Implementation of transcriptomics for developmental toxicity assessment in the zebrafish embryo model

Abstract

The zebrafish embryo is considered a promising alternative model for in vivo rodent predictive developmental toxicity testing. The objective of the research described in this thesis was to enhance the detection of embryotoxic compounds and the knowledge about their mechanisms of action by the implementation of transcriptomics in the zebrafish embryotoxicity test (ZET). The basic principle of the ZET entails the exposure of zebrafish embryos to a test compound and the subsequent morphological assessment of developmental effects. A scoring method to determine the compound-induced morphological developmental effects, GMS, was developed and tested with compounds from two chemical classes. The ranking of the compounds within two chemical classes, glycol ethers and triazoles, showed comparable ranking to in vivo studies. This indicated that the zebrafish embryo with the GMS was an efficient and useful test system for screening potentially embryotoxic compounds. To increase sensitivity and predictability of the test, we implemented transcriptomics analysis using microarray analysis to determine whether it was possible to distinguish the classes of compounds, glycol ethers and triazoles, based on their gene expression profiles. Results showed very similar profiles for the compounds within the chemical classes and different profiles between the two classes. Concentration-dependency of gene expression was demonstrated using eight concentrations of flusilazole. Genes and pathways, known to be involved in the mechanisms of action and toxicity, were regulated concentration-dependently, even at concentrations at which no morphological effects were observed. By relating gene expression profiles of five other triazoles, tested in a single concentration, to the concentration–response of flusilazole, we demonstrated that gene expression data allow discrimination of relative potencies using gene sets associated with the mechanisms of action. In a next study, two approaches of defining common gene expression signatures of different compounds based on (1) commonly expressed genes and (2) regulated gene pathways were described. Pathway enrichment identified pathways suggestive of the mode of action of the embryotoxicants. Even if pathways were commonly regulated, gene expression regulation within the pathways, in terms of the nature of the genes as well as the level of regulation, differed for each compound. With the aim to generate a generalized transcriptomic signature for developmental toxicity in the zebrafish embryotoxicity test all transcriptomic data was combined. With identified biomarker genes, regulated processes and pathways were determined and a developmental toxicity network was created. Compound-specific gene expression was induced in different parts of the network, showing overrepresentation for specific clusters of functionally linked gene-sets. This network, responsive to embryotoxicants, is proposed as a first stage reference tool for studying developmental toxicity at the molecular level. With the results described in this thesis, we have shown that the implementation of transcriptomics in the zebrafish embryotoxicity test can increase sensitivity of the test and it contributes to the detection of mechanisms involved in developmental toxicity, thereby enhancing the applicability of this model for human risk assessment. The research presented in this thesis provides a foundation for further study of the model and its use in developmental toxicity testing.