Predicting Drug-Induced Cholestasis with the Help of Hepatic Transporters—An in Silico Modeling Approach

Abstract

Cholestasis represents one out of three types of drug induced liver injury (DILI), which comprises a major challenge in drug development. In this study we applied a two-class classification scheme based on k-nearest neighbors in order to predict cholestasis, using a set of 93 two-dimensional (2D) physicochemical descriptors and predictions of selected hepatic transporters’ inhibition (BSEP, BCRP, P-gp, OATP1B1, and OATP1B3). In order to assess the potential contribution of transporter inhibition, we compared whether the inclusion of the transporters’ inhibition predictions contributes to a significant increase in model performance in comparison to the plain use of the 93 2D physicochemical descriptors. Our findings were in agreement with literature findings, indicating a contribution not only from BSEP inhibition but a rather synergistic effect deriving from the whole set of transporters. The final optimal model was validated via both 10-fold cross validation and external validation. It performs quite satisfactorily resulting in 0.686 ± 0.013 for accuracy and 0.722 ± 0.014 for area under the receiver operating characteristic curve (AUC) for 10-fold cross-validation (mean ± standard deviation from 50 iterations).

Computational Models Using Multiple Machine Learning Algorithms for Predicting Drug Hepatotoxicity with the DILIrank Dataset

“…Among the most important chemical descriptors associated with liver toxicity in our study were the mean atomic polarizability (Mp), the percentage of hydrogen atoms (H%), the Geary autocorrelation of lag 1 weighted by mass (GATS1m), normalized spectral positive sum from Burden matrix weighted by mass (SpPosA_B(m)), and Moriguchi octanol-water partition coefficient (MLOGP). Atomic polarizability, identified as important by the majority of selection algorithms used, was previously shown to associate with renal toxicity of drugs [55], but it was also used in other liver toxicity models, as was logP [17,56], the latter measuring lipophilicity, which was shown to correlate with DILI [23]. GATS1m is less intuitive than the constitutional descriptors, but it has also been reported in a different publication as an important descriptor for the liver toxicity of drugs [57].…”

Comparing the performance of meta-classifiers—a case study on selected imbalanced data sets relevant for prediction of liver toxicity

“…The external test sets for OATP1B1 and 1B3 inhibition from our previous study served as test datasets in this study [ 19 ]. The test set for human cholestasis was compiled in two stages from two previous studies [ 21 ]. The positives for human cholestasis were compiled from literature [ 22 – 25 ] and from the SIDER v2 database [ 26 , 27 ].…”

“…The positives for human cholestasis were compiled from literature [ 22 – 25 ] and from the SIDER v2 database [ 26 , 27 ]. As cholestasis is one of the three types of drug induced liver injury (DILI), and the compounds that are negative for DILI will also be negative for cholestasis, the negatives for drug-induced liver injury compiled in a previous study [ 21 ] were used as negatives for cholestasis. Overall, the external human cholestasis dataset consisted of 231 compounds.…”

Vienna LiverTox Workspace—A Set of Machine Learning Models for Prediction of Interactions Profiles of Small Molecules With Transporters Relevant for Regulatory Agencies

“…org/), then a combination of feature selection, MetaCost, and support vector machines with RBF kernel was used for learning. The cholestasis model uses data from Kotsampasakou and Ecker (2017) and a combination of MetaCost and a tree algorithm to predict whether a compound is likely to cause cholestasis or not. Finally, the DILI model is based on a 966-compound dataset carefully compiled from literature (Kotsampasakou et al, 2017c).…”