Comparison of the Zebrafish Embryo Toxicity Assay and the General and Behavioral Embryo Toxicity Assay as New Approach Methods for Chemical Screening

Abstract

The movement away from mammalian testing of potential toxicants and new chemical entities has primarily led to cell line testing and protein-based assays. However, these assays may not yet be sufficient to properly characterize the toxic potential of a chemical. The zebrafish embryo model is widely recognized as a potential new approach method for chemical testing that may provide a bridge between cell and protein-based assays and mammalian testing. The Zebrafish Embryo Toxicity (ZET) model is increasingly recognized as a valuable toxicity testing platform. The ZET assay focuses on the early stages of embryo development and is considered a more humane model compared to adult zebrafish testing. A complementary model has been developed that exposes larvae to toxicants at a later time point during development where body patterning has already been established. Here we compare the toxicity profiles of 20 compounds for this General and Behavioral Toxicity (GBT) assay to the ZET assay. The results show partially overlapping toxicity profiles along with unique information provided by each assay. It appears from this work that these two assays applied together can strengthen the use of zebrafish embryos/larvae as standard toxicity testing models.

Keywords: developmental period; phenotype; toxicity assay; zebrafish larvae.

Figure 1

Correlation analysis between assays. The correlation between EC₅₀ (A) and LC₅₀ (B) values was determined between the ZET assay (x-axis) and GBT assay (y-axis). Both axes are displayed as log (values) in for ease of visualization. Compounds with no determined toxicity values for one (Benzophenone, Permethrin, and Thiobendazole) or both (Dechlorane Plus, Amoxicillin) of the assays were omitted from the corresponding analyses. Correlation coefficients (R²) for both correlation analyses are indicated. Removal of the outlying chemicals (Benzophenone, Valproic acid, and Thiobendazole) from the EC₅₀ analysis whose points on the graph are outside the 95% confidence interval of the linear regression (denoted by “×”) results in the correlation coefficient in parentheses.

Figure 2

Correlation analysis comparing the LOEC rankings between the ZET assay and the Behavioral endpoint of the GBT assay (A) and between the ZET assay and the phenotypic endpoint from the GBT assay (B). R² values for each are embedded in the plots.

Figure 3

Prevalence of phenotypic groups observed in the ZET (6–120 hpf) and GBT (72–120 hpf) larval toxicity assays for each tested chemical. Phenotypic groups are described in Table 2. The values represent the percentage of times the specified phenotype was observed for all the phenotypic observations of all exposure replicates for each compound. Values are color coded from highest percentage to lowest percentage as red to orange to yellow to green. The total number of phenotypic counts observed for each tested chemical is listed below each column. Note: low phenotypic counts for larvae exposed to Amoxicillin and Dechlorane Plus in both assays.