Is a similar test possible for developmental toxicants? Probably not. Whereas carcinogens act in a limited number of ways, primarily by inducing mutations in the DNA of somatic cells, developmental toxicants probably act by a large variety of mechanisms involving many aspects of development. Attempts to use very simple metazoans, such as hydra, to test for general effects on development have proved to be unsuccessful, because the results obtained were not interpretable as predictive of mammalian responses. On the other hand, the rodent assays, which are now considered the most predictive of human developmental responses, are expensive and slow and hence suitable for assaying only a small number of compounds. As currently performed, they also might detect only gross effects.
From the knowledge now being gained about developmental mechanisms, it seems possible that many developmental toxicants (those that defy the drug-metabolizing defenses of the animal) will prove to act by perturbing the signaling pathways involved in the many inductive interactions between cells and tissues. (As previously emphasized, signaling pathways appear to be highly conserved among most animal phyla.) However, this hypothesis remains largely untested. Do known developmental toxicants affect signaling pathways, and if so, is this how they cause developmental defects? Pursuit of these questions is a search of mechanisms of developmental toxicity. Using the simple and relatively inexpensive animal model systems amenable to genetics, scientists should be able to answer these questions. If the answers are yes, as is the committee’s hypothesis, it should be possible to design evaluation approaches for potential developmental toxicants with the use of animals, having sensitized genetic backgrounds and reporter-gene outputs, to detect effects on specific signaling pathways, as described in further detail below. Results will have to be used with caution, so that false positives are not overinterpreted. However, a judiciously applied battery of such tests could represent a major advance in developmental toxicity testing.
Some mutations essentially shut down a pathway by completely inactivating a component. Others that produce less inactivation cause no visible phenotype, although they bring the pathway close to a threshold of function and, therefore, render it sensitive to changes of activity of other components of the pathway—changes that by themselves might be asymptomatic. In signaling pathways, such sensitization can be accomplished either by raising or by diminishing the level of activity of a particular component, depending on its activating or inhibiting contribution. Change in the activity of a second component of this pathway due to mutation would cause the threshold to be crossed to altered function and phenotypic consequences. Hence, depending on how the assay is established, a phenotype might be enhanced or suppressed by perturbing a second element. In attempting to define a pathway, genetic screens for new dominant mutations that enhance or suppress the pathway-defective phenotype in a mutagenized, geneti-