2.4. In Silico Analysis

For the analysis, compounds 3a, 3h, and 3g were chosen due to their half-maximal inhibitory concentration (IC50) measured after 72 h incubation with cancerous cell lines.

ADME analysis of compounds 3a, 3g, and 3h was performed via SwissADME [24], a freely available software provided by the Swiss Institute of Bioinformatics. The software provides information about estimated predictors, such as basic physiochemical properties, lipophilicity, water solubility, pharmacokinetics, druglikeness, and medicinal chemistry (Table 4). Physiochemical properties are relevant due to crossing biological barriers. Lipophilicity is very important for pharmacokinetics drug discovery. Orally active drugs should consist of no more than 5 hydrogen bond donors less than 10 hydrogen bond acceptors, molecular weight of less than 500 daltons, and an n-octanol and water partition coefficient lesser than 5 [25]. Water solubility is connected to the oral admission of the drug. Water solubility is the major property of medicine absorption. Being a substrate of the permeability glycoprotein is a predictor of pharmacokinetics. Knowledge about the interaction of the compound and cytochromes P450 gives information about effective drug elimination through metabolic biotransformation. The last predictor-medicinal chemistry-supports structural drug discovery. It informs about problematic structural fragments inside the investigated compound.

Estimation of probable macromolecular targets was performed using SwissTargetPrediction [26], choosing Homo sapiens as a target species. SwissTargetPrediction is a freely available web tool provided by the Swiss Bioinformatics Institute.

The protein target structures listed above were downloaded from the Protein Data Bank (PDB). The inhibitors from the complexes were removed, as well as water molecules. The pH value for the protonation of both the ligand (compounds 3a, 3g, and 3h) and the protein was set to 7.4 using PDB2PQR [27]. The input files were prepared using AutoDock tools [28]. The search space of each protein was defined from “center on hetero”. A box was placed on the geometric center of an existing ligand. The molecular docking was performed using AutoDock Vina [29]. The docking was performed for each target from the Table 5. presented in Section 3.4.