Wird KI neue Medikamente gegen Herzkrankheiten hervorbringen?

Zusammenfassung

Angesichts der umwälzenden Auswirkungen, die künstliche Intelligenz (KI) auf Wissenschaft, Medizin und darüber hinaus hat, betrachten wir hier das Potenzial von KI für die Entdeckung neuer Medikamente gegen Herzkrankheiten. Wir definieren KI im weitesten Sinne als den Einsatz von maschinellem Lernen, einschließlich Statistik und Deep Learning, um Muster in Datensätzen zu erkennen, die für Vorhersagen genutzt werden können. Jüngste Durchbrüche in der Fähigkeit, sehr große Datenmengen zu berücksichtigen, haben einen Boom in der KI-gestützten Arzneimittelentdeckung sowohl in der Wissenschaft als auch in der Industrie ausgelöst. Viele neue Unternehmen verfügen bereits über Arzneimittel-Pipelines, die bis in die klinische Erprobung reichen, aber noch keine Medikamente gegen Herzkrankheiten enthalten. Wir beschreiben hier den Einsatz von KI für die Entdeckung von niedermolekularen Medikamenten und Biologika, einschließlich therapeutischer Peptide, sowie für die Vorhersage von Wirkungen wie Kardiotoxizität. Der konzertierte Einsatz von KI zusammen mit physikbasierten Simulationen und experimentellen Rückkopplungsschleifen wird notwendig sein, um das Potenzial der KI für die Arzneimittelentdeckung und die Entwicklung von Präzisionsarzneimitteln für Herzkrankheiten voll auszuschöpfen.

Abstract

Given the transformational impact that artificial intelligence (AI) is having on science, medicine and beyond, we here consider the potential of AI for the discovery of new medicines against heart disease. We define AI broadly as the use of machine learning, including statistics and deep learning, to identify patterns in datasets that can be used to make predictions. Recent breakthroughs in the ability to consider very large amounts of data have spurred a boom in AI-driven drug discovery in both academia and industry. Many new companies already have drug pipelines extending into clinical trials, but these do not yet include drugs against heart disease. We here describe the use of AI for the discovery of low-molecular weight drugs and biologics, including therapeutic peptides, as well as for predicting effects such as cardiotoxicity. The concerted use of AI together with physics-based simulations and experimental feedback loops will be necessary to fully realize the potential of AI for drug discovery and the development of precision medicines for cardiac conditions.

Schlüsselwörter

maschinelles Lernen - molekulare Simulation - Wirkstoffdesign - Arzneimittelentdeckung - Arzneimittelentwicklung - künstliche Intelligenz (KI)

Keywords

machine learning - molecular simulation - drug design - drug discovery - drug development - artificial intelligence (AI)

• Literatur

• 1 Savarese G, Becher PM, Lund LH. et al. Global burden of heart failure: a comprehensive and updated review of epidemiology. Cardiovasc Res 2023; 118: 3272-3287
• 2 Braunwald E. SGLT2 inhibitors: the statins of the 21st century. Eur Heart J 2022; 43: 1029-1030
• 3 Sun D, Gao W, Hu H. et al. Why 90% of clinical drug development fails and how to improve it?. Acta Pharm Sin B 2022; 12: 3049-3062
• 4 Dowden H, Munro J. Trends in clinical success rates and therapeutic focus. Nat Rev Drug Discov 2019; 18: 495-496
• 5 McCarthy J. What is artificial intelligence?. 12.11.2017 Zugriff am 13. Oktober 2023 unter: https://www-formal.stanford.edu/jmc/whatisai.pdf
• 6 LeCun Y, Bengio Y, Hinton G. Deep learning. Nature 2015; 521: 436-444
• 7 Lo YC, Rensi SE, Torng W. et al. Machine learning in chemoinformatics and drug discovery. Drug Discov Today 2018; 23: 1538-1546
• 8 Muratov EN, Bajorath J, Sheridan RP. et al. QSAR without borders. Chem Soc Rev 2020; 49: 3525-3564
• 9 Vanhaelen Q, Lin YC, Zhavoronkov A. The Advent of Generative Chemistry. ACS Med Chem Lett 2020; 11: 1496-1505
• 10 Khalak Y, Tresadern G, Hahn DF. et al. Chemical Space Exploration with Active Learning and Alchemical Free Energies. J Chem Theory Comput 2022; 18: 6259-6270
• 11 Pun FW, Liu BHM, Long X. et al. Identification of Therapeutic Targets for Amyotrophic Lateral Sclerosis Using PandaOmics – An AI-Enabled Biological Target Discovery Platform. Front Aging Neurosci 2022; 14: 914017
• 12 Prasad K, Kumar V. Artificial intelligence-driven drug repurposing and structural biology for SARS-CoV-2. Curr Res Pharmacol Drug Discov 2021; 2: 100042
• 13 Yang Y, Yao K, Repasky MP. et al. Efficient Exploration of Chemical Space with Docking and Deep Learning. J Chem Theory Comput 2021; 17: 7106-7119
• 14 Beuming T, Martin H, Diaz-Rovira AM. et al. Are Deep Learning Structural Models Sufficiently Accurate for Free-Energy Calculations? Application of FEP+ to AlphaFold2-Predicted Structures. J Chem Inf Model 2022; 62: 4351-4360
• 15 Yuan JH, Han SB, Richter S. et al. Druggability Assessment in TRAPP Using Machine Learning Approaches. J Chem Inf Model 2020; 60: 1685-1699
• 16 Lemke T, Peter C. EncoderMap: Dimensionality Reduction and Generation of Molecule Conformations. J Chem Theory Comput 2019; 15: 1209-1215
• 17 Bernetti M, Bertazzo M, Masetti M. Data-Driven Molecular Dynamics: A Multifaceted Challenge. Pharmaceuticals (Basel) 2020; 13: 253
• 18 Kokh DB, Amaral M, Bomke J. et al. Estimation of Drug-Target Residence Times by tau-Random Acceleration Molecular Dynamics Simulations. J Chem Theory Comput 2018; 14: 3859-3869
• 19 Noe F, Tkatchenko A, Muller KR. et al. Machine Learning for Molecular Simulation. Annu Rev Phys Chem 2020; 71: 361-390
• 20 Jayatunga MKP, Xie W, Ruder L. et al. AI in small-molecule drug discovery: a coming wave?. Nat Rev Drug Discov 2022; 21: 175-176
• 21 Madani A, Krause B, Greene ER. et al. Large language models generate functional protein sequences across diverse families. Nat Biotechnol 2023; 41: 1099-1106
• 22 Hie BL, Shanker VR, Xu D. et al. Efficient evolution of human antibodies from general protein language models. Nat Biotechnol 2023;
• 23 Hie BL, Yang KK, Kim PS. Evolutionary velocity with protein language models predicts evolutionary dynamics of diverse proteins. Cell Syst 2022; 13: 274-285.e6
• 24 Olsen TH, Moal IH, Deane CM. AbLang: an antibody language model for completing antibody sequences. Bioinform Adv 2022; 2: vbac046
• 25 Mork SK, Kadivar M, Bol KF. et al. Personalized therapy with peptide-based neoantigen vaccine (EVX-01) including a novel adjuvant, CAF(R)09b, in patients with metastatic melanoma. Oncoimmunology 2022; 11: 2023255
• 26 Glaser M, Bruce NJ, Han SB. et al. Simulation of the Positive Inotropic Peptide S100A1ct in Aqueous Environment by Gaussian Accelerated Molecular Dynamics. J Phys Chem B 2021; 125: 4654-4666
• 27 Kehr D, Ritterhoff J, Glaser M. et al. S100A1ct: a synthetic peptide derived from human S100A1 protein improves cardiac contractile performance and survival in pre-clinical heart failure models. bioRxiv 2023;
• 28 Iftkhar S, de Sa AGC, Velloso JPL. et al. cardioToxCSM: A Web Server for Predicting Cardiotoxicity of Small Molecules. J Chem Inf Model 2022; 62: 4827-4836
• 29 Creanza TM, Delre P, Ancona N. et al. Structure-Based Prediction of hERG-Related Cardiotoxicity: A Benchmark Study. J Chem Inf Model 2021; 61: 4758-4770
• 30 Cai C, Guo P, Zhou Y. et al. Deep Learning-Based Prediction of Drug-Induced Cardiotoxicity. J Chem Inf Model 2019; 59: 1073-1084