Predicting candidate therapeutic drugs for sepsis-induced acute respiratory distress syndrome based on transcriptome profiling.

Sepsis-induced acute respiratory distress syndrome (ARDS) remains a major threat to human health without effective therapeutic drugs. Previous studies demonstrated the power of gene expression profiling to reveal pathological changes associated with sepsis-induced ARDS. However, there is still a lack of systematic data mining framework for identifying potential targets for treatment. In this study, we demonstrated the feasibility of druggable targets prediction based on gene expression data. Through the functional enrichment analysis of microarray-based expression profiles between sepsis-induced ARDS and non-sepsis ARDS samples, we revealed genes involved in anti-microbial infection immunity were significantly altered in sepsis-induced ARDS. Protein-protein interaction (PPI) network analysis highlighted TOP2A gene as the key regulator in the dysregulated gene network of sepsis-induced ARDS. We were also able to predict several therapeutic drug candidates for sepsis-induced ARDS using Connectivity Map (Cmap) database, among which doxorubicin was identified to interact with TOP2A with a high affinity similar to its endogenous ligand. Overall, our findings suggest that doxorubicin could be a potential therapeutic for sepsis-induced ARDS by targeting TOP2A, which requires further investigation and validation. The whole study relies on publicly available dataset and publicly accessible database or bioinformatic tools for data mining. Therefore, our study benchmarks a workflow for druggable target prediction which can be widely applicable in the search of targets in other pathological conditions.