Integration of synovial pathotypes and serum proteomic modules as determinants of disease progression and therapeutic response in early rheumatoid arthritis
Sagrario Corrales, Julio Osuna-Soto, Andrea Cid-Chaves, Rafaela Ortega-Castro, Lourdes Ladehesa-Pineda, Christian Merlo-Ruiz, Desirée Ruiz-Vílchez, Nuria Barbarroja, Fernando Leiva-Cepas, Carlos Perez-Sanchez, Marta Alarcón-Riquelme, Concepción Aranda-Valera, Alejandro Escudero-Contreras, Chary López-PedreraBackground
Rheumatoid arthritis (RA) is characterised by marked clinical, histological and molecular heterogeneity, complicating treatment-response prediction. This study evaluated associations between synovial pathotypes, pathotype-associated serum proteomic patterns and 6-month clinical outcomes in early RA.
Methods
A total of 147 patients with early RA were included in two independent cohorts. In cohort 1, 25 patients underwent synovial biopsy for histological classification by synovial pathotype and Krenn synovitis score. Serum samples collected at biopsy were analysed using the Olink Inflammation panel. Cohort 2, comprising 122 patients, was used to assess the directional consistency of associations between predefined serum proteomic scores and 6-month European Alliance of Associations for Rheumatology (EULAR) response.
Results
The lymphomyeloid pathotype was associated with higher disease activity, autoantibody positivity and increased synovial immune infiltration. High Krenn score was also associated with greater inflammatory burden and higher disease activity. However, no significant differences in treatment response were observed according to Krenn score, whereas synovial pathotype was associated with differences in biological or targeted synthetic disease-modifying antirheumatic drug (b/tsDMARD) initiation and exploratory 6-month response patterns. Pathotype-associated serum protein patterns identified in cohort 1 showed differential distributions across histological groups. Higher lymphomyeloid scores were observed among conventional synthetic disease-modifying antirheumatic drug responders, whereas higher diffuse-myeloid scores were observed among b/tsDMARD non-responders. These treatment-stratified associations were exploratory and directionally consistent after adjustment in cohort 2, although neither reached statistical significance.
Conclusion
Synovial pathotypes in early RA were associated with distinct circulating proteomic patterns reflecting tissue-level heterogeneity. This tissue-to-blood approach supports development of pathotype-associated serum profiles as candidate minimally invasive biomarkers and provides a promising framework for precision medicine in RA.