Background/Purpose: Rheumatoid arthritis (RA) is a remarkably heterogeneous autoimmune disease whose clinical outcomes with disease-modifying antirheumatic drugs (DMARDs) remain unpredictable. The aim of this study was to characterize the molecular landscape of RA patients by using a multi-omic approach involving transcriptomics and proteomics and their association with disease status and clinical response.

Methods: Peripheral blood mononuclear cells from 123 RA patients were profiled by RNAseq on Illumina platforms. The RA cohort included 49 patients taking conventional DMARDs and 74 biologics-naïve patients before receiving biologic (TNFi) or targeted-synthetic DMARDs (JAKinibs). Clinical outcomes were evaluated after 6 months of treatment following EULAR criteria. A consensus cluster approach was used to identify patients’ subgroups based on their transcriptomic profile. In parallel, a panel of 92 inflammatory mediators was analyzed in the RA serum using the Olink platform.

Results: Unsupervised hierarchical clustering identified 3 subgroups of RA patients displaying differential expression of 7 gene modules defining distinctive biological pathways. Cluster 1 included the inflamed-myeloid group, showing high expression in myeloid and inflammation genes modules. Cluster 2 defined the healthy-like RA patients, displaying low levels of all gene modules. Cluster 3, identified the B-cells group, exhibiting high levels of B cell gene-modules. The analysis of the serum proteome among these clusters identified 34 proteins showing differential expression levels, where a signature of chemokines, interleukins, and growth factors was found increased in C1 compared with C2 and C3.

Clinically, C1 grouped RA patients with more severe disease activity, higher number of circulating monocytes and neutrophils, and larger disease evolution compared with C2 and C3. On the contrary, C2 involved patients with the lowest disease severity and evolution time and included few patients starting b- or ts-DMARDs. Lastly, C3 included patients with disease activity and evolution time like C2 but showed the highest number of circulating lymphocytes.

Regarding the response to DMARDs, C1 grouped RA patients with the highest percentage of response to TNFi. Clinical response of patients treated with JAKinibs was independent of the cluster where patients were allocated. That data suggests that clinical response to each drug might be associated with a deregulated expression of specific modules of genes and inflammatory proteins at baseline. Accordingly, correlation studies among DAS28 variation at 6 months and basal levels of different gene modules and proteins showed specificity depending on the drug analyzed.

Conclusion:

1. RA patients conform distinctive subgroups based on altered transcriptomic and proteomic profiles, directly linked to their clinical status.
2. Clinical effectiveness of TNFi and JAKinibs was divergent among these molecular clusters and associated with specific transcriptomic and proteomic profiles before starting such therapies.

Supported by the EU/EFPIA-IMI Joint Undertaking 3TR, ISCIII (PI21/0591, CD21/00187 and RICOR-21/0002/0033), RYC2021-033828-I, and JA (P20_01367); co-financed by FEDER.

« Back to ACR Convergence 2023

ACR Meeting Abstracts - https://acrabstracts.org/abstract/the-molecular-stratification-of-rheumatoid-arthritis-using-high-throughput-technologies-is-directly-associated-with-disease-activity-and-clinical-response/