US EPA struggles to replace animal tests for pesticide toxicity

In early 2016, the US Environmental Protection Agency set an immediate goal to reduce the number of animals used to test the toxicity of pesticides. The agency claims that it is making significant progress toward meeting that goal, but manufacturers are still using tens of thousands of laboratory animals each year to demonstrate that new pesticides entering the US marketplace meet safety standards.

The primary way the EPA’s pesticide office has reduced the use of animals is by granting waivers for certain tests. The agency has been much slower to identify and validate nonanimal methods to replace in vivo animal studies.

One of the biggest hurdles to replace in vivo animal studies with in vitro human-relevant tests is a lack of correlation between the results of human and animal studies. Toxicologists know that animal studies are not always reproducible and useful for predicting adverse human health effects. Even so, animal studies are still the gold standard that scientists try to model and replicate with in vitro systems. A growing group of scientists is calling for that to change.

The EPA evaluates about 500 new pesticide formulations plus a small number of new active ingredients each year. The agency typically requires manufacturers to provide data on acute dermal, oral, and inhalation toxicity, as well as skin and eye irritation and skin sensitization. Completing all six acute mammalian toxicity tests, referred to as the six-pack, requires more than 50 laboratory animals for each formulation or active ingredient.

If the six-pack tests can be modernized, “applicants and registrants will be able to use in vitro and computational approaches that have the potential to reduce the number of animals needed for testing by over 20,000 per year,” an EPA spokesperson says. The EPA did not grant C&EN’s request to interview someone from its pesticide office. The agency instead provided written answers to questions submitted by C&EN, saying the information should be attributed to an unnamed spokesperson.

The EPA uses the six-pack studies to classify pesticides into four categories according to their toxicity. The most dangerous pesticides are category I. Products in this category have the most stringent personal protective equipment (PPE) labeling requirements to protect workers who apply the pesticides.

The six tests measure only acute effects, which occur within a short time after exposure to a chemical, as opposed to medium- or long-term effects—called subchronic and chronic, respectively—which happen after repeated exposures over several days or weeks. Evaluating subchronic and chronic effects requires additional animal tests.

To satisfy the six-pack testing requirements, manufacturers can, in some cases, use data from already-registered products that are similar in composition. Manufacturers can also request waivers for some of the tests if the pesticide meets certain criteria. The EPA grants waivers when a test is not relevant to a particular chemical, according to guidelines released by the agency in 2012. For example, the agency would consider waiving acute oral toxicity tests when a chemical exists as a vapor or gas.

In 2016, the EPA released additional guidelines, announcing that it would waive the acute dermal test for new pesticide formulations. The EPA justified its decision with an analysis comparing oral and dermal acute toxicity data previously submitted to the agency for a range of different pesticides.

If the EPA used just the oral studies, “PPE requirements on the labelling would have been equally protective or more protective” for 95% of the formulations analyzed, the EPA reported in 2016. The agency is still working on guidance for waiving the acute dermal test for active ingredients in pesticides.

Of the remaining five tests, the EPA has made the most progress toward nonanimal testing for skin sensitization—an inflammatory skin reaction caused by an allergic response after repeated skin contact with a chemical. In April 2018, the agency began accepting nonanimal approaches to test active ingredients for skin sensitization. The policy builds on methods already established by the Organisation for Economic Co-operation and Development (OECD).

Those methods use a combination of in vitro cellular assays, computational approaches, and biochemical tests, such as peptide-binding assays. None of the methods can be used alone as a replacement for the animal test. Each test targets a specific event along the pathway that leads to skin sensitization.

Such tests, however, have been validated only for active ingredients. “The assays need to be validated for formulations and mixtures, and that work is being done now,” the EPA spokesperson says.

The situation is similar for the skin and eye irritation tests. “There are OECD in vitro test methods that have been around for many years now, and they have been validated,” says Kristie Sullivan, vice president for research policy at the Physicians Committee for Responsible Medicine, an advocacy group that promotes scientific research without the use of animals.“But all of the validation was done with either cosmetics or personal care products or industrial chemicals. Maybe some active ingredient pesticides. The real gap is, again, the pesticide formulations and mixtures.”

During the past few years, companies have voluntarily provided pesticide formulations to toxicology-testing laboratories, such as Integrated Laboratory Systems and the Institute for In Vitro Sciences (IIVS). The labs are working closely with industry, the EPA, and other stakeholders to determine how well the OECD in vitro tests work on the formulations. So far, for both eye and skin irritation, “we aren’t quite finding what we are looking for,” Sullivan says.

One of the challenges is the classification system that the EPA uses for pesticides. That system is different from the Globally Harmonized System of Classification and Labelling of Chemicals (GHS) used in the European Union and by the US Occupational Safety and Health Administration.

Nonanimal test methods for skin and eye irritation have been validated for GHS categorization. They are not easily transferable to the EPA’s categories because the GHS has three categories, whereas the EPA has four, says Hans Raabe, vice president and chief operating officer at IIVS.

It is quite straightforward to predict skin irritation using the currently available in vitro test methods, Raabe says. “This is due in large part to our mechanistic understanding of cellular responses in the skin after chemical exposure and the fact that these responses can be modeled with available in vitro test methods.”

Tests using “reconstructed human epidermal tissue constructs were successfully validated for identifying chemicals that fall within the GHS classification of corrosive and skin irritant categories,” Raabe says.

The catch is that in vitro tests are better at predicting the extremes—corrosive and nonirritant. It is the EPA’s middle categories—severe and moderate irritant—that are difficult to predict with the in vitro tests.

“The EPA and GHS skin irritation category systems are not particularly well aligned,” Raabe notes. “It certainly would be easier if all regulatory systems followed one approach—for example, the GHS categorization scheme—so that the test methods validated for GHS schemes would be directly applicable to the EPA.”

Regulators in the EU and many countries that follow the GHS categorization already accept nonanimal methods for skin irritation for many purposes.

One problem that all regulators around the world face is the reproducibility of the in vivo test itself, Sullivan says. Chemicals tested multiple times for eye and skin irritation using animal tests have discordant results, according to analyses of the literature. “How do you decide what the correct answer is?” Sullivan asks.

“I think the time has come for us to break our reliance on the animal test as the reference standard and begin to rely on the categorization of the in vitro test methods to move forward for classification and labeling purposes,” Raabe says. The newer in vitro methods include tissues, organ culture systems, and chemical exposures that are more relevant for humans than the animal studies. “We spend a lot of time pondering why we can’t get 100% correlations to the animal test results when we should be asking ourselves why we assume that the animal test is always correct,” Raabe says.

For the remaining two six-pack tests—acute oral and inhalation toxicity—the EPA is considering accepting calculations used under the GHS that add up the hazards of each chemical in a mixture as an alternative to animal tests. A pilot program is underway to collect data from in vivo oral and inhalation toxicity studies on pesticide formulations to evaluate the ability of the GHS mixtures equation to predict the acute toxicity categories derived from traditional animal studies.

The EPA acknowledges that nonanimal approaches can yield more information than animal studies and improve the agency’s ability to predict the hazards of chemicals. The agency is working closely with stakeholders to find and validate nonanimal methods, but “it is a highly time-consuming process,” the EPA spokesperson says. The EPA cannot predict how long it will take to replace the six pesticide tests, but the agency says that it will continue to waive tests and accept alternative approaches “as the science is ready for use.”