New non-animal models and a weight of evidence approach could reduce the amount of animal testing required by regulators

Innovations for in vitro and in silico models for predictive toxicology are gaining ground. Typically, in order for regulatory bodies like the US Food and Drug Administration (FDA) to approve the use of new chemicals, many studies using animal and non-animal models must be run. These studies range from addressing physiological effects on people and wildlife to the ecological effects of a chemical ending up in the environment. One of the most crucial steps in the process is predictive toxicology, which researches the chemical’s effect on biological processes.

Part of what drives these innovations is our better understanding of the mechanism of action of toxic responses. Regulators are starting to trust New Approach Methodologies (basically any non-animal approach to chemical risk assessment), because our understanding of the strengths and limitations of non-animal models is improving every year.

Selection of the right non-animal model is critical and should be based on a sound scientific rationale. For example, there are a number of in vitro models for testing lung toxicity, ranging from simple in vitro 2D cell cultures through the more complex 3D tissue models and “lung on a chip” models. The favored regulatory models are the 3D tissue models which recapitulate the structure and functions of the upper and lower airway regions of the lung.

A pivotal inhalation study

In one 2018 case study by the US Environmental Protection Agency (EPA), researchers described how the pesticide, chlorothalonil, was analyzed for inhalation risk assessment for point of contact toxicity. In the pesticides original risk assessment in 2010, it was noted that a repeat dose inhalation study was not available, and the authors felt that there were too many animal welfare concerns and overestimation of toxicity. In order to avoid these issues, Syngenta suggested to the EPA that it should perform an in vitro assay to characterize the risk to human health from inhalation of chlorothalonil. Charles River performed the test for Syngenta using EpiThelix human upper airway tissue model, MucilAirTM,which is derived from primary human airway epithelial cells. As this was so successful, Syngenta incorporated this data into a New Approach Methodology with particle size distributions (PSDs), computational fluid dynamics (CFD), and worker breathing measurements to derive a point of departure for expected human equivalent concentrations (HECs) for the risk assessment. The EPA assessed the NAM and will evaluate this data for respiratory toxicity from registrants on a case by case basis.

In this case, in vitro models allowed for the replacement of in vivo models while still giving the sponsor and regulators the information they needed about the pesticide’s risks. If applied to similar problems, in vitro models like MucilAirTM could reduce the number of animals needed for predictive toxicology studies by replacing them entirely. To achieve the greatest potential of these models, regulations would need to continue to be updated to reflect the new scientific reality.

We have come a long way from the paradigm that risk assessment is simply based on animal safety studies, and the regulatory authorities are increasingly accepting of an integrated toxicology or weight of evidence approach using a broad range of different models. Better yet, in some specific, scientifically justified cases, they are not requiring animal studies at all. Does this mean that we will soon be running most preclinical safety programs without animals? No, but we are still moving in the right direction to limit their use to where they are absolutely required.

Our four-part series on predictive toxicology concludes tomorrow with some perspective from a toxicology expert on how the field should approach predictive toxicology studies going forward.