From protein binding to pharmacokinetics: a novel approach to active drug absorption prediction

TL;DR

This paper introduces a novel protein binding-based method for predicting active drug absorption, leveraging large-scale docking to identify correlations with intestinal permeability and improve absorption models.

Contribution

It presents a new paradigm for absorption prediction using protein activity data, moving beyond traditional QSPR approaches.

Findings

Identified a protein whose binding correlates with intestinal permeability.

Enhanced passive absorption models with a non-linear flux component.

Demonstrated that protein binding data can predict active absorption.

Abstract

Due to inherent complexity active transport presents a landmark hurdle for oral absorption properties prediction. We present a novel approach carrier-mediated drug absorption parameters calculation based on entirely different paradigm than QSPR. We capitalize on recently emerged ideas that molecule activities against a large protein set can be used for prediction of biological effects and performed a large scale numerical docking of drug-like compounds to a large diversified set of proteins. As a result we identified for the first time a protein, binding to which correlates well with the intestinal permeability of many actively absorbed compounds. Although the protein is not a transporter, we speculate that it has the binding site force field similar to that of an important intestinal transporter. The observation helped us to improve the passive absorption model by adding non-liner flux associated with the transporting protein to obtain a quantitative model of active transport. This study demonstrates that binding data to a sufficiently representative set of proteins can serve as a basis for active absorption prediction for a given compound.