An In Silico Investigation to Explore Anti-Cancer Potential of Foeniculum vulgare Mill. Phytoconstituents for the Management of Human Breast Cancer

Breast cancer is one of the most prevalent cancers in the world. Traditionally, medicinal plants have been used to cure various types of diseases and disorders. Based on a literature survey, the current study was undertaken to explore the anticancer potential of Foeniculum vulgare Mill. phytoconstituents against breast cancer target protein (PDB ID: 6CHZ) by the molecular docking technique. Molecular docking was done using Autodock/vina software. Toxicity was predicted by the Protox II server and drug likeness was predicted by Molinspiration. 100 ns MD simulation of the best protein-ligand complexes were done using the Amber 18 tool. The present molecular docking investigation has revealed that among the 40 selected phytoconstituents of F. vulgare, α-pinene and D-limonene showed best binding energy (−6 and −5.9 kcal/mol respectively) with the breast cancer target. α-Pinene and D-limonene followed all the parameters of toxicity, and 100 ns MD simulations of α-pinene and D-limonene complexes with 6CHZ were found to be stable. α-Pinene and D-limonene can be used as new therapeutic agents to cure breast cancer.

Palmitic acid (Hexadecanoic acid) Colon (HT29), colorectal (HCT-116) [10,11] 24. 38. γ-asarone Gastric (AGS) cells [13] 39. γ-terpinene Human prostate (LNCaP, PC-3), glioblastoma (SF-763, SF-767) cells [19] Assays have demonstrated the anti-cancer potential of some important Foeniculum vulgare constituents against a wide array of cancers including human breast cancer. Using In silico approaches, we can simultaneously screen out millions of phytocompounds and drugs against any diseases. Furthermore, molecular dynamics-based screening helps to hypothesize the efficacy, stability and toxicity of the important drug candidates or pytocompounds and to design experiments for their in vivo testing. Therefore, the present study was undertaken to propose potential cytotoxic phytoconstituents of Foeniculum vulgare using in silico approaches for developing novel therapeutics for the management of human breast cancer. An important drug target candidate in the case of human breast cancer is an Estrogen Receptor α Y537S protein (PDB ID 6CHZ) which was selected on the basis of previous literature.

Molecular Docking of Major Phytoconstituents of Foeniculum vulgare with Breast Cancer Target
Forty phytoconstituents were selected for the molecular docking studies with the Estrogen Receptor α Y537S breast cancer target protein (PDB ID 6CHZ).

Toxicity Prediction of Foeniculum vulgare Phytoconstituents
The toxicity of the selected forty phytoconstituents was analyzed by using the Protox II online tool. Toxicity data showed that among all the selected phytoconstituents, anisaldehyde, gamma-asarone, trans-anethole, estragole, p-tolualdehyde, and quercetin-3-O-beta-D-glucuronide are carcinogenic in nature, alpha-D-glucose and octadecenoic acid are hepatotoxic in nature, alpha-D-glucose, and quercetin-3-O-beta-D-glucuronide are immunotoxic in nature, and gamma-asarone is mutagenic in nature (Table 4). Molecular docking and toxicity data revealed that alpha-pinene and D-limonene are the best drug molecules for the management of human breast cancer. Alpha-pinene belongs to class 5 drugs with an LD 50 of 3700 mg/kg body weight and D-limonene is a class 4 drug with an LD 50 of 500 mg/kg body weight. The detailed in silico toxicity parameters are depicted in Table 4.

Molecular Dynamic simulation of Complexes
Based on molecular docking, and toxicity investigations, alpha-pinene and D-limonene were found to be the choicest drug candidates for managing human breast cancer. Therefore, alpha-pinene and D-limonene ligand complexes with breast cancer target protein 6CHZ were selected for MD simulation studies to further verify stability of these complexes. The RMSDs of D-limonene complex with 6CHZ C-alpha was stable from the start of simulation up to 40 ns (3 Å). It showed little fluctuation between 40-50 ns (which is in the acceptable range), and afterwards it remained stable up to 100 ns at 2.9 Å ( Figure 4A). Whereas, in the case of 6CHZ complex with alpha-pinene, RMSDs of C-alpha showed little fluctuation between 0-30 ns (3 Å), which remained stable between 30 to 100 ns thereafter at 2.7 Å ( Figure 4B). The RMSF plots of D-limonene and alpha-pinene fitted over 6CHZ protein also showed lesser residual fluctuations in the protein's secondary structures, such as the alpha helices and beta strands ( Figure 5A).
Radius of gyration of α-pinene and D-limonene in complex with 6CHZ is in the range of 18-19 Å and as shown in Figure 6A. The radius of gyration plots establishes the compactness of the α-pinene and D-limonene in complex with 6CHZ protein complex and confirms the stability of complexes. Solvent accessible range of alpha pinene and D-limonene complexed with 6CHZ protein is between 14,000-15,000 Å; as shown in Figure 6B.

Discussion
Medicinal herbs always confer beneficial effects on human health when consumed in moderate quantities. Foeniculum vulgare, especially being rich in several bioactive constituents, has been used as a food condiment and ingredient throughout the world. It has been traditionally used since ancient times to cure several human diseases including arthritis, cancers, conjunctivitis, endocrine, gastric, hepatic, insomnia, kidney, reproductive, and respiratory ailments. These studies signify that F. vulgare holds a promising future, and harnessing its hidden anti-cancer potential could be an important milestone in the field of novel drug development. However, the development of potent cytotoxic agents requires investigation of the molecular mechanisms of disease prevention to substantiate the beneficial attributes as well as to authenticate the immense pharmacological importance of Foeniculum vulgare constituents.
In the current study we found that among all forty phytocompounds, αpinene and D-limonene showed the best binding affinity with the breast cancer target. They were also found to be non-toxic in nature. In contrast to our study, Ghasemian et al. [35] reported the anticancer activity of F. vulgare against the MCF-7 breast cancer cell line by MTT assay. They also reported that the essential oil F. vulgare increases the expression of Bax and decreases the Bcl2 gene expression. Similarly, Mohamad et al. [36] reported the anticancer activity of F. vulgare methanolic extract and essential oil against the MCF-7 breast cancer cell line and Hepg-2 liver cell line by in vivo and in vitro assays. Berrington and Lall [37] also reported the anticancer activity of F. vulgare acetone extract against the Vero African green monkey kidney cell line and the cervical cancer cell line HeLa. Batool et al. [38] also reported the anticancer activity of F. vulgare methanolic and ethanolic extract against the MCF-7 breast cancer cell line by MTT assay. Both extracts are more effective against the breast cancer cell line (methanol and ethanol-40 µg/mL). Zaahkouk et al. [39] also reported the anticancer activity of methanolic extract of Foeniculum vulgare seed against breast, colon and liver cancer. Similar to our study, Hossain [40] studied the anticancer activity of Withania somnifera phytocompounds with 6CHZ protein by molecular docking. He also reported the drug likeness and ADMET activity of phytocom pounds. There are several reports on molecular docking, MD simulation, and ADMET screening of phytoconstituents [41][42][43][44][45][46].
Phytocompounds of Rheum emodi were screened for antibacterial and antiviral properties by molecular docking, and results were validated by MD simulation [41]. Similarly, Salaria et al. [42,43] studied the antimicrobial activity of T. serpyllum compounds and the antioxidant and anti-inflammatory activities of important phytoconstituents were investigated using molecular docking and MD simulation, and the results were further validated by in vitro experiments. In addition, phytocompounds containing nanoparticles were shown to provide significant anti-cancer effects against breast cancer cell lines (MCF-7, PC-12, MDA-MB-231) [47][48][49][50][51]. Letrozole incorporated folate-conjugated polymer-lipid hybrid nanoparticles were also shown to exhibit anticancer activity against the MCF-7 and PC-12 cell lines [52]. Folic acid functionalized apoferritin is a drug delivery vehicle for equirubicin against breast cancer cells (MCF-7) [53].  [54], and Discovery Studio were used in the present investigation.

Protein Preparation
Estrogen Receptor α Y537S breast cancer target (PDB ID 6CHZ) was used for molecular docking with major phytocompounds from Foeniculum vulgare Mill. The three-dimensional 6CHZ was downloaded from the protein databank (http://www.rscb.org/pdb, accessed on 20 December 2021). It consists of a dimeric structure, and its chain A was extracted for docking using Pymol software. The active site was predicted manually by grid box analysis (grid dimensions x = 72, y = 74, z = 76 Å, center at x, y, z = −28.959, −2.617, −25.683 Å).

Molecular Docking of Major Phytoconstituents of Foeniculum vulgare with Breast Cancer Target
The selected ligands were docked to the catalytic triad of proteins using AutoDock vina, which was then saved as a pdbqt file. The population of potential ligand conformations/orientations at the binding site was estimated via docking. A vina script was used to align the ligands in the same spatial coordinates [54]. After the docking search was completed, the best conformation with the lowest docked energy was chosen. Discovery Studio (https://discover.3ds.com/d, accessed on 20 December 2021) was used to study interactions between proteins and ligands in the pdb complex preparations. A negative score (kcal/mol) was used to calculate the ligand's binding strength. The equilibrium constant was calculated by using formula [55]: Ki = e −∆G/RT ∆G = Gibbs free energy; R = 1.9872 cal/mol·K; T = 298.15 • K

Drug Likeness Prediction of Foeniculum vulgare Phytoconstituents
Drug likeness prediction of F. vulgare phytoconstituents was done using the Molinspiration online server (http://www.molinspiration.com, accessed on 20 December 2021). Drug likeness is based on the Lipinski's rule of five. According to the rule of five, the number of hydrogen acceptors should be <10, the number of hydrogen donors should be <5, the molecular weight should be <500 Da, and the partition coefficient should be >5 (estimated in terms of log P) in acceptable drug molecules. In the case of variables, only one violation is acceptable [56].

Toxicity Prediction of F. vulgare Phytoconstituents by Protox II Server
The pharmacokinetics and toxicity of pharmacologically important phytoconstituents were predicted by using Protox II servers. Toxicity was estimated in terms of LD50 values ranging from ≤50 mg/kg (in the case of Class I compounds), between 50 to 500 mg/kg (in the case of Class II compounds), between 500 to 5000 mg/kg (in the case of Class III compounds), and >5000 mg/kg (in the case of Class IV compounds). Classes I, II, and III have less toxicity, whereas Class IV displays no toxicity [57]. Moreover, PROTOX is a rodent oral toxicity server that is used to determine LD50 values and toxicity classes of potentially cytotoxic agents [58]. Based on the molecular docking drug-likeness data and toxicity data, phytoconstituents were selected for further MD simulation analysis.

MD Simulation of Best Protein-Ligands Complexes
MD simulations of best protein-ligand complexes were done by using the Amber18 tool. MD simulations were performed to gain a better understanding of the binding interactions of the 6CHZ protein with the selected phytoconstituents, namely α-pinene and D-Limonene [51]. The ligands underwent an amber generalized force field (GAFF), while the protein was subjected to amber ff14SB [52,53]. Using Gaussian 09 software, the atomic charges of the ligands were computed using the restrained electrostatic potential (RESP) procedure at the HF/6-31G* level of theory 31, 32 [55]. Using an H++ server, proton transfer states of the ionizable residues in protein structures were investigated using the pKa method at a neutral pH. The tLeap application was used to create each system. Each system was solvated in a cubic water box with the TIP3P model after being treated with sodium ions. Each system was exposed to at least four minimizations in order to eliminate the worst conflicts. Initially, all of the sodium ions and water molecules were reduced using a steepest descent technique of 2000 steps, followed by a conjugate gradient algorithm of 3000 steps. The same approach was used to relax all of the hydrogen and water molecules in a row. Finally, the entire system was energy-minimized for 5000 steps of steepest descent and 5000 steps of conjugate gradients. The system was heated from 0 to 300 K while performing 200 ps of MD and then 300 ps of density equilibration at a fixed volume with position restrictions on the protein atoms. All protein-ligand complexes were stabilized with for 10 ns of MD without any structural restrictions at a constant pressure before the manufacturing process. By linking to a Langevin thermostat with a collision frequency of 2 cm −1 , the temperature was kept at 300 K. For the unpaired electron interactions, a cut off of 10 was chosen, and the Particle Mesh Ewald (PME) [53] approach and the SHAKE algorithm was used to limit the bond lengths involving hydrogen atoms. Finally, at a temperature of 300 K, MD simulations (productions) were run for 100 ns. The computed trajectories were utilized to examine activities of all the complexes in order to determine the stability of the system. Important parameters like root mean square deviation (RMSD), root mean square fluctuation (RMSF), radius of gyration (RG), and solvent accessible surface area (SASA) were used to examine deviations of the protein and protein-ligand complexes [51,52]. Furthermore, using molecular mechanics and the Poisson-Boltzmann Surface Area (MM-PBSA) method, the total free energy of binding, the free energy of solvation (polar vs. non-polar solvation energies), and the potential energy (electrostatic and van der Waal's interactions) of each protein-ligand complex were calculated. For the MM-PBSA computation, the last 10 ns of the MD trajectory were used [53].

Conclusions
Forty major phytoconstituents of Foeniculum vulgare were screened for breast cancer by molecular docking with the 6CHZ target protein. Among all of the selected phytoconstituents, D-limonene and α-pinene have the best binding affinity and follow all the parameters of toxicity. An MD simulation study validated the stability of complexes. αpinene has a lot of potential for the treatment of breast cancer, and this hypothesis can be further validated by in vitro and in vivo experiments.

Conflicts of Interest:
The authors declare that they have no conflicts of interest.
Sample Availability: Yes: D-limonene is available with Rajan Rolta and Deeksha Salaria.