Luciferase Reporter Gene Assay on Human, Murine and Rat Histamine H4 Receptor Orthologs: Correlations and Discrepancies between Distal and Proximal Readouts

The investigation of the (patho)physiological role of the histamine H4 receptor (H4R) and its validation as a possible drug target in translational animal models are compromised by distinct species-dependent discrepancies regarding potencies and receptor subtype selectivities of the pharmacological tools. Such differences were extremely pronounced in case of proximal readouts, e. g. [32P]GTPase or [35S]GTPγS binding assays. To improve the predictability of in vitro investigations, the aim of this study was to establish a reporter gene assay for human, murine and rat H4Rs, using bioluminescence as a more distal readout. For this purpose a cAMP responsive element (CRE) controlled luciferase reporter gene assay was established in HEK293T cells, stably expressing the human (h), the mouse (m) or the rat (r) H4R. The potencies and efficacies of 23 selected ligands (agonists, inverse agonists and antagonists) were determined and compared with the results obtained from proximal readouts. The potencies of the examined ligands at the human H4R were consistent with reported data from [32P]GTPase or [35S]GTPγS binding assays, despite a tendency toward increased intrinsic efficacies of partial agonists. The differences in potencies of individual agonists at the three H4R orthologs were generally less pronounced compared to more proximal readouts. In conclusion, the established reporter gene assay is highly sensitive and reliable. Regarding discrepancies compared to data from functional assays such as [32P]GTPase and [35S]GTPγS binding, the readout may reflect multifactorial causes downstream from G-protein activation, e.g. activation/amplification of or cross-talk between different signaling pathways.

However, there are also controversial reports. The administration of the H 4 R agonist 5(4)-methylhistamine was benefical in a murine asthma model [12], and JNJ 7777120 increased the ocular histamine concentration in a rat conjunctivitis model [17] (for a recent review cf. Neumann et al. [19]). Furthermore, the overall amino acid identities of H 4 R species orthologs are remarkably low (human versus mouse and rat: ,70%) compared to other histamine receptor subtypes (H 1 R, H 2 R and H 3 R) [20]. Although relatively small differences in the sequence of histamine receptor species orthologs can result in different potencies and efficacies of individual ligands, the discrepancies are exceptionally high in case of the H 4 R [21]. In various in vitro assay systems the recombinantly expressed mouse and rat H 4 R revealed substantial species-dependent differences compared to the human receptor concerning affinity, potency and quality of action of pharmacological tools, compromising the predictive value with respect to translational animal models [20][21][22][23]. For example, in comparison to the human H 4 R, UR-PI294 (N 1 -[3-(1H-imidazol-4-yl)propyl]-N 2 -propionylguanidine) and UR-PI376 (2-cyano-1-[4-(1H-imidazol-4-yl)butyl]-3-[(2-phenylthio)ethyl]guanidine) [24,25] displayed considerably lower potencies and efficacies (UR-PI376) in the [ 32 P]GTPase and [ 35 S]GTPcS binding assays on membrane preparations of Sf9 insect cells expressing the mouse or rat H 4 R [23]. Most strikingly, JNJ 7777120 exhibited stimulatory effects at the mouse and rat H 4 R in functional assays on Sf9 cell membranes [23]. Moreover, the use of JNJ 7777120 as standard antagonist in animal models was questioned due to stimulation of G-protein independent b-arrestin recruitment [26]. Biased signaling of the hH 4 R has also been shown for other H 4 R ligands [27].
The aforementioned controversial findings underline the necessity to evaluate pharmacological tools at the H 4 R species orthologs of interest using different assay systems. For this purpose, a cAMP response element (CRE) controlled luciferase reporter gene assay in HEK293T cells, stably expressing the human, the mouse or the rat H 4 R, was established. The H 4 R is Ga i/o -coupled and reduces forskolin stimulated cyclic adenosine monophosphate (cAMP) formation after agonist binding [2]. The optimal concentration of forskolin used for pre-stimulation depends on the cell type [28] and should correspond to the EC 50 of forskolin in the assay system [29]. Therefore, the potency of forskolin was determined, and the effect of the phosphodiesterase (PDE) inhibitor isobutylmethylxanthine (IBMX) was evaluated to optimize the sensitivity of the procedure. Due to the delayed onset of gene expression, incubation periods of four to six hours are required [30], increasing the risk of agonist mediated receptor desensitization, which can lead to a decrease in agonist potencies [30]. Therefore, the time course of the luciferase expression was determined to find the minimum incubation period required for appropriate signal strength. For validation, potencies and efficacies of 23 selected H 4 R ligands, comprising agonists, inverse agonists and antagonists, were determined ( Figure 1).

Ethics Statement
Human embryonal kidney (HEK293T) cells were purchased from the German Collection of Microorganism and Cell Cultures (DSMZ, Braunschweig, Germany).

Luciferase Reporter Gene Assay
Approximately 2 ? 10 5 transfected cells, suspended in DMEM supplemented with 10% (v/v) FCS, were seeded per well into flatbottomed 96-well plates (Greiner, Frickenhausen, Germany). The cells were allowed to attach for 17 h at 37uC, 5% CO 2 in a watersaturated atmosphere. A stock solution (10 mM) of forskolin (Sigma) in DMSO was used to prepare feed solutions in DMEM containing 10% (v/v) FCS (final DMSO concentration in the assay was #1%). For experiments in the presence of a PDE inhibitor, the feed solution of forskolin contained 500 mM of IBMX (Sigma).
After addition of forskolin (0.4 mM for the cells expressing the human H 4 R and 1 mM for the rat and mouse H 4 R expressing cells) alone (to determine forskolin potency) or in combination with histaminergic ligands, the cells were incubated for 5 h. In antagonist mode, the forskolin solution was supplemented with 0.10, 0.15 or 1.00 mM of histamine as the agonist for the human, mouse and rat H 4 R expressing cells, respectively. Thereafter, the medium was discarded, the cells were washed once with 100 mL of phosphate buffered saline (PBS, pH 7.4) (KCl 2.7 mM; KH 2 PO 4 1.5 mM; NaCl 137 mM; Na 2 HPO 4 5.6 mM; NaH 2 PO 4 1.1 mM in Millipore water; all chemicals were from Merck, Darmstadt,    [43], hH 4 R-GAIP+G ia2 + b 1c2 , rH 4 R or mH 4 R+G ia2 + b 1c2 + GAIP d [23], hH 4 c (a value of ST-1012 referred to thioperamide = 21.0, [39]); e calcium mobilization assay in 293-EBNA cells transiently co-expressing the hH 4 R, mH 4 R or rH 4 R with G qi5 [20]; f,g,h calcium mobilization assay in HEK293 cells stably co-expressing the hH 4 R, mH 4 R or rH 4 R with G qi5 f [46], g [44], h [45]; i,j,k,l CRE-b-galactosidase reporter gene assay in SK-N-MC cells stably co-expressing: the hH 4 R [40][41][42] or the mH 4 R k [57] with the CRE-b-galactosidase reporter gene; m CRE-luciferase reporter gene assay in HEK293T cells, transiently co-expressing the hH 4 R with the CRE-controlled luciferase reporter gene [27]; l SRE-luciferase reporter gene assay in HEK293 cells, co-expressing the human, mouse or rat H 4 R+SRE-luciferase+Ga qi chimeric G-protein [57].  [31]. Light emission was induced by the injection of 80 mL of the luciferase assay buffer into each well. Luminescence, expressed as RLUs (relative light units), was measured for 10 s. The basal luciferase activity was subtracted from each signal. EC 50 and IC 50 values were analyzed by nonlinear regression and best fitted to sigmoidal concentrationresponse curves with GraphPad Prism 5.04 (Graph Pad, San Diego (CA), USA). IC 50 values were converted to K B values using the Cheng-Prussoff equation [32]. The intrinsic activity of ligands   Cell culture and generation of high-titer recombinant baculovirus stocks as well as the co-infection of Sf9 cells with high-titer baculovirus stocks encoding Ga i2, Gb 1 c 2 and the respective H 4 R were performed as described recently [34,35]. Membrane preparations were performed according to Gether et al. (1995) [36] in the presence of 0.2 mM phenylmethylsulfonyl fluoride, 1 mM ethylenediaminetetraacetic acid (EDTA), 10 mg/mL leu-peptin and 10 mg/mL benzamidine as protease inhibitors. Prepared membranes were resuspended in binding buffer (75 mM Tris/HCl, 12.5 mM MgCl 2 ,1 mM EDTA, pH 7.4) and stored at 280uC in 0.5 or 1.0 mL aliquots.
Membranes were thawed, centrifuged for 10 min at 4uC and 13,000 g and carefully resuspended in binding buffer. Experiments were performed in 96-well plates in a total volume of 100 mL per well. Each tube contained 8-12 mg of protein, 1 mM GDP, 100 mM NaCl, 0.05% (w/v) bovine serum albumine (BSA), 20 nCi of [ 35 S]GTPcS ($0.2 nM) and ligand at various concentrations. Neutral antagonists were incubated in the presence of histamine at concentrations corresponding to the 10-fold of the EC 50 value at the respective receptor. Nonspecific binding was determined in the presence of 10 mM unlabeled GTPcS. After incubation under shaking at 200 rpm at room temperature for 2 h, bound [ 35 S]GTPcS was separated from free [ 35 S]GTPcS by filtration through glass microfibre filters using a 96-well Brandel harvester (Brandel Inc., Unterföhring, Germany). The filters were washed three to four times with cold binding buffer (4uC), dried over night and impregnated with meltable scintillation wax prior to counting with a Micro Beta2 1450 scintillation counter (Perkin Elmer, Rodgau, Germany).
Ligands were tested in triplicate. The maximal response to histamine was set to 100% and all other ligands were referenced to histamine.

Optimization of the Assay Conditions
In order to detect a Ga i -mediated inhibitory effect on the adenylyl cyclase (AC) activity, the reporter gene assay was performed in the presence of the AC stimulator forskolin. The time course of the luciferase expression upon stimulation with 10 mM forskolin is shown in Figure 2A. After a latency period of 0.5-1 h, the enzyme activity steeply increased, and a maximum was reached after 8 h. An incubation period of 5 h was sufficient to obtain 76-94% of the maximum expression. To optimize assay performance, the pEC 50 value of forskolin in the respective cAMP reporter gene assay system [29] was determined ( Figure 2). As the concentration-response curve shows an optimum ( Figure 2B), only the ascending part of the curve was considered up to a forskolin concentration of 10 mM ( Figure 2C). Interestingly, the potency of forskolin was significantly different: pEC 50 values were 6.4160.05 and 5.9560.04 in the hH 4 R and mH 4 R co-transfected cells, respectively, and 5.5060.11 in the HEK293T-CRE-Luc cells ( Figure 2C). Forskolin concentration-dependently increased the luciferase expression in HEK293T-SF-rH 4 R-His 6 -CRE-Luc cells, which was inhibited by histamine (1) ( Figure 3A) with pEC 50 values of 6.8160.11, 6.5360.04, 6.2960.07 and 5.9160.04 ( Figure 3B) at forskolin concentrations of 0.5, 1.0, 2.5 and 5 mM, respectively. Therefore, a concentration of 0.4 mM of forskolin was used for pre-stimulating the hH 4 R expressing cells, whereas 1 mM of forskolin was considered optimal for AC stimulation in mH 4 R and rH 4 R expressing cells. With respect to comparability of concentration-response curves of H 4 R ligands at H 4 R orthologs, the difference between maximum forskolin stimulation in the absence and the presence of the reference agonist histamine (100 mM) was set to 100% ( Figure 3B).
In the presence of the PDE inhibitor IBMX (50 mM) the concentration-response curve of forskolin on HEK293T-SF-hH 4 R-His 6 -CRE-Luc-cells was shifted to the left, resulting in an pEC 50 value of 6.8660.06 (N = 3). Additionally, IBMX increased the receptor-independent luciferase activity by about a factor of four (data not shown). To investigate the effect of IBMX on the concentration-response curve of the full H 4 R agonist histamine (1) and the H 4 R inverse agonist thioperamide (20) HEK293T-hH 4 R-His 6 -CRE-Luc cells were pre-stimulated with forskolin (0.5 mM) alone or in combination with IBMX (50 mM) (cf. Figure 4). The maximum responses to histamine (1) and thioperamide (20), and thus the range of the signals, were reduced in the presence of IBMX. Therefore, further experiments were performed in the absence of IBMX.
hH 4 R agonists (compounds 1-17). The endogenous agonist histamine (1) inhibited forskolin stimulated luciferase activity with pEC 50 values of 7.77, 7.06 and 6.53 in the hH 4 R, mH 4 R and rH 4 R expressing reporter cells, respectively ( Table 1). The methylsubstituted analogs of histamine (2)(3)(4)(5) acted, with the exception of 3, as full agonists at the three H 4 R orthologs. Compared to the hH 4 R, a trend towards decreased potency was detected at the rodent receptors for compounds 1-5 ( Figure 5A, B, C). Among the enantiomers 2 and 3, (R)-a-methylhistamine (2) was the eutomer at all species orthologs. Compared to immepip (6), the pyridine analog immethridine (7) showed significantly reduced potency and intrinsic activity at the hH 4 R. By contrast, immethridine (7) exhibited almost full agonist activity at both, the mouse and rat H 4 R, with similar moderate potency compared to the hH 4 R. Imetit (8) exhibited almost the same potency and efficacy at the three H 4 R orthologs. In contrast, clobenpropit (9) and iodophenpropit (10), which can be considered as analogs of imetit (8) with an increased distance between the basic moieties and a large lipophilic group in the side chain, displayed a clear decrease in potency and maximal response at the mouse and rat H 4 R compared to the hH 4 R. Clobenpropit (9) was a potent full agonist at the hH 4 R and only a moderate partial agonist at the mouse and rat H 4 R, whereas iodophenpropit (10) acted as a partial agonist at the hH 4 R and a neutral antagonist at both, the mouse and the rat H 4 R. Proxyfan (11) partially activated the three H 4 R orthologs with significantly lower potencies on the rodent receptors. Whereas H 4 R-independent effects of 11 were negligible at concentrations .10 mM, the structural analog ciproxifan (12) induced a strong increase (by up to 250%) in luciferase activity at concentrations from 1 to 100 mM in HEK293-CRE-Luc cells devoid of H 4 R expression ( Figure 6). Therefore, functional activities of 12 on H 4 R orthologs were not determined in the luciferase assay. The non-selective acylguanidine-type H 3/4 R agonist UR-PI294 (13) fully activated the human, mouse and rat H 4 R ( Figure 5A, B, C), being the most potent agonist at all three H 4 R orthologs (Table 1). In contrast, the selective cyanoguanidine-type H 4 R agonist UR-PI376 (14) acted as a potent full hH 4 R agonist, exhibited only partial agonistic activity at the mH 4 R and was devoid of agonism at the rH 4 R ( Table 1). VUF 8430 (15) had about the same potency at both, the mH 4 R and the hH 4 R, whereas the potency at the rH 4 R was distinctly lower. At all three H 4 R species orthologs, VUF8430 (15) was almost as efficacious as histamine (a = 0.96-0.98). The aminopyrimidine-type compound ST-1006 (16) exhibited pronounced differences in the quality of action at the H 4 R orthologs with nearly full agonism at the hH 4 R, partial agonism at the mH 4 R and inverse agonism at the rH 4 R. The antipsychotic drug clozapine (17) exhibited only moderate agonistic potency at the hH 4 R. However, with an a value of 1.30, clozapine was even more efficacious than histamine (1). Furthermore, clozapine (17) fully activated both, the mouse and the rat H 4 R, though with low pEC 50 values (Table 1). hH 4 R antagonists and inverse agonists (18)(19)(20)(21)(22)(23). Interestingly, VUF 5681 (18), with a spacer extended by two carbon atoms compared to the H 4 R agonist immepip (6), displayed no agonistic activity at the hH 4 R and only partial agonism at the mH 4 R. In the antagonist mode at the hH 4 R, VUF 5681 (18) inhibited the histamine-induced decrease in luciferase activity with a pK B value of 6.1660.20. JNJ 7777120 (19) behaved as neutral antagonist at the human and mouse H 4 R in the luciferase reporter gene assay with comparable pK B values of 7.8160.19 and 7.5860.13, respectively ( Figure 5A, B, D). In contrast, at the rH 4 R JNJ 7777120 (19) acted as a partial agonist (a = 0.4960.05) with a pEC 50 value of 8.2160.10 ( Figure 5C). By analogy with ciproxifan, but much less pronounced, JNJ 7777120 (19) and thioperamide (20) produced receptor-independent increases in luciferase activity at concentrations $10 mM in control experiments using cells devoid of H 4 R expression ( Figure 6). The corresponding values were therefore omitted in the construction of concentration-response curves of 19 and 20, when studied in the antagonist mode (shown for JNJ 7777120 (19) in Figure 5D). Thioperamide (20) acted as an inverse agonist, achieving comparable pEC 50 values at the human and mouse H 4 R ( Figure 5A, B, Table 1), and revealed moderate antagonistic acitivity at the rH 4 R with a pK B value of 6.8960.14. The aminopyrimidine ST-1012 (21) acted as an inverse agonist at the hH 4 R, but revealed partial agonistic activity at the mouse and the rat H 4 R. The conformationally constrained aminopyrimidines A 943931 (22) and A 987306 (23) were inverse agonists at the hH 4 R and neutral antagonists at the rH 4 R.

Assay Optimization
The pEC 50 value of forskolin varied among the different transfectants probably due to different expression levels of the CRE-controlled luciferase. The concentration-response curve revealed a decline at forskolin concentrations higher than 10 mM. This decline of the forskolin effect became already obvious at concentrations .3.2 mM in the presence of 50 mM of the PDE inhibitor IBMX (data not shown), as already described for a CRE-directed luciferase reporter gene assay in Chinese hamster ovary cells (CHO) [37]. By analogy with a report by Kemp et al. [38] an activation of the inducable cAMP early repressor (ICER) may counteract the luciferase expression in HEK293T cells. Ga i -protein mediated inhibition of the cAMP synthesis as well as the signal-to-noise ratio was lowered by increasing concentrations of forskolin and IBMX. This was reflected by smaller relative effects and potencies of histamine (1) in the presence of increasing forskolin concentrations ( Figure 3) and 50 mM of IBMX ( Figure 4). Thus, high forskolin concentrations should be avoided and the altered potency of forskolin, when used in combination with IBMX, must be considered in this assay.
The co-expression of a CRE-controlled luciferase reporter gene with the human, mouse and rat H 4 R, respectively, in HEK293T cells enabled the functional analysis of H 4 R ligands. A set of 23 imidazole and non-imidazole ligands comprising agonists, inverse agonists and antagonists was investigated for ability to effect forskolin stimulated luciferase activity. The obtained pEC 50 values or pK B values were compared with ligand activities from different functional assay systems reported in literature.

Off-target Effects
The luciferase stimulation becoming obvious at concentrations .1 mM of JNJ7777120 (19) and thioperamide (20) in cells expressing the H 4 R orthologs (cf. dashed lines in the concentration-response curves of 19 and 20 in Figure 5A-C) suggest inverse agonism. However, the investigation of selected compounds on HEK293T-CRE-Luc cells lacking the H 4 R (cf. Figure 6) revealed H 4 R-independent increase in luciferase activity. This effect was most prominent in case of ciproxifan (12), but also pronounced for 19 and 20. Therefore, off-target effects should be taken into account to avoid misinterpretation of biological responses to such compounds at concentrations $10 mM.

Activities at the Human H 4 Receptor
Except for ST-1006 (16) [39], all determined H 4 R ligand activities at the hH 4 R were in agreement with results reported in literature [20,23,[39][40][41][42]. However, a tendency toward elevated intrinsic activities was observed. Contrary to partial agonistic activity of immepip (6) and clobenpropit (9) in the [ 35 S]GTPcS binding assay on membrane preparations of H 4 R expressing Sf9 cells (a = 0.81and 0.45, respectively) ( Table 2), full agonism at the hH 4 R was determined in the luciferase assay. Iodophenpropit (10), described as a neutral antagonist [40], exerted strong partial agonistic activity at the hH 4 R in the present study. Partial agonistic activity was also determined for iodophenpropit (10) in a Ca 2+ mobilization assay in HEK293 cells, co-transfected with the hH 4 R and the chimeric G-protein G qi5 [5]. ST-1006 (16) had low intrinsic activity in the [ 32 P]GTPase and [ 35 S]GTPcS binding assay at the hH 4 R [39], but was an almost full agonist in the luciferase assay. The increased intrinsic activity was accompanied with a decrease in potency of about one order of magnitude. In case of clozapine (17), the maximal agonistic response surpassed that of histamine by 30%. In control experiments on HEK293T-CRE-Luc cells devoid of the H 4 R, clozapine (17) at concentrations as high as 100 mM caused an increase in CRE-activity by up to 17% (data not shown). The effect of clozapine on hH 4 R expressing cells was antagonized by JNJ 7777120 in a concentrationdependent manner, indicating that the (super)agonistic effect was receptor mediated (Figure 7). Using histamine or clozapine as H 4 R agonists revealed approximately the same pA 2 value for JNJ 7777120 (pA 2 values: 8.39 and 8.17). However, compared to the concentration response curve of histamine in the presence of JNJ 7777120 ( Figure 7A), the extent of rightward shift was smaller in case of clozapine ( Figure 7B), resulting in different slopes (0.83 compared to 0.45) of the corresponding Schild plots ( Figure 7C). This may be taken as a hint that histamine and clozapine activate the H 4 R not exactly in the same way. However, due to the pleiotropic character of clozapine (17), effects mediated by targets other than the H 4 R must be taken into account. Most probably, increased intrinsic activities in the luciferase assay compared to more proximal readouts are caused by amplifications in signaling downstream from G-protein activation [30,37]. For instance, in functional assays on Sf9 cell membranes, ST-1006 (16) [39] and clozapine (17) [43] showed only partial agonism ( Table 2).
The pK B values of neutral antagonists, such as iodophenpropit (10) at the mouse and rat H 4 R as well as thioperamide (20) and UR-PI376 (14) at the rH 4 R were comparable to those determined in the [ 35 S]GTPcS binding assay (Table 2). Mouse and rat H 4 Rmediated inhibition of forskolin-stimulated luciferase activity in HEK293T-CRE-Luc cells resulted in higher potencies compared to functional assays using Ga-protein activation as readout. This suggests that signal amplification or concomitant activation of different signaling pathways potentiates the inhibition of the luciferase activity. For example, the cAMP pathway may be modulated by a cross-talk with Ca 2+ signaling elicited by activation of phospholipase C (PLC) [47]. Ca 2+ is an inhibitor of (forskolin) stimulated and Ca 2+ sensitive adenylate cyclases type V/VI [48][49][50], which are endogenously expressed in HEK293T cells [51] and interact with the Ga i protein [52]. Furthermore, the relevance of this crosstalk with regard to the cAMP signaling pathway of Gprotein coupled receptors (GPCRs) was demonstrated by the inhibitory effect of the activated Ga q coupled histamine H 1 R on the cAMP level in U373 MG cells [53] and, more importantly, by a crosstalk between the Ga i coupled M 2 mACh receptor and the Ga q coupled M 3 mACh receptor. In the latter case the inhibition of forskolin-stimulated cyclic AMP accumulation was facilitated at low agonist concentrations [54]. Further studies on the influence of Ca 2+ are needed to clarify, whether only the rodent H 4 Rs are concerned, since agonist potencies at the hH 4 R were consistent with data from the [ 32 P]GTPase and [ 35 S]GTPcS binding assay ( Table 2). Very recently, investigations on human esosinophils revealed a lower Ca 2+ response to stimulation by histamine (1) and UR-PI376 (14) compared to the chemokine eotaxin via the CCR3 receptor [55]. This may be interpreted as a hint to minor contribution of Ca 2+ signaling to the overall H 4 R mediated response, at least in native human cells. The presence of a range of alternative signaling pathways for the H 4 R in living cells was underlined recently by the Ga independent ß-arrestin recruitment of several H 4 R ligands [26,27].
The results for the standard antagonist JNJ 7777120 (19) at the mouse and rat H 4 R compared with data reported for other functional assays revealed discrepancies, too. In the luciferase assay JNJ 7777120 (19) acted as a neutral antagonist at the mH 4 R, but as a potent partial agonist at the rH 4 R. Antagonistic activity at both receptors was found in a CRE-driven b-galactosidase assay in SK-N-MC cells [18] and in a Ca 2+ assay in HEK293 cells [46], whereas partial agonistic activity was determined at the mouse and rat receptor in the [ 32 P]GTPase [23] and [ 35 S]GTPcS binding assay ( Table 2). The pK B value at the mH 4 R in the luciferase assay is consistent with the pK B value in the Ca 2+ assay [46], whereas the agonistic potency at the rH 4 R is about two orders of magnitude higher compared to the [ 32 P]GTPase assay [23]. Discrepancies between the H 4 R orthologs in the different assay systems may result from differential equilibria between the active and inactive states of the H 4 R in the different assay systems as described recently [56]. In the luciferase assay, the constitutive activity, reflected by the inverse agonism of compounds 20-23, was considerably higher for the mH 4 R than for the rH 4 R. At the latter JNJ 7777120 shifted the equilibrium toward the active state, becoming obvious as agonistic activity. Inversely, ST1006 (16), a potent agonist a human and mouse H 4 R, showed considerable inverse agonism at the rH 4 R. Thus, the outcome of studies in translational animal models cannot be unequivocally predicted by in vitro experiments, but such data may help to interprete conflicting results such as the pro-inflammatory effect of JNJ 7777120 (19) in a rat conjunctivitis model [17].
In case of agonism at the human H 4 R, the data correlate very well with data provided by more proximal readouts such as GTPase activity or GTPcS binding (Table 1, Table 2, Figure 8A). This also holds for the rank order of agonists at the mouse and rat H 4 R (Table 1, Table 2, Figure 8B,C), however, the potencies are up to 100-fold higher in the luciferase assay.

Conclusions
The reporter gene (luciferase) assay in HEK293T cells allows for the quantification of agonistic, inverse agonistic and antagonistic activity at the H 4 R species orthologs in a highly sensitive and reliable manner. In view of significantly increased potencies and efficacies of agonists, especially at the rodent H 4 R orthologs, obviously, there is a positive effect on the readout by activation/ amplification of or cross-talk between different signaling pathways in the luciferase reporter gene assay compared to more proximal functional assays on Sf9 cell membranes. It has now become clear that unequivocal characterization of H 4 R ligands as agonists, antagonists or inverse agonists in assays using a single readout is impossible. Thus, ligands have to be examined in multiple assays. But at present, it seems impossible to predict the value of the in vitro data with respect to translational animal studies and their clinical relevance.