Discovery of GPR183 Agonists Based on an Antagonist Scaffold

Abstract The G protein‐coupled receptor GPR183/EBI2, which is activated by oxysterols, is a therapeutic target for inflammatory and metabolic diseases where both antagonists and agonists are of potential interest. Using the piperazine diamide core of the known GPR183 antagonist (E)‐3‐(4‐bromophenyl)‐1‐(4‐(4‐methoxybenzoyl)piperazin‐1‐yl)prop‐2‐en‐1‐one (NIBR189) as starting point, we identified and sourced 79 structurally related compounds that were commercially available. In vitro screening of this compound collection using a Ca2+ mobilization assay resulted in the identification of 10 compounds with agonist properties. To enable establishment of initial structure‐activity relationship trends, these were supplemented with five in‐house compounds, two of which were also shown to be GPR183 agonists. Taken together, our findings suggest that the agonist activity of this compound series is dictated by the substitution pattern of one of the two distal phenyl rings, which functions as a molecular efficacy‐switch.


Supplementary Figure and
1-7 Table S1 8-9 Experimental Section 10-14 1 H and 13 C NMR Spectra 15-21 __________________________________________________________________________________ Figure S1 (below). Structures of the 79 screened compounds (9-87), along with the reference antagonist NIBR189 (5), which is shown in black with the two-atom C=C spacer highlighted in grey. The structural differences between each compound 9-87 and the reference compound 5 are shown in red. The compounds are divided into active (9-18) and inactive  compounds, which in turn are grouped by type of two-atom spacer (C=C, C-C, C-O, C-S, C-N).  Table S1. SMILES and vendor IDs for the 79 screened compounds 9-87 ( Figure S1).

Experimental Section
In silico substructure screen The Enamine Screening Collection, containing 2.776.742 in stock available compounds from Enamine Ltd. (Kiev, Ukraine) was used in a substructure screen using the Markush structure 7 shown in Figure 1B as query. SurflexSim (version 2.5) was used with the optionspscreen to conduct the screening. The results were filtered based on chemical criteria (molecular weight < 500 g/mol, number of rotatable bonds < 10, logP < 5) and clustered based on a Tc threshold of 0.2. One compound was selected from each cluster. Finally, 79 compounds were manually selected for experimental screening.

Tissue culture and transfection
Chinese hamster ovary (CHO-K1) cells (ATCC® CCL-61) were grown at 5% CO2 and 37 °C in RPMI 1640 supplemented with 10% (v/v) FBS, 2 mM GlutaMAX (Gibco), 180 units ml -1 penicillin and 45 μg ml -1 streptomycin and split once confluent (around every 3 rd day). Transient transfection of CHO-K1 cells was performed by lipofection using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. Lipofectamine™ 2000 reagent and Opti-MEM were purchased from Invitrogen (Glasgow, UK). Human GPR183 contained an N-terminal M1-FLAG and was inserted into a modified pcDNA3 vector which contained an upstream sequence encoding a hemagglutinin signal peptide fused to the FLAG tag. The promiscuous chimeric G protein GαΔ6qi4myr (abbreviated Gqi4myr) was kindly provided by Evi Kostenis (Rheinische Friedrich-Wilhelm University, Bonn, Germany) and the CAMYEL sensor was kindly provided by Jonathan Javitch (Columbia University, New York, US). EC50 and pEC50 values determined in Prism using non-linear curve fit.
BRET cAMP 2 days prior to assay, CHO-K1 cells were seeded 500.000/well in 6-well plates. The following day cells were transfected using 6 µl Lipofectamine, 160 ng receptor DNA and 840 ng CAMYEL DNA in each well. On assay day, cells from each well were harvested in 3 ml PBS with 5mM glucose and added to a 96-well plate, 84 µl in each. Just before ligand addition, 10 µl coelenterazine (5 µM in well) was added followed by 1 µl 100X ligand in DMSO. 5 minutes post ligand addition 5 µl forskolin (10 µM in well) was added and at 40 min post ligand addition, luminescence (475 nm) and fluorescence (525 nm) were measured using the EnVision plate reader (PerkinElmer). For the testing of antagonism, the antagonist (100X in DMSO) was added 10 min prior to ligand addition. EC50 and pEC50 values were determined in Prism using non-linear curve fit.
Ca 2+ release assay CHO-K1 cells were seeded in 96-well clear bottom black plates at a density of 2 x 10 4 cells per well. The following day the cells were transfected with GPR183 (30 ng/well) and Gqi4myr (10 ng/well). 24 h after transfection, cells were washed once with 100 μl of HBSS pre-warmed to 37 °C supplemented with 20 mM HEPES (Invitrogen), 1 mM CaCl2, 1 mM MgCl2, and 0.7 mg ml -1 probenicid (Life Technologies, Thermo Fisher Scientific)). Afterwards, cells were incubated for 60 min at 37 °C and 5% CO2 with 50 μl of pre-warmed loading buffer per well (wash buffer supplemented with 0.2% (v/v) Fluo-4 (Life Technologies, Thermo Fisher Scientific)) and covered from light. After incubation, cells were washed twice with 75 μl of pre-warmed wash buffer, and 100 μl of 37 °C pre-warmed cell medium was added. Ligands were added and plates measured using a FlexStation 3 Multi-Mode Microplate Reader (Molecular Devices) with excitation and emission wavelengths of 485 and 525 nm, respectively. Determinations were made in duplicates.

Chemistry -General
All reagents were of commercial grade and were used as received without further purification. Anhydrous reactions were carried out in flame-dried glassware under argon atmosphere. Dry chromatographic grade dichloromethane (DCM), tetrahydrofurane (THF) and dimethylformamide (DMF) were obtained from a Waters SG solvent purification system. Thin layer chromatography (TLC) Silica gel 60 F254, Merck pre-coated plates were used, visualized under UV light (254 or 365nm), developed in the system stated for each compound. Flash chromatography of compounds was performed using silica gel 60 (40-64 μm), where loading of the compounds was done after dry mixing with Celite.
Nuclear Magnetic Resonance (NMR) spectra were recorded on 400 or 600 MHz Bruker instruments ( 1 H NMRs were obtained at 400 or 600 MHz and 13 C NMRs were obtained at 100 or 151MHz). The obtained FID files were processed with Mnova 14 software. Spectra are calibrated relative to residual solvent peaks. Multiplet patterns are designated the following abbreviations, or combinations of these: m -multiplet, s -singlet, d -doublet, t -triplet, qquartet, p -pentet, h -sextet.
Mass analysis by matrix-assisted laser desorption/ionization high-resolution mass spectrometry (MALDI-HRMS) was performed on a QExactive Orbitrap mass spectrometer (Thermo Scientific, Bremen, Germany) equipped with a SMALDI5 ion source (TransMIT GmbH, Giessen, Germany). The sample was analyzed in the positive ion mode using a peak from the DHB matrix for internal mass calibration whereby a mass accuracy of 2 ppm or better was achieved. The samples (1 mg) were mixed with a 20 mg/mL solution of 2,5-dihydroxybenzoic acid in MeOH as MALDI matrix, 2 µL of the mixture was deposited on a 30-well glass plate (Electron Microscopy Sciences, Hatfield, PA, USA) and analysis followed upon evaporation.