Screening for Endocrine Disruption

Environmentalists are increasingly concerned that pre-approval testing of chemicals, including pesticides, insufficiently addresses potential toxicity to the endocrine system(1). Unease arises from evidence in humans of decreasing sperm counts in men, and increases in birth defects and cancers affecting reproductive organs(2). Effects on wildlife provide evidence that hormonally-active compounds are present in the environment at biologically significant concentrations(3). The last issue of Pesticides News drew attention to testing inadequacies.  Marlissa Campbell points out that testing techniques are available, but a multidisciplinary approach is needed.   

Many biological activities occur in response to stimulation of a receptor by a specific signal molecule, or hormone. Molecules other than the natural hormone may be able to bind the receptor, thereby stimulating (or blocking) the activity regulated by that receptor. Possible consequences of this inappropriate binding include: altered growth and development of young organisms, abnormal development of sexual organs, infertility, impaired sexual behaviour, and increased susceptibility to certain cancers. 'Endocrine toxicity' is not generally considered as a separate endpoint, but rather a mechanism by which many toxic effects can occur.

Testing practices
Pre-approval data requirements for pesticides are similar among developed countries. In the UK, studies required to assess potential risks to human health include: acute toxicity, skin and eye irritation, allergic sensitization, neurotoxicity, mutagenicity , reproductive toxicity, long-term toxicity and/or carcinogenicity. These studies are designed to be 'apical' tests: their purpose is to detect any adverse effects on the organism, not to identify the mechanism responsible for observed toxicity.
    The aim of a multigeneration reproductive toxicity study is to observe at least one generation of animals (usually rats) which have been constantly exposed to the test chemical throughout all life stages. Parental animals (FO generation) are exposed prior to mating, the pregnant females are then exposed throughout pregnancy and lactation, and the weaned pups (F1 generation) are exposed through sexual maturity, mating and production of their own offspring (F2 generation). This design allows examination of animals exposed throughout their life-span not only for impaired reproduction, but also for parameters of growth and development—such as survival, anatomical normality, body weight at different time points, and age at sexual maturation. Reproductive organs and small endocrine glands, such as the thyroid and pituitary, are specifically evaluated for any pathology.

Multi-generaltional testing
Dr. Theo Colborn, of the World Wildlife Fund US, has stated that “Endocrine-disrupting effects are not currently considered in assessing risks to humans, domestic animals, and wildlife” 3. She expresses particular concern for effects resulting from exposures occurring early in life, but which may not become manifest until adulthood. I disagree with her comment, in general, as the multigeneration test described above specifically addresses such effects. The real problem is that comparatively few chemicals currently in the marketplace have been subjected to such detailed scrutiny. For example, it has been estimated that only 4,000 out of over 60,000 man-made chemicals currently used in the US have been tested for harm to reproduction and the developing organism(4). Attacking at least some aspects of this backlog is where short-term, in vitro, tests of hormonal activity could prove valuable.

DNA techniques may help
One suggested strategy requires developing a panel of tissue culture assays for receptor-mediated activity(5). Recom-binant DNA techniques would be used to transfect cells with the gene encoding a particular receptor along with a 'marker' gene—a gene directing synthesis of a product which is easy to measure, and which increases proportionately to stimulation of the receptor. A screening panel would consist of several engineered cell lines expressing receptors such as those for estrogen, progesterone, androgens, glucocorticoids, retinoids, thyroid hormone, peroxisome prolife-rators, etc.
    Pharmaceutical companies already use similar strategies for screening the activity of potential drugs; it is not clear that such assays have been validated for toxicity testing. For example, the human androgen receptor has been expressed in the budding yeast, Saccharomyces cerevisiae(6). Binding of androgenic substances to the human receptor stimulates production of a marker protein by the yeast organism. Measuring the concentration of the marker protein indicates the extent to which androgen receptors are activated, and hence the androgenic potency of the test chemical.
    Hormone-sensitive cell lines derived from tumours are another potential source of in vitro bioassays. Cultured cells derived from an estrogen-sensitive human breast cancer have been used to assay for estrogenic activity(7). These cells are stimulated to divide by adding estrogen, or chemicals having estrogenic activity, to the culture media. The 'E-SCREEN' successfully identified a series of chemicals known to have estrogenic activity, and did not falsely identify any non-estrogenic chemicals. There were false negatives, however. PCBs were not active in the E-SCREEN, possibly because metabolic activation is required to generate the estrogenic effects of the compounds.
    The problem of modelling metabolic activation in vitro may be addressed by an assay system using liver cells from male fish. John Sumpter (Brunel University, Uxbridge, UK) is reported to be using expression of vitellogenin by male fish as a biomarker for estrogenic compounds2. Vitellogenin is a precursor of egg yolk, which is normally detectable in the plasma of female fish; male fish produce it only when abnormally stimulated. Wild fish could provide a sensitive indicator of estrogenic pollutants in the environment, while their cultured liver cells could be used for screening chemicals in the lab.

Conclusions
Reproductive problems, birth defects, and cancer are among the toxic endpoints which can be caused by hormonal mechanisms. However, as any of these endpoints can alternatively be caused by other mechanisms, it would not be appropriate to substitute a set of specific endocrine tests for the current standard apical studies. Nor is it clear that adding such screens to pre-approval requirements would improve the accuracy of regulatory decisions. In the case of new pesticides, having reasonably comprehensive data sets, the problem is perhaps less with the type of data required than with the way the information is used. I would highlight three specific areas for improvement:

• basic research aimed at improving our ability to predict risk; 

• post-approval monitoring; and 

• communication between wildlife biologists and human toxicologists.

In addition to providing information about otherwise untested chemicals, in vitro techniques are powerful tools for basic research. Specific receptor-activating assays can be used to pin down mechanisms of toxicity, as well as to clarify relationships between structure and toxic potency. This is the type of information which will, in the long run, improve our overall ability to predict risk from chemical exposures.
    At present, there are no systematic programmes for monitoring environmental levels or chronic effects of pesticides (or other chemicals). We need to know what and how much is out there and what the effects are. We need to be prepared to take appropriate action to solve any problems identified.
    The gap between wildlife biologists and those of us specializing in assessing risks to human health is basically a historical accident. It starts with education in separate academic departments, and progresses to effectively separate assessment of environmental and human toxins. We don’t really know what data on wildlife can tell us about potential human toxicity, or vice-versa. A multidisciplinary approach is needed, such that techniques from one scientific field are brought to bear on problems identified by another.

References
1. Link, A. and D. Buffin, Endocrine Disruptors. Pesticides News, 23 March 1994.
2. Environmental Estrogens linked to reproductive abnormalities, cancers. Chemistry & Engineering News, 31 January 1993.
3. Colborn, T., Vom-Saal, F.S. et al., Developmental effects of endocrine disrupting chemicals in wildlife and humans. Environmental Health Perspectives. 101:378-374.
4. Reproductive toxicity: regs slow to change, Science. 254:25, 1991.
5. Mclachlan, J.A. Functional toxicology: a new approach to detect biologically active xenobiotics. Environmental Health Perspectives. 101:386-387.
6. Purvis, I.J., Chotai, E., et al., An androgen-inducible expression system for Saccharomyces cerevisiae. Gene. 106:35-42, 1991.
7. Soto, A.M, Lin, T.-M. et al., An “in culture” bioassay to assess the estrogenicity of xenobiotics (E-SCREEN). In: Chemically-Induced Alterations in Sexual and Functional Development: The Wildlife/Human Connection. T. Colborn & C. Clement (eds.). Princeton Scientific Publishing Co., Inc., Princeton, N.J., 1992.

Marlissa Campbell is a toxicologist, formerly working with the US EPA.

[This article first appeared in Pesticides News No. 24, June 1994, pages 8-9]