Visfatin/Pre-B-cell Colony-enhancing Factor (PBEF), a Proinflammatory and Cell Motility-changing Factor in Rheumatoid Arthritis*

Background: The adipokine visfatin exerts proinflammatory effects in synovial fibroblasts of patients with rheumatoid arthritis (RA). Results: Visfatin induces high amounts of chemokines, creating a microenvironment of enhanced fibroblast motility. Conclusion: Visfatin is therefore very capable of contributing to the inflammatory state of RA. Significance: Enlightening visfatin pathophysiology may lead to possible therapeutic targeting in the future. Adipokines such as adiponectin and visfatin/pre-B-cell colony-enhancing factor (PBEF) have been recently shown to contribute to synovial inflammation in rheumatoid arthritis (RA). In this study, we evaluated the pathophysiological implication of visfatin/PBEF in the molecular patterns of RA synovial tissue, focusing on RA synovial fibroblasts (RASFs), key players in RA synovium. Expression of visfatin/PBEF in synovial fluid and tissue of RA patients was detected by immunoassays and immunohistochemistry. RASFs were stimulated with different concentrations of visfatin/PBEF over varying time intervals, and changes in gene expression were evaluated at the RNA and protein levels using Affymetrix array, real-time PCR, and immunoassays. The signaling pathways involved were identified. The influence of visfatin/PBEF on fibroblast motility and migration was analyzed. In RA synovium, visfatin/PBEF was predominantly expressed in the lining layer, lymphoid aggregates, and interstitial vessels. In RASFs, visfatin/PBEF induced high amounts of chemokines such as IL-8 and MCP-1, proinflammatory cytokines such as IL-6, and matrix metalloproteinases such as MMP-3. Phosphorylation of p38 MAPK was observed after visfatin/PBEF stimulation, and inhibition of p38 MAPK showed strong reduction of visfatin-induced effects. Directed as well as general fibroblast motility was increased by visfatin/PBEF-induced factors. The results of this study indicate that visfatin/PBEF is involved in synovial fibroblast activation by triggering fibroblast motility and promoting cytokine synthesis at central sites in RA synovium.

Rheumatoid arthritis (RA) 2 is a chronic polyarticular disease manifesting as painful inflammation of the synovial tissues and progressive destruction of the joints. Besides macrophages and T-and B-cells, synovial fibroblasts as resident cells are key players in mediating most of the relevant pathways in this process (1,2). Activation of RA synovial fibroblasts (RASFs) in the synovium results in the production of proinflammatory cytokines and matrix metalloproteinases (MMPs). RASFs actively perpetuate inflammation as well as matrix degradation and invasion, leading to progressive destruction of the articular cartilage and the adjacent bone, thus determining the outcome of the disease (1,3,4).
The terms adipocytokine and adipokine are used for cytokine-like molecules synthesized by adipocytes, e.g. including adiponectin, leptin, resistin, and visfatin. There is growing evidence that adipose tissue is not only a repository for triglycerides or a passive connective tissue but, in fact, is an active endocrine organ regulating energy homeostasis and metabolism (5). Moreover, adipose tissue and obesity are connected with a lowgrade state of inflammation and play a role in other chronic inflammatory disorders, including RA (6,7).
Visfatin/PBEF was originally described as a cytokine involved in early B-cell development and was later renamed visfatin due to the fact that it is secreted mainly by visceral fat (20,21). Recent studies showed that visfatin/PBEF is up-regulated in activated RASFs by inflammatory stimuli such as STAT-3-dependent IL-6 trans-signaling and poly(I-C)-mediated TLR-3 activation (8,9). In turn, visfatin/PBEF induces IL-6 in RASFs, suggesting a positive feedback mechanism due to the proinflammatory activities of this protein (9). In this study, we investigated the effects of visfatin/PBEF on RASFs in detail, with the main focus specifically on the change in gene and protein expression mediated by visfatin/PBEF, signaling pathways involved in these processes, and alterations in fibroblast motility to further clarify the role of visfatin/PBEF in RA.

Tissue Specimens and Cell Culture
During routine synovectomies, synovial tissue was obtained from patients with RA and osteoarthritis (OA). Patients met the 1987 criteria of the American College of Rheumatology classification of RA and OA (22,23). Written informed consent to use the synovium or synovial fluid for research purposes was obtained from each patient. The local ethics commission of the Justus Liebig University Giessen approved the study. RASFs were cultured as described recently (24 -26). The synovial fluid of RA patients with articular effusion was aspirated. Primary human lymphocytes were isolated by Ficoll gradient centrifugation from Buffy coat samples of healthy donors. Primary human lymphocytes were cultured in RPMI 1640 medium (PAA Laboratories) supplemented with 5% human serum (PAA Laboratories), 100 units/ml penicillin, 10 g/ml streptomycin, and 10 mM HEPES (PAA Laboratories) at 37°C and 5% CO 2 .

Stimulation Assays
Recombinant human visfatin/PBEF (subsequently referred to as visfatin/PBEF; 100 ng/ml; BioVendor R&D) was used throughout the experiments if not mentioned otherwise. To exclude the effects of stimulation-related cell division, cell counting was performed.

Affymetrix GeneChipா Expression Analysis
RASFs were stimulated with 500 ng/ml visfatin/PBEF for 15 h. RNA was extracted using the RNeasy TM miniprep kit (Qiagen). Target preparation and hybridization for the Affymetrix human genome U133 Plus 2.0 GeneChip (AFFX) were performed according to the recommended protocol. The results of the two-condition design were analyzed with Gene-Spring microarray analysis software (Silicon Genetics) to obtain increase/decrease or no-change calls.

Real-time PCR
RNA was isolated using the RNeasy TM miniprep kit. Reverse transcription was performed using avian myeloblastosis virus reverse transcriptase (Promega) and random hexamer primers (Roche Applied Science). Primers (provided in the supplemental table (No. 1)) were designed, and the efficiency of each primer pair was tested using the standard curve method (E ϭ 10 Ϫ1/slope ) considering 2.00 Ϯ 0.05 as acceptable for experiments. Real-time PCR was performed using LightCycler (Roche Applied Science) with SYBR Green I (Roche Applied Science) as the detection system. Melting curve analysis was used to confirm the specificity of amplification. 18 S RNA served as a reference gene. Results were analyzed using Light-Cycler software.

Western Blot Analysis of p38 Phosphorylation
Cell extracts were harvested in 0.5% Nonidet P-40, 1 mM EDTA, 20 mM Tris-HCl (pH 8.0), 100 mM NaCl, and 10% glycerol containing 1ϫ cOmplete protease inhibitor mixture (Roche Applied Science). Protein was separated on an SDS-10% (w/v) polyacrylamide gel and blotted onto nitrocellulose membrane (Bio-Rad). Proteins were detected by antibodies against phosphorylated and total p38 (Cell Signaling Technology). Detection was performed using horseradish peroxidase-conjugated secondary antibodies (Dako) and an enhanced chemiluminescence detection kit (ECL ϩ Western blotting detection system, GE Healthcare). All Western blots were probed for cyclophilin B (Abcam) to ensure equal loading of samples.

Enzyme Immunoassay and ELISA
Commercially available ELISAs and enzyme immunoassays (R&D Systems and Phoenix Europe) were performed in accordance with the recommended protocols. Absorption was measured at 450 nm, and data were analyzed using Magellan software (Tecan).

Inhibition of Signal Transduction Pathways
Cultured RASFs were preincubated for 1.

Immunohistochemistry
Snap-frozen RA and OA synovial tissue sections were prepared. Nonspecific binding was blocked with 2% bovine serum albumin, followed by overnight incubation in a moist chamber at 4°C with rabbit anti-human visfatin antibodies (10 g/ml; Bethyl Laboratories) (9). Endogenous peroxidase activity was blocked with 0.3% H 2 O 2 in 100% methanol. Sections were then incubated with Histofine Simple Stain MAX PO (MULTI) (mouse/rabbit primary antibodies; Nichirei Biosciences) for 30 min. Color development with 3-amino-9-ethylcarbazole substrate (Vector Laboratories) at room temperature was stopped after microscopic examination, followed by counterstaining of the nuclei using hematoxylin. Rabbit isotype-matched IgG sera (Santa Cruz Biotechnology) served as an isotype control, and mouse anti-human vimentin antibody (Dako) served as a positive control. A negative control experiment without primary antibody was carried out.

Chemotaxis
Migration of RASFs was measured using a 48-well Boyden microchemotaxis chamber (Neuro Probe, Inc.). Conditioned medium was prepared incubating RASFs for 16 h in DMEM without FCS. In contrast, stimulated media were acquired by stimulation for 16 h with 1) visfatin/PBEF (100 ng/ml), 2) visfatin/PBEF (100 ng/ml) and p38 MAPK inhibitor (20 M), and 3) p38 MAPK inhibitor (20 M), all in DMEM without FCS. In addition to using different cell populations, experimental replications were also performed. 20,000 cells in 50 l of conditioned medium were placed in the upper chamber. In the lower chamber, 30 l of conditioned medium (migration base line), medium with 10% FCS and without FCS (positive and negative controls, respectively), and stimulated media (as described above) were added. Both chambers were separated by an 8-m pore size filter (GE Osmonics). After migration for 7 h, cells on the upper filter side were scraped off. Filters were fixed and stained. Of each well, six areas were photographed, the number of cells was counted, and data were expressed as the migration index (ratio of sample to base line) (28).
Migration of lymphocytes was measured using a 24-well Transwell migration assay (Corning). A total of 10 6 cells were added to the filter. The prepared conditioned medium and stimulated media (as described above) were placed into the lower well. In the control experiment, we tested visfatin/PBEF diluted in normal DMEM and 10% FCS. Cells were allowed to migrate for 4 h at 37°C. Cells in the lower well were counted with a hemocytometer.

Scrape Motility Assay
RASFs were preincubated with increasing concentrations of visfatin/PBEF (5, 25, 50, 100, and 250 ng/ml) for 1.5 h. Beginning from the center of each well, cells were removed by scraping using a 100-l pipette tip. To visualize cell motility within the scrape line, photos of the defined cell-free scraped areas were taken every 1.5 h between 8 and 17 h after scraping (29). Cells migrating into the gap were counted, and data were expressed as a comparison of the sample to base line in percentages. In addition to using different cell populations, experimental replications were performed.

Detection of Visfatin/PBEF in RA Synovial Fluid and Synovium
The mean concentration of visfatin/PBEF in RA synovial fluid was 76.26 Ϯ 7.22 ng/ml (n ϭ 24) (Fig. 1A). Strong expression of visfatin/PBEF in RA synovium (n ϭ 3), mainly in the synovial lining layer, in lymphoid aggregates and perivascular areas was observed, indicating contribution to the inflammation of the synovium by local visfatin/PBEF production independent of adipose tissue. Visfatin/PBEF protein was expressed in OA synovium (n ϭ 2) as well but to a lesser extent and mainly in regions of interstitial vessels (Fig. 1B).  2). Visfatin/PBEF protein is shown in red, and nuclei are shown in blue (magnification ϫ400). In RA synovium, synovial fibroblasts in the lining layer showed a strong expression of visfatin/PBEF (black arrow). Small interstitial vessels (black arrowhead) and lymphoid aggregates (white arrow) expressed visfatin/PBEF protein as well. In OA synovium, minor vessels of the interstitium showed visfatin/PBEF protein expression (white arrowhead). An isotype control experiment using OA synovium was performed. ss, synovial space; ll, lining layer; sl, sublining; v, vessel; la, lymphoid aggregate; i, interstitium.

Identification and Confirmation of Genes Regulated by Visfatin/PBEF in RASFs
To elucidate the effects of visfatin/PBEF on RASFs in a broad approach, the changes in gene expression and protein production were analyzed using AFFX (n ϭ 1) as a screening method and real-time PCR (n ϭ 7-9) and ELISA (n ϭ 4 -5) for confirmation. Of note, the -fold induction was different, as expected, between the non-quantitative AFFX and the quantitative realtime PCR and ELISA due to the different stimulation concentrations (500 versus 100 ng/ml).

Establishment of Dose-dependent Relationships and Time-dependent Responses for Visfatin/PBEF Stimulation of RASFs
With regard to the variability of visfatin/PBEF concentrations in the synovial fluids of RA patients, we determined whether the concentration of visfatin/PBEF plays a critical role in the production of proinflammatory markers such as IL-6 and IL-8. We further investigated whether the incubation period of RASFs with visfatin/PBEF leads to an additional increase in or suppression of cytokine production over time.
Dose-dependent Relationships-The effect of visfatin/PBEF on IL-6 and IL-8 production of RASFs was dose-dependent. The basal concentration of IL-6 was 23.3 pg/10 3 cells, which was increased up to 477.96 pg/10 3 cells by visfatin/PBEF stimulation (100 ng/ml). IL-8 was increased from 0.72 pg/10 3 cells up to 199.49 pg/10 3 cells (Fig. 2A). For subsequent stimulation experiments, 100 ng/ml visfatin/PBEF was used, which was located in the linearly increasing section of the curve and showed a significant effect on the production of proinflammatory cytokines. Further increase in the visfatin/PBEF concentration did not result in an additional increase in cytokine production and led to a plateau in the dose dependence curve ( Fig. 2A).
Time-dependent Responses-After stimulating RASFs for 15 h with visfatin/PBEF (100 ng/ml), the basal concentration of IL-6 was increased from 14.9 pg/10 3 cells up to 106.04 pg/10 3 cells. For IL-8, the basal concentration was increased from 3.97 pg/10 3 cells up to 67.42 pg/10 3 cells. Extending the incubation period up to 48 h resulted in reduction of cytokine production, perhaps due to counter-regulation (Fig. 2B). However, 15-h incubations were used throughout the stimulation experiments to avoid secondary regulatory mechanisms taking place after longer incubation periods in vitro.

Visfatin/PBEF Influences Synovial Fibroblast and Lymphocyte Motility
Next, we analyzed whether the induction of proinflammatory factors and chemokines induced in RASFs via visfatin/ PBEF leads to a cytokine environment of increased directed (chemotaxis assay) and/or general (scrape assay) cell motility of RASFs and/or lymphocytes.
Scrape Assay-The aim was to determine whether increasing concentrations of visfatin/PBEF induce cell motility not directed toward a chemokine gradient. The contribution of cell proliferation to the repopulation of the scraped zone was minimized using serum-free medium. The findings demonstrated augmented cell motility for concentrations of visfatin/PBEF of 5, 25, and 50 ng/ml (n ϭ 3) (Fig. 4, B-D). The strongest increase in motility was found for visfatin/PBEF at 25 ng/ml (ϩ13.3 Ϯ 2.7% (p ϭ 0.038) and ϩ18.4 Ϯ 3.6 (p ϭ 0.035) compared with the base line after 8 and 15.5 h) (Fig. 4C). No significant change in fibroblast motility was seen for visfatin/PBEF at 100 and 250 ng/ml (data not shown). The effects were evened out at the latest time points due to completely filling in the scraped gap. Interestingly, visfatin/PBEF at 100 ng/ml, the concentration used for the chemotaxis experiments, showed almost no altered general cell motility.
Chemotaxis Assay-Stimulation of RASFs with visfatin/ PBEF led to increased amounts of chemokines (Table 1). Migration toward these media could be detected by analyzing RASFs and lymphocytes. The migration index for RASFs was 2.73 Ϯ 0.45 (n ϭ 2) (Fig. 5A). Supplemental experiments showed that migration to visfatin/PBEF-stimulated media could be nearly reduced to the base line (1.08 Ϯ 0.59, n ϭ 2); 60% reduction) if preincubated with the p38 MAPK inhibitor. The p38 MAPK inhibitor itself reduced to some extent the migratory potential of RASFs (0.39 Ϯ 0.31; n ϭ 2) (Fig. 5A). Simultaneous with the effects of visfatin/PBEF-induced factors on RASFs, lymphocyte migration was increased but could be reduced via p38 MAPK inhibition (n ϭ 2) (Fig. 5B). To see whether the increased migration was indeed caused by visfatin/PBEF-induced factors and not by visfatin/PBEF itself, we measured lymphocyte migration toward medium containing 100 ng/ml visfatin/PBEF and found no significant change in migration (n ϭ 2) (Fig. 5C).

DISCUSSION
In this study, we investigated the potential of visfatin/PBEF to act as an effector molecule in RA. The significant changes in gene expression of RASFs mediated by visfatin/PBEF, especially the induction of a variety of chemokines and proinflammatory and matrix-degrading factors, could be confirmed at the RNA and protein levels. This supports the hypothesis that visfatin/  PBEF creates an inflammatory molecular environment of increased fibroblast and leukocyte motility within RA synovial tissue.
Inflammation of the synovium is a hallmark in RA, and activated RASFs play a central role in local pathophysiological mechanisms. In the inflamed synovium of RA patients, accumulation of visfatin/PBEF was most dominant at the site of cartilage invasion (9), in the lining layer, in lymphoid aggregates, and around interstitial vessels, confirming that fibroblasts, lymphocytes, and endothelial cells express visfatin/PBEF in the synovial tissue and are exposed to increased concentrations of visfatin/PBEF in return (8,20,34).
The lining layer contains high numbers of activated RASFs characterized by a high basal production of IL-6 and MMPs advancing chronic inflammatory responses, contributing to Tand B-cell activation and progressive cartilage destruction (35). Visfatin/PBEF enhances these effects by increasing the production of IL-6 and MMP-3, MMP-10, MMP-12, and MMP-19. Therefore, visfatin/PBEF intensifies the aggressive phenotype of RASFs. This holds also true for visfatin-mediated stimulation of chondrocytes, resulting in higher amounts of prostaglandin E 2 and MMP-3, and CD14 ϩ monocytes, expressing higher amounts of IL-6, IL-1␤, and TNF-␣, as reported recently (12,36). By elucidating the change in gene expression of RASFs after visfatin/PBEF stimulation, we showed that a broad variety of chemokines of both the CXC and CC clusters are strongly upregulated, suggesting that visfatin/PBEF mediates chemoattraction in RA synovium to a significant extent. In addition, visfatin/PBEF elevates the expression of adhesion molecules such as VCAM-1, ICAM-1, and ICAM-2, enabling RASFs to increase attachment to cartilage but also to contribute to cell migration (1). Chemoattraction is mediated via enhanced cell motility. Our results support the idea that visfatin/PBEF operates as a cell motility-increasing molecule for RASFs in vitro in a similar way as was shown for CD14 ϩ monocytes and CD19 ϩ B-cells under other pathophysiological conditions (12). The visfatin/PBEF-induced cytokine environment exerts chemoattractive properties promoting directed fibroblast motility to the site of visfatin/PBEF expression. Furthermore, the visfatin/ PBEF-induced cytokine environment contributes to lymphocyte recruitment. Therefore, visfatin/PBEF may contribute to accumulation of RASFs in the lining layer as well as at sites of cartilage invasion (1,37), resulting in increased matrix destruction and remodeling at the invasion zone, a characteristic of hyperplastic RA synovium.
The inflammatory process in RA synovium is also driven by the influx of inflammatory cells into the synovial tissue, and activated interstitial vessels play a key role in the extravasation of leukocytes. Increased expression of visfatin/PBEF around interstitial vessels and a strong induction of IL-8, VEGF, and endothelial cell growth factor, factors with a strong pro-angiogenic effect (38,39), were observed. It is likely that visfatin/ PBEF promotes also RA synovial angiogenesis to some extent. This process has been demonstrated for visfatin/PBEF-mediated activation of ERK1/2 pathways as well as for increases in FGF-2 in human endothelial cells (40 -43).
Because IL-6 and MMP-1 production is closely linked to PKC and p38 MAPK activation (30 -32, 44 -46) and because initia- tion of common pathways such as PKA and NF-B results in inflammatory responses of RASFs (30,33), we investigated whether the observed effects were mediated via those pathways. Although it was shown that visfatin/PBEF activates the NF-B pathway in RASFs (9), the only significant effects we could detect were seen upon inhibition of the p38 MAPK pathway. Of interest, p38 MAPK inhibition in animal models of arthritis leads to improvement of the disease severity, thus reflecting its central role in arthritis (47,48).
Taken together, the results of the study show that visfatin/ PBEF is a strong effector molecule in the pathophysiology of RA, specifically by contributing to the production of proinflammatory chemokines in RASFs, matrix-degrading enzymes, and pro-angiogenic molecules. The data also support the idea that the effects of adipokines are not restricted to adipose tissue but are operative in chronic inflammatory joint diseases.