Identification and Validation of Fibroblast-Associated Genes in Osteoarthritis Based on High-Dimensional Weighted...

Background: Osteoarthritis (OA) is a degenerative joint disease with articular cartilage destruction, triggering a pro-inflammatory response. The aim of this study was to screen key genes associated with fibroblasts based on single-cell transcriptomic data and explore their potential value in OA diagnosis. Methods: We obtained RNA sequencing (RNA-seq) and single-cell RNA-seq (scRNA-seq) data of OA from the Gene Expression Omnibus (GEO) database. The CellChat package for cell-to-cell communication analysis and identification of possible ligand-receptor pairs. High-dimensional weighted gene coexpression network analysis (hdWGCNA) was applied to identify the gene modules, and the key genes in the modules were identified and subjected to functional enrichment analysis. Subsequently, limma packages were used to screen for differentially expressed genes (DEGs) between OA and its control samples. Finally, the R package multipleROC was used to test the diagnostic potential of the screened key genes and to construct an OA diagnostic model using the rms package. Result: Eight cell populations were identified and annotated based on scRNA-seq and the percentage of fibroblasts was the highest. The cell-cell communication analysis has suggested that the highest communication probability was seen between mesenchymal cells/T cells and fibroblasts through the pairs of CD99-CD99. The hdWGCNA analysis suggested that genes of modules M3, M4, M5, M6, and M8 (50 genes in total) were highly expressed in fibroblasts. Thereafter, we obtained 394 DEGs in OA and its control samples and took intersections with 50 modular genes and identified seven central genes (including apolipoprotein D [APOD], biglycan [BGN], MXRA5, THY1, C1QTNF3, dermatopontin [DPT], and osteoglycin [OGN]). The constructed diagnostic models showed good predictive performance with all area under the curve (AUC) values >0.8. Finally, a satisfactory diagnostic model was established using these seven genes, and the differences in mRNA expression levels of these genes in OA and normal tissues were verified. Conclusion: For the first time, our study systematically screened and validated key genes with diagnostic potential based on fibroblast-specific single-cell data in combination with hdWGCNA, providing a new theoretical basis and research direction for molecular typing and diagnosis of OA.