LINCS dataset-based repositioning of rosiglitazone as a potential anti-human adenovirus drug

Abstract

Human adenoviruses (HAdVs) often cause mild respiratory infections. These infections, however, can potentially become fatal in immunosuppressive patients. Unfortunately, there has been no specific anti-HAdV drug approved for treatment of HAdV infections. In this study, a time-course transcriptome of HAdV-infected human lung epithelial cells (A549 cells) was performed and compared with perturbation datasets of 890 drug-treated A549 cells from the library of integrated network-based cellular signatures (LINCS) database to predict previously unknown therapeutic drug-HAdV relationships using a characteristic direction (CD) algorithm. We performed experiments to validate a prediction for the anti-diabetic drug rosiglitazone as a candidate drug for treatment of anti-HAdV both in vivo and in vitro. The Type I interferon (IFNs) signaling pathway was negatively regulated during the course of HAdV infection and rosiglitazone increased STAT1 phosphorylation for antiviral IFN response induction. Taken together, this study confirmed the prospect for re-exploitation of this FDA-approved drug as a potential therapeutic for HAdV infections.