Concerns about the widespread use of rodent models for human risk assessments of endocrine disruptors

René Habert,Marie-Justine Guerquin,Romain Lambrot,Alexandra Benachi,Tiphany Grisin,Vincent Muczynski,Abdelali Lehraiki,Virginie Rouiller-Fabre,Sébastien Messiaen,Thierry N'Tumba-Byn,Gabriel Livera,Delphine Moison,Géraldine Delbes,René Frydman,Christine Levacher

doi:10.1530/rep-13-0497

Abstract

Fetal testis is a major target of endocrine disruptors (EDs). During the last 20 years, we have developed an organotypic culture system that maintains the function of the different fetal testis cell types and have used this approach as a toxicological test to evaluate the effects of various compounds on gametogenesis and steroidogenesis in rat, mouse and human testes. We named this test rat, mouse and human fetal testis assay. With this approach, we compared the effects of six potential EDs ((mono-(2-ethylhexyl) phthalate (MEHP), cadmium, depleted uranium, diethylstilboestrol (DES), bisphenol A (BPA) and metformin) and one signalling molecule (retinoic acid (RA)) on the function of rat, mouse and human fetal testis at a comparable developmental stage. We found that the response is similar in humans and rodents for only one third of our analyses. For instance, RA and MEHP have similar negative effects on gametogenesis in the three species. For another third of our analyses, the threshold efficient concentrations that disturb gametogenesis and/or steroidogenesis differ as a function of the species. For instance, BPA and metformin have similar negative effects on steroidogenesis in human and rodents, but at different threshold doses. For the last third of our analyses, the qualitative response is species specific. For instance, MEHP and DES affect steroidogenesis in rodents, but not in human fetal testis. These species differences raise concerns about the extrapolation of data obtained in rodents to human health risk assessment and highlight the need of rigorous comparisons of the effects in human and rodent models, when assessing ED risk.

Highlights

In regulatory toxicology, the human health risk from exposure to a given endocrine disruptor (ED) is classically assessed using animal, especially rodent, models followed by an extrapolation of the data to humans
During the last 20 years, we have developed an organotypic culture system that maintains the function of the different fetal testis cell types and have used this approach as a toxicological test to evaluate the effects of various compounds on gametogenesis and steroidogenesis in rat, mouse and human testes
To define the regulatory acceptance for human health, the lowest no observed adverse effect level (NOAEL) measured in the rodent model is divided by a safety factor equal to ten to account for the difference between rodents and humans and by an additional tenfold uncertainty factor to account for interindividual differences in susceptibility

Summary

Introduction

The human health risk from exposure to a given endocrine disruptor (ED) is classically assessed using animal, especially rodent, models followed by an extrapolation of the data to humans. In the 1990s, we compared different in vitro systems and found that the best functional and histological results were obtained by placing rat fetal testis explants on a membrane that floats on the culture medium at the air–medium interface (Habert et al 1991) This organotypic culture system allows reproducing in vitro the normal development of testis somatic and germ cells without addition of any exogenous signalling factor (Habert et al 1991, Lecerf et al 1993, Olaso et al 1998, Livera et al 2000). The treated testicular explants are compared with explants issued from the same testis (when the fetal testes are large enough to be cut into different pieces, i.e. with human or older rodent testes) or with the contralateral testis from the same fetus (for 14.5–15.5 dpc rat and 12.5–13.5 dpc mouse fetuses) that are cultured with vehicle only and served as controls This paired analysis limits individual variability and increases the sensitivity of the method.

Conclusion

Discussion from meeting