (18.9 mmole), 1.5 mL of glacial acetic acid and 3.4 mL of 6 N HCl in 50 mL of water was added dropwise to a solution of 1.38 g of sodium nitrite (20 mmole) in 5 mL of water at 0°C. The mixture was stirred for 1 h and the pink precipitate was collected by filtration to give 3.17 g of 33 (yield 78%). 1 H NMR (DMSO-d 6 ) 0.87 (t, 3H, J = 7.8 Hz, -CH 3 ), 1.51 (m, 2H, -CH 2 -), 3.72 (t, 2H, J = 6.8 Hz, -NCH 2 -), 9.12 (s, 1H, -NH 2 ). 0.086 g of 33 (0.4 mmole) was hydrogenated with 10% Pd/C in 5 mL of MeOH under H 2 atmosphere (1 atm) at 25°C until the pink color disappeared (30 min). After the removal of the balloon of H 2 , 5 mL of DMF was added and the mixture was stirred for 10 min and filtered through a Celite bed. To the solution of crude 34 was added 0.078 g of methyl 4-formylphenyloxyacetate (0.4 mmole) and 0.5 mL of acetic acid. The mixture was heated at 50°C for 30 min, evaporated under reduced pressure, and suspended with 20 mL of ether. The yellow precipitate (mixture of 35 and 36) was collected by filtration, dissolved in 5 mL of DMF, and treated with 1 mL of aqueous solution of 0.085 g of sodium periodate (0.4 mmole) for 2 h. After evaporation, the product was purified by crystallization in MeOH/H 2 O to give 0.048 g of 36 (yield 34%). 1

Pharmacology
The human A 2B receptor cDNA was subcloned into the expression plasmid pDoubleTrouble [Robeva et al., 1996a]. The plasmid was amplified in competent JM109 cells and plasmid DNA isolated using Wizard Megaprep columns (Promega Corp., Madison, WI). A 2B adenosine receptors were introduced into HEK-293 cells by means of Lipofectin [Felgner et al., 1987].

Radioligand binding studies-Confluent monolayers of HEK-A 2B cells were washed
with PBS followed by ice-cold Buffer A (10 mM HEPES, 10 mM EDTA, pH 7.4) with protease inhibitors (10 mg/mL benzamidine, 100 mM phenylmethanesulfonyl fluoride, and 2 mg/mL of each aprotinin, pepstatin, and leupeptin). The cells were homogenized in a Polytron (Brinkmann) for 20 sec, centrifuged at 30,000g, and the pellets washed twice with buffer HE (10 mM HEPES, 1 mM EDTA, pH 7.4 with protease inhibitors). The final pellet was resuspended in buffer HE, supplemented with 10% sucrose and frozen in aliquots at −80°C. For binding assays, membranes were thawed and diluted 5-10-fold with HE to a final protein concentration of approximately 1 mg/mL. To determine protein concentrations, membranes, and bovine serum albumin standards were dissolved in 0.2% NaOH/0.01% SDS and protein determined using fluorescamine fluorescence [Stowell et al., 1978]. Saturation binding assays for human A 2B adenosine receptors were performed with [ 125 I-]ABOPX (2,200 Ci/mmol). To prepare [ 125 I-]ABOPX, 10 mL of 1 mM ABOPX in methanol/1 M NaOH (20:1) was added to 50 mL of 100 mM phosphate buffer, pH 7.3. One or 2 mCi of Na 125 I was added, followed by 10 mL of 1 mg/mL chloramine-T in water. After incubating for 20 min at room temperature, 50 mL of 10 mg/mL Na-metabisulfite in water was added the quench the reaction. The reaction products were applied to a C18 HPLC column using 4 mM phosphate, pH 6.0/methanol. After 5 min in 35% methanol, the methanol concentration was ramped to 100% over 15 min. Unreacted ABOPX eluted in 11-12 min; [ 125 I-]ABOPX eluted at 18-19 min in a yield of 50-60% of the initial 125 I. In equilibrium binding assays the ratio of [ 127 I/ 125 I-]ABOPX was 10-20/1. Radioligand binding experiments were performed in triplicate with 20-25 μg membrane protein in a total volume of 0.1 mL HE buffer supplemented with 1 U/mL adenosine deaminase and 5 mM MgCl 2 . The incubation time was 3 h at 21°C. Nonspecific binding was measured in the presence of 100 mM NECA. Competition experiments were carried out using 0.6 nM 125 I-ABOPX. Membranes were filtered on Whatman GF/C filters using a Brandell cell harvester (Gaithersburg, MD) and washed three times over 15-20 sec with ice-cold buffer (10 mM Tris, 1 mM MgCl 2 , pH 7.4). B max and K D values were calculated by Marquardt's nonlinear least squares interpolation for single site binding models [Marquardt, 1963]. K i values for different compounds were derived from IC 50 values as described previously [Linden, 1982]. Data from replicate experiments are tabulated as means ± SEM.
[ 3 H]CPX, 125 I-ZM 241385 and 125 I-ABA were utilized in radioligand binding assays to membranes derived from HEK-293 cells expressing recombinant human A 1 , A 2A , and A 3 adenosine receptors, respectively. Binding of [ 3 H]R-N 6 -phenylisopropyladenosine ([ 3 H]R-PIA; Amersham, Chicago, IL) to A 1 receptors from rat cerebral cortical membranes and of [ 3 H]CGS 21680 (NEN Life Sciences, Boston, MA) to A 2A receptors from rat striatal membranes was performed as described previously [Schwabe and Trost, 1980;Jarvis et al., 1989]. Adenosine deaminase (3 units/mL) was present during the preparation of the brain membranes in a preincubation of 30 min at 30°C and during the incubation with the radioligands. All nonradioactive compounds were initially dissolved in DMSO and diluted with buffer to the final concentration, where the amount of DMSO never exceeded 2%. Incubations were terminated by rapid filtration over Whatman GF/B filters using a Brandell cell harvester. The tubes were rinsed three times with 3 mL buffer each.
At least six different concentrations of competitor, spanning 3 orders of magnitude adjusted appropriately for the IC 50 value of each compound, were used. IC 50 values, calculated with the nonlinear regression method implemented in Graph-Pad Prism (San Diego, CA) were converted to apparent K i values as described by Linden [1982]. Hill coefficients of the tested compounds were in the range of 0.8-1.1.

RESULTS AND DISCUSSION
The structures of the xanthine derivatives, 4, 9-27, tested for affinity in radioligand binding assays at adenosine receptors, are shown in Table 1. Most of the xanthines are derivatives of XCC [Jacobson et al., 1985], in which an acyl-hydrazide group is present. This group was included based on the high potency in the A 2B receptor binding assay (K i value of 9.75 nM ) of an N-hydroxysuccinimide ester of XCC, 4d. The hydrazide of XCC, 4c, was acylated with a variety of mono-and dicarboxylic acids. Cyclization reactions were carried out for dicarboxylic acids, in two steps using the anhydride, 28, for acylation, leading to imide (5-or 6-membered ring) derivatives (Fig. 2). The final step of ring-closure of 29a to 29b was effected at 50 °C, using excess carbodiimide and 1-hydroxybenzotriazole as catalyst. In some cases, where symmetric dicarboxylic acids were used, it was possible to isolate both the open structure, 29a, and the cyclized imide form, 29b. Pairs of open and cyclized derivatives of symmetric dicarboxylic acids prepared include compounds 18-20. Also, the glutamic acid derivative 24a was prepared using orthogonal protecting and the corresponding imide, 24b. An 8-phenyl analog, 15, of enprofylline was synthesized by standard methods from the asymmetric urea, 30 (Fig. 3).
The initial screening utilized rat A 1 /A 2A receptors, and selected compounds were examined at the human subtypes. Selectivities for the human A 2B vs. rat A 1 /A 2A receptors were generally small (3-4-fold at best), while comparisons within the same species (human) generally lead to greater selectivities. A 1,2-dimethylmaleimide derivative, 14, bound to human A 2B receptors with a K i of 19 nM and proved to be selective vs. human A 1 /A 2A /A 3 receptors by 160-, 100-, and 35-fold, respectively.
The identification of 14 (MRS 1595) as an adenosine antagonist which is potent and selective for human A 2B receptors and should be hydrolytically stable will provide an opportunity to test the hypothesis that this subtype is involved in asthma. Further SAR studies are in progress to enhance the pharmacological profile of these xanthine derivatives as A 2B receptor antagonists.

ACKNOWLEDGMENT
We thank Melissa Marshall for technical assistance with the binding assays. Structures of xanthines and nonxanthines previously identified as antagonists at A 2B receptors.

Fig. 2.
Derivatization of a xanthine containing a hydrazide group attached through the p-position of an 8-phenyl substituent [Jacobson et al., 1985]. The hydrazide, 4c, was acylated with the anhydride, 28, of a variety of dicarboxylic acids, followed by ring closure leading to stable imide derivatives, 29b. Synthesis of xanthine derivatives containing both 8-phenyl substituents and the 1-H-3-propyl substitution present in enprofylline, 1, as potentially selective A 2B receptor antagonists.