Combinatorial and fragment-based in silico design of PI3K-alpha natural hybrid antagonists for breast cancer therapeutics

Aim: The PI3K (phosphoinositide 3-kinase)-alpha isoform is found upregulated in 30–40% of breast cancer. Currently, there are limited selective and specific drugs that target PI3K-alpha, and no natural therapeutic option is available. This study aims to develop natural hybrid antagonists of PI3K-alpha for breast cancer therapeutics. Methods: 25 pan-PI3K and PI3K-alpha targeting drugs were obtained from various sources, including the COCONUT (Collection of Open Natural Products) database. On the parent dataset, high throughput virtual screening (HTVS), standard precision (SP) docking, and extra precision (XP) docking were performed to produce Murcko scaffolds and heterogenous fragments. Murcko scaffolds are hybridized with fragments of natural compounds (Category 1) and drugs (Category 2), respectively. Hybrids are docked with HTVS, SP, and XP, followed by induced fit docking and ADME (absorption, distribution, metabolism, and excretion) prediction. MM/GBSA (molecular mechanics/generalized Born and surface area) was performed on the docked poses. Results: Highest docking scores of –13.354 kcal/mol and –12.670 kcal/mol were achieved by hybrids in Category 1 and Category 2, respectively. MM/GBSA free energy ranged from –51.14 kcal/mol to –72.66 kcal/mol. In terms of binding docking, pharmacological properties, and Lipinski’s rule of five, the natural hybrids outperformed the parent drugs. Conclusions: PI3K-alpha kinase proteins can be targeted with natural-drug hybrid antagonists for breast cancer treatment. Hybrid molecules, such as NH-01 and NH-06, show better binding with promising ADME properties. Thus, in vivo and in vitro testing is necessary to prove the value of such hybrids.