The discovery of the EDCs has added another chapter to the “low dose” issue in toxicology. This is not a new issue. The 1950s and the 1960s have witnessed some strong controversies between different groups of toxicologists. Some claimed that for most compounds there was no safe dose, even at low concentrations, while other claimed that, below a certain threshold that should be determined, most compounds were safe. The latter view prevailed but with the important exception of genotoxicant carcinogens for which it was considered that even very low doses could lead to irreversible effects (i.e. mutagenesis and long-term effects). This Yalta-like conclusion was the dawn of regulatory toxicology and of regulatory reference values below which compounds were considered essentially harmless. It should be noted that scientific foundations for the calculation of those reference values are at best controversial. EDCs have brought the dose issue to light again. First, it was observed that low doses, i.e. doses similar to a usual environmental contamination, can have significant effects in some experimental models, notably during developmental windows of vulnerability [3]. This means that some reference values that are determined based on regulatory and often non-comprehensive tests may not be protective enough. One has to keep in mind that most toxic effects related to EDCs have not been discovered through traditional regulatory tests, but rather by academic scientists exploring new mechanisms of toxicity. The second important point is that, in some cases, the dose response curve describing one toxic effect of a chemical as a function of its dose may not be monotonous. Intuitively, most of us would think that a toxic effect should increase with the dose. In reality, there are cases where effects are more potent at lower doses than at higher ones. This has been discussed at length in several conferences and papers [3,4]. The mechanisms are diverse and could be related to multiple mechanisms of action triggered at different doses or to the intrinsic properties of the endocrine system. Recently, Anses and other EU agencies have critically analyzed the literature for dose effects and concluded that, while some of the claims for non-monotonous dose response curves are overstated, there are indeed a few cases where dose response curves could confidently be considered as non-monotonous in humans [5]. One important consequence of non-monotonous curves is that regulatory tests should now encompass a much larger dose range than previously in order not to miss a specific low dose effect and that identification of reference doses may become even more difficult than in the past [6]. A reevaluation of the regulatory approaches to reference value determination appears to be required.

One of the most challenging tasks in toxicology is to understand the mechanisms of long-term effects leading to chronic diseases and to find the right models to study them. Long-term means years, decades and possibly generations! With the exception of mutagens, long-term effects were traditionally thought to be related to continuous exposure as in the case of air pollution and smoking. Toxicity related to long-term continuous exposure has some paradoxical features. Indeed, in many cases such a long-term toxicity is unexpectedly related to the adaptive metabolic pathways that are triggered by exposure to chemicals; those pathways, by allowing the elimination of chemicals are protective in the short term, but they also entail the transient production of very reactive intermediate compounds that may lead to toxicity in the long run [7]. What this is telling us is that the same pathway could be adaptive or toxic depending on the time scale that is considered. Long-term effects could also be due to the internal persistence of chemicals as in the case of Persistent Organic Pollutants (POPs) which are poorly metabolized and eliminated and which are stored in adipose tissue; the latter, in turn, becomes an internal source of continuous exposure [8]. Again, this tissue has a paradoxical effect toward POP handling; by storing those pollutants, the adipose tissue protects other sensitive organs such as the brain or gonads, but in the long run, it does constitute an internal source of chronic exposure.

With EDCs, a third mechanism was unraveled. Indeed, both experimental and epidemiological studies indicated that exposure to several EDCs at specific developmental stages was associated with an increase in the risk of disease later in life [9]. In that case, exposure can be either continuous or limited in time, but the targeted organism is in a state of high vulnerability. It is thought that vulnerability is due to the remodeling of tissues and organs during development and to limited defense mechanisms. The most likely mechanism is through the alteration of epigenetic marks that are somatically heritable and that therefore may persist for a long time [10]. Such alterations could lead to subtle changes in organ physiology that may increase the risk of disease development later in life. As an illustration, altered methylation of adipose tissue genes that are targets of the metabolically critical receptor PPARγ by certain EDCs are thought to increase the risk of adiposity in mice [11]. We still need more data linking EDC-triggered epigenetic regulation to health effects. We also need to develop the concept further and to weigh the evidence supporting the similarity between epigenetic toxicity and genotoxicity. Indeed, both can account for long-term effects, although epigenetic marks are reversible in the long run, which is not the case of mutations. Regulatory implications are important in that epigenetic toxicity may fall in the “no-threshold” toxic effect category.

Whereas traditional toxicology has primarily focused on the substance (“the dose makes the poison”), it has now become clear that the state of the targeted organism is also critical. Indeed, under certain conditions, organisms can be more vulnerable to the toxicity of chemicals [12]. This could be due, for example, to the genetic background of the contaminated individual, which has led to the development of gene–environment interactions in population studies. Another source of vulnerability is the health and social status, for example if the individual suffers from diabetes, kidney disease, or low socio-economic status. But, as stated earlier, most of the recent advances have been made on the developmental stage, since certain periods of development display a higher likelihood of toxicity in case of exposure to a chemical. Clearly, the extensive epigenetic remodeling and stem cell differentiation during development is a likely cause of vulnerability particularly when toxicity is related to a modification of programming [12].

Interest in mixture effects precedes the discovery of endocrine disruptors. It is in fact a traditional issue in pharmacology and toxicology and is particularly relevant when the identification of toxicants is related to the mode of action. For example, several dioxins, furans and PCBs share a similar mechanism of action, i.e. the activation of the ArylHydrocarbon Receptor, albeit with different potencies [13]. In addition, those compounds, which are highly lipophilic, tend to accumulate together in fat, so that co-contamination is very frequent. The toxic equivalency concept was developed to account for such combined exposures and the dose addition method was developed accordingly. Because EDCs are also identified by their mechanism of action, it was also relevant to classify them accordingly. Studies have focused on those that display a similar mode of action, for example anti-androgenic effects or xeno-estrogenic effects and, in that case, the current evidence suggests that dose addition should be used to account for mixture effects [14]. The study of the combination of EDCs with different mode of action is less advanced; authors have found different outcomes such additive, synergistic or antagonistic effects. This is a major challenge for the near future with implications in both basic and regulatory sciences.

EDCs have constituted a unique stimulus pushing toxicological concepts forward (Table 1). Together with the exposome concept [15], the endocrine disruption concept has led to a more systemic approach to toxicity, combining environmental sources of exposure, health effects and an integration of different stressors (chemical physical, biological, psychological, social) during the lifetime. There are implications both in basic sciences, since we are witnessing increasing links between toxicology, physiology, developmental biology and epigenetics as well as in regulatory sciences, which are considerable. The latter will probably change the landscape of regulation in the EU and worldwide. This is indeed a wonderful example of the challenges and opportunities of translating science into policy.