Endocrine active agents: implications of adverse and non-adverse changes

The US Environmental Protection Agency (EPA) is currently in the process of developing screening and testing methodologies for the assessment of agents that may possess endocrine-like activity--the so-called endocrine disruptors. Moreover, the EPA has signaled its intention of placing information arising from such studies on the worldwide web. This has created significant interest in how such information may be used in risk assessment and by policymakers and the public in the potential regulation or deselection of specific chemical agents. The construction of lists of endocrine disruptors, although fulfilling the requirements of some parties, is really of little use when the nature of the response, the dose level employed, and the lifestage of the test species used are not given. Thus, we have already seen positive in vitro information available on the interaction with a receptor being used as a key indicator when the results of large, high quality in vivo studies showing no adverse changes have been ignored. Clearly a number of in vitro systems are available to ascertain chemical interaction with specific (mainly steroid) hormone receptors including a number of reporter gene assays. These assays only provide indicators of potential problems and should not be, in isolation, indicators of toxicity. Likewise, short-term in vivo screens such as the uterotrophic and Hershberger studies are frequently conducted in castrated animals and thus indicate the potential for a pharmacological response in vivo rather than an adverse effect. A number of new end points have been added to standard rodent testing protocols in the belief of providing more sensitivity to detect endocrine related changes. These include the measurement of anogenital distance (AGD), developmental landmarks [vaginal opening (VO), preputial separation (PPS)], and in some studies the counting of nipples and areolae on males. AGD, VO, and PPS are all affected by the size of the pup in which they are measured and should always be compared using bodyweight as a covariate. The historical control database for such changes is gradually growing, albeit that if pups are not individually identified it becomes problematic to associate any change with a specific malformation or to assess whether a delay or advance in, for example, developmental landmarks is biologically significant. Agents that significantly reduce AGD in males (it is an androgen-dependent variable) frequently have other more adverse changes associated with this end point (eg, reproductive tract malformations), but a 2 to 3% change in AGD although measurable is unlikely to be biologically of importance and in isolation would not necessarily be considered adverse. Retention of thoracic nipples in male rat pups is also an indicator of impaired androgen status. Recent studies have also shown that this retention for some endocrine active chemicals is permanent. Thus, the presence of a permanent structural change that is rarely found in adult control animals could be considered a malformation and therefore a developmental adverse effect on which risk assessment decisions could be made. The advent of multigeneration reproduction studies as the definitive studies for the assessment of the dose-response relationships and risk assessment for endocrine disruptors has shown that current testing protocols may be inadequate to reliably detect the adverse effects of concern as only 1 adult/sex/litter is examined. A number of the effects on reproductive development although, due to an in utero exposure, will not be manifest until after puberty or at adulthood. The use of only a limited number of animals to examine such changes, particularly for weaker acting materials indicates that some agents may have been examined in well-conducted, modern protocols but have insufficient power to detect low incidence phenomena (eg, a 5% incidence of malformations).