Alternatives in Reproductive Toxicity: A Way Forward

Home / New Perspectives / Reproductive & Developmental Toxicity / Alternatives in Reproductive Toxicity: A Way Forward

Reproductive & Developmental Toxicity

Alternatives in Reproductive Toxicity: A Way Forward

George P. Daston, Procter & Gamble

Published: December 6, 2007

About the Author(s)
Dr. Daston oversees human safety research at Procter & Gamble, where he has worked for the past 22 years. He has published over 100 peer-reviewed articles, reviews and book chapters, and has edited four books. Dr. Daston has served as President of the Teratology Society (1999-2000); member of the National Academy of Sciences Board on Environmental Studies and Toxicology (1995-98); Councilor of the Society of Toxicology (2001-2003); member of the EPA Board of Scientific Counselors (2002 on); member of the National Toxicology Program Board of Scientific Counselors (2003-06; Chair in 2006); and member of the NIH National Children’s Study Advisory Committee (2003-06). He chaired the Developmental and Reproductive Toxicology Technical Committee at ILSI-HESI (1996-2004) and a similar committee at the American Industrial Health Council (1990-99). Dr. Daston is Editor-in-Chief of Birth Defects Research: Developmental and Reprodutive Toxicology, and is on the Editorial Board of Human and Ecological Risk Assessment. He served for six years as Associate Editor of Toxicologcal Sciences. Dr. Daston is an adjunct professor in the Department of Pediatrics and Developmental Biology Program at the University of Cincinnati and Children’s Hospital Research Foundation. Dr. Daston was a Visiting Scientist at the Salk Institute, Molecular Neurobiology Laboratory, 1993-94. Dr. Daston was elected a Fellow of AAAS in 1999 and of the Academy of Toxicological Sciences in 2005.

Dr. George P. Daston
Procter & Gamble
Miami Valley Laboratories
11810 E. Miami River Rd.
Cincinnati, OH 45253
E-mail: daston.gp@pg.com

The reproductive cycle is perhaps the most complex set of processes that organisms undergo. Development of a new individual from a merger of gametes, maturation of that individual, and the formation of gametes and production of the next generation involves multiple organ systems, covers a considerable amount of the lifespan of the organism, and includes processes, such as the formation of axial polarity, differentiation of organs and tissues, etc., that are not duplicated elsewhere in the life cycle.

Modeling the reproductive cycle to the extent that non-animal assays can predict human reproductive risk will be a monumental challenge. Still, a journey of a thousand miles begins with a single step. The good news is that we have started to take those steps, and work to model parts of the reproductive cycle has yielded promising results. Furthermore, developmental biology is among the most active areas of life sciences research, and a solid foundation of knowledge about the molecular control of development and the nature of interacting systems in development is being laid down.

We currently rely on animal tests to predict the potential for chemicals to cause reproductive harm in humans. Animal tests for assessing reproductive toxicity are designed to cover the entire reproductive cycle, either as a series of tests that evaluate specific segments of the reproductive cycle (reproduction/fertility, prenatal development, postnatal development), or as a single protocol (two-generation test). Because these tests evaluate structure and function from gametogenesis through embryonic and postnatal development to adulthood, they are complex in nature. They have served us well; the predictive capacity of these tests is good.

However, testing for reproductive toxicity is animal-intensive. Furthermore, the basic protocols for the tests have not changed much since they were implemented in the mid 1960s in response to the thalidomide tragedy. The opportunity exists to improve the informational output from these studies, which has the potential not only to further improve their value for prediction of risk, but also to provide a better mechanistic understanding of abnormal reproduction and development, which will be critical in developing mechanistically relevant non-animal alternatives.

At present, early embryonic development is the only phase of the reproductive cycle for which in vitro assays have been developed for the purpose of toxicity screening. A number of assays have been tried for this purpose, with mixed success. The assays fall into a few categories: intact rodent embryos in culture, primary cell cultures derived from embryos, mouse embryonic stem cell cultures, free-living embryos (vertebrate and invertebrate), and established cell cultures.

The use of established cell lines for developmental toxicity screening ended about 15 years ago, after a National Toxicology Program-sponsored validation study on two of them (human embryonic palatal mesenchyme (HEPM), and mouse ovarian tumor cells (MOT)) produced disappointing results. In retrospect, it is obvious why these assays failed to predict developmental toxicity. Each assay was capable of demonstrating only one developmentally relevant event: proliferation in the case of HEPM, cell adhesion in the case of MOT. Neither of these endpoints is highly specific for development, and given the complexity of development and the number of potential mechanisms by which developmental toxicity can occur, it isn’t a surprise that single-endpoint assays in cells that probably have few or no remaining developmental characteristics are not predictive.

Free-living embryos, at least in theory, ought to make good surrogates for mammalian development, primarily because developmental processes have been highly conserved during evolution. Relatively extensive programs to develop assays using free-living embryos to predict mammalian toxicity have been developed in three species: chick, Xenopus, and Drosophila, and less extensive research has been done using zebrafish, Japanese medaka fish, sea urchins, the nematode C. elegans, and Hydra.

Unfortunately, these programs have failed to produce results that can be used in any practical way for hazard evaluation. There a number of reasons for these failures, most of which are surmountable given enough research attention. Chief among these are 1) pharmacokinetic issues; and 2) selection of endpoints to measure that are relevant for mammalian development. There has recently been renewed interest in zebrafish as a developmental toxicity model. Zebrafish has been a leading model in developmental genetics, it has been sequenced, and large-scale programs to mutate all the developmentally relevant genes in the organism have led to the availability of sensitized models for the study of abnormal development. Recent research programs using zebrafish embryos as developmental screens have yielded promising results, and this is an area that deserves more attention.

It is also worth looking at Drosophila and C. elegans again, as the developmental genetics of these species has also been intensively studied. One possible reason for disappointing results in the past with these species is that we have been looking at the wrong readout of development. Evaluating the morphological manifestations of abnormal development may not be as easy to interpret or extrapolate. Perhaps it will be more fruitful to look at more highly conserved processes, such as signal transduction pathways or expression of developmentally conserved genes. Such an approach was suggested in a National Academy of Sciences review (NRC, 2000. Scientific Frontiers in Developmental Toxicology and Risk Assessment, NAS Press).

The situation for primary cell cultures and intact mammalian embryos is brighter, but the downside is that these assays require intact animals as the source of embryos or embryonic cells. Still, as an interim measure these assays show promise. These assays include the chick embryo retina cell culture (CERC), whole rodent embryo culture (WEC), and rat limb and CNS micromass cultures. The latter two assays have been evaluated in a European validation program and have been declared to be validated for some preliminary screening purposes. However, it is acknowledged that they cannot yet replace the traditional animal studies in regulatory decision-making, although they can be of use early in compound screening and may contribute to weight-of-evidence assessments.

Mouse embryonic stem (ES) cells were also evaluated in the European validation trial. The results were promising, but limited in scope compared to the WEC and micromass assays that were evaluated at the same time. There is considerable research in a number of labs to further validate the ES cell assay. It clearly has the potential to be a significant addition to the arsenal of in vitro tests for developmental toxicity. It has the advantage of not needing to be derived from animals for every experiment, as WEC and micromass do. Stem cells also have the capacity to differentiate into any other cell type, which may be valuable in the development of organ-specific toxicity assays. However, it will be necessary to investigate the appropriate culture conditions that will drive the stem cells towards specific developmental fates.

Research activities to model other aspects of the reproductive cycle have been limited. In vitro systems exist for some reproductive tissues and for some aspects of the reproductive cycle, but these have been used for mechanistic research and have not been optimized for toxicity testing. Most of them consist of organ cultures or primary cultures. Culture systems for pre-implantation embryos have been well established because of interest in in vitro fertilization techniques and reproductive cloning, but little attention has been paid to these as toxicology models. As for development from the early embryonic period through adulthood, there are no models available.

Progress in this area continues to be good, but we must be realistic about the fact that a much stronger fundamental mechanistic understanding of normal and abnormal reproductive development will be needed if we are to build useful models for predicting human risk. A great deal of fundamental knowledge is being generated in developmental biology and genetics, but this needs to be applied to evaluate toxicological questions.

Another unacknowledged but significant hurdle in this area is the development of methods that will allow the results of in vitro assays to be used for risk assessment, not just hazard identification. This problem is not unique to reproductive toxicology, but it needs to be acknowledged. Yes-no judgments about toxic potential tend to have limited application: they are useful in prioritizing for further testing or product development, but do little to support regulatory decision-making. There are ways to improve the design and output of the assays, but we must develop new methods and get beyond the old mindset in in vitro testing that its purpose is for initial screening and that therefore unreasonably high false positive rates are acceptable.

In summary, the way forward for alternatives in reproductive toxicity is a long one, but we have made good progress in some areas. New findings in basic developmental and reproductive biology will provide important underpinnings to the development of mechanism-based assays that will be highly predictive of human toxicity potential. We should be realistic that it will take many years to move from these basic findings to test methods that are true replacements. We should be optimistic that we have avenues to pursue that were not available to us until very recently and should support methods development and improvement.
©2007 George Daston