Urinary incontinence (UI) is a highly prevalent condition characterized by the involuntary loss of urine, which profoundly impairs physical, psychological, and social well-being. Although current evidence indicates that detrusor abnormalities and systemic pathophysiological factors contribute to the risk of UI, the exact causal roles of specific plasma proteins in its etiology remain unknown due to the inherent limitations of observational designs, such as unmeasured confounding and reverse causation. Therefore, this study aimed to systematically investigate the causal associations of plasma proteins with the risk of UI using a comprehensive two-sample Mendelian randomization (MR) framework and to prioritize high-confidence, translatable therapeutic drug targets through multi-dimensional evidence integration.

We conducted a two-sample MR study utilizing large-scale plasma protein quantitative trait loci (pQTL) data as exposures and UI genome-wide association study (GWAS) data as the outcome. To systematically evaluate and prioritize the identified causal proteins, we established a 6-tier scoring framework encompassing: colocalization test, Steiger filtering, replication using independent Fenland cohort pQTL data, functional annotation, protein-protein interaction (PPI) network analysis, and druggability assessment (Figure 1A). To assess biological relevance, proteins were mapped to Gene Ontology (GO) databases. The PPI network of the MR-significant genes was created using data from the StringDB database, and the core genes were estimated using the Cytoscape. Finally, the translational potential of these proteins was evaluated by cross-referencing with the DGIdb, DrugBank, and Pharmaprojects databases to identify established drug-gene interactions.

The primary MR analysis identified 42 candidate plasma proteins significantly associated with UI risk (Figure 2A). In the 6-tier evaluation, colocalization analysis (PP.H4 > 0.3 or PP.H4/[PP.H3+PP.H4] > 0.8) confirmed that all 42 proteins shared causal genetic variants with UI (Figure 2B), and Steiger filtering validated the correct causal direction for 32 proteins (Figure 2C). Serving as the most stringent filter in our framework, external replication using the independent Fenland GWAS successfully isolated one top-tier candidate with exceptional cross-cohort robustness. Functional annotation demonstrated that all 42 proteins mapped to explicit biological categories including NABA matrisome associated, negative regulation of response to external stimulus, regulation of nervous system development, blood vessel development and other functions (Figure 2D), and 21 core genes were estimated through PPI network analysis. Druggability assessment revealed that 8 MR-significant proteins are established drug targets, encompassing 11 distinct drug-gene interactions (Figure 2E). Through this comprehensive scoring system, KNG1 achieved the maximum score of 6, successfully passing all evaluation tests, and was identified as a robust protective factor for UI (OR = 0.37, 95% CI: 0.18-0.75, P = 0.006) (Figure 1B). Furthermore, four additional proteins achieved a high score of 5, including ANG, SELP as protective factors and HAVCR2, NRP1 as risk factors.

The integration of proteome-wide MR with systems biology and pharmacological databases effectively bridges the gap between genomic associations and clinical translation. The identification of KNG1, alongside four other highly scored proteins, highlights specific biological pathways—particularly extracellular matrix organization (NABA matrisome associated), regulation of nervous system development, and vascular remodeling (blood vessel development and endothelial cell migration) —that intrinsically drive UI pathogenesis. The presence of core genes within the PPI network and their validation against established pharmacological databases strongly suggests that these prioritized candidates possess substantial potential for successful therapeutic targeting or drug repurposing.

By integrating a proteome-wide MR design with multi-layered bioinformatics, this study demonstrates the causal associations of specific plasma proteins with UI risk and successfully prioritizes 5 high-confidence pharmacological targets, with KNG1 emerging as the most robust and highly druggable candidate. These findings provide compelling genomic and biological evidence that enhances our understanding of the pathophysiological mechanisms underlying UI, offering promising novel avenues for precision clinical management and targeted drug development.