LAUSR

AIM To generate and validate predictive models for blood brain permeation (BBB) of CNS molecules using QSPR approach. BACKGROUND Prediction of molecules crossing BBB remain a challenge in drug delivery. Predictive models are designed for evaluation of set of preclinical drugs which may serve as alternatives for determining BBB permeation by experimentation. OBJECTIVE The objective of present study was to generate QSPR models for permeation of CNS molecules across BBB and its validation using existing in-house leads. METHOD The present study envisaged the determination of set of molecular descriptors which are considered significant correlative factors for BBB permeation property. Quantitative Structure Property Relationship (QSPR) approach was followed to describe the correlation between identified descriptors for 45 molecules and highest, moderate and least BBB permeation data. The molecular descriptors were selected based on drug likeness, hydrophilicity, hydrophobicity, polar surface area, etc. of molecules which served highest correlation with BBB permeation. The experimental data in terms of logBB were collected from available literature, subjected for 2D-QSPR model generation using regression analysis method like Multiple Linear Regression (MLR). RESULT The best QSPR model (Model 3) exhibited regression coefficient as R2= 0.89, F = 36; Q2= 0.7805 and properties such as polar surface area, hydrophobic hydrophilic distance, electronegativity etc., which were considered key parameters in determination of the BBB permeability. The developed QSPR models were validated with in-house 1,5-benzodiazepines molecules and correlation studies were conducted between experimental and predicted BBB permeability. CONCLUSION The QSPR model 3showed predictive results in good agreements with experimental results for blood brain permeation. Thus, this model was found to be satisfactory in achieving goodcorrelation between selected descriptors and BBB permeation for benzodiazepines and tricyclic compounds.