High-Throughput Screening Campaign Identified a Potential Small Molecule RXFP3/4 Agonist

Relaxin/insulin-like family peptide receptor 3 (RXFP3) belongs to class A G protein-coupled receptor family. RXFP3 and its endogenous ligand relaxin-3 are mainly expressed in the brain with important roles in the regulation of appetite, energy metabolism, endocrine homeostasis and emotional processing. It is therefore implicated as a potential target for treatment of various central nervous system diseases. Since selective agonists of RXFP3 are restricted to relaxin-3 and its analogs, we conducted a high-throughput screening campaign against 32,021 synthetic and natural product-derived compounds using a cyclic adenosine monophosphate (cAMP) measurement-based method. Only one compound, WNN0109-C011, was identified following primary screening, secondary screening and dose-response studies. Although displayed agonistic effect in cells overexpressing the human RXFP3, it also showed cross-reactivity with the human RXFP4. This hit compound may provide not only a chemical probe to investigate the function of RXFP3/4, but also a novel scaffold for the development of RXFP3/4 agonists.

Although RXFP3 is attractive for drug discovery, most available ligands for this receptor are restricted to peptidic analogs of relaxin-3 [4,17,[29][30][31][36][37][38][39]. To date, only a limited number of small molecule modulators have been identified, including a series of small molecule RXFP3/4 dual agonists [28] and a selective positive allosteric modulator [40]. Since peptidic ligands for RXFP3/4 are hard to make and no selective RXFP3 agonist has been found, we conducted a high-throughput screening (HTS) campaign against 32,021 synthetic and natural product-derived compounds using a cAMP measurementbased method. Only one compound (WNN0109-C011) that cross-reacted with RXFP4 was identified following primary screening, secondary screening and dose-response studies.

Assay Validation
As RXFP3 is coupled to G i protein to trigger the inhibition of cAMP accumulation, we applied forskolin to produce cAMP signals for the detection of inhibitory effects of small molecule compounds. The optimal concentration of forskolin was selected from dose-response curves generated at different cell densities that were also optimized during assay development. Finally, 2000 cells/well and 25 µM forskolin were used for the HTS campaign. Under the optimized conditions, the pEC 50 value of R3/I5 (a RXFP3 agonist peptide) was 9.46 ± 0.02 (Figure 1a), consistent with that documented in the literature [37].
Although RXFP3 is attractive for drug discovery, most available ligands for this receptor are restricted to peptidic analogs of relaxin-3 [4,17,[29][30][31][36][37][38][39]. To date, only a limited number of small molecule modulators have been identified, including a series of small molecule RXFP3/4 dual agonists [28] and a selective positive allosteric modulator [40]. Since peptidic ligands for RXFP3/4 are hard to make and no selective RXFP3 agonist has been found, we conducted a high-throughput screening (HTS) campaign against 32,021 synthetic and natural product-derived compounds using a cAMP measurementbased method. Only one compound (WNN0109-C011) that cross-reacted with RXFP4 was identified following primary screening, secondary screening and dose-response studies.

Assay Validation
As RXFP3 is coupled to Gi protein to trigger the inhibition of cAMP accumulation, we applied forskolin to produce cAMP signals for the detection of inhibitory effects of small molecule compounds. The optimal concentration of forskolin was selected from dose-response curves generated at different cell densities that were also optimized during assay development. Finally, 2000 cells/well and 25 μM forskolin were used for the HTS campaign. Under the optimized conditions, the pEC50 value of R3/I5 (a RXFP3 agonist peptide) was 9.46 ± 0.02 (Figure 1a), consistent with that documented in the literature [37].

Assay Performance
In the HTS campaign with cAMP accumulation assay, the positive signals were evoked by R3/I5, a chimeric peptidic agonist for RXFP3, while the negative signals (background) were assessed in the presence of 1% DMSO. As shown in Figure 1b, coefficient of variation (CV) values were 7.18% (for R3/I5) and 10.37% (for background), respectively. Z' factor calculated was 0.54 with an S/B ratio of 3.18. These parameters are indicative of a high assay quality that is suitable for HTS [41].

Assay Performance
In the HTS campaign with cAMP accumulation assay, the positive signals were evoked by R3/I5, a chimeric peptidic agonist for RXFP3, while the negative signals (background) were assessed in the presence of 1% DMSO. As shown in Figure 1b, coefficient of variation (CV) values were 7.18% (for R3/I5) and 10.37% (for background), respectively. Z' factor calculated was 0.54 with an S/B ratio of 3.18. These parameters are indicative of a high assay quality that is suitable for HTS [41].

Binding Pose Prediction
To reveal the potential binding mode of WNN0109-C011 in RXFP3, we performed molecular docking and molecular dynamics (MD) simulation studies. As shown in Figure  4, WNN0109-C011 stably occupied the orthosteric binding pocket formed by transmembrane helices (TMs) 2, 3, 5, 6, and 7. Specifically, the terminal group methanesulfonamide inserted deeply to the ligand-binding pocket with the formation of a hydrogen bond with K271 5.42 (superscript refers to Ballesteros-Weinstein numbering [43]), as seen from the measured minimum distance between the specific atoms of K271 and WNN0109-C011 (Figure 4c). The middle hydroxy pointed toward the side chains of E141 2.63 with the formation of one hydrogen bond. Meanwhile, the adjacent charged amide was stabilized by the stacking interaction with W138 2.60 . Another terminal group benzene located closely to TM5 with the formation of rich stacking interactions with Y267 5.38 and H268 5

Binding Pose Prediction
To reveal the potential binding mode of WNN0109-C011 in RXFP3, we performed molecular docking and molecular dynamics (MD) simulation studies. As shown in Figure 4, WNN0109-C011 stably occupied the orthosteric binding pocket formed by transmembrane helices (TMs) 2, 3, 5, 6, and 7. Specifically, the terminal group methanesulfonamide inserted deeply to the ligand-binding pocket with the formation of a hydrogen bond with K271 5.42 (superscript refers to Ballesteros-Weinstein numbering [43]), as seen from the measured minimum distance between the specific atoms of K271 and WNN0109-C011 (Figure 4c). The middle hydroxy pointed toward the side chains of E141 2.63 with the formation of one hydrogen bond. Meanwhile, the adjacent charged amide was stabilized by the stacking interaction with W138 2.60 . Another terminal group benzene located closely to TM5 with the formation of rich stacking interactions with Y267 5.38 and H268 5 for RXFP4, and S159 3.29 , S163 3.33 , H268 5.39 and V375 7.39 for RXFP3), which may explain the reduced potency of WNN0109-C011 at RXFP4.
In this study, we established an HTS assay by detecting cAMP accumulation in CHO cells stably expressing RXFP3, a method that was used to screen a collection of 32,021 synthetic and natural product-derived compounds against RXFP3. The RXFP3 agonist peptide R3/I5 was used as a positive control to optimize the assay system, and our results (Figure 1) showed that the assay parameters, such as CV, S/B ratio and Z' factor, were all of high quality. Although the HTS campaign identified 23 initial hits (0.07%) most of them failed to show agonistic activities in secondary screening or dose response studies, leading
In this study, we established an HTS assay by detecting cAMP accumulation in CHO cells stably expressing RXFP3, a method that was used to screen a collection of 32,021 synthetic and natural product-derived compounds against RXFP3. The RXFP3 agonist peptide R3/I5 was used as a positive control to optimize the assay system, and our results (Figure 1) showed that the assay parameters, such as CV, S/B ratio and Z' factor, were all of high quality. Although the HTS campaign identified 23 initial hits (0.07%) most of them failed to show agonistic activities in secondary screening or dose response studies, leading to only one confirmed hit (WNN0109-C011). This compound, with no structural resemblance to the hits identified in the HTS campaign against human (RXFP4) [42], was subsequently found to inhibit forskolin-elicited cAMP responses in both hRXFP3-CHO and hRXFP4-CHO cells, without affecting cAMP levels in parental CHO cells, demonstrating its dual agonist (RXFP3/4) potential.

Compound Library
A collection of 32,021 synthetic and natural product-derived compounds stored at the Chinese National Compound Library [46] were used in this study. The structural diversity covers heterocycles, lactams, sulfonates, sulfonamides, amines and secondary amides, etc. All the compounds are highly pure, and the stock, pre-solubilized in 100% DMSO solution, was applied to the primary screening.

HTS Campaign
The HTS campaign was carried out in white 384-well plates by measuring the inhibition of forskolin-stimulated cAMP accumulation in hRXFP3-CHO cells. Briefly, the positive control (5 nM R3/I5) and negative control (1% DMSO) were located at the 64 wells of the outer four columns on both sides with 32 replicates, respectively. The individual screening compounds were loaded at the center twenty columns with an average final concentration of 10 µM in a single-dose format without replicates. Time-resolved fluorescence resonance energy transfer (TR-FRET) signals were read by an EnVision multilabel plate reader (PerkinElmer). Z' factor, coefficient of variation (CV) and signal-to-background ratio (S/B) were calculated according to the literature [41]. Positive control was regarded as 100% activation, while negative control represented no activation. Compounds showing greater than 50% activation were considered hits, and secondary screening using the same assay was performed to eliminate false positives.

cAMP Assay
cAMP accumulation assay was performed in hRXFP3-CHO cells at a density of 5 × 10 5 cells/mL. The cells were treated with different concentrations of compounds plus 25 µM forskolin and 500 µM IBMX for 30 min at room temperature (RT). cAMP-d2 conjugate and cryptate conjugate working solutions were then diluted and added to the plates separately. Plates were incubated at RT for 60 min and the fluorescence intensity measured at 620 nm and 665 nm by an EnVision multilabel plate reader. The parental CHO cells were applied to examine off-target effect, while hRXFP4-CHO cells were utilized for selectivity assessment.

Europium-Labelled H3 B1-22R Competition Binding
CHO-K1 cells stably transfected with RXFP3 were plated onto pre-coated poly-L-lysine 96-well plates for the binding assay performed using 5 nM europium (Eu)-H3 B1-22R as previously described [46] with triplicate determinations within assay and a minimum of three independent experiments. Data are presented as means ± SEM. pKi values were calculated using one-site fit Ki and a Kd value of 26 nM for H3 B1-22R [47].

Molecular Docking
Molecular docking was performed with Schrödinger Suite 2018−1 (New York, NY, USA). Homology model of active hRXFP3 retrieved from the GPCRdb website [48] was built mainly based on the X-ray structure of the active angiotensin II type 1 receptor (PDB code: 6DO1) with a sequence similarity of 34%. Receptor structure and ligands were prepared by Protein Preparation Wizard and LigPrep in Schrödinger Suite 2018−1 with default settings, respectively. Receptor grid was generated by Receptor Grid Generation tool in the Glide (New York, NY, USA) [49]. Ligands were docked following the XP scoring function. The initial docking poses were visualized and analyzed, while ten poses with best docking scoring scores were subjected to Prime MM-GBSA optimization by Prime MM-GBSA module (New York, NY, USA). Finally, by visual inspection of the optimized docking pose and considering the XP docking scores, two predicted binding poses of WNN0109-C011 in hRXFP3 were obtained for MD simulations.

Molecular Dynamics Simulation
MD simulation study was performed by Gromacs 2020.1 (Groningen, The Netherlands) [50]. The hRXFP3 structure was capped using CHARMM-GUI Membrane Builder v3.2.2 (Bethlehem, PA, USA) [51]. Residue D128 2.50 was deprotonated, while other titratable residues were left in their dominant state at pH 7.0. The WNN0109-C011−RXFP3 complexes were embedded in a bilayer composed of 201 POPC lipids and solvated with 0.15 M NaCl in explicit TIP3P waters using CHARMM-GUI Membrane Builder v3.2.2 (Bethlehem, PA, USA) [51]. The CHARMM36-CAMP force filed [52] was adopted for protein, lipids and salt ions. WNN0109-C011 was modelled with the CHARMM CGenFF small-molecule force field [53]. The Particle Mesh Ewald (PME) method was used to treat all electrostatic interactions beyond a cut-off of 12 Å and the bonds involving hydrogen atoms were constrained using LINCS algorithm [54]. The complex system was first relaxed using the steepest descent energy minimization, followed by slow heating of the system to 310 K with restraints. The restraints were reduced gradually over 20 ns. Finally, 250 ns restrain-free production run was carried out, with a time step of 2 fs in the NPT ensemble at 310 K and 1 bar using the Nose-Hoover thermostat and the semi-isotropic Parrinello-Rahman barostat [55], respectively.

Statistical Analysis
Dose-response data were analyzed with Prism software (GraphPad, San Diego, CA, USA) using a sigmoidal model with variable slope. Data were expressed as the means ± S.E.M of at least three independent experiments.

Conclusions
In summary, we have developed an optimized HTS method by measuring cAMP production to identify novel RXFP3 agonists. Our HTS campaign and follow-up hit characterization experiments identified WNN0109-C011 as a potential agonist for both RXFP3 and RXFP4, without any agonistic effect on parental CHO cells. Discovery of specific RXFP3 agonists is still challenging along with an urgency of in-depth evaluation of its in vivo functions associated with the druggability of this target. Determination of  Table S1: Hits from primary screening; Table S2: Binding of europium-labelled H3 B1-22R to RXFP3 in competition with various small molecules agonists.