Molecular structure-based prediction of the partition coefficients of organic chemicals for physiological pharmacokinetic models

The objectives of the present study were (1) to develop and validate a methodology for predicting tissue:air (P-t:a) and blood:air (P-b:a) partition coefficients (PCs) of organic chemicals from molecular structure information, and (2) to incorporate this methodology within physiologically based pharmacokinetic (PBPK) models to enable automated calculation of PCs from molecular structure information provided as input to the model. The proposed methodology involves (1) estimating n-octanol:water or oil:water (P-o:w) PCs and water:air (P-w:a) PCs at 25 degrees C of chemicals from their molecular structure information using previously validated fragment constant methods, (2) correcting for the temperature dependence of vapor pressures using the Clausius-Clapeyron equation to extrapolate the P-w:a to 37 degrees C, and (3) incorporating these data along with data on volumes of neutral lipids, phospholipids, and water in tissues and blood in an algorithm to predict P-t:a and P-b:a. The predictions of rat and human P-t:a (liver, muscle, and adipose tissue) and P-b:a for 17 chemicals obtained with the present methodology were, in general, within a factor of two of the corresponding experimental values obtained from the literature. Following the incorporation of the elements of this methodology within a human PBPK model for dichloromethane, the P-t:a and Pb-b:a values were automatically calculated during each run within the model from the molecular structure information provided as input. The methodological approaches proposed in this article represent a significant step toward the development of PBPK models from molecular structure information provided as the sole input.