Functional Characterization and In Silico Prediction Tools Improve the Pathogenicity Prediction of Novel Bile Acid...

The pathogenicity of cholestatic liver diseases (CLDs) remains insufficiently characterized, hindering definitive diagnosis and timely treatment. The aim of this study was to improve the pathogenicity prediction of novel bile acid (BA) transporter variants in patients with CLDs. We analyzed the clinical characteristics and genetic profiles of a CLD cohort (n = 57) using multiple in silico tools and in vitro functional assays. We identified 78 unique variants in four BA transporter genes. The predominant defects were associated with ABCC2 (57/78, 73.1%), with the most frequent being missense variants (39/78, 50.0%). Using in silico tools, we identified 47 novel variants: 12 mis-splicing, 21 deleterious missense, and 23 with altered protein stability. Of the 34 novel variants in ABCC2 identified through in vitro functional assays, seven incurred aberrant splicing, 11 missense variants resulted in MRP2 reduction, 9 missense variants resulted in abnormal N-glycosylation, 18 variants altered MRP2 localization, and 26 variants reduced organic anion transport activity. These findings indicate that a multidisciplinary approach, integrating bioinformatics and experimental data, significantly enhances the accuracy of genetic-based CLD diagnosis. It serves as a foundational study for BA transport variants pathogenicity reclassification and expands the mutation spectrum of CLDs in China.