Xenobiotics and toxins injure the liver and can result in substantial hepatic pathology. The liver metabolizes and excretes drugs and xenobiotics. All drugs have some side effects, and many affect hepatocytes including cisplatin, tegafur, cyclophosphamide (anti-cancer drugs), nefazodone (an antidepressant), and some diabetes medications [1]. Hepatocytes are also responsible for the excretion of drugs from animals [1]. Drug overdose, toxins, chemotherapeutic agents like acetamide, and a hepatotoxic agent, carbon tetrachloride (CCl₄) can damage hepatocytes and lead to liver inflammation (hepatitis) and cirrhosis [2]. Plants and their extracts have been usedto treat human diseases since ancient time. Among these, some plants have been reported to have additional medicinal value and beneficial characteristics including anti-inflammatory, immune-modulatory and anti-viral actions on hepatoprotective properties [3-5]. The secondary metabolites of herbal treatments have become more prominent for the treatment of liver disease, with evidence-based outcomes being established through promising clinical trials and validation.

Silybummarianum, commonly known as milk thistle, produces seeds that have medicinal value when ripe [6,7]. Silymarin (SLN), an important secondary metabolite of Silybummarianum, consists of a complex mixture of four isomers (flavonolignans): silybin, isosilybin, silydianin, and silychristin. SLN has anti-inflammatory, anti-lipid, immuno-modulatory, anti-oxidative, and hepatocyte-regenerating actions. However, it is not thought to be anti-viral [1,2].Glycyrrhizaglabra is a leguminous plant belonging to the Leguminosae family [8]. The root extract of Glycyrrhizaglabra contains various chemical compounds of medicinal value including saponin, triterpines, flavonoids, and other chemicals like sugars, coumarins, amino acids, choline, and tannins [4,9,10].Moreover, in Japan, glycyrrhetic acid GLN is used for the management of chronic hepatitis C [11]. Glycyrrhizin metabolism is important because its metabolites inhibit the production of aldosterone and suppress 5-β reductase commonly called hepatic pseudoaldosterone syndrome. There is an inhibition of phospholipase A2 activity which is important in various inflammatory processes. Glycyrrhizin has the ability to interfere with cyclooxygenase and prostaglandin production involved in the progression of inflammatory mechanisms in biological system [18].

In the current study, molecular docking strategy was performed to find out their respective binding energies along with the number of hydrogen bonds and other hydrophobic interactions. The study was further validated by the use of in silico ADMET prediction of both compounds (silibinin and glycyrrhetic acid) in order to check their pharmacokinetics and pharmacodynamics properties. These two compounds have already being tested as potent therapeutic compounds in various experimental trials. The compounds are being viewed as potent therapeutic molecules in the management of inflammatory disorders.

Target proteins were docked with silibinin and glycyrrhetic acid using AutoDock 4.2 and binding energies were calculated.

Structure retrieval

The three dimensional crystal structures of the three target proteins were retrieved from the Protein Data Bank. Cyclooxygenase-2 (6COX) [12], 5-beta-reductase (3BV7) [13], and phospholipase A2 (2B03) [14] were taken as targets. The ligand molecules, glycyrrhetic acid (GLN) and silibinin (SLB) were retrieved from PubChem (https://pubchem.ncbi.nlm.nih.gov) and were refined using ACD ChemSketch (http://www.acdlabs.com/resources/freeware), a tool that offers functionalities such as structure refining, optimization, etc.

Active site prediction

The active sites of all the three target proteins were identified using Q-SiteFinder [15]. Q-SiteFinder works by binding hydrophobic probes to the protein. It then finds the clusters of probes with most favorable binding region based on energy values. It ranks these clusters according to the sum of total binding energies for individual clusters in the order of likelihood of being a binding site.

Various pharmacodynamic as well as pharmacokinetic parameters were considered including human intestinal absorption (HIA), Caco-2 permeability, aqueous solubility, blood-brain barrier penetration, renal organic cation transportation, cytochrome P (CYP) inhibitory promiscuity, cytochrome P₄₅₀ inhibition, AMES toxicity, fish toxicity, rat acute toxicity, Tetrahymenapyriformis toxicity, human ether-a-go-go-related genes inhibition, mutagenic, tumorigenic and reproductive risks ().

Docking studies

Target proteins were docked with silibinin and glycyrrhetic acid using AutoDock 4.2. The free energy of binding between the ingredients of ligands and proteins were calculated. AutoDock 4.2 uses charge-based desolvation force fields and well defined improved models of the unbound state. Docking analysis attempts to bind the ligand into the obtained binding sites of the target protein and produces the best docked conformations with minimal energy, as the output. Semi-flexible docking protocol was applied, wherein the target proteins were kept rigid while the phytochemical ligands were kept flexible for being docked upon. A 5A° grid was built surrounding the binding pocket. Grid maps dimensions were set as 60 × 60 × 60 points with spacing of 0.375A° to yield the receptor model that included atoms within 0.5A° of the grid center. All the other parameters were kept at default and Lamarckian Genetic Algorithm (LGA) was chosen to predict the best conformers. The protein - ligand complexes were viewed by LigPlot viewer (http://www.ebi.ac.uk/thornton-srv/software/LIGPLOT) and Chimera (http://www.cgl.ucsf.edu/chimera).

Drug-likeness analysis

Chosen compounds were subjected to further selection on the basis of Lipinski’s rule of five (Ro5) [16]. Lipinski’s screening was performed using Molinspiration server (http://www.molins piration.com/cgi-bin/properties) and physicoche-mical properties of ligands were calculated. Drug scores were calculated using ORISIS property explorer (http://www.organic-chemistry.org/ prog/peo).

Absorption, distribution, metabolism, elimination and toxicity (ADMET) studies

The molecular structure of both ligands were submitted to ADMET-SAR server (http://lmmd. ecust.edu.cn:8000) to examine their drug likeliness and different pharmacokinetic and pharmacodynamic parameters including blood-brain barrier penetration, human intestinal absorption, Caco-2 permeability, cytochrome P450 solubility, cytochrome P (CYP) inhibitory promiscuity, renal organic cation transportation, human ether-a-go-go related genes inhibition, rat acute toxicity, fish toxicity, Tetrahymena pyriformis toxicity and Ames toxicity.

Both ligands (silibinin and glycyrrhetic acid) have shown strong binding energy with all three target proteins (-7.5 to -10.9 Kcal/mol).The said compounds followed the Lipinski’s rule in of five without any violation with respect to molecular weight (≤ 500KDa), an octanol-water partition coefficient (log P ≤ 5), molecular refractivity (40 - 130), number of H-bond acceptors (≤ 10) and number of H-bond donors (≤ 5). Lipinski’s rule of five analyzes various physiochemical properties [16]. These includes Log P anoctanol water partition coefficient which should be greater than or equal to 5, number of H-bonds donors and acceptor ≤5 and ≤10 respectively, molecular weight of greater than 500 and a molecular refractivity in the range of 40-130. The Lipinski’s screening is an essential filter that determines if a compound is suitable for drug designing. Upon docking, both silibinin and glycyrrheticacid () formed at least one hydrogen bonds with all three target proteins. Docking with cyclooxygenase-2 (6COX) hydrogen bonds formation were observed with Phe580 residue having bond length of 2.84Å (silibinin) and 3.09Å (glycyrrhetic acid) as shown in . Additionally, Asp347, Gln350, Gly354, Gln192, His351, Ser579, Phe580 were found to be common interacting residues in the target proteins for both ligands.

Significant numbers of hydrophobic interactions were observed between the two ligands and all three target proteins. When docked with 5-beta-reductase (3BV7) Pro221, Lys273, Ile271, Gly24, Tyr26, Trp230, Val309, Ser225, Ser220 residues interacted with silibinin and glycyrrhetic acid in addition to a few other residues. Similarly, Phospholipase A2 (2B03) docking analysis revealed Leu41, Tyr111, Leu19, Phe106, Ile9, Met20, Asn23, Leu31, Gly30, Leu2 and Tyr69 as common amino acid residues responsible for scoring a high binding energy of -8.8 kcal/mol and -10.9 kcal/mol for silibinin and glycyrrhetic acid respectively (). All docking results are summarized in .

Plants secondary metabolites have medicinal values and show therapeutic potential like anti-viral, anti-inflammatory and immune-modulatory effects. The interest of investigators has been moved towards herbs because they are used as functional foods and an important source for the preparation of drugs. Hence, herbal medicines have been used worldwide for the treatment of various diseases. Inflammation is protective response involving molecular mediators, immune cells and blood vessels. Phytochemicals like silymarin and glycyrrhizin in suitable combinations are used as drugs for patients with liver disorders as they possess hepatoprotective property against inflammation [11,17]. Docking methodology facilitates structure-based virtual database screening with the ability to achieve a cost effective and quick estimation of the affinity and binding mode of a ligand for drug target. In this study, glycyrrhetic acid and silibinin have been shown to possess significant anti-inflammatory activity from ADMET prediction. Moreover, the difference between binding energies can be observed in . Both the ligands occupied the same location in the protein targets and shared common amino acid residues for the inhibition of proteins’ action in inflammation.

Poor pharmacokinetics and toxicity in the biological system lead to failure in drug development. With the help of ADMET profile during the process of drug discovery one can remove incompatible compounds as well as exhibit significant role in reducing cost and efforts [19]. For a drug to be approved for use, tedious toxicological analysis are performed to ensure its safety upon ingestion. In silico ADMET analysis is a quick approach to find if a compound has acceptable pharmacokinetics and pharmacodynamics property. The toxicity risks and bioavailability of silibinin and glycyrrhetic acid were predicted based on their ADMET profile [20]. Further analysis exhibit CYP inhibitory promiscuity as silibinin inhibit two cytochromes including CYP450-2C9 and CYP450-3A4.

Depending upon the acute oral toxicity (ADMET prediction profile), compounds are categorized into four groups. Category I contains compounds with LD₅₀ values less than or equal to 50 mg/kg. Category II contains compounds with LD₅₀ values greater than 50 mg/kg but less than 500 mg/kg. Category III includes compounds with LD₅₀ values greater than 500 mg/kg but less than 5000 mg/kg. Category IV consisted of compounds with LD₅₀ values greater than 5000 mg/kg.

The role of cytochrome P₄₅₀ enzymes is very important as they are responsible for the metabolism of drug in biological system and its clearance from the body. Inhibition of any isoform of CYP leads to the malfunctioning of drug metabolism and elevation of toxicity [21].

After ADMET analysis of compounds under study it was observed that glycyrrhetic acid possesses better ADMET profile when compared to silibinin, because it can cross blood-brain barrier but not the silibinin, it is non-inhibitor for all CYP-inhibitors while silibinin is inhibitor for two CYPs like CYP-2C9 and CYP-3A4. Moreover, glycyrrhetic acid presented low CYP inhibitory promiscuity as compared to silibinin (). Both compounds were non carcinogenic and showed no AMES toxicity. Both compounds were found to be fit for drug development.

Glycyrrhetic acid possesses better ADMET profile than silibinin. Furthermore, hydrophobic interactions were significantly high between two ligands (glycyrrhetic acid and silibinin) and three target proteins (COX-2, phospholipase A2 and 5β-reductase). Ligand-receptor complex showed that strong hydrogen bonding and van der Waals interactions were formed indicating that these two phytochemicals are suitable options as anti-inflammatory agents as well as can be considered safe for development into a commercial drug when compared to other drugs in terms of their relatively fewer side-effects.