Repeated toxic injuries of murine liver are tolerated through microsteatosis and mild inflammation

The liver has a remarkable capacity to regenerate and thus compensates for repeated injuries through toxic chemicals, drugs, alcohol or malnutrition for decades. However, largely unknown is how and when alterations in the liver occur due to tolerable damaging insults. To that end, we induced repeated liver injuries over ten weeks in a mouse model injecting carbon tetrachloride (CCl4) twice a week. We lost 10% of the study animals within the first six weeks, which was accompanied by a steady deposition of extracellular matrix (ECM) regardless of metabolic activity of the liver. From week six onwards, all mice survived, and in these mice ECM deposition was rather reduced, suggesting ECM remodeling as a liver response contributing to better coping with repeated injuries. The data of time-resolved paired transcriptome and proteome profiling of 18 mice was subjected to multi-level network inference, using Knowledge guided Multi-Omics Network inference (KiMONo), identified multi-level key markers exclusively associated with the injury-tolerant liver response. Interestingly, pathways of cancer and inflammation were lighting up and were validated using independent data sets compiled of 1034 samples from publicly available human cohorts. A yet undescribed link to lipid metabolism in this damage-tolerant phase was identified. Immunostaining revealed an unexpected accumulation of small lipid droplets (microvesicular steatosis) in parallel to a recovery of catabolic processes of the liver to pre-injury levels. Further, mild inflammation was experimentally validated. Taken together, we identified week six as a critical time point to switch the liver response program from an acute response that fosters ECM accumulation to a tolerant "survival" phase with pronounced deposition of small lipid droplets in hepatocytes potentially protecting against the repetitive injury with toxic chemicals. Our data suggest that microsteatosis formation plus a mild inflammatory state represent biomarkers and probably functional liver requirements to resist chronic damage. Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=163 HEIGHT=200 SRC="FIGDIR/small/476054v1_ufig1.gif" ALT="Figure 1"> View larger version (53K): org.highwire.dtl.DTLVardef@11ecc7aorg.highwire.dtl.DTLVardef@1027df5org.highwire.dtl.DTLVardef@9ba0f7org.highwire.dtl.DTLVardef@164f80a_HPS_FORMAT_FIGEXP M_FIG C_FIG The datasets generated via transcriptomics, proteomics as well as blood, histopathological and biochemical analysis were analyzed in an independent and integrative manner. The independent analysis was performed via state-of-the-art statistical approaches i.e. differential and consistently regulated genes and proteins. Combining the results identified three fibrosis phases. Using the KiMONo algorithm, a fibrosis specific multi-omic network was inferred. Within this network we identified several nodes connecting phase III specific features forming 13 distinct multi-omic modules suggesting a tolerance scheme. Some of these modules were experimentally validated and compared to 11 independent human studies of various liver diseases.