ASSESSMENT OF ENDOCRINE ACTIVE SUBSTANCES USINGIN SILICO AND REPORTER ASSAYS IN CELL AND MICE

In the last two decades, there have been growing scientific concern, public debate, and media attention over the possible deleterious effects in humans and wildlife that may result from exposure to substances that have the potential to interfere with the endocrine system. Endocrine disrupting compounds (EDCs) encompass a variety of substance classes, including natural and synthetic hormones, plant constituents, pesticides, substances used in industry and in consumer products, pollutant. It is well documented that EDCs targets are mainly the nuclear receptors (NRs) such as the sexual hormones estrogen receptors ERs. ER is involved in a broad spectrum of physiological processes in different organs and tissues as well as in several diseases, such as breast and endometrial cancer, osteoporosis, and prostate hypertrophy, neurodegenerative diseases or in immune system activation. The regulatory agencies for the protection of human health and wildlife have been issuing the necessity to investigate and clarify the mode of action of these exogenous substances by the development of alternative toxicological methods such as in silico model and in vitro testing in order to predict the toxicity of these EDCs. This research had the ambitious aim to develop an integrated toxicological strategy based on the combination of available information of endocrine disrupting activity retrieved from the scientific literature, in silico model, imaging methodologies applied to reporter systems in vitro and in vivo and ex vivo to predict among a set of chemicals those with an endocrine disrupting activity or ability to activate other toxicological pathways such as inflammation and oxidative stress measured in the male reproductive organs and in the genital and abdominal area of mice. The selected molecules range from known (DES) to suspected (BPA) endocrine disruptors and included both synthetic (DEHP) and natural (genistein) compounds. The first step used was in silico analysis with evaluation of the possible binding of selected substances to the estrogen alpha receptor to support the hypothesis that their hormonal activity occurred through a receptorial mechanism. This approach is commonly used and is also part of the first level of investigation suggested by EFSA/ECHA in the recognition of EDCs. The computational methodology estimated different values of affinity of each ligand to hER Ligand Binding Domain (LBD). The use of two different approaches (XP GLIDE SCORE and MMGBSA dG Bind) also allowed for solvation to be taken into account. That meant that MMGBSA protocol considered both the interactions of some water molecules with the LBD and the solvent-ligand ones. The estradiol showed the best affinity values in both approaches as being the endogenous hormone was able to contrast the solvation effect. The dissociation constant (Ki) values calculated from the XP GLIDE SCORE fitted well with the Ki experimentally determined in vitro binding assay by other research groups. Thus, the computed Ki has been chosen as parameter for the prediction of putative endocrine disruptor activity. However, the lack of correlation between the (computed and experimental) Ki and in vivo experimental observed ED activity (from literature data) for all chemicals (only estradiol and zearalenone have similar affinity to ER and literature evidences of ED activity; genistein has a good Ki values but not ED activity such as BPA and methoxychlor), did not allow a prioritization of the investigated chemicals for ED activity through the results of their affinity. In the next step, to check if the receptor binding well correlated with the hormonal activity, the ER Reporter gene assay was performed, based on the ability of a compound to stimulate ER-dependent transcriptional activity in genetically engineered mammalian cells. The cell lines are MCF-7 cells which express human endogenous ERα. The cells are transformed (transfected) by introducing vectors containing DNA sequences for the receptor, along with EREs linked to a reporter gene, and the reporter gene itself. The reporter gene used in human cancer cells usually codes for luciferase (CALUX, chemically activated luciferase expression). In the transactivation the EDCs show their estrogenic potency calculated as EC50, in respect to the positive control, 17β-estradiol. This system has enabled us to evaluate the kinetic and the biological consequences of cellular activation in the same cell monolayer by bioluminescence imaging of photon emissions that were pictures of activated ER status at 6 and 48 hours after the initial treatments. From this in vitro assay three factors were taken into consideration, power, efficacy and trend over time. The in vitro dynamic ER activation showed that for some chemicals (genistein, BPA, methoxychlor), the potency (EC50) and the efficacy (fold induction) changed over time, but not for others (estradiol, zearalenone and DES). Considering that estradiol, zearalenone and DES certainly have an activity in the animal and in man as endocrine disruptors, the duration of effect parameter combined with power and efficacy were likely to be associated in predicting the activity. Together with receptor affinity and the ability to induce a biological response, it also seemed relevant how long the response was lasting. By consequence, the combination of the variation of the potency response and the efficacy, “normalized” respect to the efficacy values quantify at 48 hours, was successfully used in discriminating positive and negative compounds for their endocrine disrupting activity. By means of this analysis 17β-estradiol, DES and zearalenone were put at the top of the list (also supported by their known ED activity), genistein resulted to represent a putative threshold of no-concern for ED effect, in supporting published data, while methoxychlor and BPA were definitely not considered a priority in terms of ED activity. This in vitro classification fitted well with the in silico outcomes, since the strongest estrogen receptor binders were ranked in the first positions (17β-estradiol, DES and zearalenone). Besides, was no possible to calculate EC50s for 4-nonylphenol, DEHP and vinclozolin, not making possible to classify them as ED, totally in agreement with in silico results and in line with literature data (vinclozolin is mainly an androgen antagonist). The third step of our stepwise approach was intended to verify in vivo the interaction of selected chemicals with the estrogen receptor and in addition the activation of their pathways triggering primary harmful effects. We used three reporter mice designed in order to evaluate the effect of selected compounds to activate ER and causing oxidative stress and inflammation. We have chosen to test zearalenone (well-known endocrine disruptor and clearly identified as such by our in silico-in vitro approach), and BPA for which there are controversial data in the literature and that our approach has negatively classified as ED. In our experiments zearalenone showed to be active on ER pathway in the abdominal area and significantly activated the inflammatory pathway in the genitals (in this specific case in the prostate, result of ex vivo bioluminescence analysis). These results were perfectly in line with the literature reports, in which prostate inflammation and metaphase were detected in both mice and rats. Bisphenol A did not produce a significant activation in both the areas and in the ex vivo analysis, again in agreement with in silico/in vitro results. We used a combination of innovative approaches that led to a conclusion that in silico screening cannot be used as a stand-alone procedure due to its intrinsic lack of biological meaning, although it can be successfully used as a first prioritizing step in a tier approach. The second mandatory check for the in silico positive hits should be an in vitro evaluation procedure, in which the affinity of the positive hits is measured through a reference cellular assay. Our results showed that integrating the time variable in the evaluation of the potency, the tested compounds could be classified as ED or no-ED. The in vivo experiment highlighted that a potent estrogenic compound, as zearalenone, could also raise concern for the activation of other toxicological pathway such as the inflammatory ones. We are aware that this indication of procedure must be evaluated and validated on dozens of molecules whose in vivo activity is already known before arriving at its use to predict the possible activity of ED of unknown molecules, but we think that this approach deserves to be implemented.