Systems biology elucidates common pathogenic mechanisms between nonalcoholic and alcoholic-fatty liver disease

Abstract

The abnormal accumulation of fat in the liver is often related either to metabolic risk factors associated with metabolic syndrome in the absence of alcohol consumption (nonalcoholic fatty liver disease, NAFLD) or to chronic alcohol consumption (alcoholic fatty liver disease, AFLD). Clinical and histological studies suggest that NAFLD and AFLD share pathogenic mechanisms. Nevertheless, current data are still inconclusive as to whether the underlying biological process and disease pathways of NAFLD and AFLD are alike. Our primary aim was to integrate omics and physiological data to answer the question of whether NAFLD and AFLD share molecular processes that lead to disease development. We also explored the extent to which insulin resistance (IR) is a distinctive feature of NAFLD. To answer these questions, we used systems biology approaches, such as gene enrichment analysis, protein-protein interaction networks, and gene prioritization, based on multi-level data extracted by computational data mining. We observed that the leading disease pathways associated with NAFLD did not significantly differ from those of AFLD. However, systems biology revealed the importance of each molecular process behind each of the two diseases, and dissected distinctive molecular NAFLD and AFLD-signatures. Comparative co-analysis of NAFLD and AFLD clarified the participation of NAFLD, but not AFLD, in cardiovascular disease, and showed that insulin signaling is impaired in fatty liver regardless of the noxa, but the putative regulatory mechanisms associated with NAFLD seem to encompass a complex network of genes and proteins, plausible of epigenetic modifications. Gene prioritization showed a cancer-related functional map that suggests that the fatty transformation of the liver tissue is regardless of the cause, an emerging mechanism of ubiquitous oncogenic activation. In conclusion, similar underlying disease mechanisms lead to NAFLD and AFLD, but specific ones depict a particular disease signature that has a different impact on the systemic context.

Figure 1. Graphic illustration of genes/proteins co-occurrence and their relatedness to biological concepts with the query “alcoholic AND (steatosis OR fatty liver) NOT (non or nonalcoholic)”.

Prediction was performed by PESCADOR (available at http://cbdm.mdc-berlin.de/tools/pescador/), a web-based tool to assist large-scale integration text-mining of biointeractions extracted from MEDLINE abstracts. The graph was constructed using the free available program MEDUSA, which is a Java application for visualizing and manipulating graphs of interaction (www.bork.embl.de/medusa) , .

Figure 2. Graphic illustration of genes/proteins co-occurrence and their relatedness to biological concepts with the query “nonalcoholic OR non-alcoholic AND fatty liver OR steatosis”.

Prediction was performed by PESCADOR (available at http://cbdm.mdc-berlin.de/tools/pescador/), a web-based tool to assist large-scale integration text-mining of biointeractions extracted from MEDLINE abstracts. The graph was constructed using the free available program MEDUSA, which is a Java application for visualizing and manipulating graphs of interaction (www.bork.embl.de/medusa) , .

Figure 3. Graphic illustration of a functional modular map of the multiple gene/ protein analysis encompassing the candidate list of NAFLD and AFLD based on disease pathways.

Results of functional association analysis performed by the bioinformatics resource ToppCluster (http://toppcluster.cchmc.org) based on pathways networks showing enriched terms from Gene Ontology, Mouse Phenotype, Co-expression, microRNAs, and transcription factors for the NAFLD- and AFLD-specific gene/protein lists. Right side of the figure depicts the highly significant enrichments for sets of genes and proteins of the NAFLD term list; left side of the figure depicts the highly significant enrichments for sets of genes and proteins of the AFLD term list; and the analysis of genes and intersection of pathways between NAFLD and AFLD is shown in the center part of the figure. Terms in red represent genes/proteins, and terms in green represent disease pathways in GO terms. The graph was constructed using the free available program Cytoscape, a software project for integrating biomolecular interaction networks with high-throughput expression data and other molecular states into a unified conceptual framework .

Figure 4. Comparative co-analysis of NAFLD and AFLD data sets focused on insulin signaling.

Results of functional association analysis performed by the bioinformatics resource PESCADOR (available at http://cbdm.mdc-berlin.de/tools/pescador/), a web-based tool to assist large-scale integration text-mining of biointeractions extracted from MEDLINE abstracts with a focus in the selected terms. The graph was constructed using the free available program, MEDUSA, which is a Java application for visualizing and manipulating graphs of interaction (www.bork.embl.de/medusa) , . The thickness of the green lines signifies greater significance.

Figure 5. Functional enrichment analysis of putative miRNAs associated with NAFLD and AFLD.

The network is shown as a cytoscape graph generated from ToppCluster (available at toppcluster.cchmc.org/) network analysis.