Slit3 inhibits Robo3-induced invasion of synovial fibroblasts in rheumatoid arthritis

Synovial tissue samples were obtained from synovectomy and arthroplastic surgery from patients with RA (RASF) or from trauma surgery patients (normal SF) after informed consent and approval of the local ethics committee. All RA patients fulfilled the American College of Rheumatology 1987 criteria for the diagnosis of RA. In RA patients, material was sampled from the wrist or proximal interphalangeal joints with the joints exhibiting florid synovitis and/or arthritic destructions. Normal SF deriving from normal synovial tissue was taken from knee joints of two patients within two to four hours after knee injury. Both patients underwent surgery because of displaced tibial plateau fractures. Synovial tissue was minced mechanically, washed extensively in sterile PBS and digested with 150 mg/ml Dispase II (Boehringer Mannheim, Mannheim, Germany) for one hour at 37°C under continuous agitation. The resulting cell suspension was seeded into tissue culture dishes and cultured in DMEM (Gibco Life Technologies, Basel, Switzerland) containing 10% FCS and 100 U/ml penicillin per 100 μg/ml streptomycin in a humidified atmosphere at 37°C followed by the addition of 5% carbon dioxide. In different passages, adhering fibroblasts were washed, trypsinized, and used for RNA isolation or in the assays.

Normal dermal fibroblasts were obtained and used as described previously [12].

The number of donors used in each experiment is given in each figure legend.

Primary fibroblasts were transfected using the Amaxa Nucleofector System (Amaxa, GmbH, Cologne, Germany) with the Amaxa Normal Human Dermal Fibroblast Nucleofector^® Kit according to the manufacturer's instructions (program: U23) [13]. The transfection efficiency was about 30%, toxicity was 20%.

For analysing Robo3 effects, synovial cells in late passages (P5 and P6) or dermal fibroblasts were transiently transfected with mouse Robo3 expression construct, which was a gift from Shyng-Shiou Yuan [14]. Vector pCMX [13] was used as control. Transfection efficiency was quantified by quantitative real-time (RT)-PCR.

Genome-wide data from a recently published screen for polymorphisms associated with RA were employed to survey the complete ROBO3 gene region in silico [15]. All successfully genotyped SNPs on the Affymetrix GeneChip^® Human Mapping 500K Array Set (Santa clara, CA, USA) on chromosome 11 in the ROBO3 region with approximately 40 kb on either side of the gene (position 124,200,001 and 124,300,000; human genome build 18) were included. Wellcome Trust Case Control Consortium (WTCCC) data were accessed on by 26 June 2007 [15, 16]. Data of the HapMap phase II release 22 data were used to assess linkage disequilibrium patterns of the ROBO3 gene region [17].

Total cellular RNA was isolated from cultured cells using the RNeasy kit (QIAGEN, Hilden, Germany) and cDNAs were generated by reverse transcriptase reaction as described previously [18].

Quantitative RT-PCR was performed on a LightCycler (Roche, Mannheim, Germany) using 10 μl N', N'-dimethyl-N-[4-[(E)-(3-methyl-1,3-benzothiazol-2-ylidene)methyl]-1-phenylquinolin-1-ium-2-yl]-N-propylpropane-1,3-diamine (SYBR) MIX (TaKaRa, Shiga, Japan), 0.5 μl (20 μM) of forward and reverse primers and 1 or 2 μl cDNA template in a total of 20 μl. cDNA fragments of Robo1, Robo2, Robo3, Slit1, Slit2, and Slit3 in SF in early passages (P3 and P4) were amplified according to the following PCR program: 30 seconds at 95°C (initial denaturation); 20°C/second temperature transition rate up to 95°C for 5 seconds, 10 seconds at 55°C (Robo1, Robo2, Slit2)/56°C (Slit1, Slit3)/58°C (Robo3), 15 seconds at 72°C, 10 seconds at 82°C (Robo1, Robo2, Slit2)/86°C (Slit1, Slit3)/85°C (Robo3) acquisition mode single, repeated for 40 times (amplification).

Matrix metalloproteinase (MMP)-1 and MMP-3 quantitative RT-PCR was performed 48 hours after transfection at 95°C for 5 seconds, 3 seconds at 68°C (MMP-1) or 62°C (MMP-3), 10 seconds at 72°C and 8 seconds at 82°C (MMP-1) or 82°C (MMP-3) acquisition mode single. Annealing temperature was optimized for each primer set and the PCR reaction was evaluated by melting curve analysis following the manufacturer's instructions and checked by electrophoresis. Beta-actin mRNA was amplified to ensure cDNA integrity and to normalize expression. Each quantitative PCR was performed at least in duplicate for two sets of RNA preparations.

The following primers were used: Robo1 (for: AGG AAG AAG ACG AAG CCG AC, rev: CGA AGA ACT AAC ACT GGA GCG), Robo2 (for: GAG ACC TCA CAA TCA CCA ACA TTC AAC, rev: CAG TAA CGC TGT ACC ATC CAC TGC), Robo3 (for: GCGCTTCTCAGTGTCTCCAAG, rev: TGGTCCCTGGAGGATGACA), Slit1 (for: ACT CGC TGG TCC TCT ATG GAA, rev: CGC AAA TGA AAG GGT TCT GGG), Slit2 (for: TGC CTT TGC CCA CCT GAG TA, rev: TGT CGC AGT GTT CAC CTA CG), Slit3 (for: TGA TGG CAA CGA GGA GAG TA, rev: ACG GCT GTT AGG TGG TTT CC) MMP1 (for: TGG ACC AAG GTC TCT GAG GGT CAA, rev: GGA TGC CAT CAA TGT CAT CCT GA) and MMP3 (for: GGC ACA ATA TGG GCA CTT TAA ATG AAG C, rev: GTC TAC ACA GAT ACA GTC ACT TGT CTG).

Proliferation was measured using the Cell Proliferation Kit II (Roche,, Mannheim, Germany) according to the supplier's instructions. It is based on a colorimetric (XTT) measured non-radioactive quantification of cell proliferation performed for four days. RASF in passage 3 and 4 were treated with recombinant mouse Slit3 (R&D Systems, Minneapolis, MN, USA) in a concentration of 0.1 μg/ml. Experiments were repeated at least twice.

Migration assay were performed using Boyden Chambers containing polycarbonate filters with 8 μm pore size (Costar, Bodenheim, Germany), as described previously [18]. Briefly, filters were coated with gelatin. The lower compartment was filled with fibroblast-conditioned medium, used as a chemo-attractant. Synovial cells in early passages (P3 and P4) and in late passages (P5 and P6; as indicated) were harvested with trypsin incubation for two minutes, resuspended in DMEM without FCS at a density of 3 × 10⁴ cells/ml and injected in the upper compartment of the chamber on the filter. After incubation at 37°C for four hours, all cells attached to the upper surface of the membrane were removed. Cells adhering to the lower surface were fixed, stained and counted. Recombinant mouse Slit3 (R&D Systems, Minneapolis, MN, USA) was added either to the upper or the lower chamber of the system in a concentration of 0.1 μg/ml. All experiments were performed in triplicates and repeated at least twice.

Cartilage was obtained from donors after informed consent and approval of the local ethics committee. Normal cartilage with no significant softening or surface fibrillation was obtained at autopsy within 48 h after death. Cartilage was kept at -80°C before use. SF (50,000 cells per attempt) in early passages (P3 and P4) and synovial cells in late passages (P5 and P6) or dermal fibroblasts with Robo3 transfection (as indicated) were incubated with pieces of cartilage (8 pieces of 2 mm³ (50 mg) per attempt) in 400 μl of medium for 10 days. After 10 days, the supernatant was collected and amounts of glycosaminoglycans (GAG) were measured using the sGAG quantitative kit (Euro-Diagnostica, Malmo, Sweden). Each attempt was measured in parallel and repeated twice with different donors of SF.

To analyse Robo3 expression on RASF, cells in early passages (P3 andP4) were detached from flasks using 5 mM EDTA in PBS. Cells were resuspended in PBS and 2.5 × 10⁵ cells per approach were fixed in 100% Methanol for 10 minutes, twice washed with PBS and permeabilized with 0.1% Tween/PBS. Cells were washed twice with PBS and incubated with anti-Robo3 at room temperature for one hour (1:50; Everest Biotech, Oxfordshire, UK) in 1% BSA/PBS. After three washing steps with PBS, the cells were stained with the Cy™5-conjugated anti-goat secondary antibody (1:100, Jackson ImmunoResearch, Suffolk, UK) in 1% BSA/PBS for 30 minutes. After three washing steps the samples were resuspended in 250 μl PBS. All steps were performed at 4°C. FACS data were analyzed using the BD FACSDiva software (Becton Dickinson, Heidelberg, Germany) and WinMDI. RASF from four donors were analyzed.

Pro- and active form of MMP1 (n = 3) and MMP3 (n = 4) were analysed with RayBio Human MMP-1 and MMP-3 ELISA Kit (Norcross, GA, USA). Supernatant of mRobo3 and control transfected synovial cells used in the GAG assay was diluted 1:50 in one times assay diluent and subsequent analysed according to the manufacturer's instructions.

Calculations were performed using the GraphPad Prism software (GraphPad software Inc, San Diego, CA, USA). All results are expressed as mean +/- standard deviation (range) or in percentage. Error bars represent standard deviation. Comparison between groups was performed using the Student's t-test.

First, we characterized the expression pattern of the receptors Robo1, Robo2, Robo3, and their ligands Slit1, Slit2, and Slit3 in RASF compared with normal SF in early passages (P3 and P4; Figure 1). Robo1 and Robo2 were almost equally expressed in normal SF and RASF, whereas Robo3 mRNA was strongly enhanced in RASF compared with SF of healthy donors.

Slit1 mRNA levels were lower than Slit2 or Slit3 mRNA levels; however, only differences in expression of Slit3 were found comparing RASF and normal SF.

Next, the potential genetic association of Robo3 and RA was analysed. The WTCCC genome-wide association study on RA [15] surveyed 15 SNPs within the ROBO3 gene region on chromosome 11 (position 124,200,001 and 124,300,000; human genome build 18). Four SNPs showed nominal association with susceptibility to RA in an additive model (Table 1). The two markers with the strongest association signal (rs11604758 and rs3923890) are in weak linkage disequilibrium (r² = 0.15).

We then focussed on the significance of the strong upregulation of Robo3 in active RASF. RASF in late passages (P5 and P6) with low Robo3 were transiently transfected using nucleofection with an mRobo3 expression construct [19]. Expression of Robo3 was confirmed by quantitative RT-PCR (data not shown). In Boyden Chamber assays, Robo3-expressing RASF revealed a significant higher migratory ability compared with mock-transfected RASF (Figure 2a).

An in vitro test system for quantifying cartilage degradation by SF was used to analyse the effect of Robo3 expression in SF on cartilage destruction (Figure 2b). Here, enhanced degradation of cartilage after Robo3 expression was observed. Additionally, even in normal, dermal fibroblasts derived from skin, expression of Robo3 resulted in a high capacity to degrade cartilage (Figure 2c).

To analyse the molecular details of Robo3 function we determined the expression of MMPs as they are known to be expressed in activated SF and play an important role in cartilage destruction. Re-expression of mRobo3 in RASF in late passages resulted in strong induction of MMP1 and MMP3 mRNA (Figure 2d) whereas MMP13 mRNA stayed unchanged. We could also confirm Induction of MMP1 and MMP3 on protein level by analysing mRobo3 transfected RASF (late passages) supernatant versus control from the GAG assay (Figure 2e).

Slits have been described to have repulsive function in the developing nervous system towards growing axons expressing Robo receptors [6]. However, there are currently no data available that indicates these repellent factors are involved in the regulation of SF activity. We, therefore, analysed the effects of Slit3 (mouse Slit3; mSlit3) on SF of patients with RA.

RASF were treated continuously with Slit and proliferation of the cells was followed for five days. Slit3 treatment resulted in a weak but non-significant inhibition of proliferation in RASF (Figure 3a).

Next, we analysed the influence of Slit3 on migration of RASF. Migration of RASF in early passages (P3 and P4) was strongly inhibited by Slit3 compared with untreated cells (Figure 3b). This effect was evident not only in the presence of Slit3 in the upper chamber with the cells, but also after adding Slit to the lower chamber. This inhibitory effect of Slit3 on RASF completely vanished in later passages (P5 and P6; Figure 3c).

Slit3 treatment had no significant effect on the migration of normal SF (P3 and P4) from healthy donors (Figure 3d).

Relative migration of RASF more than doubled compared with normal SF (both early passages; Figure 3e).

As these results hint to phenotypical changes in cell culture, we were interested in the regulation of the expression pattern of Robos and Slits over the passages (P3, P5 and P10) in RASF. In RASF, Robo1 and Robo2 expression slightly decreased from P3 to P10, whereas Robo3 expression showed strong reduction (Figure 4a). Expression of Robo3 in early passages was confirmed by FACS analysis (Figure 4b). Slit1, 2 and 3 expression data showed no significant changes between the passages (Figure 4a). Interestingly, assays on cartilage destruction using RASF in early and late passages revealed reduction of aggressiveness in higher passages which correlates with loss of Robo3 (Figure 4c).