Human relevance of in vivo and in vitro skin irritation tests for hazard classification of pesticides.

Background: Test methods to inform hazard characterization and labeling of pesticides to protect human health are typically conducted using laboratory animals, and for skin irritation/corrosion the rabbit Draize test is currently required by many regulatory agencies. Although the Draize test is generally regarded to provide protective classifications for human health, new approach methodologies (NAMs) have been developed that offer more human relevant models that circumvent the uncertainty associated with species differences that exist between rabbits and humans. Despite wide applicability and use of these test methods across a broad range of chemicals, they have not been widely adopted for testing pesticides and pesticidal formulations. One of the barriers to adoption of these methods in this sector is low concordance with results from the Draize rabbit test, particularly for chemicals within the mild to moderate irritation spectrum.

Methods: This review compares and contrasts the extent to which available models used in skin irritation testing mimic the anatomy and physiology of human skin, and how each aligns with the known key events leading to chemically-induced adverse skin irritation and corrosion. Doing so fully characterizes the human relevance of each method.

Results: As alternatives to the rabbit Draize test, several protocols using ex vivo, in chemico, and in vitro skin models are available as internationally harmonized test guidelines. These methods rely on a variety of models of human skin, including excised rodent skin, synthetic biochemical models of barrier function, cell culture systems, and reconstructed human tissue models. We find these models exhibit biological and mechanistic relevance aligned with human skin irritation responses. Further, recent retrospective analyses have shown that the reproducibility of the Draize test is less than 50% for mild and moderate responses, with many of the replicate predictions spanning more than one category (e.g., a moderate response reported in one study followed by a non-irritant response reported in another study).

Conclusions: Based on this comparative evaluation, we recommend top-down and bottom-up testing strategies that use the most human relevant in vitro test methods for skin irritation and corrosion classification of pesticides and pesticide formulations. To further discriminate among mild and non-irritant formulations, optimization of a cytokine release protocol and subsequent analyses of reference formulation test results is recommended.