The use of zebrafish feeding trials for the safety assessment of genetically modified crops

Isabelle Gabriëls, MSc, PhD

Date: 6th June 2019, at 5 PM

Location: Auditorium O5, Campus Drie Eiken


In the EU, a mandatory part of the safety assessment of genetically modified crops (GMOs) consists of a 90-day rodent feeding trial. As a relevant addition to the current assessment strategy, our aim was to develop a complementary zebrafish feeding trial specifically designed to address a number of shortcomings of existing test methods, using maize as a case study.

First, to ensure the use of nutritionally balanced experimental diets, we investigated whether substitution of a non-GM maize into the zebrafish diet causes component-related effects. High dietary maize percentages resulted in an overall disturbed uptake and processing of carbohydrates, influencing processes such as growth rate, and affected mRNA transcriptional liver patterns of genes involved in key metabolic processes. Balancing the physiological tolerance of the fish and achieving a toxicologically relevant dose of the GMO, 15% was selected as the optimal maize substitution level to be used for all experiments.

Next, a subchronic zebrafish feeding trial was developed for the toxicological safety assessment of GMOs. Endpoints included general health parameters, effects on energy metabolism, and reproductive effects including transgenerational effects. As a key aspect of our experimental design, we investigated the importance of different non-GM comparators. Our results highlight the importance of distinguishing between potential effects caused by the transformation process (i.e., the process of genetically modifying the plant) and those caused by the presence of the transgenic event itself. Indeed, most differences (in general health endpoints as well as transcriptional liver profiles) were identified between the conventional wild type control (i.e., the original genetic starting material) and null segregant (i.e., the plant progeny created during the transformation process but lacking the transgenic event) on the one hand, and between wild type and GM maize on the other hand, while differences between GM maize and null segregant were limited. To assess the toxicological relevance of biological responses, we further optimized a method allowing the interpretation of results relative to the natural response variation, i.e., the normal biological range in endpoint responses of the fish. We established natural response variation datasets for various zebrafish endpoints and discussed equivalence testing as a conceptual approach using examples originating from the subchronic/transgenerational feeding trial.

Overall, this work forms a strong foundation for advancing zebrafish feeding trials and, to facilitate the development of further strategies, main conclusions and recommendations have been summarized in a Guidance Document for the safety evaluation of GM crops using the zebrafish model.