Transcriptomic and proteomic analysis stratifies patients with axial spondyloarthritis based on disease activity,...

To identify clusters of highly correlated genes enriched in biological functions and specific molecular pathways involved in the pathogenesis of radiographic damage in axial spondyloarthritis (axSpA) and to discover molecular biomarkers of radiographic progression and disease severity.

A total of 144 patients with axSpA were included. First, RNA from peripheral blood mononuclear cells was sequenced in a cohort of 24 patients with axSpA. Hub genes were measured in a n=60 validation cohort through microfluidic PCR. A 5-year follow-up enabled the classification of the patients into fast/moderate or slow progressors. Machine learning approaches were applied to identify a predictive biomarker of progression by integrating gene expression data with clinical variables. An independent cohort of 60 patients with axSpA, with spine radiographs taken 5 years prior, underwent serum proteomic analysis using a Proximity Extension Assay.

Unsupervised clustering analysis using transcriptomics revealed two distinct groups of patients with axSpA, differentiated by their clinical profiles. Weight gene correlation network analysis identified six gene modules differentially expressed between the two clusters. Patients in cluster 2 exhibited higher disease activity, greater functional impairment and more structural damage. Molecular alterations linked to structural damage revealed a specific circulating inflammatory proteome profile associated with disease severity. A predictive model composed of two genes and basal total modified Stoke Ankylosing Spondylitis Spinal Score emerged as a key biomarker for identifying moderate-to-fast radiographic progression.

This study identified molecular pathways involved in radiographic damage and discovered potential proteomic biomarkers of disease severity and transcriptomic predictors of radiographic progression in axSpA.