A Food Contact Materials 3D database for food safety using nuclear receptors (ER and AR) as biological cancer target

Abstract

Food contact chemicals (FCCs) include a very heterogeneous class of molecules and the term refers to all chemicals that come into contact with food during the entire food chain, from the raw material to the final products. FCCs mainly derived from three different sources: natural, intentional, and unintentional. Microorganisms, plants, and animals, which directly interact with food, can produce and release molecules as a natural defence mechanism. These molecules may have a toxic effect on other organisms, including humans. Mycotoxins are a typical example of food contact chemicals that naturally occur in food. Two other important sources should be considered: the molecules that are intentionally added to food and those unintentionally present. To the first class belongs all substances which are intentionally added to enhance some food properties, and to ensure and maintain food safety, like preservatives. Instead, all food contact chemicals that accidentally come into contact with food, as contaminant molecules belong to the second class. They can occur at different levels of the food chain, e.g., processing, storing, packaging, or consumption. It is often difficult to evaluate the total extent of exposure as well as their safety for humans and the environment because information on food contact chemicals is widespread across multiple online sources, which have also a certain degree of errors. Moreover, these errors propagate very easily across the internet. Therefore, the data are not standardized, and this issue limited the analyses to evaluate the safety of these compounds. To make matters worse, food contact chemicals include thousands of chemicals which make it difficult to test their safety using experimental methods. They would require a long time and high-costs to be performed. In fact, a lot of chemicals are waiting for their safety evaluation. Based on these considerations, this Doctoral Thesis aims at developing a database of all food contact chemicals (foodchem DB) and at screening their endocrine-disrupting properties using computational methods. This may overcome experimental limits (cost and time) in order to prioritize the molecules that require further investigations. In the first part of this PhD thesis, some specific case studies have been considered. First, the class of bisphenols has been evaluated for the estrogenic and androgenic pathways. Second, some pesticides have been evaluated considering an animal model, Daphnia Magna, with the aim to show the utility of computational methods to also reduce experimental tests. Once the benefit of using computational methods to identify endocrine disruptor chemicals has been evaluated, in the second part of this PhD thesis, it has been considered a huge number of molecules. Firstly, it has been developed a food contact chemical database to take into account all molecules that come into contact naturally, intentionally, or unintentionally with food. The foodchem DB lists 11059 substances divided into ten subclasses: additive, bisphenol, dioxin, flavouring, furan, mycotoxin, PCB, pesticide, phthalate, and the food packaging forum (FPF) database lists (FCCdb). To pursue the objective to include all these molecules, regulatory lists, industry inventories, and public data sources have been considered. Moreover, in this section, it is also addressed the evaluation of the endocrine-disrupting properties undertaken by food contact chemicals using computational methods. The importance of the consensus scoring approach to overcome molecular docking limitations has also been illustrated. In this study, several nuclear receptors have been considered to evaluate a general endocrine disruptor effect produced by all food contact chemicals. However, to consider specifically the involvement of these molecules in breast and prostate cancers, in Chapter 8, a little focus has been made on estrogen and androgen receptors since they are two well-known nuclear receptors involved in these cancers. The combination of molecular docking and consensus scoring techniques used in these studies highlighted how these methods are useful to study the interaction between food contact chemicals and nuclear receptors, allowing the identification of the chemicals which have most the suitable physical-chemical characteristic to interact and disrupt the endocrine pathway. On the other side, the effect of that interaction on the receptor (from a structural point of view) can be analysed using molecular dynamic simulation. Molecules can affect nuclear receptors activity by acting as agonist or antagonist compounds. However, it is not fully understood yet how molecules induce an antagonistic effect on the androgen receptor. Thus, to study the mechanism of action (MoA) of antiandrogens on the homodimer stability of the androgen receptor, in Chapter 9, molecular dynamics simulation has been carried out. This study might be useful to decipher food contact chemical effects on the androgen receptor using computational methods. Finally, in the last chapter, two future perspectives have been illustrated. In study one, it is highlighted the advantage to use the isothermal titration calorimetry (ITC) to study protein-ligand interaction for hazard identification. In study 2, it is shown the use of two computational methods (machine learning and robust consensus scoring) to identify the substances of very high concern for the endocrine system among food contact chemicals.