Sex dimorphism of IL-17-secreting peripheral blood mononuclear cells in ankylosing spondylitis based on bioinformatics analysis and machine learning

Abstract

Abstract Background Ankylosing spondylitis (AS) with radiographic damage is more prevalent in men than in women. IL-17, which is mainly secreted from peripheral blood mononuclear cells (PBMCs), plays an important role in the development of AS. Its expression is different between male and female. However, it is still unclear whether sex dimorphism of IL-17 contribute to sex differences in AS. Methods GSE221786, GSE73754, GSE25101, GSE181364 and GSE205812 datasets were collected from the Gene Expression Omnibus (GEO) database. Differential expressed genes (DEGs) were analyzed with the Gene Set Enrichment Analysis (GSEA), Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) methods. CIBERSORTx and EcoTyper algorithms were used for immune infiltration analyses. Machine learning based on the XGBoost algorithm model was used to identify the impact of DEGs. The Connectivity Map (CMAP) database was used as a drug discovery tool for exploring potential drugs based on the DEGs. Results According to immune infiltration analyses, T cells accounted for the largest proportion of IL-17-secreting PBMCs, and KEGG analyses suggested an enhanced activation of mast cells among male AS patients, whereas the expression of TNF was higher in female AS patients. Other signaling pathways, including those involving metastasis-associated 1 family member 3 (MAT3) or proteasome, were found to be more activated in male AS patients. Regarding metabolic patterns, oxidative phosphorylation pathways and lipid oxidation were significantly upregulated in male AS patients. In XGBoost algorithm model, DEGs including METRN and TMC4 played important roles in the disease process. we integrated the CMAP database for systematic analyses of polypharmacology and drug repurposing, which indicated that atorvastatin, famciclocir, ATN-161 and taselisib may be applicable to the treatment of AS. Conclusions We analyzed the sex dimorphism of IL-17-secreting PBMCs in AS. The results showed that mast cell activation was stronger in males, while the expression of TNF was higher in females. In addition, through machine learning and the CMAP database, we found that genes such as METRN and TMC4 may promote the development of AS, and drugs such as atorvastatin potentially could be used for AS treatment.

Keywords

• Sex dimorphism
• Interleukin-17
• Ankylosing spondylitis
• Machine learning
• Bioinformatics analysis