ADME-Tox Prediction and Molecular Docking Studies of Two Lead Flavonoids From the Roots of Tephrosia Egregia Sandw and the Gastroprotective Effects of Its Root Extract in Mice

Background: This study aimed to predict the pharmacokinetic and toxicological properties of lead flavonoids from the roots of T. egregia [praecansone A (1) and pongachalcone (2) ], and to assess the gastroprotective effects and possible underlying mechanisms of the root extract in mice. Methods: Quantitative and qualitative data for in silico absorption, distribution, metabolism, and excretion (ADME) analyses of the two flavonoids were acquired from the SwissADME database. Toxicity assessment was performed with the ProTox-II server. To evaluate the putative interactions of both flavonoids with opioid receptors and NO protein, we acquired structures of the targets ( μ , κ , and δ -opioid receptors, and iNOS) in Homo sapiens from https://www.rcsb.org/. For docking studies, AutoDock 4.2 was used for ligand and target arrangement, and AutoDock Vina was used for calculations. For in vivo assays, mice were pretreated ( per os ) with T. egregia (2, 20, or 200 mg/kg). After 60 min, 99.9% ethanol (0.2 mL) was injected ( per os ). At 30 min after ethanol injection, the mice were euthanized, and the gastric damage, gastric levels of hemoglobin, glutathione content, and activity of superoxide dismutase and catalase were evaluated. To elucidate T. egre- gia mechanisms, we used misoprostol, a prostaglandin analog; indomethacin, an inhibitor of prostaglandin synthesis; L-arginine, an NO precursor; L-NAME, an antagonist of NO synthase; naloxone, an opioid antag- onist; and morphine, an opioid agonist. Results: In silico results showed that flavonoids (1) and (2) had favorable ADME properties and toxicity profiles, and exhibited satisfactory binding energies data (below −6.0 kcal/mol) when docked into their targets ( μ , κ , and δ -opioid receptors, and iNOS). T. egregia decreased the ethanol-induced gastric damage and hemoglobin levels, and increased the glutathione content, and activity of superoxide dismutase and catalase. Naloxone and L-NAME, but not indomethacin, prevented T. egregia ’s effects, thus suggesting that opioid receptors and NO are involved in T. egregia ’s efficacy. Conclusions: Flavonoids (1) and (2) exhibited favorable pharmacokinetic properties, showing high lethal dose, 50% (LD 50 ; 3,800 and 2,500 mg/kg, respectively) values. Neither flavonoid was found to be hepato- toxic, carcinogenic, or cytotoxic to human cells. In vivo assays indicated that T. egregia ameliorated oxidative stress levels, and its mechanism is at least partially based on opioid receptors and NO. T. egregia may therefore be considered as a new gastroprotective strategy.


Introduction
Gastric ulcer is a common disease globally, which poses an economic burden [1,2]. Protective factors [secretion, bicarbonate production, blood flow, production of nitric oxide (NO) and prostaglandins, and enzymatic and non-enzymatic antioxidant systems] and risk factors (hydrochloric acid, pepsin, Helicobacter pylori infection, nonsteroidal anti-inflammatory drugs, psychological stress, free radicals, and excessive intake of alcohol) modulate the pathophysiology of gastric ulcers [3][4][5].
Drugs currently used to treat gastric ulcers include antibiotics (in cases of H. pylori infection), histamine receptor antagonists (ranitidine), proton-pump inhibitors (omeprazole), and cytoprotective agents (bismuth and sucralfate). However, several adverse effects and limitations of these treatments have been reported. Therefore, the management of gastric ulcers is an issue in clinical medicine, and new approaches to treat this disease are needed to overcome the adverse effects associated with conventional treatment [6,7]. In this regard, plant-based natural products used in traditional medicine have been applied in the treatment of several illnesses.
Therefore, in this study, we performed in silico assays to predict the pharmacokinetic and toxicological properties of the lead flavonoids from the roots of T. egregia [praecansone A (1) and pongachalcone (2)]. In addition, we evaluated the possible healing effects of T. egregia on gastric ulcers induced by ethanol in mice, and we assessed the possible mechanism of action, investigating the involvement of prostaglandins, NO, and opioid receptors.

In silico studies: ADME-toxicity prediction and molecular docking
During drug-development research, substantial efforts have attempted to limit animal use in determining the acute oral toxicity of new compounds, in agreement with the "3Rs" principle [21]. Therefore, to predict the pharmacokinetic and toxicological properties of T. egregia as a drug candidate, we performed in silico absorption, distribution, metabolism, excretion, and toxicity (ADME-T) studies of the two best-known and most abundant flavonoids in the roots of T. egregia: praecansone A (1) [10], and pongachalcone (2) [12]. Figure 1 shows the chemical structures of these two compounds. In addition, to evaluate the pharmacodynamic phase, we assessed the putative interactions of both flavonoids with opioid receptors (μ, κ, and δ-opioid receptors) and inducible NO synthase (iNOS) protein structures, through molecular docking assays evaluating the compound-protein interactions.
The structures of flavonoids (1) and (2) were modeled three-dimensionally in Avogadro 1.1.2 software [22] and geometrically optimized through density functional theory with the correlation functional B3LYP and base 6-31G(d) in GAMESS software [23].
The prediction of the pharmacokinetic profiles of flavonoids (1) and (2) was evaluated according to Lipinski's "rule of five" and ADME parameters. According to Lipinski's rule, in general, to be active orally, a drug must meet at least four of the following five principles: molecular mass less than 500 g/mol, high lipophilicity (octanol-water partition coefficient less than 5), fewer than five H-bond donors, fewer than ten H-bond acceptors, and molar refractivity between 40 and 130 cm 3 ·mol −1 [24]. Another related criterion, polar surface area <140 Å 2 , was added later [25]. Quantitative and qualitative values were obtained from the SwissADME online server (http://www. swissadme.ch/).
To evaluate the putative interactions of flavonoids (1) and (2) with opioid receptors and NO protein, we downloaded structures (targets) of iNOS (PDB: 3NQS) and the opioid receptors mu (μ; PDB: 4DK1), kappa (k; PDB: 4DJH), and delta (δ; PDB: 6PT3) of Homo sapiens from the Protein Data Bank (https://www.rcsb.org/). All structures were defined by the X-ray diffraction technique with a resolution between 1.20 and 3.10 Å. The positions of the coupling sockets were based on the native ligand of the macromolecules and the Web Server Computed Atlas of Surface Topography of proteins (http://sts.bioe.uic.edu/ castp/calculation.html). For molecular docking simulations,

Plant material and extract preparation from T. egregia roots
The extract of T. egregia roots used in the experiments was obtained from the Department of Organic and Inorganic Chemistry at the Federal University of Ceará. The specimen was collected in the flowering stage on July 22, 2014, at Icaraí beach (Caucaia, Ceará, Brazil; longitude −38.6563 and latitude −3.73454). Botanical identification was performed by the Department of Biology at the Federal University of Ceará. A voucher specimen has been deposited at the Herbário Prisco Bezerra of the Federal University of Ceará, under registration number 55945.
The roots of T. egregia were extracted by continual maceration with ethanol (3 × 3.0 L) at room temperature. The resulting solution was filtered and concentrated under reduced pressure with a rotary evaporator, thus resulting in 1.13 g of root extract. Fresh dilution of the macerated extract in 0.9% NaCl solution was prepared on the day of the experiment and injected orally at different doses, as described below.

Animals
Swiss mice (female; 25-30 g) from the vivarium of the Federal University of Ceará were housed under temperature (22 ± 2°C) and light control (12/12 h light/dark cycle), and were given water and food ad libitum. According to ample-size calculation w based on α = 0.05 and a study power of 0.8, the experimental protocol was performed with seven mice per group. All experimental protocols were performed to minimize animal suffering and the number of animals used, in accordance with the Guide for the Care and Use of Laboratory Animals from the Brazilian Society of Science in Laboratory Animals. The experimental protocol was approved by the Institutional Animal Care and Use Committee from the Federal University of Ceará, Campus of Sobral (process No. 01/2016).
Gastric lesions were measured and compared with the total area of each stomach in Image J 1.42q®. The method was performed as previously described [31]. The evaluation of gastric hemoglobin (Hb) levels, glutathione (GSH) content, and superoxide dismutase (SOD) and catalase (CAT) activity was performed in unfixed stomach samples, as described below.

Assessment of the antioxidant activity of T. egregia Gastric Hb determination
Gastric Hb levels were measured with a standard kit containing Drabkin's reagent, according to the manufacturer's instructions. Briefly, slices of fresh gastric tissue (50-100 mg) were homogenized in Drabkin's reagent solution (100 mg/mL) and centrifuged twice at 10,000 rpm (10 min). The supernatant was then collected, and the Hb concentration was determined on the basis of absorbance at 540 nm, as compared with a standard Hb dilution. Results are expressed in μg Hb/100 mg tissue.
The SOD activity was measured as previously described [33]. The reactant was 50 mM phosphate buffer, 0.1 mM EDTA, and 19.5 mM L-methionine, pH 7.8. A homogenate solution of 10% of gastric fresh tissue in phosphate-buffered saline was prepared. This homogenate was centrifuged twice, first for 10 min at 3,600 rpm and 4° C, and then for 20 min at 12,000 rpm and 4°C. The supernatant was removed between each centrifugation. In the dark, 30 μL of supernatant was mixed with 1 mL of reactant (described above). Then 150 μL of 750 μM NBT and 300 μL of 1 μM riboflavin were added. After 15 min of exposure to fluorescent light (15 W), the tubes with the mixture were read in a spectrophotometer at 560 nm. The results are expressed as μSOD per μg tissue protein.
The CAT activity was evaluated according to methods described previously [34]. Briefly, stomach homogenate was diluted to 10% in reaction medium (diluted H 2 O 2 , Tris-HCl buffer, 5 mM EDTA, pH 8, in milliQ H 2 O). The enzymatic activity was measured at 230 nm with a spectrophotometer (during 6 min). Results are expressed in μM per min per μg protein.

BIOI 2022
Original Article Probing T. egregia's mechanism of action

Analysis of NO involvement in T. egregiamediated gastric protection
Mice were pretreated (per os) with vehicle (0.3 mL/30 g), T. egregia (200 mg/kg), or the NOS substrate L-arginine [600 mg/kg; intraperitoneal (i.p.)]. Thirty minutes later, 99.9% ethanol (0.2 mL/animal) was administered by gavage. In another set of experiments, L-NAME (20 mg/kg), a nonspecific inhibitor of NOS, was injected (i.p.) 15 min before vehicle, T. egregia, or L-arginine injection. Thirty minutes after ethanol challenge, mice were euthanized, and gastric tissue samples were collected and analyzed as described previously.

Study of prostaglandin involvement in T. egregia-mediated gastric protection
Mice were pretreated (per os) with T. egregia (200 mg/kg), the prostaglandin analogue misoprostol (50 mg/kg), or saline (0.3 mL/30 g) 1 h before ethanol injection. Indomethacin (10 mg/kg), a prostaglandin synthesis inhibitor, was injected (per os) 2 hours before injection of T. egregia, misoprostol, or saline. Thirty minutes after ethanol administration, mice were euthanized, and gastric tissue samples were collected and analyzed as described previously.

Examination of opioid receptor involvement in T. egregia-mediated gastric protection
First, to examine the involvement of opioid receptors in the gastroprotective effect of T. egregia, we injected mice (i.p.) with naloxone (4 mg/kg), a non-selective opioid receptor antagonist. After 15 min, the mice were treated with morphine (5 mg/kg, subcutaneous), T. egregia (200 mg/kg, per os), or saline (0.3 mL/30 g; per os), followed by ethanol (0.2 mL). Thirty minutes after the ethanol challenge, the mice were euthanized, and gastric tissue samples were collected and analyzed as described previously.

Statistical analysis
Results are expressed as mean ± standard error of the mean. ANOVA followed by Bonferroni test was used to compare means. P < 0.05 was considered significant. The analyses were performed in GraphPad Prism 6, San Diego, CA, USA.

Results
In silico predicted ADME properties of two flavonoids [praecansone A (1) and pongachalcone (2)] isolated from T. egregia roots ADME assessment is a major method for characterization of pharmacokinetic profiles. Physicochemical characteristics and the ADME parameters of flavonoids (1) and (2) are shown in Tables 1 and 2.
An established method to assess drug-likeness is evaluation according to Lipinski's rule. Flavonoids (1) and (2) showed favorable results when evaluated with Lipinski's rule ( Table 1). The molecular weights of both compounds are below 500 g/mol, and the lipophilicity (log P) is less than 5 for both compounds. Flavonoids (1) and (2) have fewer than    Table 2, indicated high gastrointestinal absorption of both compounds. In addition, our data suggested that both flavonoids (1) and (2) can cross the blood brain barrier. Moreover, both compounds appear to interact with the substrate P-glycoprotein (P-gp), an efflux transporter that mediates the active transport of drugs from the intracellular to the extracellular compartment. Additionally, in silico prediction of CYP-ligand interactions was performed to evaluate the inhibitory potency of flavonoids (1) and (2) toward five human CYP isoforms (CYP1A2, 2C9, 2C19, 2D6, and 3A4). Both compounds inhibited the enzymatic activity of CYP isoforms, thus suggesting an approach for rapid prediction of CYP inhibitory activity through in silico assays. These isoforms were selected because they are responsible for almost all metabolic reactions (90%), including the metabolism of clinically used drugs and carcinogens [35]. Therefore, flavonoids (1) and (2) have favorable ADME properties.

Prediction of the toxicity of the two flavonoids (praecansone A and pongachalcone) isolated from T. egregia roots
An essential stage during drug development is the evaluation of the toxicity of new candidate drugs. The toxicity of flavonoids (1) and (2) isolated from T. egregia roots was predicted with in silico tools. The LD 50 , hepatotoxicity, carcinogenicity, immunotoxicity, mutagenicity, and cytotoxicity were predicted ( Table 3). On the basis of the consensus of the results, the compounds were determined to be positive or negative for particular toxicity endpoints. With the ProTox-II webserver, flavonoids (1) and (2) were predicted to be in toxicity class 5 (2,000 < LD 50 ≤ 5,000, indicating low toxicity) for acute oral toxicity [26]. Flavonoid (1) had a higher LD 50 (3,800 mg/kg) than flavonoid (2) (2,500 mg/kg), with a prediction accuracy of 70.1% and 73.5%, respectively ( Table 3). Toxicity prediction by ProTox-II indicated that flavonoids (1) and (2) are not hepatotoxic. The two flavonoids are not carcinogenic or cytotoxic to human cells ( Table 3).
Immunotoxicity is an adverse effect of xenobiotics on the immune system [26]. We used a model of immunotoxicity based on immune cell cytotoxicity data from the U.S. National Cancer Institute's public database. In this model, growth inhibition values <10 μM toward the B-cell line RPMI-8226 are considered to indicate toxicity of compounds [36].
In the overall immunotoxicity parameter, a structure-based probability was detected by the system of 20% chance for flavonoid (1) and 12% chance for flavonoid (2). The ProTox-II immunotoxicity prediction model has an accuracy of 74.00% in cross-validation and 70.00% in external validation [26]. Overall, both compounds displayed favorable toxicity profiles.

Effects of T. egregia in an ethanol-induced gastropathy model
The treatment with T. egregia (200 mg/kg) decreased (P < 0.05) gastric damage (9.13 ± 1.6) below that observed in groups receiving only vehicle (30.46 ± 2.74; Figure 3A). Interestingly, the findings for the groups receiving T. egregia (200 mg/kg) and ranitidine (9.83 ± 1.01) were similar (P > 0.05). Macroscopic images of the gastroprotective effects of T. egregia (200 mg/kg) against ethanol-induced gastric damage are shown in Figure 3E-c. Figure 3F illustrates the effect of T. egregia (200 mg/kg) on the histopathological (H&E; magnification ×200) features of gastric mucosa from mice after ethanol challenge. The T. egregia (200 mg/ kg) group showed a preserved epithelial lining, whereas the ethanol 99.9% group displayed loss of integrity of the epithelial lining. Table 4 shows that T. egregia (200 mg/kg) decreased (P < 0.05) the gastric Hb to levels below those in the groups receiving only the vehicle.

Evaluation of GSH content, and gastric SOD and CAT activity
The antioxidant effects of T. egregia on gastric damage induced by ethanol challenge are depicted in Table 5. T. egregia pre-treatment (200 mg/kg), compared with vehicle alone, resulted in significantly higher (P < 0.05) GSH content, and SOD and CAT activity.

Discussion
This study revealed the gastroprotective effects of the extract of T. egregia roots in mice subjected to ethanol challenge. Because a main phase in drug development involves evaluation of the pharmacokinetics, toxicology, and pharmacodynamics of a drug candidate, we performed in silico ADME and toxicity prediction of two lead flavonoids from the roots of T. egregia. In addition, we found evidence that that the gastroprotective and antioxidant activity of T. egregia is mediated at least partly by opioid receptors and NO. The Tephrosia genus (family Fabaceae) is mainly found in regions with high temperatures [37]. Phytochemical investigations revealed several phytoconstituents in the Tephrosia genus with biological activities including anti-diabetic, anti-ulcer, anti-diarrheal, wound healing, anti-inflammatory, insecticidal, anti-viral, anti-protozoal, anti-fungal, and anti-plasmodial effects [19]. However, very few data are available regarding the pharmacokinetics, toxicology, and mechanism of action of T. egregia and its phytoconstituents.
Drug discovery is a complex process requiring analyses of several parameters (safety, pharmacokinetics, and efficacy) of new drug candidates. During drug research, substantial efforts have been made to decrease animal use in determining the acute oral toxicity of new compounds, in agreement with the "3Rs" principle [21]. In this regard, computational software has contributed to predicting the features and parameters (pharmacokinetics, toxicology, and pharmacodynamics of drug candidates) needed for the development of new drug candidates. In the present study, in silico results showed that the lead flavonoids from the roots of T. egregia [praecansone A (1) and pongachalcone (2)] had favorable ADME properties and toxicity profiles, and exhibited satisfactory binding energies (below −6.0 kcal/mol) data when docked into their targets (μ-, κ-, and δ-opioid receptors, and iNOS).  Through in vivo assays, we evaluated gastric Hb levels, a sign of mucosal bleeding. In fact, gastric damage has been associated with elevated levels of Hb [38]. Herein, an increase in Hb levels was observed after ethanol challenge, but was diminished after treatment with T. egregia (200 mg/kg).
Oxidative stress plays a key role in the pathophysiology of gastric disorders [39]. Similarly to other findings [40], our results indicated the involvement of oxidative stress in gastric pathology, thereby indicating the potential role of antioxidants as gastroprotective agents for gastric ulceration [41]. Ethanol reacts with cell membranes and consequently increases lipid peroxidation and the production of hydroxyl radicals and superoxide anions, which may induce direct oxidative damage [42]. In the present study, ethanol challenge decreased the activity of the antioxidant enzymes SOD and CAT, but these effects were prevented by T. egregia (200 mg/kg) treatment. Notably, we observed no differences (P > 0.05) in the SOD and CAT values between animals treated with T. egregia and naive animals. These results suggested that T. egregia has therapeutic potential in ameliorating ethanol-induced gastric injury by alleviating oxidative stress. Intracellular anti-oxidative systems protect the gastric mucosa against oxidative stress [43].
Prostaglandins have important roles in the gastrointestinal tract, modulating mucosal defense by inhibiting acid secretion, increasing the amount of mucus, and promoting vasodilation [44]. Because prostaglandin synthesis may be blocked by nonsteroidal anti-inflammatory drugs, the administration of these agents is associated with gastric damage [45]. Our current data showed that T. egregia gastroprotection occurred independently of pretreatment with indomethacin, a non-steroidal anti-inflammatory agent, thus demonstrating that the gastroprotective effects do not occur through prostaglandins. Substantial evidence indicates that NO acts as a dual signaling molecule in both physiological and pathophysiological processes. NO's roles in gastric mucosa integrity, protection, and ulcer healing have been described [46]. In fact, NO is a vasodilatory molecule that may control gastric blood flow and gastric mucosal barrier integrity-important elements of gastric ulcer healing [47]. In the present study, L-NAME, a non-selective competitive inhibitor of NOS, decreased the gastroprotective effect of T. egregia, thus suggesting the role of NO in T. egregia gastroprotective effect.
To evaluate the putative involvement of opioid receptors in the gastroprotective effect of T. egregia, we used naloxone, a non-selective opioid antagonist. Naloxone significantly prevented the effect of morphine, the standard opioid agonist, and consequently reversed the gastroprotective effects. Similarly, in the presence of naloxone, no gastroprotective effects of T. egregia were observed, thus suggesting that T. egregia's effect in ameliorating ethanol-induced gastric damage in mice appears to depend on the activation of opioid receptors. The activation of supraspinal δand μ-opioid receptors has been associated with greater production of NO in a model of ethanol-induced gastric damage [48], thus suggesting that activation of opioid receptors and mucosal NO may be involved in the gastroprotective effect of T. egregia.

Conclusion
Our data suggest that T. egregia has gastroprotective potential in mice, which may be partly due to its antioxidant efficiency, and the involvement of opioid receptors and NO. Considering the numerous mechanisms through which the gastric mucosa may resist injury, we believe that T. egregia may serve as an innovative technology to develop new strategies for improving ulcer healing.