Background: Response to biologic disease-modifying antirheumatic drugs (bDMARDs), including anti–tumor necrosis factor (anti-TNF) agents, varies substantially among patients with immune-mediated inflammatory rheumatic diseases (IRDs), such as rheumatoid arthritis (RA) and axial spondyloarthritis (axSpA). However, prognostic factors for treatment response remain poorly understood. Currently, no validated clinical, genetic, or proteomic biomarkers are available to predict treatment response, which precludes effective patient stratification and limits the implementation of precision medicine strategies.

Objectives: To conduct a discovery proteomic analysis to identify circulating protein biomarkers predictive of response to anti-TNF therapy in inflammatory rheumatic diseases, with a focus on rheumatoid arthritis and axial spondyloarthritis.

Methods: This retrospective observational study included baseline serum samples selected from two cohorts of patients with RA and axSpA initiating anti-TNF treatment at a tertiary referral hospital in A Coruña, Spain, between 2020 and 2023. Demographic and clinical variables were collected, and treatment response was assessed after 6 months of therapy according to standardized clinical criteria, classifying patients as responders or non-responders. In RA, treatment response was defined using EULAR response criteria, excluding moderate responders. In axSpA, response was assessed using ASAS improvement criteria together with ASDAS disease activity thresholds, defining non-response as ASAS improvement <1.1 in the presence of high or very high disease activity (ASDAS ≥ 2.1).

Serum Proteomic profiling was performed using the Olink® Reveal proximity extension assay (PEA), a multiplex immunoassay quantifying 1,037 proteins predominantly related to inflammatory pathways. Protein levels were reported as Normalized Protein eXpression (NPX) values. Data analysis was conducted using MetaboAnalyst and R Software. Multivariate analyses included principal component analysis (PCA) and partial least squares-discriminant analysis (PLS-DA). Associations between protein levels and treatment response were evaluated using volcano plot analysis (fold-change ≥1.2), generalized linear models (GLM), and receiver operating characteristic (ROC) curve analysis.

Results: A total of 86 serum samples were analyzed, including 25 RA patients, 45 axSpA patients, and 16 healthy controls. Baseline demographic and clinical characteristics of RA and axSpA patients are summarized in Table 1. After 6 months of anti-TNF therapy, 15 RA patients were classified as good responders and 10 as non-responders, while 28 axSpA patients were classified as responders and 17 as non-responders. Unsupervised PCA showed no global baseline proteomic differences according to treatment response, whereas supervised PLS-DA identified a subset of proteins with high Variable Importance in Projection (VIP) scores associated with response. Volcano plot analysis identified 16 proteins associated with treatment response in the combined cohort (p<0.05). In disease-specific analyses, 22 proteins were significantly modulated in RA and 22 in axSpA. ROC curve analysis further refined these findings, highlighting one protein that was consistently associated with response across both diseases, with higher baseline levels in responders. In RA, the three proteins with the highest discriminatory performance showed strong discriminatory performance (AUC 0.85-0.86) and were consistently decreased in responders. In axSpA, the three best-performing candidates showed moderate accuracy (AUC 0.73-0.78), with one protein increased in responders and two showing the opposite pattern. GLM confirmed independent associations between the selected candidate proteins with treatment response.

Conclusions: Baseline serum proteomic profiling identified candidate protein biomarkers associated with response to anti-TNF therapy across two inflammatory rheumatic diseases, RA and axSpA. Although validation in independent cohorts is required, these findings support the potential of high-throughput serum proteomics to stratify patients prior to treatment and represent a first step towards precision medicine in inflammatory rheumatic diseases.

REFERENCES: NIL.

Acknowledgments: NIL.

Disclosure of Interests: None declared.