Blogs

Uncertainty factors were developed in the 1980s by the US Environmental Protection Agency (US EPA) to derive a margin of safety of a chemical when the data are extrapolated from animals to humans. As we move toward animal-free testing with the application of new approach methods (NAMs), animal testing is being slowly replaced by in vitro studies using human cells and organ-on-a-chip models. Are uncertainty factors required when human cells are used? This was debated during the Roundtable Session at the Virtual SOT 60th Annual Meeting and ToxExpo titled “The Future of Uncertainty Factors with In Vitro Studies Using Human Cells.”

Dr. Suzanne Fitzpatrick from the US FDA, who was chairing the session, presented the historical use of uncertainty factors and the need for regulatory risk assessments to evolve as we move toward human-relevant assays. Dr. A. Wallace Hayes of the University of South Florida highlighted publications that provided a historical perspective on uncertainty factors, acceptable daily intake, and reference dose. He emphasized the application of safety factors considering hypersensitive populations like children and the elderly. Dr. Hayes, who also was moderating the session, introduced the two debaters, Dr. Michael L. Dourson from Toxicology Excellence for Risk Assessment, who debated “for the question,” supporting the use of uncertainty factors in in vitro studies using human cells, and Dr. Lorna Ewart from Emulate Inc., who debated “opposing the question,” that uncertainty factors will more than likely not be needed in in vitro studies using human cells.

Dr. Dourson presented his standpoint that the use of different uncertainty factors for different chemicals is often necessary because the underlying experimental data are not always uniform. He spoke about the safe dose concept. As the dose increases and moves above the safe dose up to a point where adverse effects are observed, there is a “fog of uncertainty.” Dr. Dourson said even though uncertainty factors have been considered arbitrary, they are actually imprecise because the underlying biology is imprecise. He highlighted five commonly used uncertainty factors and explained their use in a dose-response assessment. There is always a human variability uncertainty factor, owing to sensitivity among populations. In the absence of human data, which is most often the case, risk assessment relies on dose-response curves from animal data and use of appropriate safety factors. However, in such cases, the risk is not reduced; we only move from the animal dose-response curve to the human dose-response curve. Risk reduction is observed only when two safety factors (i.e., UF_L – LOAEL to NOAEL and UF_H – Human variability) are used. Dr. Dourson elaborated that the human variability in toxic response in reality is well beyond 10-fold, as demonstrated through data from clinical trials and observational studies and in in vitro systems. The use of a 10-fold UF_H safety factor benefits by covering even the sensitive human population and provides a reference dose (RfD). From a future perspective, in vitro data from human cells will eliminate the need for at least one safety factor (i.e., animal to human extrapolation). However, the variability observed among in vivo databases will most likely be observed in in vitro databases, too, and uncertainty factors will be needed when there are no LOAEL or NOAEL data from in vitro studies. Also, critical effects like loss of body weight observed in vivo cannot be mimicked in vitro, leading to additional uncertainties. Dr. Dourson concluded on a lighter note that he will keep his calculator’s division key in good working order since uncertainty factors are here to stay.

Dr. Ewart presented her standpoint on how advanced in vitro technologies can reduce the need for safety factors. She stated that risk assessment is a conservative endeavor to arrive at a safe human dose; however, in today’s age, it is warranted to consider alternative and complementary approaches to achieve the same goal but with fewer uncertainties. So, to bring a greater degree of certainty, we need to break down into steps the distance between a dose and a toxic response. By assessing the “target organ dose,” “target organ metabolism,” and “target organ response,” we can potentially reduce the need for use of uncertainty factors. Dr. Ewart provided supporting evidence through experimental data generated from organ-on-a-chip models. She explained how the cellular microenvironment is recreated in organ-chips and how the tissue-tissue interface and the physical microenvironment are similar to in vivo conditions. She presented data on how organ-chips can enable better approximation of absorption and can enable the determination of target organ dose. Further, target organ-chips can determine if the dose is in the relevant toxic range and can reveal differential human sensitivities, which is lacking in other in vitro models. She also showed how liver-chips can accurately predict liver toxicity and the IC₅₀ is close to in vivo data. Also, liver-chips are more accurate than 3D hepatic spheroids. Similar trends were seen with other models, like intestine and bone-marrow-chips. Bone-marrow-chips also have been used to study the concentration-effect time relationship analogous to in vivo toxicokinetic exposure. Further, the question of species differences also has been answered by organ-chips. Dr. Ewart’s concluding remarks were that uncertainty factors are required to account for interspecies differences and for human variability; however, organ-chips can provide an alternative to safety factors.

Both presenters provided a rebuttal to their standpoints before the Q&A session. Two panel members, Dr. Silvia Barros from the Universidade de São Paulo and Dr. Brinda Mahadevan of BRINCOR Associates LLC, put forth interesting questions. Dr. Barros’s question was addressed to Dr. Ewart: How can chronic in vivo exposure studies be modeled on organ-chips? Dr. Ewart explained that cellular functionality has been studied up to 28 days on organ-chips. However, the chronicity question can be answered through a combination of 28 days of good, solid data and building a mathematical model around that to understand chronic exposure.

Dr. Mahadevan’s question was addressed to Dr. Dourson: Considering that animal data are obtained from high-dose studies and then moving on to low doses, how can this complement what is done on organ-chips? Dr. Dourson explained that the advantage of in vitro systems is that mechanisms of toxicity are better understood in vitro than in vivo, that there are limitations of observing critical effects in vivo, and that maybe in vitro can provide answers in such cases. Additionally, mixture assessments can be assessed lot quicker through in vitro systems than through in vivo.

The Roundtable discussion concluded with the panel members agreeing to the fact that collaboration is needed between traditional risk assessors with decades of experience with in vivo data and risk assessors working with modern technologies like organ-chips to find a way forward.

This blog was prepared by an SOT Reporter and represents the views of the author. SOT Reporters are SOT members who volunteer to write about sessions and events in which they participate during the SOT Annual Meeting and ToxExpo. SOT does not propose or endorse any position by posting this article. If you are interested in participating in the SOT Reporter program in the future, please email Giuliana Macaluso.

Sessions delivered during the 2021 SOT Annual Meeting and ToxExpo will be available on the Virtual Meeting Platform on demand to registrants through May 31, 2021.