TXNDC12 and GDF2 as genetically supported plasma proteins associated with knee osteoarthritis: evidence from Mendelian randomization and preliminary biological support
Background Primary knee osteoarthritis (PKOA) is a common degenerative joint disease lacking effective disease-modifying therapies. Identifying circulating proteins associated with PKOA may provide clues for therapeutic target discovery. Methods A two-sample Mendelian randomization (MR) study was conducted using protein quantitative trait locus data from deCODE Genetics, including 4,907 plasma proteins measured in 35,559 individuals, and PKOA genetic association data from the FinnGen study, comprising 18,563 cases and 363,345 controls. Single nucleotide polymorphisms were selected as instrumental variables, and sensitivity analyses were performed to assess the validity and robustness of the MR estimates. Bonferroni correction and the Benjamini- Hochberg procedure were applied for multiple testing correction. Protein–protein interaction, drug–gene interaction, and functional enrichment analyses were performed for prioritized proteins. Western blotting and enzyme-linked immunosorbent assay were used to provide preliminary biological support for MR-prioritized proteins. Results Two plasma proteins showed genetic associations with PKOA susceptibility. Genetically predicted higher levels of thioredoxin domain-containing protein 12 (TXNDC12) were associated with an increased risk of PKOA (OR = 1.38, 95%CI: 1.27–1.50, P = 9.33 × 10–15), whereas growth/differentiation factor 2 (GDF2) was associated with a reduced risk (OR = 0.93, 95%: CI; 0.89–0.98, P = 4.17 × 10–2). After multiple testing correction, TXNDC12 remained significant after Bonferroni and false discovery rate (FDR) correction, while GDF2 met the FDR threshold only and should therefore be considered a hypothesis-generating finding requiring further validation. Network analyses identified 21 functionally related genes, including GSTA3, GSTM4, ACVRL1, and BMPR2. Preliminary experimental analyses showed increased TXNDC12 expression and decreased GDF2 expression in PKOA model mice compared with sham controls, consistent with the MR results. Conclusion Our study identifies TXNDC12 as a genetically supported PKOA-associated plasma protein, while GDF2 remains a hypothesis-generating candidate requiring further validation. These findings provide preliminary associative evidence linking both proteins to PKOA and warrant further mechanistic investigation.