Background/Purpose: Rheumatoid arthritis (RA) is a chronic inflammatory disease that can lead to joint damage and disability. Although methotrexate is the standard first-line treatment, about 40% of patients fail to reach remission/low disease activity and require escalation to bDMARDs. Anti-TNF agents are frequently used, but responses vary and predicting individual treatment responses remains difficult.

Methods: This international multi-centrer prospective observational study, recruited RA patients who were non-responders to synthetic DMARDs and eligible for anti-TNF therapy. Patients were followed from Visit 1 to 5 (~20 weeks). Due to selective criteria and missing data, 42 of the 55 enrolled patients were analysed. Demographic and clinical characteristics were recorded at each visit and analysed to assess anti-TNF response in RA. In parallel, pre-treatment transcriptomic profiling (RNASeq by Quantseq method with sequencing depth of ~20M single-end reads / sample) of synovial biopsies (Visit 2) was performed to provide transcriptomic data, and anti-TNF therapy was initiated thereafter.Prediction of anti-TNF response via transcriptomic scoring will be presented separately. At baseline, the level of expression of mRNA corresponding to proteins directly involved in the mechanism of actions (Drug Target Complex or DTC (Paul et al., 2019)) of main DMARDs families (BCellInhibitors, AntiIL1, AntiIL6, JakStatInhibitors, AntiTNF, TCellInhibition) have been studied. Treatment response was assessed using the EULAR criteria based on DAS28-CRP, patients were classified as responders (good or moderate), or non-responders. Statistical analysis used Pearson’s chi-squared and Wilcoxon rank-sum tests to assess links between baseline traits and treatment response.

Results: Based on the EULAR DAS28-CRP criteria at Visit 5, 32.6% of patients were classified as non-responders and 67.4% as responders (good or moderate). No significant associations were found between treatment response and sex (p = 0.356), smoking status (p = 0.653), or age (p = 0.243). However, BMI was significantly linked to treatment response (p = 0.049), suggesting higher BMI may reduce clinical response. Transcriptomic data of DTCs revealed two main observations: (1) transcriptomic profiles of patients are highly heterogeneous, each set of mRNA related to the different DTCs appearing as predominant in similar proportions of patients. (2) two clusters emerged: patients with high B-cell DTC expression tended to show low expression of other DTCs, and vice versa. In contrast, expression levels of non–B-cell DTCs are mutually correlated.

Conclusion: The clinical characteristics and early treatment outcomes align with expectations for this population. Initial analyses suggest that BMI may be associated with therapeutic response, while sex, age, and smoking are not. Early transcriptomic profiling may reveal mechanistic subtypes of RA, with opposing expression patterns between B-cell–related and other DMARDs-related mechanisms. Future analyses will assess the predictive ability of a synovial transcriptomic signature to improve patient stratification and personalize RA treatment.

Table 1. Baseline age and BMI by EULAR response

Table 2. Smoker and gender status by EULAR response

Figure 1. Correlations between the main DMARDs families

« Back to ACR Convergence 2025

ACR Meeting Abstracts - https://acrabstracts.org/abstract/clinical-and-transcriptomic-heterogeneity-in-patients-with-ra-a-multi-centric-international-study-of-anti-tnf-response/