Synovial fibroblasts spread rheumatoid arthritis to unaffected joints

Abstract

Active rheumatoid arthritis originates from few joints but subsequently affects the majority of joints. Thus far, the pathways of the progression of the disease are largely unknown. As rheumatoid arthritis synovial fibroblasts (RASFs) which can be found in RA synovium are key players in joint destruction and are able to migrate in vitro, we evaluated the potential of RASFs to spread the disease in vivo. To simulate the primary joint of origin, we implanted healthy human cartilage together with RASFs subcutaneously into severe combined immunodeficient (SCID) mice. At the contralateral flank, we implanted healthy cartilage without cells. RASFs showed an active movement to the naive cartilage via the vasculature independent of the site of application of RASFs into the SCID mouse, leading to a marked destruction of the target cartilage. These findings support the hypothesis that the characteristic clinical phenomenon of destructive arthritis spreading between joints is mediated, at least in part, by the transmigration of activated RASFs.

Figure 1. Migration of RASFs

(a) Cartilage-sponge complexes with or without RASFs were implanted into SCID mice at opposite sites. OASFs, human synovium, necrotic human, bovine or murine cartilage served as controls. (b) Invasion scores show a deep invasion of RASFs in the primary and contralateral implant. Limited cartilage invasion by OASFs or into RASF-free cartilage was observed. *p<0.05; **p<0.001; NS: not significant. (c, d) Histology shows RASF invasion and pc degradation at the primary and contralaterally implanted cartilage (red arrows: invasion; blue arrows: pc degradation; mt: murine tissue; ca: cartilage; sp: sponge). (e, f) Human origin of invading cells was confirmed by species-specific antibodies against vimentin (e) or proMMP-13 (f) (red arrows). The invasion zone was not stained by murine-specific IL-1R antibodies (g), in contrast to murine vessels (black arrows). Phase contrast shows the non-stained invasion zone (green arrows).

Figure 2. Migratory potential of RASFs

(a) Invasion scores of cartilage after s.c., i.v., i.p. injection, implantation of whole synovium, bovine cartilage and cartilage without viable chondrocytes, showing a strong invasion in contrast to OASFs and implants without RASFs. (b) Neovascularization 7 days after implantation. I: Implant; sk: murine skin; black arrows: murine vessels. (c) Migration of human RASF out of synovial tissue into RASF-free cartilage (red arrows: invasion; blue arrow: pc degradation). (d) Invasion of implanted murine articular heads was mainly observed at the cartilage-bone junction (red arrow; blue arrow: pc cartilage degradation; ca: cartilage; b: bone; sp: sponge).

Figure 3. Migration and adhesion of RASF in vitro

(a) Several RASFs (black bars) were detectable in murine organs, particularly in the spleen and ear cartilage (grey bars: single RASFs; white bars: no RASFs). (b) Immunohistochemistry for human vimentin showed RASFs in the spleen (b) but not in the intestine (c). Left: hematoxylin-eosin staining; center: vimentin-stained RASFs (black arrows); right: EGFP-transduced RASFs (white arrows). (d) After collagenase treatment and injection of RASFs, invasive erosions (arrows) were detectable in the damaged articular cartilage surface (day 39 after injection; c: cartilage; b: bone; js: joint space). (e) Increased adhesion of RASFs seeded on collagenase-treated and non-treated human cartilage in comparison to OASFs and towards collagenase-digested cartilage when compared to intact cartilage. (f) Vimentin-positive RASFs are present in the murine blood, showing a remarkably high amount of human cells in this sample (n = 2); red arrows: RASFs; green arrows: murine cells. (g) y-chromosome-specific RASF-derived sry-fragment was detectable in isolated DNA out of murine blood and in the positive control. Estimated size of the sry-fragment: 601 bp. (h) No difference in expression of integrin-subunits between invading/non-invading or primary/contralaterally implanted RASF was observed by real time-PCR after LMM.

Figure 4. RASF-transmigration and inhibition in vitro

(a) RASFs showed an increased transmigratory and invasive potential in the TEER assay (n = 5 different primary cultures) through MDCK-C7 monolayers, comparable to Cal78 cells. HSF and OASF (n = 3 different primary cultures) showed no transmigratory behavior. p<0.05 for RASFs and Cal78 vs. HSFs and OASFs, respectively. (b) Cell adhesion was inhibited by RASF-treatment with anti-VCAM-1 antibodies. The transmigratory potential of RASFs (TEER assay) was decreased by treatment or transduction of RASFs with BB-94 (c) and TIMP-3 (d). P<0.05 for RASF transduced with BB94 and TIMP-1 vs. non-transduced RASFs.