Use of the Distribution Coefficient in Brain Polar Lipids for the Assessment of Drug-Induced Phospholipidosis Risk

In vitro safety assessment in early drug discovery represents an important step to detect potential safety-related liabilities. It reduces late stage attrition and allows candidate optimization. In this study, we report on the use of the LogDBPL assay (a recently published assay for the determination of drug distribution coefficients between an aqueous phase and porcine brain polar lipids extract) for phospholipidosis (PLD) risk evaluation. The LogDBPL parameter was first compared to the effective permeability in the parallel artificial membrane permeability assay (PAMPA), previously reported as correlating with PLD risk. Subsequently, the LogDBPL for a set of 234 drugs with known PLD effect was measured, representing the largest data set of LogDBPL data published so far, and the correlation with phospholipid accumulation was further investigated. In addition, a comparison with other in silico methods based on physicochemical parameters is reported. Results showed that LogDBPL is an efficient descriptor to assess PLD risk, especially when corrected using the pKa value of compounds, being superior to the distribution coefficient in octanol, LogDOCT, and the effective permeability in the PAMPA assay. A multivariate statistical analysis approach was finally used to better define the intrinsic features of LogDBPL, whose effect proved to be highly similar to that of volume of distribution in silico when used to predict brain distribution and PLD.