5th World Congress on Alternatives & Animal Use in the Life Sciences, Session 5.3

A common animal model of chemical hepatocarcinogenesis was used to demonstrate the utility of genomic and proteomic data to identifyearly biomarkers for conventional toxicological endpoints in toxicological studies. N-nitrosomorpholine was applied to young adult male Wistar rats for 7 weeks followed by an exposure-free period of up to 43 weeks. A second group of untreated animals served as control. Five animals in each treatment group were sacrificed at 9 time points during and after exposure. Gene and protein expression of liver tissue homogenates were analyzed by Affymetrix Rat Genome Chips (RG-U34A) and 2D electrophoresis based proteomics. Proteomic analysis was performed in two laboratories. Proteins were separated using 2D electrophoresis (pH 3-10) and gel images were analyzed to obtain intensity values for each spot. Statistical evaluation using linear model analysis and Wilcoxon rank sum tests resulted in candidate genes and proteins differentially expressed between treated and control animals at early and late intervention time points. Time course of change in expression of deregulated proteins was compared to the related gene expression profiles. One focus of the study was to determine biomarkers for carcinogenicity characterized by pronounced differential expression at late stages of hepatocarcinogenesis. Among others, glutathione S-transferase P (GSTp) was identified by both approaches. To validate this finding, GSTp-stained liver sections of all animals were evaluated morphometrically and the area fractions of GSTp-positive foci of altered hepatocytes were determined. A strong correlation between area fraction of foci and both, gene and protein expression as well as between gene and protein expression could be demonstrated.