Detection of multiple H₃ receptor affinity states utilizing [³H]A-349821, a novel, selective, non-imidazole histamine H₃ receptor inverse agonist radioligand

Abstract

1. A-349821 is a selective histamine H₃ receptor antagonist/inverse agonist. Herein, binding of the novel non-imidazole H₃ receptor radioligand [³H]A-349821 to membranes expressing native or recombinant H₃ receptors from rat or human sources was characterized and compared with the binding of the agonist [³H]N-α-methylhistamine ([³H]NαMH).

2. [³H]A-349821 bound with high affinity and specificity to an apparent single class of saturable sites and recognized human H₃ receptors with 10-fold higher affinity compared to rat H₃ receptors. [³H]A-349821 detected larger populations of receptors compared to [³H]NαMH.

3. Displacement of [³H]A-349821 binding by H₃ receptor antagonists/inverse agonists was monophasic, suggesting recognition of a single binding site, while that of H₃ receptor agonists was biphasic, suggesting recognition of both high- and low-affinity H₃ receptor sites.

4. pK_i values of high-affinity binding sites for H₃ receptor competitors utilizing [³H]A-349821 were highly correlated with pK_i values obtained with [³H]NαMH, consistent with labelling of H₃ receptors by [³H]A-349821.

5. Unlike assays utilizing [³H]NαMH, addition of GDP had no effect on saturation parameters measured with [³H]A-349821, while displacement of [³H]A-349821 binding by the H₃ receptor agonist histamine was sensitive to GDP.

6. In conclusion, [³H]A-349821 labels interconvertible high- and low-affinity states of the H₃ receptor, and displays improved selectivity over imidazole-containing H₃ receptor antagonist radioligands. [³H]A-349821 competition studies showed significant differences in the proportions and potencies of high- and low-affinity sites across species, providing new information about the fundamental pharmacological nature of H₃ receptors.

Introduction

To date, a majority of histamine H₃ receptor-binding studies have relied on the H₃ receptor-selective agonists [³H]RαMH and [³H]NαMH as radiolabels (van der Goot & Timmerman, 2000). While having high selectivity and specific binding along with favorable signal to noise ratios, binding of these agonist radioligands to H₃ receptors is sensitive to guanine nucleotides, suggesting that H₃ receptor agonist radioligands recognize both high- and low-affinity receptor sites (Arrang et al., 1990; West et al., 1990a; Kilpatrick & Michel, 1991; Clark et al., 1993). This makes characterization of H₃ receptors difficult because of assay-dependent variables, such as membrane preparation procedures (Childers & Lariviere, 1984; Kim & Neubig, 1987) and ionic composition of assay buffers that can alter receptor–G-protein interactions and affect the apparent affinity state of the receptor (Hamblin & Creese, 1982; Kilpatrick & Michel, 1991; Parkinson & Fredholm, 1992). Several radiolabelled H₃-antagonists have been described previously, such as [³H]S-methylthioperamide, [¹²⁵I]iodophenpropit, [¹²⁵I]iodoproxyfan, [³H]thioperamide, [³H]GR168320 and [³H]clobenpropit (Jansen et al., 1992; Ligneau et al., 1994; Yanai et al., 1994; Alves-Rodrigues et al., 1996; Brown et al., 1996; Harper et al., 1999), but they have not been widely used. This is attributable to a variety of limiting factors, such as low specific activity, poor H₃ receptor specificity, equivocal functional properties or lack of commercial availability. Moreover, all of the previously referenced H₃ receptor antagonist radiolabels contain an imidazole moiety, which may be responsible for off-target binding to 5-HT₃ receptors, H₄ receptors and cytochrome P450 isoenzymes (Labella et al., 1992; Leurs et al., 1995; Schlicker et al., 1995; Alves-Rodrigues et al., 1996; Harper et al., 1999; Yang et al., 2002; Esbenshade et al., 2005a; Kitbunnadaj et al., 2005). More recently, we described the first non-imidazole, H₃ receptor antagonist, [³H]A-317920 (Esbenshade et al., 2005b; Yao et al., 2005). [³H]A-317920 is a highly selective and potent radiolabel useful for detection and characterization of recombinant rat H₃ receptors expressed in C6 cells, a rat glioma cell line, or native H₃ receptors expressed in rat cortical tissue. However, this radioligand does not recognize human H₃ receptors with high affinity, and as such has limited utility.

Subsequent medicinal chemistry efforts revealed a novel non-imidazole H₃ receptor ligand, A-349821, having high potency and selectivity for both rat and human H₃ receptors in competition-binding assays and potent antagonist activity in a variety of in vitro functional assays as well as H₃ receptor-mediated in vivo models (Esbenshade et al., 2004). A-349821 also displayed highly efficacious and potent inverse agonist properties in [³⁵S]GTPγS-binding assays across species (Esbenshade et al., 2004). Competition-binding studies carried out with ∼75 G-protein coupled receptors (GPCRs) and ligand-gated ion channels showed minimal crossreactivity (IC₅₀>1 μM), with weak affinity (IC₅₀=250 nM) observed for α_2c adrenergic receptors (Esbenshade et al., 2004). With these promising binding and pharmacological properties, [³H]A-349821 was prepared and evaluated as a radiolabel for the detection of H₃ receptors in membranes expressing recombinant or native rat or human H₃ receptors. Like the agonist radioligand [³H]NαMH, [³H]A-349821 exhibited specific and saturable binding to tissues or cell lines expressing either rat or human H₃ receptors. Detailed competition-binding studies using a large panel of H₃ receptor ligands were performed to compare the pharmacology of [³H]A-349821 with that of [³H]NαMH. The effects of guanosine 5′-diphosphate (GDP) on binding were determined to investigate the contribution of low- and high-affinity binding states to the observed pharmacology. Herein, we describe the use of [³H]A-349821 as the first non-imidazole radioligand for the highly specific detection and characterization of both rat and human H₃ receptors.

Methods

Synthesis of [³H]A-349821

A-362114 ({4′-[3-(2R,5R)-dimethyl-2,5-dihydro-pyrrol-1-yl)-propoxy]-biphenyl-4-yl}-morpholin-4-yl-methanone) (10 mg, 0.024 mmol) was dissolved in ethyl acetate (2 ml). Palladium on carbon (10%, 11 mg) was added and the mixture was degassed by freeze–pump–thaw and stirred under an atmosphere of tritium gas (0.042 mmol, 2.5 Ci) at room temperature for 2.5 h. The catalyst was filtered and labile tritium was removed by three evaporations of methanol to yield 362 mCi of crude [³H]A-349821 ([dimethylpyrrolidinyl-3,4-³H₂]{4′-[3-(2R,5R-dimethyl-pyrrolidin-1-yl)-propoxy]-biphenyl-4-yl}-morpholin-4-yl-methanone) (for chemical structures, see Figure 1). One-fifth of the product was purified by reversed-phase HPLC, using a 20 × 250 mm BHK ODS-W/S column (BHK Laboratories, Chicago, IL, U.S.A.) eluted at 16 ml min⁻¹ with 43% acetonitrile and 0.1% trifluoroacetic acid. The major UV peak (254 nm) was collected in four fractions and the two major fractions were combined and lyophilized to give a white solid (50 mCi). HPLC analysis showed 84% radiochemical purity, so a second purification using the same system, but with 30% acetonitrile, ensued. The desired product eluted at 34–38 min providing 20 mCi at >99% radiochemical purity by HPLC. The identity was established by HPLC retention time, co-injecting with authentic A-349821. The specific activity was determined to be 22 Ci mmol⁻¹ by measuring the mass concentration from an HPLC standard curve followed by liquid scintillation counting to determine the radioactivity concentration.

Figure 1.

Chemical structures of the precursor, A-362114 (a) and the radiolabelled product ([³H]A-349821 ([dimethylpyrrolidinyl-3,4-3H₂]{4′-[3-(2R,5R-dimethyl-pyrrolidin-l-yl)-propoxyl-biphenyl-4-yl}-morpholin-4-yl-methanone) (b).

Preparation of membranes from rat and human cortices and C6 cells expressing recombinant rat and human H₃ receptors

Cortices were obtained from Sprague–Dawley rats (Pelfreez, Rogers, AR, U.S.A.). Human cortical tissue was obtained from Analytical Biological Sciences (Wilmington, DE, U.S.A.). C6 cell lines stably expressing full-length (445 A.A.) recombinant human or rat H₃ receptors were generated and cultured as described previously (Yao et al., 2003). Cortical tissues or C6 cells expressing recombinant H₃ receptors were homogenized in ice-cold Tris–EDTA buffer (50 mM Tris-HCl, 5 mM EDTA, pH 7.4) containing protease inhibitors (2 μg ml⁻¹ aprotinin, 1 mM benzamidine, 2 μg ml⁻¹ leupeptin and 1 μg ml⁻¹ pepstatin; Sigma, St Louis, MO, U.S.A.) using an Ultra-Turrax T25 polytron (Janke & Kunkel GMBH & Co. KG, Germany) at maximum setting (24,000 r.p.m.) for 2 × 10 s. Homogenates were centrifuged at 40,000 × g for 20 min. Membrane pellets were further processed by repeating the above homogenization and centrifugation steps. Final membrane preparations were obtained by re-homogenizing pellets by polytron in 6.25 ml g⁻¹ (volume/wet pellet weight) of ice-cold Tris–EDTA buffer. Membranes were divided into aliquots, flash frozen in liquid nitrogen, and stored at −80°C until use. Protein concentrations were determined using the BCA method with bovine serum albumin (BSA) as a standard (Smith et al., 1985).

[³H]NαMH and [³H]A-349821 – saturation studies

Binding assays were conducted as described previously (Esbenshade & Hancock, 2000). Briefly, dilutions of radioligands and competitors were made with Tris–EDTA buffer (50 mM Tris, 5 mM EDTA, pH 7.4 at room temperature). Tris–EDTA buffer with 0.2% BSA was included for all studies using [³H]A-349821 to effectively prevent adsorption to labware surfaces. Membrane preparations were thawed and homogenized (polytron) in various volumes of Tris–EDTA buffer to give protein concentrations of ∼70, ∼60, ∼600, and ∼1000 μg ml⁻¹ for membranes prepared from C6 cells stably expressing rat or human H₃ receptors, and from rat and human cortices, respectively. Assay components were combined in 96-well BioBlocks™ in duplicate to give a final volume of 0.5 ml, or for recombinant human, 1 ml. Membrane preparations were incubated with various concentrations of [³H]NαMH (∼0.1–6 nM) or [³H]A-349821 (∼0.01 to ∼6 nM) at 25°C for 30 or 60 min, respectively. Nonspecific binding was defined by the presence of 10 μM thioperamide. Binding reactions were terminated by rapid filtration through Whatman GF/B filters pre-soaked in 0.3% polyethylenimine in Tris buffer (50 mM Tris, pH 7.4 at 4°C); filters were washed (4 × 2 ml) with ice-cold Tris buffer using a Brandel Cell Harvester (Brandel Inc., Gaithersburg, MD, U.S.A.). Filters were transferred into scintillation vials, Ultima-Gold™ scintillant (PerkinElmer, Boston, MA, U.S.A.) was added, and after 2 h the bound radioactivity was determined by counting for 3 min in a Packard liquid scintillation counter.

[³H]A-349821 – kinetic studies

Kinetic studies were performed with ∼4 nM [³H]A-349821 (membranes from C6 cells stably expressing the recombinant rat H₃ receptor or rat cortices) or ∼1 nM [³H]A-349821 (membranes from C6 cells stably expressing the recombinant human H₃ receptor) as described above with minor modifications. To determine the observed association rates, triplicate membrane samples were incubated with [³H]A-349821 for increasing time intervals ranging from 0.33 to 90 min. Nonspecific binding for each time point was defined by the presence of 10 μM thioperamide. For dissociation kinetic studies, the same membrane preparation was incubated with [³H]A-349821 1 h before dissociation was initiated by exposure to 10 μM thioperamide for increasing time intervals ranging from 1 to 120 min. Filtration and counting methods were as described above.

[³H]NαMH and [³H]A-349821 – competition studies

Competition-binding assays were performed with minor modifications to the binding assays described above. Membranes were incubated with 0.5 nM [³H]NαMH or 1.5 nM [³H]A-349821 (membranes from C6 cells stably expressing rat H₃ receptors or rat cortices), or 0.5 nM [³H]A-349821 (membranes from C6 cells stably expressing human H₃ receptors) in the presence or absence of competitors. Nonspecific binding was defined by the presence of 10 μM thioperamide. Binding reactions were carried out for 30 min ([³H]NαMH) or 60 min ([³H]A-349821) at 25°C. To study the effects of GDP on binding, assays were carried out in the absence or presence of 100 μM GDP with an incubation time of 60 min for all experiments. For membranes from C6 cells stably expressing rat or human H₃ receptors, binding reactions were terminated by rapid filtration through 96-well GF/B-Unifilters™ presoaked in 0.3% polyethylenimine in Tris buffer, which were washed (4 × 2 ml) with ice-cold Tris buffer using a Packard Unifilter harvester. Filters were allowed to dry, PerkinElmer MICROSCINT™ 20 scintillant was added (40 μl well⁻¹) and bound radioactivity was determined by counting in a Packard Topcount scintillation counter for 3 min. Binding assays with membranes from rat cortices were terminated and processed using the Brandell harvester, as described above.

Data analysis

K_D values from saturation-binding assays were evaluated by one-site and two-site binding methods using Prism™ software (GraphPad, San Diego, CA, U.S.A.). K_D values reported are based on comparison of one-site and two-site curve fits with an F-test in which the sum-of-squares and degrees of freedom are compared for each fit (significance at a 95% confidence limit).

Observed rate constants (k_ob) or dissociation rate constants (k_off) were analyzed with Prism™ software using one-phase and two-phase exponential association or one-phase and two-phase exponential decay parameters. Rate constant values reported are based on comparison of one-phase and two-phase curve fits with an F-test in which the sum of squares and degrees of freedom are compared for each fit (significance at a 95% confidence limit). Association rate constants (k_on) were defined as (k_ob−k_off)/[radioligand] and dissociation constants (K_D) were defined as k_off/k_on.

Competition data were analyzed with Prism™ software using both one-site and two-site competition parameters. The number of binding sites reported is based on comparison of one-site and two-site curve fits with an F-test in which the sum of squares and degrees of freedom are compared for each fit (significance at a 95% confidence limit). IC₅₀ values were converted to K_i values using the generalized Cheng–Prussoff equation (Cheng & Prusoff, 1973; Leff & Dougall, 1993). The paired t-test (two-tailed) with significance determined at the 95% confidence interval (P<0.05) was used for all other statistical comparisons.

Drugs and chemicals

Histamine and clobenpropit were purchased from Sigma Chemical Co. (St Louis, MO, U.S.A.). (N)-α-methylhistamine (NαMH), (R)-α-methylhistamine (RαMH), (S)-α-methylhistamine (SαMH), imetit, immipep, iodophenpropit, and thioperamide were purchased from Tocris-Cookson (Ellisville, MO, U.S.A.). Ciproxifan, proxyfan, and GT-2331 ((1S, 2S)-(+)-4-[2-(5,5-dimethyl-hex-1-ynyl)-cyclopropyl]-1H-imidazole) (Liu et al., 2004) were synthesized at Abbott Laboratories. Nonlabelled A-349821 ({4′-[3-(2R,5R-dimethyl-pyrrolidin-1-yl)-propoxy]-biphenyl-4-yl}-morpholin-4-yl-methanone) and other A-compounds were prepared at Abbott laboratories as described previously (Faghih et al., 2002, 2003a, 2003b; Turner et al., 2003). Radiolabelled [³H]NαMH was purchased from Perkin-Elmer Life Sciences (Boston, MA, U.S.A.). Aprotinin, leupeptin, and pepstatin were purchased from Roche Molecular Biochemicals (Indianapolis, IN, U.S.A.). BSA (fatty acid free, fraction V) was from ICN Biomedicals (Aurora, OH, U.S.A.). Dulbeco's modified Eagle's cell culture medium, hygromycin B, and lipofectamine 2000 reagent were obtained from Invitrogen/LifeTech (Grand Island, NY, U.S.A.). The rat C6 glioma cell line was obtained from the American Type Culture Collection (Rockville, MD, U.S.A.).

Results

Saturation analysis: [³H]A-349821 and [³H]NαMH

Saturable specific binding of [³H]A-349821 (for chemical structure, see Figure 1) was observed using membranes from clonal cell lines stably expressing recombinant rat or human H₃ receptors and in membranes prepared from rat and human cortices (Figures 2a–d). Nonspecific binding, at [³H]A-349821 concentrations equivalent to the K_D, were 1.7, 4.3, 8.3, and 9.2% for membranes prepared from C6 cells stably expressing rat or human H₃ receptors, and rat and human cortices, respectively. Individual plots of specific binding versus [³H]A-349821 concentrations were, in all cases, best fit by the one-site equation over the concentration range of [³H]A-349821 tested. Scatchard plots appeared linear, also consistent with binding to a single site over the concentration range of [³H]A-349821 tested (Figure 2e–h). Calculated pK_D values for [³H]A-349821 binding to membranes prepared from C6 cells stably expressing rat H₃ receptors and rat cortices were 9.75 and 9.32, respectively, while somewhat higher pK_D values were obtained for membranes prepared from C6 cells stably expressing human H₃ receptors and human cortices (10.54 and 10.50, respectively), showing that [³H]A-349821 labels recombinant and native rat and human H₃ receptors with high affinity (Table 1). Estimated B_max values were 2190, 2240, 227, and 46.6 fmol mg protein⁻¹ for membranes prepared from C6 cells stably expressing rat or human H₃ receptors and rat and human cortices, respectively. Parallel saturation assays were conducted with the H₃ receptor agonist radioligand, [³H]NαMH (Figure 2a–d). Analysis of specific binding versus [³H]NαMH concentration showed a preference for curve fitting by the two-site equation, although the maximal concentrations of [³H]NαMH tested were not sufficient to adequately describe the low-affinity site. Scatchard transformations tended to deviate from linearity, consistent with the presence of both high- and low-affinity binding sites (Figure 2e–h, insets). Higher [³H]NαMH concentrations were not tested because of poor signal to noise ratios. Due to this limitation, binding parameters for the low-affinity component were ill-defined and only results for the one-site analysis of the saturation data are reported (Table 1). In contrast to [³H]A-349821, [³H]NαMH bound with similar affinity to all membrane preparations, giving pK_D values ranging from 9.19 to 9.31 (Table 1). Estimated B_max values were 1300, 1170, 140, and 29.8 fmol mg protein⁻¹ for membranes prepared from C6 cells stably expressing rat or human H₃ receptors and rat and human cortices, respectively, showing that, over the concentration range tested, [³H]NαMH labelled a significantly lower number of receptor sites (52–64%) compared with [³H]A-349821 (Table 1 and Figure 2a–h). Specific binding of [³H]A-349821 was not observed in membranes prepared from untransfected C6 cells, in membranes expressing human H₄ receptors, human α_2c adrenoceptors, or human 5-HT₃ receptors using concentrations up to ∼6 nM (data not shown). Specificity of [³H]A-349821 at H₃ receptors was also demonstrated since chlorpheniramine and ranitidine (H₁ and H₂ receptor ligands, respectively) had negligible affinity (pK_i<5) in competition-binding studies with [³H]A-349821 (data not shown).

Figure 2.

Representative saturation curves from a total of four experiments each for binding of either [³H]A-349821 (solid lines) or [³H]NαMH (dashed lines) to sites in membranes expressing recombinant rat H₃ receptors (a), cortical rat H₃ receptors (b), recombinant human H₃ receptors (c), or cortical human H₃ receptors (d). Panels e–h are Scatchard transformation plots of panels a–d, respectively. For [³H]A-349821 saturation studies, ∼18, ∼150, ∼15, and 250 μg of membrane protein from membranes expressing recombinant rat H₃ receptors, cortical rat H₃ receptors, recombinant human H₃ receptors, or cortical human H₃ receptors, respectively, were incubated for 60 min at 25°C, in total volumes of 0.5 or 1 ml (recombinant human), with various concentrations of [³H]A-349821 in combination with buffer (total binding) or 10 μM thioperamide (nonspecific binding). [³H]NαMH saturation-binding experiments were carried out similarly, but with a 30-min incubation time. All determinations were from duplicate tubes. Saturation-binding curves were analyzed by the one-site binding hyperbolic fit method using Prism software.

Table 1.

[³H]A-349821 and [³H]NαMH saturation-binding parameters from binding to rat or human H₃ receptors in the absence and presence of 100 μM GDP

		[3H]A-349821		[3H]NαMH
	GDP	pKD	Bmax (fmol mg protein−1)	pKD	Bmax (fmol mg protein−1)
Rat H3	−	9.75±0.12	2190±100	9.31±0.02	1300±130a
	+	9.69±0.17	2350±70	8.59±0.02b	1140±40a
Rat cortex	−	9.32±0.11	227±24	9.19±0.01	140±7a
	+	9.39±0.05	224±15	8.98±0.02b	119±4a,c
Human H3	−	10.54±0.06	2240±190	9.29±0.03	1170±80a
	+	10.59±0.06	2450±190	8.55±0.04b	1160±70a
Human cortex	−	10.50±0.02	46.6±1.8	9.19±0.07	29.8±1.6a
	+	10.49±0.02	48.6±2.0	8.96±0.03b	27.3±0.9a

Mean pK_D (−log dissociation constant) values and B_max (maximal binding density) were calculated from saturation analysis of [³H]A-349821 (0.01–6 nM) and [³H]NαMH (0.1–6 nM) binding to rat or human H₃ receptors. Data are the mean±s.e.m. of four experiments conducted with duplicate determinations.

^aSignificance of lower B_max values for [³H]NαMH versus [³H]A-349821 (P<0.05).

^bSignificance of lower pK_D values with the same radiolabel in the presence of GDP compared to its absence (P<0.05).

^cSignificance of lower B_max values in the presence of GDP compared to its absence (P<0.05).

Kinetic studies: [³H]A-349821

Association and dissociation rate profiles were measured using concentrations of [³H]A-349821 equal to ∼5 nM for membranes from C6 cells or cortices expressing rat H₃ receptors or ∼1 nM for membranes from C6 cells expressing human H₃ receptors. These concentrations of radioligand were selected to give high receptor occupancy and to maximize signals, particularly at early time points. Due to low H₃ receptor expression levels in human cortices and limited availability, kinetic studies were not carried out with these tissues. Specific binding of [³H]A-349821 in membranes prepared from C6 cells stably expressing rat H₃ receptors was extremely rapid, achieving equilibrium within 10 min, while displacement of [³H]A-349821 by 10 μM thioperamide was rapid and virtually complete within 15 min (Figure 3). Association and dissociation rate data appeared to be monophasic, consistent with binding to a single site. Rate data for membranes prepared from rat cortices and C6 cells stably expressing human H₃ receptors were similar in this regard (data not shown). Kinetic rate and dissociation constants (k_on, k_off) are summarized in Table 2. Mean rates for membranes from C6 cells expressing rat H₃ receptors and rat cortices were similar (k_on=1.79 and 0.80 min⁻¹ nM⁻¹, respectively; k_off=0.38 and 0.41 min⁻¹, respectively), while the association rate determined in membranes from C6 cells stably expressing human H₃ receptors was markedly faster (k_on=3.48 min⁻¹ nM⁻¹) and the dissociation rate was conversely slower (k_off=0.07 min⁻¹). As a result, the dissociation constants for [³H]A-349821 calculated using k_on and k_off values revealed lower affinity values in membranes from C6 cells expressing rat H₃ receptors and rat cortices (pK_D(Kinetic)=9.67 and 9.29, respectively) compared to membranes from C6 cells stably expressing human H₃ receptors (pK_D(Kinetic)=10.7). Kinetic dissociation constant values were similar to those determined by saturation-binding analyses (Table 2).

Figure 3.

Representative association–dissociation plots from a total of four experiments for specific binding of [³H]A-349821 to membranes expressing recombinant rat H₃ receptors. The association rate experiment was performed by incubating [³H]A-349821 (125 μl; 4 nM final concentration) for increasing times (0.33–90 min) in triplicate tubes containing membranes (250 μl; 0.03 mg ml⁻¹ protein final), 100 μl of buffer and 25 μl of buffer or thioperamide (10 μM final concentration) to define total and nonspecific binding, respectively (a). The dissociation rate experiment was performed by incubating [³H]A-349821 (125 μl; 2 nM final concentration) for 60 min in tubes containing membranes (250 μl; 0.03 mg ml⁻¹ protein final), 125 μl of buffer in one set of triplicate tubes to define total binding, and 100 μl of buffer and 25 μl of thioperamide (10 μM final concentration) in another set of triplicate tubes to define nonspecific binding. At this time, 25 μl of 10 μM thioperamide were added to the total binding tubes, and binding was measured over increasing times (1–120 min) (b). All experiments were carried out at 25°C in triplicate tubes. Association and dissociation rates were estimated using the ‘one-phase exponential' and ‘one-phase exponential decay' equations, respectively, using Prism software.

Table 2.

Kinetic analysis of [³H]A-349821 binding to membranes prepared from C6 cells stably expressing recombinant rat or human H₃ receptors and rat cortices

	kon (min−1 nM−1)	koff (min−1)	pKD (kinetic)	pKD (saturation)a
Rat H3	1.79±0.23	0.38±0.03	9.67±0.09	9.75±0.12
Rat cortex	0.80±0.16	0.41±0.09	9.29±0.09	9.32±0.11
Human H3	3.48±0.67	0.07±0.02	10.7±0.1	10.5±0.1

^aFor comparison, pK_D values from Table 1 are included (N=4).

Competition studies: [³H]A-349821 and [³H]NαMH

A panel of reference H₃ receptor ligands, including previously described H₃ receptor antagonists/inverse agonists that represent three lead chemical series (for chemical structures, see Figure 4) developed by Abbott Laboratories (Table 3 and Figure 5), as well as a large panel of lead series analogs (Table 4), were assessed for their ability to compete with [³H]A-349821 binding to membranes prepared from C6 cells expressing recombinant rat or human H₃ receptors or from rat cortices. Individual competition curves for H₃ receptor ligands previously characterized as full agonists (histamine, RαMH, NαMH, SαMH, immepip, and imetit) were biphasic and best described by two-site curve fits (Figure 5a–c and Table 3). Individual competition curves for proxyfan, an H₃ receptor ligand having either agonist or antagonist functional properties depending on the assay system (Morisset et al., 2000; Gbahou et al., 2003; Krueger et al., 2005), were best described by two-site curve fits in all experiments performed with membranes from C6 cells stably expressing human H₃ receptors. However, a mixture of one- or two-site curve fits was observed in individual experiments performed using membranes prepared from C6 cells expressing rat H₃ receptors, while the combined competition curves were best described by two-site curve fits (Table 3 and Figure 5d–f). Varied curve fitting was also observed in individual competition curves for GT-2331, a compound having mixed functional properties similar to proxyfan (Wulff et al., 2002; Krueger et al., 2005), while combined competition curves were overall best described by two-site curve fits for membranes from C6 cells stably expressing rat or human H₃ receptors, but one site for membranes from rat cortices (Table 3 and Figure 5d–f). For compounds best described by two-site curve fits, the average percentages of high-affinity H₃ receptor sites were 53, 46, and 38% in membranes prepared from C6 cells expressing rat H₃ receptors, rat cortices, and membranes prepared from C6 cells expressing human H₃ receptors, respectively (Table 3).

Figure 4.

Structures for lead compounds representing three chemical series for the analogues listed in Table 4: the piperazine, A-317920 (a); the biaryl, A-331440 (b); the tricyclic, A-347166 (c); and the biaryl, A-349821 (d).

Table 3.

Parameters for multiple data sets from competition experiments using [³H]A-349821 or [³H]NαMH in membranes expressing rat or human H₃ receptors

	[3H]A-349821			[3H]NαMH
Compound	Rat H3	Rat Cortex	Human H3	Rat H3	Rat Cortex	Human H3
	pKi (F)	pKi (F)	pKi (F)	pKi	pKi	pKi
	[nH]	[nH]	[nH]	[nH]	[nH]	[nH]
Histamine	8.25±0.18 (57%)	8.34±0.20 (51%)	8.21±0.14 (41%)	8.56±0.02	8.42±0.02	8.57±0.01
(ID=1)	6.60±0.26 (43%)	6.83±0.11 (49%)	6.19±0.41 (59%)	[0.94±0.01]	[0.95±0.01]	[0.95±0.01]
RαMH	9.14±0.43 (53%)	9.11±0.24 (43%)	9.10±0.09 (37%)	9.29±0.15	8.79±0.08	9.28±0.06
(ID=2)	7.40±0.42 (47%)	7.73±0.13 (57%)	7.13±0.10 (63%)	[0.91±0.05]	[0.93±0.06]	[0.92±0.03]
NαMH	9.58±0.66 (52%)	9.49±0.13 (48%)	9.23±0.14 (38%)	9.50±0.09	8.98±0.11	9.48±0.04
(ID=3)	7.58±0.36 (48%)	7.93±0.08 (52%)	7.16±0.08 (62%)	[0.84±0.04]	[0.88±0.04]	[0.94±0.03]
SαMH	7.69±0.37 (38%)	7.49±0.16 (54%)	8.39±0.26 (30%)	7.72±0.14	7.49±0.08	8.15±0.05
(ID=4)	6.24±0.21 (62%)	6.03±0.18 (46%)	6.13±0.16 (70%)	[1.04±0.04]	[0.93±0.04]	[0.83±0.05]
Immepip	9.67±0.29 (59%)	9.99±0.26 (46%)	9.54±0.12 (43%)	9.93±0.06	9.29±0.15	9.83±0.03
(ID=5)	8.10±0.78 (41%)	8.43±0.26 (54%)	7.84±0.10 (57%)	[1.05±0.06]	[1.15±0.21]	[1.05±0.05]
Imetit	9.84±0.26 (53%)	10.1±0.1 (40%)	9.44±0.16 (35%)	10.1±0.1	9.53±0.09	9.59±0.09
(ID=6)	8.23±0.20 (47%)	8.68±0.09 (60%)	7.65±0.05 (65%)	[0.95±0.05]	[0.96±0.04]	[1.00±0.03]
Proxyfan	8.40 (39%)a	8.21 (73%)a	8.26±0.14 (32%)	8.62±0.10	8.51±0.05	8.35±0.06
(ID=7)	7.29 (61%)a	7.37 (27%)a	6.57±0.08 (68%)	[0.88±0.04]	[0.89±0.03]	[0.95±0.04]
GT-2331	9.95 (38%)a	9.22±0.13	7.96±0.15 (32%)	10.0±0.1	9.63±0.06	8.39±0.07
(ID=8)	9.14 (62%)a	[1.27±0.06]	6.63±0.04 (68%)	[1.03±0.04]	[1.04±0.03]	[1.03±0.02]
Clobenpropit	10.2±0.1	10.0±0.2	9.91±0.30	9.82±007	9.45±0.08	9.47±0.04
(ID=9)	[1.29±0.16]	[1.21±0.03]	[1.05±0.11]	[1.08±0.03]	[0.98±0.03]	[1.01±0.03]
Thioperamide	8.74±0.22	8.43±0.23	7.54±0.21	8.63±0.05	8.20±0.07	7.24±0.07
(ID=10)	[0.95±0.06]	[1.12±0.02]	[1.00±0.11]	[0.81±0.03]	[0.89±0.03]	[0.90±0.03]
Ciproxifan	9.53±0.41	9.17±0.35	7.21±0.36	9.38±0.07	9.22±0.05	7.20±0.04
(ID=11)	[1.15±0.01]	[1.12±0.06]	[0.80±0.12]	[0.90±0.02]	[0.89±0.03]	[0.84±0.03]
Iodophenpropit	10.1±0.1	9.69±0.05	9.61±0.23	9.83±0.08	9.39±0.08	9.14±0.05
(ID=12)	[1.33±0.03]	[1.16±0.02]	[1.01±0.06]	[1.03±0.08]	[1.06±0.07]	[0.92±0.04]
A-317920b	9.95±0.31	9.31±0.24	7.51±0.31	9.24±0.08	9.13±0.04	7.02±0.04
(ID=13)	[1.00±0.05]	[1.20±0.02]	[0.78±0.12]	[0.85±0.02]	[0.92±0.03]	[0.89±0.03]
A-331440c	8.53±0.12	7.74±0.11	9.00±0.14	8.27±0.05	7.75±0.06	8.50±0.04
(ID=14)	[1.00±0.10]	[1.14±0.07]	[1.10±0.23]	[0.90±0.05]	[0.81±0.04]	[0.91±0.03]
A-347166d	9.02±0.09	8.69±0.09	9.04±0.12	8.95±0.7	8.60±0.05	8.80±0.07
(ID=15)	[1.09±0.08]	[1.01±0.02]	[1.14±0.25]	[0.80±0.07]	[0.81±0.03]	[0.84±0.02]
A-349821e	9.73±0.20	9.33±0.10	10.5±0.4	9.01±0.12	8.77±0.05	9.46±0.07
(ID=16)	[1.07±0.09]	[1.17±0.07]	[1.13±0.18]	[0.88±0.04]	[0.85±0.03]	[0.93±0.02]

^aBinding parameters based on curve fitting to compiled competition curves.

^bFaghih et al. (2003a) (Compd 8). For chemical structure, see Figure 4a.

^cFaghih et al. (2002) (Comp 31). For chemical structure, see Figure 4b.

^dTurner et al. (2003) (Compd 23). For chemical structure, see Figure 4c.

^eFaghih et al. (2003b) (Compd A6). For chemical structure, see Figure 4d.

Mean pK_i (−log inhibition constant) for compounds displaying two-site competition with high-affinity values (upper line) and low-affinity values (lower line), as well as the percentage fraction, F (in parenthesis), are shown. For compounds displaying one-site competition, affinities are expressed as mean pK_i±s.e.m., with mean n_H±s.e.m. shown on the second line in brackets (N=4–451).

Figure 5.

Displacement curves for [³H]A-349821 in competition with a panel of reference H₃ receptor agonists (a–c) and selected H₃ receptor ligands with mixed pharmacology (proxyfan and GT-2331) or antagonists/inverse agonists (d–f) in membranes expressing recombinant rat H₃ receptors (a and d), rat cortical H₃ receptors (b and e) or recombinant human H₃ receptors (c and f). s.e.m. bars have been omitted for the purpose of clarity. s.e.m. values were typically less than ±10% of the mean. [³H]A-349821 concentrations were ∼1.8, ∼1.6 and ∼0.6 nM for membranes expressing recombinant rat H₃ receptors, rat cortical H₃ receptors, or recombinant human H₃ receptors, respectively. Nonspecific binding was defined by the presence of 10 μM thioperamide. All experiments were carried at 25°C for 60 min with duplicate tubes. Data represent mean±s.e.m. of three or four experiments where each point was determined in duplicate. The curves shown superimposed on the mean data points were described with the best fit (F-test at 95% confidence) by either the ‘one-site competition' or ‘two-site competition' equations using Prism software.

Table 4.

Parameters for multiple data sets from competition experiments using [³H]A-349821 or [³H]NαMH in membranes expressing human or rat H₃ receptors

		[3H]A-349821			[3H]NαMH
Compound	Chemical	Rat H3	Rat Cortex	Human H3	Rat H3	Rat Cortex	Human H3
	Series	pKi	pKi	pKi	pKi	pKi	pKi
		[nH]	[nH]	[nH]	[nH]	[nH]	[nH]
1aA-347187	Tricyclic	6.81±0.08	6.53±0.09	6.42±0.14	6.57±0.06	6.40±0.13	6.12±0.07
		[0.83±0.19]	[1.08±0.03]	[0.86±0.11]	[0.75±0.05]	[0.77±0.05]	[0.92±0.09]
1bA-349951	Tricyclic	6.93±0.35	6.53±0.30	7.26±0.30	7.10±0.04	7.01±0.15	7.29±0.10
		[1.03±0.26]	[0.87±0.06]	[0.87±0.16]	[0.81±0.02]	[0.71±0.10]	[0.83±0.07]
2A-311346	Biaryl	7.77±0.08	6.90±0.19	8.09±0.12	7.74±0.08	7.20±0.33	8.06±0.14
		[1.30±0.10]	[1.03±0.20]	[0.89±0.10]	[0.94±0.10]	[0.78±0.11]	[0.90±0.07]
3aA-349762	Biaryl	8.71±0.17	8.50±0.09	9.55±0.17	8.48±0.02	8.55±0.07	9.08±0.12
		[1.10±0.17]	[1.32±0.10]	[1.09±0.17]	[0.97±0.05]	[0.70±0.02]	[0.88±0.06]
3bA-349763	Biaryl	8.70±0.06	8.43±0.15	9.23±0.26	8.81±0.09	8.42±0.07	8.68±0.12
		[0.92±0.17]	[1.22±0.08]	[1.45±0.32]	[0.78±0.07]	[0.79±0.01]	[0.97±0.14]
3cA-349765	Biaryl	7.35±0.64	7.11±0.72	7.87±0.69	6.59±0.06	6.39±0.10	6.96±0.09
		[1.12±0.24]	[1.23±0.05]	[1.10±0.19]	[0.80±0.07]	[0.62±0.04]	[0.75±0.05]
3dA-349768	Biaryl	8.71±0.07	8.56±0.06	9.42±011	8.97±0.08	8.69±0.07	9.10±0.09
		[1.21±0.11]	[1.09±0.04]	[1.26±0.19]	[0.76±0.08]	[0.79±0.07]	[0.86±0.05]
3eA-349771	Biaryl	7.45±0.12	7.26±0.06	8.08±0.11	7.52±0.04	7.38±0.18	8.01±0.14
		[1.13±0.11]	[1.33±0.14]	[1.18±0.19]	[0.80±0.01]	[0.68±0.06]	[0.82±0.11]
3fA-349774	Biaryl	7.86±0.01	7.55±0.08	8.59±0.10	7.55±0.05	7.67±0.09	8.20±0.12
		[1.05±0.15]	[1.26±0.01]	[1.14±0.16]	[0.94±0.07]	[0.82±0.10]	[0.95±0.06]
3gA-349777	Biaryl	6.84±0.16	6.69±0.04	7.29±0.16	6.54±0.02	6.96±0.09	7.61±0.11
		[1.20±0.27]	[1.17±0.11]	[1.27±0.14]	[0.85±0.09]	[0.68±0.09]	[0.78±0.09]
3hA-349778	Biaryl	6.39±0.15	6.24±0.26	6.68±0.13	6.32±0.07	6.66±0.18	6.64±0.11
		[0.90±0.17]	[1.15±0.07]	[0.99±0.10]	[0.82±0.11]	[0.56±0.20]	[0.73±0.11]
3iA-349785	Biaryl	6.77±0.08	6.23±0.07	7.00±0.15	6.62±0.11	6.33±0.15	6.74±0.05
		[1.03±0.19]	[1.28±0.27]	[1.08±0.21]	[1.03±0.07]	[0.74±0.13]	[0.86±0.09]
3jA-349786	Biaryl	8.39±0.13	7.79±0.20	8.98±0.11	8.55±0.11	8.08±0.13	8.46±0.12
		[1.17±0.34]	[1.23±0.12]	[1.22±0.24]	[0.86±0.17]	[0.79±0.11]	[0.96±0.06]
3kA-349816	Biaryl	7.29±0.33	6.62±0.22	7.62±0.29	7.26±0.03	7.14±0.07	7.63±0.17
		[1.05±0.16]	[1.49±0.19]	[1.23±0.19]	[0.93±0.04]	[0.83±0.02]	[0.90±0.12]
3lA-349822	Biaryl	7.94±0.16	8.03±0.17	8.36±0.15	8.02±0.05	7.97±0.10	8.13±0.11
		[1.43±0.22]	[1.24±0.10]	[1.32±0.17]	[0.93±0.02]	[0.81±0.08]	[1.00±0.08]
3mA-350810	Biaryl	7.53±0.11	7.25±0.09	8.36±0.12	7.38±0.07	6.76±0.26	7.39±0.15
		[1.16±0.36]	[1.29±0.08]	[1.25±0.31]	[0.78±0.05]	[0.69±0.14]	[0.72±0.05]

¹For chemical structures, see Turner et al. (2003) (1a=Compd. 27, 1b=Compd. 28).

²For chemical structures, see Faghih et al. (2002) (2=Compd. 29).

³For chemical structures, see Faghih et al. (2003b) (3a=Compd. F6, 3b=Compd. F5, 3c=Compd. G1, 3d=Compd. G6, 3e=Compd. D1, 3f=Compd. D3, 3g=Compd. D7, 3h=Compd. E1, 3i=Compd. B1, 3j=Compd. B2, 3k=Compd. B7, 3l=Compd. A5, 3m=Compd. C7).

Mean pK_i (−log inhibition constant)±s.e.m., with mean n_H±s.e.m. in brackets shown below (N=3–31).

By contrast, competition curves for thioperamide and clobenpropit, both reported to have antagonist/inverse agonist attributes in functional studies (Wieland et al., 2001), were monophasic and best described by one-site curve fits (Table 3 and Figure 5d–f). Similarly, competition curves for four Abbott compounds, representing three distinct chemical series (for chemical structures, see Figure 4), and previously shown to have antagonist/inverse agonist functional properties (Esbenshade et al., 2003; 2004; Turner et al., 2003; Hancock et al., 2004), were monophasic and best described by one-site curve fits (Table 3 and Figure 5d–f). The imidazoles thioperamide and clobenpropit displaced [³H]A-349821 to the same level as the non-imidazole, A-317920 (Figure 5d–f). In addition, a large panel of Abbott compounds from these three chemical series (previously described as H₃ receptor antagonists/inverse agonists), which displayed a wide range of potencies in the displacement of [³H]NαMH binding (Faghih et al., 2002, 2003a, 2003b; Turner et al., 2003), were tested in competition with [³H]A-349821. For these compounds as well, competition curves displayed Hill slopes close to or slightly above unity and were best described by the one-site analysis (Table 4).

For comparison, parallel studies were carried out for the complete panel of compounds in competition-binding assays with the H₃ receptor agonist radioligand [³H]NαMH (Tables 3 and 4). Competition curves for all compounds displayed Hill slopes close to or slightly below unity and were best described by the one-site curve fit (Tables 3 and 4).

pK_i values for antagonists (denoted by pK_H) and high- and low-affinity pK_i values for agonists (denoted by pK_H and pK_L, respectively) determined by displacement of [³H]A-349821 binding to membranes from C6 cells expressing rat H₃ receptors (Figure 6a), human H₃ receptors (Figure 6c), or rat cortices (Figure 6b) were correlated to pK_i values for antagonists and agonists obtained by displacement of [³H]NαMH binding. Two distinct regression lines with slopes close to unity were revealed (Figure 6a–c). One regression line was close to the line of unity, as described by pK_i values for displacement of [³H]NαMH (pK_i) versus high-affinity displacement of [³H]A-349821 from a high-affinity site recognized by both antagonists and agonists (pK_H), while another was frame shifted to the right, as described by pK_i values for displacement of [³H]NαMH (pK_i) versus displacement of [³H]A-349821 from a low-affinity site recognized by agonists (pK_L) (Figure 6). In all instances, regression lines were highly correlated (R²⩾0.82).

Figure 6.

Regression lines showing mean affinities for H₃ receptor ligands estimated from competition with [³H]NαMH (pK_i) or [³H]A-349821 (pK_H and pK_L) in membranes expressing recombinant rat H₃ receptors (a), rat cortical H₃ receptors (b), and recombinant human H₃ receptors (c). The solid line is the linear regression for pK_i versus pK_H, the dashed line is the linear regression for pK_i versus pK_L, and the dotted line is the line of identity. Compounds of interest are notated and can be identified by the corresponding compound ID values and compound names included in Table 3. Compounds 1–6 are reference agonists, compounds 7 and 8 (proxyfan and GT-2331, respectively) display mixed pharmacology and compounds 9–16 are antagonists/inverse agonists. Symbols not marked by a number refer to the compounds listed in Table 4.

Saturation studies using [³H]A-349821 and [³H]NαMH: the effect of GDP

The effect of GDP (100 μM) on saturation binding of [³H]A-349821 and [³H]NαMH was investigated, since guanine nucleotides are known to affect agonist affinities for many GPCRs. K_D and B_max values for [³H]A-349821 binding were largely insensitive to GDP in all membranes tested (Table 1 and Figure 7). By contrast, K_D values for [³H]NαMH binding to all membranes were significantly reduced by addition of GDP (P<0.05) (Table 1 and Figure 7). Similarly, B_max values tended to be reduced by addition of GDP, but they were marginally significant only for membranes from rat cortices (P=0.048).

Figure 7.

Representative saturation curves from a total of four experiments for binding of either [³H]A-349821 (open symbols) or [³H]NαMH (solid symbols) to sites in membranes expressing recombinant rat H₃ receptors (a), cortical rat H₃ receptors (b), recombinant human H₃ receptors (c), or cortical human H₃ receptors (d) in the absence or presence of 100 μM GDP. Panels e–h are Scatchard transformation plots of panels a–d, respectively. All determinations were from duplicate tubes. Saturation studies and data analysis were performed as described in Figure 2, but include simultaneous determinations in the presence of GDP.

[³H]A-349821 displacement by histamine and clobenpropit: the effect of GDP

Displacement of [³H]A-349821 binding by the endogenous H₃ receptor agonist, histamine, displayed shallow slopes that were best fitted by the two-site model, whereas competition curves performed in the presence of GDP were shifted to the right, showing an overall reduction in histamine affinity and an increase in Hill slope values in all membranes tested (Figure 8a–c). However, the displacement curves in the presence of GDP were still best fit by the two-site model. The GDP effect was most pronounced when membranes expressing recombinant human H₃ receptors were utilized (Figure 8c). Competition binding with the H₃ receptor antagonist/inverse agonist clobenpropit was, in all cases, monophasic and completely insensitive to GDP (Figure 8d–f).

Figure 8.

Representative [³H]A-349821 displacement curves from a total of three experiments for histamine (a–c) or clobenpropit (d–f) in membranes expressing recombinant rat H₃ receptors (a and d), rat cortical H₃ receptors (b and e), and recombinant human H₃ receptors (c and f), with (open circles) and without GDP (closed circles). All determinations were from duplicate tubes. Competition studies and data analysis were performed as described in Figure 5, but included addition of either buffer or GDP (100 μM, final concentration) immediately prior to addition of [³H]A-349821.

Discussion

The present study describes the characterization of a novel and highly specific, non-imidazole-containing, H₃ receptor antagonist/inverse agonist radioligand, [³H]A-349821, and its use in comparing the pharmacology displayed by membranes expressing recombinant or native H₃ receptors from rat or human sources. Specific [³H]A-349821 binding was saturable and displayed high affinity in these membranes. [³H]A-349821 appeared to saturate a single pool of receptors and exhibited very low nonspecific binding, even in membranes from rat cortices (Figure 2), where previously described imidazole-based H₃ receptor antagonist radioligands have been reported to significantly label H₄ receptors, 5-HT₃ receptors or cytochrome P450 sites (Labella et al., 1992; Leurs et al., 1995; Schlicker et al., 1995; Alves-Rodrigues et al., 1996; Harper et al., 1999; Yang et al., 2002). Specific binding of [³H]A-349821 was absent in membranes from the parent cell line and from cell lines expressing α_2c adrenoceptors, H₄, or 5-HT₃ receptors, while specific binding was insensitive to competition by H₁ and H₂ ligands in membranes expressing H₃ receptors (data not shown). Thus, [³H]A-349821 demonstrated high specificity for the H₃ receptor, consistent with previous studies in which unlabelled A-349821 used in competition assays showed low affinity (μM) for a large number of GPCRs and ion channels, although weak potency towards α_2c adrenoceptors (IC₅₀=250 nM), was observed (Esbenshade et al., 2004). While this potency was similar to the potency of [¹²⁵I]Iodoproxyfan for α₂ adrenoceptors (pK_i=162 nM) (Schlicker et al., 1995), disparate interaction of these radioligands with cytochrome P450 sites was found. Previous studies (Yao et al., 2005) showed that [¹²⁵I]iodoproxyfan binding was displaced differentially by thioperamide, which competed for a low-affinity binding site that was sensitive to the cytochrome P450 inhibitor metyrapone, whereas the non-imidazole A-317920 did not interact with this site. In the current study, [³H]A-349821 binding was displaced equally by both thioperamide and A-317920 (Figure 5d–f) and saturation analysis with [³H]A-349821 (0.06–6 nM) showed identical inhibition of binding by thioperamide and A-317921 (data not shown), suggesting that, unlike [¹²⁵I]Iodoproxyfan, [³H]A-349821 does not label cytochrome P450 sites.

While [³H]A-349821 bound to an apparent single class of receptors, [³H]NαMH saturation binding suggested interaction with both high- and low-affinity sites, although accurate estimation of the low-affinity binding parameters was not possible and, as such, saturation data are reported for single-site occupancy. By contrast, while [³H]NαMH showed decreased specific binding, binding affinity (pK_D), and B_max values upon exposure to GDP, which converts GPCRs from the high- to low-affinity state for agonists, binding of [³H]A-349821 was largely insensitive to GDP (Table 1 and Figure 7), consistent with its previously reported antagonist/inverse agonist properties. Interestingly, the binding of [¹²⁵I]Iodoproxyfan, a compound originally described as an antagonist radioligand, like that of [³H]NαMH, has been reported to be sensitive to guanylnucleotides (Ligneau et al., 1994). This observation is consistent with several publications describing agonist properties associated with iodoproxyfan (Schlicker et al., 1995; 1996; Wulff et al., 2002).

Saturation-binding studies utilizing membranes prepared from cortical tissue or from C6 cells stably expressing either rat or human H₃ receptors further revealed that [³H]A-349821 labelled a larger population of receptors compared to [³H]NαMH (Figure 2). This may be accounted for by [³H]A-349821 stabilizing the inactive H₃ receptor conformation, to which it binds with high affinity, or by [³H]A-349821 binding with equivalent affinity to either the inactive or active state of the receptor. Conversely, [³H]NαMH binds with high affinity only to the active state of the receptor, while the receptors in the inactive state would remain largely undetected over the concentrations of [³H]NαMH that are useable in saturation studies. Consistent with the observation that [³H]NαMH detected roughly 50% of the receptor population identified by [³H]A-349821 in the saturation-binding studies (Table 1), the population of high-affinity sites detected with agonists in competition with [³H]A-349821 was found to be roughly 50% (Table 3). Of note is that B_max values reported for [¹²⁵I]iodoproxyfan binding in rat cortical membranes were similar to those reported for [³H]NαMH or [³H]RαMH binding (Arrang et al., 1987; West et al., 1990a, 1990b; Ligneau et al., 1994), suggesting that [¹²⁵I]iodoproxyfan, like other agonist radioligands, detected only a fraction of the total H₃ receptors. Thus, [³H]A-349821 may provide a more accurate assessment of H₃ receptor density, which may be helpful for membrane-binding and autoradiographic studies, particularly in native tissue. Although [³H]A-349821 is a useful radiolabel for detection of H₃ receptors, modest specific activity, in combination with very high affinity, requires careful modifications of assay conditions to avoid ligand depletion, a limitation that may be overcome by improving the current radiolabelling process to increase specific activity.

In kinetic studies, [³H]A-349821 bound rapidly to a single site, with equilibrium achieved in all membranes within 5–15 min (Figure 3). Dissociation constants derived from kinetic studies agreed with values obtained in saturation studies and showed that [³H]A-349821 bound with approximately 10-fold higher affinity towards human compared to rat H₃ receptors due to a combination of faster on-rates and slower off-rates (Table 2).

Studies using a large panel of H₃ receptor ligands in competition with [³H]A-349821 gave potencies highly correlated with those observed for [³H]NαMH (R²⩾0.82), consistent with labelling of the H₃ receptor by [³H]A-349821 (Figure 6). Competition for [³H]A-349821 binding by H₃ receptor ligands was biphasic for compounds previously shown to be agonists, but monophasic for antagonists/inverse agonists (Tables 3 and 4 and Figure 5). GDP induced rightward shifts and increased Hill slopes for displacement curves with the H₃ receptor agonist histamine, but had no effect on displacement by the H₃ receptor antagonist clobenpropit (Figure 8). By contrast, competition against [³H]NαMH was monophasic for all ligands, presumably owing to the labelling of predominantly high-affinity sites at the concentrations of [³H]NαMH used for competition studies (∼0.5 nM). The findings that (1) [³H]A-349821 labelled predominantly one site while [³H]NαMH labelled two sites; (2) specific binding, B_max and pK_D values determined for [³H]NαMH but not [³H]A-349821 were affected by GDP; and (3) histamine displacement of [³H]A-349821 was shifted to the right by GDP, are consistent with G-protein modulation of interconvertable high- and low-affinity states for H₃ receptors.

Inspection of high- and low-affinity binding parameters estimated from displacement of [³H]A-349821 by reference H₃ receptor agonists revealed a subtle species difference (Table 3). Mean fractions of high-affinity sites for agonists were higher at recombinant or cortical rat H₃ receptors compared to recombinant human H₃ receptors (52±15, 47±11, and 37±7%, respectively, P<0.05). Another interesting observation is that the differences between mean potencies for high- and low-affinity states (pK_H−pK_L=ΔpK_i; derived from Table 3) for agonists were smaller in membranes expressing recombinant or cortical rat H₃ receptors compared to membranes expressing recombinant human H₃ receptors (1.67±0.19, 1.48±0.08, and 1.97±0.20, respectively; P<0.05). This trend is reflected in the correlation plots that show a greater frame shift with respect to high- and low-affinity sites for the human H₃ receptor compared to rat (Figure 6). Taken together, these results provide a body of evidence supporting the idea that discrimination of agonist binding to high- and low-affinity sites is less pronounced in rat versus human H₃ receptors.

Pharmacological comparisons utilizing [³H]A-349821 further revealed that although expression levels of H₃ receptors were 10-fold higher in membranes expressing recombinant rat H₃ receptors compared to those expressing cortical rat H₃ receptors (2190 and 227 fmol mg protein⁻¹, respectively), the mean proportion of high-affinity sites recognized by all agonists tested remained unchanged (49±9 and 51±11%, respectively), as did the differences between mean pK_H and pK_L values (1.67±0.19 and 1.48±0.08, respectively) (Tables 1 and 3). Furthermore, histamine competition curves against [³H]A-349821 showed a similar degree of sensitivity to GDP in membranes expressing recombinant rat H₃ receptors compared to those from rat cortices (Figure 8). Taken together, these results suggest that the proportion of high-affinity sites, as well as the ability of agonists to differentiate between high- and low-affinity sites, is largely independent of expression level and the milieu in which the rat H₃ receptor is expressed. Since multiple rat H₃ receptor isoforms are present in the cortex (Drutel et al., 2001), these data further indicate that, under these conditions, multiple H₃ receptor isoforms do not affect the apparent proportions of affinity states, either because of lower expression levels and/or similar binding properties.

It is notable that full agonists showed a clear discrimination between recognition of high- and low-affinity binding sites, while ligands having only antagonist and/or inverse agonist properties did not. Proxyfan and GT-2331, two compounds that have been previously shown to be functionally similar across a range of assays, displayed differential pharmacological classification ranging from antagonism to full agonism, depending on G-protein subtypes and the system in which the functional measurements were made (Krueger et al., 2005). Of interest, proxyfan has been previously described as a ‘protean agonist', being able to act as an agonist in systems where the constitutive activity is low or as an inverse agonist in systems where the constitutive activity is high (Gbahou et al., 2003). Consistent with this differential pharmacological behavior, competition-binding studies utilizing proxyfan and GT-2331 showed varied ability to discriminate between high- and low-affinity binding sites, different from the results seen with full agonists or antagonist/inverse agonists. While several pharmacological properties are similar for proxyfan and GT-2331, it should be mentioned that significant differences between these two compounds are apparent as GT-2331 binds with higher affinity to rat compared to human H₃ receptors, while the affinity of proxyfan is roughly the same (Table 3).

In conclusion, the novel non-imidazole-containing antagonist/inverse agonist radioligand, [³H]A-349821, labelled both rat and human H₃ receptors with high affinity and selectivity. The results described herein show that [³H]A-349821 labels interconvertable high- and low-affinity H₃ receptor sites, revealing the aspects of H₃ receptor affinity states with improved definition. Competition data using this radioligand indicate subtle species differences with respect to the pharmacology of receptor states by suggesting that rat H₃ receptor activation may be more facile compared to that of human. Investigation of high- and low-affinity H₃ receptor states using [³H]A-349821 in competition assays also allowed for the delineation of compounds displaying differential pharmacology from those displaying only full agonism, suggesting a relationship between binding properties and intrinsic activity. Application to membranes expressing recombinant and native H₃ receptors from various species could provide further comparative insights into H₃ receptor pharmacology.

Abbreviations

GPCR

G-protein-coupled receptor

HEK-293

human embryonic kidney cells

k_ob

observed rate constant

k_off

dissociation rate constant

k_on

association rate constant

NαMH

(N)-α-methylhistamine

RαMH

(R)-α-methyl-histamine

SαMH

(S)-α-methylhistamine