Fluorescent H₂ Receptor Squaramide-Type Antagonists: Synthesis, Characterization, and Applications

Abstract

Fluorescence labeled ligands have been gaining importance as molecular tools, enabling receptor–ligand-binding studies by various fluorescence-based techniques. Aiming at red-emitting fluorescent ligands for the hH₂R, a series of squaramides labeled with pyridinium or cyanine fluorophores (19–27) was synthesized and characterized. The highest hH₂R affinities in radioligand competition binding assays were obtained in the case of pyridinium labeled antagonists 19–21 (pK_i: 7.71–7.76) and cyanine labeled antagonists 23 and 25 (pK_i: 7.67, 7.11). These fluorescent ligands proved to be useful tools for binding studies (saturation and competition binding as well as kinetic experiments), using confocal microscopy, flow cytometry, and high content imaging. Saturation binding experiments revealed pK_d values comparable to the pK_i values. The fluorescent probes 21, 23, and 25 could be used to localize H₂ receptors in HEK cells and to determine the binding affinities of unlabeled compounds.

The histamine H₂ receptor (H₂R), an aminergic GPCR, is one of the histamine receptor subtypes (H_1–4R) which mediate the action of the biogenic amine histamine (1). Activation of H₂R results, e.g., in gastric acid secretion,^1,2 and positive inotropic and chronotropic effects.³ In humans, the H₂R is located on parietal cells in the stomach,² in the brain,^4,5 on neutrophils and eosinophils,⁶ as well as on smooth muscle cells.⁷ However, the (patho)-physiological role of the H₂R, especially in the brain, is still far from being understood.

During the past few decades, fluorescence labeled GPCR ligands have gained increasing importance as molecular tools for the investigation of ligand-receptor-interactions as they represent a complement or even an attractive alternative to radioligands with respect to waste disposal, safety protocols, and costs.⁸⁻¹¹ Various fluorescent ligands for aminergic GPCRs have been reported, for example for muscarinic,¹²⁻¹⁴ adrenergic,^15,16 and histamine H₁^17,18 and H₃¹⁹⁻²¹ receptors. There are different fluorophore core structures available for labeling,²² e.g., BODIPY, rhodamine, dansyl/NBD, or cyanine.⁸ Most of the fluorescent ligands reported for the H₂R are emitting at wavelengths below 550 nm.^23,24 The majority of reported fluorescent H₂R ligands consist of a piperidinomethylphenoxypropylamino pharmacophore, derived either from roxatidine (2), iodoaminopotentidine (3), or BMY 2536 (4), which were linked to the fluorophore by an alkyl chain (Figure 1). Labeling with relatively small chromophores such as the N-methylanthranilic acid amide or NBD-analogs resulted in fluorescent ligands with high affinity for the H₂R (5, 6: pA₂ or pK_i > 7.0) (Figure 1).^23,24 However, 5 and 6 were inapplicable for cell-based methods e.g. flow cytometry due to the high cellular autofluorescence at the emission wavelength, giving low signal-to-noise ratios.^23,24 In order to expand the range of applications, well characterized fluorescent H₂R ligands labeled with red-emitting fluorophores (emission wavelength >600 nm) are required. First attempts to the development of such ligands were already made by labeling piperidinomethylphenoxypropylamino pharmacophores with the cyanine dye S0436 (7²⁵ and 8(26), BODIPY 650/665 (9²⁵), or BODIPY 630/650 (10²⁷) (Figure 1). AB118175, an aminopotentidine derivative labeled with BODIPY 630/650 (chemical structure not disclosed, alleged structure of 10 was deducted by the provided molecular formula (C₄₈H₅₆BF₂N₉O₄S)²⁷), is the only commercially available fluorescent ligand for H₂R, but the exact chemical structure is elusive and the high costs compromise its appropriateness as a H₂R fluorescent ligand. Additionally, AB118175 and the other red-emitting H₂R fluorescent ligands lack a comprehensive pharmacological characterization (only functional data available, no subtype selectivity) and the suitability for binding studies by fluorescence-based techniques (e.g., confocal microscopy, flow cytometry) was only partially reported.²⁵⁻²⁷ Furthermore, ligands 8 and 9 were reported to have only moderate H₂R activities (pK_b: 6.89²⁶ and 6.59²⁵).

Figure 1.

Chemical structures of histamine (1), reported H₂R standard antagonists (2–4), fluorescent H₂R ligands (5–10), and the radioligand [³H]UR-DE257 (11).^5,23−29

Here, we describe the synthesis and characterization of high affinity fluorescent H₂R antagonists with improved optical and physicochemical properties to gain access to a wide range of potential applications, in particular to confocal microscopy and to high throughput/content imaging.

Replacing the propionic acid amide of the high affinity radioligand [³H]UR-DE257 (11) (K_d value: 31 nM),²⁸ a squaramide derived from BMY 2536 (4),²⁹ by red-emitting fluorophores was the starting point for the development of high affinity fluorescent hH₂R ligands. As the physicochemical properties of the fluorescent labels can considerably effect, e.g., nonspecific binding or internalization,³⁰ we chose two different types of red-emitting fluorophores with various electrical charge: the positively charged pyrilium dye (Py-5, 12) and differently charged cyanine dyes (positive: S0223 (13a), neutral: S0436 (14a), or negative: S0386 (15a), free acids) (Scheme 1).

Scheme 1. Chemical Structures of Py-5 (12), the Free Acids S0223 (13a), S0436 (14a), and S0387 (15a), and the Succinimidyl Esters S2197 (13b), S0S0536 (14b), and S0586 (15b), as Well as the Synthesis of the Fluorescent Ligands 19–27.

Reagents and conditions: (i) DMF, TEA or DIPEA, rt, 90–120 min, 24–32%; (ii) DMF, DIPEA, rt, 45–90 min, 18–44%.

The synthesis of the amine precursors 16–18 and BMY 2536 (4) according to previously published procedures^29,31 is described in the Supporting Information. The Py-5 labeled fluorescent ligands 19–21 were synthesized by direct coupling of Py-5 (12, chameleon label³²) with the respective amine precursor 16–18 under basic conditions (Scheme 1). The reaction progressed rapidly accompanied by an immediate change in color from blue to red. The cyanine labeled fluorescent ligands 22–27 were derived from the amine precursors 16–17 by amide coupling using succinimidyl esters of the respective fluorescent dyes (13b–15b).

The fluorescence quantum yields of representative compounds (20, 21, 23, 25, 27) were determined in PBS at pH 7.4 and PBS containing 1% (w/v) of BSA (Table S1, Figure S61, Supporting Information). For all investigated compounds, fluorescence quantum yields, determined in PBS containing 1% BSA, were higher compared to the quantum yields determined in PBS devoid of protein. Fluorescence is strongly dependent on the environment of the fluorophore; this phenomenon can be explained by intermolecular hydrophobic and electrostatic interactions of the fluorescent ligand with proteins.³³ Additionally, ligands are flexible in solution and become more rigid upon binding, which generally leads to an increase in quantum yield.³³ Therefore, fluorescent ligands are not suitable for the determination of absolute values (e.g., number of specific binding sites B_max) but are valuable tools for the determination of dissociation constants (pK_d and pK_i), which are accessible by measuring relative fluorescence intensities, by analogy with the determination of count rates (cpm/cps) in competition radioligand binding. The Py-5 labeled compounds 20 and 21 showed an excitation maximum at 481 nm and an emission maximum at 646 nm in the presence of BSA. The cyanine labeled compounds 23 and 25 showed an excitation maximum at 663–667 nm and an emission maximum at 672–676 nm in the presence of BSA. Therefore, the Py-5 labeled compounds 19–21 can be excited with an argon laser at 488 nm and the cyanine labeled compounds 22–27 by a red laser at 635 nm.

The fluorescent ligands 19–27, BMY2536 (4), and the amine precursors 16–18 were investigated in equilibrium competition binding experiments on membrane preparations from Sf9 insect cells expressing the hH₂R-G_sαS fusion protein, coexpressing hH₁R and RGS4, or coexpressing either the hH₃R or the hH₄R and G_αi2 and G_β1γ2 proteins (Table 1; Table S2, Figures S62 and S63, Supporting Information).³⁴ Radioligand competition binding experiments revealed that most of the fluorescent labels were tolerated with no or only a slight decrease in affinity. Exceptions were the cyanine labeled ligands 26 and 27, in which the introduction of the S0387 fluorophore with a negative charge resulted in a decrease in hH₂R affinity (pK_i: 26, 5.69; 27, 5.88) compared to the corresponding amine-precursors (pK_i: 16, 6.52; 17, 7.87). The Py-5 labeled ligands 19–21 showed, irrespective of linker length, high hH₂R affinities (pK_i: 7.71–7.76) in the same range as the parent compound BMY 25368 (4) (pK_i: 7.80). In the case of 19, the hH₂R affinity even increased with labeling (pK_i of amine precursor 16: 6.52, pK_i of 19: 7.75). In the cyanine dye series, ligand 23, precursor 17 (n = 6, Table 1) labeled with fluorophore S0223 (positive charge), and ligand 25, precursor 17 (n = 6, Table 1) labeled with fluorophore S0436 (no charge), showed the highest hH₂R affinity (pK_i: 23, 7.67; 25, 7.11). Interestingly, labeling with Py-5, S0223, and S0436 led to an increase in hH₁R and hH₃R receptor affinity up to 2 orders of magnitude compared to the corresponding amine precursor (Table S2, Figure S62, Supporting Information). The pyridinium labeled ligands 19–21 and the cyanine labeled ligands 23 and 25 showed a slight preference for the hH₂R over the hH₁R and hH₃R. In the case of the cyanine labeled ligand 24, precursor 16 (n = 4) labeled with fluorophore S0436 (no charge), the preference changed in favor of the hH₃R. Interestingly, cyanine ligand 22, precursor 16 (n = 4) labeled with fluorophore S0223 (positive charge), showed a high hH₁R affinity (pK_i of 7.88) with a 20-fold selectivity for the hH₁R over the hH₂R. Compounds 19–21, 23, and 25, which were investigated for hH₄R affinity, showed a 70- to 150-fold selectivity for the hH₂R over hH₄R (Table S2, Figure S63, Supporting Information).

Table 1. H₂R Affinities and Antagonistic Activities as Well as HR Selectivity Profiles.

			hH2R
					pKd high content imagingd			HR selectivity profile
compd.	na	Fluorophore	pKib	pKd flow cytom.c	automated cell imaging	imaging flow cytometry	pKb (pEC50)e	ratios of KiH1R/H2R/H3R/H4Rf
His (1)			6.53 ± 0.04	n.a.	n.a.	n.a.	(5.80 ± 0.06)	–/1/0.5/0.1
4			7.80 ± 0.01g	n.a.	n.a.	n.a.	7.03 ± 0.02	–/1/1300/–
16	4		6.52 ± 0.04	n.a.	n.a.	n.a.	5.76 ± 0.22	33/1/36/33
17	6		7.87 ± 0.02	n.a.	n.a.	n.a.	6.73 ± 0.08	740/1/620/740
18	7		7.86 ± 0.02	n.a.	n.a.	n.a.	7.06 ± 0.13	380/1/550/720
19	4	Py-5	7.75 ± 0.02	7.55 ± 0.02	7.13 ± 0.01	7.06 ± 0.03	n.d.	6/1/20/70
20	6	Py-5	7.71 ± 0.04	7.84 ± 0.07	n.d.	n.d.	7.21 ± 0.04	5/1/4/120
21	7	Py-5	7.76 ± 0.01	7.73 ± 0.09	7.05 ± 0.04	7.19 ± 0.03	7.85 ± 0.10	4/1/6/160
22	4	S0223	6.57 ± 0.02	6.54 ± 0.05	6.92 ± 0.06	n.d.	n.d.	1/20/8/180
23	6	S0223	7.67 ± 0.07	7.89 ± 0.07	7.82 ± 0.10	n.d.	7.73 ± 0.04	2/1/3/50
24	4	S0436	6.49 ± 0.04	6.25 ± 0.21	6.77 ± 0.09	n.d.	n.d.	2/2/1/–
25	6	S0436	7.11 ± 0.01	7.32 ± 0.02	7.79 ± 0.13	n.d.	6.49 ± 0.03	6/1/3/130
26	4	S0387	5.69 ± 0.08	n.d.	n.d.	n.d.	n.d.	–/1/2/–
27	6	S0387	5.88 ± 0.09	n.d.	6.35 ± 0.04	n.d.	n.d.	n.d.

^aLength of the linker given as the number of carbon atoms.

^bDetermined by radioligand competition binding with [³H]UR-DE257 (K_d = 12.2 nM, c = 20 nM) at membrane preparations of Sf9 insect cells expressing the hH₂R-G_sαS fusion protein; mean values ± SEM from three independent experiments (each performed in triplicate).

^cDetermined by flow cytometric saturation binding at HEK293T-hH₂R-qs5 cells; mean values ± SEM from three independent experiments (each performed in duplicate).

^dDetermined by high-content imaging saturation binding at HEK293T-hH₂R-qs5 cells; mean values ± SEM from 2 to 3 experiments (each performed in duplicate).

^eDetermined by GTPγS binding assay at membrane preparations of Sf9 insect cells expressing the hH₂R-G_sαS fusion protein; pK_b values of neutral antagonists were determined in the antagonist mode versus histamine (c = 1 μM); mean values ± SEM from 2 to 3 independent experiments (each performed in triplicate).

^fCalculated from the K_i values obtained by conversion of the respective pK_i value (Table S2, Supporting Information).

^gData was previously reported as K_i value by Baumeister et al.²⁸ and the raw data was reanalyzed to give the pK_i values.

^c,dNonspecific binding was determined in the presence of famotidine (300-fold excess). The incubation period was 60 min (b, d, e, f) or 90 min (c). n.d.: not determined. n.a.: not applicable.

The amine precursors 16–18, BMY 25368 (4), and representative fluorescent ligands (20, 21, 23, and 25) were functionally characterized in a GTPγS binding assay on membrane preparations of Sf9 insect cells expressing the hH₂R-G_sαS fusion protein (Table 1; Figure S64, Supporting Information).³³ The investigated compounds proved antagonists with the calculated pK_b values being in good agreement with the pK_i values from radioligand competition binding.

Furthermore, representative fluorescent ligands (20, 21, 23, and 26) were investigated for hH₂R agonism in a β-arrestin2 recruitment assay on HEK293T-β-Arr2-hH₂R cells (Figure S65, Supporting Information). None of the investigated ligands showed any β-arrestin2 recruitment, indicating that no β-arrestin2 mediated internalization of the receptor–ligand-complex took place.

The hH₂R binding of the fluorescent ligands 21, 23, and 25 was also investigated by confocal microscopy (Figure 2). After 20 min of incubation, all investigated ligands were still localized at the cell membrane of HEK293T-hH₂R-qs5 cells.

Figure 2.

Visualization of binding of the fluorescent ligands 21, 23, and 25 (all 100 nM) to the membrane of HEK293T-hH₂R-qs5 cells determined by confocal microscopy after 20 min of incubation at rt. Nonspecific binding was determined in the presence of famotidine (300-fold excess). Images were acquired with a Zeiss Axiovert 200 M microscope equipped with the LSM 510 laser scanner. A 63x/1.40 oil immersion objective was used.

The ligands 21, comprising a positively charged pyridinium moiety, and 25, comprising an electroneutral cyanine moiety, showed low nonspecific binding at a concentration of 100 nM in the presence of famotidine (300-fold excess), whereas 23, comprising a positively charged cyanine moiety, showed higher nonspecific binding (Figure 2).

Fluorescent ligands 19–25, which showed moderate to high hH₂R affinity (pK_i > 6.0) were used for flow cytometric binding studies at HEK293T-hH₂R-qs5 cells (Table 1; Figure S66–S68, Supporting Information). All investigated ligands afforded pK_d values which were in good agreement with the pK_i values. Within the pyridinium labeled ligands 19–21, nonspecific binding was low (<10% of total binding around the K_d). The cyanine labeled ligands 22 and 23 with a positive charge of the fluorophore showed slightly higher nonspecific binding (∼20%). The introduction of a sulfonic acid group into the cyanine moiety resulting in an uncharged fluorophore (24, 25) had a positive effect regarding nonspecific binding (∼10%).

The association and dissociation kinetics of 21, 23, and 25 were determined on HEK293T-hH₂R-qs5 cells at 37 °C using flow cytometry (Figure S68, Table S3, Supporting Information). The fluorescent ligands showed a fast (k_on (21): 0.0043 min^–1 nM^–1) or moderate (k_on (23, 25): 0.0009–0.002 min^–1 nM^–1) association rate and incomplete dissociation. A similar dissociation behavior was reported for the closely related radioligand [³H]UR-DE257.²⁸ A possible explanation for the (pseudo)irreversible binding of these ligands is a slow dissociation from the receptor or binding to two different receptor states: a fast reversible and a tight-binding state (for more detailed information, see the Supporting Information).²⁸ The estimated pK_d(kin) values of 21, 23, and 25 were consistent with the pK_d values determined in saturation binding experiments, suggesting that binding followed at least in part the law of mass action. The (pseudo)irreversibility of binding might be suboptimal for, e.g., competition binding experiments but might be advantageous in microscopy or even in in vivo (imaging) experiments. The fluorescent ligands 19, 21–25, and 27 were also analyzed by automated cell imaging, enabling measurement of hH₂R binding on live and adherent HEK293T-hH₂R-qs5 cells. Figure 3 shows representative images after incubation of the cells with the fluorescent ligands 21 (250 nM), 23 (75 nM), 25 (75 nM), and 27 (500 nM) at room temperature for 60 min, followed by a washing step. All fluorescent ligands were localized at the cell membrane. The pK_d values determined by saturation binding by automated cell imaging were generally in good agreement with those obtained by flow cytometry (Table 1; Figures S69–S70, Supporting Information).

Figure 3.

Visualization of binding of the fluorescent ligands 21 (250 nM), 23 (75 nM), 25 (75 nM), and 27 (500 nM) to the membrane of HEK293T-hH₂R-qs5 cells (red) determined by automated cell imaging after 60 min of incubation at rt. Hoechst 33342 was used as a nuclear stain (blue). Fluorescence was mainly associated with the plasma membrane. Nonspecific binding was determined in the presence of famotidine (300-fold excess). Images were acquired with a IN Cell Analyzer 2000.

The background fluorescence (binding to plastics) was low in the case of fluorescent ligands with no or with negative charge of the fluorophore (24, 25, 27: <15% of total binding around the K_d) but relatively high if the ligand contained a fluorophore with positive charge (21: 25%; 22, 23: 30–40%). An exception was compound 19, which was also labeled with a positively charged Py-5 but showed a lower background fluorescence than 21 (∼12%).

Investigation of the association and dissociation kinetics of the cyanine labeled ligand 25 with automated cell imaging revealed a much faster association (k_on: 0.0098 min^–1 nM^–1) compared to flow cytometry but also an incomplete dissociation (Figures S71–S72, Table S3, Supporting Information). The residual fluorescent ligand was preferentially located at the cell membrane, and there was only very low fluorescence in the cytoplasm.

Results from saturation binding studies with 19 and 21 on suspended HEK293T-hH₂R-qs5 cells using an imaging flow cytometer were in good agreement with the results from automated cell imaging (Table 1; Figures S73–74, Supporting Information). The applicability of fluorescent ligands 21 and 25 for the determination of binding affinities of unlabeled ligands was demonstrated by flow cytometric and high content imaging competition binding assays (a more detailed description is provided in the Supporting Information, Figure S75, Table S4).

For none of the investigated fluorescent ligands was a negative influence on cell viability during the incubation periods of the applied assays observed (flow cytometry and high content imaging (Figures S76–S78, Supporting Information).

Taken together, we showed that the presented fluorescent ligands are useful molecular tools for nonradioactive binding studies using different techniques such as confocal microscopy, (imaging) flow cytometry, and automated cell imaging. The cyanine dye labeled ligands proved to be useful candidates for in vivo imaging as a high tissue permeability requires excitation and emission wavelengths in the far-red/near-infrared (>650 nm)³⁵ and the pyridinium labeled ligands 19–21 might be suitable molecular tools (BRET acceptors) for Nano-BRET assays.³⁶

Glossary

Abbreviations

BSA

bovine serum albumin

BRET

bioluminescence energy transfer

cpm/cps

counts per minute/counts per second

DIPEA

diisopropylethylamine

GPCR

G-protein coupled receptor

GTPγS

guanosine 5′-thiotriphosphate

HEK293T

human embryonic kidney 293T

NBD

4-nitrobenzo-2-oxa-1,3-diazole

PBS

phosphate buffered saline

pK_b

negative logarithm of the dissociation constant obtained from functional assays

pK_d

negative logarithm of the dissociation constant obtained from saturation binding experiments

pK_i

negative logarithm of the dissociation constant obtained from competition binding experiments

Sf9

Spodoptera frugiperda cells

TEA

triethylamine

TFA

trifluoroacetic acid

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.0c00033.

• Synthesis of the intermediates 30–36, BMY 25368 (4), and the amine precursors 16–18 (including Schemes S1 and S2); experimental section (including Figures S1–S9); ¹H and ¹³C NMR spectra of 30–36, 4, and 16–27 (including Figures S10–S37); HPLC chromatograms of 16–27 and BMY 2536 (4) (including Figures S38–S40); HRMS spectra of compounds 30–36, 4, and 16–27 (including Figures S41–S60); determination of quantum yields (including Figure S61 and Table S1); radioligand competition binding studies and GTPγS binding assay (including Figures S62–S64 and Table S2); investigation of β-arrestin2 recruitment by a split luciferase assay at HEK-hH₂R cells (including Figure S65); binding studies of 19–25 at HEK293T-hH₂R-qs5 cells studied by flow cytometry (including Figures S66–S68 and Table S3); binding of 19, 21–25, and 27 at HEK293T-hH₂R-qs5 cells studied by automated cell imaging (including Figures S69–S72); saturation binding of 19 and 21 at HEK293T-hH₂R-qs5 cells studied by imaging flow cytometry (including Figure S73 and S74); competition binding studies with 21 and 25 (including Figure S75 and Table S4); and rough estimate of cell viability (including Figures S76–S78) (PDF)

• Molecular formula strings (XLSX)

Author Present Address

^# (S.B.) Agrolab Labor GmbH, 84079 Bruckberg, Germany.

Author Present Address

^∇ (D.E.) Zentrale Arzneimittelüberwachung Bayern, Germany.

Author Present Address

^○ (M.C.) Sanofi Investment Co., Ltd., Chaoyang District, Beijing, China.

Author Present Address

^□ (T.L.) AbbVie Deutschland GmbH & Co. KG, 67061 Ludwigshafen am Rhein, Germany.

Author Contributions

S.B., D.E., K.T., and M.C. performed the syntheses. S.B., I.A., T.L., A.F.M., and M.G-L. performed functional and binding studies. M.T. cloned the vector hH₂R-ELucC/ELucN-β-arrestin2 under supervision of T.O. G.B. and A.B. initiated and planned the project. G.B., M.K., M.L., and A.B. supervised the research. S.B. and G.B. wrote the manuscript. All authors have given approval to the final version of the manuscript.

This work was funded by the Graduate Training Programs GRK760 (D.E., G.B., and A.B.) and GRK1910 (S.B., K.T., M.C., T.L., M.K., and A.B.) of the Deutsche Forschungsgemeinschaft (DFG) and the Elite Network of Bavaria (T.L. and A.B.).

The authors declare no competing financial interest.

Author Status

^∥ (A.B.) Deceased, July 18, 2017.