Rheumatoid arthritis (RA) is the most prevalent autoimmune mediated joint disease. A complex dysfunction of the immune response is considered central for its pathogenesis as well as a dysregulated crosstalk between fibroblast-like synoviocytes (FLS) and monocytes.
Objectives
This project aims to examine and validate the anti-inflammatory effects of the natural glucose metabolite itaconate-derivative 4-octyl itaconate (4-OI) on stromal and immune cells from patients with active RA and evaluate the downstream effector molecule Heme-Oxygenase 1 (HO-1) as a potential biomarker.
Methods
Synovial fluid mononuclear cells (SFMC) from patients with chronic RA (cRA) were used to harvest RA FLS. Monocultures of synovial fluid derived FLS (SF-FLS) and autologous co-cultures of SF-FLS and peripheral blood mononuclear cells (PBMC) were cultured with and without 4-OI, corticosteroids, and anti-TNF (n=7). Subsequently, analyses by flow cytometry, Western Blotting, MTT assay, and ELISA were performed. The 4-OI target protein Nrf2 was knocked out in immortalized FLS (FLS-KO) generated by Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR). Furthermore, we measured HO-1 by ELISA in plasma samples from newly diagnosed, active and treatment-naïve early RA (eRA) patients at baseline and after 6 months of intensive treatment (the OPERA trial, n=80)[1] with a follow-up period of 24 months and plasma and in synovial fluid samples from patients with active, chronic RA (cRA, n=20) during disease activity flares. Plasma samples from age approximated and gender matched healthy controls (HC) (n=35) were also included.
Results
In vitro 4-OI significantly inhibited MCP-1 secretion in cultured SFMC, THP-1 (monocyte cell line), and FLS and resulted in a concomitant increased HO-1 production. 4-OI reduced the MCP-1 secretion by 84%, anti-TNF by 41% and corticosteroids by 54% (all p<0.01) in co-cultures consisting of autologous RA SF-FLS and PBMCs. In FLS we observed a linear decrease in pro-inflammatory cytokine production. 4-OI did not decrease collagen production, even in dosage of 500 µn FLS we observed a linear decrease in pro-inflammatory cytokine production. 4-OI did not decrease collagen production, even in dosage of 500 1 production. 4-OI reduced the MCP. In eRA plasma levels of HO-1 was significantly increased at baseline (2039 pg/mL, 95% CI [1619, 2460]) vs HC (1209 pg/mL, 95% CI [1047, 1370]). Treatment with MTX or MTX + anti-TNF, lowered levels of HO-1 (1704 pg/mL, 95% CI [1377, 2031]) after 6 months irrespective addition of anti-TNF treatment, reducing the difference to a no longer significantly different level from HCs. HO-1 was associated with number of swollen joints (r=0.25, p<0.02), tender joints (r=0.24, p<0.02) and Clinical Disease Activity Index (r=0.23, p<0.02). Baseline HO-1 predicted radiographic joint damage progression (change in Total Sharp-van der Heijde score (Δn eRA plasma levels of HO-1 was significantly increased at baseline (2039 pg/mL, 95% CI [1619, 2460]) vs HC (1209 pg/mL, 95% CI [1047, 1370]). Treatment with MTX or MTX + anti-TNF, lowered levels of HO-1 (1704 pg/mL, 95% CI [1377, 2031]) after 6 months irund (r=0.75, p<0.001). High levels of HO-1 in SF from inflamed joints were associated with CRP (r=0.52, p<0.05) during flares in cRA.
Conclusion
In vitro 4-OI decreased MCP-1 secretion and increased the HO-1 expression in cultures of monocytes and in RA FLS and PBMC without decrease in collagen production or cell growth. In total, 4-OI decreased MCP-1 in both immune and stromal dominated RA pathobiology. In eRA patients, pre-treatment-plasma HO-1 levels were increased and correlated with parameters for disease activity, treatment-effect after 6 months, and predicted radiographic progression after 24-month.
Reference
[1] Hørslev-Petersen K et al. Ann Rheum Dis. 2013
Acknowledgements
We thank medical doctors and nurses at the Department of Rheumatology, Aarhus University Hospital for helping to collect the patient samples. The authors kindly acknowledge a generous grant from the Danish Rheumatoid Association (R177-A6156).
Background: Bone erosions in rheumatoid arthritis (RA) is a major complication. Despite improved treatment, erosions still occur and progress. Therefore, a continuous investigation of the interplay between bone regulation and immune activity is needed. Co-inhibitory receptors, like CTLA-4, participate in modulating osteoclast activity 1 , and blocking these receptors in cancer treatment results in autoimmune disease 2 . Programmed death 1 (PD-1) is a central co-inhibitory receptor, also present in a soluble (s) form. We have previously shown that sPD-1 is associated with disease activity and reduced bone erosions in RA 3 . PD-1 and its ligands are glycosylated and the glycosylation affects signaling through the PD-1 pathway 4,5 . Galectins (GAL) bind to specific glycosylation patterns on glycoproteins by their carbonrecognition domain. Gal-3 can bind to multiple immune receptors shaping the immunological response 6 . Objectives: We aimed to investigate if PD-1 and GAL-3 modulate osteoclastogenesis in RA. Methods: Plasma and synovial fluid (SF) samples were collected from patients with chronic (c) RA (>8 years of disease, n=14) SF and blood samples were obtained when patients presented with disease flare. Soluble PD-1 and GAL-3 were investigated by ELISA. GAL-3:PD-1 complexes were captured in an optimized ELISA using both GAL-3 and PD-1 antibodies. Surface plasmon resonance was used to evaluate on the binding properties between GAL-3 and PD-1, on a Biacore 3000. Lactose was used to block the potential binding. Cells from the synovial fluid (SFMC) were differentiated into osteoclasts with M-CSF and RANKL, recombinant human (rh) PD-1 and rhGAL-3 were added and osteoclast formation evaluated by TRAP in the supernatant. Results: Soluble PD-1 and GAL-3 were present in both plasma and SF from cRA patient, and the ratio (PD-1/GAL-3) was increased in SF. PD-1:GAL-3 complexes were detected in both plasma and SF, with the highest concentration in SF. The binding between PD-1 and GAL-3 was confirmed by surface plasmon resonance analyses, with an estimated Kd of 5uM. Binding could be blocked by addition of lactose, confirming the binding to be glycan dependent. In SFMC osteoclast cultures, rhPD-1 and rhGAL-3 slightly decreased osteoclast formation evaluated by TRAP. However, combining rhPD-1 and rhGAL3 further potentiated the reduction in osteoclast formation by 37%. Conclusion: We confirm glycan depended binding between the co-inhibitory receptor PD-1 and GAL-3. Both sPD-1 and Gal-3, and the PD-1:GAL-3 complexes, are upregulated in the inflamed joint at site of erosions. I n vitro RA culture demonstrates that GAL-3 potentiate the functions of PD-1 and reduces osteoclastogenesis. These findings indicate that the binding between Gal-3 and PD-1 could provide a novel target to control erosions in RA. Future in vivo studies on this interaction is needed. References: [1]Axmann, R. et al. CTLA-4 directly inhibits osteoclast formation. Ann Rheum Dis 67, 1603–1609 (2008). [2]Friedman, C. F., Proverbs-Singh, T. A. & Postow, M. A. Treatment of the Immune-Related Adverse Effects of Immune Checkpoint Inhibitors. JAMA Oncol 2, 1346–8 (2016). [3]Greisen, S. et al. Increased soluble programmed death-1 (sPD-1) is associated with disease activity and radiographic progression in early rheumatoid arthritis. Scand J Rehabil Med 43, 101–108 (2013). [4]Verdura, S. et al. Resveratrol targets PD-L1 glycosylation and dimerization to enhance antitumor T-cell immunity. Aging (Albany NY ) 12, 8–34 (2020). [5]Wang, M. et al. Identification of a monoclonal antibody that targets PD-1 in a manner requiring PD-1 Asn58 glycosylation. Com Biol 1–10 (2019). doi:10.1038/s42003-019-0642-9 [6]de Oliveira, F. L. et al. Galectin-3 in autoimmunity and autoimmune diseases. Exp Biol Med (Maywood ) 240, 1019–1028 (2015). Disclosure of Interests: None declared
4-1BB is a T cell costimulatory receptor and a member of the tumor necrosis factor receptor superfamily. Here, we show that Galectin-3 (Gal-3) decreases the cellular response to its ligand (4-1BBL). Gal-3 binds to both soluble 4-1BB (s4-1BB) and membrane-bound 4-1BB (mem4-1BB), without blocking co-binding of 4-1BBL. In plasma, we detected complexes composed of 4-1BB and Gal-3 larger than 100 nm in size; these complexes were reduced in synovial fluid from rheumatoid arthritis. Both activated 4-1BB + T cells and 4-1BB-transfected HEK293 cells depleted these complexes from plasma, followed by increased expression of 4-1BB and Gal-3 on the cell surface. The increase was accompanied by a 4-fold decrease in TNFα production by the 4-1BB high Gal-3 + T cells, after exposure to 4-1BB/Gal-3 complexes. In RA patients, complexes containing 4-1BB/Gal-3 were dramatically reduced in both plasma and SF compared with healthy plasma. These results support that Gal-3 binds to 4-1BB without blocking the co-binding of 4-1BBL. Instead, Gal-3 leads to formation of large soluble 4-1BB/Gal-3 complexes that attach to mem4-1BB on the cell surfaces, resulting in suppression of 4-1BBL’s bioactivity.
Targeting intracellular pathways with oral small molecules is an attractive therapeutic approach for treating immune mediated inflammatory diseases. The mitogen-activated protein kinase (MAPK) pathway is activated by environmental stressors, growth factors and inflammatory cytokines. However, the inhibition of central MAPK proteins has so far had undesirable side effects. The MAPK-activated protein kinase 2 (MK2) is a downstream mediator in the MAPK signalling pathway and could therefore be inhibited without the same side effects (see figure 1).
Objectives
The objective of this study was to study the effects of a small molecule inhibiting MK2 on inflammation and structural changes in ex vivo models of immune mediated inflammatory arthritis.
Methods
Synovial fluid mononuclear cells (SFMCs), fibroblast like synovial cells (FLSs) and peripheral blood mononuclear cells (PBMCs) were obtained from a study population consisting of patients with active RA or peripheral SpA with at least one swollen joint (for obtaining synovial fluid) (n=14). SFMCs were cultured for 48 hours with and without addition of a MK2 inhibitor (Celgene) at 1000 nM, 333 nM and 111 nM and supernatants were analysed by the Olink proseek multiplex interferon panel and commercially available ELISA assays. Because FLSs are only found in small amounts among SFMCs, autologous co-cultures of FLS and PBMCs and SFMCs were also used. SFMCs cultured for 21 days were used to study inflammatory macrophage differentiation and osteoclastogenesis.
Results
In SFMCs cultured for 48 hours, the MK2 inhibitor decreased the production of CXCL9 (p<0.001), CXCL10 (p<0.01), HGF (p<0.01), CXCL11 (p<0.01), TWEAK (p<0.05), and IL-12B (p<0.05) and increased the production of CXCL5 (p<0.0001), CXCL1 (p<0.0001), CXCL6 (p<0.001), TGFα (p=0.01), MCP-3 (p<0.01), LAP TGFβ (p<0.05) dose-dependently after Bonferroni correction (all corrected P values). At the highest concentration, the MK2 inhibitor also decreased MCP-1 production (p<0.05). In FLS-SFMC co-cultures, the MK2 inhibitor decreased MCP-1 production (p<0.05) but did not change the production of DKK1 and MMP3. In FLS-PBMC co-cultures, the MK2 inhibitor decreased the production of MCP-1 (p<0.0001), increased MMP3 production (p<0.05) but did not change DKK1 production. In SFMCs cultured for 21 days as a model of inflammatory macrophage differentiation and osteoclastogenesis, the MK2 inhibitor decreased the production of MCP-1 (p<0.05) and tartrate-resistant acid phosphatase (TRAP) (p<0.05) but did not change the production of IL-10.
Conclusions
This study reveals the effects of a MK2 inhibitor in ex vivo models of immune mediated inflammatory arthritis. The MK2 inhibitor changed the secretory profile of SFMCs and decreased inflammatory osteoclastogenesis. Taken together, this points to a role of this MK2 inhibitor in attenuating inflammatory and destructive arthritis.
Disclosure of Interest
T. Kragstrup: None declared, A. Mellemkjær: None declared, M. Nielsen: None declared, L. Heftdal: None declared, P. Schafer Shareholder of: Celgene, Employee of: Celgene, B. Deleuran: None declared
Immune mediated inflammatory arthritis including rheumatoid arthritis (RA), psoriatic arthritis (PsA) and spondyloarthritis (SpA) all characterised by joint synovitis. Disease-modifying antirheumatic drugs (bDMARDs) targeting specific components of the pathogenesis have radically improved the treatment of the diseases. However, a fair proportion of patients are non-responders. Today, the first choice of DMARD is dependent on market pricing, regardless of the immunological target. This is due to the rather similar efficacy profile of the different DMARDs. Therefore, there is a need for stratification of patients suffering from Immune mediated inflammatory arthritis in order to reduce the fraction of DMARD non-responders.
Objectives
The objective of this study was to study the effects of various DMARDs on different synovial cell subsets using several human ex vivo models of immune mediated inflammatory arthritis. This could potentially guide future studies of personalising DMARDs in these diseases.
Methods
Synovial fluid was obtained from a study population of patients with active rheumatoid arthritis (RA) or peripheral spondyloarthritis (SpA). Synovial fluid mononuclear cells (SFMCs) containing primarily synovial monocytes and lymphocytes cultured for 48 hours ("Macrophage and Lymphocyte model") were used to study the effect of different biological agents on secretion of monocyte chemoattractant protein-1 (MCP-1) (n=14). Further, fibroblast-like synovial cells (FLSs) were co-cultured with autologous PBMCs ("FLS model") to study the effects of the same biological agents (n=6) in cultures dominated by synovial FLSs. Finally, SFMCs cultured for 21 days ("Osteoclast model") were studied to assess the effects on inflammatory osteoclastogenesis (n=10) measured by tartrate-resistant acid phosphatase (TRAP). The DMARDs investigated are shown in table 1.
Results
"Macrophage and Lymphocyte model"; In SFMCs cultured for 48 hours, all DMARDs included, except anakinra, had the ability to decrease the production of MCP-1. The two TNF inhibitors (adalimumab and etanercept) (p<0.05 and p<0.01) and baricitinib (p<0.05) had the most pronounced effects and reduced the production of MCP-1 by approximately 25%. Tocilizumab had in this culture a non-significant reduction of MCP-1 production. "FLS model"; In the FLS+PBMCs cultured for 48 hours, tocilizumab (p<0.001) and the two JAK inhibitors (tofacitinib and baricitinib, p<0.05 and p<0.05) were exclusive in decreasing the cytokine production of MCP-1 by around 50%. "Osteoclast model"; In SFMCs cultured for 21 days, only the two TNF inhibitors, adalimumab and etanercept were able to significantly reduce the secretion of TRAP from adherent macrophage like synovial cells by roughly 25% (p<0.01, p<0.001).
Conclusions
This study reveals that most DMARDs have effects in the "Macrophage and Lymphocyte model" whereas tocilizumab, tofacitinib and baricitinib were superior in the "FLS model" and only the two TNF inhibitors were effective in the "Osteoclast model". The findings in the "FLS model" reveals a possible beneficial effect of tocilizumab and JAK inhibitors to patients with fibroblast dominated arthritis. This study could potentially guide future studies of personalising DMARDs to treat immune mediated inflammatory arthritis.
This study explores the early identification of rheumatoid arthritis (RA) patients at elevated risk of progression. Haem-oxygenase-1 (HO-1) is a marker of oxidative stress in inflammation. Here, we investigate HO-1 as a biomarker of oxidative stress and its association with clinical disease activity and radiographic progression in RA.
Background: Apremilast (Otezla ® ) is a phosphodiesterase 4 (PDE4) inhibitor approved for the treatment of psoriasis and psoriatic arthritis (PsA), but the reason why apremilast shows clinical effect is not fully understood. The objective of this study was to study the downstream effects of apremilast on cells of inflamed joints in immune-mediated inflammatory arthritis. Methods: Synovial fluid was obtained from patients with active rheumatoid arthritis (RA), PsA or peripheral spondyloarthritis (SpA; n = 18). The in vitro models consisted of synovial fluid mononuclear cells (SFMCs) or fibroblast-like synovial cells (FLSs) cultured for 48 h, SFMCs cultured for 21 days, an osteoclast pit formation assay, and a mineralization assay. Results: In SFMCs cultured for 48 h, apremilast decreased the production of interleukin (IL)-12/IL-23p40 (the shared subunit of IL-12 and IL-23), colony-stimulating factor 1, CD6, and CD40 and increased the production of C-X-C motif chemokine 5 dose-dependently. Apremilast had a very different response signature compared with the tumor necrosis factor alpha inhibitor adalimumab with a substantially greater inhibition of IL-12/IL-23p40. In SFMCs cultured for 21 days, apremilast increased the secretion of IL-10. In FLS cultures, apremilast decreased matrix metalloproteinase-3 production. Apremilast decreased osteoclastogenesis but did not affect mineralization by human osteoblasts. Conclusion: This study reveals the downstream effects of apremilast in ex vivo models of arthritis with a strong inhibition of IL-12/IL-23p40 by SFMCs. Our findings could explain some of the efficacy of apremilast seen in IL-12/IL-23-driven immune-mediated inflammatory diseases such as psoriasis and PsA.
Background: Fibroblast-like synoviocytes (FLS) are central cellular components in persistent inflammatory joint diseases such as rheumatoid arthritis (RA). Pathological subsets of FLS have been identified from synovial tissue. However, the synovial tissue obtained from arthroplasty procedures is acquired at late disease stages and the cellular yield obtained from synovial tissue biopsies is fairly low. Collectively, challenging the robustness of human RA in vivo and in vitro models. FLS obtained from the synovial fluid (SF-FLS) are proposed as an alternative source of FLS, but a detailed phenotypical and functional characterization of FLS subsets from the synovial fluid has not been performed. Objectives: The aim of this study was to determine the phenotypical and functional characteristics of synovial fluid-derived fibroblast-like synoviocytes in rheumatoid arthritis. Methods: In the present study, paired peripheral blood mononuclear cells (PBMC) and SF-FLS from patients with RA were obtained (n=7). FLS were isolated from the synovial fluid by a strict trypsinization protocol 1 and their cellular characteristics and functionality were evaluated at passage 4. Monocultures (SF-FLS) and autologous co-cultures (SF-FLS and PBMC) were established from five patients with RA and subsequently evaluated by flow cytometry, Western blotting and multiplex immunoassays. Human cartilage-sponges (n=3) with SF-FLS and without SF-FLS (n=3) were co-implanted subcutaneously in SCID mice (n=15), mice with only cell-free human cartilage-sponges were used as controls (n=12). After 45 days, the implants were evaluated using stained sections to determine the SF-FLS invasion score based on perichondrocytic cartilage degradation. Data are expressed as median (25-75 percentile). P-values <0.05 were considered statistically significant. Results: The homogeneous subpopulations of FLS, isolated from the synovial fluid, were negative for CD34 and CD45 [98.9%, (97.5-99.7]) and positive for Thy-1 and PDPN [94.6%, (79.9-97.4]). Without stimulation, RA SF-FLS showed high and comparable levels of NFκB related pathway proteins and secreted multiple pro-inflammatory cytokines and chemokines dominated by IL-6 [2648 pg/mL, (1327-6116)] and MCP-1 [2458 pg/mL, (692-8719)]. SF-FLS increased their ICAM-1 and HLA-DR expression after encountering autologous PBMCs (p<0.01), (p<0.05). Further, SF-FLS and PBMC interacted synergistically in a co-culture model of RA and significantly increasing the secretion of several cytokines (IL-1β, IL-2, IL-6, (p<0.01)) and a chemokine (MCP-1, (p<0.01)). The invasion score of the human SF-FLS in vivo was at primary site, [1.6, (1.3-1.7)] and contralateral implantation site [1.5, (1.1-2.2)]. The invasion score of the human SF-FLS-containing implants both at primary and contralateral site were significantly higher compared with cartilage-sponges evaluated from SF-FLS-free control mice (p<0001). Conclusion: This phenotypical and functional characterization of SF-FLS, acquired and activated at the site of pathology, lays a foundation for establishing in vivo and in vitro FLS models. These FLS models will be beneficial in our understanding of the role of this cellular subset in arthritis and for characterization of drugs specifically targeting this pathological RA FLS subset. References: [1]Nielsen M. A. et al. Responses to Cytokine Inhibitors Associated with Cellular Composition in Models of Immune-Mediated Inflammatory Arthritis. ACR Open Rheumatology, 2(1):3-10. http://doi.org/10.1002/acr2.11094 Disclosure of Interests: Ditte Køster: None declared, Johanne Hovgaard Egedal: None declared, Malene Hvid: None declared, Martin Roelsgaard Jakobsen: None declared, Ulf Müller-Ladner Speakers bureau: Biogen, Bent Deleuran: None declared, Tue Wenzel Kragstrup Shareholder of: iBio Tech ApS, Consultant of: Bristol-Myers Squibb, Speakers bureau: TWK has engaged in educational activities talking about immunology in rheumatic diseases receiving speaking fees from Pfizer, Bristol-Myers Squibb, Eli Lilly, Novartis, and UCB., Elena Neumann: None declared, Morten Aagaard Nielsen: None declared
