Antagonism of Human Formyl Peptide Receptor 1 (FPR1) by Chromones and Related Isoflavones

Abstract

Formyl peptide receptors (FPRs) are G protein-coupled receptors (GPCRs) expressed on a variety of cell types. Because FPRs play an important role in the regulation of inflammatory reactions implicated in disease pathogenesis, FPR antagonists may represent novel therapeutics for modulating innate immunity. Previously, 4H-chromones were reported to be potent and competitive FPR1 antagonists. In the present studies, 96 additional chromone analogs, including related synthetic and natural isoflavones were evaluated for FPR1 antagonist activity. We identified a number of novel competitive FPR1 antagonists that inhibited fMLF-induced intracellular Ca²⁺ mobilization in FPR1-HL60 cells and effectively competed with WKYMVm-FITC for binding to FPR1 in FPR1-HL60 and FPR1-RBL cells. Compound 10 (6-hexyl-2-methyl-3-(1-methyl-1H-benzimidazol-2-yl)-4-oxo-4H-chromen-7-yl acetate) was found to be the most potent FPR1-specific antagonist, with binding affinity K_i~100 nM. These chromones inhibited Ca²⁺ flux and chemotaxis in human neutrophils with nanomolar-micromolar IC₅₀ values. In addition, the most potent novel FPR1 antagonists inhibited fMLF-induced phosphorylation of extracellular signal-regulated kinases (ERK1/2) in FPR1-RBL cells. These antagonists were specific for FPR1 and did not inhibit WKYMVM/WKYMVm-induced intracellular Ca²⁺ mobilization in FPR2-HL60 cells, FPR3-HL60 cells, RBL cells transfected with murine Fpr1, or interleukin 8-induced Ca²⁺ flux in human neutrophils and RBL cells transfected with CXC chemokine receptor 1 (CXCR1). Moreover, pharmacophore modeling showed that the active chromones had a significantly higher degree of similarity with the pharmacophore template as compared to inactive analogs. Thus, the chromone/isoflavone scaffold represents a relevant backbone for development of novel FPR1 antagonists.

1. Introduction

Formyl peptide receptors (FPRs) are G protein-coupled receptors (GPCR) that play important roles as sensors of pathogen- and host-derived products and recruit leukocytes to sites of infection where these cells exert microbicidal effector functions and clear cellular debris [1, 2]. In humans, there are three FPR isoforms: FPR1, FPR2, and FPR3 [3]. These receptors are expressed on a variety of cell types, including macrophages, neutrophils, T lymphocytes, epithelial cells, dendritic cells, fibroblasts, and astrocytes (reviewed in [3, 4]). Being expressed in the majority of white blood cells, FPRs play an important role in the regulation of inflammatory reactions and cellular dysfunction and, thereby, represent an attractive family of pharmacological targets for therapeutic development [4–6].

FPR1 exhibits high affinity for formyl peptides, which are produced by bacteria and can also be released from damaged mitochondria during tissue injury [7, 8]. In addition, FPR1 has recently been reported to contribute to disease pathogenesis. For example, interaction of endogenous annexin A1 with FPR1 leads to transactivation of the receptor for epithelial growth factor (EGFR), which promotes invasion and growth of glioma cells [2]. Likewise, Cheng et al. [9] reported that FPR1 expression is associated with tumor progression and survival in gastric cancer. Thus, bioactive ligands acting as FPR1 antagonists might serve as useful therapeutics in host defense in order to reduce detrimental effects associated with inflammation and cancer.

The receptor-specific and most potent FPR1 antagonists described so far are the fungal hydrophobic cyclic peptides, cyclosporines A and H [10]. Although cyclosporine H blocked N-formyl-Met-Leu-Phe (fMLF)-induced analgesia [11] and attenuated the acute inflammatory response evoked by cigarette smoke [12], in vivo studies of cyclosporines should be interpreted carefully because their main therapeutic effects appear to involve signaling pathways unrelated to FPR1 [13–16]. Other known peptide FPR antagonists are Boc-1 (Boc-MLF) and Boc-2 (Boc-FLFLFL), and there are several reports of in vivo application of Boc-2 [17–20]. Recently, analogs of Boc-2 were reported as FPR1 antagonists [21, 22]. Several non-steroidal anti-inflammatory drugs (NSAIDs), including diclofenac, piroxicam, sulfinpyrazone, and tenoxicam have been reported as low activity FPR1 antagonists [23–25]. However, because NSAIDs exhibit a variety of pharmacological properties, these drugs are not suitable for in vivo studies designed to probe the physiological roles of FPR1.

Growing evidence supporting the anti-inflammatory and tissue-protective effects of FPR antagonists led to the screening of commercial libraries for novel small-molecule FPR antagonists. As result of these screening efforts and/or structure–activity relationship (SAR)-directed design and synthesis, a number of synthetic non-peptide FPR1/FPR2 antagonists with a wide range of chemical diversity have been identified ([26–33]). Structures of the most potent small-molecule FPR1 antagonists are shown in Figure 1. Among these competitive FPR1 antagonists are some compounds with a 4H-chromen-4-one scaffold (Figure 1, compounds 1-4) [26, 27]. However, activities of these chromones in primary cells, SAR analysis of related chromones and isoflavones, as well as molecular modeling have not been described.

Figure 1.

Structures of previously reported small-molecule FPR1 antagonists.

In the present study, we evaluated 96 4H-chromen-4-ones, including synthetic and naturally occurring isoflavones, for their ability to antagonize FPR-dependent signaling in neutrophils and FPR-transfected cells and identified novel and potent FPR1-specific antagonists. These antagonists were specific for FPR1 and did not inhibit FPR2-, FPR3-, or CXCR1-dependent responses. SAR analysis of these compounds revealed the importance of a small hydrophobic group at position 2 and the type of substituent at position 7 of the 4H-chromen-4-one scaffold. In addition, molecular modeling showed a high degree of similarity for low-energy conformations of these antagonists to the pharmacophore model for FPR1 antagonists. Overall, the isoflavone scaffold represents an appropriate backbone to develop novel FPR1 antagonists.

2. Materials and Methods

2.1. Materials

Phorbol-12-myristate-13-acetate (PMA), dimethyl sulfoxide (DMSO), fMLF, HEPES, Percoll, and Histopaque 1077 were from Sigma Chemical Co. (St. Louis, MO). RPMI 1640 medium and penicillin-streptomycin solution were from Mediatech (Herdon, VA). Fetal bovine serum (FBS) was from Atlas Biologicals (Fort Collins, CO). Peptides WKYMVm and WKYMVM were from Calbiochem (San Diego, CA) and Tocris Bioscience (Ellisville, MO), respectively. Human interleukin-8 (IL-8) was from Peprotech Inc (Rocky Hill, NJ). Hanks’ balanced salt solution (HBSS), Fluo-4 AM, and G418 were from Life Technologies (Grand Island, NY). HBSS containing 1.3 mM CaCl₂ and 1.0 mM MgSO₄ is designated as HBSS⁺; HBSS without ions Ca²⁺ and Mg²⁺ is designated as HBSS⁻. Selected 4H-chromen-4-ones were purchased from ChemDiv (San Diego, CA), ChemBridge (San Diego, CA), Princeton BioMolecular Research (Monmouth Junction, NJ), Indofine (Hillsborough, NJ), Otava (Toronto, Canada), Vitas-M Laboratory (Moscow, Russia), and InterBioScreen (Moscow, Russia). Fluorescein isothiocyanate (FITC) was conjugated to the lysine residue of the WKYMVm peptide to produce a fluorescent ligand (WKYMVm-FITC) that binds to both FPR1 and FPR2 (custom synthesis by Bachem, Torrance, CA).

2.2. Cell Culture

Human promyelocytic leukemia HL-60 cells stably transfected with FPR1 (FPR1-HL60 cells) or FPR2 (FPR2-HL60 cells) were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated FBS, 10 mM HEPES, 100 μg/ml streptomycin, 100 U/ml penicillin, and G418 (1 mg/mL), as described previously [34]. Rat basophilic leukemia (RBL-2H3) cells transfected with human FPR1 (FPR1-RBL), human CXCR1, or mouse Fpr1 (Fpr1-RBL) were cultured in DMEM supplemented with 20% (v/v) heat-inactivated FBS, 10 mM HEPES, 100 μg/ml streptomycin, 100 U/ml penicillin, and G418 (250 μg/ml), as described previously. Although stable cell lines are cultured under G418 selection pressure, G418 may affect some assays, so it was removed in the last round of culture before assays. Wild-type HL-60 and RBL-2H3 cells were cultured under the same conditions, but without G418.

2.3. Isolation of Human Neutrophils

Blood was collected from healthy donors in accordance with a protocol approved by the Institutional Review Board at Montana State University. Neutrophils were purified from the blood using dextran sedimentation, followed by Histopaque 1077 gradient separation and hypotonic lysis of red blood cells, as described previously [35]. Isolated neutrophils were washed twice and resuspended in HBSS⁻. Neutrophil preparations were routinely >95 % pure, as determined by light microscopy, and > 98 % viable, as determined by trypan blue exclusion.

2.4. Isolation of Bone Marrow Leukocytes and Murine Neutrophils

All animal use was conducted in accordance with a protocol approved by the Institutional Animal Care and Use Committee at Montana State University. Bone marrow leukocytes were flushed from tibias and femurs of BALB/c mice with HBSS, filtered through a 70-μm nylon cell strainer to remove cell clumps and bone particles, and resuspended in HBSS⁻ at 1 × 10⁶ cells/ml. Bone marrow neutrophils were isolated from the bone marrow leukocyte preparations as described previously [36]. Briefly, bone marrow leukocytes were resuspended in 3 ml of 45% Percoll solution and layered on top of a Percoll gradient consisting of 2 ml each of 50, 55, 62, and 81% Percoll solutions in a conical 15-ml polypropylene tube. The gradient was centrifuged at 1600g for 30 min at 10°C, and the cell band located between the 62 and 81% Percoll layers was collected. The cells were washed, layered on top of 3 ml of Histopaque 1119, and centrifuged at 1600g for 30 min at 10°C to remove contaminating red blood cells. The purified neutrophils were collected, washed, and resuspended in HBSS⁻.

2.5. Ca²⁺ Mobilization Assay

Changes in intracellular Ca²⁺ were measured with a FlexStation II scanning fluorometer (Molecular Devices, Sunnyvale, CA) in human neutrophils, HL-60 cells, and RBL cells, as described previously [34]. The cells, suspended in HBSS⁻ containing 10 mM HEPES, were loaded with Fluo-4 AM dye (Invitrogen) (1.25 μg/mL final concentration) and incubated for 30 min in the dark at 37 °C. After dye loading, the cells were washed with HBSS⁻ containing 10 mM HEPES, resuspended in HBSS⁺ containing 10 mM HEPES, and aliquotted into the wells of flat-bottom, half-area-well black microtiter plates (2 × 10⁵ cells/well). For evaluation of direct agonist activity, the compounds of interest were added from a source plate containing dilutions of test compounds in HBSS⁺, and changes in fluorescence were monitored (λ_ex = 485 nm, λ_em = 538 nm) every 5 s for 240 s at room temperature after automated addition of compounds.

Antagonist activity and selectivity were evaluated after 5–30 min pretreatment with test compounds at room temperature, followed by addition of peptide/chemokine agonist (5 nM fMLF, 5 nM WKYMVm, 10 nM WKYMVM, or 25 nM IL-8). In some experiments, a range of fMLF concentrations was used. Maximum change in fluorescence during the first 3 min, expressed in arbitrary units over baseline, was used to determine a response. Responses for FPR1 antagonists were normalized to the response induced by 5 nM fMLF for FPR1-HL60 cells and neutrophils, which were assigned a value of 100%. Curve fitting (5–6 points) and calculation of median effective inhibitory concentrations (IC₅₀) were performed by nonlinear regression analysis of the dose-response curves generated using Prism 6 (GraphPad Software, Inc., San Diego, CA). Efficacy is expressed as % inhibition by an antagonist of the response induced by 5 nM fMLF at the maximal applied concentration of an antagonist (~ 50 μM).

2.6. Chemotaxis Assay

Human neutrophils were suspended in HBSS⁺ containing 2% (v/v) heat-inactivated FBS (2 × 10⁶ cells/ml), and chemotaxis was analyzed in 96-well ChemoTx chemotaxis chambers (Neuroprobe, Gaithersburg, MD), as described previously [35] with modifications. In brief, neutrophils were preincubated with the indicated concentrations of the tested compounds or DMSO for 30 min at room temperature and added to the upper wells of the ChemoTx chemotaxis chambers. The lower wells were loaded with 30 μl of HBSS⁺ containing 2% (v/v) heat-inactivated FBS with the indicated concentrations of test compounds plus 1 nM fMLF, DMSO plus 1 nM fMLF (positive control), or DMSO alone (negative control). Neutrophils were allowed to migrate through the 5.0-μm pore polycarbonate membrane filter for 60 min at 37 °C and 5% CO₂. The number of migrated cells was determined by measuring ATP in lysates of transmigrated cells using a luminescence-based assay (CellTiter-Glo; Promega, Madison, WI), and luminescence measurements were converted to absolute cell numbers by comparison of the values with standard curves obtained with known numbers of neutrophils. Curve fitting (at least eight to nine points) and calculation of median effective concentration values (IC₅₀) were performed by nonlinear regression analysis of the dose-response curves generated using Prism software.

2.7. Competition Binding Assay

Dose-response assays were performed to measure test compound competition with the high-affinity fluorescent ligand WKYMVm-FITC for binding to human FPR1 in HL-60 and RBL transfected cells, as described previously [27] with modifications. Briefly, FPR1-HL60 or FPR1-RBL cells were preincubated with different concentrations of test compound for 30 min at 4°C. WKYMVm-FITC (0.5 nM) was added, and after incubation for an additional 30 min at 4°C, the samples were immediately analyzed using flow cytometry (LSRII; BD Biosciences, San Jose, CA) without washing. The assay response range was defined by replicate control samples containing 1 μM of unlabeled fMLF (positive control) or buffer (negative control). In an individual dose-response experiment, each compound was tested in duplicate, resulting in 9 data points. The ligand competition curves were fitted by Prism software using nonlinear least-squares regression in a sigmoidal dose-response model to determine the concentration of added test compound that inhibited fluorescent ligand binding by 50% (i.e., IC₅₀). In equilibrium binding experiments with the labeled ligand, the K_d values for FPR1 in HL60-FPR1 and RBL- FPR1 cells were found to be ~0.5 and 0.4 nM, respectively. K_i values were calculated from IC₅₀, as reported previously [27].

2.7. ERK Phosphorylation Assay

Phosphorylation of p44/42 mitogen-activated protein kinases (ERK1/2) was determined based on activation-associated phosphorylation. Cells cultured in six-well plates were serum starved for 4 hr before stimulation. Samples were treated with the indicated compounds at 50 μM or 0.5% DMSO (vehicle) for 10 min and then stimulated with fMLF (20 nM) for various time periods (0 – 15 min) or 200 ng/ml of PMA. The reactions were terminated by adding 150 μl of ice-cold SDS-PAGE loading buffer. Samples were analyzed by SDS-PAGE and Western blotting using rabbit anti-ERK1/2 (9102) and rabbit anti-ERK1/2 phosphorylated at Thr202 and Tyr204 (9101) (1:1000; Cell Signaling Technology, Inc., Danvers, MA). Horseradish peroxidase-conjugated AffiniPure goat anti-rabbit IgG (1:3000; Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) was used as secondary antibody. The immunoblots were visualized using a SuperSignal West Pico Chemiluminescence kit (Thermo Scientific, Rockford, IL) according to the manufacturer’s instructions. The results of Western blotting were quantified using ImageJ software from the National Institutes of Health.

2.8. Compound Cytotoxicity

Cytotoxicity was analyzed with a CellTiter-Glo Luminescent Cell Viability Assay Kit (Promega, Madison, WI), according to the manufacturer’s protocol. Briefly, HL-60 cells (wild type) were cultured at a density of 1×10⁵ cells/well with different concentrations of compound under investigation for 18 h at 37 °C and 5% CO₂. Following treatment, the cells were allowed to equilibrate to room temperature for 30 min, substrate was added, and the samples were analyzed with a Fluoroscan Ascent FL microplate reader.

2.9. Molecular Modeling

Four FPR1 antagonists (compounds 1, 5-7) with different chemical scaffolds and relatively high FPR1 antagonist activity were chosen as reference compounds for pharmacophore modeling (structures shown in Figure 1). We used a ligand-based approach for molecular modeling known as field point methodology [37, 38], as described previously [39]. The structures of the compounds were pre-optimized by the semi-empirical PM3 method using HyperChem 8.0 software. The Polak-Ribiere conjugate gradient method was applied for the minimization with a RMS gradient less than 0.01 kcal·mol⁻¹·Å⁻¹ as a convergence condition. The structures in Tripos MOL2 format were then imported into the FieldTemplater program (FieldTemplater Version 2.0.1; Cresset Biomolecular Discovery Ltd., Hertfordshire, UK). The conformation hunter algorithm was used to generate representative sets of conformations corresponding to local minima of energy calculated within the extended electron distribution force field [37, 40]. This algorithm incorporated in the FieldTemplater and FieldAlign software allowed us to obtain up to 200 independent conformations that were used in further calculation of field points surrounding each conformation of each molecule. To decrease the number of rotatable bonds during the conformation search, the “force amides trans” option was enabled in the program. For the generation of field point patterns, probe atoms having positive, negative, and zero charge were placed in the vicinity of a given conformation, and the energy of their interaction with the molecular field was calculated using the extended electron distribution parameter set. Positions of energy extrema for positive probes give “negative” field points, whereas energy extrema for negative and neutral probe atoms correspond to “positive” and steric field points, respectively. Hydrophobic field points were also generated with neutral probes capable of penetrating into the molecular core and reaching extrema in the centers of hydrophobic regions (e.g., benzene rings). The size of a field point depends on magnitude of an extremum. The field points are colored according to the following convention: blue, electron-rich (negative); red, electron-deficient (positive); yellow, van der Waals attractive (steric); and orange, hydrophobic. A detailed description of the field point calculation procedure has been published elsewhere [41]. A clique-matching algorithm with further simplex optimization was applied to obtain the conformations of four active molecules giving good mutual overlays in terms of geometric and field similarity. However, it was possible to obtain a satisfactory superimposition of three FPR1 antagonists (1, 5, and 7). The best overlay was taken as a template representative of a FPR1 antagonist bioactive conformation.

Structures of additional FPR1 antagonists and inactive analogs were pre-optimized by the PM3 method using the approach described above, converted into Tripos MOL2 format, and imported into FieldAlign program (FieldAlign Version 2.0.1; Cresset Biomolecular Discovery Ltd., Hertfordshire, UK) together with the template. The conformation hunter algorithm was used to generate up to 200 conformations for each ligand, which were then superimposed onto the template, taking into account mutual correspondence of field points. Superimpositions with the best similarity scores were collected. Statistical analysis of the similarity scores was performed with the use of ANOVA methodology as implemented in STATISTICA 8.0 software.

The FPR1 homology model was created using the crystal structure of bovine rhodopsin, as reported previously [42]. Structure of compound 10 was pre-optimized by the semi-empirical PM3 method using HyperChem software, saved in Tripos MOL2 format, imported into the Molegro Virtual Docker (MVD) program (MVD 2010.4.2, MolegroApS), and docked as reported previously [43].

3. Results

3.1. SAR Analysis of 4H-Chromen-4-one FPR1 Antagonists

Based on structures of four published FPR1 antagonists 1-4 with a 4H-chromen-4-one scaffold [26, 27] (Figure 1), 96 additional chromen-4-one analogs and related isoflavones were selected and evaluated for FPR1 antagonist activity in FPR1-HL60 cells by monitoring effects on fMLF-induced Ca²⁺ mobilization. These analogs included 25 derivatives of compound 1 (benzimidazoles 8-32, series A), 22 derivatives of compounds 2 and 3 (furoyl and thienoyloxy derivatives 33-54, series B), nine 7-benzoyloxy-substituted isoflavones (55-63, series C), 26 derivatives of compound 4 (2-trifluoromethyl-2′-methoxy isoflavones 64-89, series D), and 14 additional natural and synthetic isoflavones (90-103, series E). Structures of these analogs and reference compounds 1-4 are presented in Table 1.

Table 1.

FPR1 Antagonist activity of 4H-chromen-4-ones

A. 3-(Benzimidazol-2-yl)-4H-chromen-4-ones (1, 8-32)a
Compd	R1	R2	R3	IC50 (μM) and efficacy (%)in FPR1-HL60	Sim.b
1	CH2CH3	OCOCH3	CH3	1.4 ± 0.24 (100)	0.676
8	H	OCOCH3	H	N.A.	0.549
9	CH2CH3	OCOCH3	H	N.A.	0.562
10	(CH2)5CH3	OCOCH3	CH3	0.31 ± 0.13 (100)	0.609
11	CH2CH3	OCO-tBu	CH3	N.A.	0.552
12	H	OCO-tBu	CH3	N.A.	0.613
13	CH2CH3	OCO-tBu	(CH2)2COOH	N.A.	0.562
14	H	OCO-tBu	iPr	N.A.	0.538
15	H	OH	CH3	24.7 ± 3.2 (80)	0.666
16	CH2CH3	OH	H	N.A.	0.569
17	CH2CH3	OH	CH3	N.A.	0.566
18	CH2CH3	OH	iPr	N.A.	0.541
19	CH2CH3	OH	CF3	N.A.	0.551
20	H	OH	CH2CH3	N.A.	0.563
21	H	OH	iPr	N.A.	0.543
22	H	OH	(CH2)2COOH	N.A.	0.569
23	CH2CH3	OH	(CH2)2COOH	N.A.	0.600
24	CH2CH3	OH	COOCH2CH3	N.A.	0.560
25	CH2CH3	OH	CH2OCH2COOH	N.A.	0.591
26	CH2CH3	OCO-2-Furyl	H	N.A.	0.606
27	H	OCO-2-Furyl	CH3	3.1 ± 1.1 (100)	0.620
28	CH2CH3	OCO-C6H4Cl-o	CH3	5.2 ± 1.8 (90)	0.634
29	CH2CH3	OCO-Ph	H	N.A.	0.538
30	CH2CH3	OCO-C6H4OMe-p	H	N.A.	0.564
31	H	OCH2-C6H4NO2-p	CH3	N.A.	0.547
32	CH2CH3	OSO2CH3	H	N.A.	0.554

B. 7-(2-Furoyloxy)- and 7-(2-thienoyloxy)-substituted isoflavones (2,3,33-54)
Compd	X	R3	R4	R5	R6	IC50 (μM) and efficacy (%)in FPR1-HL60	Sim.
2	O	CF3	OCH3	H	H	0.44 ± 0.13 (100)	0.576
3	S	CF3	Cl	H	H	0.42 ± 0.11 (100)	0.567
33	S	CF3	OCH3	H	H	N.A.	0.566
34	O	CF3	H	H	H	N.A.	0.579
35	S	H	OCH3	H	H	N.A.	0.558
36	O	CF3	Cl	H	H	0.38 ± 0.17 (100)	0.559
37	O	H	Cl	H	H	N.A.	0.563
38	S	CF3	H	H	H	3.9 ± 1.3 (90)	0.574
39	O	CF3	H	H	Cl	N.A.	0.558
40	S	CF3	H	H	OCH3	N.A.	0.556
41	S	CH3	H	H	OCH3	N.A.	0.539
42	S	H	H	H	H	N.A.	0.572
43	S	CH3	H	H	Cl	N.A.	0.541
44	S	CH3	H	OCH3	OCH3	N.A.	0.559
45	O	CH3	OCH3	H	H	1.9 ± 0.24 (100)	0.580
46	S	CH3	OCH3	H	H	2.6 ± 0.41 (100)	0.571
47	O	CH3	H	H	H	2.9 ± 0.32 (100)	0.581
48	S	CH3	H	H	H	7.2 ± 1.8 (95)	0.577
49	O	CH3	H	H	Br	N.A.	0.549
50	O	H	H	H	OCH3	N.A.	0.566
51	O	H	H	H	Cl	N.A.	0.565
52	O	H	H	H	Br	N.A.	0.551
53	O	COOCH2CH3	H	H	OCH3	N.A.	0.542
54	O	COOCH2CH3	OCH3	H	H	N.A.	0.536

C. 7-Benzoyloxy-substituted isoflavones (55-63)
Compd	R3	R4	R7	R8	R9	IC50 (μM) and efficacy (%)in FPR1-HL60	Sim.
55	CF3	OCH3	H	H	H	1.3 ± 0.19 (90)	0.562
56	CF3	OCH3	H	H	Cl	N.A.	0.548
57	CF3	OCH3	Cl	H	H	N.A.	0.551
58	CF3	OCH3	H	H	OCH3	N.A.	0.537
59	CF3	OCH3	H	OCH3	H	N.A.	0.549
60	CF3	OCH3	OCH3	H	H	N.A.	0.559
61	CF3	Cl	Br	H	H	N.A.	0.555
62	CF3	Cl	NO2	H	H	N.A.	0.549
63	H	OCH3	H	H	H	N.A.	0.562

D. 2-Trifluoromethyl-2′-methoxy isoflavones (4,64–89)
Compd	R2	IC50 (μM) and efficacy (%)in FPR1-HL60	Sim.
4	OCO-iPr	1.4 ± 0.3 (100)	0.591
64	OH	3.9 ± 1.2 (70)	0.600
65	OCH3	N.A.	0.530
66	OCH2CH3	N.A.	0.549
67	OCH2CN	N.A.	0.552
68	OCO-cycloPr	0.8 ± 0.3 (100)	0.590
69	OCO-tBu	N.A.	0.564
70	OCOCH-Et2	N.A.	0.540
71	OCOCH2-tBu	N.A.	0.561
72	OCO-cycloC6H11	N.A.	0.577
73	OCOOCH2CH3	2.2 ± 0.6 (80)	0.596
74	OCH2COOH	6.1 ± 1.7 (100)	0.592
75	OCH2CONH2	15.6 ± 3.3 (65)	0.595
76	OCH2COOCH3	N.A.	0.562
77	OCH2COOCH2CH3	N.A.	0.580
78	OCH2COO(CH2)2CH3	N.A.	-a
79	OCH2COOCH(CH3)2	N.A.	-
80	OCH2COOCH2CH(CH3)2	N.A.	-
81	OCH2COO(CH2)3CH3	N.A.	-
82	OCH(CH3)COOCH3	N.A.	-
83	OCH2C=C(CH3)2	N.A.	-
84	OCH2(CH3)(=CH2)	N.A.	-
85	OCH2COO(CH2)4CH3	N.A.	-
86	OCH(CH3)COOCH(CH3)2	N.A.	-
87	OCH(CH3)COO(CH2)3CH3	N.A.	-
88	OCH(CH3)COOCH2CH(CH3)2	N.A.	-
89	OCH(CH2CH3)COOCH2CH3	N.A.	-

E. Miscellaneous natural and synthetic isoflavones (90-103)
Compd	R2	R3	R4	R10	R11	IC50 (μM) and efficacy (%)in FPR1-HL60	Sim.
90a	OH	H	H	H	H	N.A.	0.530
91	OH	H	H	OH	H	N.A.	0.569
92	OH	H	H	OCH3	H	N.A.	0.561
93	OH	H	H	OH	OH	N.A.	0.566
94	OH	H	H	OCH3	OH	10.3 ± 1.7 (80)	0.551
95	OH	CF3	H	OCH3	OH	2.6 ± 0.8 (100)	0.591
96	OH	CH3	H	OCH3	H	N.A.	0.558
97	OH	CF3	H	OCH3	H	N.A.	0.554
98	OH	CH3	OCH3	H	H	23.1 ± 1.8 (50)	0.608
99	OH	CH3	OCH3	H	OH	N.A.	0.584
100	OCOCH3	CH3	OCH3	H	H	6.9 ± 2.7 (100)	0.596
101	OCH2COOCH2CH3	CH3	OCH3	H	H	N.A.	0.588
102	OCH2COOH	CH3	H	OCH3	H	N.A.	0.553
103						N.A.	0.593

^aSubstituents at similar positions of the molecules in series A–E have the same numbers in the enumeration scheme chosen in this paper.

^bSimilarity (Sim.) between aligned molecule and the template was calculated using the FieldAlign program.

N.A.: No activity was observed at the highest tested concentration (50 μM).

N.A.: No activity was observed at the highest tested concentration (50 μM).

N.A.: No activity was observed at the highest tested concentration (50 μM).

^aNot calculated.

N.A.: No activity was observed at the highest tested concentration (50 μM).

^aCommon names of the natural isoflavones are: 90, 7-hydroxy isoflavone; 91, daidzein; 92, formononetin; 93, genistein; 94, biochanin A; 103, afrormosin.

N.A.: No activity was observed at the highest tested concentration (50 μM).

In previous studies of potential FPR1 antagonists where inhibition of fMLF-induced Ca²⁺ mobilization was evaluated, cells were pretreated for 5 to 30 min with putative antagonists prior to fMLF addition [26, 27]. Here, we found that pretreatment of FPR1-HL60 cells with novel chromone analogs for 5 min at room temperature blocked fMLF-induced Ca²⁺ flux up to 90% of control level (see example in Figure 2A). However, the response was completely blocked only after 30-min pretreatment (Figure 2B). Thus, we used a 30-min pretreatment period for further screening of chromone analogs and dose-response analysis. As an example, a representative dose-response curve for inhibition of fMLF-induced Ca²⁺ mobilization in FPR1-HL60 cells by compound 10 is shown in Figure 2C. Antagonist activity (expressed as IC₅₀) of all chromone analogs tested from scaffolds A–E is presented in Table 1.

Figure 2. Inhibition fMLF-induced Ca²⁺ mobilization in FPR1-HL60 but not FPR2-HL60 cells by compound 10.

Panel A. FPR1-HL60 cells were preincubated for 5 min (at 25 °C) with 1 μM of compound 10 or DMSO (vehicle control) and treated with 5 nM fMLF or DMSO (vehicle control). Kinetics of Ca²⁺ mobilization were monitored for 5 minutes. Panel B. FPR1-HL60 cells were preincubated for the indicated times (at 25 °C) with 1 μM of compound 10 or DMSO (vehicle control) and treated with 5 nM fMLF or DMSO (vehicle control). Inhibition of fMLF-induced response was calculated as indicated under Materials and Methods (mean ± S.D.; n=3). *Significance difference from 100% inhibition (p <0.05). Panel C. FPR1-HL60 cells (●) and FPR2-HL60 cells (□) were preincubated with the indicated concentrations of compound 10 for 30 min and 25 °C, and the cells were stimulated with 5 nM of fMLF or WKYMVM, respectively. The response induced by peptide agonist alone was assigned a value of 100%, and % of control responses are shown. In each panel, values are samples from one experiment that is representative of three independent experiments.

All benzimidazole derivatives from series A that inhibited fMLF-induced Ca²⁺ flux in FPR1-HL60 cells (compounds 1, 10, 15, 27, and 28) contained a CH₃ group at position 2 of the chromone scaffold (R₃) (Table 1). This methyl moiety was essential for antagonist activity, as elimination of this group led to inactive compounds (compare active 1 and inactive 9). Substitution at position 7 of the chromone scaffold (R₂) also had effects on activity, but a wide range of modifications was tolerated. Although substitution of OCOCH₃ in reference compound 1 with tert-butyl-carboxyl or hydroxyl groups led to inactive compounds 11 and 17, compound 28 with a bulky ortho-chloro-benzoyloxy (OCO-C₆H₄Cl-o) substituent at R₂ retained FPR1 antagonist activity. Other active FPR1 antagonists from series A contained hydroxyl (15) or furan-2-carboxylate (27) groups at this position.

All active FPR1 antagonists in series B (compounds 2, 3, 36, 38, and 45-48) contained CH₃ or CF₃ groups at position 2 of the chromone heterocycle (R₃), supporting the importance of a small hydrophobic group at this position for antagonist activity, which was noted above for series A compounds (Table 1). Indeed, elimination of the CF₃ group in compound 36 or substitution of CF₃ in reference compound 2 with an ethyl-carboxylate group resulted in inactive compounds 37 and 54, respectively. Although replacement of CF₃ by CH₃ resulted in decreased activity for some analogs (compare 2 and 45 or 38 and 48), this same replacement converted inactive compound 34 into active 47. Replacement of the furan ring by thiophene in the aroyloxy group led to variable effects on activity, depending on the presence of substituents at other positions in a given molecule. Most active derivatives within this series contained Cl (compounds 3 and 36) or OCH₃ (compounds 2, 35, 45, and 46) in the ortho position of the benzene ring (R₄). However, the presence of a substituent at this position was not absolutely essential for antagonist activity, as compounds 38 and 47 were also active.

From series C, only compound 55 had FPR1 antagonist activity (Table 1). Addition of substituents in the benzoyloxy phenyl ring led to complete loss of activity among the 2-trifluoromethyl derivatives (56-62). Elimination of the CF₃ group in compound 55 resulted in inactive compound 63, again supporting the importance of this group at position 2 of the chromone scaffold for FPR1 antagonist activity. Based on analysis of benzoyloxy-containing compounds 28-31 (A series) and 55-63 (C series), the presence of a benzoyloxy group at R₂ is not an essential feature for antagonist activity, probably because of the bulkiness and/or high hydrophobicity of this substituent. Hence, we did not consider further compounds with a benzoyloxy moiety at R₂ in subsequent screening of 4H-chromen-4-ones.

Keeping substituents CF₃ and OCH₃ at positions R₃ and R₄, respectively, since these groups led to the best FPR1 antagonist activity, we next evaluated a variety of alkoxy, acyloxy, and other oxygen-containing groups at position 7 of the chromone scaffold (R₂) (Table 1, series D). Five novel FPR1 antagonists (64, 68, and 73-75) were identified in the series of 2-trifluoromethyl-2′-methoxy isoflavones, with the most potent being compound 68. It is interesting to note that replacement of cyclopropyl in compound 68 with isopropyl decreased antagonist activity by ~20-fold (reference compound 4), while substitution of the cyclopropyl group with tert-butyl resulted in inactive compound 69. A possible reason for this decrease is the smaller effective size of the cyclopropyl group compared to the isopropyl and tert-butyl substituents. Other compounds with antagonist activity were molecules bearing hydroxyl (64), OCOOCH₂CH₃ (73), OCH₂COOH (74), and OCH₂CONH₂ (75) groups. On the other hand, longer R₂ chains in compounds 76-89 led to loss of activity, probably due to steric hindrance within the binding pocket.

Four novel FPR1 antagonists, including the naturally occurring isoflavone biochanin A (compound 94), were identified among the 14 natural and synthetic isoflavones tested (series E). Note, however, that the closely related isoflavone genistein (compound 93) was inactive. Removal of a hydroxyl group at position 5 of the isoflavone scaffold (R₁₁) resulted in loss of antagonist activity (compare compounds 95 and 97), suggesting this OH group may increase tolerance for a modification at position 2 (R₃), because removing CF₃ in compound 95 led to decreased but not complete loss of antagonist activity (compound 94), which was a feature of the compounds in series A-C. On the other hand, introduction of a hydroxyl group at position 5 of compound 98 resulted in inactive compound 99. Thus, the effect of this hydroxyl substitution may depend on the presence of substituents at other positions of a molecule. Overall, no clear SAR emerged from other modifications in this series, and additional evaluation of synthetic and natural occurring isoflavones, especially 5-hydroxy isoflavones, will be necessary.

All 100 chromone derivatives were evaluated for FPR-independent activation in non-transfected HL-60 cells, and only compounds 94 and 95 directly activated Ca²⁺ mobilization in wild-type cells (data not shown), indicating their apparent FPR antagonist activity could be due to receptor cross-desensitization or some other nonspecific effect on the cell. Thus, these compounds were not included in further studies of specific antagonists.

3.2. Competition Binding of Selected FPR1 Antagonists

Compounds that inhibited fMLF-induced Ca²⁺ mobilization in FPR1-HL60 cells with IC₅₀ <5 μM and efficacy >75% were analyzed for their ability to compete with WKYMVm-FITC for binding to FPR1 in FPR1-HL60 and FPR1-RBL cells, as described previously [44]. Values of IC₅₀ for the competitive binding assay and calculated inhibition constants (K_i) are presented in Table 2. As an example, a representative dose–response curve of competitive inhibition of WKYMVm-FITC binding by compound 10 in FPR1-RBL transfected cells is shown in Figure 3A. Compound 10 had the highest binding affinity (K_i =110 and 90 nM in FPR1-HL60 and FPR1-RBL cells, respectively) among all chromones evaluated, including previously reported chromone FPR1 antagonists 1-4 [27, 45]. Notably, compound 10 had a 5.7-fold higher binding affinity and ~4.5-fold greater antagonist activity than compound 1, which was the most potent, selective small-molecule FPR1 antagonist previously reported. Overall, the binding affinity (K_i) of the active compounds in FPR1-HL60 cells exhibited a good linear correlation (r =0.92, n =11) with potency (IC₅₀) of the ligands in fMLF-stimulated FPR1-HL60 cells, with one exception (compound 3) (Figure 3B).

Table 2.

Antagonist activity and binding affinity of selected 4H-chromen-4-ones^a

Compd	Competition Binding		Calcium Flux					Chemotaxis
	FPR1-RBL	FPR1-HL60	FPR2-HL60	FPR3-HL60	Fpr1-RBL	mPMN	hPMN	hPMN
	Ki(μM)		IC50 (μM)
1	0.85	0.65	N.A.	N.A.	N.A.	N.A.	1.3	0.027
2	2.5	1.2	N.A.	N.A.	N.A.	N.A.	0.48	0.18
3	2.9	6.1	N.A.	N.A.	N.A.	N.A.	0.15	0.08
4	2.6	1.6	N.A.	N.A.	N.A.	N.A.	1.1	0.20
10	0.09	0.11	N.A.	N.A.	N.A.	N.A.	0.12	0.024
27	3.5	5.7	N.A.	N.A.	N.A.	N.A.	5.1	0.41
36	6.1	0.79	N.A.	N.A.	N.A.	N.A.	0.65	0.05
38	3.8	6.5	N.A.	N.A.	N.A.	N.A.	0.69	0.64
45	3.6	1.7	N.A.	N.A.	N.A.	N.A.	3.5	0.43
46	4.0	4.1	N.A.	N.A.	N.A.	N.A.	2.6	0.32
47	2.0	4.0	N.A.	N.A.	N.A.	N.A.	20.2	0.88
55	(45%)b	(30%)	N.A.	N.A.	N.A.	N.A.	4.5	0.12
68	(45%)	(30%)	N.A	N.A.	N.A.	N.A.	0.9	0.36
73	5.6	5.1	N.A	N.A.	N.A.	N.A.	1.7	1.0

^aAntagonist activity was evaluated as inhibition of Ca²⁺ mobilization induced by 5 nM WKYMVM in FPR2-HL60 cells, 10 nM WKYMVM in FPR3-HL60 cells, 5 nM WKYMVm in Fpr1-RBL cells and murine neutrophils (mPMN), or 5 nM fMLF in human neutrophils (hPMN). Inhibition of chemotactic activity in human neutrophils was evaluated in the presence of 1 nM fMLF.

^bBinding did not exceed 50% inhibition at the highest tested concentration (50 μM), so K_i values for these compounds were not calculated. Chemical names for the selected compounds are: 1, 6-ethyl-2-methyl-3-(1-methyl-1H-benzimidazol-2-yl)-4-oxo-4H-chromen-7-yl acetate; 2, 3-(2-methoxyphenyl)-4-oxo-2-(trifluoromethyl)-4H-chromen-7-yl 2-furoate; 3, 3-(2-chlorophenyl)-4-oxo-2-(trifluoromethyl)-4H-chromen-7-yl thiophene-2-carboxylate; 4, 3-(2-methoxyphenyl)-4-oxo-2-(trifluoromethyl)-4H-chromen-7-yl 2-methylpropanoate; 10, 6-hexyl-2-methyl-3-(1-methyl-1H-benzimidazol-2-yl)-4-oxo-4H-chromen-7-yl acetate; 27, 2-methyl-3-(1-methyl-1H-benzimidazol-2-yl)-4-oxo-4H-chromen-7-yl furan-2-carboxylate, 36, 3-(2-chlorophenyl)-4-oxo-2-(trifluoromethyl)-4H-chromen-7-yl furan-2-carboxylate; 38, 4-oxo-3-phenyl-2-(trifluoromethyl)-4H-chromen-7-yl thiophene-2-carboxylate; 45, 3-(2-methoxyphenyl)-2-methyl-4-oxo-4H-chromen-7-yl furan-2-carboxylate; 46, 3-(2-methoxyphenyl)-2-methyl-4-oxo-4H-chromen-7-yl thiophene-2-carboxylate, 47, 2-methyl-4-oxo-3-phenyl-4H-chromen-7-yl furan-2-carboxylate; 55, 3-(2-methoxyphenyl)-4-oxo-2-(trifluoromethyl)-4H-chromen-7-yl benzoate; 68, 3-(2-methoxyphenyl)-4-oxo-2-(trifluoromethyl)-4H-chromen-7-yl cyclopropanecarboxylate; 73, ethyl 3-(2-methoxyphenyl)-4-oxo-2-(trifluoromethyl)-4H-chromen-7-yl carbonate.

N.A.: No activity was observed at the highest tested concentration (50 μM).

Figure 3. Selected antagonists compete with WKYMVm-FITC for binding to FPR1.

Panel A. FPR1-RBL cells were incubated with the indicated concentrations of compound 10, followed by addition of WKYMVm-FITC, and bound FITC was analyzed by flow cytometry, as described under Materials and Methods. Values are samples from one experiment that is representative of three independent experiments. Panel B. Plot of IC₅₀ values for inhibition of fMLF-induced Ca²⁺ mobilization in FPR1-HL60 cells versus K_i values reflecting binding affinity to the receptor. IC₅₀ and K_i values were selected from Tables 1 and 2, respectively.

In contrast to the other compounds tested, compound 94 did not displace WKYMVm-FITC (data not shown), again supporting the conclusion that inhibition of fMLF-induced Ca²⁺ mobilization in FPR1-HL60 cells by this isoflavone involves an FPR1-independent mechanism.

3.3. Selectivity of FPR1 Antagonists

The selected FPR1 antagonists (Table 2) were evaluated for their ability to activate and inhibit Ca²⁺ mobilization in FPR1-, FPR2, or FPR3-transfected HL60 cells, Fpr1-transfected RBL cells, and CXCR1-transfected RBL cells, as well as in primary human and murine neutrophils. Importantly, all but one of these competitive antagonists did not directly activate Ca²⁺ flux in any of the cell lines or primary cells, supporting the conclusion that they are competitive receptor antagonists. The one exception was compound 2, which did not have an agonist effect in FPR-transfected cells or human neutrophils but did have agonist activity in murine neutrophils. The reason for this activity is currently not understood; however, we have observed differing response patterns to some agonists and/or antagonists in human versus murine neutrophils, respectively (data not shown).

To evaluate receptor selectivity of the antagonists, FPR2-HL60, FPR3-HL60, Fpr1-RBL, and CXCR1-RBL cells were pretreated with the selected compounds for 30 min, followed by stimulation with WKYMVm (FPR2-HL60 or Fpr1-RBL cells), WKYMVM (FPR3-HL60 cells), or IL-8 (CXCR1-RBL cells). None of the selected FPR1 antagonists inhibited FPR2-, FPR3-, Fpr1-, or CXCR1-dependent Ca²⁺ flux (Table 2), demonstrating specificity for FPR1. Likewise, these compounds also did not inhibit human FPR2-dependent or murine Fpr1-dependent neutrophil responses induced by WKYMVM (Table 2 and Figure 2C) and WKYMVm (Table 2), respectively. Moreover, the selected FPR1 antagonists did not inhibit Ca²⁺ flux induced in human neutrophils by IL-8 (data not shown). Thus, these antagonists are FPR1-specific and do not target human FPR2, human FPR3, mouse Fpr1, or other chemokine GPCRs, such as human CXCR1. On the other hand, all selected antagonists inhibited fMLF-induced FPR1 activation in human neutrophils, further supporting their specificity for FPR1 (Table 2). As an example, a representative dose-response curve for the inhibition of fMLF-induced Ca²⁺ flux in human neutrophils by compound 10 is shown in Figure 4A. Since the IC₅₀ value is relative and depends on agonist concentration, we also evaluated inhibition of fMLF-induced Ca²⁺ flux in human neutrophils by compound 10 at different fMLF concentrations and found that IC₅₀ values for the antagonist were ~0.032, 0.12, and ~2.0 μM, at fMLF concentrations of 1, 5, and 25 nM, respectively (Figure 4A). Furthermore, activity of these FPR1 antagonists in fMLF-stimulated human neutrophils generally had a good linear correlation (r =0.77, n=12) with IC₅₀ values in fMLF-stimulated FPR1-HL60 cells, with only two exceptions (compounds 38 and 47) (Figure 4B). Although naturally occurring isoflavones including genistein, daidzein, afrormosin, formononetin, and 7-hydroxyisoflavone were inactive in FPR1-HL60 cells, these compounds were also tested in human neutrophils. None of these compounds inhibited fMLF-induced Ca²⁺ mobilization in these primary cells at the tested concentrations (≤50 μM) (data not shown).

Figure 4. Inhibition of Ca²⁺ mobilization by selected FPR1 antagonists.

Panel A. Human neutrophils (hPMN) were preincubated with the indicated concentrations of compound 10 for 30 min at 25 °C and then stimulated with 1 nM (●), 5 nM (○), or 25 nM (■) of fMLF or 10 nM WKYMVM (△). The response induced by peptide agonist alone was assigned a value of 100%. Panel B. Plot of IC₅₀ values for inhibition of fMLF-induced Ca²⁺ mobilization in FPR1-HL60 cells versus IC₅₀ values for inhibition of fMLF-induced Ca²⁺ mobilization in the human neutrophils. IC₅₀ values for inhibition of Ca²⁺ flux in FPR1-HL60 cells and human neutrophils were selected from Tables 1 and 2, respectively. Compound numbers are indicated for each symbol. Panel C. Human neutrophils were preincubated with the indicated concentrations of the compound 10 or DMSO for 30 min at room temperature, and chemotaxis toward 1 nM fMLF was measured as described under Materials and Methods. The results are expressed as the number of migrated cells per well (mean ± S.D.; n=3). In the panels A and C, values are samples from one experiment that is representative of three independent experiments.

To further investigate the effects of competitive FPR1 antagonists on human neutrophil function, we also evaluated the ability of these compounds to inhibit chemotaxis. All 14 compounds dose-dependently inhibited fMLF-induced neutrophil migration, with IC₅₀ values of 24 nM to 4.5 μM (Table 2). As an example, a representative dose–response curve for the inhibition of fMLF-induced neutrophil chemotaxis by compound 10 is shown in Figure 4C.

fMLF stimulation results in phosphorylation of ERK1/2 in human neutrophils and FPR1-transfected cells [46, 47]. Thus, we evaluated whether selected FPR1 antagonists altered phosphorylation of ERK1/2 in FPR1-RBL cells. Stimulation of FPR1-RBL cells with fMLF resulted in rapid phosphorylation of ERK1/2 (Figure 5A–C). Pretreatment with two of the most potent FPR1 antagonists (10 and 36) effectively inhibited fMLF-induced phosphorylation of ERK1/2 (Figure 5A,B). This effect was specific for fMLF activation, as compounds 10 and 36 did not significantly alter PMA-induced phosphorylation of ERK in untransfected RBL cells (Figure 5D). Importantly, untransfected RBL cells did not respond to fMLF with ERK1/2 phosphorylation (Figure 5D). Consistent with the relative potency of FPR1 antagonists (Table 1), low activity antagonists 15 and 38 only weakly inhibited ERK1/2 phosphorylation, while non-active compound 35 also had no effect on ERK1/2 phosphorylation induced by fMLF (Figure 5).

Figure 5.

Effect of the FPR1 antagonists on ERK1/2 phosphorylation. FPR1-RBL cells (Panels A, B, and C) or un-transfected RBL-2H3 cells (Panel D) were serum-starved for 4 h and then stimulated with the indicated compounds at 50 μM for 10 min. The samples were then stimulated with 20 nM fMLF for different time points. Phosphorylated ERK1/2 (top panel) as well as total ERK1/2 (middle panel) were determined by Western blotting, as described. Relative ERK1/2 phosphoryaltion level was quantified using ImageJ software (bottom panel). In Panel D, RBL-2H3 cells were pretreated with compounds 10 and 36, as indicated, followed by stimulation with 200 ng/ml PMA for 0, 5, 15 min or 20 nM fMLF for 5 min. The data shown are representative of two independent experiments.

To ensure that the results were not influenced by possible compound toxicity, cytotoxicity of the most potent FPR1 antagonists was evaluated at various concentrations (≤25 μM) in wild-type HL60 cells. None of the active FPR1 antagonists affected cell viability at the highest tested concentrations (data not shown), thereby verifying that these compounds were not cytotoxic, at least during the 18-hr incubation period.

3.4. Molecular Modeling of FPR1 Antagonists

For development of the ligand-binding site template for FPR1 antagonists, four potent small-molecule FPR1 antagonists with different chemical scaffolds were selected from the literature (reviewed in [33]). These molecules were identified previously using high-throughput screening and/or lead optimization and included compounds 1 and 7 [27], 5 [28], and 6 [30]. For all independent conformations generated for these compounds, a search for optimal superimpositions was performed. We found that only three of these compounds (1, 5, and 7) could be overlaid simultaneously with good correspondence of the molecular fields, and the superimposition with the highest similarity score calculated by FieldTemplater was chosen for further investigation as a pharmacophore model/template. This pharmacophore template has three hydrophobic sites (protrusion and two areas I and II), one compact grouping (area III) of H-bond donors or positively charged groups in a proposed receptor, and one region of H-bond acceptors (area IV) or negatively charged centers (Figure 6A). The template is slightly curved in shape with the protrusion where CH₃ and CF₃ groups of compounds 1 and 7 are located together with the methylthioethyl substituent of molecule 5. On this protrusion, four hydrophobic (orange) pharmacophore points are positioned according to the nature of these substituents. Another compact grouping of hydrophobic points lies in area I, where fused nitrogen-containing heterocycles of compounds 1 and 5 are located. Molecule 7 has an ortho-chlorophenyl substituent in this region. A more diffuse group of hydrophobic field points (area II) corresponds to the benzofuran moiety of compound 5, substituted benzene ring of the chromone moiety in molecule 1, and benzyl group of compound 7, which are superimposed within the template. Additionally, several negative (blue) and positive (red) points lay very near heteroatoms and H-bond donors/acceptors (areas III and IV, respectively). A few blue field points are also surrounding area II, because it contains electronegative oxygen atoms.

Figure 6.

Multi-molecule template for FPR1 antagonists and superimposition of the compound 10 on the template and docking pose of compound 10 into homology model of FPR1 binding site. Panel A. Compound 10 (magenta backbone, spherical field points) superimposed onto 3-molecular template (grey backbones, icosahedral field points), developed from compounds 1 (light blue), 5 (pink), and 7 (violet). Field points are colored as follows: blue = electron-rich (negative); red = electron-deficient (positive); yellow = van der Waals attractive (steric); and orange = hydrophobic. Panel B. Model of antagonist docking to FPR1. Geometry of the hydrophobic field surface of the compound 10 (left side) matches to the binding site geometry of FPR1 (right side). FPR1 antagonist can approach the FPR1 binding site, which includes two channels (A and C), two cavities (B and E), and the bottom (D). These key features of the FPR1 binding site are designated in accordance to our previous publication [43] and indicated with light-blue arrows. Matching key ligand areas to key areas of the FPR1 binding site is indicated by green arrows under the Panel B. Surface coloring in the FPR1 was made according to electrostatic properties, whereby negatively and positively charged areas are shown in red and blue, respectively. It should be noted, that red (negatively charged) surface areas of the receptor correspond to red field points obtained with negative probe atom and blue (positively charged) surface areas of the receptor correspond to blue field points obtained with positive probe atom.

A visual inspection of the molecule overlays on the 3-molecule FPR1 template revealed that chromone-derived FPR1 antagonists have very similar alignments on the template, with the R₃ group (CH₃ or CF₃) lying within the protrusion. Substituent R₂ coincides with area II, while a moiety at position 3 of the chromone heterocycle (i.e., benzimidazole or substituted phenyl ring) falls into area I. As an example, the most potent FPR1 antagonist (compound 10) superimposed onto the 3-molecule template is shown in Figure 6A.

Since the geometry and location of the FPR1-binding site were not considered explicitly in the field point methodology described above, a homology model of FPR1 was applied. Several subpockets of the FPR1 binding site, including two channels (A and C), two cavities (B and E), and the bottom (D) were recently described for this model [43], and the same designation of the subpockets (A-D) is used here for comparison (Figure 6B). To determine if these subpockets could accommodate the FPR1 antagonist pharmacophore, receptor-docking poses of compound 10 were determined. A pose with minimal energy score of −114.7 kcal mol⁻¹ had a geometry close to its best-fit configuration in the FPR1 antagonist pharmacophore template. A slight difference was found between conformations of the flexible hexyl group (R₂) in the docking pose and field point superimposition. According to our docking study, the pharmacophore area I with hydrophobic area III are located near the bottom D of the FPR1 binding site, the protrusion is fitted in cavity B, and area II is located in cavity E with the long hexyl chain positioned in channel C (Figure 6B). FPR1 residues corresponding to these areas of the binding pocket were enumerated previously [43].

We next explored optimal alignments of all active and most inactive 4H-chromen-4-ones onto the template with the use of FieldAlign program. The alignments with higher similarity scores were analyzed in terms of molecular moieties coincidence with key regions in the pharmacophore model described above. Software options of the calculations are described in Materials and Methods. The values of similarity scores obtained for active and inactive 4H-chromen-4-ones with different chemical scaffolds are presented in Table 1. According to Field Point methodology [37], the template reflects the main geometrical and electronic features of a receptor binding site, hence active and inactive compounds should differ in their level of similarity, and a statistical analysis of similarity scores using ANOVA was performed in order to estimate the difference between the antagonists and inactive compounds in terms of the score values. The data were analyzed for the total set of compounds and separately for compounds with common scaffolds, with the exception of compounds with scaffold C, because this set had only one active analog (compound 55). Although there was no a linear correlation between the similarity scores and inverse (1/IC₅₀) values of FPR1 antagonist activity (r =0.23), the ANOVA gave a statistically significant difference (p<0.001) between groups of active and inactive compounds with respect to their similarity scores (Figure 7A). Within individual chemical scaffolds classes evaluated, the similarity scores also differed significantly between FPR1 antagonists and inactive analogs (Figure 7B).

Figure 7.

Similarity scores of FPR1 antagonists and inactive analogs in whole set of 4H-chromen-4-ones (Panel A) and in sub-sets of the analogs with related chemical scaffolds (Panel B). Selected molecules with different chemical scaffolds (see Table 1) were superimposed onto 3-molecular template for FPR1 antagonists and similarity score between aligned molecule and the template was calculated using the MVD program. Sub-set of the chemical scaffold C was excluded, because the set contains only one active compound. Position of biochanin A (compound 94) with minimal similarity score with the FPR1 template among active compounds is indicated on the Panel A. Statistical analysis of the similarity scores was performed with the use of ANOVA.

Absence of a correlation between the similarity score and antagonist activity can be partially explained by variations in a binding mode to FPR1. For example, inactive compound 63 has a reverse mode of superimposition onto the template, as compared with its active trifluoromethyl analogue 55, although their level of similarity is equal (Table 1C). Compound 63 does not occupy the protrusion, while antagonist 55 has a CF₃ group (R₃) in this area, which is similar to other active compounds. Clearly, a region corresponding to the protrusion is crucial for antagonist activity. Another important feature influencing activity was the presence of bulky substituents that encountered a wall of the proposed binding site. For example, inactive compound 60 has a large o-methoxy group outside of the template boundaries, although the protrusion of the FPR1 pharmacophore template is occupied by the desired CF₃ group. Besides the global measure of similarity to the template, some local similarity can also be important for emergence of FPR1 antagonist activity. “Local” representation of chemical structures can be encoded by other descriptors, such as frontal polygons [48] and/or atom pairs [49]. Further studies using more sophisticated SAR analysis and comprehensive molecular docking will be necessary to address these questions in the future.

4. Discussion

4H-Chromones are an important class of synthetic and natural compounds that exhibit a broad range of biological activities. Their closely related derivatives are isoflavones, a large family of secondary plant metabolites with several variants of a heterocyclic ring substitution pattern. These compounds, along with their synthetic analogues, possess a wide variety of biological activities, including antiinflammatory, antiproliferative, antiosteoporotic, antihyperglycemic, antifungal, antiviral, antiparasitic, antioxidant, and cardiovascular effects [50–55]. Although reactive oxidant scavenging activity of flavonoids/isoflavones may contribute to modulation of phagocyte functional activity [56, 57], these compounds also interact with other cellular targets, including receptors, enzymes, and other macromolecules (for review [58–60]). Such properties include actions upon peroxisome proliferator-activated receptor isoforms, estrogen receptors, tyrosine kinases, DNA topoisomerase II, and NF-κB activation [61–63]. There are a few reports about regulation of GPCRs by isoflavones, although most describe effects on estrogen receptor 1 (GPER1/GPR30) [64, 65].

Previously, several 4H-chromen-4-one derivatives were identified as small-molecule competitive FPR1 antagonists [26, 27]. In the current work, we report further characterization and development of related analogs with improved potency and FPR1 antagonist activity in functional tests using transfected cells and primary neutrophils. Screening of a library of 96 4H-chromen-4-ones, which are structural derivatives of four known chromone-derived competitive FPR1 antagonists 1-4, resulted in the discovery of novel potent FPR1 antagonists. We also showed that these compounds can compete with FITC-labeled peptide WKYMVm for binding with FPR1 in both FPR1-HL-60 and FPR1-RBL transfected cells. Compound 10 was the most potent FPR1 antagonist and exhibits the highest binding affinity among all known chromone analogs, including previously reported compounds 1-4.

A review of the current literature regarding small-molecule FPR1 antagonists [66] showed that pyrazole-4-carboxamides are even more active at inhibition of fMLF-induced Ca²⁺ flux in human neutrophils, with the most potent compound 6 (see the structure in Figure 1), but competition binding at FPR1 for these compounds was not evaluated [29]. In addition, all independent conformations generated for molecule 6 in our modeling experiments did not exhibit good correspondence with the molecular fields of the other three antagonists with different chemical scaffolds (1, 5, and 7) that were used for building the FPR1 pharmacophore template. Thus, inhibition of fMLF-induced Ca²⁺ flux in human neutrophils by pyrazole-4-carboxamides may involve FPR1-independent mechanisms.

Compound 10 and several additional FPR1 antagonists identified here specifically blocked fMLF-induced responses mediated via FPR1 in FPR1-HL60 cells and human neutrophils, but not responses mediated via FPR2 or FPR3 (in human neutrophils and transfected HL60 cells) or Fpr1 (murine neutrophils and transfected RBL cells). These compounds also did not inhibit IL-8-induced Ca²⁺ mobilization in RBL cells transfected with CXCR1 and human neutrophils. The selected chromone/isoflavone derivatives (Table 2) were also inactive as direct agonists of FPR2 and Fpr1 and did not activate Ca²⁺ flux in murine and human neutrophils. Although there are several examples in which compounds derived from similar chemical scaffolds have opposite functional activities (agonist vs. antagonist) at FPR1 and FPR2 (for example, see [31, 67]), no agonists of human FPR1 with a 4H-chromen-4-one scaffold have been reported. Thus, 4H-chromen-4-one may represent a unique chemical scaffold for development of specific FPR1 antagonists that do not have undesirable off-target effects.

In addition to mobilizing calcium from intracellular stores, fMLF activates the mitogen-activated protein kinases ERK1/2 [46, 47]. We evaluated this response to test the relative potency of selected FPR1 antagonists. Compounds 10 and 36, which exhibited high potency in the Ca²⁺ flux assay, also inhibited fMLF-induced ERK phosphorylation. In comparison, antagonists with low potency in the Ca²⁺ flux assay, such as 15 and 38, or the inactive compound 35 had much less or no effect on ERK1/2 phosphorylation. These results verify the Ca²⁺ flux data and together established a dose-effect relationship for these FPR1 antagonists.

Comprehensive SAR analysis of all 100 analogs was performed mainly at positions 2 and 7 of the chromone scaffold due to the apparent importance of these regions of the molecules. Our analysis suggests the importance of a small hydrophobic group (CH₃/CF₃) at position 2 of the heterocycle (R₃ functionality) for FPR1 antagonist activity. Substitution at position 7 of the chromone scaffold also had effects on antagonist activity, but a wider range of modifications was tolerated. Although several substituents at positions 6 and 2′ of the chromone and isoflavone scaffolds were evaluated, modification of these regions will be necessary for further optimization of chromone-based FPR1 antagonists.

Our molecular modeling showed a noticeably higher degree of similarity of the active molecules to the pharmacophore model, which was developed based on the structure of previously reported potent FPR1 antagonists with diverse chemical scaffolds, including chromone 1, methionine-derived benzimidazole 5, and pyrazolo[1,5-a]pyrimidin-7(1H)-one 7. This reflects the main hypothesis of molecular recognition using a field point approach [38]: molecules of different chemical groups have the same biological action if they have similar geometrical arrangement of key pharmacophore features, including hydrophobic, electronegative, or electropositive features. Indeed, a significant statistical difference between groups of FPR1 antagonists with different sub-scaffolds and their inactive derivatives with respect to their similarity scores compared to the 3-molecule FPR1 antagonist pharmacophore was observed. The docking study of the most potent FPR1 antagonist 10 showed that the subpockets of FPR1 binding site could also fully accommodate the FPR1 antagonist pharmacophore. Further studies are clearly required to define the nature of molecular interactions between the chromone/isoflavone based ligands and FPR1 binding site.

Although biochanin A (compound 94) inhibited fMLF-induced Ca²⁺ flux in FPR1-HL60 cells and human neutrophils, this natural occurring isoflavone did not displace WKYMVm-FITC from FPR1-HL60 cells. In addition, biochanin A and its synthetic 2-trifluoromethyl derivative 95 had direct agonist activity in cell lines and primary neutrophils. Thus, compounds 94 and 95 may activate Ca²⁺ flux via FPR-independent pathways, which leads to cross-desensensitization of FPR1 and/or cross-talk with the receptor. Indeed, several other synthetic and natural flavonoids (wogonin, apigenin, genistein, kazinol B, and ugonin U) have been reported to activate intracellular Ca²⁺ flux in various cell lines [68, 69] or neutrophils [70, 71]. Clearly, additional research is required to define the mechanisms of biochanin A action on neutrophils.

In the present studies, we found that genistein did not antagonize fMLF-induced Ca²⁺ mobilization in FPR1-HL60 cells and human neutrophils. Thus, the previously reported inhibitory effect of genistein on fMLF-stimulated degranulation, aggregation, and reactive oxygen species (ROS) production in human neutrophils [47, 72, 73] may be related to pathways downstream of Ca²⁺-mobilization, such as tyrosine kinase-dependent signaling [46, 47] and ROS scavenger activity [74]. Indeed, the broad modulatory activity of genistein on neutrophil activity suggests a non-receptor mode of action. For example, genistein inhibited neutrophil adhesion/adherence activated by agonists of protease-activated receptor 2 (PAR2) [75] or TNF [76]. Moreover, pretreatment of neutrophils with genistein potentiated fMLF-induced cell transmigration [77]. Another isoflavone, afrormosin, inhibited fMLF-stimulated neutrophil degranulation at very high concentrations only (IC₅₀ ~67 μM) [78]. Thus, our current studies suggest that FPR1-independent molecular mechanisms could play a role in the previously reported inhibition of fMLF-induced neutrophil responses by genistein and afrormosin.

Although isoflavones are abundant in the members of the Fabaceae (i.e., Leguminosae) [79], these compounds were also isolated from species belong to other families. Dozens of novel isoflavonoids are discovered every year (for review, see [79]) and further broad screening of natural isoflavones for FPR antagonist activity and ant-inflammatory activity should now be considered. Furthermore, previously reported inhibitory effects of 4H-chromen-4-ones and related flavonoids (e.g., [80]) on fMLF-induced neutrophil responses should be reevaluated in light of our current findings because their biological effects may involve FPR1 antagonist activity.

In conclusion, we have identified a number of specific, competitive FPR1 antagonists with isoflavone backbones. Several of these 4H-chromen-4-one derivatives could represent important leads for therapeutic development focused on FPR1 function and attenuation of neutrophil-mediated inflammatory diseases. These compounds can also serve as scaffolds for the development of additional potent and selective FPR1 antagonists. Furthermore, characterization of this class of antagonists and analysis of additional derivatives should provide important clues to understanding the molecular structural requirements of FPR1 antagonists.