A Systems Biology Approach to Identify Pathways Implicated in Liver Radiation Injury Using Integrated Epigenetic and Expression Profiling

Abstract

It is suspected that liver irradiation induces many genome level events, but little is known about the precise nature of these changes and how they contribute to Radiation Induced Liver Disease (RILD). We hypothesize that radiation exposure in normal liver tissue induces genome-wide epigenetic effects which mediate the long-term phenotypic consequence of liver fibrosis. We sought to use a network-based approach to identify significant epigenetic changes and link them with the progression of liver injury after irradiation. In 2 experiments, anesthetized C57BL/6 mice (N=4, N=3) were exposed to 50 Gy single-dose whole-liver irradiation with a control group (N=3, N=3) receiving no irradiation. At one month post-treatment all animals were euthanized and liver sections were obtained for RNA and DNA extraction. Gene expression profiles were assessed with Affymetrix Mouse Gene ST 2.0 arrays, and genome-wide DNA methylation profiles were assayed using HELP-tagging, a high-throughput sequencing approach. Gene enhancers were identified using H3K4me1 mouse liver ChIP-seq experiments from a publicly available database (ENCODE). A network-based analysis, SMITE, was applied to identify subnetworks enriched in genomic deregulation. ConsensusPathDB-mouse, an integrated network of gene interactions was used to identify subnetworks enriched for both epigenetic and expression events. Curated gene networks (KEGG and REACTOME) were used to validate identified network properties. Enrichment statistics were computed using Chi-square and Fisher exact tests. The largest significantly enriched module (Chi-square p=0.012) centered around STAT3, an immune system modulator and contained 80 interconnected genes with diverse functions. Within the module, 28 of the genes significantly enrich (p=2.93e-27) in human REACTOME Immune System Pathway further indicating an immune system defect after radiation exposure. 15 genes within the module significantly enrich in KEGG Pathways in Cancer (p=2.01e-17), with several genes unrelated to the Immune System Pathway, such as Pias3, Itga3, Itga6, Pdgfr, and Vegfb. Multiple DNA methylation events occurred in the gene bodies and gene enhancers of three histone deacetylase (HDAC) genes, signifying a relationship between epigenetic signaling and immune system deregulation within the context of liver fibrosis. The network analysis also implicated highly connected genes involved in fibrosis without direct evidence of DNA methylation or expression changes, such as JAK2. 15 genes significantly enrich (p=2.28e-22) in the human KEGG JAK-STAT Pathway providing evidence of deregulation of the JAK-STAT regulatory network. Integrated epigenetic and expression analysis identified multiple, novel dysregulated networks connecting immune system changes with epigenetic modification. These findings implicate STAT3 as regulator of RILD and suggest a role for aberrant histone deacetylation in its progression.