회선신경망 기반 SARS-CoV-2 (COVID-19) 치료제 약물 재창출 및 세포 실험 결과로서의 Peroxisome Proliferator-Activated Receptors-Gamma Agonist의 효과

Abstract

Background: Acute respiratory distress syndrome resulting from coronavirus (COVID-19) infection is triggered by cytokine storms, so activation of inhibitory modulators of inflammatory pathways has become a new candidate modality for COVID-19 treatment. This study utilized artificial intelligence (A.I.) to search databases, and compiled a list of 50 drugs deemed plausible candidates for COVID-19 treatment. We then designed a cellbased in vitro assay to evaluate the efficacy of PPAR-γ agonists against viral induced inflammation. Methods: We applied RNN screening to Drugbank and CORD-19 databases, and selected as the top 50 drug candidates those compounds that have the highest docking energy with the main protease produced by SARS-CoV-2 infected cells. We then designed an in vitro study including chloroquine, lopinavir, and remdesivir treated cells as controls, and cells treated with two PPAR-γ agonists as experimental groups. SARS-CoV-2 infected cells were administered a range of concentrations of each drug, and inhibition-normalized infection ratios were derived using an immunofluorescence method. Results: The positive control groups’ SI’s were >1 (chloroquine SI=9.28, remdesivir SI=4.56, lopinavir SI=3.5), confirming their inhibitory effects against SARS-CoV-2 infection. However, chloroquine and lopinavir displayed high cytotoxicity, and Remdesivir displayed low cytotoxicity. The two PPAR-γ agonist SIs indicated that they possess no inhibitory effect against SARS-CoV-2 infection, but are clinically safe. Conclusion: The PPAR-γ agonists did not reduce numbers of SARS-CoV-2 infected cells. Nevertheless, this study has significance in that we introduced the use of A.I. for rapid new drug development during the COVID pandemic.