Compared pharmacology of human histamine H3 and H4 receptors: structure–activity relationships of histamine derivatives

Authors

  • Florence Gbahou,

    1. INSERM, Unité de Neurobiologie et Pharmacologie Moléculaire (U573), Centre Paul Broca, 2 ter rue d'Alésia, 75014 Paris, France
    Search for more papers by this author
    • 2

      These authors contributed equally to this study.

  • Ludwig Vincent,

    1. INSERM, Unité de Neurobiologie et Pharmacologie Moléculaire (U573), Centre Paul Broca, 2 ter rue d'Alésia, 75014 Paris, France
    Search for more papers by this author
    • 2

      These authors contributed equally to this study.

  • Marie Humbert-Claude,

    1. INSERM, Unité de Neurobiologie et Pharmacologie Moléculaire (U573), Centre Paul Broca, 2 ter rue d'Alésia, 75014 Paris, France
    Search for more papers by this author
  • Joel Tardivel-Lacombe,

    1. INSERM, Unité de Neurobiologie et Pharmacologie Moléculaire (U573), Centre Paul Broca, 2 ter rue d'Alésia, 75014 Paris, France
    Search for more papers by this author
  • Claude Chabret,

    1. INSERM, Unité de Neurobiologie et Pharmacologie Moléculaire (U573), Centre Paul Broca, 2 ter rue d'Alésia, 75014 Paris, France
    Search for more papers by this author
  • Jean-Michel Arrang

    Corresponding author
    1. INSERM, Unité de Neurobiologie et Pharmacologie Moléculaire (U573), Centre Paul Broca, 2 ter rue d'Alésia, 75014 Paris, France
      INSERM, Unité de Neurobiologie et Pharmacologie Moléculaire (U573), Centre Paul Broca, 2 ter rue d'Alésia, 75014 Paris, France. E-mail: arrang@broca.inserm.fr
    Search for more papers by this author

INSERM, Unité de Neurobiologie et Pharmacologie Moléculaire (U573), Centre Paul Broca, 2 ter rue d'Alésia, 75014 Paris, France. E-mail: arrang@broca.inserm.fr

Abstract

  • Various histamine derivatives were investigated at the human H3 receptor (H3R) and H4 receptor (H4R) stably expressed in human embryonic kidney (HEK)-293 cells using [125I]iodoproxyfan and [3H]histamine binding, respectively.

  • In Tris buffer, [3H]histamine binding to membranes of HEK(hH4R) cells was monophasic (KD of 3.8±0.8 nM). In phosphate buffer, the Hill coefficient was decreased (nH=0.5±0.1) and a large fraction of the binding was converted into a low-affinity component (KD=67±27 nM).

  • The inhibition of [3H]histamine binding by two agonists, a protean agonist and five antagonists/inverse agonists confirms that the potency of many H3R ligands is retained or only slightly reduced at the H4R.

  • Histamine derivatives substituted with methyl groups in α, β or Nα position of the side chain retained a nanomolar potency at the H3R, but their affinity was dramatically decreased at the H4R. With relative potencies to histamine of 282 and 0.13% at the H3R and H4R, respectively, (±)-α,β-dimethylhistamine is a potent and selective H3R agonist.

  • Chiral α-branched analogues exhibited a marked stereoselectivity at the H3R and H4R, the enantiomers with a configuration equivalent to L-histidine being preferred at both receptors.

  • The methylsubstitution of the imidazole ring was also studied. The relative potency to histamine of 4-methylhistamine (4-MeHA) at the H4R (67%) was similar to that reported at H2 receptors but, owing to its high affinity at the H4R (Ki=7.0±1.2 nM) and very low potency at H1- and H3-receptors, it can be considered as a potent and selective H4R agonist.

  • On inhibition of forskolin-induced cAMP formation, all the compounds tested, including 4-MeHA, behaved as full agonists at both receptors. However, the maximal inhibition achieved at the H4R (∼−30%) was much lower than at the H3R (∼−80%). Thioperamide behaved as an inverse agonist at both receptors and increased cAMP formation with the same maximal effect (∼+25%).

  • In conclusion, although the pharmacological profiles of the human H3R and H4R overlap, the structure–activity relationships of histamine derivatives at both receptors strongly differ and lead to the identification of selective compounds.

British Journal of Pharmacology (2006) 147, 744–754. doi:10.1038/sj.bjp.0706666

Abbreviations:
EtHA

ethylhistamine

FSK

forskolin

HEK

human embryonic kidney

hH3R

human H3 receptor

hH4R

human H4 receptor

MAP kinase

mitogen-activated protein kinase

MeHA

methylhistamine

Introduction

The H3 receptor (H3R) was detected in the 1980s as an autoreceptor controlling histamine synthesis and release in the rat and human brain (Arrang et al., 1983; 1987; 1988). The cDNAs encoding the H3R from various species, including human (Lovenberg et al., 1999) and rat (Lovenberg et al., 2000; Morisset et al., 2000; Drutel et al., 2001), were cloned recently. Screening of human libraries and genome databases by various groups led to the cloning and preliminary characterization of a receptor closely related to the H3R, the H4 receptor (H4R) (Nakamura et al., 2000; Oda et al., 2000; Liu et al., 2001a; Morse et al., 2001; Nguyen et al., 2001; Zhu et al., 2001). The H4R has about 40% sequence homogy to the H3R (58% in transmembrane domains) and both receptors display similar genomic structures with two introns and three exons (Coge et al., 2001b; Tardivel-Lacombe et al., 2001). In addition, some reports suggest that the recombinant H4R, like the H3R (Lovenberg et al., 1999), couples to Gi/Go proteins and inhibits forskolin-induced cAMP formation (Zhu et al., 2001). In contrast to these structural similarities, the expression patterns of both receptors strongly differ. Whereas the H3R is predominantly localized in the brain, where it is present on many neuronal perikarya, dendrites and projections (Pillot et al., 2002), the H4R is mainly expressed on hematopoietic cells (Oda et al., 2000; Liu et al., 2001a; Morse et al., 2001; Zhu et al., 2001). The presence of the H4R on leukocytes and mast cells suggests that it plays an important role in immune responses and inflammation (Gantner et al., 2002; Buckland et al., 2003; Hofstra et al., 2003; Takeshita et al., 2003; Bell et al., 2004; Thurmond et al., 2004).

The pharmacological profiles of H3Rs (Ligneau et al., 2000; Lovenberg et al., 2000) and H4Rs (Liu et al., 2001b) show strong species differences, but early studies indicated that they overlap, several H3R agonists or antagonists having appreciable activity at the H4R (Hough, 2001). Few selective H4R ligands are yet available (de Esch et al., 2005). The first potent and selective H4R antagonist is a non-imidazole derivative that was designed recently and used for the characterization of the recombinant and native H4R (Ling et al., 2004; Thurmond et al., 2004). Among agonists, the atypical neuroleptic clozapine has been shown to fully activate the H4R, although it displays a moderate (submicromolar) affinity (Oda et al., 2000; Liu et al., 2001a, 2001b; Buckland et al., 2003). The two methylcyanoguanidine derivatives of imifuramine, OUP-13 and OUP-16, also act as full agonists at the human H4 receptor (hH4R), but display a higher potency with a 40-fold selectivity over the human H3 receptor (hH3R) (Hashimoto et al., 2003).

In the present study, the potencies of various chiral and nonchiral histamine derivatives were determined at the recombinant human H4R on inhibition of [3H]histamine binding (in different experimental conditions). For comparison, and because most of these histamine derivatives had been previously studied only at the rat H3 autoreceptor, their potencies have also been determined at the recombinant hH3R on inhibition of [125I]iodoproxyfan binding (Ligneau et al., 1994). In addition, some selected compounds have been studied at the human H3R and H4R on inhibition of forskolin-induced cAMP formation.

Methods

Cloning of the hH3R and hH4R cDNAs

The hH3R was cloned by screening of a human striatum cDNA library as described (Ligneau et al., 2000). cDNAs corresponding to the full-length coding sequence of the human H4 receptor (hH4R) were cloned by PCR. Human bone marrow Marathon-Ready cDNAs (Clontech, Basingstoke, U.K.) were amplified for 40 cycles (94°C, 55°C and 72°C for 30 s each) using AmpliTaq Gold polymerase (Perkin-Elmer Life Sciences, Boston, MA, U.S.A.) and primers based on the N-terminal and C-terminal regions of the hH4R sequence (forward primer: 5′-ATGCCAGATACTAATAGCACAATCAATTTATC-3′ and reverse primer: 5′-TTAAGAAGATACTGACCGACTGTGTTGT-3′). PCR products were electrophoresed on a 1% agarose gel, subcloned and sequenced.

Stable transfection of human embryonic kidney (HEK)-293 cells

cDNAs corresponding to the full-length coding sequences of the hH3R and hH4R were ligated into the mammalian expression vector pCIneo (Promega, Charbonnières, France). HEK-293 cells were transfected using PolyFect (Qiagen, Courtaboeuf, France). Stable transfectants were selected with 2 mg ml−1 of geneticin (G418, Invitrogen Gibco BRL, Cergy-Pontoise, France) and tested for [125I]iodoproxyfan binding (hH3R) (Ligneau et al., 1994) or for [3H]histamine binding (hH4R). Several clones named HEK(hH3R) or HEK(hH4R) were selected for further characterization and maintained in the presence of 0.2−1 mg ml−1 of G418.

Binding assays

Binding assays on the hH3R were performed as described previously (Ligneau et al., 1994). Aliquots of membrane suspensions from HEK(hH3R) cells (10−20 μg of protein) were incubated for 60 min at 25°C with 25 pM [125I]iodoproxyfan alone or together with competing drugs in phosphate buffer (Na2HPO4/KH2PO4 50 mM, pH 7.5) (200 μl final volume). The nonspecific binding was determined using imetit (1 μM).

For binding assays on the hH4R, HEK(hH4R) cells were washed and homogenized with a Polytron in ice-cold Tris buffer (Tris-HCl 50 mM, pH 7.5). After centrifugation (12,000 × g for 30 min at 4°C), the pellet was suspended in 1 ml of the same ice-cold binding buffer. Aliquots of the membrane suspension (10−20 μg of protein) were incubated for 60 min at 25°C with [3H]histamine alone or together with competing drugs (1 ml final volume). The nonspecific binding was determined using imetit (1 μM). In some saturation studies, the membrane fraction and binding assay was performed in phosphate buffer (Na2HPO4/KH2PO4 50 mM, pH 7.5).

cAMP accumulation

HEK(hH3R) or HEK(hH4R) cells were incubated for 10 min at 37°C with 1 μM forskolin and, when required, 10 μM of the various ligands. cAMP was extracted and measured by radioimmunoassay (Perkin-Elmer Life Sciences, Boston, MA, U.S.A.). Statistical evaluation of the results was performed by one-way ANOVA followed by Newman–Keuls test.

Analysis of data

The saturation and inhibition curves were analyzed with an iterative least-squares method derived from that of Parker & Waud (1971). Computer analysis was performed by nonlinear regression using a one-site cooperative model, except in some saturation experiments in which a significant improvement of the analysis was obtained by resolution of the data in two components with a two-site model. The method provided estimates for KD and Bmax values of [125I]iodoproxyfan (hH3R) and [3H]histamine (hH4R) and for IC50 values of competing drugs. Ki values of the latter were calculated from their IC50 values assuming a competitive antagonism and by using the relationship Ki=IC50/(1+S/KD), where S represents the concentration of the radioligand (25 pM [125I]iodoproxyfan or 2 nM [3H]histamine) and KD its apparent dissociation constant (Cheng & Prussoff, 1973).

Radiochemicals and drugs

[125I]Iodoproxyfan (2000 Ci mmol−1) was prepared as described (Krause et al., 1997). [3H]Histamine (51 Ci mmol−1) was from Amersham Pharmacia Biotech (Les Ulis, France). Thioperamide and ciproxifan were from Bioprojet (Paris, France). Clobenpropit was from Tocris (Bristol, U.K.). Proxyfan, FUB 349 (Ligneau et al., 2000), FUB 465 (Morisset et al., 2000), Nα-ethylhistamine, β-methylhistamine (β-MeHA), α,α-, α,β- and β,β-dimethylhistamine (α,α-diMeHA, α,β-diMeHA and β,β-diMeHA), (R)- and (S)-α-methylhistamine (α-MeHA), (R)- and (S)-α-chloromethylhistamine (α-ChloroMeHA) and (R)- and (S)-α-hydroxymethylhistamine (α-HydroxyMeHA) were provided by W. Schunack (Freie Universität Berlin, Germany). Imetit, Nα-MeHA, Nα,Nα-diMeHA, 2- and 4-MeHA were provided by C.R. Ganellin (University College, London, U.K.).

Results

Characterization of [3H]histamine binding to the hH4R

[3H]histamine binding to membranes of HEK(hH4R) cells in Tris-HCl buffer (50 mM, pH 7. 5) was saturable and analysis of the data using an one-site cooperative model indicated that the Hill coefficient was not significantly different from unity (nH=0.9±0.1), with a KD value of 3.8±0.8 nM and a Bmax value of 2400±200 fmol mg−1 protein. Scatchard analysis of saturation binding data also disclosed a single population of sites with a KD value of 3.5±0.4 nM (Figure 1a). In the presence of 100 μM GTPγS, no significant specific binding could be detected when the membranes were incubated in Tris-HCl buffer with 25 nM [3H]histamine (data not shown).

Figure 1.

Saturation of [3H]histamine binding to the recombinant hH4R. Membranes of HEK(hH4R) cells were incubated with [3H]histamine in increasing concentrations for 60 min at 25°C in 50 mM Tris-HCl buffer, pH 7.5 (a) or 50 mM Na2HPO4/KH2PO4 buffer, pH 7.5 (b). Specific binding was defined as that inhibited by 1 μM imetit. The insets show the Scatchard transformation of the data. The lines in (b) were drawn after analysis of the data by nonlinear regression using a least-square curve fitting procedure for a two-site model. Each point represents the mean of 4–8 determinations from two separate experiments.

When the experiments were performed with the same membrane samples in phosphate buffer (50 mM Na2HPO4/KH2PO4, pH 7.5), analysis of the data using an one-site cooperative model indicated that the Hill coefficient was greatly decreased as compared with a Tris-HCl buffer (nH=0.5±0.1). The Bmax was also significantly decreased and the data fitted significantly better to a two-site model analysis. The Scatchard plot could also be resolved in a high-affinity population of sites (KD=0.56±0.12 nM and Bmax=154±64 fmol mg−1 protein, i.e., 15% of maximal specific binding) and a low-affinity population with KD and Bmax values of 67±27 nM and 883±72 fmol mg−1 protein, respectively (Figure 1b). When the membranes were incubated in phosphate buffer with 25 nM [3H]histamine in the presence of 100 μM GTPγS, a partial but significant decrease of specific binding was observed (−29±5%), and this decrease corresponded to the density of the high affinity population of sites (−144±28 fmol mg−1 protein) (data not shown).

Results from competition studies using Tris-HCl (50 mM, pH 7.5) buffer show that [3H]histamine binding to membranes of HEK(hH4R) cells was inhibited in a concentration-dependent manner by a range of H3-receptor ligands. All of the compounds inhibited the binding with Hill coefficients not significantly different from unity, although the coefficient found for thioperamide in the one-site model tended to be slightly lower (nH=0.82±0.03). The deduced Ki values for each compound are given in Table 1. Unlabelled histamine inhibited the binding with a Ki value similar to its KD value (4.7±0.3 and 3.8±0.8 nM, respectively). The affinities of the compounds at the hH4R were also compared to their affinities that we recently reported at the hH3R (Ligneau et al., 2000). The agonist imetit and the protean agonist proxyfan (Gbahou et al., 2003) were about two- and 13-fold less potent at the H4R, respectively. Among the H3-receptor antagonists/inverse agonists, thioperamide displayed similar potencies at human H4Rs and H3Rs, whereas clobenpropit, FUB 349, FUB 465 and ciproxifan were about 2–13-fold less potent at the H4R (Table 1).

Table 1. Compared potencies of H3-receptor ligands at the human H4 and H3 receptors (H4Rs and H3Rs)
AgenthH4RhH3RSelectivity (ratio)
  1. Values (Ki, nM) regarding the H3 receptor are taken from Ligneau et al. (2000), except for imetit (Wulff et al., 2002), histamine and FUB 465.

Histamine4.7±0.311±2H4 (2.3)
Imetit1.6±0.10.7±0.1H3 (2.3)
Proxyfan34±12.7±0.1H3 (13)
Thioperamide43±360±12H4 (1.4)
Ciproxifan612±3246±4H3 (13)
Clobenpropit4.3±0.22.4±0.6H3 (1.8)
FUB 465704±74188±12H3 (3.7)
FUB 3499.5±0.22.1±0.2H3 (4.5)

Potencies of Nα-substituted histamine derivatives, and of 2- and 4-MeHA at the human H3Rs and H4Rs

The potencies of three Nα-substituted histamine derivatives at the hH3R were determined by the [125I]iodoproxyfan-binding assay (Ligneau et al., 1994) on membranes of HEK(hH3R) cells stably expressing the receptor. Specific [125I]iodoproxyfan binding was monophasic and saturable, and analysis by nonlinear regression using a one-site cooperative model led to a KD value of 82±12 pM, in close agreement with the value that we previously reported on membranes of CHO(hH3R) cells (50±7 pM, Ligneau et al., 2000). The inhibition curves for the agonists Nα-MeHA and Nα,Nα-dimethylhistamine (Nα,Nα-diMeHA), were found to be shallow, their pseudo-Hill coefficients being close to 0.7 (Figure 2). A similar observation could be made for the inhibition curve of histamine ((Ligneau et al., 2000) and Figure 2). The apparent Ki values that were deduced from the mean IC50 values are given in Table 2. The potencies of the agonists relative to histamine (=100) calculated from these Ki values indicated that Nα-MeHA and Nα,Nα-diMeHA were about four- and two-fold more potent than histamine itself, respectively, whereas Nα-ethylhistamine (Nα-EtHA) was about four-fold less potent than histamine at the hH3R (Table 2). Both 2-MeHA and 4-MeHA displayed a much lower affinity with Ki values in the micromolar range leading to potencies relative to histamine less than 0.1% (Table 2).

Figure 2.

Effect of Nα-substituted histamine derivatives and of 2- and 4-MeHA on [125I]iodoproxyfan binding to the hH3R and on [3H]histamine binding to the hH4R. Membranes of HEK(hH3R) cells were incubated with 25 pM [125I]iodoproxyfan and drugs as described (Ligneau et al., 1994). Membranes of HEK(hH4R) cells were incubated in Tris-HCl buffer with 2 nM [3H]histamine and drugs in increasing concentrations. Each point represents the results from two independent experiments with triplicate determinations.

Table 2. Compared potencies of Nα-substituted histamine derivatives and of 2- and 4-methylhistamine at the human H4 and H3 receptors (H4Rs and H3Rs)Thumbnail image of

The potencies of the histamine derivatives were then determined at the hH4R by the [3H]histamine-binding assay performed in 50 mM Tris-HCl buffer (pH 7.5) on membranes of HEK(hH4R) cells. The three Nα-substituted compounds inhibited the binding with Hill coefficients not significantly different from unity and displayed a decreased potency relative to histamine at H4Rs (Figure 2). Nα-MeHA was about 10-fold less potent than histamine. Nα-ethylhistamine was about four-fold less potent than its methyl analogue. Nα,Nα-diMeHA inhibited [3H]histamine binding with a very low (micromolar) affinity, leading to a potency relative to histamine of only 0.4% (Table 2). 2-MeHA was also much less potent than histamine with a relative potency of 1.9%. In contrast, 4-MeHA displayed a high affinity at the H4R with a Ki value of 7.0±1.2 nM leading to a potency relative of histamine of 67% (Table 2).

Potencies of non-chiral α- and/or β-methylated histamine derivatives at the human H3Rs and H4Rs

Substitution with methyl groups in α- and/or β-position of the side chain of the histamine molecule leads to compounds that displaced specific [125I]iodoproxyfan binding to hH3Rs (nH=0.7–0.8) with a high affinity (Figure 3). (±)-β-MeHA and (±)-α,β-diMeHA displayed a nanomolar affinity and were two- to three-fold more potent than histamine. The affinities of α,α-diMeHA and β,β-diMeHA were lower, these two derivatives being about three- to four-fold less potent than histamine (Table 3).

Figure 3.

Effect of nonchiral α- and/or β-methylated histamine derivatives on [125I]iodoproxyfan binding to the hH3R and on [3H]histamine binding to the hH4R. Each point represents the results from two independent experiments with triplicate determinations.

Table 3. Compared potencies of nonchiral α- and/or β-methylated histamine derivatives at the human H4 and H3 receptors (H4Rs and H3Rs)Thumbnail image of

These histamine derivatives displaced specific [3H]histamine binding to hH4Rs with a low affinity and with Hill coefficients not significantly different from unity (Figure 3). (±)-β-MeHA displayed a potency relative to histamine of 4.3% on membranes of HEK(hH4R) cells. The affinity of the dimethyl analogues was even lower with Ki values in the micromolar range leading to potencies relative to histamine below 0.5% (Table 3).

Potencies of chiral histamine derivatives at the human H3Rs and H4Rs

[125I]iodoproxyfan binding to membranes of HEK(hH3R) cells was inhibited in a concentration-dependent and stereoselective manner (nH=0.6–0.7) by the enantiomeric pairs of three chiral α-branched histamine derivatives, that is, α-MeHA, α-ChloroMeHA and α-HydroxyMeHA. As expected, the binding of the compounds was stereoselective and the enantiomer with the same spatial configuration as L-histidine, that is, (R)-α-MeHA, (S)-α-ChloroMeHA and (S)-α-HydroxyMeHA, was in each case preferred at hH3Rs (Table 4). The potent agonist (R)-α-MeHA displaced specific binding with a nanomolar affinity similar to that previously reported in the same binding test (Ligneau et al., 2000; Coge et al., 2001a; Uveges et al., 2002; Wulff et al., 2002). (R)-α-MeHA was 17-fold more potent than (S)-α-MeHA, the potencies relative to histamine of the two enantiomers being of 611 and 37%, respectively (Table 4). The enantiomers of α-ChloroMeHA and α-HydroxyMeHA displayed a lower affinity but a marked stereoselectivity was again observed with these two analogues. The isomer with the same relative configuration as L-histidine, and therefore (R)-α-MeHA, in their case, the S-isomer, was preferred at hH3Rs. The Ki values obtained for each enantiomeric pair yielded a ratio S/R of 4.3 for the isomers of α-ChloroMeHA and 7.1 for the isomers of α-HydroxyMeHA at hH3Rs (Table 4).

Table 4. Compared potencies of chiral histamine derivatives at the human H4 and H3 receptors (H4Rs and H3Rs)Thumbnail image of

These chiral derivatives displayed a lower potency relative to histamine (below 4%) on [3H]histamine binding to membranes of HEK(hH4R) cells, but a marked difference in affinity was also observed between enantiomers. Although it was about 25-fold less potent than histamine, the affinity of (R)-α-MeHA was 17-fold higher than that displayed by (S)-α-MeHA (Table 4). A marked stereoselectivity was also observed with the two enantiomers of α-ChloroMeHA and again, the isomer corresponding to L-histidine, in that case the S-isomer, was the most potent at hH4Rs (ratio S/R of 4.4). The potencies relative to histamine of the R- and S-enantiomers of α-HydroxyMeHA were not significantly different (0.13 and 0.08%), indicating that this chiral analogue binds to H4Rs without any clear stereoselectivity (Table 4).

The affinities of the ligands at the hH4R were also evaluated when [3H]histamine binding to membranes of HEK(hH4R) cells was performed in the phosphate buffer. When [3H]histamine was added at a concentration (10 nM) selected to ensure a maximal specific binding at the high affinity site, without any significant binding to the low affinity site, the Ki values calculated from the obtained IC50 values indicated that the affinity of the H3R ligands (see Table 1), histamine itself and the various histamine derivatives, was very similar to that found in Tris buffer, the Ki values being about 2–6 fold higher.

Effect of histamine derivatives on forskolin-induced cAMP accumulation in HEK(hH3R) and HEK(hH4R) cells

The functional properties of some of the various derivatives described above were investigated on forskolin-induced cAMP accumulation in the HEK-293 cells expressing the human H3Rs and H4Rs. In agreement with previous findings obtained in the same cells (Wulff et al., 2002; Takahashi et al., 2003), histamine itself used at a maximal concentration strongly inhibited cAMP formation in HEK(hH3R) cells (−86±2%, P<0.001). Nα-MeHA, (R)-α-MeHA, β-MeHA, α,β-diMeHA (10 μM) all induced the same maximal inhibition of forskolin-induced cAMP accumulation, indicating that they all behave as full agonists at the hH3R. In agreement with its low affinity at the H3R, a very low inhibition (−18±4%) was induced by 4-MeHA used at the same concentration (10 μM) (Figure 4). Thioperamide, a standard inverse agonist, significantly increased cAMP formation in HEK(hH3R) cells (+23±8%) (Figure 4).

Figure 4.

Effect of histamine derivatives on forskolin-induced cAMP accumulation in HEK(hH3R) (white bars) and HEK(hH4R) (black bars) cells. Cells were incubated with 1 μM forskolin (FSK) and, when required, drugs at a 10 μM final concentration. Results are expressed as the percent change of the FSK-evoked response which represented 7.5±0.5 pmol (hH3R) and 2.1±0.1 pmol (hH4R), and are the mean values from two to three (hH3R) and four to nine (hH4R) separate experiments with 4–10 determinations. *P<0.05; **P<0.001 vs FSK.

Histamine also significantly inhibited forskolin-induced cAMP accumulation in HEK(hH4R) cells. However, the maximal inhibition was much lower than that observed in HEK(hH3R) cells (−34±3%) (Figure 4). Such a low inhibition was also observed with all the substituted derivatives tested. In addition, all the derivatives appeared to behave as full agonists at the hH4R, exhibiting a maximal inhibitory effect similar to that of histamine (around −30%) (Figure 4). In contrast, thioperamide significantly increased cAMP formation in HEK(hH4R) cells (+25±4%) (Figure 4).

Discussion

As frequently reported for 3H-agonist binding to various G-protein-coupled receptors, the binding characteristics of [3H]histamine to H4Rs varied with the incubation medium, being markedly altered by the presence of monovalent cations. In Tris buffer, that is, in the absence of monovalent cations, the specific binding of [3H]histamine to membranes of HEK(hH4R) cells occurred with a nanomolar affinity to a single high affinity population of sites. When the experiments were performed with the same membranes in sodium/potassium phosphate buffer, the apparent maximal specific binding was decreased by ∼50%, a finding probably resulting from the conversion by sodium ions of a fraction of the sites to low-affinity sites, no longer detectable in our conditions. Moreover, the remainder of the [3H]histamine-binding sites that displayed an affinity high enough to be observed in sodium/potassium phosphate buffer was heterogeneous. The population displaying the highest affinity (in the nanomolar range) was similar to that found in Tris buffer. In agreement, the KD value of [3H]histamine and the Ki values of all competing ligands at this high affinity site were only 2–6-fold lower in phosphate buffer than in Tris buffer. In addition, in both conditions, the binding to this high affinity site was entirely abolished by GTPγS, indicating that it represents an active state of the receptor coupled to G proteins. The presence of monovalent cations in the phosphate buffer promoted the conversion of a large fraction of sites into a component displaying an ∼120-fold lower affinity (KD=67±24 nM). This low-affinity site remained apparently unaltered by GTPγS and therefore presumably corresponded to an uncoupled state of the receptor.

This strong heterogeneity of [3H]histamine binding at the hH4R revealed in phosphate buffer probably accounts for the large differences found in the various estimations of the KD value of [3H]histamine at the high-affinity conformations (from 0.56 nM (the present study in sodium/potassium phosphate buffer) to 17.6 nM (Zhu et al., 2001)). The existence of these different conformations also presumably accounts for the lower KD and Ki values of histamine in binding tests as compared to its EC50 value in some functional tests (Liu et al., 2001a; Oda et al., 2002). The modulation of 3H-agonist binding to native H3Rs by sodium and calcium ions (Arrang et al., 1990; Kilpatrick & Michel, 1991; Clark & Hill, 1995), as well as the shallow inhibition curves observed for various agonists at recombinant H3Rs from various species (Ligneau et al., 2000; Morisset et al., 2001; Uveges et al., 2002; Rouleau et al., 2004) indicated a similar heterogeneity among agonist-binding sites at the H3R. However, the pattern of 3H-agonist binding to H3 and H4Rs is clearly different. The binding of both [3H](R)-α-MeHA and [3H]Nα-MeHA in sodium/potassium phosphate buffer apparently occurred to a single high-affinity H3-receptor conformation and could therefore be successfully used for the pharmacological characterization of the H3R (Arrang et al., 1990; Clark & Hill, 1995).

The inhibition of [3H]histamine binding by various H3-receptor ligands was also consistent with the labelling of a single conformation of H4Rs in Tris buffer. Moreover, in agreement with previous studies, the inhibition induced by two H3-receptor agonists (histamine and imetit), a H3-receptor protean agonist (proxyfan) and five H3-receptor antagonists/inverse agonists (thioperamide, ciproxifan, clobenpropit, FUB 349 and FUB 465) further showed that the pharmacological profiles of the human H3Rs and H4Rs strongly overlap (Hough, 2001). The potencies of the compounds at the hH4R were consistent with those previously found in various cells (Liu et al., 2001a, 2001b; Morse et al., 2001; Zhu et al., 2001; Oda et al., 2002; O'Reilly et al., 2002; Esbenshade et al., 2003) and, for most of them, were close to their potencies previously reported at the hH3R (Lovenberg et al., 1999; 2000; West et al., 1999; Ligneau et al., 2000; Coge et al., 2001a; Ireland-Denny et al., 2001; Liu et al., 2001a; Wieland et al., 2001; O'Reilly et al., 2002; Uveges et al., 2002; Wulff et al., 2002; Esbenshade et al., 2003; Yao et al., 2003). For example, the potency of imetit at the hH4R (Ki=1–6 nM) was only five-fold lower than its mean potency at the hH3R (Ki=0.2–1.6 nM). The protean agonist proxyfan had a 10-fold lower potency at the hH4R than at the hH3R (Ki=34 vs 2.7–5.0 nM). Among H3-receptor antagonists/inverse agonists, the potency of clobenpropit at the hH4R was in close agreement with that previously reported in various cells (Ki=5–42 nM) and was only 2–10-fold lower than at the hH3R (Ki=0.4–5.7 nM). FUB 349 and FUB 465, two imidazole antagonists exhibiting a nanomolar and submicromolar affinity, respectively, at the hH3R (this study and Ligneau et al., 2000) were also only four-fold less potent at the hH4R. Ciproxifan exhibited only a micromolar potency at the hH4R (Ki of 612 nM in the present study and 1.86 μM in (Esbenshade et al., 2003)), and tended therefore to be significantly less potent than at the hH3R (Ki=46–180 nM). In the studies mentioned above, a strong difference (up to 20-fold) was curiously found between the various determinations of the Ki value of thioperamide at hH4Rs (27 to 519 nM), and it was also suggested that it may be less potent at hH4Rs than at hH3Rs (Hough, 2001). However, the potency found in the present study at H4Rs (Ki=43±3 nM) was similar to that previously found at the hH3R (Ki=25–200 nM). Taken together, the present results added to the preliminary pharmacology previously obtained with other imidazole compounds, confirm that the potency of many H3R ligands is retained or only slightly reduced at the hH4R.

The present study shows that methylsubstitution in the side chain of the histamine molecule in Nα, α, or β position is much less tolerated by the H4R than by the H3R. The corresponding methylsubstituted derivatives display a high agonist potency at H3 autoreceptors regulating histamine release in the rat brain (Arrang et al., 1991; Lipp et al., 1991). In the present study, they also retained a high potency at the hH3R but their potency at hH4Rs stably expressed in the same cells, that is, HEK-293 cells, was dramatically decreased.

The negative influence of Nα-substitution at the H4R became obvious when the methyl and dimethyl substituents were compared. In agreement with previous binding studies, Nα-MeHA exhibited a nanomolar affinity at the hH3R and was about four-fold more potent than histamine, a relative potency in the same range as that found at the rat H3R (Arrang et al., 1983; 1990). Its affinity at the hH4R was about 20-fold lower, with a relative potency to histamine of about 10% and its maximal effect on the inhibition of cAMP formation confirmed that it acts as a full agonist. Nα,Nα-diMeHA was also slightly more potent than histamine at the rat (Arrang et al., 1983) and hH3Rs, but its affinity became about 200-fold lower at the hH4R than at the H3R, with a relative potency to histamine less than 1%.

Substitution with methyl groups in α- and/or β-position of the side chain leads to similar observations. The corresponding compounds retained a high potency at the H3R, but their affinity was dramatically decreased at the H4R, although they still behaved as full agonists on cAMP formation. (±)-β-MeHA, which was about two-fold more potent than histamine at the rat (Arrang et al., 1991; 1992; Lipp et al., 1991; 1992b) and hH3R, was 15-fold less potent at the hH4R than at the hH3R. β,β-diMeHA displayed a micromolar affinity at the hH4R, 35-fold lower than at the hH3R. A high selectivity ratio was found with α,α-diMeHA which exhibited a negligible affinity at the hH4R, 700-fold lower than at the hH3R. Interestingly, α,α-diMeHA and β,β-diMeHA exhibited similar potencies at the hH3R, whereas α,α-diMeHA was 75-fold more potent than β,β-diMeHA at the rat H3 autoreceptor (Arrang et al., 1991; 1992; Lipp et al., 1991; 1995). This finding may result from the distinct pharmacological profiles of the rat and hH3Rs (Arrang et al., 1987; 1988; West et al., 1999; Ligneau et al., 2000; Lovenberg et al., 2000; Ireland-Denny et al., 2001; Wulff et al., 2002; Yao et al., 2003). Alternatively, these two compounds may discriminate distinct conformations of the receptor in the functional assay (modulation of histamine release) and the [125I]iodoproxyfan-binding assay, inasmuch as agonists, including α,α-diMeHA and β,β-diMeHA, inhibit the binding with the Hill coefficients significantly lower than unity, revealing an heterogeneity among agonist-binding sites at the H3R.

Among the α and/or β-substituted compounds tested, (±)-α,β-diMeHA was the most potent H3R agonist, both in the rat (Lipp et al., 1991; 1992a) and human. It also displayed the highest selectivity for the H3R, its affinity being 900-fold lower at the hH4R. As it also displays a very low potency at H1 and H2 receptors (Lipp et al., 1991; 1992a), (±)-α,β-diMeHA can, therefore, be considered as a potent and highly selective H3-receptor agonist.

As expected (Arrang et al., 1985; 1987; 1990), the binding of the chiral α-branched ligands at the hH3R exhibited a pronounced stereoselectivity, and in all cases the enantiomer with a configuration equivalent to L-histidine was preferred, as already observed for the rat receptor (Arrang et al., 1985; Lipp et al., 1992b). (R)-α-MeHA, which exhibited an expected nanomolar affinity, was 17-fold more potent than (S)-α-MeHA, a ratio similar to that previously found at the human receptor (West et al., 1999; Ireland-Denny et al., 2001; Wulff et al., 2002), but lower than that reported at the rat receptor (Arrang et al., 1990; Ligneau et al., 1994; Wulff et al., 2002). Compared to the R-isomer, the higher potency of the S-isomer of α-ChloroMeHA and α-HydroxyMeHA, in that case the isomer with the relative configuration corresponding to L-histidine, and therefore to (R)-α-MeHA, also revealed a marked stereoselectivity with these two chiral analogues.

In spite of much lower affinities of the derivatives, a marked stereoselectivity was also observed with the chiral α-branched ligands at the hH4R, and the enantiomers preferred were the same as those preferred at the hH3R. In agreement with previous reports (Liu et al., 2001a, 2001b; Morse et al., 2001; Zhu et al., 2001; Oda et al., 2002; O'Reilly et al., 2002), (R)-α-MeHA still acted as a full agonist on cAMP formation but was about 60-fold less potent at the hH4R than at the H3R, confirming that the methylation of the side chain of the histamine molecule is not well tolerated by the H4R. In agreement with previous functional studies (Shin et al., 2002), it was more potent than (S)-α-MeHA. (S)-α-ChloroMeHA was also more potent than the corresponding R-isomer. In addition, the stereoselectivity ratios between the two isomers of both chiral derivatives were similar to those found at the hH3R. However, the marked stereoselectivity found between the isomers of α-HydroxyMeHA at the hH3R was no longer observed at the hH4R, indicating a different degree of stereoselectivity of both receptors for some compounds.

In addition to the effect of methylsubstitution in the side chain of the histamine molecule, the effect of methylsubstitution of the imidazole ring at human H3Rs and H4Rs was also studied with 2- and 4-MeHA, two compounds known to display a relative selectivity for H1 and H2 receptors, respectively (Black et al., 1972; Ganellin, 1982; Hill et al., 1997). As expected from our previous studies at the rat H3 autoreceptor (Arrang et al., 1983), both compounds displayed a very low affinity at the hH3R. Although its affinity was about 150-fold higher at the hH4R than at the H3R, 2-MeHA remained much less potent than histamine at the hH4R. In contrast, 4-MeHA, which acted as a full agonist on the inhibition of cAMP formation, displayed a potency similar to that of histamine at the hH4R, with an affinity in the nanomolar range, 2700-fold higher than at the hH3R. 4-MeHA is very poorly active at H1 receptors but it has also long been considered as a relatively selective H2-receptor agonist. Its potency has been reported to be about 50% that of histamine at H2 receptors in various systems and species (Black et al., 1972; Ganellin, 1982; Hill et al., 1997). Therefore, its potency relative to histamine that we report here at H4Rs (67%) is much higher than at H1Rs and H3Rs but is in the same range as that found at H2 receptors, which may indicate a rather limited selectivity. However, the binding affinity of histamine at H4Rs being several orders of magnitude higher than at either the H1- or H2 receptors (Hill et al., 1977; Arrang et al., 1990; Ruat et al., 1990), 4-MeHA is likely to activate H4Rs at concentrations much lower than those required to activate the three other subclasses of histamine receptors. Therefore, although 4-MeHA acts as an agonist and the potency of agonists is well known to be dependent on the test system used, it can be considered as selective for H4Rs. While this work was in progress, Lim et al. (2005) have also identified 4-MeHA as a potent agonist at hH4Rs expressed in SK-N-MC cells, although its affinity was found to be lower (Ki=50 nM).

Like the recombinant H3R (Lovenberg et al., 1999; Morisset et al., 2000; Drutel et al., 2001), the recombinant H4R has been reported to be coupled to Gi/o proteins and inhibition of cAMP formation. However, in spite of a high expression level, our data show that the maximal inhibition achieved by various full agonists in HEK(hH4R) cells was much lower than that obtained in HEK(hH3R) cells, making quantification of the modulation of cAMP levels problematic. In agreement, in other studies, direct coupling of the H4R to inhibition of adenylate cyclase could not be detected (Morse et al., 2001) or was also found to be weak (Nakamura et al., 2000; Oda et al., 2000; Liu et al., 2001a, 2001b). This might suggest that inhibition of adenylate cyclase is not a primary transduction pathway of the H4R or that it involves a low coupling efficacy of the receptor to the G proteins.

Recently, we showed that the recombinant rat and hH3Rs expressed at moderate densities display constitutive activity (Morisset et al., 2000; Rouleau et al., 2002). This constitutive activity was clearly evidenced by the enhancement of cAMP accumulation induced in HEK(hH3R) cells by thioperamide acting as an inverse agonist. Interestingly, in agreement with a previous study based upon [35S]GTPγ[S] binding and MAP kinase activity (Morse et al., 2001), constitutive activity of the H4R could also be easily detected by the enhancement of cAMP formation induced by thioperamide in HEK(hH4R) cells. Consistent with the physiological relevance of the phenomenon, we demonstrated constitutive activity of the native rat and mouse H3Rs and showed that it controls histaminergic neuron activity in rodent brain in vivo (Morisset et al., 2000; Rouleau et al., 2002). Whether native H4Rs also display constitutive activity remains to be established. However, it is worth noting that, whereas the maximal effect of the agonists at the H4R was much lower than at the H3R, the intrinsic efficacy of thioperamide acting as an inverse agonist was similar at H3Rs and H4Rs, indicating a relatively high coupling efficacy of constitutively active conformations of the H4R compared to that of ligand-selective active conformations (Gbahou et al., 2003).

In conclusion, the present data confirm that the affinity of many H3R ligands, including standard agonists and antagonists, is retained or only slightly reduced at the hH4R. However, the structure–activity relationships of histamine derivatives at human H3Rs and H4Rs strongly differ and lead to the identification of selective compounds. Derivatives methylsubstituted in Nα, α- or β-position of the side chain retain a high affinity at the H3R but their affinity at the H4R is strongly decreased. Among them, (±)-α,β-diMeHA is a potent and selective H3-receptor agonist, with a 900-fold lower affinity at H4Rs and a very low potency at H1 and H2 receptors. In contrast, 4-MeHA displays a much higher affinity at hH4Rs than at H3Rs and can be considered as a potent and selective H4-receptor agonist.

Acknowledgments

We thank Pr W. Schunack and Pr C.R. Ganellin for providing many of the drugs used in the present study.

Ancillary