Gene expression patterns associated with histopathology in toxic liver injury

Predicting liver injury after exposure to toxic industrial chemicals is complicated by the large number of potential environmental contaminants, mixtures, and exposure dose and route scenarios. Identifying indicators of end organ injury can complement exposure-based assays and improve predictive power. A multiplexed approach was used to experimentally evaluate a panel of 67 genes predicted to be fibrogenic by computationally mining DrugMatrix, a publicly available repository of gene microarray data. Five-day oral gavage studies in male Sprague-Dawley rats dosed with varying concentrations of three fibrogenic compounds (allyl alcohol, carbon tetrachloride, and 4,4’-methylenedianiline) and two non-fibrogenic compounds (bromobenzene and dexamethasone) were conducted. Fibrosis was definitively diagnosed by histopathology. Transcriptomics data matched the predictions made using the DrugMatrix data with greater than 90% accuracy. Microarray data were verified using a 67-plex panel Bioplex® assay, confirming that the 67-plex panel constituted a biomolecular signature of hepatic fibrosis (Figure). Necrosis and inflammatory infiltration were comorbid with fibrosis. Interaction analysis identified 24 genes specific for the fibrosis phenotype. The protein product of the gene most strongly correlated with the fibrosis phenotype (Pcolce) was dose-dependently elevated in plasma from animals administered fibrogenic chemicals (p<0.05). PCOLCE is a novel biomarker candidate of fibrotic injury. These results support the development of gene panels for liver injury and may suggest bridging biomarkers for molecular mediators linked to histopathology. To induce the pathophysiological effects of exposure to 5 toxicants, rats were dosed by oral gavage for 5 consecutive days (+ 10 minutes from the previous day's dose administration time) with either one of 5 toxicants or the vehicle control. Volumes were based on daily body weight of the animal. Twenty-four hours after the final dose administration, animals were euthanized and livers collected. Exposures for allyl alcohol, bromobenzene, and carbon tetrachloride each had 4 doses (low, mid, mid-high, high) and a vehicle control with an n=5 for a total of 25 animals per toxicant. For dexamethasone, only a high dose was performed with an n=5 for an additional 5 animals. A vehicle control was shared with 4,4’-methylenedianiline (MDA) since the exposure was concurrent with the dexamethasone exposure. For MDA, 3 doses (low, mid, high) were performed along with the shared vehicle control with an n=5 for an additional 20 animals. Liver tissue was analyzed from each animal for a total of 100 arrays. Four arrays did not meet QC criteria and were excluded from analysis. Therefore, only 96 were included and reported here: 22 allyl alcohol, 24 bromobenzene, 25 carbon tetrachloride, 5 dexamethasone, and 20 MDA.