Proteome Profiling Identifies Serum Biomarkers in Rheumatoid Arthritis

Abstract

Rheumatoid arthritis (RA) causes serious disability and productivity loss, and there is an urgent need for appropriate biomarkers for diagnosis, treatment assessment, and prognosis evaluation. To identify serum markers of RA, we performed mass spectrometry (MS)-based proteomics, and we obtained 24 important markers in normal and RA patient samples using a random forest machine learning model and 11 protein-protein interaction (PPI) network topological analysis methods. Markers were reanalyzed using additional proteomics datasets, immune infiltration status, tissue specificity, subcellular localization, correlation analysis with disease activity-based diagnostic indications, and diagnostic receiver-operating characteristic analysis. We discovered that ORM1 in serum is significantly differentially expressed in normal and RA patient samples, which is positively correlated with disease activity, and is closely related to CD56dim natural killer cell, effector memory CD8⁺T cell, and natural killer cell in the pathological mechanism, which can be better utilized for future research on RA. This study supplies a comprehensive strategy for discovering potential serum biomarkers of RA and provides a different perspective for comprehending the pathological mechanism of RA, identifying potential therapeutic targets, and disease management.

Keywords: ORM1; biomarker; proteomics; rheumatoid arthritis; serum.

Figure 1

Schematic diagram about the analysis of RA potential proteins based on LC-MS proteomics. (A) Identification of differentially expressed proteins. (B) Multiple-analysis methods to assess the importance of differentially expressed proteins.

Figure 2

Identification of differentially expressed proteins. (A) Random forest machine learning method combined with the PPI network importance analysis algorithm to screen differential proteins. (B) Enrichment results of 24 differential proteins. The blue size represents the number of counts corresponding to the result.

Figure 3

Combine other proteomics datasets to analyze differential proteins. (A) Crossover situation in 5 proteomics datasets. (B) Visually display differential proteins co-expressed (greater than 2) in multiple proteomics datasets. Different shapes and colors represent different datasets. (C, D) Enrichment results of 30 differential proteins in 5 proteomics datasets. The green size represents the number of counts corresponding to the result.

Figure 4

Multi-angle analysis of 24 differentially expressed proteins. (A) Immune infiltration analysis protein expression in 23 immune cells in healthy controls and RA patient. (B) Correlation analysis of 24 proteins with statistically significant immune cells. (C) Tissue specificity of 24 differential proteins. (D) Subcellular localization of 24 differential proteins. *p < 0.05, **p < 0.01, ***p < 0.001. ns, no significance.

Figure 5

Correlation analysis between RA laboratory diagnostic indicators and disease activity. Positive and negative signs indicate the direction of the correlation. A positive sign indicates a positive correlation (red), and a negative sign indicates a negative correlation (blue). *p < 0.05, **p < 0.01.

Figure 6

Summarize a variety of analysis methods to select the most meaningful protein as the key protein of RA.