Using a technique known as drug repositioning, researchers at the Institute of Biomedical Sciences at University of Sao Paulo (ICB-USP) found seven possible drugs to inhibit the replication of coronavirus. The products are already approved by the Food and Drug Administration (FDA, regulatory agency of the United States), which would facilitate the advancement to clinical trials if their effectiveness is proven in vitro, in tests performed with cell lines.
The results of the research – supported by FAPESP through three projects – were published in the Journal of Biomolecular Structure and Dynamics.
The investigation focused on the SARS-CoV-2 enzyme 3CLpro, considered essential for virus replication. Using machine learning techniques, the scientists tested more than 11,000 molecules and selected those that showed greater affinity with the target molecule, as well as greater stability within the active site – the region of the protein where the chemical reaction takes place.
“These computational predictions we made selected seven molecules that could be promising for testing in cells. If they work in vitro, we can see them being tested in humans. The advantage of testing drugs that already exist on the market is that the effects of toxicity and side effects are already widely known”, highlights Cristiane Guzzo, study coordinator. “Therefore, after validation in in vitro assays, it would be possible to carry out clinical tests in patients with covid-19.”
The finding may also be important in establishing the criteria and properties that the drug must have to inhibit the 3CLpro enzyme. “We’ve seen that the best compounds are those that interact favorably with the enzyme’s five specific amino acid residues. Therefore, these residues can be used to discover other inhibitors”, explains postdoctoral fellow Anacleto Silva de Souza, first author of the article.
Coronavirus 3CLpro is a protease responsible for breaking a chain of viral proteins (polyprotein) into its functional subunits, thus allowing RNA replication and assembly of new viral particles that will infect other cells. The hypothesis is that, by inhibiting 3CLpro, it cannot perform its function and the virus stops replicating and proliferating, reducing the viral load and the severity of the disease.
This is a different target from other studies, many of them focused on the spike protein, which is on the surface of the virus and is responsible for the interaction with the host cell. “We chose 3CLpro because there was already a considerable amount of information about it, due to more than 15 years of research with SARS-CoV’s 3CLpro [coronavírus que causou a epidemia de síndrome respiratória aguda grave entre 2002 e 2003]”, says Guzzo.
To simulate the interaction of drugs with the protease, the researchers developed three mathematical models, using artificial neural networks and two regression models. These quantitative models, known as QSAR, were based on information from the literature on molecules that were already known to have inhibitory properties against the 3CLpro enzyme.
With the QSAR models, an affinity prediction of 11,000 molecules was made using the database drugbank, a repository that combines detailed chemical and pharmacological data on approximately 4,300 drugs. Based on the computational results, 2,500 molecules were discarded for having low affinity with the enzyme. Of the 8,500 compounds that followed, 14 molecules with different pharmacological properties (drugs for the treatment of migraine, respiratory diseases, antimicrobial action and natural products) were selected for computational simulation between these drugs and the enzyme.
“To inhibit the enzyme, it is not enough just to have the predicted high affinity, it is necessary for the molecule to be able to remain attached to the target. Otherwise, the enzyme function is reactivated. We then simulated how these 14 molecules would behave within the active site of 3CLpro. We carry out a docking [acoplamento entre moléculas] followed by molecular dynamics to verify if the compound is stable inside the enzyme, and that leaves seven candidates”, the researchers explain.
Scientists now intend to confirm these predictions in biochemical experiments, through cloning of 3CLpro. If the discovery is validated, the expectation is to establish partnerships with other ICB laboratories to test the drugs in vitro.
This text was originally published by FAPESP Agency according to the Creative Commons license CC-BY-NC-ND.