Background: Intact signaling of cortisol through the glucocorticoid receptor (GR) in synovial cells is essential for the suppression of experimental inflammatory arthritis [1]. On the other hand, blood cells overexpress NR3C1 , the gene encoding the GR, during hypocortisolemia [2], while lower circulating cortisol levels in patients with rheumatoid arthritis (RA) have been associated with worse treatment responses [3]. Data on GR expression in the rheumatoid synovium, as well as its association with local proinflammatory responses in RA patients, are limited.

Objectives: To examine whether GR is aberrantly expressed in rheumatoid synovium and the possible association of its expression with the synovial activation status.

Methods: We defined the cellular pattern of NR3C1 expression in the synovium using single-cell RNA-sequencing (scRNA-seq.) data derived from synovial tissue of 6 RA patients and 6 osteoarthritis patients [4], as well as from synovial fibroblasts (SFs) of arthritic and normal (wild-type) mice [5]. We also examined bulk RNA-sequencing data derived from synovial biopsies of 151 patients with early or established RA versus 22 osteoarthritis patients, and 28 apparently healthy controls with no cartilage damage or synovitis on knee arthroscopy [6]. Early RA (n=57) was defined as within 12 months of diagnosis without prior treatment, while established RA (n=94) was defined as disease duration > 12 months and previous treatment with disease-modifying antirheumatic drugs (6).

Results: In the human scRNA-seq. synovial dataset we observed that NR3C1 was expressed in all detected cell populations including lining and sublining SFs, venous, arterial, and lymphatic endothelial cells, vascular smooth muscle cells, pericytes, leukocytes and a cluster of unassigned cells. Similarly, analysis of mouse SFs at the single-cell level revealed the ubiquitous expression of Nr3c1 in all 9 identified SF subsets. In the large bulk RNA-seq. dataset including 201 synovial biopsies we observed that GR expression was higher in RA than in healthy and osteoarthritic tissue (mean±SD normalized counts 24,764±4,929, 15,633±9,058 and 19,280±8,862, respectively; both P<0.001), regardless of disease duration or prior treatment. Given that GR expression varied among RA patients, we searched for differences between RA patients with higher vs lower GR expression. Indeed, the synovial transcriptome of RA patients with high compared to low GR expression (first quartile: 30,517±4,876 normalised counts vs fourth quartile: 19,382±2,523 normalised counts) was enriched for proinflammatory gene sets, including ‘IL6/JAK/STAT3 signaling’ and ‘IFN-γ response’. Similar results were obtained when examining early, untreated RA patients alone. High synovial GR expression was also associated with increased JAK2 and PTPRK expression, which have been recently described as markers of activated sublining SFs [7]. In contrast, low GR expression was associated with increased COMP and COL6A2 , denoting a resting synovial state [7].

Conclusion: GR expression is increased in the synovium of some RA patients, where it may act as a compensatory mechanism sensitizing synovial tissue to glucocorticoid action. However, when cortisol is inadequate, the partially liganded GR is probably unable to suppress synovial inflammation in these patients.

REFERENCES: [1] Koenen M. et al. , Ann Rheum Dis . 2018;77:1610–8

[2] Hagendorf A. et al., J Clin Endocrinol Metab . 2005;90:6237–43

[3] Yavropoulou M. et al. RMD Open, 2024 (in press )

[4] Wei K. et al., Nature . 2020;582:259–64

[5] Armaka M. et al. , Genome Med . 2022;14:78

[6] Guo Y. et al. , J Immunol . 2017;198:4490–501

[7] Smith M.H. et al., Nat Immunol 2023;24:1200–1210

Acknowledgements: NIL.

Disclosure of Interests: None declared.