Joined Workshop between the Institute of Crystallography-CNR (IC-CNR) and the Maastricht MultiModal Molecular Imaging Institute (M4i), Maastricht University, The Netherlands

Machine learning-based prediction of hERG-mediated cardiotoxicity: a structure-based investigation

7 Mar 2023, 13:15

30m

Webinar

Session

Speaker

Dr Giuseppe Felice Mangiatordi (IC-CNR (Bari))

Description

Giuseppe Felice Mangiatordi(1) Pietro Delre(1,2) Teresa Maria Creanza(3), Nicola Ancona,(3) Giovanni Lentini,(4) Michele Saviano(1).

1. Istituto di Cristallografia, Consiglio Nazionale delle Ricerche, Via G. Amendola 122/O, 70126 Bari.
2. Chemistry Department, University of Bari “Aldo Moro”, via E. Orabona, 4, I-70125 Bari, Italy.
3. CNR - Institute of Intelligent Industrial Technologies and Systems for Advanced Manufacturing, Via Amendola 122/o, 70126 Bari, Italy.
4. Dipartimento di Farmacia—Scienze del Farmaco, Università̀ degli Studi di Bari “Aldo Moro”, via E. Orabona, 4, I-70125 Bari, Italy.

Prioritizing drug candidates based on their human ether-à-go-go-related gene potassium channel (hERG) blocking potential is a mandatory step in the early preclinical stage of a drug discovery program. The hERG blockade is, in fact, considered the main cause of cardiotoxicity in post-marketing surveillance. Several ligand-based approaches were therefore developed in the last years and are currently employed in a drug discovery process for in silico cardiac safety assessment of drug candidates. Herein, the first structure-based classifiers able to discern hERG binders from non-binders will be presented.(1) LASSO regularized Support Vector Machines were applied to integrate docking scores and protein-ligand interaction fingerprints. 396 models were trained and validated based on: i) high-quality experimental bioactivity information returned by 8,337 curated compounds extracted from ChEMBL (version 25(2)) and ii) structural predictor data. Molecular docking simulations were performed by using GLIDE and GOLD software programs and different hERG structural models, namely the recently published structures obtained by cryo-electron microscopy (PDB codes: 5VA1(3 )and 7CN1(4)) and two published homology models selected for comparison. Remarkably, some models return performances comparable to ligand-based classifiers in terms of accuracy (AUCMAX = 0.86±0.01) and negative predictive values (NPVMAX = 0.81±0.01) thus putting forward the herein developed computational workflow as a valuable tool for predicting hERG mediated cardiotoxicity without the limitations of ligand-based models, typically affected by low interpretability and a limited applicability domain. The study highlights the importance of using hERG structural models accounting for ligand-induced fit effects and allowed us to select the best performing protein conformation to be employed for a reliable structure-based prediction of hERG-related cardiotoxicity.

References
1. Creanza, T.M.; Delre, P.; Ancona, N.; Lentini, G.; Saviano, M.; Mangiatordi, G.F. Structure-Based Prediction of hERG Related Cardiotoxicity: A Benchmark Study. Journal of Chemical Information and Modeling 2021, 61 (9), 4758-4770
2. Davies, M.; Nowotka, M.; Papadatos, G.; Dedman, N.; Gaulton, A.; Atkinson, F.; Bellis, L.; Overington, J. P. ChEMBL Web Services: Streamlining Access to Drug Discovery Data and Utilities. Nucleic Acids Res 2015, 43 (W1), W612–W620
3. Wang, W.; MacKinnon, R. Cryo-EM Structure of the Open Human Ether-à-Go-Go-Related K+ Channel HERG. Cell 2017, 169 (3), 422-430
4. Asai, T.; Adachi, N.; Moriya, T.; Oki, H.; Maru, T.; Kawasaki, M.; Suzuki, K.; Chen, S.; Ishii, R.; Yonemori, K.; Igaki, S.; Yasuda, S.; Ogasawara, S.; Senda, T.; Murata, T. Cryo-EM Structure of K+-Bound HERG Channel Complexed with the Blocker Astemizole. Structure 2021, 29 (3), 203-212.