Identification of novel protein biomarkers for knee osteoarthritis by integrating human plasma proteome: evidence from Mendelian randomization and preliminary in vitro investigation

Objective This study sought to identify candidate plasma proteins with potential causal associations with knee osteoarthritis (KOA) through Mendelian randomization, and to provide preliminary biological evidence through in vitro experiments. Methods Two-sample Mendelian randomization (MR) was conducted utilizing genetic instruments on 4,489 plasma proteins as exposures. Protein quantitative trait locus (pQTL) data were sourced from three large-scale studies. Genetic associations with KOA were obtained from a genome-wide association study (GWAS) of the European ancestry (accession ID: ebi-a-GCST007090; n = 403,124). The primary MR method was inverse variance weighting (IVW), supplemented by MR-Egger, weighted median, and weighted mode. Pleiotropy and reverse causality were assessed in sensitivity analyses, along with co-localization and protein–protein interaction (PPI) analyses. Preliminary in vitro exploration was conducted using macrophages (MØs) and synovial fibroblasts (SF) to assess biological plausibility. Results Galectin-3 (encoded by LGALS3 [prot-b-6]) was identified as a biomarker of KOA (odds ratio [OR] = 1.07, 95% confidence interval [CI]:1.03–1.11, p = 0.00048), with moderate-to-strong co-localization support for a shared causal variant (posterior probability of H4 [PPH4] = 77.3%, lead single nucleotide polymorphism [SNP]: rs9323280). Exploratory in vitro experiments revealed that galectin-3 stimulation upregulated the expression of inflammatory cytokines (tumor necrosis factor alpha, interleukin-1β, and interleukin-18) and related inflammatory mediators (proteinase 3 [encoded by PRTN3] and myeloperoxidase [encoded by MPO]) in MØs, suggesting a potential pro-inflammatory association. Conclusion This study identifies galectin-3 as a preferred plasma protein biomarker of KOA. The in vitro experiments provide preliminary biological plausibility for galectin-3-mediated inflammatory responses, suggesting potential involvement of myeloperoxidase and proteinase 3. These findings offer novel etiological insights and position galectin-3 as a therapeutic target, which warrants further exploration.