After manual visual inspection of 117 compounds, only 18 compounds were found to be unpatented and compliant with molecular weights under 350 g/mol and LogP less than 3 when one significant figure is considered. PubChem database [10] was used for compound retrieval and for confirmation that the 18 compounds have not been patented. These 18 compounds were selected for ADMET studies and molecular docking. The International Union of Pure and Applied Chemistry (IUPAC) names and the chemical structures of the 18 compounds are shown in Table 1 and Figure 2, respectively.

Table 1. IUPAC names of compounds 1 - 18.

Figure 2. Structures of compounds 1 - 18.

All compounds were shown to have %ABS greater than or equal to 82.53% and high gastrointestinal absorption, making them suitable for oral ingestion and blood circulation. The bioavailability of the compounds was found to be moderate with a value of 0.55, which will impact dosage. The predicted ADMET profile of all 18 compounds is shown in Table 2. The molar refractivity (MR) of the compounds ranges from 73.76 to 96.72, (cm³∙mol⁻¹) with compound 18 having the highest value, suggesting that they have higher probability of being non-toxic in vivo, which is desirable for drug-likeness [6]. Consensus Log P values produced using SwissADME for each compound was less than or equal to 3. The predicted toxicity data in Table 3 showed compounds 3 and 4 to be active (≤0.55). The remaining 16 compounds were predicted to be inactive in hepatotoxicity, cytotoxicity, Androgen receptor and ATPASE AAA domain containing protein (Table 3). However, as evident in etoposide, anticancer agents do not necessarily require cytotoxicity [11]. Androgen receptors have been a long-time target for anticancer agents, but studies have shown that the development of resistance to androgen receptor-targeted therapies, such as enzalutamide and abiraterone, in advanced prostate cancer require new treatment approaches [12]. In addition, ATPase domain containing protein, which is responsible for various cellular processes that are dysregulated in cancer, such as autophagy and the proteasome pathway [13], the selected compounds showed moderate toxicity. Due to the unexpectedly low toxicity of the selected compounds, molecular docking data was needed to determine each compound’s binding affinity to topoisomerase IIα.

Table 2. ADME profile of compounds 1 - 18.

Table 3. Toxicity data of selected compounds and the template molecule.

The binding affinity for topo IIα of the template compound ARN-21934 −9.5 kcal/mol versus −10.3 kcal/mol for 18. Of the 18 docked compounds, compound 18 had the highest binding affinity for topo IIα, as shown in Table 4. Compounds 5, 7, 13, and 14 were not far behind the template compound, having a range of −0.5 to −0.1 kcal/mol difference. The superior binding affinity of 18 confirmed it to be a viable lead compound for optimization. Compounds 14 and 18 are commercially available and have been purchased for invitro testing and validation. The result will be published in a future article.

Compound 18 exhibited the highest molecular docking binding affinity to topo IIα (PDB 1ZXM). The 2D interaction diagram of the complex generated with Biovia Discovery Studio 2025 is shown in Figure 3 below.

Table 4. Binding affinities of compounds 1 - 18.

Figure 3. 2D interaction diagram of compound 3 in the active site of Topo IIα.

The template compound was searched in PubChem database [10] using its common name, ARN-21934. Using the “Similar Structures Search”, the settings were adjusted to 90% Tanimoto threshold. The filters were then adjusted from the defaults to only include compounds with molecular weights under 350 g/mol and LogP less than 3. If the entire database is searched, this search will yield over 1200 compounds, so for this round of screening, only the first 117 compounds, sorted by relevance, were used.

To obtain the ADME data for each compound, they were entered into SwissADME [14]. To use this electronic resource, the compounds must be in SMILES format. Using the sketching tool in SwissADME, each compound was drawn individually to be converted to SMILES format. After SMILES were generated for each compound, a complete data summary was generated. %ABS was derived from TPSA (obtained from SwissADME) using the formula % A B S = 109 − 0.345 T P S A [15]. For toxicity data, the SMILES structures of the 18 compounds were entered into ProToxII [16] and filtered to display hepatotoxicity, cytotoxicity, androgen receptor, and ATPase AAA domain containing protein.

Computer-based molecular docking can expedite the early stages of drug discovery via a systematic prescreening of small molecule ligands for shape and energetic compatibility with a receptor before experimental evaluation [17]-[19]. In addition, docking studies provide insight into the binding mode of potential drugs at atomic level [17]-[19]. We, therefore, commenced on the molecular docking of the 18 compounds as delineated below.

VirtualScreening: 18 compounds were screened virtually using PyRx 0.98 [20]. PyRx is an open-source program that combines the functionalities of Auto Dock, Vina, Auto Dock 4.2, and Open Babel [21].

Proteinpreparation: The crystal structure of human topo IIα ATPase/ADP was retrieved from the RCSB protein data bank (PDB ID: 1zxm) and was used as target for the screening of all 18 compounds. 1ZXM is a well-established target for catalytic inhibitors of topoisomerase. The above target was chosen because the template molecule was found to be a selective catalytic inhibitor of topoisomerase II. All water molecules were removed. PyRx automates many intermediate steps including addition of hydrogens steps, energy minimization, removal of bound ligands and heteroatoms, including setting of auto-grid dimensions. The target was initially prepared and saved in pdb file formats but was later converted to a pdbqt file format prior to docking simulation. Ligandpreparation: For the construction of ligand files from chemical structures, all eighteen compounds were drawn using ChemSketch, an open source program available as a free download from the developers (ACD Labs) [22]. Once the structure is drawn it is saved as a mol file and then imported into OpenBabel. OpenBabel is a versatile program that converts structural files between a volume of different formats. The ligand file is saved as a pdb file before importing into PyRx, which converts the pdb file into pdbqt file format for docking simulation.