In our research part of the doctoral dissertation we focused on the search for potential endocrine disruptors and other toxic properties of the ingredients used in cosmetic products, and the possible interactions of these properties. In the first part of the doctoral dissertation we tested the hypotheses whether the standard and three-dimensional QSAR models, which are freely accessible on the internet, predict toxic properties of a selected series of compounds. We studied randomly selected compounds from the CosIng (Cosmetic Ingredients) database. With QSAR methodology, implemented in the program package CAESAR, we predicted toxic properties for mutagenicity, carcinogenicity and developmental toxicity of 558 compounds. The obtained predictions were subsequently analyzed by a well-established chemometric method of Kohonen self-organizing artificial neural networks. With this method we classified compounds in clusters, according to similarity, and determined the similarity indicators of individual clusters. These results give us a general overview of the data processed, and the way of systematic processing of a much larger data set. In the following studies, we upgraded the predictions for a selected set of compounds to three-dimensional QSAR models software package Endocrine Disruptome, which is based on molecular docking method. This online tool was used to predict binding affinity of 558 selected compounds on the binding domain of 12 nuclear receptors with a known crystal structure. With binding of these compounds to the receptor a receptor response occurs, which is reflected in the malfunction of the endocrine hormonal system, and in such a way that the tool could be used to identify potential endocrine disruptors. The results gave us predictions that at least one fifth of the studied compounds is capable of acting on endocrine hormone systems, since they show the binding affinity on at least one of the 12 nuclear receptors. The most common target among these receptors are estrogen, mineralocorticoid, glucocorticoid, thyroid, and androgen receptors. In the second part of doctoral dissertation we tried to verify the results of molecular docking with experimental methods. We made in vitro screening tests on MDA-KB2 cell line, which has been recombinantly modified with luciferase reporter plasmid, and it expresses both, glucocorticoid and androgen receptor, and screening tests on the GH3.TRE-Luc cell line with expressed thyroid receptor. Of 558 randomly selected compounds from the CosIng database we excluded 10 compounds, of which predictions are pointing to the ability of binding to multiple receptors at the same time, have a high free binding energy, a high value of logP, and have a greater ability to pass through the cell membrane to the binding sites of nuclear receptors. Nine of these ten compounds have the ability to bind to at least one of the selected receptors. In the final phase we compared all the obtained results and confirmed the predictive method of molecular docking, and outlined the parallels between endocrine disruptors and toxic properties of mutagenicity, carcinogenicity and developmental toxicity.
|
