Discovery, Synthesis, and In Vitro Characterization of 2,3 Derivatives of 4,5,6,7-Tetrahydro-Benzothiophene as Potent Modulators of Retinoic Acid Receptor-Related Orphan Receptor γt

Abstract

Retinoic acid receptor-related orphan receptor γt (RORγt) is a nuclear receptor that is expressed in a variety of tissues and is a potential drug target for the treatment of inflammatory and auto-immune diseases, metabolic diseases, and resistant cancer types. We herein report the discovery of 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene modulators of RORγt. We also report the solubility in acidic/neutral pH, mouse/human/dog/rat microsomal stability, Caco-2, and MDR1-MDCKII permeabilities of a set of these derivatives. For this group of modulators, inverse agonism by steric clashes and push–pull mechanisms induce greater instability to protein conformation compared to agonist lock hydration. Independent of the two mechanisms, we observed a basal modulatory activity of the tested 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene toward RORγt due to the interactions with the Cys320-Glu326 and Arg364-Phe377 hydrophilic regions. The drug discovery approach reported in the current study can be employed to discover modulators of nuclear receptors and other globular protein targets.

Introduction

Retinoic acid receptor-related orphan receptor γt with its globular 12-helix ligand-binding domain (LBD) is a nuclear receptor of a compelling drug target profile in cancer and autoimmune/inflammatory diseases.¹⁻⁸ The receptor has been extensively investigated in the past decade for its structural features, biological functions, and most importantly for its candidacy as a drug target.^1,2,7−16 Several X-ray crystallography and NMR structural studies have identified numerous ligands, agonists, inverse agonists, and antagonists, with over 100 published crystal structures.^1,9,16−18 These studies offered an opportunity for researchers to generate novel ligand design concepts. However, during development, most of the preclinical and clinical candidates were rejected based on safety and side effect profiles.^{3,8,12,13,19,20} Despite these unsuccessful trials, the interest in developing RORγt modulators did not fade away, and more studies were released reporting vastly diverse binding modes of new ligands that differ significantly in their chemical structures, in addition to the identification of allosteric and sub-orthosteric pockets.^9,21−23 It has become apparent that both typical and so-called atypical ligands are much more diverse and can confer varied conformational changes at the structural level in terms of protein stability, helix–helix interactions, LBD coactivator, or corepressor recruitment. These are differentially reflected in biological function in vivo and comprise a large group of modulators with a corresponding range of potency and efficacy.^21,22,24,25 Crystallography studies on RORγt have revealed substantial structural information about the binding pocket and its key features.^9,16,17,25 The orthosteric ligand-binding pocket of RORγt carries special characteristics in its cavity size, contour, and lipophilicity. Its size is one of the largest among all nuclear receptors. The volume of the pocket is 940 ų.^9,16,17,23 The lipophilic volume of the receptor is 575 ų (61%), while hydrophilic, basic, and acidic volumes are 161 ų (16%), 178 ų (19%), and 26 ų (3%), respectively.^23,25 The pocket preserves plasticity regardless of the molecular volume occupied with agonists/inverse agonists, which vary in size from 200 to 800 g/mol.^9,16,17 One of the key substructures of the orthosteric pocket is the agonist lock His479-Tyr502-Phe509, which lays on the outer end of the ligand pocket and the back wall of coactivator-/corepressor-binding pocket and is considered a key target for synthetic agonists.^{9,21,22,24,26} The interruption/dissociation of the His479-Tyr502 hydrogen bond results in a subsequent disruption of the H11–H12 helix–helix interaction and eventually further impedes H12 coactivator interaction. On the other hand, conformational changes in H3 and H5 helices, which interface with both ligand-binding pocket and the back wall of coactivator-binding pocket, can disrupt protein–coactivator interactions. Such disruptions lead to the reduction or complete loss of RORγt activity.^{9,16,17,21,22,24,27} Molecular dynamics (MD) simulation studies have shown that the chemical structure of an inverse agonist and its mode of binding can greatly affect protein dynamics and result in great variability in protein folding and conformation of the helices in solution.^21,22,24 Such changes can be crucial for activity, and their early prediction can improve computer-aided drug design and selection of candidates for further testing.^22,24 Variabilities seen in the degree of water trapping during crystallization of RORγt ligand–protein complexes indicate a variability of pocket hydration induced by the binding of different agonists.^9,24 Hydration, on the other hand, can induce great distortion of RORγt folding that cannot be captured by crystallography.²⁴ In a few examples, where the rmsds of RORγt crystals greatly deviate (>3 Å) from their known average position, the deviation has been linked to the binding of larger inverse agonists.^9,16,28 Other than the deviations arising from the ligand size, the rmsds between RORγt crystals are subtle and cannot be correlated with other types of ligand–protein binding interactions without a thorough analysis and additional experimentation. Because of the those limitations, MD simulation studies have become a useful tool to characterize and investigate the binding modes of agonists/inverse agonists.¹⁶ For example, direct and controlled comparisons between ligands from different classes could be made using MD simulations where the stabilization/disruption of the H11–H12 triangular net was better judged during motion.²⁴ Cholesterol derivatives such as 22, 24, 25, and 27-hydroxycholesterols are the early identified classes of ligands that bind the RORγt orthosteric pocket. Steroid conformations of most of the identified ligands are trans–trans–trans, and they behave as agonists.^7,9,12,29 7α-Hydroxycholesterol, 7β-hydroxycholesterol, and 7-ketocholesterol were considered inverse agonists; however, subsequent studies showed that they behave as agonists or partial agonists, like all trans cholesterol derivatives.^12,30,31 Cardiac glycosides and ursolic acid, on the other hand, are inverse agonists and have a cis conformation at ring A or at both ring A and D of the sterol core.^1,32 The range of synthetic RORγt agonists/inverse agonists is large and includes sulfonamides, pyrazole amides, biaryl amides, tertiary amines, benzoxazepines, tetrahydroquinoline sulfonamides, tetrahydro-naphthyridine derivatives, n-thiazolyl acetamides, quinoline sulfonamides, n-aryl sulfonyl indolines, phenyl (3-phenylpyrrolidin-3-yl) sulfones, hexafluoro isopropanol aryl sulfonamides, imidazopyridines, pyrimidines, quinazolinedione derivatives, thiazole-bis-amides, triazolopyridines, imidazolopyridines, carbazole carboxamides, diphenyl propanamides, 1,3,5-triazine derivatives, tetrahydrobenzo[4,5]thieno[2,3-d]pyrimidines, triazolopyridines, dihydroimidazole-tethered imidazolinethiones, sulfoximines, phenyl glycinamides, oxadiazole-substituted thiazoles, and n-indanyl benzamides.^{18,26,27,33−57} Herein, we report the discovery of 4,5,6,7-tetrahydro-benzothiophene modulators of RORs. We show the results of “in silico/in vitro” screening, molecular docking, and MD simulation studies in comparison with selected reference RORγt modulators from different classes. In the current study, we tailored a drug discovery cycle that is suitable for nuclear receptors and similar globular protein drug targets, in which the processes of the identification, evaluation, and optimization of novel modulators can be all carried out contemporaneously.

Results

Discovery workflow of 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene and results of modeling steps are summarized in Figure 1. The RORγt modulator molecular model was designed starting from 25-hydroxycholesterol (25HC) and using Autodock Vina for ideal pose prediction, as shown in Figure 1A,B. A two-dimensional virtual screen using the molecular model as a fingerprint with the Tanimoto coefficient cutoff of 0.7 returned a library of a thousand of 4,5,6,7-tetrahydro-benzothiophene-based virtual hits. Two- and three-dimensional processing with Ligprep produced 1500 conformers. Ranking conformers based on their rmsd_scr/ref from the best scoring conformer of the top ranked virtual hit in a 2D screen created three clusters (<1, 2–4, and >5 Å). The energy of conformers (E_conf) was 5–15 kcal/mol in the range, Figure 1C. A subsequent three-dimensional phase screen returned 167 virtual hits. Clustering those hits based on their rmsd_align, rmsd_scr/ref, and the relative energy of the conformer in binding position (ΔE_bconf) revealed that the rmsd_align values for the hits were 0.3–1.7 Å. Additionally, there was a variability in binding modes, and the rmsd_scr/ref was 0–7 Å, while their ΔE_bconf was in the range of 0–40 kcal/mol, Figure 1D. Molecular docking of low rmsd_scr/ref, rmsd_align, and ΔE_bconf hits using Glide HTVS and SP docking determined the top conformers that aligns with the original design and the ideal pose of the initial Autodock Vina screening, Figure 1E. MD simulations (50 ns) of selected hits showed stability of the protein–ligand complex in their binding poses (rmsd <4 Å), Figure 1F. The hits were more stable through the simulation compared to gsk2981278 in its proposed binding pose.⁵⁸ Calculated ADME/PK revealed drug-like properties of many of the top hits, Figure 1G. The stepwise computer-aided evaluation process resulted in a selected set of 23 virtual hits, which varied in MW from 350 to 515 g/mol. The selected hits, except 3, followed Lipinski’s rule of five and had high oral bioavailability >0.55 and their calculated ilogP⁵⁹ or logD (pH = 7.4) was 0.4–5. The TPSA⁶¹ of the 23 hits was 80–140 Ų. The predicted solubility of the hits ranged from poor to soluble using ESOL model logS.⁶⁰ Following synthesis, kinetic solubility evaluation results showed moderate solubility of the compounds except 2, 4, and 23 which had poor solubility (≤1 μM, pH = 7.4). Free-energy perturbation (FEP) results showed slight differences in binding free energies (ΔΔG) of the virtual hits (0.4–5 kcal/mol). This was reflected by closely comparable experimental free binding energies (ΔG_EXP) in fluorescence polarization (FP) competitive RORγt binding assay against 25-NBD cholesterol (25NC) and in time-resolved fluorescence energy transfer (TR-FRET) assay. Furthermore, the lipophilic efficiency (LipE) was calculated using logD and ilogP (LipE_logD and LipE_ilogP), Table 1.

Figure 1.

Discovery workflow of 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene modulators of RORγt. (A,B) Linking and substitution starting from 25HC to design a fingerprint for 2D screening; (C) scatter plot of 1000 virtual hits showing maximum common structure rmsd_scr/ref screening hit (scr) to the reference structure (ref), the energy of conformer (E_conf) and Tanimoto coefficient; (D) scatter plot of the hits of phase screen showing rmsd_scr/ref, relative energy of bound conformer ΔE_bconf, and rmsd_align; (E) predicted active poses of the top virtual hits inside the RORγt orthosteric ligand-binding pocket (PDB ID: 5APH) minimized without crystal water; (F) plots of short (50 ns) MD simulations of synthesis candidates chosen from HTVS, SP, and XP docking screens; (G) radar plot of the calculated ADME/PK properties of an optimized structure of one of the phase screen hits; and (H) synthesis and biological evaluation.

Table 1. Calculated Physicochemical Properties, Kinetic Solubility, Docking Scores, Calculated, and Experimental Binding Free Energies, IC₅₀ Values in TR-FRET Assay and Lipophilic Efficiency of 2,3 Derivatives of 4,5,6,7-Tetrahydro-Benzothiophene Modulators of RORγt.

2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene form non-covalent interactions with various residues of the RORγt orthosteric ligand-binding pocket. Extra precision (XP) docking results show that the 23 compounds of the selected set vary slightly in their binding modes and alignment inside the pocket; however, they were all characterized by hydrogen bonds and/or water bridge contacts between their amide bridges and Cys320-Glu326 and Arg364-Phe377 hydrophilic regions. The saturated tetrahydro-benzene ring substructures were buried in the hydrophobic region of the pocket. Examples of binding poses are shown in Figure 2A–D. With their aromatic rings linked to position 3 of 4,5,6,7-tetrahydro-benzothiophene through amide bridges, compounds 1, 2, 7–11, 13, 16, and 19 form a π–π bonds with His479 and Trp317. In addition, compound 19 forms a hydrogen bond with His479 through its triazole ring. Superposed binding poses of compounds 2, 10, and 13 are shown in Figure 2A. When compound 10, in its predicted pose, was aligned with structurally different natural and synthetic ligands 25HC, gsk2981278, and the inverse agonist iag (Figure S1) described by Saen-Oon et al., 2019,²⁴ it showed good alignment with docked gsk2981278 and iag with rmsd <1 and <2 Å, respectively, based on selected atom pairs, Figure 2B. On the other hand, the ring substituents linked to position 2 of 4,5,6,7-tetrahydro-benzothiophene, except compound 7, either formed a hydrogen bond or were in proximity to form a water bridge with Gln286 such as in case of compounds 9, 10, and 11 that formed hydrogen bonds between their piperidine nitrogen and Gln286. A deviation (>2 Å) from the ideal pose occurred when the 4,5,6,7-tetrahydro-benzothiophene ring was substituted at position 6 with an alkyl group such as in case of compounds 3–6 and 18. This was not reflected by a reduction or an increase in activity in either TR-FRET or FP assay, Table 1. Superposition of compound 3 and 2 shows the extent of this deviation, where the 1,1 dioxide tetrahydro-thiophene ring of 3 forms a hydrogen bond with Gln286 while the methyl-thio-pyrimidine ring linked to position 2 of the 4,5,6,7-tetrahydro-benzothiophene forms a hydrogen bond with Phe377, Figure 2B. This had no effect on activity in TR-FRET or FP assay, Table 1. Reducing the length of the bridge linking ring substitution to position 3 of 4,5,6,7-tetrahydro-benzothiophene ring to one atom as in case of compounds 12, 15, 22, and 23 resulted in a deviation (>2 Å) from ideal pose and a reduction of the activity in non-cellular binding assays, Table 1. Similarly compounds 14 and 17 deviated from the ideal pose, and there was no interaction with the agonist lock; however, they formed hydrogen bonds through their hydroxyl groups with Gln286. Their poses are stabilized by intramolecular hydrogen bond between their hydroxyl groups and the amide bridge of position 2 of 4,5,6,7-tetrahydro-benzothiophene. The activity of the compound 14 in TR-FRET or FP was not lost in such a pose despite that there was no significant interaction with the agonist lock from the ring linked to position 3 of 4,5,6,7-tetrahydro-benzothiophene or other substructures of the molecules, Table 1 and Figure 2D. Compound 17, on the other hand, showed a lower docking score and an increase in deviation (>2 Å) toward the hydrophobic region of the pocket. This was reflected by a reduced activity in both TR-FRET and FP assays, Table 1. Finally, increasing the amide bridge linked to position 2 of 4,5,6,7-tetrahydro-benzothiophene ring such as in case of compounds 8 and 16 had a slight effect on binding poses and no effect on activity in binding assays.⁶⁴

Figure 2.

Predicted active poses of the 4,5,6,7-tetrahydro-benzothiophene derivatives superimposed with natural and synthetic ligands of RORγt (PDB ID: 5APH): (A) compounds 2 (orange), 10 (cyan), and 13 (green) (Ligrmsd <1 Å) in their active pose superimposed on 25HC; (B) 10 (cyan), superimposed on gsk2981278 and iag, forms a hydrogen bond with Gln286:δCO and π–π stacking between its phenyl group and His479; (C) 3 (cyan) superimposed on 2 (orange) and showing a deviation (Ligrmsd > 2 Å) of the predicted pose and forming hydrogen bond with Gln286:δCO through the 3-amide substitution ring and a hydrogen bond through the 2-amide substitution ring; (D) 14 (fuchsia) superimposed on 13 (green) showing a deviation of the 3-amide substitution ring and in proximity with Gln286:δCO and Hid323:CO (Ligrmsd = 1.77 Å).

The stability of 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene ligand protein complexes with RORγt, from their initial binding poses shown in Figure 2, was confirmed with 50 and 250 ns MD simulations. Compared to 25HC–, iag– and ago– (described by Saen-Oon et al., 2019) RORγt, 2–, 3–, 10–, and 14–RORγt complexes were equally stable over the length of trajectory, and the change in ligand protein rmsd was <1 Å. However, compound 13 moved away from its initial binding position <2 Å, and no additional shift beyond that occurred at 40 ns, which was comparable with gsk2981278–RORγt complex, Figure 3A. The observed magnitude of H11–H12 atomic fluctuations in 4,5,6,7-tetrahydro-benzothiophene–RORγt complexes was 0.1–0.5 Å higher compared to all the reference ligands, Figure 3B. The magnitude of fluctuations in the H12 region is >1.5 Å in case of 3–, 13–, and 14–RORγt complexes compared to apoprotein, and all ligand–protein complexes were evaluated except 25HC–RORγt complex which showed an equally elevated fluctuation at the H12 region. Additionally, the observed magnitude of fluctuations in the H6-loop-H7 region, Leu391-Glu395 residues, of 2– and 10–RORγt complexes was 0.3–1 Å higher compared to apoprotein and all other simulated ligand–RORγt complexes, while the fluctuations in this region were only slightly elevated in case of 14–RORγt complexes compared to apoprotein. Similarly, 25HC–, ago–, and gsk2981278–RORγt complexes showed a moderate increase of the magnitude of fluctuations 0.1–0.2 Å in the H6-loop-H7 region compared to apoprotein. There was also a 0.2–0.3 Å reduction in magnitude of fluctuations of the H6-loop-H7 region in trajectories of 3– and 13–RORγt complexes compared to apoprotein. Similarly, the iag–RORγt complex showed a 0.2–0.3 Å reduction in fluctuations in the H6-loop-H7 region compared to apoprotein. Metadynamics simulations based on CV distance from Gln286:δCO to His479:Nε (14–15 Å) revealed deep energy minima in pocket center in case of 2–, 10–, and 13–RORγt complexes compared to apoprotein. Those energy minima were shallow in case of RORγt ligand–protein complexes with 3 and 14 or completely absent for all reference compounds, Figure 3C,D. Radial distribution function (RDF) analysis of explicit water solvent around agonist lock Tyr502:OH showed an increase in hydration in case of 4,5,6,7-tetrahydro-benzothiophene–RORγt compared to apoprotein or ago–, iag–, and gsk2981278–RORγt complexes. Similarly, there was an increase in the hydration around agonist lock of 2HC–RORγt compared to apoprotein, ago–, iag–, and gsk2981278–RORγt complexes. 13–RORγt complex was the only among all the evaluated 4,5,6,7-tetrahydro-benzothiophene derivative complexes with RORγt that showed higher hydration around the agonist lock compared to the 25HC–RORγt complex. The hydration around the Gln286:δCO hydrophilic center showed a reduction in case of 2–, 3–, 10–, and 14–RORγt compared to apoprotein or ago– and iag–RORγt complexes. Similarly, the hydration around Hid323:CO decreased in case of 2–, 10–, and 13–RORγt complexes compared to apoprotein and ago–RORγt complexes. The hydration around the Phe377:CO hydrophilic center decreased in case of 2–, 3–, 10–, and 14–RORγt compared to all other reference ligand–RORγt complexes, Figure 3E. The results of RDF analysis of hydration show that the pocket hydration patterns of the 13–RORγt complex were different from the other four 4,5,6,7-tetrahydro-benzothiophene–RORγt complexes, Figure 3A–E, while differences in pocket hydration between the four RORγt complexes were insignificant.

Figure 3.

MD simulation studies performed on 2–, 3–, 10–, 13–, and 14–RORγt (PDB ID: 5APH) complexes: (A) rmsd plots of ligand fit protein superposed showing only small shifts for all compounds compared to reference RORγt agonists and inverse agonists; (B) rmsf of H11–H12 region showing an increase in atomic fluctuations for all the compounds 2–14 compared to reference ligands; (C) metadynamics simulation plots showing the change of free energy surface (FES) in terms of distance CV. Compounds 2–14 show deeper energy troughs in the pocket region compared to reference ligands; (D) 3D ribbon diagram of RORγt showing H11–H12 region in blue and orange colors. CV distance direction from Gln286:δCO to His479:Nε (14–15 Å) is shown by a red arrow; (E) RDF of water around the three hydrophilic centers Gln286:δCO, Hid323:CO, Phe377:CO and the agonist lock Tyr502:OH of 2−, 3−, 10−, 13−, and 14−RORγt compared to apoprotein or 25HC−, ago−, iag−, and gsk2981278−RORγt complexes.

ADME properties of 167 hits from 3D screening were calculated to exclude compounds that violate the Lipinski rule of five. The compounds were selected for synthesis based on their docking scores, ligand–protein complex stability, drug-likeness, and synthetic feasibility. Compounds 1–23 can be obtained through amide formation from starting reactants, Scheme 1. Compound 1 was synthesized by the amide formation of 3-carboxylic acid of 4,5,6,7-tetrahydro-benzothiophene 2-amide derivative and 1-benzylpiperazine in the presence of 1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxid hexafluorophosphate, N-[(dimethylamino)-1H-1,2,3-triazolo-[4,5-b]pyridin-1-ylmethylene]-N-methylmethanaminium hexafluorophosphate N-oxide (HATU) and N,N-diisopropylethylamine (DIPEA), Scheme 1. Compounds 2, 12, 14, 15, 17, and 19–21, on the other hand, were obtained through the amide formation reaction between their corresponding 2-amino 4,5,6,7-tetrahydro-benzothiophene 3-amide derivative and carboxylic acids in the presence of 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) and dimethyl sulfoxide (DMSO), Scheme 1. Similarly, compound 3, the 2-amide bond of 4,5,6,7-tetrahydro-benzothiophene 3-amide derivative was formed with the carboxylic acid in the presence of HATU and DIPEA, Scheme 1. For compounds 4–6, the 2-amide bond was formed between the corresponding 2-amino 4,5,6,7-tetrahydro-benzothiophene 3-amide derivatives and acyl chlorides in the presence of NaOH, Scheme 1. In case of compound 7, the 2-amide bond was formed between 2-amino 4,5,6,7-tetrahydro-benzothiophene 3-amide derivative and trifluoroacetic anhydride (TFAA) at 60 °C, Scheme 1. The 2-amide bond for compounds 8, 9, and 11 were formed between the corresponding 2-amino 4,5,6,7-tetrahydro-benzothiophene 3-amide derivatives and carboxylic acids in the presence of methane-sulfonyl chloride (MsCl) and triethylamine (NEt₃), Scheme 1. Compound 10 was obtained from 2-amide formation of 2-amino 4,5,6,7-tetrahydro-benzothiophene 3-amide derivatives and carboxylic acid in the presence of K₂CO₃, Scheme 1. The final amide formation reaction to derive compound 13 was between the 3-carboxylic acid of 4,5,6,7-tetrahydro-benzothiophene 2-amide derivative and amine in the presence of 1,1′-carbonyldiimidazole (CDI), Scheme 1. N-{3-[(2R,4R)-2,4-Dimethylpiperidine-1-carbonyl]-4,5,6,7-tetrahydro-1-benzothiophen-2-yl}pyridine-3-carboxamide was separated from racemate by chiral chromatography, Figure S50. Enantiomeric excess (ee %) was >98% when compared to racemate at standard conditions. The (R,R)-configuration was confirmed by X-ray crystallographic analysis. In the asymmetric unit of the crystal, two independent molecules were found. The 2,4-dimethylpiperidine ring of both molecules had a twisted boat conformation with the two methyl groups positioned in the cis–cis configuration. They had close conformational characteristics and differed only in the conformation of the cyclohexene ring of 4,5,6,7-tetrahydro-benzothiophene fragment. The differences in geometric characteristics were insignificant. In both molecules, intramolecular interactions between the sulfur atom of the thiophene fragment and CO amide oxygen in position 2 lead to stabilization of cis-location of substituents with short S–O contacts of 2.768(6) and 2.804(6) Å, respectively, which are shorter than the sum of van der Waals radii for S and O atoms of 3.32 Å, Figures S51 and S52. Compound 16 was synthesized in the presence of DIPEA and DMSO starting from of 2-amino 4,5,6,7-tetrahydro-benzothiophene 3-amide derivative and carboxylic acid, Scheme 1. Compounds 18, 22, and 23 were derived by amide formation between the corresponding 2-amino of 4,5,6,7-tetrahydro-benzothiophene 3-derivatives and carboxylic acids in the presence of DIPEA and 2-chloro-1-methylpyridinium iodide at 100 °C, Scheme 1.

Scheme 1. Synthesis Scheme of Compounds 1–23.

R = fluoro-2λ³-benzene for 1, 5, and 6; 5-chloro-2-(methylthio)-4λ³-pyrimidine for 2 and 3; 1-methyl-1H-5λ³-pyrazole for 4; trifluoro-λ³-methane for 7; 2-((λ³-methyl)thio)-4-methylpyrimidine for 8; 1-(λ³-methyl)piperidine for 9; 1-(λ³-methyl)-4-methylpiperazine for 10; 1-(λ³-methyl)-4-methylpiperidine for 11; 3λ³-pyridine for 12–15, 17, 19–21; 2-((λ³-methyl)thio)-1-methyl-1H-imidazole for 16; 2λ³-pyrazine for 18 and 22; 5-methyl-2-(methylsulfonyl)-4λ³-pyrimidine for 23. R₁ = 1-benzyl-4λ²-piperazine for 1; benzyl-λ²-azane for 2, 7–11 and 16; 3-(λ²-azaneyl)tetrahydrothiophene 1,1-dioxide for 3; o-tolyl-λ²-azane for 4; 1-methyl-4λ²-piperazine for 5 and 6; 2-methyl-1λ²-pyrrolidine for 12; (2S,4S)-2,4-dimethyl-1λ²-piperidine for 13; (1R,4R)-4-(λ²-azaneyl)cyclohexan-1-ol for 14; 3-ethyl-1λ²-pyrrolidine for 15; (1λ²-pyrrolidin-3-yl)methanol for 17; 4λ²-morpholine for 18; 3-((λ²-azaneyl)methyl)-1-methyl-1H-1,2,4-triazole for 19; (R)-2-(λ²-azaneyl)propanenitrile for 20; 4-((λ²-azaneyl)methyl)piperidine for 21; 1λ³-benzene for 22 and 23. R₂ = H for 1, 2, 7–17, and 19–23; λ³-methane for 3, 6, and 18; 1λ³-ethane for 4; 2-methyl-2λ³-propane for 5. R₃ = H for 1–17 and 19–23; λ³-methane for 18. Reagents and conditions: (a) amine, DMF, HATU, DIPEA, rt, 32% for 1; (b) amine, DMF, CDI, 70 °C, 24% for 13; (c) carboxylic acid, EDC, DMSO, rt, 58% for 2, 67% for 12, 63% for 14, 65% for 15, 59% for 17, 57% for 19, 67% for 20 and 63% for 21; (d) carboxylic acid, DMF, HATU, DIPEA, 50 °C, 31% for 3; (e) acyl chloride, NaOH, rt, 3% for 4, 3% for 5 and 14% for 6; (f) TFAA, 60 °C, 52% for 7; (g) carboxylic acid, dioxane, MsCl, NEt₃, rt, 3% for 8, 12% for 9, 10% for 11; (h) carboxylic acid, DMF, K₂CO₃, 50 °C, 42% for 10; (i) carboxylic acid, DIPEA, DMSO, 100 °C, 65% for 16; (j) carboxylic acid, DIPEA and 2-chloro-1-methylpyridinium iodide at 100 °C, 56% for 18, 58% for 22, 62% for 23.

In vitro potency of the compounds was evaluated in two binding assays: TR-FRET and FP competitive assay. First, TR-FRET assay confirmed that 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene are potent inverse agonists of RORγt; however, there was no significant difference between the 23 compounds, and their IC₅₀ was in range 0.5–5 nM, Table 1. Experimental values were in good concordance with the SP docking scores. Lipophilicity, poor solubility, or high TPSA of some compounds such as 1, 2, and 16 did not have a significant impact on IC₅₀ determined in TR-FRET assay. Second, FP assay confirmed the findings of TR-FRET. The experimental binding free energies (ΔG_EXP) of 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene–RORγt complexes were comparable to XP docking scores and those obtained in FEP simulations (ΔG_FEP), Figure 4A,B. To confirm in vitro cellular binding, we tested compounds 2, 3, 10, and 13 in human GAL4-RORγ reporter assay that utilizes human cells that provide high-level expression of a hybrid human RORγ. In the reporter assay, the N-terminal DNA-binding domain (DBD) of the native RORγ receptor has been substituted with that of the yeast GAL4-DBD.^62,63 Inverse agonists that bind to RORγ (LBD) decrease the affinity of GAL4-RORγ to DNA site. The IC₅₀ measured in human GAL4-RORγ reporter assay of compounds 2, 3, 10, and 13 were 10 times lower compared ursolic acid (90 nM), Figure 4C.

Figure 4.

In vitro binding assay results: (A) linear regression of experimental binding free energies (ΔG_EXP) compared to Docking scores, R² = 0.58, P < 0.01; (B) ΔG_EXP compared to ΔG_FEP, R² = 0.75, P < 0.01; (C) concentration response GAL4-RORγ reporter assay of compound 2, 3, 10, 13 compared to ursolic acid (ctrl), IC₅₀ = 0.75, 2.19, 0.82, 1.61, and 99.92 nM, respectively.

Furthermore, we performed an in vivo Th17 polarization assay to compare the effect of 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene on IL17A production. IL-17A gene expression is regulated by RORγt binding to ROR response element (RORE). Hence, IL17A protein can be used as a marker for RORγt activity.⁶⁴ Compounds 1–23 were tested for their ability to inhibit Th17 polarization of human peripheral blood mononuclear cells (PBMCs). PBMCs treated with compounds 2, 9, 10, 13, and 14 showed the greatest reduction in IL17A+ cells compared to control (Figure 5A). Compound 3 showed a 30% reduction in percentage of IL17A+ cells (Figure 5A), compared to 2, 10, 13, and 14 which reduced IL17A+ cells by 40–60%. Compounds 2, 10, and 13 decreased IL17A and IL21 secretion from polarized PBMCs compared to the vehicle, (Figure 5B,C). We then sought to determine IC₅₀ of compound 2 in Th17 polarization as a representative of the set of 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene. The IC₅₀ was 0.38 nM in the cellular assay compared to 0.39 nM in TR-FRET assay. We further tested compound 2 on Th17 polarization of PBMCs from multiple donors (n = 9), which under the same experimental condition express different levels of IL17A. Compound 2 reduced the percentage of both IL17A+ and IL21+ cells from all donors by 20–40%, Figure 5D,E. There was no sign of cell death or decrease in cell viability when the cells were evaluated by flow cytometry. 2,3 Derivatives of 4,5,6,7-tetrahydro-benzothiophene did not increase cell proliferation or apoptosis and did not affect the percentage of IFNγ+ or IL22+ T cells. The solubility of 12 compounds was further assessed in acidic pH (1.5 and 5). The solubility of 1, 6, 10–13, 16, and 17 was inversely proportional to changes in pH. Compound 3 was unstable in acidic pH as the maximum solubility decreased from 40 μM (pH = 7.4), Table 1, to 9.72 μM (pH = 5) to 1.7 μM (pH = 1.5), Table 2. The solubility of 4 and 14 was not affected by pH changes, Table 2. Mouse/human/rat/dog microsomal stability studies showed that the intrinsic clearance (Cl_in) is high for compounds 1, 11, and 16; moderate for 2, 3, 6, 10, and 13; and low for 4, 12, and 14, Table 2. The determination of two direction Caco-2 permeability, like microsomal stability, revealed that the compounds are diverse in membrane permeability profiles where compounds 6, 12, 13, and 16 exhibited high permeability, while 1, 2, 10, and 17 were moderate and 3, 4, and 11 showed low permeability in the A–B direction. Compounds 3 and 17 undergo active efflux, Table 2. Finally, canine MDR1 knock-out, human MDR1 knock-in MDCKII (MDR1-MDCKII) cell permeabilities of 6, 12, 13, and 16 were high, while 1, 2, 3, 4, 10, and 17 were moderate. Compounds 11 and 14 showed low permeability in the A–B direction. Compounds 3 and 17 undergo active efflux and are P-gp substrates, Table 2. The efflux ratio of compound 3 was decreased partially by verapamil in Caco-2 and completely blocked by cyclosporine A in MDR1-MDCKII cells, Table 2.

Figure 5.

Th17 polarization assays: (A) effect of 10 nM of gsk2981278 and compounds 1–23 on Th17 polarization of PBMCs, *P < 0.05 compared to vehicle (ctrl), one-way ANOVA, and Bonferroni’s multiple comparisons test; (B) effect of 10 nM of gsk2981278, 2, 10, and 13 on levels of IL17A in the supernatant of Th17-polarized PBMCs compared to vehicle (ctrl), *<0.05, one-way ANOVA, and Dunnett’s multiple comparisons test; (C) effect of 10 nM of gsk2981278, 10, and 13 on IL21 levels in the supernatant of Th17-polarized PBMCs compared to vehicle (ctrl), *P < 0.05, one-way ANOVA, and Dunnett’s multiple comparisons test; (D) effect of 2 (10 nM) on percentage of CD4+ IL17A+ cells in Th17 polarization of PBMCs derived from multiple donors (n = 9), *P < 0.05 compared to ctrl, paired t-test; and (E) effect of 2 (10 nM) on percentage of CD4+ IL21+ cells in Th17 polarization of PBMCs derived from multiple donors (n = 9), *P < 0.05 compared to ctrl, paired t-test.

Table 2. Solubility, Microsomal Stability, Caco-2, and MDR1-MDCKII Permeability of 2,3 Derivatives of 4,5,6,7-Tetrahydro-benzothiophene RORγt Modulators.

			microsomal stability	Caco-2 permeability			MDR1-MDCKII permeability
ID	solubility μM pH = 5.0, T = 298 K, UV–vis	solubility μM pH = 1.6, T = 298 K, LC/MS	Clin (μL/min/mg) mouse/human/rat/dog	Papp (AB), 10–6cm/s M ± SD	efflux	efflux verapamil	Papp (AB), 10–6cm/s M ± SD	efflux	efflux cyclosporine A
1	34	291	2015/1915/641/540	5.8 ± 0.2	1.3	1.7	7.9 ± 2.1	0.4	0.9
2	ND	ND	127/126/267/115	4.2 ± 0.1	0.3	0.4	8.0 ± 0.4	0.3	0.3
3	9.72	1.70	322/36/118/41	1.4 ± 0.1	19.4	4.5	2.5 ± 0.1	8.7	1.1
4	8.13	2.80	11/22/29/24	1.8 ± 0.2	0.4	0.6	3.2 ± 1.5	0.3	0.3
6	349.39	≥400	546/91/116/60	28.2 ± 2.0	0.6	1.0	28.9 ± 4.1	0.7	0.6
10	357.28	≥400	319/214/115/132	4.6 ± 0.4	1.3	1.6	6.8 ± 2.4	1.1	1.2
11	3.02	364	2345/510/942/260	1.2 ± 0.3	0.3	1.2	1.3 ± 0.2	0.8	0.7
12	391.51	≥400	71/67/97/53	34.2 ± 2.2	0.8	0.9	29.0 ± 4.7	0.8	0.8
13	23.72	≥200	414/404/507/369	26.7 ± 1.3	0.9	1.2	26.7 ± 3.4	0.7	1.0
14	33.70	16	50/35/5/16	2.0 ± 0.5	0.9	1.0	1.2 ± 0.1	1.3	0.9
16	13.56	267	2142/954/887/332	18.0 ± 0.2	0.7	0.7	27.8 ± 5.2	0.8	1.1
17	365.21	330	46/23/7/11	4.4 ± 0.1	7.1	1.1	5.2 ± 0.3	4.7	1.4

Discussion

In their pocket orientation, 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene modulators of RORγt conform to structural features of previously discovered RORγt inverse agonists. The saturated tetrahydro-benzene ring structure is the hydrophobic pocket interacting moiety, the 3-amide ring substitution comprised the head group while the 2-amide ring substitution is the tail group. These structural features kept the ligand pose within rmsd <2 Å compared to reference ligands. On the other hand, the amide arms that replace rings B and C of cholesterol core per initial design hypothesis provided two hydrophilic surfaces that align in the two hydrophilic regions of the RORγt pocket Cys320-Glu326 and Arg364-Phe377. They specifically interact with Hid323 and Phe377 backbone amides. Those interactions stabilized the molecular complex in MD simulations and lowered the free-energy surface of this region of the ligand pocket. As such the 2,3-amide derivatives of 4,5,6,7-tetrahydro-benzothiophene complexes with RORγt were energetically more stable compared to RORγt complexes with other reference agonists/inverse agonists evaluated. Alkyl substitution of the tetrahydro-benzothiophene ring at position 6 resulted in a deviation from the active pose. In the case of one methyl or one ethyl substitutions, there was no effect on the docking score and no reduction in activity in vitro. Substituting the tetrahydro-benzothiophene ring with a bulkier group such as in case of compound 5 resulted in a larger deviation (>3 Å) and a lower docking score; however, there was no reduction in activity in vitro. The addition of a second alkyl such as di-methyl substitution in the case of compound 18 resulted in both a lower docking score and a reduction of in vitro activity. This can be explained by the preservation of the chair conformation in the cases of a single substitution, while adding two alkyls forces the tetrahydro-benzene ring to adopt a boat or twisted boat conformations and the creation of an additional perpendicular plane. In this case, the substitution can affect both the ligand entry and the stability of the ligand inside the pocket. It should be expected that a similar substitution at 4, 5, or 7 positions can lead to similar effects. As the results showed, the length of the bridge connecting ring substitution to position 3 of 4,5,6,7-tetrahydro-benzothiophene is another important determinant of the activity. In the case of bridges that are one carbon long such as the case of compounds 22 and 23, the linked ring structure becomes sterically restricted and adopts a perpendicular conformation with respect to the tetrahydro-benzothiophene ring. As such, the ligand entry and pose are both affected. Extending the bridge beyond four atoms such as in case of compound 1 creates a similar perpendicular orientation of the substituent, however more distant from the tetrahydro-benzothiophene core. This modification does not seem to affect the entry; however, it affects the adoption of the ideal pose, resulting in a lower docking score and a reduction of potency in vitro. Structure–activity relationship (SAR) analysis shows that the adequate length of the bridge at position 3 is three atoms. A planar phenyl ring substitution linked to position 3 of the tetrahydro-benzothiophene ring through a three atoms bridge, such as in case of compound 2 and 7–11, showed a better affinity as determined by docking score, an increased stability as determined by MD simulation and higher potency in vitro. In those six compounds, ring planarity and bridge length allow for parallel or convex planes between the tetrahydro-benzothiophene and phenyl ring substitution. Such a conformation allows for stacking and the reduction of PSA as calculated based on the atomic partitioning of B3LYP/6311G* wavefunctions (data not shown) compared to their TPSA. These π–π intramolecular interactions contribute to improved activity in cellular assays compared to the compounds with two-atom bridges such as 17, 18, 22, and 23. An exception was compound 21, which had flexible piperidine ring substitution. This substructure combined to a three-atom bridge led to the creation of an additional perpendicular plane like that seen in 1 and the compounds with two-atom bridges. On the other hand, compounds with two-atom bridges at position 3 were active when the substitution was a flexible saturated ring that can adopt a chair conformation such as in case of compound 13. Increasing the length of amide bridge at position 2 of 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene which constitutes the tail group was less critical to affect the binding pose or in vitro activity. We increased the length of amide bridge at position 2 of the tetrahydro-benzothiophene ring up to four atoms, such as in case of compounds 8 and 16 where the interactions with Gln286 and other residues at the pocket entry were maintained. Moreover, we evaluated several ring substitutions linked to positions 2 and 3 of 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene. Contrary to our initial assumptions, the incorporation of a hydroxyl group such as in case of compounds 14 and 17 resulted in a reduction of docking score and in vitro activity compared to 13 and 15, respectively. These effects can be due to the intramolecular interactions between the hydroxyls and the amide bridges on position 2 as well as the nitrogen of the pyridine ring because these interactions can change the distribution of dipole moment along the molecules, which affects their ability to form hydrogen bond interactions with the hydrophilic regions of the pocket through amide arms. Physicochemical studies revealed that all the evaluated 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene except 2, 3, 4, and 23 have good solubility in neutral and acidic pH. They have diverse microsomal stability and Caco-2 permeability profiles that allow for a wide range of formulation options and organ-specific targeting. Similarly, they have different potentials to cross blood–brain barrier, as determined by MDR1-MDCKII permeability, which allows for targeting the central nervous system (CNS) inflammatory conditions such as multiple sclerosis and brain tumors. It is worth mentioning that recovery values of compounds 2, 4, 10, and 11 in Caco-2 permeability assay were low (<80%), which can be attributed to the low solubility of compounds 2 and 4 or increased metabolism by Caco-2 cells for 10 and 11, so the permeability values for those compounds are approximate. The structure metabolism relationship analysis, on the other hand, reveals that the compounds that had an alkyl substitution in position 6 of 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene were more stable in microsomal stability assay such as compound 3, 4, and 6. The hydroxyl substitutions and resulting intramolecular interactions also protected the compounds from microsomal metabolism such as in case of compound 14 and 17. Increasing the length of amide bridge at position 3 such as in case of compound 1 or position 2 as in case of compound 16 resulted in an accelerated metabolism in microsomal stability assay. As aforementioned, these compounds adopt an open conformation because of the weak intramolecular interactions, where their ring substitutions at position 3 align to a perpendicular plane far from the thiophene ring due to the increased bridge length. Such open conformations increase compound’s susceptibility to enzymatic cleavage of the amide bond. Compound 12 has improved stability and was effective in Th17 cellular assay; however, this compound remains less attractive for clinical development. This is because of the small size of pyrrolidine ring substitution at position 3 renders compound 12 less target specific which is also the case for 20 with its propanenitrile. The same applies to compound 7 which is lacking a ring structure at the 2-position substitution with its trifluoromethyl group. Those three compounds maintain an activity toward RORγt; however, they are less specific in terms of target selectivity which we observed in “in silico” off-target binding studies (data not shown). Additionally, they adopt other poses with orientations that differ significantly from the active pose. These findings collectively set forth the importance of the central region of the pocket between Cys320-Glu326 and Arg364-Phe377 hydrophilic centers, where 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene modulators of RORγt that occupy and strongly interact with this region showed a potent inverse agonist effect in vitro in both cell free and cellular assays. This potent inverse agonist effect was seen for all 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene modulators of RORγt evaluated regardless of their conformity with the active pose defined by the compounds that interact with the agonist lock such as 1, 2, 9–11, 13, 16, and 19. In the current study, the availability of a fluorescent cholesterol 25NC provided a chance to perform FP assay which is superior to other binding assays such as TR-FRET, SPR, or SPA as it allows for ΔH and ΔS determination. 25NC was selected as a fluorescent tracer based on docking studies, molecular weight, and solubility profile. In addition, the availability of approximately 100 crystal structures of RORγt provided a large pool for sampling and testing with MD simulations with great accuracy, which renders molecular structure studies for confirmation of binding poses unnecessary; however, future studies that employ X-ray crystallography can be performed to confirm the findings of the current study. Furthermore, in vivo chronic toxicity studies are required to understand the long-term effects of the modulators on body organs especially liver and lymphatic tissues which have been shown to be adversely affected by chronic administration of RORγt inverse agonists. Finally, in vivo evaluation of 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene pharmacokinetics in rodents as well as RORγt disease animal models is a suitable next step. Compounds 10 and 13 are considered best lead candidates based on physicochemical properties and in vitro efficacy; however, metabolite profiling studies are required to identify the active metabolites of the two leads that can partially contribute to the activity toward RORγt. Such studies can give an explanation to the superior in vitro potency of compounds 10 and 13 despite their relatively higher intrinsic clearance compared to other 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene that were evaluated in this study.

Conclusions

We identified 2,3 derivatives of 4,5,6,7-tetrahydro-benzothiophene as potent modulators of RORγt. The modulators efficiently attenuate Th17 polarization and can be used in vivo to block T-cell differentiation to Th17. The compounds can be developed to treat conditions associated with increased IL17A, IL21, and other cytokines regulated by RORγt. The compounds described in this study are bioavailable and can be developed for organ-specific therapies. The compounds that have high potential to penetrate blood–brain barrier can be used to target brain and CNS conditions. The use of in silico studies in combination with in vitro studies provided an accelerated alternative to high-throughput in vitro screening discovery and optimization studies. The efficiency of the approach is dependent on the availability of high-quality structural data of the target protein. Additionally, where multiple of such high-quality structure determination studies have been performed, adequately reviewed, and publicly disclosed, we conclude that docking and MD simulations can effectively determine binding poses of new ligands without the need for a confirmation through NMR or X-ray crystallography studies.

Experimental Section

2D Fingerprint Design

A review was conducted to collect data about the structure of positive hits identified in published RORα, γ, γt high-throughput, medium-throughput screening bioassays, counter screens, and final screens. All positive hits were ranked by potency and activity. Active hits were visually inspected to identify common structure features, (1) number of rings, (2) degree of branching, (3) linker length, and (4) functional groups. The collected information was used to create intermediate design ideas starting from 25HC. The intermediate designs were tested by Autodock Vina against three RORγt crystal structures (PDB ID: 3KYT, 3L0L, 3L0J). Several rounds of substitutions were carried out to reach final design concepts, which were then used as screening fingerprints.

Virtual Screening

2D screenings were performed using ChemSpider, REAL, PubChem, Mcule, and other databases that allow for structure search. Initial screens were based on 2D designs derived from 25HC linking and substitution, and final screens were based on structures of compound 1 and 2. The minimum Tanimoto coefficient of similarity was set at 0.7 for all 2D screens. 3D screenings were performed using phase screening and Glide HTVS dockings (PDB ID: 5APH). A ligand binding hypothesis was created based on the pocket of the best resolution RORγt X-ray crystal (PDB ID: 5APH) that is currently published.⁵² Library enrichment and diversification were performed at all stages by small rounds of 2D screenings of chemical catalogues of virtual and enumerated database of synthetically feasible molecules against selected scaffolds. The 23 compounds were selected individually from small, clustered databases generated in parallel for each hit from phase screening or HTVS. Parallel rounds of screenings were performed using Glide, ASE-dock, GBVI/WSA ΔG Gold, London ΔG and Autodock Vina against RORγt (PDB ID: 5APH). Rmsd_scr/ref was based on the maximum common structure. Rmsd_align was computed using PhaseScreen.

Molecular Docking

All docking calculations were carried out with Glide software using its standard precision (SP) and XP scoring functions. Compounds were processed with Ligprep, to generate 3D energy-minimized molecular structures with correct tautomeric states in OPLS3e force field. Epik was used to enumerate tautomers and protonation states at pH 7 ± 2. RORγt (PDB ID: 5APH) was pre-processed and refined with protein preparation wizard to correct for deficiencies such as missing side chains, loops, hydrogen atoms; to check for protonation states of ionizable amino acids (at physiological pH); to flip wrongly assigned Asn and Gln side chains. The resulting structures were minimized with protein wizard. Active pose docking scores are reported in the SAR table. Rmsds of bound ligands were computed with Maestro superposition utility or Ligrmsd.

MD Simulations

Initial screening (50 ns) of MD simulations was conducted in OPLS3e force field, Desmond software. Longer simulations (250 ns) of ligand protein complexes were carried out in OPLS2005. Explicitly solvated systems were simulated in the NPT ensemble in an orthorhombic box of sizes 79 Å × 56 Å × 75 Å, 34,100 atoms (total), and 9900 water molecules (TIP3P). Systems were brought to neutrality with the addition of sodium and chlorine ions (physiologic pH, 0.15 M NaCl), minimized using a hybrid method of steepest decent and limited-memory algorithms. Simulations were performed under NPT ensemble implementing Berendsen thermostat/barostat methods, under constant temperature of 300 K, Nose–Hoover thermostat, and Martyna–Tobias–Klein barostat algorithms to maintain 1 atm of pressure, respectively. Short-range Coulombic interactions were analyzed using a cut-off value of 9 Å using the short-range method. A smooth particle mesh Ewald method was used for handling long-range Coulombic interactions. A tolerance value of 1 × 10^–9 was set for computations implemented by the SHAKE algorithm. Simulations were carried out for 25-HC, gsk2981278, 2, 3, 10, 13, and 14 for 50 and 250 ns, and trajectory sampling was done at 50 and 5 ps, respectively.

FEP Ligand Mutation Analysis

FEP calculations were carried out using the GPU enabled FEP methodology, and Desmond 6.9. REST algorithm was used for locally enhanced sampling. Error estimates are based on standard deviation of repeat simulations and the mean unsigned error (MUE). Systems were assigned OPLS2005 force field, solvated in an orthogonal box of SPC water molecules with buffer width of 5 Å for complex and 10 Å for solvent simulations. Systems were brought to neutrality with the addition of sodium and chlorine ions (physiologic pH, 0.15 M NaCl). Systems were relaxed and equilibrated using the Desmond relaxation protocol, consisting of a 100 ps energy-minimization with restraints on solute using Brownian dynamics NVT, then a 12 ps simulations at T = 10 K using an NVT ensemble followed by a 12 ps simulation using NPT ensemble. Systems were equilibrated at room temperature for 24 ps using the NPT ensemble with restraints. A 240 ps room temperature NPT ensemble simulation, no restraints, was conducted. Finally, 5 ns hopping simulations were carried out in the NPT ensemble, for both complex and solvent systems.

Metadynamics Simulations

The solvated complex was loaded into the metadynamics panel. The CV distance is defined between Gln286:δCO to His479:Nε. The Gaussian width and height are set to 0.05 Å and 0.03 kcal/mol, respectively, and a wall was set at 30 Å. Systems were relaxed and equilibrated using the Desmond relaxation protocol, consisting of a 100 ps energy-minimization with restraints on the solute heavy atoms using Brownian dynamics NVT, then a 12 ps NVT ensemble simulations at T = 10 K, small timesteps, and restraints on solute heavy atoms followed by a 12 ps simulation using NPT ensemble, similar to restraints on solute heavy atoms. The solvated complexes were equilibrated at room temperature for 1 ps using NPT ensemble with restraints. This was followed by 24 ps simulation at room temperature, NPT ensemble with no restraints. Finally, 25 ns metadynamics simulations were carried out in the NPT ensemble.

Synthesis

Compounds 1–23 can be obtained through amide formation from starting reactants provided by Enamine Ltd, Kyiv, Ukraine. All compounds are >95% pure by HPLC analysis. HPLC/UPLC traces and ¹H NMR spectra of example compounds are included in the Supporting Information. Analytical LC/MS instrument specifications: Agilent 1100 series LC/MSD system with DAD/ELSD and Agilent LC/MSD VL (G1956A), SL (G1956B) mass-spectrometers, Agilent 1200 series LC/MSD system with DAD/ELSD and Agilent LC/MSD SL (G6130A), SL (G6140A) mass-spectrometers. LC/MS data were obtained using positive/negative mode switching; column: SB-C18 1.8 μm 4.6 × 15 mm rapid resolution cartridge; mobile phase: (A) acetonitrile, 0.1% formic acid, (B) water, 0.1% formic acid; flow rate: 3 mL/min; gradient: 0 min—100% B, 0.01 min—100% B, 1.5 min—0% B, 1.8 min—0% B, 1.81 min—100% B, injection volume: 1 μL; ionization mode: atmospheric pressure chemical ionization (APCI); scan range: m/z 80–1000. ¹H NMR; the spectra were obtained using different spectrometers, mainly with an operating proton frequency of 400 MHz in DMSO-d₆. Calibration was carried out according to internal standard DMSO = 2.5 ppm. General synthetic procedure (method 1): an amine (100 mg), a carboxylic acid (1.1 mol equiv to the amine), and 0.5 mL of DMSO were placed into a 4 mL capped glass vial, and the mixture was stirred for 30 min. Then, EDC (1.2 mol equiv to the amine) was added, and the mixture was stirred for 1 h. If the solution was transparent, the mixture was left overnight at room temperature; otherwise, the vial was placed in the ultrasonic bath and left overnight. The solution was filtered, and the solvent and volatile components were evaporated under reduced pressure to give the crude product. The product was further purified by HPLC. General synthetic procedure (method 2): an amine (100 mg), DIPEA (1.2 mol equiv to the amine), and DMSO (0.5 mL) were placed into a 4 mL capped glass vial and stirred for 30 min. After the addition of an alkyl halide (1.2 mol equiv to the amine), the vial was stirred for 1 h at room temperature. Then, the vial was placed into a thermostat (set to 100 °C) for 9 h. After cooling, the mixture was filtered; the solvent and volatile components were evaporated under reduced pressure to give the crude product. The product was further purified by HPLC. General synthetic procedure (method 3): an amine (100 mg), a carboxylic acid (1.2 mol equiv to the amine), DIPEA (1.2 mol equiv to the amine), and acetonitrile (0.5 mL) were placed into a 4 mL capped glass vial and stirred for 30 min. After the addition of 2-chloro-1-methylpyridine-1-ium iodide (1.44 mol equiv to the amine), the vial was stirred for 1 h at room temperature and placed into a thermostat (set to 100 °C) for 6 h. After cooling, the mixture was filtered; solvent and volatile components were evaporated under reduced pressure to give the crude product. The product was further purified by HPLC.

N-(3-(4-Benzylpiperazine-1-carbonyl)-4,5,6,7-tetrahydro-benzo[b]thiophen-2-yl)-2-fluorobenzamide (1)

Step A: 2 M aq NaOH solution (91 mL, 182 mmol) was added to a suspension of 2-amino-4, 5,6,7-tetrahydro-benzo[b]thiophene-3-carboxylic acid methyl ester 1 (13.1 g, 60.8 mmol) in EtOH (100 mL). The reaction mixture was heated to 60 °C and stirred for 4 h. The temperature was raised to 80 °C, and the reaction was stirred overnight. The reaction was left stirring at 80 °C for a further 7 h and was then cooled to room temperature and stirred overnight. The reaction mixture was concentrated under reduced pressure to remove the EtOH. The aqueous solution was washed with DCM (3 × 200 mL) and then acidified to pH 2 with concentrated hydrochloric acid. A brown solid was formed, filtered off, washed with water, and then with hexane. The solid was dried to afford the title compound (6 g, 50%) as a dark brown solid. Step B: 2-fluorobenzoic acid (500 mg, 3.5 mmol) and (COCl)₂ (10 mL) were heated at 50 °C for 30 min. The reaction mixture was concentrated under reduced pressure. The residue was dissolved in 25 mL of dioxane. NaOH (3.5 mmol), and 10 mL of water were added at 0 °C. 2-Amino-4,5,6,7-tetrahydro-1-benzothiophene-3-carboxylic acid (3.5 mmol) was added, and the mixture was stirred at room temperature for 12 h, and 2 mL of HCl was added. Then, the precipitate was filtered and washed with water. Yield: 340 mg, 44%. Step C: 2-(2-fluorobenzamido)-4,5,6,7-tetrahydro-1-benzothiophene-3-carboxylic acid (340 mg, 1 mmol), HATU (1 mmol), and DIPEA (2 mmol) were dissolved in DMF (25 mL). 1-Benzylpiperazine (1 mmol) was added, and the mixture was stirred for 24 h at 50 °C and cooled to room temperature. The reaction mixture was concentrated under reduced pressure to remove the DMF. 50 mL of water was added, 2 mL of HOAc was added, and the precipitate was filtered and washed with water. The product was purified with the use of HPLC. Yield: 149 mg, 32%. Yellow solid. Purity, >95% (assessed by LC/MS). LC/MS (APSI) m/z: [M + H] calcd for C₂₇H₂₈FN₃O₂S, 478.2; found, 478.2.

N-(3-(Benzylcarbamoyl)-4,5,6,7-tetrahydrobenzo[b]thiophen-2-yl)-5-chloro-2-(methylthio) Pyrimidine-4-carboxamide (2)

(Method 1) 2-amino-N-benzyl-4,5,6,7-tetrahydrobenzo[b] thiophene-3-carboxamide (100 mg), 5-chloro-2-(methylthio) pyrimidine-4-carboxylic acid (1.1 mol equiv to the amine), and 0.5 mL of DMSO were placed into a 4 mL capped glass vial, and the mixture was stirred for 30 min. EDC (1.2 mol equiv to the amine) was added. The mixture was stirred for 1 h and left overnight at room temperature. The solution was filtered, and the solvent and volatile components were evaporated under reduced pressure to give the crude product. The product was further purified by HPLC. Yield: 58%; purity, >95% (assessed by LC/MS). ¹H NMR (400 MHz, DMSO-d₆): δ 13.07 (s, 1H), 8.91 (s, 1H), 8.05 (s, 1H), 7.50–7.12 (m, 5H), 4.48 (d, J = 5.7 Hz, 2H), 2.86–2.56 (m, 6H), 1.77 (s, 4H). LC/MS (APSI) m/z: [M + H] calcd for C₂₂H₂₁ClN₄O₂S₂, 473.2; found, 473.2.

5-Chloro-N-(3-{[(1,1-dioxidotetrahydro-3-thienyl)amino]-carbonyl}-6-methyl-4,5,6,7-tetrahydro-1-benzothien-2-yl)-2-(methylthio)-4-pyrimidine-carboxamide (3)

Step A: 4-methylcyclohexan-1-one (5 g, 30 mmol) and methyl 2-cyanoacetate (5 g, 30 mmol) were mixed with NEt₃ (20 mL) and sulfur (powder, 2.5 g). The mixture was heated at 50 °C for 24 h, cooled, and the reaction mixture was concentrated under reduced pressure. The residue was recrystallized from iPrOH, the precipitate was filtered, washed with iPrOH, and dried. Yield: 3 g, 35%. Step B: 2 M aq NaOH solution (91 mL, 182 mmol) was added to a suspension of methyl 2-amino-6-methyl-4,5,6,7-tetrahydro-1-benzothiophene-3-carboxylate 1 (14 g, 60.8 mmol) in EtOH (100 mL). The reaction mixture was heated to 60 °C and stirred for 4 h. The temperature was raised to 80 °C, and the reaction was stirred overnight. The reaction was left stirring at 80 °C for a further 7 h and was then cooled to room temperature and stirred overnight. The reaction mixture was concentrated under reduced pressure to remove the EtOH. The aqueous solution was washed with DCM (3 × 200 mL) and then acidified to pH 2 with concentrated hydrochloric acid. A brown solid was formed, then filtered off, and washed with water and hexane. Then, the solid was dried to afford the title compound (4 g, 40%) as a dark brown solid. Step C: 2-amino-6-methyl-4,5,6,7-tetrahydro-1-benzothiophene-3-carboxylic acid (500 mg, 2.4 mmol), HATU (2.4 mmol), and DIPEA (7.5 mmol) were dissolved in DMF (25 mL). 3-Aminotetrahydrothiophene 1,1-dioxide hydrochloride (2.4 mmol) was added, and the mixture was stirred for 24 h at 50 °C and cooled to room temperature. The reaction mixture was concentrated under reduced pressure to remove the DMF, 50 mL of water was added, 2 mL of HOAc was added, and the precipitate was filtered and washed with water. The product was purified with the use of HPLC. Yield: 129 mg, 15%. Step D: 2-amino-N-(1,1-dioxo-1λ⁶-thiolan-3-yl)-6-methyl-4,5,6,7-tetrahydro-1-benzothiophene-3-carboxamide (129 mg, 0.4 mmol), 5-chloro-2-(methylsulfanyl) pyrimidine-4-carboxylic acid (0.4 mmol), HATU (0.4 mmol), and DIPEA (0.8 mmol) were dissolved in DMF (15 mL). The mixture was stirred for 24 h at 50 °C and cooled to room temperature. The reaction mixture was concentrated under reduced pressure to remove the DMF, 50 mL of water was added, 2 mL of HOAc was added, and the precipitate was filtered and washed with water. The product was purified with the use of HPLC. Yield: 62 mg, 31%. Brown solid. Purity, >95% (assessed by LC/MS). LC/MS (APSI) m/z: [M + H] calcd for C₂₀H₂₃ClN₄O₄S₃, 515.0; found, 515.0.

N-(6-Ethyl-3-{[(2-methylphenyl)amino]carbonyl}-4,5,6,7-tetrahydro-1-benzothien-2-yl)-1-methyl-1H-pyrazole-5-carboxamide (4)

Step A: 2-cyano-N-(o-tolyl) acetamide (5 g, 30 mmol) and 4-ethylcyclohexan-1-one (2 g, 11.5 mmol) and 4-ethylcyclohexan-1-one (11.5 mmol) were mixed with NEt₃ (10 mL) and sulfur (powder, 1.5 g). The mixture was heated at 50 °C for 24 h, cooled, and the reaction mixture was concentrated under reduced pressure. The residue was recrystallized from iPrOH, the precipitate was filtered, washed with iPrOH, and dried. Yield: 1.4 g, 41%. Step B: 1-methyl-1H-pyrazole-5-carboxylic acid (100 mg, 0.8 mmol) and (COCl)₂ (10 mL) were heated at 50 °C for 30 min. The reaction mixture was concentrated under reduced pressure. The residue was dissolved in 25 mL of dioxane. NaOH (0.8 mmol) and 10 mL of water were added at 0 °C. 2-Amino-6-ethyl-N-(2-methylphenyl)-4,5,6,7-tetrahydro-1-benzothiophene-3-carboxamide (0.8 mmol) was added, and the mixture was stirred at room temperature for 12 h. 1 mL of HCl was added to the precipitate which was filtered and washed with water. The product was purified with the use of HPLC. Yield: 12 mg, 3%. Beige solid. Purity, >95% (assessed by LC/MS). LC/MS (APSI) m/z: [M + H] calcd for C₂₃H₂₆N₄O₂S, 423.2; found, 423.2.

N-{6-tert-Butyl-3-[(4-methyl-1-piperazinyl) carbonyl]-4,5,6,7-tetrahydro-1-benzothien-2-yl}-2-fluorobenzamide (5)

Step A: 3-(4-methylpiperazin-1-yl)-3-oxopropanenitrile (2 g, 12 mmol) and 4-(tert-butyl) cyclohexan-1-one (12 mmol) were mixed with NEt₃ (10 mL) and sulfur (powder, 1.5 g). The mixture was heated at 50 °C for 24 h, cooled, and the reaction mixture was concentrated under reduced pressure. The residue was recrystallized from iPrOH, the precipitate was filtered, washed with iPrOH, and dried. Yield: 2 g, 50%. Step B: 2-fluorobenzoic acid (100 mg, 0.7 mmol) and (COCl)₂ (10 mL) were heated at 50 °C for 30 min. The reaction mixture was concentrated under reduced pressure. The residue was dissolved in 25 mL of dioxane. NaOH (0.7 mmol) and 10 mL of water were added at 0 °C. 2-6-tert-Butyl-3-(4-methylpiperazine-1-carbonyl)-4,5,6,7-tetrahydro-1-benzothiophene-2-amine (0.7 mmol) was added, and the mixture was stirred at room temperature for 12 h, 1 mL of HCl was added, and the precipitate was filtered and washed with water. The product was purified with the use of HPLC. Yield: 11 mg, 3%. Yellow oil. Purity, >95% (assessed by LC/MS). LC/MS (APSI) m/z: [M + H] calcd for C₂₅H₃₂FN₃O₂S, 458.2; found, 458.2.

2-Fluoro-N-{6-methyl-3-[(4-methyl-1-piperazinyl) carbonyl]-4,5,6,7-tetrahydro-1-benzothien-2-yl} Benzamide (6)

Step A: 3-(4-methylpiperazin-1-yl)-3-oxopropanenitrile (5 g, 30 mmol) and 4-methylcyclohexan-1-one (5 g, 30 mmol) were mixed with NEt₃ (20 mL) and sulfur (powder, 2.5 g). The mixture was heated at 50 °C for 24 h, cooled, and the reaction mixture was concentrated under reduced pressure. The residue was recrystallized from iPrOH, the precipitate was filtered, washed with iPrOH, and dried. Yield: 3.5 g, 40%. Step B: 2-fluorobenzoic acid (500 mg, 3.5 mmol) and (COCl)₂ (10 mL) were heated at 50 °C for 30 min. The reaction mixture was concentrated under reduced pressure. The residue was dissolved in 25 mL of dioxane. NaOH (3.5 mmol) and 10 mL of water were added at 0 °C. 6-methyl-3-(4-methylpiperazine-1-carbonyl)-4,5,6,7-tetrahydro-1-benzothiophen-2-amine (3.5 mmol) was added, and the mixture was stirred at room temperature for 12 h. 2 mL of HCl was added, the precipitate was filtered and washed with water. The product was purified with the use of HPLC. Yield: 25 mg, 14%. Yellow solid. Purity, >95% (assessed by LC/MS). LC/MS (APSI) m/z: [M + H] calcd for C₂₂H₂₆FN₃O₂S, 416.2; found, 416.0.

N-Benzyl-2-[(trifluoroacetyl)amino]-4,5,6,7-tetrahydro-1-benzothiophene-3-carboxamide (7)

2-amino-N-benzyl-4,5,6,7-tetrahydro-1-benzothiophene-3-carboxamide (100 mg, 0.35 mmol) was mixed with TFAA (20 mL), and the mixture was heated at 60 °C for 5 h. The reaction mixture was concentrated under reduced pressure, and the residue was purified with the use of HPLC. Yield: 55 mg, 52%. Beige solid. Purity, >95% (assessed by LC/MS). LC/MS (APSI) m/z: [M + H] calcd for C₁₈H₁₇F₃N₂O₂S, 383.1; found, 383.0.

N-Benzyl-2-({[(4-methyl-2-pyrimidinyl)thio]acetyl}amino)-4,5,6,7-tetrahydro-1 Benzothiophene-3-carboxamide (8)

2-[(4-methylpyrimidine-2-yl) sulfanyl] acetic acid (0.35 mmol) was dissolved in dioxane (15 mL), and MsCl (0.4 mmol) and NEt₃ (0.4 mmol) were added at 0 °C. The mixture was stirred at room temperature for 1 h. 2-Amino-N-benzyl-4,5,6,7-tetrahydro-1-benzothiophene-3-carboxamide (100 mg, 0.35 mmol) was added and stirred for 12 h at room temperature. The reaction mixture was concentrated under reduced pressure. Yield: 2 mg, 3%. Brown solid. Purity, >95% (assessed by LC/MS). LC/MS (APSI) m/z: [M + H] calcd for C₂₃H₂₄N₄O₂S₂, 453.1; found, 453.0.

N-Benzyl-2-[(1-piperidinylacetyl) amino]-4,5,6,7-tetrahydro-1-benzothiophene-3-carboxamide (9)

2-piperidine-1-ylacetic acid (0.35 mmol) was dissolved in dioxane (15 mL), MsCl (0.4 mmol), and NEt₃ (0.4 mmol) was added at 0 °C. The mixture was stirred at room temperature for 1 h, 2-amino-N-benzyl-4,5,6,7-tetrahydro-1-benzothiophene-3-carboxamide (100 mg, 0.35 mmol) was added and stirred 12 h at room temperature. The reaction mixture was concentrated under reduced pressure. Yield: 16 mg, 12%. Brown solid. Purity, >95% (assessed by LC/MS). LC/MS (APSI) m/z: [M + H] calcd for C₂₃H₂₉N₃O₂S, 412.2; found, 412.2.

N-Benzyl-2-{[(4-methyl-1-piperazinyl) Acetyl] amino}-4,5,6,7-tetrahydro-1-benzothiophene-3-carboxamide (10)

N-benzyl-2-(2-chloroacetamido) −4,5,6,7-tetrahydro-1-benzothiophene-3-carboxamide (compound 1) (3.63 g, 10 mmol) and 1-methylpiperazine (compound 2) (1.00 g, 10 mmol) were dissolved in 45 mL of DMF, and K₂CO₃ (2.07 g, 15 mmol) was added to the mixture. After that, the mixture was stirred for 12 h at 50 °C. Then, the mixture was cooled to room temperature, and cold water (250 mL) was added. The product was extracted with EtOAc (3 × 50 mL), washed with brine (3 × 50 mL), and the organic layer was dried with Na₂SO₄. The solvent was removed, and the residue was purified using HPLC. Brown oil. Yield: 42%; purity, >95% (assessed by LC/MS). ¹H NMR (400 MHz, DMSO-d₆): δ 11.84 (s, 1H), 7.95 (s, 1H), 7.32 (h, J = 6.6 Hz, 4H), 7.23 (t, J = 6.9 Hz, 1H), 4.46 (d, J = 5.8 Hz, 2H), 3.14 (s, 2H), 2.72 (d, J = 6.5 Hz, 2H), 2.60 (d, J = 6.6 Hz, 2H), 2.45 (s, 3H), 2.34 (s, 4H), 2.14 (s, 3H), 1.77–1.70 (m, 5H). LC/MS (APSI) m/z: [M + H] calcd for C₂₃H₃₀N₄O₂S, 427.2; found, 427.2.

N-Benzyl-2-{[(4-methyl-1-piperidinyl) Acetyl] amino}-4,5,6,7-tetrahydro-1-benzothiophene-3-carboxamide (11)

2-(4-methyl piperidine-1-yl) acetic acid (0.35 mmol) was dissolved in dioxane (15 mL), MsCl (0.4 mmol), and NEt₃ (0.4 mmol) was added at 0 °C. The mixture was stirred at room temperature for 1 h. 2-amino-N-benzyl-4,5,6,7-tetrahydro-1-benzothiophene-3-carboxamide (100 mg, 0.35 mmol) was added and stirred 12 h at room temperature. The reaction mixture was concentrated under reduced pressure. Yield: 15 mg, 10%. Light brown solid. Purity, >95% (assessed by LC/MS). LC/MS (APSI) m/z: [M + H] calcd for C₂₄H₃₁N₃O₂S, 426.2; found, 426.2.

N-(3-(2-Methylpyrrolidine-1-carbonyl)-4,5,6,7-tetrahydro-benzo[b]thiophen-2-yl) Nicotinamide (12)

(Method 1) (2-amino-4,5,6,7-tetrahydrobenzo[b]thiophen-3-yl) (2-methylpyrrolidin-1-yl)methanone (100 mg), nicotinic acid (1.1 mol equiv to the amine), and 0.5 mL of DMSO were placed into a 4 mL capped glass vial, and the mixture was stirred for 30 min. EDC (1.2 mol equiv to the amine) was added. The mixture was stirred for 1 h and left overnight at room temperature. The solution was filtered, and the solvent and volatile components were evaporated under reduced pressure to give the crude product. The product was further purified by HPLC. Yield: 67%; purity, >95% (assessed by LC/MS). ¹H NMR (500 MHz, DMSO-d₆): δ 11.33 (s, 1H), 10.95 (s, 1H), 8.96 (d, J = 2.3 Hz, 1H), 8.73 (d, J = 4.8 Hz, 1H), 8.20–8.14 (m, 1H), 2.64 (d, J = 6.3 Hz, 2H), 1.75 (s, 2H), 1.51 (s, 1H), 1.15 (s, 1H), 0.80 (s, 1H). LC/MS (APSI) m/z: [M + H] calcd for C₂₀H₂₃N₃O₂S, 370.2; found, 370.2.

N-(3-((2R,4R)-2,4-Dimethylpiperidine-1-carbonyl)-4,5,6,7-tetrahydrobenzo[b]thiophen-2-yl) Nicotinamide (13)

Step A: methyl 2-amino-4,5,6,7-tetrahydro-1-benzothiophene-3-carboxylate (21.1 g, 100 mol) was dissolved in dry DMF (150 mL) followed by addition of HATU (100 mmol), pyridine-3-carboxylic acid (13.5 g, 110 mmol), and DIPEA (300 mmol) at room temperature. The reaction mixture was stirred at ambient temperature for 12 h. LC/MS analysis showed complete conversion, and the mixture was concentrated under reduced pressure. The residue was redissolved in EtOAc (1 L), and the organic phase was washed with 0.1 M HCl (aqueous) and brine, dried (MgSO₄), and filtered. Evaporation of the solvent under reduced pressure gave compound 3 which was purified by chromatography. Yield: 65%. Step B: 2-(pyridine-3-amido)-4,5,6,7-tetrahydro-1-benzothiophene-3-carboxylic acid (50 mmol) and CDI (60 mmol) were dissolved in dry DMF (100 mL). The mixture was stirred at room temperature for 30 min and then at 70 °C for 1 h. Rac-(2R,4R) −2,4-dimethylpiperidine (60 mmol) was added. The mixture was heated while stirring for 12 h at 50 °C, cooled to room temperature, and water (300 mL) was added. The residue was filtered out, washed with water (3 × 100 mL), dried at 60 °C, and purified using HPLC. Yield: 38%. Step C: the mixture of rac-N-{3-[(2R,4R)-2,4-dimethylpiperidine-1-carbonyl]-4,5,6,7-tetrahydro-1-benzothiophen-2-yl}pyridine-3-carboxamide and 2-(pyridin-3-yl)-5,6,7,8-tetrahydro-4H-benzo[4,5]thieno[2,3-d][1,3]oxazin-4-one (25 mmol) and rac-(2R,4R)-2,4-dimethylpiperidine (50 mmol) were dissolved in dry DMF (50 mL), and the mixture was stirred at 100 °C for 24 h (LC\MS control). The reaction mixture was concentrated under reduced pressure. The residue was mixed with water (100 mL), and iPrOH (30 mL) was added. The mixture was stirred at room temperature for 2 h, filtered, and washed with water (3 × 50 mL), and residue was purified using HPLC. Yield: 24%. Step D: for preparative separation, 120 mg of rac-N-{3-[(2R,4R)-2,4-dimethylpiperidine-1-carbonyl]-4,5,6,7-tetrahydro-1-benzothiophen-2-yl}pyridine-3-carboxamide was dissolved in methanol (0.9 mL), degassed by an ultrasonic bath. Separations were performed in the isocratic mode using an Agilent 1260 HPLC instrument, Chiralpak IA (250 × 20 mm i.d; 5 μm) with guard column Chiralpak IA (50 × 20 mm i.d; 5 μm) stationary phase and MeOH/IPA mixtures with 70% of MeOH mobile phase. Enantiomers were collected and evaporated to obtain solid samples. The enantiomeric excess (ee %) of R,R isomer solution in methanol was determined in comparison with racemate at 21 °C and 589 nm. Yield: 37%. ¹H NMR (400 MHz, DMSO-d₆): δ 10.91 (d, J = 82.0 Hz, 1H), 9.01 (dd, J = 18.9, 2.3 Hz, 1H), 8.75 (d, J = 4.8 Hz, 1H), 8.22 (dd, J = 21.5, 8.2 Hz, 1H), 7.55 (dd, J = 8.0, 4.9 Hz, 1H), 4.13 (dd, J = 13.0, 6.9 Hz, 1H), 3.62–3.41 (m, 1H), 3.22–3.04 (m, 1H), 2.67 (d, J = 5.2 Hz, 3H), 2.34 (d, J = 7.7 Hz, 2H), 1.73 (ddd, J = 32.9, 16.5, 10.7 Hz, 8H), 1.14 (dd, J = 11.7, 6.6 Hz, 4H), 0.95 (d, J = 6.1 Hz, 1H), 0.84 (d, J = 6.5 Hz, 2H). LC/MS (APSI) m/z: [M + H] calcd for C₂₂H₂₇N₃O₂S, 398.0; found, 398.0. [α]₂₁^D −71.67 (c 0.2, CH3OH). Crystal data: C₂₂H₂₇N₃O₂S, M = 397.52, orthorhombic, space group P2₁2₁2₁, a = 15.766(2), b = 16.081(3), c = 16.835(3) Å, V = 4268.3(11) ų, Z = 8, d_c = 1.237 g·cm^–3, μ = 0.173 mm^–1, F(000) = 1696, crystal size ca. 0.17 × 0.19 × 0.55 mm. All crystallographic measurements were performed at 173 K on a Bruker Smart Apex II diffractometer operating in the ω-scans mode with graphite mono-chromated Mo Kα radiation (λ = 0.71078 Å). The intensity of 23,644 reflections (7816 unique reflections, R_merg = 0.0765) was collected within the θ_max ≤ 25.4°. The structure was solved by direct methods and refined by the full-matrix least-squares technique in the anisotropic approximation for non-hydrogen atoms using the Bruker SHELXTL program package. All CH hydrogen atoms were placed at calculated positions and refined as a “riding” model. The NH hydrogen atoms were found in the difference Fourier syntheses and refined iso-tropically. Convergence was obtained at R₁ = 0. 0668 and wR₂ = 0.1374 for 5179 observed reflections with I ≥ 2σ(I), R₁ = 0. 1117, wR₂ = 0.1601, GOF = 1.000 for 7816 independent reflections, 513 parameters, and the largest and minimal peaks in the final difference map 0.57 and −0.26 e/ų. Full crystallographic details have been deposited at Cambridge Crystallographic Data Centre (CCDC). Any request to the CCDC for these materials should quote the full literature citation and reference number CCDC 2257780.

N-(3-(((1R,4R)-4-Hydroxycyclohexyl)carbamoyl)-4,5,6,7-tetrahydrobenzo[b]thiophen-2-yl) Nicotinamide (14)

(Method1) 2-amino-N-((1R,4R)-4-hydroxycyclohexyl)-4,5,6,7-tetrahydrobenzo [b]thiophene-3-carboxamide (100 mg), nicotinic acid (1.1 mol equiv to the amine), and 0.5 mL of DMSO were placed into a 4 mL capped glass vial, and the mixture was stirred for 30 min. EDC (1.2 mol equiv to the amine) was added. The mixture was stirred for 1 h and left overnight at room temperature. The solution was filtered, and the solvent and volatile components were evaporated under reduced pressure to give the crude product. The product was further purified by HPLC. Yield: 63%; purity, >95% (assessed by LC/MS). ¹H NMR (500 MHz, DMSO-d₆): δ 12.15 (s, 1H), 9.03 (d, J = 2.3 Hz, 1H), 8.78 (s, 1H), 8.20 (dt, J = 8.1, 2.0 Hz, 1H), 7.61 (t, J = 6.3 Hz, 1H), 7.32 (s, 1H), 4.55 (d, J = 4.3 Hz, 1H), 3.80–3.63 (m, 1H), 2.66 (dt, J = 13.4, 6.0 Hz, 4H), 1.89–1.68 (m, 9H), 1.40–1.16 (m, 4H). LC/MS (APSI) m/z: [M + H] calcd for C₂₁H₂₅N₃O₃S, 400.2; found, 400.2.

N-(3-(3-Ethylpyrrolidine-1-carbonyl)-4,5,6,7-tetrahydro-benzo[b]thiophen-2-yl)nicotinamide (15)

(Method 1) (2-amino-4,5,6,7-tetrahydrobenzo[b]thiophen-3-yl) (3-ethylpyrrolidin-1-yl)methanone (100 mg), nicotinic acid (1.1 mol equiv to the amine), and 0.5 mL of DMSO were placed into a 4 mL capped glass vial, and the mixture was stirred for 30 min. EDC (1.2 mol equiv to the amine) was added. The mixture was stirred for 1 h and left overnight at room temperature. The solution was filtered, and the solvent and volatile components were evaporated under reduced pressure to give the crude product. The product was further purified by HPLC. Yield: 65%; purity, >95% (assessed by LC/MS). ¹H NMR (500 MHz, DMSO-d₆): δ 11.09 (s, 1H), 8.97 (s, 1H), 8.73 (dd, J = 4.7, 1.6 Hz, 1H), 8.22–8.13 (m, 1H), 7.55 (dd, J = 8.0, 4.9 Hz, 1H), 3.52 (s, 2H), 2.93 (d, J = 58.9 Hz, 1H), 2.64 (t, J = 6.0 Hz, 2H), 2.36 (s, 1H), 1.99 (s, 2H), 1.82–1.57 (m, 4H), 1.44 (t, J = 10.0 Hz, 1H), 1.30 (s, 3H), 0.80 (s, 2H). LC/MS (APSI) m/z: [M + H] calcd for C₂₁H₂₅N₃O₂S, 384.2; found, 384.2.

N-Benzyl-2-(2-((1-methyl-1H-imidazole-2-yl)thio)acetamido)-4,5,6,7-tetrahydrobenzo[b]thiophene-3-carboxamide (16)

(Method 2) 2-amino-N-benzyl-4,5,6,7-tetrahydrobenzo[b] thiophene-3-carboxamide (100 mg), DIPEA (1.2 mol equiv to the amine), and DMSO (0.5 mL) were placed into a 4 mL capped glass vial and stirred for 30 min. 2-((1-methyl-1H-imidazole-2-yl)thio)acetyl chloride (1.2 mol equiv to the amine) was added, and the vial was stirred for 1 h at room temperature. Then, the vial was placed into a thermostat (set to 100 °C) for 9 h. After cooling, the mixture was filtered; the solvent and volatile components were evaporated under reduced pressure to give the crude product. The product was further purified by HPLC. Yield: 65%; purity, >95% (assessed by LC/MS). ¹H NMR (500 MHz, DMSO-d₆): δ 11.75 (s, 1H), 8.03 (t, J = 6.0 Hz, 1H), 7.30 (d, J = 6.6 Hz, 4H), 7.22 (s, 1H), 7.17 (d, J = 1.4 Hz, 1H), 6.86 (d, J = 1.3 Hz, 1H), 4.44 (d, J = 5.9 Hz, 2H), 3.96 (s, 2H), 3.54 (s, 3H), 2.64 (s, 2H), 2.59 (d, J = 5.6 Hz, 2H), 1.70 (s, 4H). LC/MS (APSI) m/z: [M + H] calcd for C₂₂H₂₄N₄O₂S₂, 441.2; found, 441.2.

N-(3-(3-(Hydroxymethyl)pyrrolidine-1-carbonyl)-4,5,6,7-tetrahydrobenzo[b]thiophen-2-yl) Nicotinamide (17)

(Method 1) (2-amino-4,5,6,7-tetrahydrobenzo[b]thiophen-3-yl) (3-(hydroxymethyl)pyrrolidin-1-yl)methanone (100 mg), a carboxylic acid (1.1 mol equiv to the amine), and 0.5 mL of DMSO were placed into a 4 mL capped glass vial, and the mixture was stirred for 30 min. EDC (1.2 mol equiv to the amine) was added. The mixture was stirred for 1 h and left overnight at room temperature. Solution was filtered, and the solvent and volatile components were evaporated under reduced pressure to give the crude product. The product was further purified by HPLC. Yield: 59%; purity, >95% (assessed by LC/MS). ¹H NMR (500 MHz, DMSO-d₆): δ 11.10 (s, 1H), 8.97 (d, J = 2.3 Hz, 1H), 8.73 (dd, J = 4.8, 1.6 Hz, 1H), 8.18 (dt, J = 7.9, 1.9 Hz, 1H), 7.54 (dd, J = 8.0, 4.8 Hz, 1H), 4.63 (s, 1H), 3.52 (s, 1H), 2.63 (t, J = 5.9 Hz, 2H), 2.43–2.14 (m, 3H), 1.97–1.41 (m, 6H). LC/MS (APSI) m/z: [M + H] calcd for C₂₀H₂₃N₃O₃S, 386.2; found, 386.2.

N-(6,6-Dimethyl-3-(morpholine-4-carbonyl)-4,5,6,7-tetrahydrobenzo[b]thiophen-2-yl) Pyrazine-2-carboxamide (18)

(Method 3) (2-amino-6,6-dimethyl-4,5,6,7-tetrahydrobenzo[b] thiophen-3-yl) (morpholino)methanone (100 mg), pyrazine-2-carboxylic acid (1.2 mol equiv to the amine), DIPEA (1.2 mol equiv to the amine), and acetonitrile (0.5 mL) were placed into a 4 mL capped glass vial and stirred for 30 min. After the addition of 2-chloro-1-methylpyridine-1-ium iodide (1.44 mol equiv to the amine), the vial was stirred for 1 h at room temperature. Then, the vial was placed into a thermostat (set to 100 °C) for 6 h. After cooling, the mixture was filtered; the solvent and volatile components were evaporated under reduced pressure to give the crude product. The product was further purified by HPLC. Yield: 56%; purity, >95% (assessed by LC/MS). ¹H NMR (500 MHz, DMSO-d₆): δ 10.94 (s, 1H), 9.27 (d, J = 1.5 Hz, 1H), 8.95 (d, J = 2.4 Hz, 1H), 8.83 (dd, J = 2.5, 1.5 Hz, 1H), 3.63–3.35 (m, 9H), 3.12 (d, J = 8.0 Hz, 1H), 2.45 (d, J = 15.5 Hz, 4H), 1.49 (t, J = 6.4 Hz, 2H), 1.23 (q, J = 7.3 Hz, 6H), 0.97 (s, 7H). LC/MS (APSI) m/z: [M + H] calcd for C₂₀H₂₄N₄O₃S, 401.2; found, 401.2.

N-(3-(((1-Methyl-1H-1,2,4-triazol-3-yl)methyl)carbamoyl)-4,5,6,7-tetrahydrobenzo[b]thiophen-2-yl)nicotinamide (19)

(Method 1) 2-amino-N-((1-methyl-1H-1,2,4-triazol-3-yl)methyl)-4,5,6,7-tetrahydrobenzo[b]thiophene-3-carboxamide (100 mg), nicotinic acid (1.1 mol equiv to the amine), and 0.5 mL of DMSO were placed into a 4 mL capped glass vial, and the mixture was stirred for 30 min. EDC (1.2 mol equiv to the amine) was added. The mixture was stirred for 1 h and left overnight at room temperature. The solution was filtered, and the solvent and volatile components were evaporated under reduced pressure to give the crude product. The product was further purified by HPLC. Yield: 57%; purity, >95% (assessed by LC/MS). LC/MS (APSI) m/z: [M + H] calcd for C₁₉H₂₀N₆O₂S, 397.0; found, 397.0.

(S)-N-(3-((1-Cyanoethyl)carbamoyl)-4,5,6,7-tetrahydro-benzo[b]thiophen-2-yl)nicotinamide (20)

(Method 1) (R)-2-amino-N-(1-cyanoethyl)-4,5,6,7-tetrahydrobenzo[b]thiophene-3-carboxamide (100 mg), nicotinic acid (1.1 mol equiv to the amine), and 0.5 mL of DMSO were placed into a 4 mL capped glass vial, and the mixture was stirred for 30 min. EDC (1.2 mol equiv to the amine) was added. The mixture was stirred for 1 h and was left overnight at room temperature. The solution was filtered, and the solvent and volatile components were evaporated under reduced pressure to give the crude product. The product was further purified by HPLC. Yield: 67%; purity, >95% (assessed by LC/MS). LC/MS (APSI) m/z: [M + H] calcd for C₁₈H₁₈N₄O₂S, 355.0; found, 355.0.

N-(3-((piperidin-4-ylmethyl)carbamoyl)-4,5,6,7-tetrahydro-benzo[b]thiophen-2-yl) Nicotinamide (21)

(Method 1) 2-amino-N-(piperidin-4-ylmethyl)-4,5,6,7-tetrahydrobenzo[b] thiophene-3-carboxamide (100 mg), nicotinic acid (1.1 mol equiv to the amine), and 0.5 mL of DMSO were placed into a 4 mL capped glass vial, and the mixture was stirred for 30 min. EDC (1.2 mol equiv to the amine) was added. The mixture was stirred for 1 h and left overnight at room temperature. The solution was filtered, and the solvent and volatile components were evaporated under reduced pressure to give the crude product. The product was further purified by HPLC. Yield: 63%; purity, >95% (assessed by LC/MS). ¹H NMR (500 MHz, DMSO-d₆): δ 9.09 (d, J = 2.1 Hz, 1H), 8.54 (d, J = 4.7 Hz, 1H), 8.20 (dd, J = 7.9, 2.1 Hz, 1H), 7.41 (dd, J = 7.9, 4.8 Hz, 1H), 3.25–3.13 (m, 4H), 2.85–2.69 (m, 3H), 2.52 (d, J = 6.5 Hz, 2H), 1.79 (t, J = 10.4 Hz, 3H), 1.71–1.62 (m, 4H), 1.31 (q, J = 13.6, 12.9 Hz, 2H). LC/MS (APSI) m/z: [M + H] calcd for C₂₁H₂₆N₄O₂S, 399.2; found, 399.2.

N-(3-Benzoyl-4,5,6,7-tetrahydrobenzo[b]thiophen-2-yl)pyrazine-2-carboxamide (22)

(Method 3) (2-amino-4,5,6,7-tetrahydrobenzo[b]thiophen-3-yl) (phenyl)methanone (100 mg), pyrazine-2-carboxylic acid (1.2 mol equiv to the amine), DIPEA (1.2 mol equiv to the amine), and acetonitrile (0.5 mL) were placed into a 4 mL capped glass vial and stirred for 30 min. After the addition of 2-chloro-1-methylpyridine-1-ium iodide (1.44 mol equiv to the amine), the vial was stirred for 1 h at room temperature, then placed into a thermostat (set to 100 °C) for 6 h. After cooling, the mixture was filtered; the solvent and volatile components were evaporated under reduced pressure to give the crude product. The product was further purified by HPLC. Yield: 58%; purity, >95% (assessed by LC/MS). ¹H NMR (500 MHz, DMSO-d₆): δ 12.58 (s, 1H), 9.31 (s, 1H), 8.96 (d, J = 2.5 Hz, 1H), 8.81 (d, J = 2.6 Hz, 1H), 7.63–7.45 (m, 5H), 2.70 (t, J = 6.4 Hz, 2H), 1.96 (t, J = 6.1 Hz, 2H), 1.73 (t, J = 5.9 Hz, 2H), 1.50 (d, J = 7.6 Hz, 2H). LC/MS (APSI) m/z: [M + H] calcd for C₂₀H₁₇N₃O₂S, 364.0; found, 364.0.

N-(3-Benzoyl-4,5,6,7-tetrahydrobenzo[b]thiophen-2-yl)-5-methyl-2-(methylsulfonyl) Pyrimidine-4-carboxamide (23)

(Method 3) (2-amino-4,5,6,7-tetrahydrobenzo[b]thiophen-3-yl) (phenyl)methanone (100 mg), 5-methyl-2-(methylsulfonyl)pyrimidine-4-carboxylic acid (1.2 mol equiv to the amine), DIPEA (1.2 mol equiv to the amine), and acetonitrile (0.5 mL) were placed into a 4 mL capped glass vial and stirred for 30 min. After the addition of 2-chloro-1-methylpyridine-1-ium iodide (1.44 mol equiv to the amine), the vial was stirred for 1 h at room temperature. Then, the vial was placed into a thermostat (set to 100 °C) for 6 h. After cooling, the mixture was filtered; the solvent and volatile components were evaporated under reduced pressure to give the crude product. The product was further purified by HPLC. Yield: 62%; purity, >95% (assessed by LC/MS). ¹H NMR (500 MHz, DMSO-d₆): δ 12.82 (s, 1H), 9.21 (s, 1H), 7.63–7.45 (m, 5H), 3.57 (s, 3H), 2.70 (s, 5H), 1.94 (s, 2H), 1.72 (s, 2H), 1.50 (s, 2H). LC/MS (APSI) m/z: [M + H] calcd for C₂₂H₂₁N₃O₄S₂, 456.1; found, 456.1.

TR-FRET Co-activator Recruitment Assay

5 μL of each compound in an assay buffer is added to 15 μL of detection mix for 20 μL of total assay volume (50 mM Tris-HCl pH = 7.0, 150 mM NaCl, 50 mM KCl, 5 mM MgCl₂, 1 mM DTT, 0.1% BSA, 0.001% triton X); 1.5 nM GST–RORγt (LBD) expressed in insect cells (creative biomart, Shirley, NY, USA); 90 nM biotinylated RIP140 (biotinyl-NH-Ahx-NSHQKVTLLQLLLGHKNEEN-CONH2) coactivator peptide (CPC Scientific, San Jose, CA, USA); 50 nM streptavidin-APC; 1.5 nM Eu-anti GST IgG. Reactions were read with an EnSpire plate reader, (excitation: 320 nm, emission A: 615 nm; and emission B: 699 nm) to measure the emission ratio acceptor/donor. Final compound concentrations range from 0.004 nM to 100 μM. 100% activity is represented by average DMSO controls. Zero % activity is average of two wells of 100 μM gsk2981278 controls.

FP Competitive Binding Assay

Competition assays were performed with 10 nM full length RORγ, expressed in HEK293T cells (OriGene Technologies, Rockville, MD, USA), loaded with 5 nM 25NC. The mixture was incubated with serial dilutions of compounds (0.004 nM to 100 μM). Parallel and perpendicular fluorescence intensity were measured with synergy 2 plate reader (Excitation 497 nm and emission 551 nm), and polarization was reported in milli polarization. Ki values were determined by the Cheng–Prusoff equation.⁶⁵ ΔH and ΔS were determined by the Van’t Hoff equation. ΔG was determined by the Gibbs free-energy equation.⁶⁶

RORγ Reporter Assay

The RORγ reporter assay kit (INDIGO Biosciences, State College, PA, USA) was used to measure the inhibitory activity of the compounds compared to standard RORγ inverse agonist ursolic acid. 200 μL of reporter cells was dispensed into each well of the assay plate. The cells were preincubated for 4–6 h. Following the pre-incubation period, culture media were discarded, and 200 μL/well of the prepared 1× concentration treatment media are added. Following 24 h incubation, treatment media were discarded, and luciferase detection reagents were added to each well. The intensity of light emission in units of relative light units (RLUs) from each well was quantified using a plate-reading luminometer.

Th17 Polarization

PBMCs were plated in complete RPMI-1640 medium (10% FBS) and incubated at 37 °C in a humidified, 5% CO₂ atmosphere, treated with CD3/CD28 (2 μL: 80,000 cells) for 12 days in the presence of IL6 10 ng/mL, IL1B 10 ng/mL, TGFβ 10 ng/mL, and IL23 10 ng/mL. Compounds were added to the cells on day 12 for 48 h. On day 14, cells were treated with cell stimulation cocktail for 5 h at the end of 48 h incubation. PBMCs were stained cells with live/dead (455 UV) viability dye. Anti-human antibodies for IL17A, RORγt, IL21, IL22, and IFNγ were added to wells at 1:1000 concentration in permeabilization buffer. Cell viability and intracellular cytokines were analyzed by flowcytometry.

Kinetic Solubility

Compound dilutions were prepared in duplicates in phosphate-buffered saline (PBS, pH = 7.4), 100 mM citrate buffer (pH = 5.0), or fasted state simulated gastric fluid (FaSSGF, pH = 1.6) with 2% DMSO. The dilutions in PBS were allowed to equilibrate at 25 °C on a thermostatic shaker for 2 h and then filtered through HTS filter plates. The filtrates of test compounds were diluted two-fold with acetonitrile with 2% DMSO before measuring. Compound dilutions in 50% acetonitrile/PBS were prepared (2% final DMSO) to generate calibration curves. Ondansetron was used as a reference compound to control the proper assay performance. For PBS and citrate buffer solubility, 200 μL of each sample was transferred to a 96-well plate and measured by an UV spectrometer/reader in the 260–300 nm range and absolute absorbance unit values for the minimum and maximum concentration points within the 0–3 OD range. For FaSSGF solubility, samples were analyzed using the HPLC system coupled with a tandem mass spectrometer.

Microsomal Stability

Microsomal incubations were carried out in 96-well plates in five aliquots of 30 μL each (one for each time point). Liver microsomal incubation medium comprised phosphate buffer (100 mM, pH = 7.4), MgCl₂ (3.3 mM), NADPH (3 mM), glucose-6-phosphate (5.3 mM), and glucose-6-phosphate dehydrogenase (0.67 units/mL) with 0.42 mg of liver microsomal protein per milliliter. In the control reactions, the NADPH cofactor system was substituted with phosphate buffer. Compounds (2 μM) were incubated with microsomes at 37 °C, shaking at 100 rpm. Each reaction was performed in duplicates. Five time points over 40 min were analyzed. Reactions were stopped by adding methanol containing internal standard, followed by protein sedimentation by centrifuging. Supernatants were analyzed using the HPLC system coupled with tandem mass spectrometer. Intrinsic clearance (Cl_in) was determined in plot of ln (AUC) versus time, using linear regression analysis according to the equation

Caco-2 Permeability

Caco-2 cells were suspended in the DMEM complete medium to a final concentration 6 × 10⁵ cells/mL. 0.4 mL of the cell suspension was added to each well of the HTS 24 multi-well insert system, and 25 mL of pre-warmed DMEM complete medium was added to the feeder tray. The plates were incubated for 6–10 days before the transport experiments, then placed in sterile 24-well transport analysis plates. The inserts were washed with PBS after medium aspiration. Ketoprofen, atenolol, quinidine, and digoxin were used as reference compounds. To determine the rate of compounds transport in apical (A) to basolateral (B) direction, 0.3 mL of compounds solutions in transport buffer (Hank’s BSS (9.5 g/L), NaHCO₃ (0.35 g/L), 0.81 mM MgSO₄, 1.26 mM CaCl₂, 25 mM HEPES, pH = 7.4) were added into the filter wells and 1 mL of transport buffer was added to transport analysis plate wells. To determine transport rates in the B–A direction, 1 mL of compound solutions was added into the wells of the transport analysis plate, and the wells in the filter plate were filled with 0.3 mL of buffer (apical compartment). P-gp-mediated transport of the compounds was assessed by determining the bidirectional transport in the presence or absence of verapamil. Plates were incubated for 90 min at 37 °C under continuous shaking. Aliquots were taken from the donor and receiver compartments for LC–MS/MS analysis. Samples were mixed with acetonitrile followed by protein sedimentation by centrifuging. Supernatants were analyzed using the HPLC system coupled with a tandem mass spectrometer.

MDR1-MDCKII Permeability

MDR1-MDCKII cells (Sigma-Aldrich, Saint Louis, MO, USA) were cultured in 75 cm² flasks to 80–90% confluency in a humidified atmosphere at 37 °C and 5% CO₂. Cells were detached with trypsin/EDTA solution and resuspended in the cell culture medium to a final concentration of 4 × 105 cells/mL. Then, 400 μL of the cell suspension was added to each well of the HTS multiwell insert system, and 25 mL of prewarmed complete medium was added to the feeder tray. Confluent MDR1-MDCKII monolayers expressing P-gp were obtained after 4–5 days post seeding. The integrity of the cell monolayers was determined by measuring the trans-epithelial electrical resistance (TEER, Ω × cm²) using an epithelial voltammeter. A mature MDR1-MDCKII cell monolayer exhibited TEER values >100 Ω cm². A 24-well insert plate was removed from its feeder tray and placed in a new sterile 24-well transport analysis plate. The inserts were washed with PBS after medium aspiration. To determine the rate of compound transport in the A–B direction, 300 μL of the test compound dissolved in transport buffer (9.5 g/L Hanks’ BSS and 0.35 g/L NaHCO₃ with 0.81 mM MgSO₄, 1.26 mM CaCl₂, 25 mM HEPES, pH adjusted to 7.4) was added into the filter wells; 1000 μL of transport buffer was added to transport analysis plate wells. Ketoprofen, atenolol, digoxin, and quinidine were used as reference compounds. To determine transport rates in the B–A direction, 1000 μL of the test compound solutions were added into the wells of the transport analysis plate, and the wells in the filter plate were filled with 300 μL of buffer (apical compartment). The final concentration of the test compounds was 10 μM. The effect of the inhibitor on the P-gp-mediated transport of the tested compounds was assessed by determining the bidirectional transport in the presence or absence of cyclosporine A. The MDR1-MDCKII cells were preincubated for 30 min at 37 °C with 10 μM of cyclosporine A in both apical and basolateral compartments. After removal of the preincubation medium, the test compounds (final concentration 10 μM) with cyclosporine A (10 μM) in transport buffer were added in donor wells, while the receiver wells were filled with the appropriate volume of transport buffer with 10 μM cyclosporine A, respectively. The plates were incubated for 90 min at 37 °C under continuous shaking at 100 rpm. After this, 75 μL of aliquots was taken from the donor and receiver compartments for LC–MS/MS analysis. All samples were mixed with two volumes of acetonitrile followed by protein sedimentation by centrifuging at 10,000 rpm for 10 min. Supernatants were analyzed using the HPLC system coupled with a tandem mass spectrometer.

Glossary

Abbreviations

25HC

25-hydroxycholesterol

25NC

25-NBD cholesterol

APCI

atmospheric pressure chemical ionization

APSI

appearance potential soft ionization

CDI

1,1′-carbonyldiimidazole

Cl_in

intrinsic clearance

DBD

DNA-binding domains

DIPEA

N,N-diisopropylethylamine

E_bconf

energy of conformer (pocket bound)

E_conf

energy of conformer

EDC

1-ethyl-3-(3-dimethylaminopropyl) carbodiimide

FEP

free-energy perturbations

FES

free-energy surface

FP

fluorescence polarization

G

Gibbs energy

G_EXP

experimental Gibbs energy

G_FEP

FEP calculated Gibbs energy

HATU

1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxid hexafluorophosphate, N-[(dimethylamino)-1H-1,2,3-triazolo-[4,5-b]pyridin-1-ylmethylene]-N-methylmethanaminium hexafluorophosphate N-oxide

HTVS

high-throughput virtual screening

LipE

lipophilic efficiency

LipE_ilogP

lipophilic efficiency based on ilogP

LipE_logD

lipophilic efficiency based on log D

MDR1-MDCKII

canine MDR1 knock-out, human MDR1 knock-in MDCKII

MsCl

methane-sulfonyl chloride

MUE

mean unsigned error

NEt₃

triethylamine

RDF

radial distribution function

RLU

relative light units

rmsf

root-mean-square fluctuation

RORγt

retinoic acid receptor-related orphan receptor γt

SP

standard precision

TPSA

topological polar surface area

TR-FRET

time-resolved fluorescence energy transfer

XP

extra precision

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jmedchem.3c00021.

• LC/MS traces, ¹H NMR spectra of compounds 1–23, detailed synthesis schemes of select compounds, chiral separation, and X-ray crystal structure of compound 13 (PDF)

• Molecular formula/SMILES of compounds 1–23 (CSV)

• Molecular formula/SMILES of 1500 conformers’ library (CSV)

• Docking poses of 2, 10, and 13 superimposed on 25HC (PDB ID: 5APH)(PDB)

• Docking poses of 10 superimposed on gsk2981278 & iag (PDB ID: 5APH) (PDB)

• Docking poses of 2 and 3 (PDB ID: 5APH) (PDB)

• Docking poses of 13 and 14 (PDB ID: 5APH) (PDB)

The authors declare no competing financial interest.