Causal Effects of Gut Microbiota on Osteoarthritis: Insights from Genome-Wide Mendelian Randomization, Bulk RNA, and Single-Cell RNA

Background/Objectives: Alterations in the gut microbiota (GM) have been linked to osteoarthritis (OA), yet the causal relationship and fundamental mechanisms remain unclear. This study conducted a comprehensive analysis to prioritize OA-related candidate genes through integration of GM-associated genetic signals and transcriptomic data and to explore their potential biological relevance. Methods: A Mendelian randomization (MR) investigation was performed to assess potential causal links between GM composition and OA, leveraging publicly available genome-wide association study (GWAS) datasets. Differentially expressed genes (DEGs) from GSE169077 and microbial taxonomic group (MTG)-associated SNP-mapped genes were intersected to identify SNP-mapped OA DEGs, and OA-related candidate genes were further prioritized through cross-dataset expression assessment. In addition, a nomogram was constructed, functional analysis was conducted to identify relevant enriched pathways, followed by database-based TF–gene network analysis. In addition, GSE255460 was analyzed to perform ppseudotimeand cell communication analyses. The expression levels of OA-related candidate genes were assessed by real-time quantitative polymerase chain reaction (RT-qPCR) in knee synovial tissue samples from the OA group and the tibial plateau fracture control group. Results: In total, 10 SNP-mapped OA DEGs were identified, among which, NELL1 and MPST were prioritized as OA-related candidate genes based on their concordant expression patterns across the GSE169077 and GSE114007 datasets. In addition, an exploratory nomogram based on NELL1 and MPST was constructed using GSE169077. GSEA identified BH-adjusted significant enrichment of the ribosome and spliceosome pathways associated with MPST and the ECM–receptor interaction pathway associated with NELL1. Database-based network analysis predicted 67 putative TFs, including MYC and EZH2, with potential links to the OA-related candidate genes. Moreover, pseudotime analysis suggested variation in MPST expression along the inferred cell-state trajectories of inflammatory chondrocytes (InfCs) and reparative chondrocytes (RepCs), while CellChat analysis predicted relatively strong communication between these two cell populations. In addition, RT-qPCR analysis showed that the expression of NELL1 was significantly upregulated, while the expression of MPST showed the opposite trend in OA, which was consistent with the prediction results of the database. Conclusions: Our integrative analysis identified potential associations between selected gut microbial taxa and OA and highlighted NELL1 and MPST as OA-related candidate genes. Evidence from bulk transcriptomics, single-cell analyses, and synovial RT-qPCR provided convergent but exploratory support for their relevance to OA. These findings provide a basis for further mechanistic and clinical validation of GM-associated molecular features in OA.