Related Experiment Videos
QSAR analysis of interstudy variable skin permeability based on the "latent membrane permeability" concept
Shin-Ichi Fujiwara1, Fumiyoshi Yamashita, Mitsuru Hashida
1Department of Drug Delivery Research, Graduate School of Pharmaceutical Sciences, Kyoto University, Yoshidashimoadachi-cho, Sakyo-ku, Kyoto 606-8501, Japan.
Journal of Pharmaceutical Sciences
|September 23, 2003
Summary
This study developed a unified quantitative structure-activity relationship (QSAR) model for skin permeability, integrating multiple datasets. The model effectively predicts compound skin penetration using octanol/water partition coefficient and molecular weight.
Area of Science:
- Pharmacokinetics
- Computational Chemistry
- Dermatology
Background:
- Quantitative structure-activity relationship (QSAR) models for skin permeability often lack consistency due to interspecies and interlaboratory variations.
- Existing models struggle to reconcile diverse datasets, hindering reliable predictions of transdermal drug delivery.
Purpose of the Study:
- To develop a robust and unified QSAR model for predicting skin permeability by integrating multiple experimental datasets.
- To identify essential molecular descriptors that consistently explain skin permeability across different species and experimental conditions.
Main Methods:
- Collected and analyzed ten diverse skin permeability datasets from existing literature, encompassing 111 permeability coefficients for 94 compounds across human, hairless mouse, and hairless rat skin.
- Employed a novel statistical approach, adapting Potts and Guy's methodology, to simultaneously analyze all datasets, assuming a shared underlying structure-permeability relationship.
- Utilized octanol/water partition coefficient (log P) and molecular weight (MW) as key molecular descriptors.
Main Results:
- Achieved high determination coefficients (R2) across all ten datasets, with an average R2 of 0.815, demonstrating the model's strong predictive power.
- The unified model showed comparable performance to individual analyses (average R2 = 0.825), despite limitations in degrees of freedom.
- Indicated a nearly equal contribution ratio (approximately 1:1) between log P and MW in explaining skin permeability.
Conclusions:
- A consistent and essential QSAR model for skin permeability can be extracted by integrating multiple datasets using a novel statistical approach.
- Octanol/water partition coefficient and molecular weight are crucial and balanced predictors of skin permeability across various models.
- This unified approach offers improved reliability for predicting compound skin penetration in drug development and toxicological assessments.