Pharmacological Characterization of Human Histamine Receptors and Histamine Receptor Mutants in the Sf9 Cell Expression System

Abstract

A large problem of histamine receptor research is data heterogeneity. Various experimental approaches, the complex signaling pathways of mammalian cells, and the use of different species orthologues render it difficult to compare and interpret the published results. Thus, the four human histamine receptor subtypes were analyzed side-by-side in the Sf9 insect cell expression system, using radioligand binding assays as well as functional readouts proximal to the receptor activation event (steady-state GTPase assays and [³⁵S]GTPγS assays). The human H₁R was co-expressed with the regulators of G protein signaling RGS4 or GAIP, which unmasked a productive interaction between hH₁R and insect cell Gα_q. By contrast, functional expression of the hH₂R required the generation of an hH₂R-Gsα fusion protein to ensure close proximity of G protein and receptor. Fusion of hH₂R to the long (Gsα_L) or short (Gsα_S) splice variant of Gα_s resulted in comparable constitutive hH₂R activity, although both G protein variants show different GDP affinities. Medicinal chemistry studies revealed profound species differences between hH₁R/hH₂R and their guinea pig orthologues gpH₁R/gpH₂R. The causes for these differences were analyzed by molecular modeling in combination with mutational studies. Co-expression of the hH₃R with Gα_i1, Gα_i2, Gα_i3, and Gα_i/o in Sf9 cells revealed high constitutive activity and comparable interaction efficiency with all G protein isoforms. A comparison of various cations (Li⁺, Na⁺, K⁺) and anions (Cl⁻, Br⁻, I⁻) revealed that anions with large radii most efficiently stabilize the inactive hH₃R state. Potential sodium binding sites in the hH₃R protein were analyzed by expressing specific hH₃R mutants in Sf9 cells. In contrast to the hH₃R, the hH₄R preferentially couples to co-expressed Gα_i2 in Sf9 cells. Its high constitutive activity is resistant to NaCl or GTPγS. The hH₄R shows structural instability and adopts a G protein-independent high-affinity state. A detailed characterization of affinity and activity of a series of hH₄R antagonists/inverse agonists allowed first conclusions about structure/activity relationships for inverse agonists at hH₄R. In summary, the Sf9 cell system permitted a successful side-by-side comparison of all four human histamine receptor subtypes. This chapter summarizes the results of pharmacological as well as medicinal chemistry/molecular modeling approaches and demonstrates that these data are not only important for a deeper understanding of H_xR pharmacology, but also have significant implications for the molecular pharmacology of GPCRs in general.

Parts of this chapter were previously used in a habilitation thesis (EHS).

1 Principles of GPCR Analysis in the Sf9 Cell Expression System

1.1 The Sf9 Cell Expression System

Pharmacological characterization of GPCRs is commonly performed in transfected mammalian cells or in cells that endogenously express the receptor of interest (Kenakin 1996). There are, however, several problems of mammalian cell systems. First, mammalian cells normally express various additional GPCRs, which may result in GPCR heteromerization or signaling crosstalk (Breitwieser 2004; Prezeau et al. 2010; Gomes et al. 2016). For example, signaling crosstalk between GPCRs has been described for the Gα_i-coupled GABA_BR and the Gα_q-coupled mGlu_1AR (Rives et al. 2009). Another example is ACKR1 (atypical chemokine receptor 1), which has been shown to functionally antagonize CCR5 by forming ACKR1/CCR5 heterodimers (Chakera et al. 2008). Second, the presence of other constitutively active receptors may interfere with the analysis of agonist-independent activity of the receptor of interest. For example, the inverse FPR1 agonist cyclosporin H failed to inhibit basal Gα_i protein activity in HL-60 cells, indicating that these cells additionally express other constitutively active receptors different from FPR1 (Wenzel-Seifert and Seifert 1993; Seifert and Wenzel-Seifert 2003). Third, promiscuous G protein coupling of GPCRs in the presence of several G protein subtypes may preclude the analysis of GPCR-G protein selectivity (Woehler and Ponimaskin 2009). Finally, some GPCRs are only expressed at low levels in mammalian cells, rendering it difficult to obtain a sufficiently high signal-to-noise ratio in functional and ligand binding assays.

As discussed in a comprehensive review article (Schneider and Seifert 2010c), the problems listed above are effectively addressed by using the Sf9 cell expression system. Sf9 cells are derived from the Sf21 cell line, which had been originally isolated from the pupal ovarian tissue of the American fall army worm (Spodoptera frugiperda). The protein of interest is expressed by infecting Sf9 cells with baculoviruses encoding the corresponding gene. Although Sf9 cells express Gα_i-, Gα_q-, and Gα_s-like proteins, insect cell Gα_i is not activated by mammalian GPCRs. This renders Sf9 cells a functionally “Gα_i-free” system and permits the analysis of Gα_i-coupled receptors without the necessity of pertussis toxin (PTX)-mediated GPCR/Gα_i uncoupling. Also, PTX would not be active in Sf9 cells, because it does not enter the cells (Wenzel-Seifert et al. 1998). By contrast, uncoupling of Gα_i-coupled GPCRs by PTX in mammalian cells is problematic. Despite entering mammalian cells, PTX is not capable of completely inactivating all Gα_i proteins (Wenzel-Seifert and Seifert 1990).

Moreover, Sf9 cells do not express constitutively active GPCRs and therefore provide a low-background environment for the analysis of agonist-independent receptor activity. Furthermore, the highly efficient baculovirus promoters lead to very high expression levels of GPCRs in Sf9 cells. This results in high signal-to-noise ratios in binding assays and allows the purification of receptor protein, e.g. for crystallization purposes. Finally, as explained below, Sf9 cell membranes expressing large amounts of GPCRs and G proteins can be used to study G protein activation in steady-state GTPase assays and experiments with [³⁵S]GTPγS ([³⁵S]-labeled guanosine 5′-O-[γ-thio]triphosphate).

For the preparation of baculoviruses encoding the gene of interest, several straightforward methods are established. The Sf9 cell studies discussed in this chapter were performed by using the BaculoGold™ kit from Invitrogen. As explained in Fig. 1, the gene of interest (in this example hH₄R) is cloned into a pVL1392 baculovirus transfer vector, which is transfected into Sf9 cells together with the missing part of the baculovirus genome (BaculoGold™ DNA).

Fig. 1

Preparation of baculoviruses for the expression of a GPCR (example for H₄R-Gα_i2 fusion protein). The gene of interest (in this case a FLAG-tagged hH₄R fused to Gα_i2 via a His₆ linker) is cloned into a pVl1392 transfer vector. The plasmid and the missing part of the baculovirus DNA (BaculoGold™ DNA) are co-transfected into Sf9 cells. The full baculovirus genome is completed in Sf9 cells by homologous recombination. The cells start to release virus particles into the surrounding medium

After that, the full baculovirus genome with the integrated receptor gene is reconstituted in the host cells by homologous recombination. The cell then releases virus particles into the surrounding medium which is harvested and used for further infections. A detailed protocol for the production and maintenance of genetically modified baculoviruses was published in Methods in Enzymology (Schneider and Seifert 2010a). Numerous examples of the characterization of Gα_q-, Gα_s-, and Gα_i-coupled receptors reconstituted in Sf9 insect cells were documented by Schneider and Seifert (2010c). In this chapter, an in-depth discussion of the pharmacological characterization of histamine receptors in Sf9 cell membranes is provided.

1.2 Methods for the Characterization of Histamine Receptors in Sf9 Cell Membranes

1.2.1 The G Protein Cycle

The G protein activation cycle (Gilman 1987; Oldham and Hamm 2008), which is explained in the following, is the basis for the methods used to generate the functional histamine receptor data discussed in this chapter. When histamine binds to the hH₄R, the receptor protein undergoes a conformational change and interacts with an inactive GDP-bound heterotrimeric G protein (Fig. 2 step 1). This induces GDP release and the formation of the so-called ternary complex, which contains agonist, receptor and guanine-nucleotide-free G protein (Fig. 2, step 2). It is generally accepted that a GPCR exhibits its highest agonist-binding affinity, when it is part of the ternary complex. The interaction between agonist-bound GPCR and G protein promotes GTP binding to the Gα-subunit. This weakens the intermolecular interactions in the G protein and in the ternary complex, breaking the complex up into agonist and GPCR as well as Gα- and Gβγ subunit (Fig. 2, step 3).

Fig. 2

Stimulation of Gα_i-proteins by the histamine H₄R and resulting G protein cycle. The numbers designate the different stages of the cycle and are explained in detail in Sect. 1.2.1

After their dissociation from the receptor, the active GTP-loaded Gα subunit and the Gβγ part interact with various effector proteins (Fig. 2, step 4) and induce numerous biochemical processes. Such effects include activation (Gα_s) or inhibition (Gα_i) of membranous adenylyl cyclase (AC), modulation of ion channel activity (Gβγ, Gα_i) or stimulation of phospholipase C (PLC) activity followed by intracellular Ca²⁺ mobilization (Gβγ, Gα_q). As long as GTP is bound to Gα, the Gα and Gβγ subunits are active. To terminate signaling, the Gα subunit inactivates itself by its intrinsic GTPase activity, resulting in conversion of the bound GTP to GDP and release of inorganic phosphate (Fig. 2, step 5). The inactive GDP-bound Gα subunit re-associates with Gβγ and becomes available for another cycle (Fig. 2, step 6).

1.2.2 High Affinity Radioligand Binding

High affinity radioligand binding with histamine receptors is performed with radiolabeled agonists, e.g. tritiated histamine ([³H]histamine). Normally, agonists show their highest affinity to the ternary complex (Fig. 2, step 2) and stabilize the active receptor conformation. Thus, agonistic radioligands preferentially label the G protein-coupled high-affinity receptor population. When two populations of GPCRs with different G protein coupling states occur simultaneously, the saturation or competition curves with agonistic radioligands may become biphasic, which allows the determination of high-affinity and a low-affinity binding constants. This was, e.g., demonstrated for histamine H₂R (Houston et al. 2002) as well as for the β₂-adrenergic receptor (β₂AR) or the dopamine D₁R (Gille and Seifert 2003). By contrast, inverse agonists interact preferentially with the inactive receptor state and therefore show increased affinity to uncoupled GPCRs. Neutral antagonists do not discriminate between active and inactive receptor states and label both receptor conformations with comparable affinity.

For some experiments it may be required to convert GPCRs to their inactive conformation by disrupting receptor-G protein interactions. This is achieved by addition of GTPγS (guanosine 5′-O-[γ-thio]triphosphate), which binds to the Gα-subunit like GTP (Fig. 2, step 3), but cannot be hydrolyzed by the Gα subunit (Gilman 1987). Thus, no GDP-loaded G protein is available anymore for the formation of new ternary complexes resulting in uncoupling of the entire GPCR population. This is normally reflected by a dramatic reduction in the binding affinity of agonistic radioligands. A detailed protocol for high-affinity agonist binding assays as well as example data for various receptor/G protein systems is provided in book chapters about GPCR/G protein co-expression and fusion protein systems in Sf9 cell membranes (Schneider and Seifert 2010a, b).

1.2.3 Steady-State GTPase Assays

In steady-state GTPase assays, the intrinsic GTPase activity (Fig. 2, step 5) of the active GTP-bound Gα-subunit is determined (Gilman 1987; Schneider and Seifert 2010a). This is achieved by quantitating radioactive inorganic phosphate released after Gα-mediated hydrolysis of [γ-³²P]GTP. The steady-state GTPase assay represents a very proximal readout of GPCR activation, which directly reflects GPCR-mediated G protein stimulation. By contrast, functional assays analyzing more distal parameters (e.g., Ca²⁺-, cAMP- or reporter gene assays) are often influenced by signal amplification processes, making valid conclusions about the original extent of receptor activation difficult. Technical details of the steady-state GTPase assay were explained in two book chapters about GPCR/G protein co-expression and fusion protein systems in Sf9 cells (Schneider and Seifert 2010a, b).

Steady-state GTPase assays can be used for the functional characterization of ligands in medicinal chemistry projects. In addition, these assays provide information about the efficacy of receptor-G protein interactions. In Michaelis–Menten kinetics experiments with increasing concentrations of the substrate [γ-³²P]GTP, the K _M and V _max value of the Gα-GTPase can be determined (Schneider and Seifert 2009, 2010a). Subtraction of the GTPase activity in the presence of a full inverse agonist from the activity elicited by a full agonist yields the total receptor-regulated GTPase activity (ΔV _max). Dividing the ΔV _max value by B _max (maximum number of radiolabeled receptor proteins) provides the so-called turnover number, which signifies the number of GTP molecules hydrolyzed per minute, resulting from the activation of a single GPCR protein (Schneider and Seifert 2010a).

1.2.4 [³⁵S]GTPγS Binding Assays

The [³⁵S]GTPγS binding assay is another method to determine the functional effect of a ligand at a very proximal level of GPCR signal transduction. As depicted in Fig. 2, GTPγS binds to the activated Gα-subunit instead of GTP, resulting in the dissociation of the ternary complex (Fig. 2, step 3). Unlike GTP, however, GTPγS cannot be hydrolyzed by the intrinsic GTPase activity of Gα (Gilman 1987), resulting in an accumulation of GTPγS-bound Gα-subunits. When radiolabeled [³⁵S]GTPγS is used, the amount of activated Gα subunits can be quantitated by scintillation counting, allowing the characterization of Gα activation kinetics (time course of [³⁵S]GTPγS-Gα accumulation) and the determination of agonist- and inverse-agonist modulated Gα activation. When the total ligand-regulated Gα activation (maximum effect of full agonist minus activation level in the presence of a full inverse agonist) is divided by the B _max value from radioligand binding, the so-called coupling factor is obtained. Similar to the aforementioned turnover number, the coupling factor provides information about the number of Gα subunits stimulated by a single GPCR protein.

Furthermore, saturation binding experiments with increasing concentrations of [³⁵S]GTPγS yield information about alterations of Gα affinity to [³⁵S]GTPγS under various conditions (e.g., constitutive receptor activity, agonist- or inverse agonist-induced effects). Finally, [³⁵S]GTPγS binding assays are useful to pharmacologically characterize new ligands synthesized during the course of medicinal chemistry projects. A detailed experimental protocol of [³⁵S]GTPγS binding assays as well as an explanation of how to analyze and interpret the data is provided in comprehensive book chapters about the characterization of GPCR/Gα co-expression and fusion protein systems in Sf9 cell membranes (Schneider and Seifert 2010a, b).

1.2.5 Fusion Protein Systems

Mammalian GPCRs and G protein Gα and Gβγ subunits can be readily co-expressed in the baculovirus/Sf9 cell system yielding useful systems for the pharmacological characterization of GPCR ligands and receptor-G protein interactions. However, sometimes co-expression systems produce only insufficient GPCR-mediated Gα activation (Seifert et al. 1998a; Gille and Seifert 2003). Specifically, Gα_s proteins rapidly dissociate from the plasma membrane (Yu and Rasenick 2002) and therefore cannot be efficiently activated by a co-expressed GPCR. This problem is solved by constructing GPCR-Gα fusion proteins (Fig. 3) that guarantee close proximity of receptor and G protein.

Fig. 3

Structure of a GPCR-Gα fusion protein. The GPCR is N-terminally tagged with a FLAG epitope, which allows detection by an anti-FLAG antibody, and connected to the N-terminus of a Gα-subunit via a His6 linker. Gα proteins are anchored in the plasma membrane via their acylation sites. The interaction between Gα_s proteins and the plasma membrane is only weak in co-expression systems, but can be significantly improved in GPCR-Gαs fusion proteins

This approach was successfully used for the pharmacological characterization of Gα_s-coupled receptors like the β₂AR (Bertin et al. 1994; Seifert et al. 1998a) or the histamine H₂R (Wenzel-Seifert et al. 2001). GPCR-Gα fusion proteins of β₂AR, FPR1 or dopamine D₁R allowed a detailed examination of Gα-isoform specificity of these receptors (Wenzel-Seifert et al. 1999; Wenzel-Seifert and Seifert 2000; Gille and Seifert 2003). GPCR-Gα fusion proteins are also useful controls to exclude activation of Sf9 cell G proteins by a specific mammalian GPCR. Normally, the turnover number from steady-state GTPase assays or the coupling factor from [³⁵S]GTPγS binding experiments should be around unity in fusion protein systems, corresponding to linear signaling. A coupling factor >1 in a GPCR-Gα fusion protein system, however, indicates additional activation of insect cell proteins.

The fusion protein approach can also be applied to generate GPCR-RGS fusion proteins. RGS proteins (regulators of G protein signaling) activate the intrinsic GTPase activity of Gα proteins. GPCR-RGS fusion proteins bring the RGS protein in close proximity to receptor and G protein. This may enhance signal intensity in steady-state GTPase assays. The first GPCR-RGS fusion proteins were constructed in 2003 (Bahia et al. 2003). A detailed discussion of various aspects of co-expression and fusion protein systems was provided in Methods in Enzymology (Schneider and Seifert 2010a, b).

2 Pharmacological Characterization of Human Histamine Receptors in Sf9 Insect Cells

The biogenic amine histamine is formed by histidine decarboxylase (HDC)-mediated decarboxylation of the precursor amino acid histidine. Histamine is stored in granula of mast cells and basophils and occurs in enterochromaffin-like cells of the stomach (Panula et al. 2015). Moreover, by means of a highly sensitive HPLC-MS/MS-based detection method, histamine was identified in lymph nodes and thymus of C57Bl/6 and Balb/c mice (Zimmermann et al. 2011). In the central nervous system (CNS), histamine occurs as a neurotransmitter. It is synthesized in histaminergic neurons that emerge from the tuberomamillary nucleus (TMN) in the posterior hypothalamus and spread to numerous regions throughout the brain (Schneider et al. 2014a, b; Panula et al. 2015). The distribution of histamine in the body indicates its most important functions, namely the regulation of inflammatory/allergic reactions, stimulation of gastric acid secretion and neurotransmission. Most of the histamine effects are mediated by four G protein-coupled receptors, H₁R, H₂R, H₃R, and H₄R (Seifert et al. 2013; Panula et al. 2015). Additionally histamine acts on some non-histaminergic targets, e.g. at NMDA receptors (Vorobjev et al. 1993; Panula et al. 2015), which, however, is not in the focus of mainstream histamine research.

This section addresses the results obtained from the pharmacological characterization of the four human histamine receptor isoforms in Sf9 cells. Other species variants will only be mentioned, when this is required by the context (e.g., comparisons of human and guinea pig H₁R or H₂R). Moreover, data from the characterization of ligands in medicinal chemistry projects will only be discussed, when they lead to new insights about structure and conformation of the corresponding receptor. Finally, publications that contain only “in silico” results without experimental verification will be omitted, since the purpose of this chapter is specifically the expression and characterization of human histamine receptors in the Sf9 cell system. For detailed information on the analysis of histamine receptor species variations in Sf9 cells or for the characterization of histamine receptor subtypes in cellular systems other than insect cells, the reader is referred to comprehensive review articles (Seifert et al. 2013; Strasser et al. 2013; Panula et al. 2015).

2.1 The Histamine H₁ Receptor

2.1.1 General Information About the Histamine H₁R

The H₁R is ubiquitously expressed, specifically in lung, CNS, and blood vessels. It preferentially couples to Gα_q/11 proteins, causing PLC and protein kinase C (PKC) activation as well as inositol-1,4-5-trisphosphate (IP₃) formation and intracellular Ca²⁺ release (Seifert et al. 2013; Panula et al. 2015). The typical signs of a type I allergic reaction like pruritus, increased vascular permeability, and edema are caused by H₁R activation. Therefore, administration of H₁R antagonists (so-called antihistamines) belongs to the most important anti-allergic therapeutic interventions (Simons and Simons 2011), e.g. for the treatment of allergic rhinitis. The H₁R is expressed on various types of immune cells, specifically on T cell subsets and dendritic cells and influences T cell polarization (Neumann et al. 2014). Moreover, as indicated by results from H₁R-deficient mice, the H₁R plays a role in various models of inflammatory diseases, e.g. nasal allergy, Th2-driven allergic asthma, atopic dermatitis or experimental autoimmune encephalitis (EAE) (Neumann et al. 2014). In the CNS, H₁R is involved in the regulation of locomotor activity, emotions, cognitive functions, arousal, sleep and circadian rhythm or pain perception (Schneider et al. 2014a). Moreover, the H₁R participates in the modulation of energy consumption, food intake, and respiration. H₁R blockade with antagonists increases susceptibility to seizures (Schneider et al. 2014a). Sedation, the most important side effect of brain-penetrating first-generation antihistamines, is caused by antagonism at H₁R in the CNS (Simons and Simons 2011; Neumann et al. 2014). The human H₁R (hH₁R) is endogenously expressed by various human cell lines. HeLa cervix carcinoma cells as well as U373 MG astrocytoma cells are used since more than two decades to study hH₁R pharmacology and signal transduction (Seifert et al. 2013). In the following, the results from the characterization of the human H₁R in the Sf9 insect cell expression system will be discussed.

2.1.2 Characterization of the hH₁R in Sf9 Cell Membranes

The hH₁R was extensively characterized in Sf9 cells with regard to ligand pharmacology, and activation of G proteins. Moreover, the pharmacological differences between the hH₁R and its guinea pig orthologue (gpH₁R) were addressed by mutational and molecular modeling studies. An overview of the most important results is provided in Table 1.

Table 1 Overview on the pharmacological characterization of the human histamine H₁R in the Sf9 cell expression system

Although Sf9 cells contain an endogenous PLC-stimulating Gα_q-like protein (Hepler et al. 1993), histamine does not induce a significant rise in steady-state GTPase activity in Sf9 cell membranes expressing the hH₁R alone (Houston et al. 2002). Only co-expression of the hH₁R with the regulators of G protein signaling RGS4 and GAIP (G-alpha-interacting protein, RGS19) unmasks an interaction of hH₁R with insect cell Gαq, resulting in histamine-induced stimulatory effects of 142% (RGS4) and 126% (GAIP) (Houston et al. 2002). These results indicate that the intrinsic GTPase activity of Sf9 cell Gα_q is rate-limiting for hH₁R-mediated G protein activation in Sf9 cell membranes. This is probably due to a low number of G proteins relative to hH₁R molecules. RGS proteins commonly accelerate the intrinsic GTPase activity of Gα proteins, which results in a higher turnover and in increased availability of inactive GDP-bound Gα subunits (Fig. 2).

Due to its favorable properties, the Sf9 cell hH₁R/RGS protein co-expression system is routinely used to characterize affinity (radioligand binding), activity (steady-state GTPase assays), and binding mode of hH₁R ligands in medicinal chemistry projects. This revealed major pharmacological differences between H₁R species isoforms. Specifically, some agonistic bulky 2-phenylhistamines and histaprodifens exhibited increased efficacy and up to tenfold higher potency at gpH₁R as compared to hH₁R (Seifert et al. 2003). Such differences were also observed for antagonists. Most notably, the potency of several arpromidine-type H₁R antagonists was up to tenfold higher at gpH₁R than at hH₁R (Seifert et al. 2003). Mutagenesis experiments were performed to elucidate the molecular basis of these pharmacological species differences. Basing on the hypothesis that smaller amino acid substitutions render the gpH₁R binding pocket more flexible than the corresponding site at the hH₁R, the amino acids 153 or 433 of the hH₁R were mutated into “gpH₁R direction” (Phe-153 → Leu 153 or Ile-433 → Val 433) (Seifert et al. 2003). Although this attempt was unsuccessful in terms of generating gpH₁R-like pharmacology, the mutations dramatically decreased hH₁R receptor expression, function, electrophoretic mobility as well as [³H]mepyramine (tritiated 2-((2-(Dimethylamino)ethyl)(p-methoxybenzyl)amino)-pyridine) affinity, suggesting that these amino acid positions are essential for correct folding and expression of the H₁R (Seifert et al. 2003). In addition, the hH₁R-F153L/I433V double mutant was studied. Although this protein was excellently expressed in Sf9 cell membranes, there were only partial changes in pharmacology. Thus, Phe-153 and Ile-433 cannot fully explain the species difference between hH₁R and gpH₁R (Seifert et al. 2003).

A series of chiral histaprodifens was pharmacologically characterized at hH₁R and gpH₁R as well as rat (r) and bovine (b) H₁R, revealing differential interaction with H₁R species isoforms. Two of the compounds showed agonism at gpH₁R, but were antagonists at hH₁R, bH₁R, and rH₁R. The histaprodifens followed the rank order of potency hH₁R < bH₁R < rH₁R < gpH₁R. The hH₁R was pharmacologically and structurally similar to bH₁R, while gpH₁R resembled rH₁R (Strasser et al. 2008a). Docking studies with an active-state model of the gpH₁R and dimeric histaprodifen revealed multiple interaction sites, involving hydrogen bonds and electrostatic interactions with Asp-116, Ser-120, Lys-187, Glu-190 and Tyr-432 (Strasser et al. 2008a).

Since the amino acid sequence of the N-terminus and the second extracellular loop (ECL2) exhibit major differences between hH₁R and gpH₁R, it was hypothesized that these structures may be responsible for the preferred binding of bulky agonists to gpH₁R as compared to hH₁R. To address this hypothesis, wild-type hH₁R and gpH₁R as well as the chimeric receptors h(gpE2)H₁R (hH₁R with ECL2 from gpH₁R) and h(gpNgpE2)H₁R (hH₁R with N-terminus and ECL2 from gpH₁R) were co-expressed with RGS4 in Sf9 cells and compared in radioligand binding and steady-state GTPase assays (Strasser et al. 2008b). A small inverse agonistic effect of mepyramine suggests that all four receptors show only low constitutive activity. Histamine potency in steady-state GTPase assays decreased in the series hH₁R > h(gpE2)H₁R > h(gpNgpE2)H₁R. Maximum G_q-protein activation by histamine and the ΔV _max/B _max ratio (turnover number) was significantly enhanced at h(gpNgpE2)H₁R as compared to hH₁R, gpH₁R, and h(gpE2)H₁R, despite a very low expression level of h(gpNgpE2)H₁R. This indicates that histamine induces a h(gpNgpE2)H₁R conformation which is specifically efficient at activating G proteins (Strasser et al. 2008b). Molecular dynamics simulations suggest that the replacement of N-terminus and ECL2 affect the network of hydrogen bonds between N-terminus, ECL1 and ECL2 and alter the conformation and flexibility of ECL2. Thus, either the replacement of the N-terminus or the combined exchange of N-terminus and ECL2 induces conformational alterations that increase the stimulatory effect of histamine and reduce its potency (Strasser et al. 2008b).

The hypothesis that major differences of N-terminus and ECL2 cause the distinct pharmacology of hH₁R and gpH₁R, however, had to be rejected, since neither binding assays nor steady-state GTPase assays revealed more pronounced “gpH₁R-like” properties of h(gpNgpE2)H₁R and h(gpE2)H₁R (Strasser et al. 2008b). Instead, three members of a new class of histaprodifens (phenoprodifens) even exhibited a reduction of pK _i and pEC₅₀ values in the series hH₁R > h(gpE2)H₁R > h(gpNgpE2)H₁R (Strasser et al. 2008b). Previous molecular dynamics simulations with these compounds had suggested that they can adopt two distinct orientations in the gpH₂R binding pocket (Strasser et al. 2008a). Thus, the data may be explained by a change in ligand orientation in the series hH₁R – h(gpE2)H₁R – h(gpNgpE2)H₁R. Such changes, however, are probably determined early in ligand binding, which can only be addressed by kinetic binding studies (Strasser et al. 2008b).

Such experiments were performed with the antagonist [³H]mepyramine and the partial agonist phenoprodifen using Sf9 cell membranes expressing RGS4 together with hH₁R, gpH₁R as well as the chimeric receptors h(gpNgpE2)H₁R and h(gpE2)H₁R (Wittmann et al. 2011). With regard to the association rate constant, h(gpNgpE2)H₁R significantly differed from both hH₁R and gpH₁R. Molecular dynamics simulations helped to explain, how the extracellular surface of the H₁R influences ligand binding kinetics, recognition of the ligand and guiding of the ligand into the binding pocket (Wittmann et al. 2011).

There are also exceptions, where bulky agonists do not interact more efficiently with gpH₁R than with hH₁R. Specifically, N ^G-acylated imidazolylpropylguanidines (AIPGs) are partial H₁R agonists that exhibit higher efficacies at hH₁R as compared to gpH₁R (Xie et al. 2006a, b). Moreover, another study addressing the pharmacology of phenylhistamines and phenoprodifens at human, guinea pig, bovine, and rat H₁R identified bulky phenylhistamines with higher potency and affinity at hH₁R as compared to gpH₁R (Strasser et al. 2009). A comparison of the hypothesized binding modes of these compounds with the binding mode of the previously characterized N ^G-acylated imidazolylpropylguanidine UR-AK57 (N ¹-(3-Cyclohexylbutanoyl)-N ²-[3-(1H-imidazol-4-yl)propyl]guanidine) (Xie et al. 2006b) suggests that the higher potency at the hH₁R is caused by a more pronounced van der Waals interaction with Asn^2.61 of hH₁R as compared to Ser^2.61 of gpH₁R (Strasser et al. 2009). Moreover, phenoprodifens seem to adopt two distinctly oriented binding modes that cause the highly conserved Trp^6.48, which is part of the toggle switch mechanism of GPCR activation (Shi et al. 2002), to assume either an active or an inactive conformation (Strasser et al. 2009).

2.2 The Histamine H₂ Receptor

2.2.1 General Information About the Histamine H₂R

The H₂R is ubiquitously expressed, most importantly in stomach, heart, and CNS (Seifert et al. 2013; Schneider et al. 2014a; Panula et al. 2015). Agonist binding to this receptor results in activation of Gα_s-proteins that stimulate the adenylyl-cyclase-mediated production of the second messenger cAMP (Panula et al. 2015). The central role of the H₂R in the regulation of gastric acid production is the basis for the therapeutic use of H₂R antagonists to treat gastroesophageal reflux disease (Schubert and Peura 2008). The function of the H₂R in the brain is less well documented as for H₁R, but includes, e.g. modulation of cognitive processes and of circadian rhythm (Schneider et al. 2014a). Moreover, H₂R influences glucose metabolism and food intake (Schneider et al. 2014a).

Experiments with knockout mice have revealed that the histamine H₂R is involved in the regulation of immune responses, specifically in the modulation of Th1- or Th2-cell polarization. It should be noted, however, that the analysis of H₂R-deficient mice yields conflicting results, probably because of the variability of the disease models studied (Neumann et al. 2014). The human histamine H₂R (hH₂R) has been pharmacologically characterized in both human cells and in the Sf9 cell expression system (Seifert et al. 2013). Neutrophils are specifically well suited for the analysis of hH₂R pharmacology, because they are primary cells that can be easily isolated from human blood in large numbers. The hH₂R inhibits superoxide anion production induced by chemotactic peptides in neutrophils (Burde et al. 1989, 1990; Reher et al. 2012a) and eosinophils (Reher et al. 2012a). Moreover, H₂R activation induces functional differentiation of HL-60 promyelocytes (Klinker et al. 1996). Furthermore, it is discussed that decreased hH₂R function may contribute to inflammation in bronchial asthma (Seifert et al. 2013).

2.2.2 Characterization of the hH₂R in Sf9 Cell Membranes

The hH₂R was extensively characterized in Sf9 cells with regard to ligand pharmacology, and activation of G proteins. Moreover, the pharmacological differences between the hH₂R and its guinea pig orthologue (gpH₂R) were addressed by mutational and molecular modeling studies. An overview of the most important results is provided in Table 2.

Table 2 Overview on the pharmacological characterization of the human histamine H₂R in the Sf9 cell expression system

Functional expression of the human hH₂R in Sf9 cells requires Gα_s proteins as intracellular coupling partners. Indeed, Sf9 cells express endogenous Gα_s proteins and activation of Sf9 cell Gα_s has been reported for mammalian GPCRs, e.g. the bradykinin B2 receptor (Shukla et al. 2006), the LH/CG receptor (Narayan et al. 1996), or the histamine H₂R (Kühn et al. 1996). Mostly, however, the interaction of mammalian GPCRs with Sf9 cell Gα_s shows only low productivity, which is most likely due to rapid dissociation of the activated Gα_s subunit from the plasma membrane. Redistribution of stimulated Gα_s proteins has been investigated in more detail in S49 lymphoma cells treated with the β-AR agonist isoproterenol (Ransnäs et al. 1989).

Fusion of a GPCR to Gα_s keeps the G protein at the cell membrane and largely enhances G protein activation. This approach was used for the human histamine H₂R, which was expressed in Sf9 cells as a fusion protein with the long (G_sαL) or short (G_sαS) splice variant of Gα_s (Wenzel-Seifert et al. 2001). Both fusion proteins were expressed at a similar level in Sf9 cell membranes and the affinity of the radiolabeled H₂R agonist [³H]tiotidine (tritiated 1-cyano-3-[2-[[2-(diaminomethylideneamino)-1,3-thiazol-4-yl]methylsulfanyl]ethyl]-2-methyl-guanidine) was comparable (~ 32 nM) for hH₂R-Gsα_L and hH₂R-Gsα_S (Wenzel-Seifert et al. 2001). Unexpectedly, the B _max values of ligand-regulated [³⁵S]GTPγS binding for hH₂R-Gsα_L or hH₂R-Gsα_S exceeded the B _max value from [³H]tiotidine binding by ~tenfold, which suggests that a large subpopulation of fusion proteins is not labeled by the radioligand (Wenzel-Seifert et al. 2001).

G_sαL exhibits lower GDP affinity than Gsα_S, and therefore, the β₂AR-G_sαL fusion protein shows higher constitutive activity than β₂AR-Gsα_S (Seifert et al. 1998b). Similarly, the hH₂R-Gsα_L fusion protein exhibited a faster GDP/GTPγS exchange than hH₂R-Gsα_S. Surprisingly, however, unlike the corresponding β₂AR fusion proteins, hH₂R-Gsα_L and hH₂R-Gsα_S showed similar constitutive activity and comparable pharmacological properties of partial agonists and inverse agonists in steady-state GTPase and [³⁵S]GTPγS binding assays (Wenzel-Seifert et al. 2001). This illustrates that the GDP affinity of G proteins does not influence the constitutive activity of all GPCRs to the same extent (Wenzel-Seifert et al. 2001).

It has been reported that the rH₂R couples to insect cell Gα_q and increases intracellular Ca²⁺ in Sf9 cells (Kühn et al. 1996). However, this effect could not be confirmed and was also not observed with hH₂R or gpH₂R (Houston et al. 2002). Moreover, co-expressed GAIP did not unmask a potential interaction of hH₂R with insect cell Gα_q (steady-state GTPase assays) although this approach was successful with hH₁R (Houston et al. 2002). The hH₂R did not even activate mammalian Gα_q co-expressed in Sf9 cells or fused to the hH₂R (Ca²⁺ assays, high-affinity agonist binding and [³⁵S]GTPγS binding) (Houston et al. 2002). Surprisingly, not even the hH₁R was able to activate co-expressed mammalian Gα_q in Sf9 cells. Thus, mammalian Gα_q was probably inactive in Sf9 cells, despite high expression levels, and therefore, Sf9 cells are not suited to investigate the interaction of GPCRs with mammalian Gα_q (Houston et al. 2002).

When only hH₂R was expressed in Sf9 cells, no ternary complex formation with insect cell Gα_s was observed in high-affinity agonist binding with [³H]tiotidine (effect of GTPγS on histamine competition curve) and in [³⁵S]GTPγS binding (characterization of the stimulatory effect of histamine). Surprisingly, however, AC assays clearly indicated hH₂R-mediated activation of insect cell Gα_s. Thus, AC assays probably exhibit higher sensitivity than [³H]tiotidine high-affinity agonist binding or [³⁵S]GTPγS binding and detect even very low insect cell Gα_s stimulation (Houston et al. 2002). Co-expression of hH₂R with mammalian Gsα_S resulted in efficient G protein interaction (high-affinity agonist binding, [³⁵S]GTPγS binding, AC assays). A further increase in interaction efficiency was observed for the hH₂R-Gsα_S fusion protein (Houston et al. 2002).

The fusion protein approach was also used for the pharmacological comparison of hH₂R and gpH₂R (Kelley et al. 2001). In [³H]tiotidine radioligand binding assays, the hH₂R-Gsα_S fusion protein expressed in Sf9 cells bound large guanidine-type agonists with lower affinity than gpH₂R-Gsα_S. Moreover, GTPγS disrupted high-affinity binding of guanidine-type agonists at hH₂R-Gsα_S more efficiently than at gpH₂R-Gsα_S. This indicates that the guanidine-stabilized conformation of gpH₂R interacts more tightly with the tethered G protein than the corresponding conformation of hH₂R (Kelley et al. 2001). In steady-state GTPase assays, the potencies and efficacies of guanidines were also higher with gpH₂R-Gsα_S than with hH₂R-Gsα_S. However, the species isoforms did not differ in case of small agonists or antagonists (Kelley et al. 2001).

Based on molecular modeling data (bovine rhodopsin-based alignment), it was hypothesized that the high potency of guanidine-type agonists at gpH₂R is caused by the non-conserved Asp-271 in TM7 (Ala-271 in hH₂R). This hypothesis was tested by expressing the mutant hH₂R-A271D-Gsα_S as well as the chimeras NgpChH₂R-Gsα_S (N-terminus – TM3 from gpH₂R and TM4-C-terminus from hH₂R, containing Ala-271) and NhCgpH₂R-Gsα_S (N-terminus – TM3 from hH₂R, and TM4-C-terminus from gpH₂R, containing Asp-271) in Sf9 cell membranes (Kelley et al. 2001). In fact, steady-state GTPase assay data clearly showed increased potency of guanidines at both hH₂R-A271D-Gsα_S and NhCgpH₂R-Gsα_S, confirming the importance of Asp-271 in the gpH₂R for guanidine binding. Unexpectedly, the efficacies of guanidine-type agonists at hH₂R-Gsα_S and NgpChH₂R-Gsα_S as well as the more “gpH₂R-like” constructs hH₂R-A271D-Gsα_S and NhCgpH₂R-Gsα_S were lower than at gpH₂R. This demonstrates that potency and efficacy are independent properties of the H₂R. The modeling and experimental data suggest that an interaction between TM1 (Tyr-17) and TM7 (Asp-271) is important for the stabilization of the guanidine-induced agonistic conformation of the gpH₂R and therefore for guanidine efficacy. This interaction is absent in hH₂R and in the other constructs analyzed by Kelley et al. (2001).

The hypothesis that a Tyr-17/Asp-271 interaction in the gpH₂R molecule stabilizes an active receptor conformation and increases efficacy of guanidine-type agonists was tested by characterizing the mutant fusion proteins hH₂R-C17Y-Gsα_S and hH₂R-C17Y-A271D-Gsα_S (Preuss et al. 2007b). As expected, the potencies and efficacies of guanidines in the steady-state GTPase assay were higher at the hH₂R-C17Y-A271D-Gsα_S double mutant as compared to the wild-type hH₂R-Gsα_S fusion protein, but they were still below the values determined for wild-type gpH₂R-Gsα_S. Thus, the Tyr-17/Asp-271 interaction is probably not solely responsible for the different pharmacology of hH₂R and gpH₂R (Preuss et al. 2007b). Moreover, the data suggest the stabilization of ligand-specific receptor conformations by agonists and inverse agonists in wild-type and mutant hH₂R-Gsα_S fusion proteins (Preuss et al. 2007b).

The results from the analysis of the hH₂R-C17Y-Gsα_S single mutant support the notion that an H-bond between Tyr-17 and Asp-271 stabilizes an active receptor conformation (Preuss et al. 2007b). The hH₂R-C17Y-Gsα_S fusion protein exhibits lower basal AC and decreased agonist-induced GTPase activities (Preuss et al. 2007b), indicating impaired G protein coupling. One possible explanation may be degradation of the hH₂R-C17Y-Gsα_S fusion protein in the Sf9 cells. This is suggested by the apparent molecular mass of 40 kDa instead of the expected ~80 kDa in Western blots (Preuss et al. 2007b).

In bovine rhodopsin (Palczewski et al. 2000) as well as in various aminergic GPCRs, e.g. dopamine D₂R (Shi and Javitch 2002), adenosine A_2aR (Kim et al. 1996), or muscarinic M₃ receptor (Scarselli et al. 2007), residues in the second extracellular loop, ECL2, probably contribute to ligand binding. Thus, it was hypothesized that differences in e2 may also determine the distinct pharmacology of hH₂R-Gsα_S and gpH₂R-Gsα_S (Preuss et al. 2007c). This hypothesis was addressed by generating mutant fusion proteins with the four e2 amino acids of hH₂R exchanged by the corresponding residues of gpH₂R (hH₂R-gpE2-Gsα_S) and vice versa (gpH₂R-hE2-Gsα_S). Steady-state GTPase assays, however, revealed that this exchange of ECL2 did not significantly alter the pharmacology of the receptors. Thus, the mutated residues most likely do not interact with the guanidine-binding pocket (Preuss et al. 2007c).

In both hH₂R and gpH₂R, Cys-174 probably forms a disulfide bond with Cys-91 in TM3 and is framed by two lysines in position 173 and 175 (Preuss et al. 2007c). A homology model of the hH₂R predicted that these two lysines are located close to the binding site of guanidine-type agonists and are involved in agonist binding (Preuss et al. 2007c). Thus, the two mutated fusion proteins hH₂R-K173A-Gsα_S and hH₂R-K175A-Gsα_S were expressed in Sf9 cells and analyzed in steady-state GTPase activity assays. The results, however, indicate that these mutations were ineffective at altering potency or efficacy of small as well as bulky H₂R agonists (Preuss et al. 2007c). Interestingly, the effect of histamine on steady-state GTPase activity of both hH₂R-K173A-Gsα_S and hH₂R-K175A-Gsα_S was reduced, which suggests that the lysines in positions 173 and 175 increase the efficiency of hH₂R-coupling to Gα_s (Preuss et al. 2007c).

2.3 The Histamine H₃ Receptor

2.3.1 General Information About the hH₃R

The Gα_i/o-coupled histamine H₃R is mainly expressed on neurons and acts as a presynaptic auto- and heteroreceptor. It inhibits the release of histamine (Arrang et al. 1983, 1985), but also of other neurotransmitters such as acetylcholine, noradrenaline, dopamine, or glutamate (Haas et al. 2008). Additionally, there is increasing evidence that H₃R is expressed postsynaptically (Ellenbroek and Ghiabi 2014), where it regulates, e.g. dopamine D₁R signaling (Ferrada et al. 2008; Brabant et al. 2009). Knockout mouse models demonstrate that the H₃R regulates numerous behaviors like locomotor activity, pain perception, food intake, memory, circadian rhythm, cognition, and anxiety (Schneider et al. 2014b). Moreover, H₃R-deficiency reduces addictive behavior in mouse models of ethanol consumption, which is probably due to the reward-inhibiting function of an increased histamine release (Vanhanen et al. 2013; Schneider et al. 2014b). This renders the H₃R an interesting target for the treatment of alcohol addiction (Nuutinen et al. 2012). Despite the decade-long research on H₃R pharmacology, only the inverse H₃R agonist pitolisant is currently used as an orphan drug to treat narcoleptic patients (Dauvilliers et al. 2013). Mouse models suggest that, in contrast to the other three histamine receptor subtypes, the H₃R does not seem to play a major role in immunological processes and inflammation (Neumann et al. 2014).

2.3.2 Characterization of the hH₃R in Sf9 Cell Membranes

There is no standard human cell culture model available that endogenously expresses hH₃R. Thus, expression and characterization of hH₃R and its species orthologues in the Sf9 insect cell system is of major importance (Schnell et al. 2010a, b; Schnell and Seifert 2010; Seifert et al. 2013; Strasser et al. 2013). Sf9 cells do not express endogenous Gαi-like protein that could interact with the corresponding mammalian GPCRs. It is, therefore, required to co-express the receptor of interest with mammalian Gα_i and Gβγ subunits. This, however, provides the unique opportunity to freely combine Gα_i-coupled receptors with any Gα_i/o isoform, allowing the characterization of Gα_i isoform specificity of GPCRs. As described in the following sections, the pharmacology of the hH₃R was extensively characterized in Sf9 cells. An overview of the most important results is provided in Table 3.

Table 3 Overview on the pharmacological characterization of the human histamine H₃R in the Sf9 cell expression system

2.3.2.1 Specificity of the hH₃R for Gα_i/o Isoforms and Investigation of Protean Agonism

The hH₃R was co-expressed in Sf9 cells with Gβ₁γ₂ and Gα_i1, Gα_i2, Gα_i3, or Gα_o. All hH₃R/G protein combinations could be readily expressed in Sf9 cells, and a semiquantitative analysis of expression levels by Western blot (purified Gα_i2 and Gα_o as reference) yielded receptor-to-G protein ratios between 1:50 and 1:100 (Schnell et al. 2010a). The receptor expression levels determined by Western blot were confirmed by radioligand saturation binding assays with the antagonist [³H]JNJ-7753707 ((4-Fluorophenyl)(1-methyl-2-{[1-(1-methylethyl)piperidin-4-yl]methoxy}-1H-imidazol-5-yl)methanone). By contrast, quantitation of the total number of activated Gα_i/o proteins in [³⁵S]GTPγS binding assays revealed a much lower amount of [³⁵S]GTPγS binding sites as compared to the Western blot results, yielding hH₃R/Gα_i isoform coupling ratios between 1:2 (hH₃R/Gα_i1) and 1:11 (hH₃R/Gα_o) (Schnell et al. 2010a).

Potencies and efficacies of the physiological agonist histamine and the inverse agonist thioperamide (N-Cyclohexyl-4-(imidazol-4-yl)-1-piperidinecarbothioamide) were determined in steady-state GTPase assays for all hH₃R/Gα_i/o combinations (Schnell et al. 2010a). When hH₃R was expressed in Sf9 cell membranes without any mammalian G protein, the signals induced by histamine and thioperamide were only small, indicating that hH₃R-mediated stimulation of insect cell G proteins was virtually absent (Schnell et al. 2010a). A comparison of all five expression systems (hH₃R alone and combined with Gα_i1, Gα_i2, Gα_i3, or Gα_o) revealed that the relative stimulatory signal of histamine and the relative inhibitory signal of thioperamide were comparable, indicating that the constitutive activity of hH₃R does not depend on the type of co-expressed Gα_i/o protein (Schnell et al. 2010a). Overall, the constitutive activity of the hH₃R was similar to the basal activity of the hH₄R (Schneider et al. 2009) (see following section).

Steady-state GTPase experiments were also performed with various hH₃R standard ligands in all hH₃R/Gα_i/o co-expression systems. N ^α-methylhistamine (NAMH) and (R)-α-methylhistamine (RAMH) turned out to be full agonists under all conditions and imetit almost reached full efficacy. Proxyfan (4-[3-(Phenylmethoxy)propyl]-1H-imidazole) and impentamine (4-(5-Aminopentyl)imidazole) were partial agonists with comparable efficacy under all conditions. Ciproxifan (cyclopropyl-(4-(3-(1H-imidazol-4-yl)propyloxy)phenyl) ketone), clobenpropit (N-(4-Chlorobenzyl)-S-[3-(4(5)-imidazolyl)propyl]isothiourea), and thioperamide exhibited inverse agonism in all systems, but efficacies were significantly different between the various Gα_i/o proteins. Nevertheless, the rank orders of potency and efficacy of the ligands remained unaltered. Taken together, these experiments again confirm the notion that the hH₃R exhibits similar pharmacological properties independently of the co-expressed Gα_i/o isoforms (Schnell et al. 2010a).

As mentioned above, the hH₃R/G protein ratios ranged between 1:2 and 1:11, indicating that it is difficult to exactly control the expression levels of receptor and G proteins. Thus, the fusion protein approach was used to ensure a 1:1 coupling ratio of hH₃R and Gα subunit. The hH₃R was fused to Gα_i2 and Gα_o, because these two Gα_i/o isoforms exhibit the lowest structural similarity. The pharmacological properties of the standard ligands histamine, imetit, proxyfan, clobenpropit, and thioperamide were similar in steady-state GTPase assays with hH₃R-Gα_i2 and hH₃R-Gα_o. This indicates again that the hH₃R pharmacology is largely independent of the type of co-expressed or fused Gα subunit (Schnell et al. 2010a).

Previously published studies about hH₃R pharmacology had reported that, depending on the expression system and the functional readout, proxyfan can be a full, a partial, or even an inverse agonist (Gbahou et al. 2003; Krueger et al. 2005). This was explained by the phenomenon of “protean agonism,” which is the ability of a ligand to induce GPCR conformations with lower G protein-coupling efficiency than the agonist-stimulated or constitutively active receptor (Gbahou et al. 2003). It has been hypothesized that protean agonism of proxyfan is due to functional selectivity, i.e. G protein coupling of the proxyfan-bound hH₃R differentiates between various Gα_i/o isoforms. The data reported by Schnell et al. (2010a), however, strongly suggest that neither proxyfan nor any other of the tested hH₃R ligands exhibits this kind of functional selectivity, at least when the hH₃R is co-expressed with or fused to various Gα_i/o isoforms in Sf9 cell membranes. One reason for this discrepancy could be the influence of different types of Gβγ subunits, which was not systematically investigated in Sf9 cells, because in the experiments performed by Schnell et al. (2010a) all hH₃R/Gα_i/o combinations were uniformly co-expressed with Gβ₁γ₂. Moreover, specific combinations of various Gα_i/o isoforms or cross-talk between signaling pathways could have influenced the results reported by Gbahou et al. (2003) and Krueger et al. (2005).

2.3.2.2 Species Differences Between Human and Rat Histamine H₃R

As discussed in the preceding section, the study of Gbahou et al. (2003) suggested that proxyfan shows protean agonism, which, however, was not confirmed in the Sf9 cell system (Schnell et al. 2010a). One of the reasons for this discrepancy could be a pharmacological difference in H₃R isoforms. Gbahou et al. (2003) used rat H₃R (rH₃R), while the experiments of Schnell et al. (2010a) were performed with hH₃R. To test this hypothesis, both species isoforms were directly compared in the Sf9 cell expression system (Schnell et al. 2010b).

Similar to the human isoform (Schnell et al. 2010a), the rH₃R was also co-expressed with Gβ₁γ₂ and the Gα_i/o isoforms Gα_i1, Gα_i2, Gα_i3, or Gα_o. A quantitation of rH₃R binding sites by radioligand binding with [³H]JNJ-7753707 and of receptor-coupled Gα subunits by [³⁵S]GTPγS binding revealed a rH₃R/G protein stoichiometry between 1:2 and 1:7 (Schnell et al. 2010b), which is comparable to the properties of the corresponding hH₃R membranes (Schnell et al. 2010a). Moreover, similar to the hH₃R, the rH₃R showed similar high constitutive activity with each of the four Gα_i/o subunits as indicated by comparable relative effects of the agonist histamine and the inverse agonist thioperamide (Schnell et al. 2010b). The independence of rH₃R pharmacology of the co-expressed Gα_i/o type was confirmed by steady-state GTPase experiments.

Several H₃R standard ligands were characterized at rH₃R (+ Gα_i2Gβ₁γ₂) and hH₃R (+ Gα_i2Gβ₁γ₂) in [³H]NAMH radioligand binding assays. The affinities of histamine, N ^α-methylhistamine, (R)-α-methylhistamine, imetit, proxifan, and clobenpropit did not differ between species isoforms, while the affinities of impentamine, imoproxifan, ciproxifan, and thioperamide were increased at the rH₃R (Schnell et al. 2010b). The radioligand binding results were largely confirmed on the functional level by steady-state GTPase experiments. Histamine, N ^α-methylhistamine, RAMH, imetit, and clobenpropit did not show species selectivity. Impentamine, however, was more potent at rH₃R than at hH₃R. Additionally, ciproxifan and thioperamide exhibited higher potency but less efficacy at rH₃R as compared to hH₃R (Schnell et al. 2010b). The hypothesis that the protean agonism of proxyfan reported by Gbahou et al. (2003) was characteristic for the rat H₃R orthologue had to be rejected, because proxyfan acted as a strong partial agonist at rH₃R expressed in Sf9 cells, independently of the co-expressed G protein (Schnell et al. 2010b).

A striking difference between hH₃R and rH₃R was observed for the H₃R ligand imoproxifan, which acted as a nearly full agonist at the hH₃R, but exhibited inverse agonism at the rat orthologue (Schnell et al. 2010b). To explain this switch in quality of action, molecular modelling studies were performed by docking imoproxifan into the binding site of the active hH₃R and the inactive rH₃R. The simulations revealed different electrostatic surfaces between TM V and TM III. While the hH₃R shows a positive surface potential in this region (NH moiety of Trp^6.48), the corresponding part of the rH₃R is slightly negatively charged (OH moiety of Thr^6.52), which results in different orientations of the ligand at both receptors. Moreover, hH₃R differs from rH₃R in amino acid position 3.37. Thr^3.37 of the hH₃R interacts with Glu^5.46, making Glu^5.46 pointing away from the binding pocket, which creates a binding site for the imoproxifan methyl moiety (Schnell et al. 2010b). By contrast, an alanine in position 3.37 of the rH₃R precludes any electrostatic interaction between Glu^5.46 and position 3.37.

Ala^3.40 of hH₃R is replaced by the bulkier Val^3.40 in rH₃R. Thus, the imoproxifan oxime moiety points downward towards Ala^3.40 in hH₃R and stabilizes Trp^6.48 in its horizontal conformation via a hydrogen bond. By contrast, the oxime moiety is directed upwards in rH₃R and interacts with Thr^6.52, while the methyl group of imoproxifan fits into a pocket between Val^3.40 and Trp^6.48. This stabilizes Trp^6.48 of rH₃R in its vertical conformation. According to the rotamer toggle switch mechanism of GPCR activation (Shi et al. 2002), the horizontal conformation of Trp^6.48 corresponds to the active state, while the vertical conformation stabilizes the inactive receptor state. Thus, this model explains the different quality of action of imoproxifan at hH₃R and rH₃R (Schnell et al. 2010b).

Interestingly, in case of imoproxifan, a comparison of steady-state GTPase assay and [³H]NAMH radioligand binding data revealed that the pEC₅₀ values at hH₃R and rH₃R were significantly higher than the corresponding pK _i values. This suggests that both hH₃R and rH₃R can adopt conformations with low affinity to partial/inverse agonists that nevertheless exhibit efficient G protein interaction (Schnell et al. 2010b).

2.3.2.3 Influence of Monovalent Ions on hH₃R Function

According to the (simplifying) two-state model of receptor activation (Fig. 4), GPCRs can adopt an active or an inactive conformation (Leff 1995). The equilibrium between both receptor states is shifted to the active side by (partial) agonists and/or interaction with G proteins. The inactive state, however, is stabilized by (partial) inverse agonists (Schneider et al. 2010b; Sato et al. 2016). The degree of constitutive activity depends on the intrinsic tendency of the receptor protein to occur in the active state. It is well established that ions are able to modulate GPCR function (Strasser et al. 2015). Specifically, sodium represents an allosteric stabilizer of the inactive receptor conformation and inhibits constitutive activity, which was, e.g., demonstrated for chemoattractant receptors (Seifert and Wenzel-Seifert 2001, 2003).

Fig. 4

Two-state model of receptor activation and factors stabilizing the active (R*) and inactive (R) receptor conformation. Every GPCR population exists in an equilibrium of active and inactive receptor conformations. Full agonists produce a maximum shift towards the active side, while inverse agonists cause a maximum stabilization of the inactive GPCR conformation. Partial agonists and partial inverse agonists induce only an incomplete shift towards either side. Neutral antagonists bind to all receptor states with the same affinity and therefore do not change the equilibrium. G proteins stabilize the active conformation, while sodium ions usually uncouple GPCRs from their G proteins by shifting the equilibrium towards the inactive side. It should be noted that, despite its usefulness, the two-state model is very simplistic and does not account for the numerous distinct ligand- and G protein-specific receptor conformations occurring in reality. Adapted from Schneider and Seifert (2010a)

As discussed above, the hH₃R exhibits high constitutive activity. Thus, hH₃R represents an interesting model for the detailed investigation of the activity-modulating effects of ions. The hH₃R was co-expressed with Gα_i2 and Gβ₁γ₂ in Sf9 cells and the influence of 100 mM of NaCl on [³H]NAMH high-affinity agonist binding and on GTP hydrolysis in the steady-state GTPase assay was investigated. Unexpectedly, in contrast to the data reported for other Gα_i/o-coupled receptors like FPR1 (Seifert and Wenzel-Seifert 2003), the affinity of the hH₃R to the radioligand was not significantly reduced by NaCl. Moreover, most surprisingly, the B _max value was even increased by NaCl. The NaCl resistance of the hH₃R in the [³H]NAMH radioligand binding assays is not fully explained yet, but may be caused by the extremely high constitutive activity of the hH₃R (Schnell and Seifert 2010).

The resistance of the hH₃R to the effect of NaCl in radioligand binding was not reflected by the data from steady-state GTPase experiments. In the presence of 100 mM of NaCl, the efficacy of histamine (full agonist) was increased and the pEC₅₀ value of histamine was reduced from 8.01 to 7.53. By contrast, the pIC₅₀ value of thioperamide (inverse agonist) was increased from 7.15 to 7.43 by NaCl, while the efficacy of thioperamide was reduced. This clearly indicates that NaCl stabilizes the inactive state of the hH₃R and reduces the constitutive activity of the system, which agrees with the predictions of the two-state model system of receptor activation (Schnell and Seifert 2010).

Since NaCl does not only contain sodium cations but also chloride anions, it is not clear if the effect of NaCl on hH₃R constitutive activity is mediated by Na⁺, by Cl⁻ or by both ions. To address this question, a profile of the effects of various monovalent cations (Li⁺, Na⁺, and K⁺) as well as of different anions (Cl⁻, Br⁻, and I⁻) was determined in steady-state GTPase assays with membranes expressing hH₃R plus Gα_i2 and Gβ₁γ₂. The rank order of efficacy was Li⁺ ~ Na⁺ ~ K⁺ < Cl⁻ < Br⁻ < I⁻. This indicates a direct proportionality between anion radii and reduction of basal hH₃R activity and shows that anions contribute more to the salt-induced reduction of constitutive activity than cations. Moreover, the different efficacies of the anions exclude the possibility that an increased osmolality may be responsible for the effect on constitutive activity (Schnell and Seifert 2010). Similar results had been previously obtained with the hβ₂AR-Gsα_L fusion protein, and it had been hypothesized that anions may enhance GDP affinity to the G protein, reducing the ability of the receptor to promote GDP dissociation (Seifert 2001). Interestingly, a comparison of the NaCl effect on hH₃R basal activity in membranes co-expressing Gβ₁γ₂ and various Gα_i/o subunits (Gα_i1, Gα_i2, Gα_i3 or Gα_i/o) revealed the strongest NaCl-mediated reduction of constitutive activity in the presence of Gα_i3 (Schnell and Seifert 2010).

It is generally assumed that the highly conserved Asp^2.50 acts as a Na⁺ binding site in GPCRs (Horstman et al. 1990; Wittmann et al. 2014). Thus, the functional consequences of a charge-neutralizing mutation from Asp^2.50 to Asn^2.50 in the hH₃R protein were investigated. In the absence of sodium, the D2.50N mutant (co-expressed with Gα_i2 and Gβ₁γ₂) exhibited a reduced number of [³H]NAMH binding sites and an affinity reduction of [³H]NAMH by about 90% as compared to the wild-type hH₃R (Schnell and Seifert 2010). Constitutive activity in steady-state GTPase assays was completely eliminated by the D2.50N mutation (co-expressed with Gα_i2 and Gβ₁γ₂) and consequently, neither thioperamide nor NaCl further inhibited basal activity. Interestingly, however, the stimulatory effect of histamine at the D2.50N mutant was highly sensitive to NaCl and was completely eliminated at NaCl concentrations > 90 mM. Most surprisingly, the D2.50N mutation introduced G protein selectivity, as the mutant did not productively interact any more with Gα_i3, but still activated Gα_i1, Gα_i2, and Gα_o1. Thus, Asp^2.50 seems to play a decisive role in the hH₃R/Gα_i3-interaction (Schnell and Seifert 2010). In summary, the characterization of the hH₃R in the Sf9 cell expression system by Schnell and Seifert (2010) revealed that Gα_i3 interacts with hH₃R in a very distinct manner as compared to the other tested Gα_i/o isoforms (stronger NaCl effect on activity of wild-type hH₃R and complete inactivity of the hH₃R-D2.50N mutant). Interestingly, the D2.50N mutant was not completely NaCl-insensitive, which indicates that the interaction between ions and hH₃R is more complex and cannot be explained by a single interaction site (Schnell and Seifert 2010).

In contrast to the hH₃R, the structurally similar hH₄R (see Sect. 2.4) exhibits completely NaCl-resistant constitutive activity (Schneider et al. 2009). A potential explanation for this discrepancy was recently offered by Wittmann et al. (2014). A comparison of various human aminergic GPCRs revealed that in the majority of receptors, glycine is the most abundant (80%) amino acid in the sodium binding channel between the ligand binding site and the sodium binding region (Wittmann et al. 2014). This is, however, not the case for hH₃R and hH₄R. Moreover, in hH₄R the glutamine in position 7.42 disrupts a water chain, which is extending from Asp^3.32 (orthosteric binding site) to Asp^2.50 (allosteric binding site). This might kinetically prevent sodium from binding to the allosteric binding site (Wittmann et al. 2014).

2.4 The Histamine H₄ Receptor

The fourth histamine receptor couples to PTX-sensitive Gα_i proteins, specifically to Gα_i2 and shows high constitutive activity (Schneider et al. 2009). The H₄R is a chemotactic receptor mainly expressed on hematopoietic cells, specifically on eosinophils (O’Reilly et al. 2002; Buckland et al. 2003; Reher et al. 2012b). Human eosinophils belong to the best characterized primary cells endogenously expressing hH₄R, but it is difficult to isolate this rare cell type in sufficiently high purity and numbers from healthy volunteers (Seifert et al. 2013). Moreover, H₄R is expressed on mast cells (Hofstra et al. 2003; Jemima et al. 2014) as well as dendritic cells (Gutzmer et al. 2005; Damaj et al. 2007; Bäumer et al. 2008; Gschwandtner et al. 2011) and expression on natural killer cells has been reported, too (Damaj et al. 2007). The presence of the H₄R on monocytes is discussed controversially (Damaj et al. 2007; Gschwandtner et al. 2013; Werner et al. 2014). Data from a comprehensive analysis of hH₄R expression on various myeloid cell types have been published very recently (Capelo et al. 2016). H₄R knockout mouse models suggest that this receptor plays a role in the pathophysiology of itch, experimental asthma and EAE (Neumann et al. 2014).

The H₄R represents an interesting target for anti-inflammatory drugs. For example, the H₄R regulates eosinophilic inflammation in a mouse model of ovalbumin-induced allergic asthma (Hartwig et al. 2015). Moreover, the hH₄R seems to be a key player in pruritus during inflammatory reactions (Bell et al. 2004; Dunford et al. 2007; Rossbach et al. 2011). However, studies with mouse models should be interpreted with caution, because H₄R pharmacology strongly differs between various species (Strasser et al. 2013). For example, the “prototypical” hH₄R antagonist JNJ7777120 (1-[(5-Chloro-1H-indol-2-yl)carbonyl]-4-methylpiperazine) is an inverse agonist at the hH₄R, but a partial agonist at the rat, mouse, and canine orthologues (Schnell et al. 2011; Strasser et al. 2013). Another caveat is H₄R-induced G protein-independent β-arrestin signaling. Although JNJ-7777120 is an inverse H₄R agonist with regard to G protein activation, it exhibits agonistic effects on H₄R-dependent β-arrestin signaling (Rosethorne and Charlton 2011; Seifert et al. 2011; Nijmeijer et al. 2013). Recently, the H₄R antagonist JNJ 39758979 ((R)-4-(3-amino-pyrrolidin-1-yl)-6-isopropyl-pyrimidin-2-ylamine) was shown to be safe and efficacious at reducing histamine-induced pruritus in a phase 1 clinical study (Kollmeier et al. 2014).

2.4.1 Successful Reconstitution of Functional Human Histamine H₄R (hH₄R) in Sf9 Cells

The N-terminally FLAG-tagged and C-terminally His-tagged wild-type hH₄R was co-expressed with Gαi₂ and Gβ₁γ₂ in Sf9 cells. Binding studies with [³H]histamine revealed a K _D value of ~10 nM (Schneider et al. 2009), which fits well to the literature range (5–20 nM). Steady-state GTPase and [³⁵S]GTPγS binding experiments confirmed the high constitutive activity of the hH₄R, which was effectively inhibited by the inverse agonist thioperamide (Schneider et al. 2009). Surprisingly, thioperamide was not able to suppress [³⁵S]GTPγS binding in the co-expression system (hH₄R + Gαi₂ + Gβ₁γ₂) to the level of control membranes expressing only Gαi₂ and Gβ₁γ₂ (Schneider et al. 2009). This strongly indicates that thioperamide is only a partial H₄R inverse agonist and not, as originally suggested in the literature (Lim et al. 2005), a full inverse agonist. The Sf9 cell system provides a “clean” background devoid of mammalian Gα_i proteins and their cognate GPCRs. Thus, expression of mammalian G proteins without GPCRs in Sf9 cells provides a valid control for baseline Gα activity and for the maximum possible effect of a full inverse agonist. In the following, the most important results from the pharmacological characterization of the hH₄R in Sf9 cell membranes are discussed. An overview of the most important results is provided in Table 4.

Table 4 Overview on the pharmacological characterization of the human histamine H₄R in the Sf9 cell expression system

2.4.2 G Protein-Independent High-Affinity-State of the hH₄R

According to the ternary complex model (De Lean et al. 1980), a GPCR shows its highest agonist affinity, when it is part of the ternary complex (Sect. 1.2.1, Fig. 2). Ternary complex formation, however, is prevented in the presence of GTPγS which binds to the Gα subunit like GTP (Gilman 1987), but cannot be hydrolyzed. Thus, GTPγS disrupts the G protein cycle, resulting in the accumulation of uncoupled inactive GPCRs with reduced agonist affinity. Surprisingly the hH₄R shows an active state which is completely independent of G proteins (Schneider et al. 2009). This is supported by the following four observations: First, high-affinity [³H]histamine binding (K _D and B _max) to membranes expressing hH₄R, Gα_i2, and Gβ₁γ₂ was retained in the presence of GTPγS. Second, [³H]histamine binding affinity was almost identical in the hH₄R/Gα_i2/Gβ₁γ₂ co-expression system and in Sf9 cell membranes expressing hH₄R in the absence of mammalian G proteins. Third, the K _i values of the inverse hH₄R agonists thioperamide and JNJ-7777120 were unaltered in membranes expressing only hH₄R, although the two-state model of receptor activation (Fig. 4) suggests that inverse agonist affinity increases, when the receptor is not coupling to G proteins and assumes an inactive state. Finally, steady-state GTPase assays with membranes co-expressing hH₄R, Gα_i2 and Gβ₁γ₂ revealed that the constitutive activity of the hH₄R is insensitive to sodium ions. According to the standard two-state model of receptor activation depicted in Fig. 4, however, it is expected that Na⁺ stabilizes the inactive state of a GPCR. This has been shown previously, e.g. for FPR-26 (Wenzel-Seifert et al. 1998; Seifert and Wenzel-Seifert 2001) or the α₂-adrenoceptor (Tian and Deth 2000).

2.4.3 Analysis of hH₄R-G Protein Coupling

Analysis of hH₄R activation in the steady-state GTPase assay in membranes co-expressing hH₄R with Gβ₁γ₂ and a specific Gα subunit (Gα_i1, Gα_i2, Gα_i3 or Gα_o) revealed that Gα_i2 was most effectively stimulated by the hH₄R. By contrast, the hH₄R hardly activated Gα_o proteins (Schneider et al. 2009). Since Gα_o is the main G protein subtype in the brain, this result suggests that the hH₄R is not of major importance in the CNS. We have seriously questioned the widespread but largely unfounded notion of functional hH₄R expression on neurons (Schneider and Seifert 2016).

The stoichiometry of the receptor-G protein interaction can be calculated by dividing the total number of receptor-regulated G proteins (from GTPγS binding assays) by the number of receptors per cell (B _max from radioligand binding or from Western blot). When co-expressed with Gα_i2 and Gβ₁γ₂ in Sf9 cell membranes, the hH₄R catalytically activates up to five Gα_i2 subunits simultaneously (Schneider et al. 2009). The affinity of [³⁵S]GTPγS to the Gα subunit (K _D value) reflects efficiency of G protein activation. The inverse agonistic character of thioperamide was confirmed in [³⁵S]GTPγS assays with membranes co-expressing hH₄R, Gα_i2 and Gβ₁γ₂. While the [³⁵S]GTPγS K _D value was 3.4 nM in the presence of histamine, it was about threefold increased by thioperamide (Schneider et al. 2009), indicating reduced [³⁵S]GTPγS affinity of the Gα subunit due to uncoupling from the hH₄R.

2.4.4 Conformational Instability of hH₄R

As demonstrated for the constitutively active mutant of the β₂-adrenoreceptor (β₂AR_CAM) (Gether et al. 1997), constitutive activity of a GPCR increases conformational flexibility and favors denaturation. By contrast, ligand binding reduces conformational flexibility and stabilizes the receptor. Thus, addition of ligands to a cell culture expressing β₂AR_CAM increased the B _max value of this receptor (Gether et al. 1997). This effect was caused by both agonists and inverse agonists, suggesting that it is the switch between different activation states rather than the nature of the activation state, which destabilizes the receptor.

The high constitutive activity of the hH₄R prompted us to investigate its conformational stability and the stabilizing effect of ligands. In fact, addition of histamine (10 μM) or thioperamide (1 μM) to Sf9 cells co-expressing hH₄R, Gα_i2 and Gβ₁γ₂ significantly increased the B _max value in histamine high-affinity agonist binding assays (Schneider et al. 2009). Interestingly, this effect was not visible in immunoblots, indicating that histamine and thioperamide mainly support the correct folding of hH₄R in the cell membrane, but not during intracellular protein synthesis. This was confirmed in experiments, where denaturation of hH₄R (co-expressed with Gα_i2 and Gβ₁γ₂) was induced by incubation of the membranes at 37°C. After 120 min, almost 70% of the histamine binding sites in the ligand-free control were lost, but only 35% in the presence of histamine. Most surprisingly, however, thioperamide increased the B _max by 30–40%, suggesting that it did not only prevent hH₄R denaturation, but even re-folded a priori misfolded receptors. This intriguing “refolding” effect of the inverse agonist thioperamide was confirmed in a two-step assay, during which the receptor was first denatured and then incubated with thioperamide.

2.4.5 Characterization of the hH₄R-Gα_i2 Fusion Protein

To analyze the interaction of the hH₄R with Gα_i2, the C-terminus of the receptor was fused to the N-terminus of the G protein by using a His₆ linker (Fig. 3). The hH₄R-Gα_i2 protein co-expressed with Gβ₁γ₂ in Sf9 cell membranes exhibited linear signaling with a coupling factor of ~1 in [³⁵S]GTPγS binding assays and a turnover number of ~1 in steady-state GTPase assays. Thus, hH₄R exclusively activates the tethered mammalian G protein but not the insect cell G proteins (Schneider et al. 2009). This was additionally supported by the lack of [³⁵S]GTPγS binding in membranes expressing non-fused hH₄R in the absence of mammalian G proteins (Schneider et al. 2009). The K _D value of [³⁵S]GTPγS in the presence of the full hH₄R agonist histamine or the inverse agonist thioperamide in membranes co-expressing hH₄R-Gα_i2 and Gβ₁γ₂ was significantly reduced as compared to the coexpression system, indicating enhanced efficiency of G protein activation (Schneider et al. 2009). A higher GTP affinity of Gα_i2 in the fusion protein was also reflected by a significantly decreased K _M value in the presence of histamine in steady-state GTPase assays. Moreover, a slight increase of constitutive activity in steady-state GTPase assays additionally demonstrates the increased efficiency of G protein activation in the fusion protein system (Schneider et al. 2009). Interestingly, the B _max value of the hH₄R-Gα_i2 fusion protein in immunoblots and [³H]histamine binding assays was increased as compared to the non-fused receptor (Schneider et al. 2009). This suggests a chaperone-like stabilizing effect of Gα_i2, favoring membrane insertion of the receptor protein. Incubation of the cell culture with histamine or thioperamide did not further enhance the B _max value of the fusion protein in [³H]histamine binding (Schneider et al. 2009), suggesting that the fusion of hH₄R to Gα_i2 induces already the maximum possible number of correctly folded receptors. An overview of the most important features of the hH₄R-Gα_i fusion protein in comparison to the co-expression system (hH₄R + Gα_i2 + Gβ₁γ₂) is provided in Table 4.

2.4.6 Role of Glycosylation for hH₄R Expression and Function

Western blotting of hH₄R-expressing Sf9 cell membranes revealed two bands at 43 and 46 kDa. Incubation of the baculovirus-infected Sf9 cell culture with the glycosylation inhibitor tunicamycin removed the 46 kDa band, indicating that this is most likely a glycosylated H₄R species (Schneider et al. 2009). Although the total protein amount on the Western blot was comparable for both untreated and tunicamycin-treated H₄R protein (2.5–3 pmol/mg as assessed by using FLAG-β₂AR standard membranes with known receptor expression levels), the B _max value in [³H]histamine binding was reduced by 75% after tunicamycin treatment. Nevertheless, the K _D value of [³H]histamine remained unchanged (Schneider et al. 2009). Thus, hH₄R deglycosylation does not significantly affect the [³H]histamine binding site of functional hH₄R, although it significantly reduces the amount of correctly folded receptor protein.

The activation of Gα_i2 proteins by deglycosylated hH₄R was investigated in [³⁵S]GTPγS saturation binding and steady-state GTPase assays. Even in the presence of histamine, the deglycosylated hH₄R in the tunicamycin-treated membranes activated Gα_i2 less efficiently than the glycosylated H₄R (increased K _D value of [³⁵S]GTPγS) (Schneider et al. 2009). Thus, proper glycosylation of hH₄R seems to be a prerequisite for efficient G protein coupling. By contrast, determination of the K _M value of GTP at the Gα_i2 subunit in steady-state GTPase assays only revealed a non-significant trend towards an increased K _M-value in the tunicamycin-treated membranes (Schneider et al. 2009).

In [³⁵S]GTPγS binding assays, deglycosylation reduced the constitutive activity of H₄R coexpressed with Gα_i2 and Gβ₁γ₂ from 70 to 40% (Schneider et al. 2009). Neither the coupling factor from [³⁵S]GTPγS binding assays nor the turnover number from steady-state GTPase assays changed significantly, when hH₄R was deglycosylated (Schneider et al. 2009). This suggests that deglycosylation of hH₄R reduces efficacy of Gα activation without affecting the total number of activated G proteins.

2.4.7 Reasons for the High Constitutive Activity of hH₄R

The inactive state of GPCRs is established by intramolecular interactions that conformationally restrain the receptor. Data obtained from the rhodopsin molecule have led to the assumption that the so-called ionic lock is highly important for the inactivation of GPCRs (Palczewski et al. 2000; Vogel et al. 2008). The ionic lock is a salt bridge between a highly conserved glutamate in position 6.30 of TM6 and the arginine of the DRY motif located on the bottom of TM3 (position 3.50). The importance of the ionic lock for the regulation of odorant GPCR activity has been shown recently (de March et al. 2015). However, some receptors do not form an ionic lock, despite the presence of the required amino acids. This has been reported, e.g. for the human β₂AR (Cherezov et al. 2007; Rasmussen et al. 2007; Rosenbaum et al. 2007) or the human A_2A adenosine receptor (Jaakola et al. 2008), both of which show considerable constitutive activity.

The hH₄R is the only histamine receptor with an alanine in position 6.30, which precludes ionic lock formation (Schneider et al. 2010a) and possibly explains the observed high G protein-independent activity of the hH₄R (Schneider et al. 2009). To test this hypothesis, the TM6 part of the potential ionic lock was reconstituted by introducing the A6.30E mutation, and the resulting mutant was analyzed in the Sf9 cell expression system. Immunoblots and [³H]histamine saturation binding indicated comparable expression levels of the mutant and the wild-type hH₄R. Unexpectedly, the pharmacological properties of hH₄R-A6.30E (co-expressed with Gα_i2 and Gβ₁γ₂) in radioligand binding, steady-state GTPase assay and [³⁵S]GTPγS binding assays were basically unaltered as compared to the wild-type hH₄R (Schneider et al. 2010a). The replacement of alanine 6.30 by glutamate resulted in a slight but non-significant reduction of coupling factor ([³⁵S]GTPγS binding), turnover number (steady-state GTPase assay) and constitutive activity ([³⁵S]GTPγS binding and steady-state GTPase assay). This indicates that the ionic lock interaction was either not fully reconstituted or not sufficient to stabilize the inactive conformation of hH₄R (Schneider et al. 2010a). An overview of the most important features of the hH₄R-A6.30E mutation in comparison to the wild-type hH₄R is provided in Table 4.

Molecular modeling studies revealed potential interactions that may stabilize the active conformation despite the presence of the reconstituted ionic lock. The hH₄R active state was modeled in complex with the C terminus of Gα_i2 by using the crystal structures of the turkey β₁AR (Warne et al. 2008) and the human adenosine A_2A receptor (Jaakola et al. 2008) as templates. This revealed an additional salt bridge between D5.69 at the N-terminus of the second cytoplasmic loop (CL3) and R6.31, which may stabilize an active receptor conformation (Schneider et al. 2010a). Since D5.69 is nearly unique among the GPCRs for biogenic amines, this salt bridge may be at least partly responsible for the high constitutive activity of hH₄R and should be analyzed in future studies.

Recently, the reasons for the high constitutive activity of hH₄R were further elucidated (Wifling et al. 2015a, b). These studies made use of the large pharmacological differences between human and rodent H₄R (Schnell et al. 2011; Strasser et al. 2013). For example, constitutive activity of mH₄R and rH₄R is strongly reduced as compared to hH₄R (Schnell et al. 2011) and the inverse hH₄R agonist JNJ7777120 exhibits partial agonism at mH₄R and rH₄R. Moreover, the potency of the agonist histamine is lower for the rodent orthologues as compared to hH₄R (Schnell et al. 2011). Mutational studies indicate that position 169 of the second extracellular loop is an important determinant of the distinct agonist binding properties of human and mouse H₄R (Lim et al. 2008). The F169 of the hH₄R is replaced by a V169 in the mH₄R. Thus, Wifling et al. (2015b) performed a detailed analysis of the “mouse-like” hH₄R-F169V mutant in the Sf9 cell system. In fact, hH₄R-F169V exhibited decreased constitutive activity as compared to wild-type hH₄R, resulting in an increased agonistic effect of histamine. Moreover, histamine binding affinity as well as the inverse agonistic effect of thioperamide was reduced (Wifling et al. 2015b). The second key amino acid identified by Wifling et al. (2015b) was S179, which is replaced by methionine in the mH₄R and by alanine in the rH₄R. The double mutants hH₄R-F169V+S179A and hH₄R-F169V+S179M showed an even stronger reduction of constitutive activity as compared to the hH₄R-F169V single mutant (Wifling et al. 2015b). These results suggest that the constitutively active state of hH₄R at least partly depends on hydrophobic interactions between the extracellular domains of TM 5, 6, and 7 and ECL2. A hydrogen bond between S179 and T323 additionally stabilizes the agonist-free active state of the hH₄R (Wifling et al. 2015b).

These mutations, however, did not completely eliminate the constitutive activity of hH₄R. A total loss of constitutive activity was only achieved by introducing the F168A mutation (Wifling et al. 2015a). This indicates that – despite the strong reduction of constitutive activity in the hH₄R-F169V mutation – the adjacent amino acid in the FF motif, F168, is the key residue responsible for the high constitutive activity of hH₄R (Wifling et al. 2015a). An FF motif in ECL2 is also present in other GPCRs, e.g. β₂AR, hH₃R and M₂R, suggesting a similar role of the ECL2 conformation on constitutive activity of these receptors.

2.4.8 The Role of the DRY Motif in G Protein Activation by the Human hH₄R

The arginine R3.50 of the DRY motif at the bottom of TM3 stabilizes the inactive receptor state by forming a salt bridge with the adjacent D/E3.49 residue (Nygaard et al. 2009). Therefore, we analyzed the effect of the hH₄R-R3.50A mutation on constitutive activity and ligand binding in membranes co-expressing hH₄R-R3.50A, Gα_i2 and Gβ₁γ₂. Surprisingly, the R3.50A exchange totally eliminated G protein coupling as indicated by the complete absence of receptor-regulated steady-state GTPase activity (Schneider et al. 2010a). Moreover, the hH₄R-R3.50A mutant adopted an inactive state with reduced affinity of the agonist histamine and increased affinity of the inverse agonist thioperamide (Schneider et al. 2010a). However, introduction of the R3.50A mutation reduced histamine affinity only by 50% and did not affect B _max. This suggests that the hH₄R-R3.50A mutant still adopts a “residual” G protein-independent high-affinity state.

To explain the total loss of G protein coupling of the hH₄R-R3.50A mutant, molecular modelling studies were performed using the active-state of the hH₄R in complex with the C-terminus of Gα_i2. This analysis revealed that R3.50 of the hH₄R may interact with the backbone oxygens of C352 and G353 in the Gα_i2 C-terminus (Schneider et al. 2010a). This supports the adoption of the Gα_i2 conformation, which is required for interaction with TM6 of the receptor. Thus, the R3.50A mutation hampers G protein recognition by hH₄R. Nevertheless, the hH₄R-R3.50A mutant is still able to form the salt bridge between D5.69 and R6.31, which stabilizes an active state. This could explain why hH₄R-R3.50A still exhibits relatively high histamine affinity (Schneider et al. 2010a). However, the effect of mutations in the E/DRY motif is not disrupting G protein coupling in all GPCRs. Rovati et al. (2007) described two phenotypes P1 and P2 that are produced by mutations of the E/D3.49- or the R3.50-residue. While in P1-type receptors high-affinity agonist binding and G protein coupling are retained after mutating position R3.50, P2-type receptors show a disrupted receptor-G protein interaction and reduced agonist binding affinity (Rovati et al. 2007). Accordingly, the hH₄R belongs to the group of P2-type GPCRs. An overview of the most important features of the hH₄R-R3.50A mutation in comparison to the wild-type hH₄R is provided in Table 4.

2.4.9 Pharmacological Characterization of hH₄R Ligands

As explained above, co-expression of the hH₄R and its cognate mammalian G proteins in Sf9 cells results in high constitutive activity (Schneider et al. 2009). This reduces the maximum available signal range, yielding a very low signal-to-noise ratio. Even in the presence of 100 mM of NaCl, the full agonist histamine produced only a signal intensity of ~30% (related to baseline) (Schneider and Seifert 2009). The expression of an hH₄R-Gα_i2 fusion protein did not improve the signal-to-noise ratio, but resulted in even higher constitutive activity and reduced relative intensity of histamine-induced signals (Schneider et al. 2009). Thus, the properties of the hH₄R/G protein co-expression system and the hH₄R-Gα_i2 fusion protein are rather unfavorable for the characterization of hH₄R ligands.

This prompted us to perform a closer investigation of the effects of regulators of G protein signaling (RGS proteins). A common feature of RGS proteins is the 120 amino acid RGS domain, which interacts with Gα subunits and increases their intrinsic GTPase activity (Willars 2006). RGS proteins are classified in eight subfamilies that differ from each other by protein size and the presence of additional functional domains. They regulate the activity of Gα_i/o- or Gα_q proteins, but no RGS protein-mediated activation of Gα_s has been reported to date. Due to their mechanism of action, RGS proteins should enhance signal intensity in steady-state GTPase assays. In fact, fusion of the α₂AR C-terminus to the RGS4 N-terminus significantly increased α₂AR-mediated stimulation of GTPase activity (Bahia et al. 2003).

For the experiments with the hH₄R, the two RGS proteins RGS4 and GAIP (Gα-interacting protein; also known as RGS19) were selected. RGS4 and GAIP both exhibit a simple protein structure without additional functional domains. Therefore, only activation of Gα_i GTPase activity is expected. Both RGS proteins were fused to the hH₄R via a His₆ linker (Fig. 5), very similar to the previously described hH₄R-Gα_i2 fusion protein approach (Fig. 3). The hH₄R-RGS fusion proteins were co-expressed with Gα_i2 and Gβ₁γ₂ in Sf9 cell membranes. The corresponding co-expression system was characterized by infecting Sf9 cells with baculoviruses encoding hH₄R, Gα_i2, Gβ₁γ₂ and RGS4 or GAIP.

Fig. 5

Schematic depiction of the hH₄R-RGS fusion protein. The C-terminus of the hH₄R is fused to the N-terminus of the RGS protein by a hexahistidine linker. This brings the RGS protein into close proximity to the heterotrimeric G protein. Adapted from Schneider and Seifert (2010c)

Both RGS4 and GAIP, irrespective of whether they were co-expressed or fused to hH₄R, increased the apparent K _M value of Gα_i2 in the presence of histamine in steady-state GTPase assays. This effect reached significance for the co-expressed GAIP and the hH₄R-RGS4 fusion protein (Schneider and Seifert 2009). By contrast, there was no effect of RGS proteins on the K _M value in the presence of the inverse agonist thioperamide (Schneider and Seifert 2009). This suggests that GPCR-mediated activation of the G protein is a prerequisite for the RGS protein effect.

Compared to the RGS4-free co-expression system (hH₄R + Gα_i2 + Gβ₁γ₂), both the quadruple expression system (hH₄R + Gα_i2 + Gβ₁γ₂ + RGS4) and the fusion protein system (hH₄R-RGS4 + Gα_i2 + Gβ₁γ₂) yielded a significantly increased relative steady-state GTPase signal of the inverse agonist thioperamide, while the histamine-induced signal remained unaffected (Schneider and Seifert 2009). The only major difference between co-expressed and hH₄R-attached RGS4 was an increased baseline steady-state GTPase activity in the hH₄R-RGS4 fusion protein system, but an unaltered baseline, when RGS4 was co-expressed (Schneider and Seifert 2009).

Co-expression of GAIP with hH₄R, Gα_i2 and Gβ₁γ₂ had no significant effect on baseline activity or thioperamide- and histamine-induced signals in steady-state GTPase assays. However, when GAIP was fused to hH₄R, the histamine-induced relative signal in steady-state GTPase assays was significantly increased by ~69% and the thioperamide-induced signal was enhanced by ~45%. The baseline activity of the GAIP-hH₄R fusion protein system, however, remained unaffected (Schneider and Seifert 2009). Thus, in contrast to hH₄R-RGS4, the hH₄R-GAIP fusion protein (co-expressed with Gα_i2 and Gβ₁γ₂) enhanced the absolute histamine-induced signal without changing baseline activity. Therefore, the relative stimulatory effect of histamine was increased (Schneider and Seifert 2009).

The different behavior of RGS4 and GAIP in the fusion proteins is surprising, because both RGS proteins have a similar RGS domain and no additional functionalities. Possibly, the differences are caused by distinct G protein affinities of these RGS proteins. According to the UniProtKB database entry P49795, GAIP binds to Gα_i proteins in the rank order Gα_i3 > Gα_i1 > Gα_o >> Gα_z/Gα_i2. Thus, among the Gα_i isoforms, Gα_i2 is the one with the lowest affinity to GAIP. This means that the effect of GAIP may only become visible, when the number of activated Gα_i2 subunits exceeds a certain threshold. While under basal conditions the number of activated Gα_i2 subunits is too low for a visible hH₄R-GAIP-mediated effect, stimulation by histamine increases the number of active Gα_i2 to a level, where the GAIP-mediated effect becomes visible. By contrast, RGS4 may exhibit a higher Gα_i2 affinity than GAIP and therefore show already an effect under basal conditions. This hypothesis, however, should be tested by a side-by-side comparison of the Gα_i2 protein affinity of RGS4 and GAIP.

Co-expression of the hH₄R-GAIP fusion protein with Gα_i2 and Gβ₁γ₂ produces a system with improved signal-to-noise ratio as compared to the standard co-expression system (hH₄R + Gα_i2 + Gβ₁γ₂). A comparison of hH₄R-GAIP and wild-type hH₄R (both co-expressed Gα_i2 and Gβ₁γ₂) in steady-state GTPase assays revealed comparable pharmacological properties. First, potency and efficacy of selected hH₄R standard ligands were unaltered. Second, similar to wild-type hH₄R, the hH₄R-GAIP fusion protein exhibited sodium chloride-insensitive constitutive activity (Schneider and Seifert 2009). Third, the hH₄R-GAIP fusion protein showed an unchanged G protein selectivity profile as compared to the unmodified hH₄R protein (Schneider et al. 2009; Schneider and Seifert 2009). The unaltered G protein profile is surprising, because GAIP shows distinct affinities to different Gα_i isoforms, which should theoretically influence the interaction between hH₄R-GAIP and the G protein. The results, however, indicate that the G-protein-specificity of the hH₄R-GAIP fusion protein is governed by the properties of the receptor rather than by the RGS protein part. In summary, the hH₄R-GAIP fusion protein (co-expressed with Gα_i2 and Gβ₁γ₂) can fully replace the standard co-expression system (hH₄R + Gα_i2 + Gβ₁γ₂) in steady-state GTPase assays and allows the functional characterization of new hH₄R ligands with higher sensitivity and signal-to-noise ratio. The hH₄R-GAIP fusion protein approach was successfully used to evaluate a new class of N ^G-acylated imidazolylpropyl-guanidine-derived hH₄R agonists (Ghorai et al. 2008; Igel et al. 2009b) or of cyanoguanidine-related hH₄R agonists (Igel et al. 2009a; Geyer et al. 2016). An overview of the most important features of the various H₄R/RGS fusion protein and co-expression approaches in comparison to the “standard” co-expression system (hH₄R + Gα_i2 + Gβ₁γ₂) is provided in Table 4.

2.4.10 Structure-Activity Relationships of hH₄R Inverse Agonists

The high constitutive activity of hH₄R significantly reduces the signal-to-noise ratio in steady-state GTPase assays and reduces the sensitivity of agonist assays. However, this feature becomes an advantage, when inverse agonists are characterized. The hH₄R may maintain its constitutive activity under physiological conditions, because it is resistant to high sodium concentrations. As hypothesized by Schneider et al. (2009), on the one side, inverse agonists could be therapeutically advantageous in case of pathophysiologically increased constitutive H₄R activity, because they may exert a stronger anti-pruritic effect than neutral antagonists. On the other side, the re-folding of misfolded H₄R protein observed with the inverse agonist thioperamide (Schneider et al. 2009) (Sect. 2.4.4) may be a general effect of inverse H₄R agonists. Thus, inverse agonist-mediated upregulation of intact H₄R protein may result in rebound effects after drug discontinuation (Schneider et al. 2009). Although these hypotheses were not proven yet under physiological conditions, they illustrate the potential importance of characterizing inverse H₄R agonism during drug development. Therefore, structure-activity relationships for hH₄R inverse agonism should be established.

A series of 25 previously described (Venable et al. 2005) H₄R ligands (indoles, benzimidazoles, and thienopyrroles; Fig. 6) structurally derived from the prototypical H₄R antagonist JNJ7777120 (Thurmond et al. 2004) was characterized in [³H]histamine binding assays and steady-state GTPase assays using membranes expressing hH₄R + Gα_i2 + Gβ₁γ₂. The steady-state GTPase assays were performed in the absence of sodium chloride to obtain maximum constitutive activity.

Fig. 6

Scaffold structure of three classes of H₄R antagonists/inverse agonists. The numbers in brackets indicate the number of compounds tested

The steady-state GTPase assay data reveal that most of the compounds were inverse agonists with a lower efficacy than thioperamide. Only three of the 25 compounds (~12%) were neutral antagonists (Schneider et al. 2010b). This confirms a previous analysis of literature data on 380 antagonists binding to 73 GPCRs. Only 15% of these compounds were neutral antagonists (Kenakin 2004). Thus, neutral antagonism seems to be a rare phenomenon.

In general, the pK _b values from steady-state GTPase assays in the presence of histamine fit very well to the pK _i values from [³H]histamine binding. In a subset of compounds, the pEC₅₀ values determined in the absence of histamine were significantly lower than the pK _i and/or pK _b values. Such discrepancies have been reported before for inverse agonists, e.g. at the hH₄R (Smits et al. 2008) or the β₂AR (Chidiac et al. 1994). Maybe, this subset of hH₄R antagonists discriminates between the agonist-free constitutively active receptor and the histamine-activated receptor state (Schneider et al. 2010b). These observations confirm the insufficiency of the two-state model of receptor activation and point to the existence of ligand-specific receptor states.

The potential binding mode of inverse hH₄R agonists of the indole series was analyzed by molecular dynamics simulations with the completely unsubstituted indole compounds (R4-7 = H; Fig. 6). The positively charged piperazine amino group interacts electrostatically with the highly conserved Asp^3.32. Moreover, both the carbonyl moiety and the indole NH of the ligand establish an interaction with the side chain of the uncharged Glu^5.46. The indole moiety of the ligand shows a hydrophobic interaction with the indole part of Trp^6.48 (Schneider et al. 2010b). Trp^6.48 is a key player in the so-called rotamer toggle switch mechanism of receptor activation, which had been previously postulated for the β₂AR (Shi et al. 2002). The stabilization of Trp^6.48 in its vertical conformation by the indole-derived ligand is a typical feature of the inactive receptor conformation and may explain the inverse agonism of such compounds. The benzimidazole-related structures bind in a similar way, but, in contrast to the indole-derived compounds, they form two tautomers with distinct binding modes (Schneider et al. 2010b).

Replacement of the R5/R7 hydrogen of the indole derivatives by the more space-filling chlorine increases H₄R binding affinity. Molecular dynamics simulations suggest that two small binding pockets in the H₄R protein may be filled by these chlorine residues, which increases the ligand-receptor contact area (Schneider et al. 2010b). Substitution of R5 by -OCH₃ reduces binding affinity, suggesting that larger substituents may be unfavorable. However, there is no significant correlation between molar volume and affinity of a series of indole compounds, suggesting that the volume of R5 may not be the only descriptor that influences binding affinity (Schneider et al. 2010b). By contrast, the size of R5 correlates excellently with the inverse agonistic efficacy of a subset of eight indole-derived compounds with varying R5 substituents. A calculation of the descriptors logP, molar refractivity, molar volume, polarizability, refraction index and polar surface area revealed that inverse agonistic efficacy solely depended on molar volume, but not on the other factors. The inverse agonistic efficacy of these compounds was inversely correlated to the molar volume of the substituent R5 (Schneider et al. 2010b).

In summary, despite the limited number of compounds and substitution patterns available, in this study the first structure-activity relationships for inverse H₄R agonism were identified. It was, however, not possible to predict all changes in binding mode and receptor conformation that result from small structural alterations of the ligand. Moreover, a general model that applies to structurally distinct classes of hH₄R inverse agonists could not be established yet. In the future, the hH₄R should be co-crystallized with various inverse agonists to elucidate the exact binding mode of these compounds. Although this would be a very ambitious project, the numerous crystallized ligand-receptor complexes published in the recent years (Cherezov et al. 2007; Rasmussen et al. 2007, 2011; Shonberg et al. 2015) demonstrate that this is not impossible.

3 Summary and Outlook

In this chapter, the results from the characterization of all four histamine receptor subtypes in the Sf9 insect cell system were summarized. On the one hand, it might be argued that insect cells do not represent physiological conditions as well as primary cells. On the other hand, it is difficult to isolate primary cells in sufficiently high numbers. Moreover, a side-by-side comparison of receptor isoforms or species orthologues in a defined environment is virtually impossible in primary cells. Since cells from different tissues have to be used, cell type-specific properties like crosstalk with other receptors or special features of the signaling pathways can lead to heterogeneous results, even for the same receptor isoform. Also, for some receptors like H₃R, no suitable primary cell system is available (Seifert et al. 2013).

Thus, for a comparison of the intrinsic properties of GPCR isoforms, e.g. G protein affinity/selectivity or constitutive activity, Sf9 cells represent a superior option. As explained in this chapter, Sf9 cells do not contain background GPCR activity and do not produce endogenous agonists activating mammalian GPCRs. Moreover, Sf9 cells allow the co-expression of defined mammalian Gα_s or Gα_i protein subunits on a “clean” signaling background. This was demonstrated by the analysis of the hH₂R interaction with long and short Gα_s splice variants or by in-depth studies of hH₃R/hH₄R Gα_i isoform specificity and ion sensitivity. Table 5 shows numerous aspects of histamine receptor pharmacology addressed by using the Sf9 insect cell expression system.

Table 5 Various aspects of histamine receptor pharmacology and medicinal chemistry investigated in the Sf9 insect cell expression system

The ligand binding studies and the G protein activation assays discussed in this chapter were all performed with radiolabeled reagents. Radioactivity-based assays, however, are increasingly hampered by legal overregulation and growing waste disposal costs. In this situation, fluorescence-based GPCR ligand binding and G protein activation assays could represent interesting alternatives. Unfortunately, many histamine receptor ligands are rather small molecules and easily lose binding affinity when coupled to a bulky fluorophore. Nevertheless, some progress has been made during the past years. For example, a cyanine dye-labeled aminopotentidine derivative exhibited nanomolar hH₂R potency (Xie et al. 2006c). Moreover, fluorescent pyrylium- or cyanine-labeled dimeric carbamoylguanidines were synthesized, but these compounds failed in binding assays due to intracellular accumulation and the resulting high fluorescence background (Kagermeier et al. 2015). A high-affinity fluorescent H₁R antagonist was obtained by labeling mepyramine with a BODIPY (4,4-difluoro-4-bora-3a,4a-diaza-s-indacene)-derived dye (Rose et al. 2012). Fluorescent hH₃R-selective ligands were developed by using the chalcone partial structure (Tomasch et al. 2012). Moreover, a compound named “Bodilisant,” which has been reported recently, is a BODIPY-labeled non-imidazole ligand with nanomolar hH₃R affinity (Tomasch et al. 2013). Some progress has also been made in the field of fluorescence-based G protein activation assays. For example, a europium-labeled non-hydrolysable GDP derivative can replace [³⁵S]GTPγS in GTPγS binding assays (Koval et al. 2010). This enables a time-resolved fluorescence-based assay that is, e.g., suited for the functional characterization of hH₃R ligands (Singh et al. 2012).

The functional assays described in this chapter focused on the determination of GPCR-mediated G protein activation (steady-state GTPase and [³⁵S]GTPγS binding assays). However, GPCRs can additionally activate G protein-independent signaling mechanisms, most importantly through β-arrestin recruitment (Lefkowitz and Shenoy 2005; Shukla et al. 2014). The hH₄R ligand JNJ-7777120, which acts as an inverse hH₄R agonist in G protein activation assays (Schneider et al. 2009), unexpectedly turned out to be an agonist with regard to hH₄R-mediated β-arrestin recruitment (Rosethorne and Charlton 2011). This phenomenon is also known as “biased signaling” or “functional selectivity” and has important implications for drug development (Seifert et al. 2011; Nijmeijer et al. 2013). In future studies, biased signaling of hH₁R, hH₂R, or hH₃R and functional selectivity of the corresponding ligands should be investigated in more detail.

However, the most important, but also most ambitious, goal in future studies would be the crystallization of all four histamine receptor subtypes. To date, only the crystal structure of the hH₁R has been resolved (Shimamura et al. 2011). The crystal structures of the histamine receptors are required to answer several still unresolved questions. For example, exact knowledge of the hH₄R conformation could help to explain, why this receptor shows such a high constitutive activity (Schneider et al. 2009; Wifling et al. 2015a). Moreover, a crystal structure of the hH₃R may provide important information about the hH₃R-G protein interaction interface and possibly answer the question, why the hH₃R discriminates between Gα_i3 and other Gα_i/o isoforms (Schnell and Seifert 2010). Furthermore, an hH₃R crystal may lead to the identification of the anion binding sites responsible for the monovalent anion-mediated reduction of constitutive hH₃R activity (Schnell and Seifert 2010). Finally, the knowledge of H_xR crystal structures could lead to the development of compounds that alter H_xR function as allosteric modulators. The concept of GPCR modulation by allosteric ligands is well established, and such ligands have been identified, e.g. for dopamine, muscarinic, adenosine, or chemokine receptors (Christopoulos 2014). By contrast, to the best of our knowledge, to date nothing is known about allosteric modulation of histamine receptors.

As a prerequisite for the preparation of H_xR crystals, high amounts of receptor protein have to be expressed, e.g. in Sf9 cells. After purification and solubilization, the physical properties of the receptors can be investigated, e.g. with fluorescence-based methods. Such studies have been previously performed with the β₂AR (Gether et al. 1995; Kobilka 1995; Neumann et al. 2002) and were important steps towards the final goal of receptor crystallization (Cherezov et al. 2007; Rasmussen et al. 2007, 2011).