Integrative analysis of single-cell and bulk RNA sequencing data using machine learning identifies PTGDS as a key subsynovial fibroblast gene in osteoarthritis.

Synovial lesions significantly contribute to osteoarthritis (OA) pathogenesis; nonetheless, the specific roles played by synovial cells in OA progression remain inadequately understood. The current study combined single-cell RNA sequencing (scRNA-seq) data with bulk RNA sequencing to investigate the functional involvement of subsynovial fibroblasts (SSF) within the synovial microenvironment of OA. An ensemble machine learning (ML) approach consisting of 113 algorithmic combinations identified four key genes associated with SSF. Pathway analysis indicated that SSF exhibited increased angiogenesis-related pathways and decreased inflammatory response pathways. Consensus clustering classified OA patients into two molecular subtypes (C1 and C2). Immune infiltration analysis identified PTGDS and SCG2 as immune-associated genes differentiating these subgroups. Mendelian randomization (MR) analysis indicated a potential causal relationship between elevated PTGDS expression and OA risk. To validate the biological significance of PTGDS in SSF, a knee OA (KOA) rat model was established. Histopathological analyses using Masson staining (MASSO) and hematoxylin-eosin (HE) staining revealed significant morphological alterations in cartilage and synovial tissues. Additionally, immunohistochemistry (IHC), western blotting (WB), and quantitative real-time PCR (RT-qPCR) demonstrated increased expression of PTGDS and SSF markers (WISP2 and THY1) at both protein and mRNA levels in the KOA group. Crucially, double immunofluorescence (IF) staining provided clear histological evidence that PTGDS is predominantly expressed in SSF rather than in synovial intimal fibroblasts (SIF), thus linking this molecular signature explicitly to a distinct synovial cell subset. This integrative strategy combining transcriptomic analysis with histological validation presents a valuable approach for elucidating OA pathogenesis and identifying novel therapeutic targets.