Identification of candidate sex hormone-associated genes and immune infiltration characteristics in osteoarthritis based on bioinformatics analysis and machine learning.

BACKGROUND

Sex hormones play critical roles in the pathogenesis and progression of osteoarthritis (OA), yet the hormone-related molecular networks remain poorly defined. This study aimed to identify candidate sex hormone-associated genes in OA and to explore their potential functional enrichment and immune-related characteristics using bioinformatics analysis.

METHODS

OA gene expression data were obtained from the GEO database and integrated with candidate sex hormone-associated genes retrieved from GeneCards. The R package "limma" was then used to identify differentially expressed genes (DEGs) and sex hormone-associated DEGs (SADEGs). OA-associated SADEGs, termed OA-SADEGs, were selected using weighted gene co-expression network analysis (WGCNA), and their potential biological functions and pathways were explored by GO and KEGG enrichment analyses. Hub genes were identified using three machine learning models. xCell analysis was used to estimate immune infiltration and its associations with hub genes, and hub gene expression was further evaluated in external datasets and peripheral blood samples.

RESULTS

We identified 32 sex hormone-associated genes in OA, enriched in extracellular matrix remodeling, receptor signaling, and antigen presentation pathways. Three candidate hub genes (LOXL1, HLA-DRA, and CYBB) were consistently upregulated in OA and showed significant correlations with immune infiltration scores. xCell analysis identified 13 differentially enriched immune cell types, of which three were associated with hub genes. External dataset analysis and peripheral blood qRT-PCR showed upregulation of LOXL1, HLA-DRA, and CYBB in OA samples.

CONCLUSION

This study integrated bioinformatics and immune analyses to identify candidate sex hormone-associated genes in OA. These findings provide associative bioinformatics evidence for sex hormone-associated molecular features in OA.