Design of novel PI3Kα and PI3Kγ inhibitors for cancer treatment using pharmacophore, protein-ligand contacts, and...

Area of Science:

• Oncology and Bioinformatics
• Computational Chemistry
• Drug Discovery

Background:

• The phosphatidylinositol 3-kinase (PI3K)/AKT pathway is frequently altered in various cancers, driving cell survival and metastasis.
• This pathway represents a critical therapeutic target for developing novel anti-cancer treatments.
• Identifying effective inhibitors requires understanding the molecular interactions between drugs and PI3K pathway proteins.

Purpose of the Study:

• To develop predictive machine learning models for classifying and quantifying the binding affinity of PI3K inhibitors.
• To identify key molecular features and interactions that contribute to effective PI3K pathway inhibition.
• To guide the rational design of new anti-cancer therapeutics targeting the PI3K pathway.

Main Methods:

• Utilized machine learning models, including regression and classification approaches, trained on PLIP, PRODIGY, and RDKit-derived molecular features.
• Analyzed 136 PI3Kα and PI3Kγ co-crystallized ligands obtained from the Protein Data Bank (PDB).
• Performed pharmacophore mapping and molecular docking to investigate structure-activity relationships and binding interactions.

Main Results:

• Four regression models (Linear regression, SMOreg, MLP, Gaussian processes) demonstrated high predictive performance (MCC of 0.9).
• Identified favorable physicochemical properties for inhibitors, including heavy atom count (>25), rotatable bonds (>4), molecular weight (>400 Da), and log P (>2).
• Non-bonding interactions, such as CC, CO, and CX contacts, significantly influenced binding affinity.

Conclusions:

• Machine learning models effectively predict the efficacy of PI3K inhibitors based on molecular features.
• Specific molecular characteristics and non-bonding interactions are crucial for achieving high binding affinity to PI3K targets.
• These findings provide valuable insights for the design of potent and selective anti-cancer drugs targeting the PI3K pathway.