In a recent study published in Communication about natureresearchers used artificial intelligence (AI)-driven virtual control to discover ‘F10’, a novel heart-specific myosin inhibitor for potential heart disease and heart failure treatments.
Study: Discovery of a novel cardiac myosin modulator using artificial intelligence-based virtual control. Image credit: PopTika/Shutterstock.com
Record
The activation of cardiac myofibrils by Ca2+ subject to contraction-relaxation cycles, during which power shocks generated by the attachment and detachment of myosin heads from thick filaments to thin actin filaments lead to muscle movement, or the availability of myosin during this process is complicated by its interaction with proteins such as titin and cardiac myosin binding protein-C (cMyBP-C).
Myosin can also form a “super relaxed state” (SRX) that minimizes adenosine triphosphatase (ATPase) activity for better energy efficiency.
Dysfunction of this structure could lead to the development of cardiovascular disease and suggests an opportunity to design a new treatment for heart failure that targets the functions of dysmyosin.
Unlike traditional treatments that focus on symptoms, myosin modulators directly address the underlying causes, potentially with fewer side effects.
Further research is needed to deepen the understanding of the complex interplay between the structural and functional states of cardiac myosin and to optimize novel myosin modulators such as ‘F10’ for more effective and minimized side effects treatments of heart disease and heart failure.
About the study
In this study, researchers used an AI-based virtual high-throughput screening (VHTS) method using Atomwise’s AtomNet® platform to screen a curated library of more than 4 million small molecules.
Human beta-cardiac myosin was the primary focus of these molecules to test their efficacy towards the Omecamtiv Mercarbil binding site.
The top 200 selected molecules from this huge collection satisfied Lipinski’s Rule of Five and were thus selected as drug-like substances. Myosin modulators were identified by testing these selected compounds using a biochemical assay.
Biochemical assays were performed using bovine S1 cardiac myosin and rabbit skeletal F-actin. These test compounds were mixed separately with an enzyme mixture of lactate dehydrogenase, bovine cardiac myosin S1, and pyruvate kinase in a plain black half surface 96-well plate.
Similarly, the assay plates had a negative control (dimethyl sulfoxide (DMSO) only) and a positive control (blebbistatin). A substrate mixture was used to initiate these reactions and their extent was determined through NADH intensity measurements at different time points.
This process allowed the identification of compounds that modulate cardiac myosin ATPase activity.
In addition, demembranated myofibrils prepared from bovine ventricles were further used to evaluate the ATPase activity of the selected compounds. These myofibrils were tested in an identical assay setup, providing information on the effects of the compounds on myofibrillar ATPase activity at steady state.
This integrated approach, combining AI screening with biochemical validation, allowed researchers to efficiently identify novel cardiac myosin regulators with potential therapeutic applications.
Study results
In the present study, the researchers used artificial intelligence to screen a virtual library of approximately four million compounds for potential cardiac myosin regulators. This method developed a new compound called F10, significantly inhibiting ATPase activity in cardiac myosin.
F10 was selected after analysis of its interaction with the Omecamtiv Mecarbil binding site on human β-cardiac myosin, taking into account factors such as hydrogen bond donors and acceptors and hydrophobic characteristics.
However, additional evaluation revealed that 10 μmol L−1 of F10 inhibited ATPase activity in bovine cardiac myosins by approximately 44%. Dose-response analysis showed that it was effective at 21 μmol L−1(IC50).
This compound did not resemble known myosin effectors and appeared to be a novel chemical scaffold. Intriguingly, F10 significantly reduced the maximal rate of ATP hydrolysis without affecting the affinity of myosin S1 for F-actin.
This specificity was further highlighted by the differential effect of F10 on ATPase activity in various myosin isoforms in different muscle types, underscoring its cardiac myosin specificity.
The present study also investigated the mechanism of F10 inhibition. In this case, single nucleotide recycling experiments showed that the release of nucleotides from cardiac myosin was slowed due to the effect produced by F10 in stabilizing the SRX state of myosin.
The constitutive effect of F10 on demembranated rat ventricular trabeculae, which reduced maximal active isometric tension and altered the orientation of myosin heads, suggested that these proteins were stabilized in the OFF state.
In addition, they observed that F10 reduced the left ventricular systolic pressure of Langendorff-perfused rat hearts almost immediately, but caused no change in heart rate or coronary perfusion. Of note, the effects of F10 were reversible and faster in onset and offset compared to Mavacamten, another myosin inhibitor.
The structure-activity relationship analysis of the study provided insight into the binding and inhibition mechanism of F10. Computational docking suggested multiple potential interactions of F10 within the myosin motor domain.
Furthermore, variations in the chemical structure of F10 led to differences in inhibitory activity, reinforcing the idea of the OM binding site as a target for the development of myosin regulators.
This research is a testament to the potential of artificial intelligence in drug discovery, especially for cardiac myosin regulators.
The findings introduce F10 as a novel inhibitor of cardiac myosin and pave the way for the development of novel therapeutic agents targeting cardiac myosin for the treatment of heart disease.
The study highlights the utility of artificial intelligence in identifying new compounds and provides a framework for future exploration in this area.
