The comparative analysis of in vivo and in vitro transcriptome data based on systems biology

Abstract

We investigated whether an in vitro cellbased system can represent toxicity of an in vivo organ using gene expression profiles. We performed an analysis of differentially expressed transcripts, pathway analysis, and Gene Ontology (GO) grouping with the toxicity-induced data following treatments with diclofenac, sulindac, itraconazole, and ketoconazole. The number of genes regulated in vitro and in vivo were much more than the randomly sampled number, but no significant correlation was observed between the in vitro and in vivo experiments. We performed a pathway analysis and GO grouping to expand the approach. In the pathway analysis, we narrowed down the hepatotoxic pathways to focus on toxicity. Then, the percentages of overlapping pathways increased. We found pathways associated with liver function in all experiments, such as the adipocyte signaling pathway, JAKSTAT signaling pathway, and peroxisome proliferatoractivated receptor (PPAR) signaling pathway. In the GO grouping analysis, the clusters obtained were primarily related to lipid metabolic and synthetic processes. The percentage of common GO terms between in vitro and in vivo experiments also increased. Therefore, we identified than an analysis performed at the systems biology level was more correlated for in vitro and in vivo systems.