Druggable Mendelian randomization prioritizes CDH2 and supports finerenone as a candidate therapeutic strategy for diabetic retinopathy.

Diabetic retinopathy remains a major cause of vision loss, and therapeutic strategies beyond anti-vascular endothelial growth factor treatment are still needed. This study aimed to identify genetically supported druggable targets for diabetic retinopathy and to evaluate finerenone as a candidate therapeutic intervention in experimental retinopathy. We performed druggable Mendelian randomization by integrating druggable-gene resources, blood cis-expression quantitative trait locus data, and a large genome-wide association dataset for diabetic retinopathy. Significant genes were further evaluated using colocalization analysis, functional enrichment, protein-protein interaction network analysis, drug prediction, and molecular docking. Finerenone was subsequently assessed in db/db mice and in the oxygen-induced retinopathy model. Thirty candidate druggable genes were associated with diabetic retinopathy after false discovery rate correction. Among them, CDH2 was the only candidate showing significant colocalization with diabetic retinopathy risk, with a posterior probability for a shared causal variant of 0.85. Protein-protein interaction analysis showed relatively high connectivity of CDH2 within the candidate network, and molecular docking suggested a favorable predicted interaction between finerenone and N-cadherin. In vivo, finerenone reduced avascular and neovascular areas in oxygen-induced retinopathy retinas and decreased retinal expression of TNF-α, IL-1β, and N-cadherin in db/db mice. These findings prioritize CDH2/N-cadherin as a genetically supported candidate target in diabetic retinopathy and support finerenone as a potential therapeutic candidate. The retinal protective effects of finerenone may be associated with suppression of inflammatory responses and downregulation of N-cadherin, although the direct mechanistic link requires further validation.