Cross-Domain Text Mining of Pathophysiological Processes Associated with Diabetic Kidney Disease

Abstract

Diabetic kidney disease (DKD) remains a significant burden on the healthcare system and is the leading cause of end-stage renal disease worldwide. The pathophysiology of DKD is multifactorial and characterized by various early signs of metabolic impairment, inflammatory biomarkers, and complex pathways that lead to progressive kidney damage. New treatment prospects rely on a comprehensive understanding of disease pathology. The study aimed to identify signaling drivers and pathways that modulate glomerular endothelial dysfunction in DKD via cross-domain text mining with SemNet 2.0. The open-source literature-based discovery approach, SemNet 2.0, leverages the power of text mining 33+ million PubMed articles to provide integrative insight into multiscalar and multifactorial pathophysiology. A set of identified relevant genes and proteins that regulate different pathological events associated with DKD were analyzed and ranked using normalized mean HeteSim scores. High-ranking genes and proteins intersecting three domains--DKD, immune response, and glomerular endothelial cells--were analyzed. The top 10% of ranked concepts mapped to the following biological functions: angiotensin, apoptosis, cell-cell function, cell adhesion, chemotaxis, growth factor signaling, vascular permeability, nitric oxide response, oxidative stress, cytokine response, macrophage signaling, NF{kappa}B factor activity, TLR signaling, glucose metabolism, inflammatory response, ERK/MAPK signaling, JAK/STAT signaling, T-cell mediated response, WNT signaling, renin angiotensin system, and NADPH response. High-ranking genes and proteins were used to generate a protein-protein interaction network. This comprehensive analysis identified testable hypotheses for interactions or molecules involved with dysregulated signaling in DKD, which can be further studied through biochemical network models.