An integrative machine learning and structure-driven drug repositioning strategy for human IRAK4-targeted cancer...

Area of Science:

• Oncology
• Computational Chemistry
• Drug Discovery

Background:

• Interleukin-1 receptor-associated kinase 4 (IRAK4) is upregulated in many cancers, promoting tumor growth.
• Existing IRAK4 inhibitors have issues with selectivity and toxicity.
• There is a need for safer, targeted IRAK4 inhibitors.

Purpose of the Study:

• To develop a computational strategy for identifying non-toxic, mechanistically focused IRAK4 inhibitors.
• To repurpose existing FDA-approved drugs for cancer therapy targeting IRAK4.
• To validate potential drug candidates using in silico methods.

Main Methods:

• Employed a machine learning (ML) pipeline with structure-driven drug repositioning.
• Trained a random forest (RF) regression model on an IRAK4-targeting dataset.
• Screened a library of 1040 FDA-approved drugs, followed by molecular docking and molecular dynamics (MD) simulations.
• Assessed toxicity using graph neural network (GNN) analysis.

Main Results:

• The ML model demonstrated predictive accuracy in identifying potential IRAK4 inhibitors.
• Molecular docking confirmed binding efficacy for udenafil, linagliptin, benflumetol, and nimodipine.
• MD simulations and GNN analysis indicated thermodynamic stability and low toxicity for these candidates.
• Identified four FDA-approved drugs as credible candidates for IRAK4 inhibition.

Conclusions:

• Established a scalable computational strategy for identifying repurposable oncology therapeutics.
• Proposed udenafil, linagliptin, benflumetol, and nimodipine as promising candidates for experimental validation.
• Highlighted the potential of computational drug repositioning for developing safer cancer treatments.