Predicting blood-brain barrier permeation from three-dimensional molecular structure.

Area of Science:

• Computational chemistry
• Medicinal chemistry
• Pharmacology

Background:

• Predicting blood-brain barrier (BBB) permeation is a significant challenge in drug discovery.
• Experimental determination of BBB partitioning for numerous preclinical candidates is impractical.
• Computerized models offer a desirable alternative for evaluating drug candidates' BBB permeability.

Purpose of the Study:

• To demonstrate the utility of 3D molecular field descriptors for estimating BBB permeation.
• To develop a simple mathematical model for external prediction of BBB permeability.
• To provide insights for drug design and screening processes.

Main Methods:

• Utilized the VolSurf method to transform 3D molecular fields into predictive descriptors.
• Employed a discriminant partial least squares procedure to correlate descriptors with experimental BBB permeation data.
• Developed a computational model for automated and rapid analysis.

Main Results:

• The developed model achieved over 90% accuracy in predicting BBB permeation for a large set of compounds.
• The model quantifies favorable and unfavorable contributions of physicochemical and structural properties.
• The computational procedure is fully automated and fast.

Conclusions:

• The 3D molecular field-based model is a valuable tool for predicting BBB permeation in drug design.
• The model provides insights into structure-permeability relationships, aiding drug optimization.
• This approach enhances virtual screening efficiency for prioritizing drug candidates.