Aromatase inhibitors in breast cancer: the discovery of new computational design and biochemical evaluation

Abstract

Continuous exposure to high levels of endogenous estrogens is associated with increased risks of developing breast cancer. In this sense, aromatase, the cytochrome P450 enzyme involved in the conversion of androgens, testosterone and androstenedione, into estrogens, estradiol and estrone, is an important target for the endocrine treatment of breast cancer in postmenopausal women. Aromatase inhibition is achieved either with compounds structurally related to the androstenedione substrate or with non-steroid inhibitors. Moreover, increasing evidence suggests that compounds mimicking the products of catalysis, such as endogenous estrogens as well as natural polyphenols, are able to bind into the aromatase active site as competitive inhibitors. Computer-assisted drug design refers to the application of informatics on the discovery and optimization of biologically active compounds. In this work we took advantage of several molecular modelling tools, combined with a fast and accurate biochemical evaluation assay, to the discovery and rational at an atomic level, of the anti-aromatase properties of new molecules. The first objective of this study was to explore the structure-activity relationships of emerging new classes of aromatase inhibitors such as estrogens and their endogenous metabolites, and natural polyphenols. The compounds were tested on a biochemical assay with aromatase extracted from human term placenta and their activities compared to that of reference compounds. Catechol estrogens as well as polyphenols from several natural sources and variable degree of functionalization, including flavones, flavanones, resveratrol and oleuropein, were found to be strong aromatase inhibitors. The physicochemical determinants for their productive binding to the active site of the enzyme were characterized through a combination of molecular modelling techniques, such as electrostatic surface potential calculations, three-dimensional quantitative structure-activity relationships, molecular interaction field mapping at the active site of a homology model of the enzyme, and docking experiments. A virtual receptor site for the binding of natural polyphenols was proposed and new structure-activity rules, useful to the rational design of synthetic derivatives were derived. Moreover, the involvement of catechol estrogens in a control mechanism of estrogen production was discussed, along with the clinical implications to breast cancer development in postmenopausal women. The results of these preliminary studies motivated a ligand-based virtual screening strategy for new potent compounds. In the absence of a high resolution X-ray structure of the enzyme, the vast information on strong aromatase inhibitors reported in the literature, combined with some of the most active polyphenols identified in this study, provided a good starting point. Three pharmacophore models were built based on the common physicochemical features of distinct sets of aromatase inhibitors, namely, azole non-steroid, steroid and polyphenol compounds. Molecular shape and information about their active site binding mode was also included in the models. Thirty six promising compounds were extracted from the large National Cancer Institute database and tested experimentally. New potent aromatase inhibitors were identified, active at very low nanomolar concentrations, comparable to second and third generation reference compounds. These molecules are competitive inhibitors and one of them was identified as a mechanism-based inactivator. In summary, the results of this work provided new data on the role of catechol estrogens in breast cancer and identified new potent aromatase inhibitors, highlighting the importance of a combined use of computational studies and and biochemical assays for the discovery of new aromatase inhibitors.