Molecular Modelling Prediction of Estrogen Mimicry and its Biological Consequences in Estrogen Receptor Positive Breast Cancer Cells

Abstract

Environmental estrogens are a diverse group of natural and synthetic compounds that can interact with estrogen receptors (ERs) in animals and humans. Interactions with ERs relies on key structural features of estrogen molecules and if these similarities are found on other molecules, they are likely to have estrogen mimicry and elicit a response. The structural diversity of these 17β-estradiol (E2) mimics makes it difficult to predict whether they will act as E2 mimics and cannot be identified by structure alone. This thesis explores the structure activity relationship of environmentally prevalent E2 mimics and the implications of exposure to these compounds.

E2 mimics were selected for this study based on their environmental prevalence. They were subject to MELN assays (a gene reporter assay based on expression of ERs) and MCF-7 (cultured ER positive breast cancer cells) proliferation studies. Studies were carried out with E2 and E2 mimics to investigate the structure activity relationship between estrogenicity values (e.g., EC50 values) and molecular modelling DockScores. This work further supports published literature and confirms the ER- driven bioactivity and DockScores have a linear relationship (R2 = 0.75, MELN assay; R2 = 0.81, MCF-7 cell proliferation studies).

In silico molecular modelling studies (Schrödinger platform) highlight the complex structural architecture of the ER and the interactions between ligands and the LBC; this indicates the potential effects of common structural features of ligands (e.g., isoflavone backbone, synthetic estrogens, parabens) on their binding energies with ERs and their estrogenicity – resulting in differing ER-driven bioactivities. Further studies using molecular dynamics focusing on ERβ investigates the protein unfolding from the agonist conformations bound with genistein and E2, and the antagonist conformation bound with 4-hydroxytamoxifen. The ligands were removed before the simulation, resulting in new conformational freedom due to absence of van der Waals contacts between the ligands and the receptor, showing that van der Waals forces are crucial to hold the ligand binding domain in either the agonist or antagonist conformation.

The interactions of E2 and E2 mimics with ERs are well documented, however, very few studies have investigated the interactions and effects these compounds have on plasma-steroid binding proteins. Sex hormone binding globulin (SHBG) is a glycoprotein and is the main, specific transport system for endogenous sex steroid hormones, the other being albumin. E2 mimics can interact with SHBG displacing endogenous sex hormones (e.g., E2, testosterone) and could lead to modulating their bioavailability. In this study a sandwich SHBG ELISA assay used to determine SHBG in plasma was utilised to investigate the effect E2 and select E2 mimics had on levels of SHBG levels in MCF-7 cell cultures. This could have implications on human exposure to E2 mimics if high enough concentrations in the blood are reached.

Further research will increase our understanding of ER/ligand interactions, and our ability to predict their biological activity and consequent human health risks. Linking artificial intelligence with biological responses is the next step forward.