Use of laser capture microdissection to generate a translational map of gene expression patterns in the liver

Abstract

Listen

Translate article

Animal models are widely used in toxicology studies of new chemical compounds. The metabolism of drugs occurs largely in the liver, but interspecies differences in the molecular mechanisms at the basis of drug metabolism have been widely demonstrated. These differences may limit the translatability of the results to human, and compromise the safety assessment for new compounds. Urge therefore a more in-depth understanding of the molecular mechanisms of liver for the selection of suitable animal models, which are the most representative for the human. In addition an in-depth knowledge of the cells specific transcriptome may drastically improve the identification of the target organ of toxicity. Accordingly, in this experimental work we focused our attention in generating a molecular map of liver gene expression for different animal models, taking advantage from the use of laser capture microdissection for obtaining cell-specific transcriptomes, from rat, dog and monkey liver slices. We developed LCM and immuno-LCM protocols for liver zone I, liver zone III, bile ducts, hepatic arteries and portal veins specimen collection and we profiled the transcriptomes of these samples by oligonucleotide arrays. The analysis confirmed the enrichment of specific cells/tissues and allowed the identification of known and unknown tissue-specific genes. We have then investigated the impact of sexual dimorphism on the expression of zone specific genes. We showed that sexual dimorphism in liver zonation is a phenomenon interesting only rats and we couldn’t observe any significant differentially expressed gene in other species. The cross species comparison of gene signature for liver zone I, III and bile ducts, revealed a surprising low number of common genes. However this data might be dependent from technical limitations more than from actual biological diversity of similar tissues in different species. The specificity of our gene signatures has been proved, at protein level, by IHC, resulting in all the cases valid and revealing additional species specificities unknown so far. We have applied the knowledge generated in the first phase of this project to a retrospective case study on toxic effects of Methapyrilene in rat liver. The histopathology report described minimal histological changes in treated animals, including single cell necrosis, increase of mitotic figures and bile duct hyperplasia. At transcriptomic level, the exposure to Methapyrilene induced a mild impact on the global liver gene expression, and a specific gene expression pattern of toxicity was not detectable. The analysis of differentially expressed genes, revealed 1500 up regulated genes in treated animals vs. controls and the comparison of those genes with the bile duct gene signature revealed a constant number (120-130) of common genes at all the time points. Using two markers belonging to rat bile duct signature (Cldn7 and Krt19), we were able to quantitatively evaluate the bile duct hyperplasia, by measuring the area covered by labeled cells. However the revealed differences were statistically significant at day 7 and 14, but not at day 3. The generation of a liver gene expression map and of tissue specific signatures allowed an “informed” whole liver transcriptome analysis. This uncommon approach revealed subtle drug induced molecular changes and histological alterations, otherwise difficult to identify. Interestingly, the transcriptomic profiling resulted more sensitive and statistically powerful than the histological approach. We can then conclude that, although a mathematical model describing the empirical approach used in this project is still missing, the knowledge generated by our work improved the interpretation of a retrospective case study, allowing the identification of a specific toxicity pattern. Additionally, the possibility to apply the same approach in prospective studies and to different species, represents a powerful contribution for the understanding of liver pathophysiology and DILI.