Multiomics Identifies Potential Biomarkers in Ankylosing Spondylitis Bone Formation

Area of Science:

• Genetics and Genomics
• Immunology
• Computational Biology

Background:

• Ankylosing spondylitis (AS) is a chronic inflammatory disease with complex genetic factors, and current treatments are insufficient to halt its progression.
• Identifying novel therapeutic targets is crucial for improving AS management and patient outcomes.

Purpose of the Study:

• To discover potential therapeutic targets for AS by integrating Mendelian Randomization (MR), transcriptomics, and machine learning.
• To identify key genes and their causal relationships with AS for future therapeutic interventions.

Main Methods:

• Differential gene expression analysis (GEO database) and cis-eQTL data (eQTLGen Consortium) were used to identify AS-associated genes.
• Mendelian Randomization (MR) and Summary Data-Based Mendelian Randomization (SMR) analyses screened for causal relationships.
• Machine learning models identified key feature genes, followed by phenome-wide association studies (PheWAS) and molecular docking.
• A collagen-induced AS mouse model (CAIA) was used for in vivo validation via RT-qPCR and histological analysis.

Main Results:

• 1607 differentially expressed genes (DEGs) were identified, with 33 showing a causal relationship with AS after MR analysis.
• Four machine learning algorithms identified RIOK1, FUCA2, COL9A2, USP16, and TTC16 as top feature genes.
• In the CAIA mouse model, FUCA2 and USP16 expression was significantly elevated, while TTC16 was reduced in joint tissues.
• PheWAS indicated potential beneficial or harmful effects of these genes on various disease phenotypes.

Conclusions:

• FUCA2, USP16, and TTC16 show altered expression in an AS mouse model and may serve as valuable biomarkers for ankylosing spondylitis.
• These findings provide a foundation for developing novel therapeutic strategies targeting specific genes in AS treatment.