SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Objective

Rheumatoid arthritis (RA) is associated with increased production of adipokines, which are cytokine-like mediators that are produced mainly in adipose tissue but also in synovial cells. Since RA synovial fibroblasts (RASFs), lymphocytes, endothelial cells, and chondrocytes are key players in the pathophysiology of RA, this study was undertaken to analyze the effects of the key adipokine adiponectin on proinflammatory and prodestructive synovial effector cells.

Methods

Lymphocytes were activated in part prior to stimulation. All cells were stimulated with adiponectin, and changes in gene and protein expression were determined by Affymetrix and protein arrays. Messenger RNA and protein levels were confirmed using semiquantitative reverse transcription–polymerase chain reaction (PCR), real-time PCR, and immunoassays. Intracellular signal transduction was evaluated using chemical signaling inhibitors.

Results

Adiponectin stimulation of human RASFs predominantly induced the secretion of chemokines, as well as proinflammatory cytokines, prostaglandin synthases, growth factors, and factors of bone metabolism and matrix remodeling. Lymphocytes, endothelial cells, and chondrocytes responded to adiponectin stimulation with enhanced synthesis of cytokines and various chemokines. Additionally, chondrocytes released increased amounts of matrix metalloproteinases. In RASFs, adiponectin-mediated effects were p38 MAPK and protein kinase C dependent.

Conclusion

Our previous findings indicated that adiponectin was present in inflamed synovium, at sites of cartilage invasion, in lymphocyte infiltrates, and in perivascular areas. The findings of the present study indicate that adiponectin induces gene expression and protein synthesis in human RASFs, lymphocytes, endothelial cells, and chondrocytes, supporting the concept of adiponectin being involved in the pathophysiologic modulation of RA effector cells. Adiponectin promotes inflammation through cytokine synthesis, attraction of inflammatory cells to the synovium, and recruitment of prodestructive cells via chemokines, thus promoting matrix destruction at sites of cartilage invasion.

Rheumatoid arthritis (RA) is a chronic inflammatory joint disease characterized by synovial hyperplasia. Inflammation and degradation within the joints of RA patients are generally accompanied by elevated levels of cytokines, chemokines, and matrix-degrading enzymes (1–5).

Originally thought to be solely responsible for energy storage and acting as structural connective tissue for organs and padding gaps, adipose tissue is increasingly being regarded as an immunoendocrine organ (3, 6–9). In addition, cytokine-like mediators produced by the major cell type of adipose tissue, the adipocyte, have been identified and termed adipocytokines or, for short, adipokines (7). In RA, the production of adipokines is increased, especially within, but not limited to, periarticular adipose tissue (10–14). However, the detailed biologic function of adipokines in RA still needs to be fully elucidated.

In this study, we focused on the adipokine adiponectin, a C1q/tumor necrosis factor α (TNFα) homolog (15, 16), which is increased in the synovial fluid of RA patients compared with osteoarthritis (OA) patients (11). Adiponectin is the adipokine with the highest concentration in human sera and synovial fluids (in the μg/ml range) (11, 17). Adiponectin was previously thought to be secreted by adipocytes only (18), but has been shown to be expressed by other cell types, including osteoblasts and synovial cells such as RA synovial fibroblasts (RASFs) (19, 20). Besides its numerous functions in energy metabolism, adiponectin plays a role in the cardiovascular system (21) and in the immune system (22, 23).

In this study, we investigated the differential effects that adiponectin might have within the RA joint, specifically on effector cells in RA pathophysiology. RASFs, aggressive cartilage-invading cells, play a central role in RA (24, 25). They are exposed in vivo to highly increased concentrations of adiponectin in the synovial tissue and in the synovial fluid of RA patients. We therefore investigated the effects of adiponectin on the transcriptome and secretome of RASFs with special interest in proinflammatory and matrix-degrading molecules involved in the pathophysiology of RA. In previous experiments (20), we showed that adiponectin strongly induced interleukin-6 (IL-6) and pro–matrix metalloproteinase 1 (proMMP-1) in RASFs, suggesting that adiponectin has proinflammatory and prodestructive properties in RA as opposed to the antiinflammatory and antiatherogenic properties observed in cardiovascular and metabolic diseases (21, 26–28). In addition, we studied the effects of adiponectin on other cell types relevant to RA pathophysiology, including lymphocytes, chondrocytes, and endothelial cells, to further elucidate the adiponectin-mediated effects in RA joints. Adiponectin-induced signaling in RASFs was also analyzed.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Cell culture.

Primary synovial fibroblasts were cultured in Dulbecco's modified Eagle's medium (DMEM; PAA Laboratories) containing, unless stated otherwise, 10% heat-inactivated fetal bovine serum (FBS; Sigma), 12.5 mM HEPES (PAA Laboratories), and 100 units/ml of penicillin/streptomycin (PAA Laboratories), referred to hereafter as supplemented DMEM, and cultured for a maximum of 8 passages at 37°C in 10% CO2. Bovine chondrocytes were cultured in supplemented DMEM containing 2.5 μg/ml of amphotericin (Sigma). Human primary chondrocytes (PromoCell) from knee joint cartilage were cultured in Chondrocyte Growth Medium (PromoCell) containing 100 units/ml of penicillin/streptomycin at 37°C in 5% CO2. For subculturing, cells were detached using a Detach Kit (PromoCell) based on trypsin–EDTA. Jurkat cells (clone E6-1; American Type Culture Collection) were seeded into culture plates at a density of 1 × 106 cells/ml and maintained in RPMI 1640 (PAA Laboratories) containing 10% heat-inactivated FBS, 12.5 mM HEPES, and 100 units/ml of penicillin/streptomycin at 37°C in 5% CO2. Human primary lymphocytes (2 × 106 cells/ml) were cultured in RPMI 1640 containing 5% heat-inactivated autologous plasma, 12.5 mM HEPES, and 100 units/ml of penicillin/streptomycin at 37°C in 5% CO2. Cell culture medium for endothelial cells from varicose veins consisted of DMEM containing 20% heat-inactivated FBS, 12.5 mM HEPES, 100 units/ml of penicillin/streptomycin, and 100 μg/ml of endothelial cell growth supplement (BD Biosciences). For harvesting or subculturing, cells were detached using trypsin–EDTA (PAA Laboratories) unless stated otherwise.

Isolation of synovial fibroblasts.

Synovial tissue samples were obtained from synovial biopsy specimens from RA and OA patients who were undergoing joint surgery. All specimens were obtained with the approval of the Ethics Committee of the Justus-Liebig-University of Giessen. All patients gave informed consent and fulfilled the criteria of the American College of Rheumatology (formerly, the American Rheumatism Association) (29, 30). Following enzymatic digestion (31, 32), primary synovial fibroblasts were isolated and cultured in supplemented DMEM as described previously.

Stimulation of RASFs and OASFs.

RASFs and OASFs (13 total RASF populations and 12 total OASF populations) from passages 3–8 were grown to 70–80% confluency and stimulated with 25 μg/ml of human adiponectin (protein recombinantly produced in HEK 293 cells; BioVendor) for 15 hours. The stimulation time was chosen based on preliminary experiments that demonstrated optimal response after 15 hours (20). Unstimulated RASFs and OASFs were used as negative controls. Additionally, RASFs were stimulated with 25 μg/ml of adiponectin that had been digested with 1,000 μg/ml of proteinase K for 15 hours at 37°C prior to use. Proteinase K was subsequently heat-inactivated for 20 minutes at 75°C. Lipopolysaccharide (LPS) from Escherichia coli K235 (Sigma) was treated in the same manner. RASFs were stimulated with 10 ng/ml of treated and untreated LPS, respectively, as well as 0.1 ng/ml of untreated LPS. Supernatants were collected and frozen at −20°C until further evaluation. Cell lysates were prepared for RNA isolation.

Isolation of bovine chondrocytes.

Bovine chondrocytes were isolated from the humeri of adult cattle. To obtain chondrocytes for use in cell culture, bovine cartilage was digested with Pronase (0.4%; Calbiochem) for at least 90 minutes at 37°C, filtered, and washed with phosphate buffered saline (PBS). Following Pronase digestion, all subsequent steps were performed under sterile conditions. The predigested cartilage was further digested with collagenase (0.025%; Roche, Mannheim, Germany) for 15 hours at 37°C. Cell strainers with 100-μm and 40-μm nylon meshes were used to remove larger tissue particles.

Stimulation of bovine chondrocytes.

After isolation, bovine chondrocytes were grown to 70–80% confluency and stimulated with 25 μg/ml of human adiponectin for 15 hours. Unstimulated bovine chondrocytes were used as negative controls. Cell lysates were prepared for RNA isolation.

Stimulation of human chondrocytes.

Human primary chondrocytes were grown up to 70–80% confluency and stimulated with 25 μg/ml of human adiponectin for 15 hours. Unstimulated human chondrocytes were used as negative controls. Human chondrocytes were characterized by positive immunocytochemical staining for type II collagen according to the vendor's certificate of analysis.

Isolation of lymphocytes from human whole blood.

Heparinized peripheral blood was obtained from healthy adult volunteers after informed consent, and peripheral blood mononuclear cells (PBMCs) were isolated on a Ficoll 400–based (Lymphocyte Separation Medium LSM 1077; PAA Laboratories) density gradient. Residual erythrocytes were removed by hypotonic lysis using EP-Grade Cell Culture Water (PAA Laboratories). For removing monocytes or other adhering cells, 5 × 107 PBMCs resuspended in 10 ml of Iscove's modified Dulbecco's medium (PAA Laboratories) containing 2% heat-inactivated autologous plasma and 100 units/ml of penicillin/streptomycin were cultured for at least 2 hours at 37°C in 5% CO2. The cells that did not adhere to the culture flask were harvested, centrifuged, and resuspended in RPMI 1640 containing 5% heat-inactivated autologous plasma, 12.5 mM HEPES, and 100 units/ml of penicillin/streptomycin. Lymphocyte subpopulations including CD4+, CD8+, and CD19+ cells were separated in a BD FACSAria II (BD Biosciences) or a Beckman Coulter Epics Altra (Beckman Coulter). In parallel, lymphocyte subpopulations including CD3+ and CD19+ cells were isolated using the Dynabeads Untouched Human T Cells kit and B Cells kit, respectively (Invitrogen). The purity of the lymphocyte subpopulations and the composition of total lymphocytes were analyzed by flow cytometry. (Results are available from the corresponding author upon request).

Stimulation of Jurkat cells and primary human lymphocytes.

Like RASFs, Jurkat cells (clone E6-1) were stimulated with 25 μg/ml of adiponectin for 15 hours. Because of the high rate of proliferation, Jurkat cells were seeded at a density of 1 × 106 cells/ml. Since primary human lymphocytes responded more quickly and strongly to external stimuli, stimulation was performed for 6 hours with 10 μg/ml of adiponectin. Total lymphocytes were additionally activated by 100 ng/ml of LPS or left unactivated. LPS was added 1 hour before adding adiponectin. Primary lymphocytes from human whole blood as well as lymphocyte subpopulations were seeded at a density of 2 × 106 cells/ml. Supernatants were collected and frozen at −20°C until further evaluation.

Isolation of endothelial cells from human varicose veins.

Freshly extracted human varicose veins were flushed internally with sterile PBS to remove all blood from the vascular lumen. Leaking branches were sealed. One end of the vessel was clamped. The lumen was filled with collagenase H (1 mg/ml; Roche). Then, the other vessel end was clamped, and the vein was incubated for 1 hour at 37°C in a glass container filled with PBS in order to prevent drying. The endothelial cells that detached from the vessel were removed by squeezing the vessel with blunt tweezers, withdrawing the cell suspension using a pipette, and flushing the lumen twice with culture medium. The cell suspension was centrifuged, and the cells were resuspended in culture medium. Culture plates were preincubated with rat tail type I collagen (92 μg/ml; BD Biosciences) for 1 hour at 37°C. Afterward, the collagen solution was removed, culture medium was added, and the cultures were stored at 37°C in 10% CO2 for no longer than 1 day. The endothelial cells were seeded into the pretreated culture plates and grown to 100% confluency before subculturing. Endothelial cells were characterized morphologically and by positive immunocytochemical staining for CD31. (Results are available from the corresponding author upon request.)

Stimulation of human endothelial cells.

Human endothelial cells (passage 2) were grown to 90–100% confluency and stimulated with 25 μg/ml of adiponectin for 15 hours. Unstimulated endothelial cells from the same source were used as negative controls. Cell lysates were prepared for RNA isolation.

Inhibition of signal transduction in RASFs using chemical signaling inhibitors.

RASFs (in passage 7) were preincubated for 2 hours with each of the following intracellular signaling inhibitors: p38 MAPK inhibitor SB203580 (Calbiochem) at a final concentration of 20 μM and 3 μM; cell-permeable myristoylated protein kinase A (PKA) inhibitor 14–22 (Calbiochem) at a final concentration of 1 μM; cell-permeable myristoylated PKC inhibitor 20–28 (Calbiochem) at a final concentration of 40 μM; and cell-permeable NF-κB inhibitor SN50 (Calbiochem) at a final concentration of 18 μM. Following preincubation, the fibroblasts were stimulated with 25 μg/ml of human adiponectin for 15 hours. Adiponectin-stimulated fibroblasts without inhibitors and unstimulated fibroblasts incubated with or without inhibitors were used as negative controls. As control for the influence of culture conditions, RASFs (in passage 7) were stimulated for 6 hours in DMEM containing 10% FBS and, in parallel, for 15 hours in DMEM containing 2% FBS.

Affymetrix gene chips.

RASFs (in passage 5) were stimulated for 15 hours with 25 μg/ml of adiponectin. Total RNA was isolated with the RNeasy Mini Prep kit (Qiagen) including treatment with RNase-free DNase. The quality of the RNA was controlled using the Agilent 2100 Bioanalyzer (Agilent Technologies). Quality control data are provided in a supplementary file available from the corresponding author upon request. Double-stranded complementary DNA (cDNA) was synthesized from 5 μg of RNA using the SuperScript cDNA synthesis customer kit (Invitrogen). Conversion to biotin-labeled complementary RNA was performed with the BioArray HighYield RNA labeling kit (Enzo Diagnostics). The labeled RNA was hybridized on human genome U133 Plus 2.0 oligonucleotide probe arrays (Affymetrix), according to standard protocols. Data were normalized and analyzed with the GeneSpring microarray analysis software (Silicon Genetics). Filters were set to a minimum of 2-fold regulation, expression had to be >200, and flags had to be present or marginal in at least 1 of 2 compared samples.

Protein arrays.

The Human Chemokine Antibody Array I kit and a custom human antibody array (both from RayBiotech) were used to quantify human cytokines, chemokines, MMPs, tissue inhibitors of metalloproteinases, growth factors, and growth factor binding proteins secreted by RASFs. RASFs (in passage 6 for the Human Chemokine Antibody Array I and in passage 5 for the custom human antibody array) were stimulated for 15 hours with 25 μg/ml of adiponectin. The array experiments were performed according to the recommendations of the manufacturer. Chemoluminescence was detected using the VersaDoc imaging system (Bio-Rad), and spot intensities were quantified numerically with the Quantity One software (Bio-Rad). The numerical data were further evaluated with Microsoft Excel. For each protein, a fold change was calculated. The following conditions had to be fulfilled for potentially regulated proteins: mean spot intensities had to be at least twice the background level for a minimum of 1 experimental condition, and the fold changes had to be ≤−2-fold or ≥2-fold.

Real-time polymerase chain reaction (PCR).

RNA was isolated from synovial fibroblasts and endothelial cells using the RNeasy Mini Kit, according to the recommendations of the manufacturer (Qiagen). RNA was reverse-transcribed into cDNA according to a standard protocol using avian myeloblastosis virus reverse transcriptase (Promega) and random hexamer primers (Roche). After denaturation (2 minutes at 70°C) and immediate cooling down on ice, reverse transcription was performed for 30 minutes at 42°C, 30 minutes at 55°C, and 10 minutes at 70°C. Complementary DNA samples were analyzed by real-time PCR in a LightCycler (Roche) with SYBR Green I as the detection system. Real-time PCR cycling conditions were 15 minutes at 95°C, 50 cycles of 15 seconds at 95°C, 35 seconds at 55–65°C (depending on the primer pair), and 35 seconds at 72°C, and were finished using a melting curve analysis program. The reference gene for normalization was 18S ribosomal RNA (18S rRNA). Primer sequences are provided in a supplementary file available from the corresponding author upon request. The results were analyzed with the Roche LightCycler software and Microsoft Excel.

Semiquantitative RT-PCR.

RNA was isolated from cultured bovine chondrocytes as described above for real-time RT-PCR analysis of synovial fibroblasts. Complementary DNA samples were amplified for 25 cycles using Taq PCR MasterMix (Qiagen) and primer pairs for bovine sequences (Bos taurus) of IL-6, MMP-3, and RANTES. Primer sequences are provided in a supplementary file available from the corresponding author upon request. The product sizes were 208 bp for IL-6, 153 bp for MMP-3, and 189 bp for RANTES. The standard PCR cycling conditions were 3 minutes at 95°C, 25 cycles of 30 seconds at 94°C, 60 seconds at 55°C, 60 seconds at 72°C, and a final step of 10 minutes at 72°C. The reference gene was 18S rRNA. Products were analyzed by agarose gel electrophoresis.

Immunoassays.

The cytokine, chemokine, MMP, and adiponectin levels in cell culture supernatants were measured using commercially available enzyme-linked immunosorbent assays (ELISAs). Adiponectin was quantified using an ELISA from BioVendor; all other ELISAs were from R&D Systems.

Statistical analysis.

Biologic or experimental replicates were used to calculate arithmetic means and SEM. Data are presented as the mean ± SEM. In order to assess the significance of differences, a Student's 2-tailed t-test was performed on the data that were either normally distributed or belonged to normally distributed populations. P values less than 0.05 were considered significant. Statistical calculations were performed using Microsoft Excel.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Adiponectin-induced expression of mRNA for chemokines, cytokines, and matrix-degrading enzymes in cultured RASFs.

Based on our previous finding that adiponectin stimulation of RASFs induced the secretion of IL-6 and proMMP-1 (20), we investigated the effects of adiponectin on RASF gene expression in detail. Adiponectin-stimulated RASFs and unstimulated controls were analyzed using Affymetrix microarrays to allow selection of differentially expressed genes that could be used as parameters for further investigation. The Affymetrix microarray analysis showed that adiponectin induced the expression of distinct genes in RASFs. Predominantly, a large number of chemokines, but also cytokines and matrix-degrading enzymes, were induced by adiponectin (Table 1). Specifically, CCL20 (fold change +1,424), CXCL10 (fold change +161), CXCL11 (fold change +131), MMP3 (fold change +63), and MMP10 (fold change +89) were highly up-regulated.

Table 1. Analysis of gene expression at the mRNA level by Affymetrix microarray analysis and real-time PCR of functionally classified genes induced by adiponectin in RASFs*
Gene name/symbolGenBank accession numberFold change (Affymetrix)Mean ± SEM fold change (real-time PCR)
  • *

    Rheumatoid arthritis synovial fibroblasts (RASFs) that were treated with 25 μg/ml of adiponectin for 15 hours and untreated controls were analyzed by Affymetrix oligonucleotide microarrays (GeneChip HG U133A). A selection of functionally classified genes that were induced by adiponectin at the messenger RNA (mRNA) level according to the microarrays are shown. The cutoff value for fold changes was ≤−2 (repression) or ≥2 (induction); however, strongly repressed genes were rare and mainly of no known or well-defined function. For verification and further analysis of adiponectin-mediated effects, selected genes were quantified by real-time polymerase chain reaction (PCR) in a Roche LightCycler. The Affymetrix microarray results are based on 1 population of RASFs, while real-time PCR fold changes are the mean ± SEM of 4 different RASF populations (indicating the biologic variability).– = not performed.

Chemokines   
 Chemokine (C-C motif) ligand 2 (CCL2)/MCP-1NM_002982.34.811.5 ± 7.7
 Chemokine (C-C motif) ligand 5 (CCL5)/RANTESNM_002985.224.0
 Chemokine (C-C motif) ligand 7 (CCL7)/MCP-3NM_006273.2101.316.3 ± 9.1
 Chemokine (C-C motif) ligand 8 (CCL8)/MCP-2NM_005623.225.8
 Chemokine (C-C motif) ligand 20 (CCL20)/MIP-3αNM_004591.11,424.0
 Chemokine (C-X-C motif) ligand 1 (CXCL1)/GROαNM_001511.129.428.2 ± 11.5
 Chemokine (C-X-C motif) ligand 2 (CXCL2)/GROβNM_002089.333.520.3 ± 6.5
 Chemokine (C-X-C motif) ligand 3 (CXCL3)/GROγNM_002090.299.624.6 ± 7.4
 Chemokine (C-X-C motif) ligand 5 (CXCL5)/ENA-78NM_002994.337.857.2 ± 9.8
 Chemokine (C-X-C motif) ligand 6 (CXCL6)/GCP-2NM_002993.25.5112 ± 46
 Chemokine (C-X-C motif) ligand 8 (CXCL8)/IL-8NM_000584.243.4
 Chemokine (C-X-C motif) ligand 9 (CXCL9)/MIGNM_002416.19.3
 Chemokine (C-X-C motif) ligand 10 (CXCL10)/IP-10NM_001565.2161.03,803 ± 2,054
 Chemokine (C-X-C motif) ligand 11 (CXCL11)/I-TACNM_005409.3130.92,491 ± 603
Cytokines   
 Interleukin-6/IL-6NM_000600.26.4
 Interleukin-11/IL-11NM_000641.224.3
Miscellaneous inflammatory molecules   
 Prostaglandin E synthase/PTGESNM_004878.33.4
 Prostaglandin G/H synthase and cyclooxygenase 2/PTGS2NM_000963.119.9
Pre–B cell growth and B cell activation   
 Bone marrow stromal cell antigen 2/BST2NM_004335.2200.8
Receptors   
 Interleukin-7 receptor/IL7RNM_002185.25.6
 Interleukin-17 receptor B/IL17RBNM_018725.32.6
Proteinases and peptidases   
 Matrix metalloproteinase 1 (interstitial collagenase)/MMP1NM_002421.211.6
 Matrix metalloproteinase 3 (stromelysin 1 progelatinase)/MMP3NM_002422.362.5
 Matrix metalloproteinase 10 (stromelysin 2)/MMP10NM_002425.188.7
 Matrix metalloproteinase 12 (macrophage elastase)/MMP12NM_002426.349.9
Bone metabolism   
 Stanniocalcin 1/STC1NM_003155.219.88.8 ± 4.1
 Stanniocalcin 2/STC2NM_003714.22.8
Growth factors   
 Fibroblast growth factor 10/FGF10NM_004465.15.0
 Fibroblast growth factor 13/FGF13NM_004114.23.4

The results of the Affymetrix microarray analysis for genes that were strongly regulated and were of interest for further analysis were verified by real-time PCR (Table 1). The fold changes as determined by real-time PCR differed from the fold changes as determined by microarray analysis, but all of them corresponded to increases in gene expression. For the chemokines growth-related oncogene α (GROα; CXCL1), GROβ (CXCL2), GROγ (CXCL3), epithelial neutrophil–activating peptide 78 (ENA-78; CXCL5), and granulocyte chemotactic protein 2 (GCP-2; CXCL6), we also investigated whether there were notable differences in the adiponectin-mediated responses of RASFs compared with OASFs (Figure 1A, left panel). Biologic variability was examined by evaluating data from several different cell populations (n = 4–8). From these experiments, 2 key observations could be made. First, adiponectin-dependent chemokine induction was stronger in RASFs than in OASFs. Second, the biologic variability between cell populations from different patients was high.

thumbnail image

Figure 1. A, Adiponectin-mediated changes in expression of mRNA for and secretion of chemokines, cytokines, and matrix metalloproteinases (MMPs) by rheumatoid arthritis synovial fibroblasts (RASFs) and osteoarthritis synovial fibroblasts (OASFs). Gene expression in synovial fibroblasts that were stimulated with human adiponectin or left unstimulated was analyzed by real-time polymerase chain reaction and enzyme-linked immunosorbent assay (ELISA). Changes in mRNA expression and protein secretion are indicated as fold changes relative to control. Bars show the mean and SEM of multiple cell populations. ∗ = P < 0.05; ∗∗ = P < 0.01; ∗∗∗ = P < 0.001, versus unstimulated controls. B, Reduction in adiponectin-induced secretion of interleukin-6 (IL-6) by intracellular p38 MAPK and protein kinase C (PKC) signaling inhibitors. RASFs that were preincubated with p38 MAPK inhibitor SB203580 or with cell-permeable myristoylated PKC inhibitor 20–28 were stimulated with human adiponectin, appropriate controls were included, and RASFs were analyzed by ELISA. Bars show the mean ± SEM of 5 different RASF populations. C, Reduction in adiponectin-induced secretion of IL-8 by proteinase K digestion. RASFs were stimulated with undigested or digested human adiponectin or left unstimulated (n = 3). In addition, RASFs were stimulated with 10 ng/ml or 0.1 ng/ml of undigested lipopolysaccharide (LPS) or LPS treated in the same manner as digested adiponectin (n = 3). Bars show the mean and SEM. GROα = growth-related oncogene α; ENA-78 = epithelial neutrophil–activating peptide 78; GCP-2 = granulocyte chemotactic protein 2; MCP-1 = monocyte chemotactic protein 1.

Download figure to PowerPoint

Altered secretion of proteins, including chemokines, cytokines, and matrix-degrading enzymes, by adiponectin-stimulated RASFs.

The analysis of gene expression at the mRNA level revealed the regulation of several gene groups that were of interest for further analysis in additional cell types. We examined the secretion of proteins by cultured RASFs after adiponectin stimulation in supernatants from adiponectin-stimulated RASFs and unstimulated controls. Changes in protein secretion were screened using antibody-based protein arrays (n = 1 for each protein array). The group of differentially regulated proteins consisted of chemokines, cytokines, soluble IL-6 receptor (sIL-6R), MMP-3, and 2 insulin-like growth factor binding proteins (Table 2).

Table 2. Analysis of adiponectin-mediated changes in protein secretion in RASFs by antibody-based protein arrays*
Protein name/symbolFold changeArray
  • *

    Chemokine Antibody Array I (1) and a custom human antibody array (2) (both from RayBiotech) were used to analyze protein secretion in rheumatoid arthritis synovial fibroblasts (RASFs) that were treated with 25 μg/ml of adiponectin and in untreated controls. The proteins listed in the table fulfilled the following conditions: they were secreted into the cell culture supernatant at a detectable level (defined as being at least twice the background level for at least 1 experimental condition), and the concentrations changed by a factor of more than +2 or −2 according to the protein arrays. GROα = growth-related oncogene α; IP-10 = interferon-γ–inducible 10-kd protein; MCP-1 = monocyte chemotactic protein 1; I-TAC = interferon-inducible T cell α chemoattractant; IL-8 = interleukin-8; GCP-2 = granulocyte chemotactic protein 2; ENA-78 = epithelial neutrophil–activating peptide 78; NAP-2 = neutrophil-activating peptide 2; MIP-1α = macrophage inflammatory protein 1α; SDF-1β = stromal cell–derived factor 1β; TARC = thymus and activation–regulated chemokine; TECK = thymus-expressed chemokine.

Chemokines  
 Chemokine (C-X-C motif) ligand 1 (CXCL1)/GROα22.41
 Melanoma growth-stimulatory activity (MGSA)/GRO19.91
 Chemokine (C-X-C motif) ligand 10 (CXCL10)/IP-1019.11
 Chemokine (C-C motif) ligand 2 (CCL2)/MCP-111.21
 Chemokine (C-X-C motif) ligand 11 (CXCL11)/I-TAC3.11
 Chemokine (C-X-C motif) ligand 8 (CXCL8)/IL-82.72
 Chemokine (C-X-C motif) ligand 6 (CXCL6)/GCP-22.61
 Chemokine (C-X-C motif) ligand 5 (CXCL5)/ENA-782.41
 Chemokine (C-X-C motif) ligand 7 (CXCL7)/NAP-22.01
 Chemokine (C-C motif) ligand 15 (CCL15)/MIP-1α−2.11
 Chemokine (C-C motif) ligand 19 (CCL19)/MIP-3β−2.21
 Chemokine (C-X-C motif) ligand 12 (CXCL12)/SDF-1β−2.91
 Chemokine (C-C motif) ligand 17 (CCL17)/TARC−3.41
 Chemokine (C-C motif) ligand 25 (CCL25)/TECK−3.41
 Chemokine (C-C motif) ligand 28 (CCL28)/CCL28−3.81
 Chemokine (C-C motif) ligand 23 (CCL23)/Ckβ 8-1−4.51
Cytokines  
 Interleukin-6/IL-63.72
 Interleukin-7/IL-73.82
 Interleukin-10/IL-102.62
Receptors  
 Soluble interleukin-6 receptor/sIL-6R3.32
Proteinases and peptidases  
 Matrix metalloproteinase 3 (stromelysin 1 progelatinase)/MMP32.62
Growth factor binding proteins  
 Insulin-like growth factor binding protein-2/IGFBP-24.72
 Insulin-like growth factor binding protein-3/IGFBP-33.22

The results of the Affymetrix microarrays and the protein arrays for selected proteins that were relevant for further analysis in additional cell types were verified by immunoassays. We studied the response of both RASFs and OASFs to adiponectin stimulation and examined the biologic variability by analyzing multiple cell populations (Table 3). Selected proteins of the major groups regulated by adiponectin are presented in Figure 1A (right panel). Adiponectin stimulation of RASFs in vitro increased the secretion of proteins, including chemokines (e.g., monocyte chemotactic protein 1 [MCP-1]), cytokines (e.g., IL-6), and MMPs (e.g., MMP-3) (Table 3). The strongest induction was observed for IL-8, which was increased by a factor of >800, which is consistent with previously published results (33).

Table 3. Analysis of adiponectin-mediated changes in protein secretion in RASFs and OASFs by immunoassay*
Protein name/symbolMean ± SEM fold changeP
  • *

    Levels of proteins secreted by RASFs and osteoarthritis synovial fibroblasts (OASFs) upon adiponectin stimulation were quantified by commercially available immunoassays (R&D Systems) using the cell culture supernatants. The fold changes in protein concentration between stimulated and unstimulated cell cultures were calculated for multiple cell populations. See Table 2 for other definitions.

Chemokines  
 Chemokine (C-X-C motif) ligand 1 (CXCL1)/GROα  
  RASFs (n = 9)125.1 ± 43.00.024
  OASFs (n = 8)21.5 ± 7.30.027
 Chemokine (C-X-C motif) ligand 5 (CXCL5)/ENA-78  
  RASFs (n = 13)22.5 ± 6.20.005
  OASFs (n = 8)6.9 ± 2.10.023
 Chemokine (C-X-C motif) ligand 6 (CXCL6)/GCP-2  
  RASFs (n = 9)77.6 ± 36.50.069
  OASFs (n = 8)18.2 ± 6.70.037
 Chemokine (C-X-C motif) ligand 8 (CXCL8)/IL-8  
  RASFs (n = 8)805.8 ± 231.60.0103
  OASFs (n = 8)54.6 ± 23.20.054
 Chemokine (C-C motif) ligand 2 (CCL2)/MCP-1  
  RASFs (n = 13)15.8 ± 3.60.001
  OASFs (n = 12)8.0 ± 0.90.00003
 Chemokine (C-C motif) ligand 5 (CCL5)/RANTES  
  RASFs (n = 8)44.4 ± 25.10.127
  OASFs (n = 4)4.2 ± 2.40.282
Cytokines  
 Activin A/inhibin βA  
  RASFs (n = 9)15.1 ± 6.00.047
  OASFs (n = 8)4.5 ± 0.70.002
 Interleukin-6/IL-6  
  RASFs (n = 8)45.3 ± 11.70.007
  OASFs (n = 8)12.3 ± 2.60.003
Proteinases and peptidases  
 Matrix metalloproteinase 1, propeptide/proMMP-1  
  RASFs (n = 10)15.4 ± 10.40.202
  OASFs (n = 8)3.6 ± 1.60.139
 Matrix metalloproteinase 3/MMP-3  
  RASFs (n = 10)10.3 ± 3.50.025
  OASFs (n = 8)3.7 ± 1.00.035
 Matrix metalloproteinase 10/MMP-10  
  RASFs (n = 6)6.6 ± 1.40.009
  OASFs (n = 4)1.7 ± 0.20.045

The following proteins of interest that were not shown to be regulated by adiponectin in the screening assays (Affymetrix microarray and protein array) were measured by ELISA but did not exceed the detection level of the respective immunoassay with or without adiponectin stimulation: interferon-γ (IFNγ; <15.6 pg/ml), IL-1β (<3.9 pg/ml), IL-12 (<7.8 pg/ml), IL-16 (<31.2 pg/ml), and TNFα (< 15.6 pg/ml). The following proteins were found to be regulated by adiponectin in the screening arrays but were not detectable by immunoassays: sIL-6R (<31.2 pg/ml), IL-10 (<7.8 pg/ml), interferon-inducible T cell α chemoattractant (I-TAC; <62.5 pg/ml), and macrophage inflammatory protein 3α (MIP-3α; <7.8 pg/ml).

Although transforming growth factor β1 (TGFβ1) and vascular endothelial growth factor (VEGF) were found to be secreted at low levels in RASFs, no changes in protein secretion were detected upon adiponectin stimulation (∼0.9 ng/ml for TGFβ1 and ∼0.2 ng/ml for VEGF in both stimulated and unstimulated RASFs). Immunoassay analysis of cell lysates from adiponectin-stimulated RASFs showed that intracellular protein levels were not increased for I-TAC (CXCL11) or MIP-3α (CCL20). Although they were highly up-regulated at the mRNA level (Table 1), protein levels of I-TAC and MIP-3α were below the detection limit in cell culture supernatants and cell lysates. Similar to findings at the mRNA level, adiponectin stimulation increased protein secretion in OASFs but to a much lesser extent than in RASFs (Table 3).

In order to verify that the adiponectin-mediated effects were not serum-dependent, stimulation of RASFs with adiponectin was repeated under serum-free conditions using culture medium without FBS. We selected 5 proteins for analysis and found that adiponectin still induced the secretion of IL-6, IL-8, MCP-1, proMMP-1, and MMP-10. Mean fold changes for cell culture medium containing 10% FBS versus 0% FBS (where n is the number of cell populations) were as follows: 45.3 (n = 8) versus 33.8 (n = 1) for IL-6; 805.8 (n = 8) versus 1,087.0 (n = 1) for IL-8; 15.8 (n = 13) versus 13.2 (n = 1) for MCP-1; 15.4 (n = 10) versus 5.9 (n = 1) for proMMP1; and 6.6 (n = 6) versus 15.6 (n = 1) for MMP-10.

Regulation of adiponectin-induced intracellular effects by p38 MAPK and PKC.

In order to analyze adiponectin-induced signaling in RASFs, the following 4 signaling inhibitors were selected to inhibit important cellular signaling pathways: p38 MAPK inhibitor SB203580, cell-permeable myristoylated PKA inhibitor 14–22, cell-permeable myristoylated PKC inhibitor 20–28, and cell-permeable NF-κB inhibitor SN50. Five different RASF populations were analyzed. IL-6 secretion, which was strongly up-regulated by adiponectin, was significantly down-regulated by the PKC inhibitor (Figure 1B). As shown previously by both our group (20) and Tang et al (34), inhibition of p38 MAPK significantly down-regulated IL-6 secretion. Results of inhibition with 20 μM SB203580 are shown in Figure 1B (n = 5). At 3 μM SB203580, IL-6 was still notably down-regulated by a factor of 2.5 (n = 1; data not shown).

In addition, adiponectin-induced MCP-1 secretion was down-regulated by p38 MAPK and PKC inhibitors; the differences, however, were not statistically significant due to the high biologic variability within RASF populations (n = 5; data not shown). No changes were observed for PKA and NF-κB inhibitors (data not shown). Basal levels of IL-6 and MCP-1 were very low and remained unchanged by the inhibitors. To test for FBS-mediated effects, the experiment was performed in parallel with DMEM containing 2% FBS instead of 10% FBS (n = 2). The effects remained unchanged under these conditions (data not shown).

Mediation of the induction of protein secretion in RASFs by adiponectin but not by LPS.

Human adiponectin was efficiently digested using the nonspecific protease proteinase K. Culture supernatants of RASFs (n = 3) that were stimulated with 25 μg/ml of adiponectin still had adiponectin concentrations in the μg range (21.9 μg/ml) after 15 hours of incubation, whereas digested adiponectin could only be detected at a concentration of <0.2 μg/ml, and unstimulated RASFs did not produce any detectable adiponectin (data not shown). IL-8 secretion increased from 25 pg/ml to 7,337 pg/ml in RASFs stimulated with undigested adiponectin, while IL-8 levels remained as low as 98 pg/ml in RASFs stimulated with digested adiponectin. Stimulation with 10 ng/ml of LPS that was treated in the same manner as digested adiponectin or was left untreated resulted in an increase in IL-8 secretion to 14,309 pg/ml and 14,966 pg/ml, respectively; stimulation with 0.1 ng/ml of treated or untreated LPS resulted in IL-8 levels of 67 pg/ml and 69 pg/ml (Figure 1C). Analogous results were obtained for IL-6 levels. (Results for IL-6 are available from the corresponding author upon request.)

Enhanced expression of chemokines, IL-6, and matrix-degrading enzymes in adiponectin-stimulated chondrocytes.

Adiponectin stimulation of cultured bovine chondrocytes induced the expression of mRNA for IL-6, RANTES, and MMP-3. Unstimulated bovine chondrocytes showed no detectable expression of mRNA for IL-6, RANTES, or MMP-3 (Figure 2A).

thumbnail image

Figure 2. A, Induction of IL-6, RANTES, and MMP-3 mRNA expression in bovine chondrocytes by adiponectin. Cultured bovine chondrocytes were stimulated with human adiponectin (+) or left unstimulated (−). The expression of mRNA for IL-6, RANTES, and MMP-3 was analyzed by semiquantitative reverse transcription–polymerase chain reaction (RT-PCR) and subsequent agarose gel electrophoresis. (PCR amplicon sizes were 208 bp for IL-6, 153 bp for MMP-3, 189 bp for RANTES, and 187 bp for 18S ribosomal RNA [18S rRNA].) M = marker (DNA ladder). B, Adiponectin-induced secretion of IL-6, GROα, IL-8, MCP-1, proMMP-1, and MMP-3 by human chondrocytes. Cultured human chondrocytes (PromoCell) were stimulated with human adiponectin (solid bars) or left unstimulated (open bars). Cell culture supernatants were collected and analyzed by ELISA (n = 5). Bars show the mean ± SEM. ∗∗ = P < 0.01; ∗∗∗ = P < 0.001. See Figure 1 for other definitions.

Download figure to PowerPoint

Since bovine chondrocytes responded to adiponectin by increased production of proinflammatory and prodestructive factors, human chondrocytes were stimulated with adiponectin. Cultured human chondrocytes (n = 5) that were stimulated with adiponectin did not produce detectable amounts of MMP-9 (<0.312 ng/ml), MMP-10 (<78.1 pg/ml), platelet-derived growth factor BB (<31.2 pg/ml), RANTES (<31.2 pg/ml), or TNFα (<15.6 pg/ml). However, levels of IL-6 and IL-8 were strongly up-regulated by adiponectin (21-fold for IL-6 and 10-fold for IL-8) (Figure 2B). Adiponectin also significantly stimulated the secretion of other proteins (Figure 2B) but to a lesser degree, particularly the chemokines GROα (3.7-fold) and MCP-1 (2.2-fold), and the matrix-degrading enzymes proMMP-1 (2.1-fold) and MMP-3 (2.5-fold). Lago et al (35) found similar results with murine chondrocytes for the proteins IL-6, MMP-3, and MCP-1.

Increased secretion of TNFα, IL-6, IL-8, and RANTES by adiponectin-stimulated primary human lymphocytes.

Immunoassay analysis of culture supernatants showed that neither unstimulated nor adiponectin-stimulated Jurkat cells produced any detectable IFNγ (<15.6 pg/ml), IL-1β (<3.9 pg/ml), IL-2 (<31.2 pg/ml), IL-6 (<3.12 pg/ml), IL-8 (<31.2 pg/ml), IL-10 (<7.8 pg/ml), MCP-1 (<31.2 pg/ml), RANTES (<31.2 pg/ml), TGFβ1 (<31.2 pg/ml), or TNFα (<15.6 pg/ml). Since cell lines often do not respond to factors in the same manner as primary cells, human lymphocytes were isolated for stimulation with adiponectin. Secretion of TNFα, IL-6, and IL-8 by primary human lymphocytes (n = 3 for stimulated lymphocytes and n = 4 for unstimulated lymphocytes) was strongly up-regulated by adiponectin (295 fold for TNFα, 143 fold for IL-6, and 4.5 fold for IL-8) (Figure 3A). In preactivated lymphocytes, however, we observed no significant effect of adiponectin on IL-8 secretion. Secretion of TNFα and IL-6 was still increased in preactivated lymphocytes, although not to the same degree as in unactivated lymphocytes (2.8-fold up-regulation for both TNFα and IL-6). RANTES, however, was significantly up-regulated in activated lymphocytes (1.9-fold) but not in unactivated lymphocytes (Figure 3A). There were no significant changes in the secretion of IFNγ, IL-1β, IL-10, or MCP-1 (data not shown). Neither unstimulated nor adiponectin-stimulated lymphocytes produced detectable amounts of IL-2 (<31.2 pg/ml) or IL-16 (<31.2 pg/ml).

thumbnail image

Figure 3. A, Adiponectin-induced secretion of tumor necrosis factor α (TNFα), IL-6, IL-8, and RANTES by primary human lymphocytes. Unactivated primary human lymphocytes and LPS-activated primary human lymphocytes were stimulated with human adiponectin (solid bars) or left unstimulated (open bars). Cell culture supernatants were collected and analyzed by ELISA. Bars show the mean and SEM (n = 3 for stimulated lymphocytes and n = 4 for unstimulated lymphocytes). B, Adiponectin-induced secretion of IL-6, GROα, IL-8, MCP-1, and RANTES by primary human endothelial cells from varicose veins. Cultured primary human endothelial cells from varicose veins were stimulated with human adiponectin (solid bars) or left unstimulated (open bars). Cell culture supernatants were collected and analyzed by ELISA (n = 5). Bars show the mean ± SEM. ∗ = P < 0.05; ∗∗ = P < 0.01; ∗∗∗ = P < 0.001. NS = not significant. C, Adiponectin-mediated up-regulation of intercellular adhesion molecule 1 (ICAM-1) and vascular cell adhesion molecule 1 (VCAM-1) mRNA expression in primary human endothelial cells from varicose veins. Cultured primary human endothelial cells from varicose veins were stimulated with adiponectin (solid bars) or left unstimulated (open bars). Expression of mRNA was analyzed by real-time polymerase chain reaction. The control (open bars) was set to a value of 1 (i.e., mRNA levels relative to control indicate fold changes in mRNA expression; n = 2). See Figure 1 for other definitions.

Download figure to PowerPoint

Lymphocyte subpopulations obtained by negative magnetic bead selection (CD3+ and CD19+) or by fluorescence-activated cell sorting (FACS; CD4+, CD8+, and CD19+) were also stimulated with adiponectin (n = 3) and analyzed for TNFα, IL-6, IL-8, and RANTES secretion. CD8+ T cells isolated by FACS and CD19+ B cells isolated by either method did not respond to adiponectin stimulation and produced only low absolute levels of TNFα, IL-6, IL-8, and RANTES, which were either below or within the range of the levels detected in unstimulated total lymphocytes (data not shown). Adiponectin induced the secretion of TNFα, IL-6, and IL-8 in CD3+ T lymphocytes as well as in CD4+ T lymphocytes, yet the amounts produced by these subpopulations were much lower than the amounts produced by adiponectin-stimulated total lymphocytes (pg/ml range versus ng/ml range). (Data are available from the corresponding author upon request). The changes in RANTES production were only minimal for both CD3+ and CD4+ lymphocytes.

Up-regulation of the secretion of IL-6, IL-8, GROα, MCP-1, and RANTES and expression of mRNA for intercellular adhesion molecule 1 (ICAM-1) and vascular cell adhesion molecule 1 (VCAM-1) in adiponectin-stimulated human macrovascular endothelial cells.

In endothelial cells isolated from varicose veins, adiponectin induced a significantly increased secretion of IL-6 (30-fold), IL-8 (9.8-fold), GROα (8.5-fold), MCP-1 (7.1-fold), and RANTES (9.7-fold) (n = 5) (Figure 3B). The expression of mRNA for ICAM-1 and VCAM-1 was up-regulated by a factor of 36.4 and 18.8, respectively (Figure 3C). In contrast, endothelial cells stimulated with adiponectin did not produce detectable levels of angiogenin (<78.1 pg/ml), sE-selectin (<0.125 ng/ml), sVCAM-1 (<6.25 ng/ml), VEGF (<15.6 pg/ml), MMP-9 (<0.312 ng/ml), or ENA-78 (<31.2 pg/ml). Although endothelial cells secreted human endothelin 1 (∼0.8 ng/ml), proMMP-1 (∼5 ng/ml), and TGFβ1 (∼20 ng/ml), adiponectin stimulation did not modulate the protein levels.

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

In the present study, we closely analyzed the effect of adiponectin on the gene expression profile of RASFs (20) and investigated the effect of adiponectin on other key cell types involved in the pathogenesis of RA, namely, lymphocytes, chondrocytes, and endothelial cells. The aggressive phenotype of RASFs is characterized by driving inflammation, invasion, degradation, and migration (25, 36). The process of inflammation, primarily within the synovial tissue, is characterized by the recruitment of inflammatory cells and the release of proinflammatory mediators and chemokines. MMPs are key factors in perpetuating invasion into and degradation of articular cartilage (5, 37, 38). Chemokines cause migration of further inflammatory cells and RASFs into the synovial tissue and toward the invasion zone. The endogenously activated RASFs (39) are further stimulated by inflammatory immune cells and numerous cytokines.

In this study, we showed that adiponectin stimulated the synthesis of chemokines in particular, but also of cytokines, MMPs, genes involved in inflammation and bone metabolism, and receptors and growth factors in RASFs. Adiponectin, which is found at high concentrations in the synovial fluid of RA patients (11, 17), promotes proinflammatory and prodestructive pathways in RA as opposed to its “protective role” in metabolic and cardiovascular diseases (21, 26–28). Numerous changes were observed in RASFs upon adiponectin stimulation. Proinflammatory factors, including IL-6 and IL-11 (40, 41), were strongly induced in RASFs. IL-11, like IL-6, belongs to a cytokine family that shares the cytokine receptor subunit gp130. Soluble IL-6R forms a complex with membrane-bound gp130 and thus enables cells that do not express IL-6R themselves to respond to IL-6. It is hence assumed to act as an IL-6 agonist (41, 42). Activin A induces cell proliferation of RASFs (43) and is also involved in inflammatory processes (32, 44).

A very important group of genes and proteins that are regulated by adiponectin is the group of MMPs, specifically, MMP-1, MMP-3, MMP-10, and MMP-12. Protein secretion of proMMP-1, MMP-3, and MMP-10, as well as levels of MMP-12 mRNA, were considerably increased in RASFs by adiponectin stimulation. All of these MMPs are known to be involved in cartilage degradation in inflammatory arthritis (5, 37, 38). The direct involvement of MMPs in the aggressive phenotype of RASFs has been shown in previous studies and results in the invasion and destruction of adjacent cartilage by RASFs (1, 45).

Another crucial ontologic group that was highly regulated in RASFs by adiponectin consisted of chemokines. The chemokines GROα, ENA-78, GCP-2, IL-8, and MCP-1, which were produced by RASFs upon adiponectin stimulation, attract inflammatory cells to the synovium to further increase inflammation (4). Besides being a chemoattractant, IL-8 is a potent angiogenic factor (46–48). Of note, distinct chemokines that appeared to be highly regulated at the mRNA level, such as MIP-3α and I-TAC, were not secreted by RASFs, suggesting a posttranscriptional regulation mechanism. Cellular lysates of adiponectin-stimulated and unstimulated RASFs did not show detectable amounts of MIP-3α and I-TAC either, which excluded the possibility that the proteins are not secreted but instead retained within the cell.

Bone marrow stromal cell antigen 2 (BST2) has been suggested to be involved in pre–B cell growth and is expressed on RA-derived synovial cell lines and other fibroblast cell lines (49). However, its specific function has not yet been determined. Interestingly, adiponectin induced strong expression of BST2 in RASFs, underlining the potential pathophysiologic role in RA. Up-regulated fibroblast growth factors 10 and 13 may contribute to RASF proliferation.

Due to concerns that LPS contamination found in recombinantly produced proteins may contribute to the effects observed for adiponectin, we performed experiments to verify that the effects were adiponectin mediated. LPS, a polysaccharide, is unaffected by proteases, in contrast to adiponectin. Therefore, any effect left after proteolytic digestion of adiponectin can be attributed to LPS, provided that the digestion was effective, which we successfully confirmed. The abrogation of the adiponectin-induced IL-8 and IL-6 secretion by RASFs when using digested adiponectin showed that the effects were actually adiponectin mediated. Both treated and untreated LPS had a strong effect on IL-8 secretion by RASFs, demonstrating that LPS activity is not destroyed by the digestion procedure. Of note, the low LPS concentration (0.1 mg/ml) usually expected in recombinant proteins containing minor LPS contaminations caused only a minimal increase in IL-8 secretion, suggesting that the adiponectin used was not significantly contaminated.

Induction of gene expression and protein secretion by adiponectin is not restricted to RASFs. We showed that other central cell types involved in RA pathogenesis also respond to adiponectin. Factors that are induced in RASFs are similarly induced upon adiponectin stimulation in bovine and human chondrocytes. IL-6, RANTES, and MMP-3 were induced in bovine chondrocytes, and secretion of IL-6, IL-8, GROα, MCP-1, proMMP-1, and MMP-3 was increased in human chondrocytes. In primary human lymphocytes, adiponectin stimulation resulted in the synthesis of TNFα, IL-6, IL-8, and RANTES, indicating that the currently successfully targeted cytokines TNFα and IL-6 are at least in part regulated by adiponectin. TNFα, IL-6, and IL-8 contribute to the initiation and perpetuation of immune responses by activation of the cells of the immune system. RANTES secretion was up-regulated in activated but not in unactivated lymphocytes. Activation of lymphocytes may “unlock” cellular signaling pathways and thus allow adiponectin to affect the secretion of RANTES, a chemokine responsible for the recruitment of additional lymphocytes to the inflamed tissue.

Interestingly, none of the selected CD3+, CD4+, CD8+, and CD19+ lymphocyte subpopulations responded to adiponectin as strongly as total lymphocytes (at equal cell densities). CD8+ and CD19+ cells did not respond at all to stimulation with adiponectin, while CD3+ and CD4+ cells responded by significant up-regulation of TNFα, IL-6, and IL-8. The interaction of lymphocytes and/or further subpopulations in unseparated lymphocytes is most likely responsible for these findings. Endothelial cells that were stimulated with adiponectin displayed enhanced secretion of IL-6, IL-8, GROα, MCP-1, and RANTES. Also, mRNA for adhesion molecules and for the endothelial cell activation markers ICAM-1 and VCAM-1 were up-regulated. The adiponectin-induced changes may contribute to endothelial activation, leukodiapedesis, thus increasing the transmigration of leukocytes from the blood into RA synovium.

The aforementioned results indicate that adiponectin-induced chemokines from RASFs, chondrocytes, and lymphocytes contribute to the activation and recruitment of inflammatory cells into the synovial tissue. Adiponectin is predominantly expressed in the lining layer and sites of synovial invasion as well in the perivascular area and around leukocyte infiltrates (20). The additional adiponectin-induced chemokine expression around vessels and in leukocyte infiltrates may increase the influx of inflammatory cells into the synovium and recruit additional RASFs to the sites of invasion. Here, adiponectin is strongly expressed, and may increase production of MMPs by RASFs as well as chondrocytes, contributing to the perpetuation of inflammation and to its becoming chronic as well as to matrix destruction in RA.

Since adiponectin levels are generally higher in RA than in OA (20), we were interested in whether adiponectin contributes to the aggressive cellular phenotype of RASFs (25, 50). OASFs, which do not display such an activated phenotype, were used as a control. For all analyzed parameters, RASFs revealed a stronger response to adiponectin than did OASFs. Therefore, adiponectin probably plays a more significant role in RA than in OA, especially with regard to the chronic inflammation and the aggressive phenotype of RASFs.

Further analyses showed that p38 MAPK (20) and PKC were involved in adiponectin-mediated signaling in RASFs, in contrast to PKA and NF-κB. Hence, at least for IL-6 and MCP-1, the signaling molecules p38 MAPK and PKC are very likely to be part of adiponectin-induced signaling.

Our data show that adiponectin significantly affects the gene and protein expression profiles of RASFs, lymphocytes, chondrocytes, and endothelial cells in a manner that promotes inflammation and matrix destruction in RA. These findings indicate that adiponectin appears to be a major player in the pathogenesis of RA.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Neumann had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Frommer, Schäffler, Gay, Müller-Ladner, Neumann.

Acquisition of data. Frommer, Zimmermann, Meier, Schröder, Heil, Büchler, Steinmeyer, Brentano, Neumann.

Analysis and interpretation of data. Frommer, Brentano, Gay, Müller-Ladner, Neumann.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

We thank Ümit Gürler and Rosel Engel as well as the Cell Sorting Core Facility (University Hospital Giessen and Marburg) for their excellent technical assistance and help. We are grateful to Massimilano Vasile and Carina Schreyäck for donating their blood, and to Matthias Geyer and Robert Dinser for obtaining the blood samples.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES