Systematic Investigation of the Molecular Descriptors for the Bioavailability of Oral Drugs

Abstract

The molecular descriptors for predicting the drug-likeness of a small molecule was first proposed by Lipinski. While the lipinski rule of five describes four key physicochemical properties or molecular descriptors such as molecular weight, partition coefficient (Log P), hydrogen bond donors and hydrogen bond acceptors for good absorption and permeability of oral drugs, other rules such as the Veber’s, Ghose, Egan’s and Muegge’s suggest many other key descriptors for better absorption of oral drugs.  The aim of this study is to understand the correlation between the various molecular descriptors or a combination of descriptors and their influence on the bioavailability of oral drugs. For this study, structurally diverse, orally administered drugs from diverse therapeutic categories were selected primarily from the National list of Essential Medicines 2022, India [https://www.who.int/publications/m/item/india--national-list-of-essential-medicines-2022-english)] for the first time.  As the dataset included neutral, acidic and basic oral drugs, in addition to log P, pH-dependent distribution coefficient (log D) at different physiologically relevant pH are also considered. Using Fraction Lipophilicity Index (FLI) which combines log P and log D as a new measure of permeability and log S as a measure of solubility, an attempt to predict Biopharmaceutical Classification System (BCS) classes of the oral drugs from the national list of essential medicines resulted in successful prediction of 70% of oral drugs into class 1 and class 2. Such an attempt is expected to provide early bioavailability profiles of drug-like small molecules and eventually narrow the search space for experimental validation.