A regression analysis using simple descriptors for multiple dermal datasets: Going from individual membranes to the...

Area of Science:

• Pharmacokinetics and exposure science
• Computational toxicology
• Dermal absorption modeling

Background:

• Accurate in silico methods are crucial for predicting chemical exposure through skin.
• The Potts-Guy equation, a regression model, uses logP and molecular weight (MW) to estimate skin permeability (Kp).
• Existing models show limitations with different skin layers, particularly dermatomed and full skin.

Purpose of the Study:

• To develop improved regression equations for estimating chemical skin absorption.
• To evaluate the performance of the Potts-Guy equation with recent human skin datasets.
• To explore a combined modeling approach integrating mechanistic and structural activity relationships.

Main Methods:

• Collected and analyzed human skin datasets for environmental and cosmetic chemicals.
• Applied multi-linear regression analysis based on the Potts-Guy equation.
• Developed an updated regression model by combining fitted permeability values from a 2D compartmental model.

Main Results:

• Regression fits for the Potts-Guy equation were stronger for epidermis-only datasets.
• Permeability estimates for dermatomed and full skin showed weaker regression coefficients.
• The updated regression equation (logKp = -2.55 + 0.65logP - 0.0085MW, R² = 0.91) demonstrated improved accuracy.

Conclusions:

• Combining mechanistic and structural activity relationships offers an advantage for skin absorption modeling.
• The developed model provides a more accurate prediction of skin permeability compared to the Potts-Guy approach alone.
• Future work will incorporate multi-compartment models for enhanced accuracy.