Comparative functional genomics of autoimmune hepatitis; mechanistic and clinical implications

Autoimmune hepatitis (AIH) is a major chronic liver disease characterized by diverse and pleiotropic spectrum of different phenotypes. Despite recent progress in clinical management of AIH patients, the exact molecular pathogenesis remains largely unknown. Here, we used a comparative genomic approach to define molecular characteristics and test the predictive value of gene expression profiles in this important liver disease. Gene expression profiles of 124 patients with different chronic liver diseases (AIH, HBV, HCV) were generated from FFPE samples using Illumina DASL assay and computationally integrated with expression data from a T-cell induced mouse model of autoimmune liver disease.

A total of 2439 genes were significantly different between AIH patients and patients with chronic viral hepatitis. Unsupervised cluster analyses revealed three distinct clusters of samples with AIH specimens predominantly located in two of these main clusters. GO pathway analyses demonstrated that key pathways associated with immune response (e.g. T-cell apoptosis, natural killer cell activation), cytoskeleton organization and DNA integration were active in AIH patients. While no chromosomal hotspots could be identified, pathways involved in demethylation and histone modification were also activated in AIH patients indicating that epigenetic mechanisms might contribute to the observed changes in gene expression. Application of comparative functional approach using an AIH mouse model further demonstrated high concordance with related molecular pathways. Furthermore, unsupervised cluster analyses revealed that AIH mouse samples were closely linked to a subpopulation of human AIH patients, indicating that the model could be useful to study molecular alterations of these patients.

Here we provide first in-depth molecular analyses of human AIH patients and defined distinct AIH subclasses that might be relevant for diagnostic and therapeutic interventions. Furthermore, cross-species comparison identified a mouse model that closely mimics the genomic characteristics of human AIH. The obtained results might be helpful to identify novel biomarkers for diagnosis, allow precise risk stratification and, improve existing therapies that target relevant genes and pathways in this important liver disease.