Integrative Data Mining, Scaffold Analysis, and Sequential Binary Classification Models for Exploring Ligand Profiles of Hepatic Organic Anion Transporting Polypeptides

Hepatocellular
organic anion transporting polypeptides (OATP1B1,
OATP1B3, and OATP2B1) are important for proper liver function and
the regulation of the drug elimination process. Understanding their
roles in different conditions of liver toxicity and cancer requires
an in-depth investigation of hepatic OATP–ligand interactions
and selectivity. However, such studies are impeded by the lack of
crystal structures, the promiscuous nature of these transporters,
and the limited availability of reliable bioactivity data, which are
spread over different data sources in the open domain. To this end,
we integrated ligand bioactivity data for hepatic OATPs from five
open data sources (ChEMBL, the UCSF–FDA TransPortal database,
DrugBank, Metrabase, and IUPHAR) in a semiautomatic KNIME workflow.
Highly curated data sets were analyzed with respect to enriched scaffolds,
and their activity profiles and interesting scaffold series providing
indication for selective, dual-, or pan-inhibitory activity toward
hepatic OATPs could be extracted. In addition, a sequential binary
modeling approach revealed common and distinctive ligand features
for inhibitory activity toward the individual transporters. The workflows
designed for integrating data from open sources, data curation, and
subsequent substructure analyses are freely available and fully adaptable.
The new data sets for inhibitors and substrates of hepatic OATPs as
well as the insights provided by the feature and substructure analyses
will guide future structure-based studies on hepatic OATP–ligand
interactions and selectivity.