Abstract
The farnesoid X receptor (FXR) is a crucial therapeutic target for treating non-alcoholic steatohepatitis (NASH). Although obeticholic acid (OCA) as a FXR agonist presents good efficacy, the safety data such as severe pruritus should be carefully considered. To discover new medications, we screen and choose the optimal compounds from ZINC15 database that may agonistically interact with FXR. We utilized the DS19 software to assist us in conducting the computer-aided structure based virtual screening to discover potential FXR agonists. After LibDock scores were determined by screening, their absorption, distribution, metabolism, excretion and toxicity predictions were examined. To determine the binding affinity between the chosen drugs and FXR, molecule docking was utilized. Molecular dynamics simulation was utilized to evaluate the stabilization of the ligand-FXR complex in its native environment. Higher binding affinity and stability with FXR were observed for ZINC000013374322 and ZINC000006036327, as two novel natural compounds, with lower rodent carcinogenicity, Ames mutagenicity, no hepatotoxicity and non-inhibitors of CYP2D6. They could stably exist in the environment, possess favorable potential energy and exert pharmacological effects at lower doses. Furthermore, ZINC000006036327 had lower skin irritancy and sensitization potential compared to OCA, also suggest the possibility of improved skin itching occurrence. ZINC000013374322 and ZINC000006036327 were found to be the best leading compounds to be FXR agonists. They are chosen as safe candidates for FXR target medicine, which play comparable pharmacological effects at lower doses.
Keywords: Non-alcoholic steatohepatitis (NASH), Discovery studio, Farnesoid X receptor (FXR), Obeticholic acid (OCA), Novel compounds
Subject terms: Medical research, Drug discovery, Virtual screening
Introduction
Non-alcoholic fatty liver disease (NAFLD) has an annual incidence of 6–35% worldwide1,2. Approximately 15% to 25% cases with NAFLD will develop into non-alcoholic steatohepatitis (NASH), which subsequently resulting in cirrhosis and hepatocellular carcinoma. NAFLD also augments the susceptibility to obesity, metabolic syndrome (MetS), type 2 diabetes (T2DM), cardiovascular diseases, and non-hepatic neoplasms3. The complicated pathogenesis of NASH is primarily responsible for the challenges encountered in drug research and development. Its pathogenesis involving the interaction of various routes, targets and cells, has not yet been determined. Delightfully, the FDA approves Resmetirom on March 14, 2024, a selective thyroid hormone receptor-β (THR-β) agonist as the first drug for NASH patients. However, only 25.9–29.9% of patients had NASH resolution without their fibrosis getting worse and the percentage of patients who showed at least one stage of improvement in their fibrosis without the NAFLD activity score worsening was only 24.2–25.9% in the resmetirom group4. Moreover, Resmetirom exhibited an adverse event rate nearly twice that of the placebo group in Phase III trials, primarily attributed to occurrences of diarrhea and nausea5. Importantly, other pharmacological treatments are being developed for the treatment of NASH, including fibroblast growth factor 21, glucagon-like peptide-1 analogues, and farnesoid X receptor (FXR) agonists6.
FXR as the bile acid receptor, is a crucial member of the nuclear receptor superfamily, affecting the metabolism of bile acids as well as in the metabolism of carbohydrates, lipids, and sterols7,8. Upon binding with its ligand, FXR forms a heterodimer with the retinoid X receptor (RXR), subsequently activates the small heterodimer partner (SHP), which exerts agonistic effects on downstream target genes through dimerization with various nuclear receptors, including hepatocyte nuclear factor 4α (HNF4α), liver X receptor (LXR) and peroxisome proliferators-activated receptors (PPARs)9–13. This regulatory mechanism plays a key role in maintaining lipid, glucose, energy, and immune homeostasis14. Consequently, FXR agonists are of significant therapeutic value in the field of NASH15,16. Obeticholic acid (OCA) as a FXR agonists is the first NASH medication in history to show promise in Phase III trials with safety debate17. Presently, several FXR agonists have been already advanced into clinical trial phases18,19.
OCA, as the second-line treatment for primary biliary cholangitis, acts by activating FXR in the liver and small intestine, which not only reduces the transformation of cholesterol to bile acids and the uptake of bile acids in the ileum, but also significantly improves the biochemical and histological characteristics of patients with NASH, including alleviating hepatic steatosis, inflammation, and fibrosis20. However, pruritus and abnormal low-density lipoprotein cholesterol (LDL-C), are the common adverse effects of high doses OCA21,22.
Several structural biology and chemical techniques, including as virtual screening, molecular docking, and toxicity prediction, were utilized during this study to discover and analyze compounds that might have agonist effects on FXR. Our discoveries established a strong basis for furthering the study of FXR agonists by supplying candidate compounds and their pharmacological properties from the ZINC15 database.
Materials and methods
Software for docking and ligand library
Discovery Studio 19 software (BIOVIA, San Diego, California, USA) facilitates drug design and development through validated 3D modeling and simulation tools. Its suite for simulating various molecule systems, is utilized for screening, designing, and refining potential drugs based on protein structure and biochemical characteristics. This approach has enabled the identification and enhancement of numerous candidate drugs and lead compounds, advancing drug discovery efforts efficiently. The LibDock module was utilized for virtual screening, which is an algorithm developed by Diller and Merz for docking small molecules into active receptor sites. It uses protein site features known as HotSpots, which are categorized into polar and apolar types. The receptor HotSpot map is calculated prior to the docking procedure, although predefined or user-adjusted HotSpot files can also be used. The CDOCKER module for docking studies and the ADME (absorption, distribution, metabolism and excretion) module for pharmacologic property analysis. The Find Similar Molecules by Fingerprints protocol finds ligands in an input library that are similar to the reference ligands by calculating the Tanimoto similarity coefficient. To screen for FXR agonists, we utilized the ZINC15 database's natural products repository, provided by the Irwin and Shoichet Laboratories at UCSF (University of California, San Francisco, CA, USA).
Structure-Guided Virtual Screening with LibDock
To search for potential FXR agonists, we selected the binding pocket of FXR and OCA as the docking site, adjusting the sphere diameter according to the pocket size. Utilizing the LibDock module for virtual screening, a grid and polar and apolar probes were placed to calculates protein hotspots. Subsequently, ligands were further aligned to create favorable interactions. The CHARMM force field (Cambridge, MA, USA) and Smart Minimizer algorithm were employed for ligand minimization. Ligands were ranked based on their scores using a ligand scoring function. OCA (ZINC000014164617) agonists were gained from the Zinc15 database, and the 3.20-Å crystal structure of human FXR (PDB ID: 3GD2) was acquired from the Protein Data Bank (PDB). Figure 1 illustrates the molecular structure of FXR. Protein preparation involved steps such as removing surrounding heteroatoms and crystal waters, adding hydrogens, protonation, ionization, and energy minimization utilizing the Smart Minimizer method and CHARMM force field. Subsequently, all ligands were virtually screened at specific protein sites. Each docking site and compound were ranked based on LibDock scores.
Figure 1.
Molecular structure of FXR. (A) The molecular structure of FXR is depicted, with the addition of a surface representation to enhance visualization. (B) The complex structure of FXR with OCA is illustrated, along with a surface representation for better understanding. Blue color indicates positive charge, while red represents negative charge.
ADME and toxicity prediction
The ADME properties of several identified compounds were assessed using the ADME module in Discovery Studio 19, encompassing factors such as aqueous solubility, human intestinal absorption, hepatotoxicity, cytochrome P450 2D6 (CYP2D6), plasma protein binding (PPB), and blood–brain barrier (BBB) permeability. In addition, we employed the TOPKAT (Toxicity Prediction by Computer Assisted Technology) module to assess various toxicological properties, including AMES (Ames mutagenicity), DTP (developmental toxicity potential), rat oral median lethal dose (LD50), chronic oral lowest observed adverse effect level (LOAEL), skin irritation and skin sensitization. Subsequent to the selection of FXR drug candidates, these pharmacological characteristics were thoroughly considered in the decision-making process.
Tanimoto similarity test
Tanimoto fingerprints, also known as Tanimoto similarity, are a method used to calculate structural similarity between molecules. They represent molecules as bit strings or binary vectors based on their structural information, where each bit represents specific structural features or substructures within the molecule. Tanimoto fingerprints assess the similarity between molecules by comparing these bit strings. The method calculates the ratio of the common bits (AND bits) between two molecules to the total bits (OR bits). The function used to calculate the similarity between two fingerprints. The functions are defined as functions of their contributions.
Molecule docking and pharmacophore prediction
Molecular docking experiments were conducted using the CDOCKER module in the CHARMM force field. The ligand maintained its flexibility while the receptor remained stiff during the docking process. The CHARMM force field was utilized to compute the energy of each complex pose and the corresponding interaction energy representing ligand binding affinity. The crystal structure of the OCA-FXR complex was obtained from the PDB. Considering the potential influence of fixed water molecules on the conformation of the complex, water molecules were removed during the docking process. To validate the ligand-receptor interaction, the agonist OCA was extracted from the original complex and redocked into the FXR structure. The binding site sphere of FXR was defined as the region within a 5-Å radius of the geometric center of the OCA ligand. During docking, the ligand was allowed to interact with protein residues within the binding site sphere. Different poses of each ligand-FXR complex were built and analyzed using CDOCKER interaction energies.
Molecular dynamics simulation
To ascertain the optimal binding conformations for subsequent molecular dynamics simulations, molecular docking analysis was performed on the ligand-FXR complexes. The system was simulated in an orthorhombic box with solvated water and sodium chloride to mimic the physiological environment. Following this, energy minimization utilizing the CHARMM force field was conducted to optimize the system, resulting in a final root mean square gradient of 0.227. Equilibration simulations were conducted for 5 ps, gradually transitioning the system from an initial temperature of 296 K to a target temperature of 302 K for 2 ps. Subsequently, a 500 ps molecular dynamics simulation with a time step of 2 fs was executed in the production module. Throughout the simulation, a standard pressure and temperature system maintained a consistent temperature of nearly 300 K. Long-range electrostatics were computed using the particle mesh Ewald technique, and hydrogen-containing bonds were fixed employing the linear constraint solver algorithm.Trajectory analysis in DS 19 allowed for the determination of structural characteristics, potential energy, and RMSD trajectory, utilizing the original complex settings as a reference point.
Results
Virtual screening of natural products database against FXR
The ligand-binding domain (LBD) of FXR serves as a pivotal site for binding with its ligands and interacting with coregulator proteins23. These coregulatory proteins modulate interactions with the basal transcriptional machinery, bringing about either the repression or activation of transcription. Following activation, FXR forms a heterodimeric complex with the retinoid X receptor (RXR), leading to the induction of the small heterodimer partner (SHP) gene. The FXR ligand-binding pocket area was chosen as the binding location to screen possible agonistic drugs. Furthermore, FXR exerts effects on the sodium taurocholate cotransporting polypeptide (NTCP) through an SHP-dependent mechanism, consequently repressing the hepatic uptake of bile acids24.This specific pocket region was identified as a reference site for further investigation. The ZINC15 database had 17,931 biogenic-for-sale-named product compounds in total. The pharmacologic characteristics of various substances were tested using the chemical structure of FXR, the receptor protein. OCA, an agonist, was used as the reference material. Following screening, 7158 compounds had the potential to bind steadily with FXR; of those, 450 compounds (LibDock score: 119.523) had higher scores than OCA. Table 1 lists the compounds that rank in the top 20.
Table 1.
Top 20 ranked compounds with higher LibDock.
| Number | Compounds | LibDock Score |
|---|---|---|
| 1 | ZINC000034944434 | 159.46 |
| 2 | ZINC000008220036 | 158.37 |
| 3 | ZINC000017044426 | 158.16 |
| 4 | ZINC000230075702 | 156.95 |
| 5 | ZINC000017044430 | 156.42 |
| 6 | ZINC000004097705 | 155.96 |
| 7 | ZINC000006129939 | 153.37 |
| 8 | ZINC000003947504 | 152.62 |
| 9 | ZINC000100822245 | 152.28 |
| 10 | ZINC000014951190 | 150.66 |
| 11 | ZINC000004016719 | 150.12 |
| 12 | ZINC000002033589 | 150.12 |
| 13 | ZINC000013328774 | 149.98 |
| 14 | ZINC000013374322 | 149.58 |
| 15 | ZINC000002526389 | 149.54 |
| 16 | ZINC000008662732 | 148.32 |
| 17 | ZINC000011666970 | 148.14 |
| 18 | ZINC000002526388 | 148.09 |
| 19 | ZINC000002528486 | 148.03 |
| 20 | ZINC000006036327 | 147.96 |
| 21 | OCA | 119.52 |
ADME and toxicity prediction
The ADME module of Discovery Studio 19 was utilized to predict the pharmacologic properties of all 20 selected ligands with OCA. These properties included aqueous solubility level, human intestinal absorption level, hepatotoxicity, CYP2D6 binding, PPB properties, and BBB level (Table 2). ZINC000008220036 and ZINC000006129939 were insoluble in water according to the aqueous solubility prediction, which was specified in water at 25 °C. In contrast, 18 other compounds were soluble in water. With the exception of 13 molecules, other ligands had comparable to human intestinal absorption levels than OCA (ZINC000014951190, ZINC000013374322, ZINC000002526389, ZINC000011666970, ZINC000002526388, ZINC000002528486, ZINC000006036327). There are 13 compounds including ZINC000013374322 and ZINC000006036327, as well as OCA, had no hepatotoxicity. Only 8 compounds (ZINC000006129939, ZINC000003947504, ZINC000014951190, ZINC000013328774, ZINC000013374322, ZINC000008662732, ZINC000011666970, ZINC000006036327) were anticipated to be CYP2D6 inhibitors. According to plasma protein binding properties, 7 compounds had good absorption, while other 13 compounds including ZINC000013374322 had weak absorption. 16 compounds demonstrated reduced blood–brain barrier penetration in comparison to OCA.
Table 2.
Adsorption, distribution, metabolism, and excretion properties of compounds.
| Number | Compounds | Solubility level | Absorption level | Hepatotoxicity | CYP2D6 | PPB level | BBB level |
|---|---|---|---|---|---|---|---|
| 1 | ZINC000034944434 | 2 | 2 | 0 | 1 | 0 | 4 |
| 2 | ZINC000008220036 | 0 | 3 | 0 | 1 | 1 | 4 |
| 3 | ZINC000017044426 | 2 | 3 | 0 | 1 | 0 | 4 |
| 4 | ZINC000230075702 | 2 | 2 | 0 | 1 | 0 | 4 |
| 5 | ZINC000017044430 | 2 | 3 | 0 | 1 | 0 | 4 |
| 6 | ZINC000004097705 | 2 | 2 | 0 | 1 | 0 | 4 |
| 7 | ZINC000006129939 | 0 | 3 | 0 | 0 | 1 | 4 |
| 8 | ZINC000003947504 | 4 | 3 | 0 | 0 | 0 | 4 |
| 9 | ZINC000100822245 | 2 | 3 | 0 | 1 | 0 | 4 |
| 10 | ZINC000014951190 | 2 | 0 | 1 | 0 | 0 | 2 |
| 11 | ZINC000004016719 | 2 | 3 | 0 | 1 | 0 | 4 |
| 12 | ZINC000002033589 | 2 | 3 | 0 | 1 | 0 | 4 |
| 13 | ZINC000013328774 | 3 | 3 | 1 | 0 | 0 | 4 |
| 14 | ZINC000013374322 | 2 | 0 | 0 | 0 | 0 | 2 |
| 15 | ZINC000002526389 | 2 | 0 | 1 | 1 | 1 | 4 |
| 16 | ZINC000008662732 | 4 | 3 | 1 | 0 | 0 | 4 |
| 17 | ZINC000011666970 | 3 | 1 | 1 | 0 | 0 | 4 |
| 18 | ZINC000002526388 | 2 | 0 | 1 | 1 | 1 | 4 |
| 19 | ZINC000002528486 | 2 | 0 | 1 | 1 | 1 | 2 |
| 20 | ZINC000006036327 | 2 | 0 | 0 | 0 | 1 | 1 |
| 21 | OCA | 2 | 0 | 0 | 0 | 1 | 2 |
Aqueous-solubility level: 0 (extremely low); 1 (very low, but possible); 2 (low); 3 (good); 4 (optimal).
Human-intestinal absorption level: 0 (good); 1 (moderate); 2 (poor); 3 (very poor).
Hepatotoxicity: 0 (Nontoxic); 1 (Toxic).
Cytochrome P450 2D6 level: 0 (Non-inhibitor); 1 (Inhibitor).
Plasma Protein Binding: 0 (Absorbent weak); 1 (Absorbent strong).
Blood Brain Barrier level: 0 (Very high penetrant); 1 (High); 2 (Medium); 3 (Low); 4 (Undefined).
We utilized various toxicity prediction methods in the TOPKAT module of Discovery Studio 19 to assess the safety of the top 20 ranked compounds. (Table 3). These indicators included skin irritation and skin sensitization, Ames mutagenicity, developmental toxicity potential properties, and rodent carcinogenicity (based on the U.S. National Toxicology Program dataset). According to the results, 5 compounds did not exhibit developmental toxicity potential (DTP), while 15 compounds were determined to be nonmutagenic. The reference OCA would have high DTP and low rodent carcinogenicity. Regarding the difference between skin irritation and skin sensitization, 13 compounds had no appreciable skin irritation, whereas 6 compounds showed no skin irritation in comparison to OCA.
Table 3.
Toxicities of compounds.
| Number | Compounds | Mouse NTP | Rat NTP | AMES | DTP | Skin Irritation | Skin Sensitization | ||
|---|---|---|---|---|---|---|---|---|---|
| Female | Male | Female | Male | ||||||
| 1 | ZINC000034944434 | 0 | 0 | 1 | 0.02 | 0 | 1 | 0.003 | 1 |
| 2 | ZINC000008220036 | 0 | 1 | 1 | 0 | 0.064 | 0 | 1 | 1 |
| 3 | ZINC000017044426 | 0 | 1 | 1 | 0.051 | 0.238 | 1 | 0.077 | 1 |
| 4 | ZINC000230075702 | 0 | 0 | 1 | 0.02 | 0 | 1 | 0.003 | 1 |
| 5 | ZINC000017044430 | 0 | 1 | 1 | 0.051 | 0.238 | 1 | 0.077 | 1 |
| 6 | ZINC000004097705 | 0 | 0 | 1 | 0.02 | 0 | 1 | 0.003 | 1 |
| 7 | ZINC000006129939 | 1 | 0 | 0 | 0.195 | 0 | 0 | 1 | 0.795 |
| 8 | ZINC000003947504 | 0.963 | 1 | 0.087 | 1 | 0.81 | 1 | 0.005 | 0.762 |
| 9 | ZINC000100822245 | 0 | 1 | 1 | 0.051 | 0.238 | 1 | 0.077 | 1 |
| 10 | ZINC000014951190 | 0 | 1 | 0.996 | 0.148 | 0 | 0.024 | 0.996 | 1 |
| 11 | ZINC000004016719 | 0 | 1 | 1 | 0.05 | 0.265 | 1 | 0.023 | 1 |
| 12 | ZINC000002033589 | 0 | 1 | 1 | 0.05 | 0.265 | 1 | 0.023 | 1 |
| 13 | ZINC000013328774 | 0.865 | 1 | 1 | 1 | 0.674 | 1 | 0 | 0.01 |
| 14 | ZINC000013374322 | 0.002 | 0 | 1 | 0.015 | 0 | 0.095 | 0.804 | 1 |
| 15 | ZINC000002526389 | 0.999 | 0.036 | 0 | 0.999 | 0.999 | 0.769 | 0.956 | 0.01 |
| 16 | ZINC000008662732 | 0.98 | 1 | 0.326 | 1 | 0.036 | 1 | 0.001 | 0.621 |
| 17 | ZINC000011666970 | 1 | 1 | 1 | 1 | 0 | 1 | 0 | 0 |
| 18 | ZINC000002526388 | 0.999 | 0.041 | 0 | 0.999 | 0.999 | 0.745 | 0.953 | 0.013 |
| 19 | ZINC000002528486 | 0.603 | 0.001 | 0 | 0.535 | 0.996 | 0.019 | 0.932 | 0.074 |
| 20 | ZINC000006036327 | 1 | 0.005 | 0 | 0.034 | 0 | 0.352 | 0 | 0.002 |
| 21 | OCA | 0 | 0 | 0 | 0 | 0.001 | 1 | 0.325 | 0.963 |
The best leading compounds were determined to be ZINC000013374322 and ZINC000006036327 when all the data in Table 3 were included. They were also less carcinogenic to rodents, non-CYP2D6 inhibitors, Ames mutagenic, and developmental toxicity. In addition, ZINC000006036327 were less likely to cause skin irritation, and skin sensitization than other compounds. OCA and compound ZINC000013374322 and ZINC000006036327, which have various multiple reactive oxygens in their chemical structures, are fairly similar in their chemical constructions, as illustrated in Fig. 2. As a result, ZINC000013374322 and ZINC000006036327 were chosen for further investigation after being verified as safe medication candidates.
Figure 2.
The structures and new compounds of OCA were screened by virtual screening. (A) ZINC000013374322; (B) ZINC000006036327; (C) OCA.
Tanimoto similarity test for molecular similarity
We used OCA as a reference compound and assessed the similarity of ZINC000013374322 and ZINC000006036327 using the Find Similar Molecules by Fingerprints module. As depicted in Table 4, their Tanimoto coefficients were calculated as follows: ZINC000006036327 exhibited a similarity score of 94.40%, while ZINC000013374322 showed a slightly lower score of 69.65%.
Table 4.
Compounds scored by Tanimoto similarity coefficient.
| SA | SB | SC | Similarity | |
|---|---|---|---|---|
| OCA | Reference | Reference | Reference | Reference |
| ZINC000006036327 | 928 | 45 | 10 | 94.40% |
| ZINC000013374322 | 677 | 182 | 113 | 69.65% |
| Tanimoto: SA/(SA + SB + SC) | ||||
SA: The number of AND bits. Bits present in both the target and the reference, SB: The number of bits in the target but not the reference, SC: The number of bits in the reference but not the target.
Analysis of ligand binding and ligand pharmacophore
To determine the binding modes of ZINC000013374322 and ZINC000006036327 with FXR, molecular docking was conducted using the CDOCKER module with FXR molecular structure. The resulting CDOCKER potential energy was calculated and presented in Table 4. The CDOCKER module exhibits remarkable dependability in accurately replicating experimental results, as evidenced by the 0.75 Å RMSD between the docking poses of the OCA and FXR complex and the crystal structure. Table 5 displays the CDOCKER potentials of ZINC000013374322 and ZINC000006036327, which were significantly lower than the reference ligand OCA's value of 56.0158 kcal/mol. This suggests that ZINC000013374322 and ZINC000006036327 exhibit higher binding affinity to FXR than OCA.
Table 5.
CDOCKER potential energy of compounds with FXR.
| Compounds | CDOCKER interaction energy (kcal/mol) |
|---|---|
| ZINC000013374322 | − 58.36 |
| ZINC000006036327 | − 57.48 |
| OCA | − 56.02 |
The hydrophobic (including Pi-Sigma, Pi-Pi Stacked, Pi-Alkyl, and Alkyl) and hydrogen bonds of ligand-FXR complexes were all subjected to structural investigation (Figs. 3, 4 and Table 6,7). ZINC000013374322 and FXR had two sets of hydrogen bonds through the compound's O3 and A:HIS294:ND1 of FXR, as well as through the compound's O16 and A:SER332:CB of FXR. Two hydrogen bonds were established with FXR in relation to ZINC000006036327 (ZINC000006036327:O9: A:PHE284, ZINC000006036327:O9: A:TRP454, respectively). Three hydrogen bonds were established with OCA and FXR (A:TYR361:OH: OCA:O4, A:TYR369:OH: OCA:O9, OCA:O4: A:PHE329, respectively).
Figure 3.
Schematic diagram of intermolecular interactions predicting binding patterns. (A) ZINC000013374322; (B) ZINC000006036327; (C) OCA.
Figure 4.
Schematic drawing of interactions between ligands and FXR. (A) ZINC000013374322; (B) ZINC000006036327; (C) OCA.
Table 6.
Hydrogen bond interaction parameters for each compound and FXR residues.
| Receptor | Compound | Donor atom | Receptor atom | Distances (Å) |
|---|---|---|---|---|
| FXR | ZINC000013374322 | A:HIS294:ND1 | ZINC000013374322:O3 | 3.09 |
| A:SER332:CB | ZINC000013374322:O16 | 3.18 | ||
| ZINC000006036327 | ZINC000006036327:O9 | A:PHE284 | 4.14 | |
| ZINC000006036327:O9 | A:TRP454 | 3.75 | ||
| OCA | A:TYR361:OH | OCA:O4 | 3.03 | |
| A:TYR369:OH | OCA:O9 | 3.05 | ||
| OCA:O4 | A:PHE329 | 3.54 |
Table 7.
Hydrophobic interaction parameters for each compound and FXR residues.
| Compound | Type | Donor atom | Receptor atom | Distances (Å) |
|---|---|---|---|---|
| ZINC000013374322 | Pi-Sigma | ZINC000013374322:C1 | A:HIS294 | 3.81 |
| Pi-Pi Stacked | ZINC000013374322 | ZINC000013374322 | 3.93 | |
| Pi-Alkyl | ZINC000013374322 | A:MET265 | 4.05 | |
| ZINC000013374322 | A:ILE335 | 4.25 | ||
| ZINC000013374322 | A:LEU287 | 5.26 | ||
| ZINC000013374322 | A:ALA291 | 4.05 | ||
| ZINC000013374322 | A:MET328 | 4.99 | ||
| ZINC000013374322 | A:MET365 | 5.17 | ||
| ZINC000006036327 | Alkyl | A:LEU287 | ZINC000006036327 | 5.25 |
| A:ALA291 | ZINC000006036327 | 5.01 | ||
| ZINC000006036327 | A:MET290 | 4.59 | ||
| ZINC000006036327:C25 | A:ARG331 | 4.16 | ||
| ZINC000006036327 | A:MET328 | 4.77 | ||
| ZINC000006036327:C30 | A:MET328 | 4.73 | ||
| Pi-Alkyl | A:PHE329 | ZINC000006036327:C1 | 4.22 | |
| A:TYR361 | ZINC000006036327:C1 | 5.36 | ||
| A:HIS447 | ZINC000006036327:C1 | 5.37 | ||
| OCA | Alkyl | A:ALA291 | OCA | 5.14 |
| A:ALA291 | OCA | 5.41 | ||
| A:ALA291 | OCA:C21 | 4.10 | ||
| A:MET328 | OCA | 4.64 | ||
| OCA | A:MET328 | 5.22 | ||
| OCA | A:MET290 | 5.07 | ||
| OCA:C20 | A:MET290 | 3.57 | ||
| OCA:C21 | A:LEU287 | 4.41 | ||
| OCA:C23 | A:MET265 | 4.93 | ||
| OCA:C23 | A:ARG331 | 4.28 | ||
| OCA:C27 | A:LEU287 | 4.87 | ||
| OCA:C27 | A:ILE352 | 4.72 | ||
| OCA:C27 | A:ILE357 | 4.45 | ||
| OCA:C27 | A:MET365 | 5.06 | ||
| Pi-Alkyl | A:HIS294 | OCA:C23 | 4.12 |
Nonetheless, the ZINC000013374322-FXR complex generated one pair of Pi-Sigma interactions, two pairs of Pi-Pi Stacked interactions and six pairs of Pi-Alkyl interactions. ZINC000006036327 generated six pairs of Alkyl contacts with FXR and three pair of Pi-Alkyl interactions. Additionally, OCA and FXR established fifteen pairs of interactions, including fourteen pairs of Alkyl interactions and one pair of Pi-Alkyl interactions.
Molecular dynamics simulation
We used the stability of the ligand-FXR complexes under ambient circumstances using the molecular dynamics simulation module test. Figure 5 shows the potential energy chart and RMSD curves for each complex using the CDOCKER module. The original conformations from the molecular docking experiment served as the basis for these. After 500 ps, the trajectories of all the complexes achieved equilibrium, and over time, their RMSD and potential energy tend to stabilize. The hydrogen bonds and other interactions that the chemicals and FXR generated affected the stability of these complexes. ZINC000013374322 and ZINC000006036327 could interact with FXR, their complexes with FXR may then progressively persist in a natural environment and operate as FXR regulating agents, similarly to OCA.
Figure 5.
Results of molecular dynamics simulation of the compounds ZINC000013374322 and ZINC000006036327. (A) Potential energy. RMSD, root-mean-square deviation. (B) Average backbone root-mean-square deviation.
Discussion
In this study, ZINC000013374322 and ZINC000006036327 were verified as the best candidate compounds due to their high intestine absorption level, without CYP2D6 inhibition and Ames mutagenicity. In addition, compared to other compounds, ZINC000006036327 were anticipated to have less developmental toxicity, skin irritation and skin sensitization. Of particular note is the higher binding affinity of our compounds compared to OCA, allowing for comparable pharmacological effects at lower doses in human body. Given that the side effects of OCA are dose-dependent, our compound may reduce the incidence of adverse effects. ZINC000006036327 with lower skin irritancy and sensitization potential compared to OCA, may reduce skin itching occurrence. Pruritus was the most prevalent adverse event described in REGENERATE study, which showed 51% occurrence in the group of 25 mg OCA and resulted in a drug discontinuation of 9%25. The underlying mechanism of pruritus is complex and mutifactorial. The cholestasis-associated pruritus was considered as non-histaminergic itch. Some receptors activation was previously reported related to pruritus occurrence26. As for the negative effect of OCA on LDL-C, low-dose atorvastatin can compensate this defect27.
In this work, ADME, TOPKAT, CDOCKER, and molecular dynamics simulation were used to virtually screen 17,931 biogenic-for-sale-named product compounds gained from the ZINC15 database. First of all, compounds with higher scores on LibDock shown better energy optimization and more stable conformations than those with lower scores. Following computation by the LibDock module, 7158 molecules were shown to have stable binding capabilities with FXR. In addition, 450 of these ligands had LibDock scores that were higher than OCA's (LibDock score: 119.523), suggesting a more stable complex with FXR and greater energy optimization than OCA. The top 20 natural chemicals were selected and combined for additional research based on their LibDock scores. Secondly, to evaluate the pharmacologic characteristics of the chosen drugs, ADME and toxicity predictions were performed. ZINC000013374322 and ZINC000006036327 had comparable to or better human intestinal absorption levels than OCA without interaction with CYP2D6 and hepatotoxicity. And their higher similarity Tanimoto coefficients may elucidate the common therapeutic targets shared between these compounds and OCA. Thirdly, the mechanism of ligand binding and chemical interactions between potential chemicals and FXR were uncovered by CDOCKER module calculation. The two compounds’ CDOCKER interaction energy was clearly lower than OCA. This implies that compared to OCA, both them might have a better binding affinity for FXR. Fourthly, the chemical structures of the two compounds were examined using molecular structure inspection. There were more chemical bonds existing in the two complexes than OCA, suggesting that these two chemicals may connect with FXR more stably. Therefore, they may contribute to the competitive activation of FXR activity, thereby increasing the efficiency of bile acid transport and enhancing the efficacy of NASH treatment. Fifthly, utilizing a molecular dynamics simulation, their stability in the natural environment were evaluated. The trajectories of these ligand-FXR complexes reached equilibrium after 500 ps, according to computational studies of the RMSD and potential energy. These two complexes' RMSD and potential energy show a tendency toward stability over time, suggesting that they might coexist peacefully in the natural world.
Comparative experiments with OCA confirmed the excellence of our compounds. However, due to limitations with our graphics card and GPU, our study was constrained to a 500 ps molecular dynamics simulation. Despite this, we conducted thorough analysis within this timeframe to ensure the accuracy and reliability of our results. Our compounds showed high LibDock scores binding to FXR, indicating stable binding and potential strong affinity. Furthermore, interactions with PPAR and LXR suggest further validation of multi-target effects is needed for future studies in NAFLD drug development.
Above all, these findings could guide future therapeutic development and designation efforts, such as the modification and refinement in increasing the stability of the ligand-receptor combination. In future studies, other compound qualities should be further evaluated. Additional investigations, such as animal testing, are necessary to more firmly corroborate our results.
Conclusions
In order to find the best leading compounds that target FXR, using computer-assisted structural and chemical analysis technologies, ZINC000013374322 and ZINC000006036327 were chosen as safe candidates of FXR targeting drugs. Additionally, this study offered a list of potential drugs along with their pharmacologic characteristics, which could help in the design and development of medications for FXR or other proteins.
Author contributions
Xindan Hu wrote the main manuscript text, Junliang Ge and Ying Wen reviewed and edited the main manuscript text. Three authors read and edited the final version of the manuscript.
Funding
The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.
Data availability
The datasets analysed during the current study are available in the PDB (protein data bank) repository with the Protein Data Bank identifier:3GD2 (https://www.rcsb.org/structure/3GD2) and the ZINC15 database: https://zinc15.docking.org/substances/subsets/for-sale/. All data generated during this study are included in this published article.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Junliang Ge, Email: gejl21@mails.jlu.edu.cn.
Ying Wen, Email: wenying666466@163.com.
References
- 1.Vernon, G., Baranova, A. & Younossi, Z. M. Systematic review: the epidemiology and natural history of non-alcoholic fatty liver disease and non-alcoholic steatohepatitis in adults. Aliment. Pharmacol. Ther.34(3), 274–285 (2011). 10.1111/j.1365-2036.2011.04724.x [DOI] [PubMed] [Google Scholar]
- 2.Rinella, M. E. Nonalcoholic fatty liver disease: A systematic review. Jama313(22), 2263–2273 (2015). 10.1001/jama.2015.5370 [DOI] [PubMed] [Google Scholar]
- 3.Chalasani, N. et al. The diagnosis and management of nonalcoholic fatty liver disease: Practice guidance from the American Association for the Study of Liver Diseases. Hepatology (Baltimore, Md)67(1), 328–357 (2018). 10.1002/hep.29367 [DOI] [PubMed] [Google Scholar]
- 4.Harrison, S. A. et al. A phase 3, randomized, controlled trial of resmetirom in NASH with liver fibrosis. N.Engl. J. Med.390(6), 497–509 (2024). 10.1056/NEJMoa2309000 [DOI] [PubMed] [Google Scholar]
- 5.Harrison, S. A. et al. Resmetirom for nonalcoholic fatty liver disease: A randomized, double-blind, placebo-controlled phase 3 trial. Nat. Med.29(11), 2919–2928 (2023). 10.1038/s41591-023-02603-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Sanyal, A. J. et al. Tropifexor for nonalcoholic steatohepatitis: An adaptive, randomized, placebo-controlled phase 2a/b trial. Nat. Med.29(2), 392–400 (2023). 10.1038/s41591-022-02200-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Ding, L. et al. Bile acid nuclear receptor FXR and digestive system diseases. Acta Pharmaceut. Sin. B5(2), 135–144 (2015). 10.1016/j.apsb.2015.01.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Xi, Y. & Li, H. Role of farnesoid X receptor in hepatic steatosis in nonalcoholic fatty liver disease. Biomed. Pharmacother. Biomed. Pharmacother.121, 109609 (2020). 10.1016/j.biopha.2019.109609 [DOI] [PubMed] [Google Scholar]
- 9.Fiorucci, S. et al. Bile acid modulators for the treatment of nonalcoholic steatohepatitis (NASH). Expert Opinion Investig. Drugs29(6), 623–632 (2020). 10.1080/13543784.2020.1763302 [DOI] [PubMed] [Google Scholar]
- 10.Kim, M. K. et al. Targeting orphan nuclear receptor SHP in the treatment of metabolic diseases. Expert Opinion Therap. Targets14(4), 453–466 (2010). 10.1517/14728221003652463 [DOI] [PubMed] [Google Scholar]
- 11.Lee, J. H. et al. Small heterodimer partner SHP mediates liver X receptor (LXR)-dependent suppression of inflammatory signaling by promoting LXR SUMOylation specifically in astrocytes. Sci. Signal.9(439), ra78 (2016). 10.1126/scisignal.aaf4850 [DOI] [PubMed] [Google Scholar]
- 12.Lee, J. H. et al. A novel role for the dioxin receptor in fatty acid metabolism and hepatic steatosis. Gastroenterology139(2), 653–663 (2010). 10.1053/j.gastro.2010.03.033 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Lee, Y. S. et al. Orphan nuclear receptor SHP interacts with and represses hepatocyte nuclear factor-6 (HNF-6) transactivation. Biochem. J.413(3), 559–569 (2008). 10.1042/BJ20071637 [DOI] [PubMed] [Google Scholar]
- 14.Thompson, M. D. et al. β-Catenin regulation of farnesoid X receptor signaling and bile acid metabolism during murine cholestasis. Hepatology (Baltimore, Md)67(3), 955–971 (2018). 10.1002/hep.29371 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Fang, Y. et al. Recent advances in the medicinal chemistry of farnesoid X receptor. J. Med. Chem.64(24), 17545–17571 (2021). 10.1021/acs.jmedchem.1c01017 [DOI] [PubMed] [Google Scholar]
- 16.Tacke, F. & Weiskirchen, R. Non-alcoholic fatty liver disease (NAFLD)/non-alcoholic steatohepatitis (NASH)-related liver fibrosis: mechanisms, treatment and prevention. Ann. Transl. Med.9(8), 729 (2021). 10.21037/atm-20-4354 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Talukdar, S. & Kharitonenkov, A. FGF19 and FGF21: NASH we trust. Mol. Metab.46, 101152 (2021). 10.1016/j.molmet.2020.101152 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Patel, K. et al. Cilofexor, a nonsteroidal fxr agonist, in patients with noncirrhotic NASH: A phase 2 randomized controlled trial. Hepatology (Baltimore, Md)72(1), 58–71 (2020). 10.1002/hep.31205 [DOI] [PubMed] [Google Scholar]
- 19.Ratziu, V. et al. Hepatic and renal improvements with FXR agonist vonafexor in individuals with suspected fibrotic NASH. J. Hepatol.78(3), 479–492 (2023). 10.1016/j.jhep.2022.10.023 [DOI] [PubMed] [Google Scholar]
- 20.Harrison, S. A. et al. A structurally optimized FXR agonist, MET409, reduced liver fat content over 12 weeks in patients with non-alcoholic steatohepatitis. J. Hepatol.75(1), 25–33 (2021). 10.1016/j.jhep.2021.01.047 [DOI] [PubMed] [Google Scholar]
- 21.Modica, S., Gadaleta, R. M. & Moschetta, A. Deciphering the nuclear bile acid receptor FXR paradigm. Nucl. Recept. Signal.8, e005 (2010). 10.1621/nrs.08005 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Younossi, Z. M. et al. Current and future therapeutic regimens for nonalcoholic fatty liver disease and nonalcoholic steatohepatitis. Hepatol. (Baltimore, Md)68(1), 361–371 (2018). 10.1002/hep.29724 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Wang, N. et al. Ligand binding and heterodimerization with retinoid X receptor α (RXRα) induce farnesoid X receptor (FXR) conformational changes affecting coactivator binding. J. Biol. Chem.293(47), 18180–18191 (2018). 10.1074/jbc.RA118.004652 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Hoeke, M. O. et al. Low retinol levels differentially modulate bile salt-induced expression of human and mouse hepatic bile salt transporters. Hepatology (Baltimore, Md)49(1), 151–159 (2009). 10.1002/hep.22661 [DOI] [PubMed] [Google Scholar]
- 25.Younossi, Z. M. et al. Obeticholic acid for the treatment of non-alcoholic steatohepatitis: interim analysis from a multicentre, randomised, placebo-controlled phase 3 trial. Lancet (London, England)394(10215), 2184–2196 (2019). 10.1016/S0140-6736(19)33041-7 [DOI] [PubMed] [Google Scholar]
- 26.Beuers, U., Wolters, F. & Oude Elferink, R. P. J. Mechanisms of pruritus in cholestasis: Understanding and treating the itch. Nat. Rev. Gastroenterol. Hepatol.20(1), 26–36 (2023). 10.1038/s41575-022-00687-7 [DOI] [PubMed] [Google Scholar]
- 27.Pockros, P. J. et al. CONTROL: A randomized phase 2 study of obeticholic acid and atorvastatin on lipoproteins in nonalcoholic steatohepatitis patients. Liver Int. Off. J. Int. Assoc. Study Liver39(11), 2082–2093 (2019). [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The datasets analysed during the current study are available in the PDB (protein data bank) repository with the Protein Data Bank identifier:3GD2 (https://www.rcsb.org/structure/3GD2) and the ZINC15 database: https://zinc15.docking.org/substances/subsets/for-sale/. All data generated during this study are included in this published article.