Identification of putative estrogen receptor-mediated endocrine disrupting chemicals using QSAR- and structure-based virtual screening approaches

Abstract

Identification of Endocrine Disrupting Chemicals is one of the important goals of environmental chemical hazard screening. We report on the development of validated in silico predictors of chemicals likely to cause Estrogen Receptor (ER)-mediated endocrine disruption to facilitate their prioritization for future screening. A database of relative binding affinity of a large number of ERα and/or ERβ ligands was assembled (546 for ERα and 137 for ERβ). Both single-task learning (STL) and multi-task learning (MTL) continuous Quantitative Structure-Activity Relationships (QSAR) models were developed for predicting ligand binding affinity to ERα or ERβ. High predictive accuracy was achieved for ERα binding affinity (MTL R2=0.71, STL R2=0.73). For ERβ binding affinity, MTL models were significantly more predictive (R2=0.53, p<0.05) than STL models. In addition, docking studies were performed on a set of ER agonists/antagonists (67 agonists and 39 antagonists for ERα, 48 agonists and 32 antagonists for ERβ, supplemented by putative decoys/non-binders) using the following ER structures (in complexes with respective ligands) retrieved from the Protein Data Bank: ERα agonist (PDB ID: 1L2I), ERα antagonist (PDB ID: 3DT3), ERβ agonist (PDB ID: 2NV7), ERβ antagonist (PDB ID: 1L2J). We found that all four ER conformations discriminated their corresponding ligands from presumed non-binders. Finally, both QSAR models and ER structures were employed in parallel to virtually screen several large libraries of environmental chemicals to derive a ligand- and structure-based prioritized list of putative estrogenic compounds to be used for in vitro and in vivo experimental validation.

Identifying potential endocrine disruptors among industrial chemicals and their metabolites – development and evaluation of in silico tools

“…This is in agreement with earlier studies showing that non-linear methods (especially k-NN which is implemented within the ASNN method) result in the highest prediction accuracy for E-, A-, and T-binding Gramatica, 2010b, 2010a;Papa et al, 2013). Obtained statistics for the EAT models with highest balanced accuracy corresponded or were slightly higher than for recently published models using similar data (Li and Gramatica, 2010b;Jensen et al, 2011;Papa et al, 2013;Zhang et al, 2013). The ASNN models are, however, difficult to interpret and thus as suggested by Vorberg and Tetko (2014), PLS models were used to increase our understanding on fundamental structural characteristics of EAT binders and non-binders.…”

“…Many of these models can be applied only to a limited number of compounds. Models with relatively wide applicability have been developed for predicting binding affinity to the estrogen Serafimova et al, 2007;Li and Gramatica, 2010b;Toropov et al, 2012;Yi and Zhang, 2012;Zhang et al, 2013) and androgen receptors Li et al, 2009Li et al, , 2013aLi and Gramatica, 2010a;Jensen et al, 2011;Todorov et al, 2011). To our knowledge, studies on the use of QSAR models for prioritizing potential endocrine disruptors considering multiple endpoints were limited to two mechanisms of action (Li and Gramatica, 2010b) or to one group of compounds sharing similar structural features (Juberg et al, 2013), for example brominated flame retardants (Kovarich et al, 2011).…”

Docking-Based Classification Models for Exploratory Toxicology Studies on High-Quality Estrogenic Experimental Data

“…All four possible ER classes were considered: ER-α bound to agonist, ER-α bound to antagonist, ER-β bound to agonist and ER-β bound to antagonist. For each of these, two crystal structures were retrieved from PDB: one previously considered by Zhang et al [28] for the identification of new EDs based on docking simulations, and the other recently selected by Kolšek et al [17] based on their capability to provide highly predictive docking-based classification models. On the other hand, the choice of eight different x-ray solved crystals (comprising ER-α and ER-β complexes for agonists and antagonists) was made to possibly cover a broader spectrum of possible biological actions of compounds comprised within the EPA training dataset.…”

Identification of structural requirements of estrogen receptor modulators using pharmacoinformatics techniques for application to estrogen therapy

“…The proper application of this technique can develop molecules with optimistic efficacy. Keeping in mind the objective of immense utility of SERMs for the treatment of post-menopausal diseases, researchers in academia as well as in industry are dedicated to the development of synthetic therapeutic SERMs for estrogen therapy (Brogia et al 2013;Chang et al 2013;Lewis et al 1995;Mukherjee et al 2005;Smith et al 2007;Zhang et al 2013). …”