SEARCH

SEARCH BY CITATION

Abstract

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

Objective

To identify the genes up-regulated by interleukin-1 (IL-1) in combination with oncostatin M (OSM) in chondrocytes that may be involved in mechanisms of cartilage repair and degradation.

Methods

Gene microarray and real-time polymerase chain reaction (PCR) experiments were performed using RNA from SW1353 chondrocytes and primary human articular chondrocytes. Sections prepared from murine joints, injected with adenovirus vectors overexpressing IL-1 and/or OSM, were analyzed by immunohistochemistry for selected proteins.

Results

The combination of IL-1 and OSM markedly up-regulated the expression of various genes, including matrix metalloproteinases (MMPs), cytokines, chemokines, extracellular matrix components, and genes involved in signal transduction. Real-time PCR confirmed a synergistic induction of several MMPs, activin A, pentraxin 3 (PTX-3), and IL-8. The in vivo findings further indicated that stimulation with IL-1 plus OSM induced protein expression of activin A, PTX-3, and KC (the murine homolog of IL-8), as compared with the changes induced by individual cytokine treatment and unstimulated controls.

Conclusion

The results confirm that the potent proinflammatory cytokine combination of IL-1 plus OSM synergistically and coordinately up-regulates many genes and several MMPs. Moreover, chondrocytes exhibit a potential repair response following this procatabolic stimulus such that the repair mechanisms are ultimately overwhelmed by degradative processes in the cartilage. This gene-profiling study provides insight into the complex processes that mediate joint disease in the inflammatory arthritides through the coordinated expression of multiple genes.

Rheumatoid arthritis (RA) is characterized by a loss of joint function resulting from proteolytic degradation of articular cartilage. Chondrocytes synthesize and maintain the extracellular matrix (ECM) of cartilage, which is composed primarily of proteoglycan (aggrecan) and collagen. These structural components provide resistance to compressive forces and give the tissue its tensile strength. Cartilage degradation is mediated predominantly by the matrix metalloproteinases (MMPs), a family of potent enzymes that, collectively, can degrade all ECM components and that have been strongly implicated in arthritic joints (1). Aggrecanolysis is considered to be mediated by the ADAMTS proteinases (2), although this ECM component can be replaced relatively rapidly once the stimulus, such as interleukin-1 (IL-1), has been removed (3). In contrast, collagen is much less readily released, but when degradation does occur, tissue integrity is irreversibly lost (4). The collagenolytic MMPs (MMPs 1, 8, and 13) have all been implicated in pathologic collagenolysis (1) and require activation of their latent proforms via proteolytic removal of the propeptide, which can be MMP-mediated (5). Indeed, this activation has been shown to be a key step in cartilage collagenolysis (6).

We have previously shown that the combination of IL-1 and oncostatin M (OSM), cytokines known to be elevated in RA synovial fluid (7, 8), promotes the synergistic loss of collagen (and proteoglycan) from cartilage in vitro (7). Furthermore, we have also demonstrated that this combination induces a marked inflammatory arthritis in vivo, which is characterized by pronounced synovial hyperplasia, increased inflammatory infiltrate, marked cartilage and bone erosions, and elevated MMP expression (9, 10). In IL-1 plus OSM–treated human chondrocytes, the most striking observation is a pronounced induction of MMP-1 (10, 11), as well as other MMPs, ADAM proteinases, and ADAMTS proteinases (11). Traditionally, inflammation and destruction of bone and cartilage have been linked, although this may not be the case (12). These processes are complex and multifactorial. Studies to date have been restricted to a relatively small subset of those metalloproteinases considered to be important in cartilage ECM degradation, rather than focusing on the diversity of proteinases and other molecules that can be expressed by chondrocytes. Indeed, cytokine-induced cartilage catabolism is multifarious, and sequential proteolysis of ECM components can occur. It is known, for example, that aggrecanolysis is an early event, whereas collagenolysis takes place much later in the disease course (7, 13). Interestingly, aggrecan has been suggested to assist in maintenance of the ECM by protecting collagen fibrils from collagenolysis (14).

Most studies have focused on the genes that help mediate the destructive response following a procatabolic stimulus, such as that provided by IL-1 plus OSM, and it is apparent that a number of genes are likely to be regulated coordinately. Little, however, is known about the repair responses that chondrocytes may initiate following such stimuli. Herein we report the findings of microarray analyses of IL-1 plus OSM–treated chondrocytes. Our results provide new insight into the mechanisms by which this potent cytokine combination can affect the breakdown of cartilage as well as the repair responses initiated by chondrocytes.

MATERIALS AND METHODS

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

Materials.

Chemicals were obtained from the following suppliers. IL-1α was a generous gift from Dr. K. Ray (GSK, Stevenage, UK), and recombinant human OSM was from Prof. J. K. Heath (University of Birmingham, UK). All polyclonal antibodies were purchased from R&D Systems (Abingdon, UK) unless otherwise stated. Vectastain Elite ABC kits were from Vector (Burlingame, CA). Superscript II reverse transcriptase was obtained from Invitrogen (Paisley, UK). Real-time polymerase chain reaction (PCR) Master Mix reagents were obtained from Roche (Lewes, UK). All other chemicals were commercially available and of analytic grade, as described previously (7, 10).

Cell culture and RNA extraction.

Human articular cartilage was obtained with consent from 4 donors who were patients with osteoarthritis undergoing joint replacement surgery at a local hospital. Chondrocytes were isolated from the tissue by sequential proteolysis, as described previously (15). Cells were seeded into 75-cm2 tissue culture flasks (Corning/Costar, High Wycombe, UK) at 1 × 106 cells per flask and grown to 85% confluence in Dulbecco's modified Eagle's medium (DMEM) containing 25 mM HEPES supplemented with 10% fetal calf serum (FCS) (Invitrogen). Cells were then washed with Dulbecco's phosphate buffered saline (PBS); the medium was then replaced with serum-free DMEM and incubated overnight, prior to stimulation with serum-free medium containing IL-1 (0.02 ng/ml) and/or OSM (10 ng/ml).

Human chondrosarcoma cells (SW1353) were purchased from American Type Culture Collection (catalog no. HTB-94; Rockville, MD). Cells were cultured in DMEM/Ham's F-12 (Invitrogen) supplemented with 1% glutamine, 1% nonessential amino acids (Invitrogen), penicillin (100 IU/ml), and streptomycin (100 μg/ml) with 10% FCS, until 85% confluent. Cells were washed with Dulbecco's PBS and then cultured overnight in serum-free medium, prior to cytokine stimulation with IL-1 (0.2 ng/ml) and/or OSM (10 ng/ml); a higher concentration of IL-1 was used with the human chondrosarcoma cells because these types of cells are less responsive to IL-1 than primary chondrocytes (data not shown).

Serum was excluded from the stimulated chondrocytes, since it can markedly alter cell metabolism in the absence of exogenous cytokines (16). Because we were using a model representative of cartilage breakdown, we avoided using serum (6, 7, 15), which contains chondroprotective agents, such as insulin-like growth factor 1 (17), that can block cartilage collagenolysis (18). The absence of serum does not affect cell viability (data not shown), and previous studies have shown that cartilage in serum-free culture for up to 8–9 days can respond to serum and other growth factors (19).

Following stimulation, cells were lysed with RNeasy lysis buffer (Qiagen, Crawley, UK), and total cellular RNA was isolated in accordance with the manufacturer's instructions. DNA was removed using the DNase I set (Qiagen) during the isolation, and RNA was quantified using Ribogreen reagent (Molecular Probes, Eugene, OR). The quality of the RNA was assessed spectrophotometrically, and only samples with a 260 nm:280 nm ratio of higher than 1.8 were used in subsequent experiments. Isolated RNA was stored at −80°C until used.

Human cytokine macroarrays.

Panorama human cytokine arrays (Sigma-Genosys, The Woodlands, TX) were used in accordance with the manufacturer's instructions. Briefly, total RNA derived from primary human articular chondrocytes (2 μg) was reverse-transcribed using α32P-dCTP. The radiolabeled complementary DNA was denatured at 95°C for 10 minutes, added to hybridization solution (5 ml) that was prewarmed to 65°C, and kept at 65°C for 16 hours. Arrays were then washed twice with 80 ml 0.5× saline–sodium phosphate–EDTA (SSPE) (90 mM NaCl, 5 mM sodium phosphate, 0.5 mM EDTA, pH 7.7) and 1% sodium dodecyl sulfate (SDS) for 20 minutes, and washed twice with 80 ml 0.1× SSPE and 1% SDS for 20 minutes, before being exposed to Phosphor screens (Molecular Dynamics, Chesham, UK) for 16 hours. Images were visualized with a Storm 860 PhosphorImager (Molecular Dynamics), and following image acquisition, scanned arrays were analyzed using Phoretix array software (Nonlinear Dynamics, Newcastle, UK) to quantify individual spot intensities, which were normalized to the signal for GAPDH present on the same array.

Microarray analysis.

Human HG-U133 A and B arrays (Affymetrix) were probed with biotin-labeled complementary RNA prepared from 15 μg of total RNA in accordance with the manufacturer's protocol. Affymetrix Microarray Suite, version 5.0 (Affymetrix, Santa Clara, CA), was used to generate a P value for detection and to assign a present, marginal, or absent call. The array hybridized with RNA from unstimulated control cells was designated as the baseline, and this was used for comparison with the arrays of stimulated cells.

Statistical analysis of microarray data.

Significant differences in gene-expression levels obtained using the Affymetrix genechips were estimated with Wilcoxon's rank sum test. The signal log ratio was used as an estimation of the magnitude and direction of change of a transcript when individual arrays were compared. Since the log scale used was base 2, a signal log ratio of 1.0 indicated a 2-fold increase in the transcript level. A one-step Tukey's Biweight method was used to obtain ratios for each probe set.

Real-time PCR.

Oligonucleotide primers were designed using Primer Express software version 1.0 (Applied Biosystems, Warrington, UK). To prevent amplification of any residual genomic DNA present, the primers were placed within different exons close to or spanning an intron–exon boundary. The sequences of the primers used (all human) were as follows: MMP-1, forward 5′-CGACTCTAGAAACACAAGAGCAAGA-3′ and reverse 5′-TTCAACTTGCCTCCCATCATT-3′; MMP-3, forward 5′-AGTCTTCCAATCCTACTGTTGCTGTG-3′ and reverse 5′-TTCTAGATATTTCTGAACAAGGTTCATGCT-3′; MMP-8, forward 5′-TCTCCCTGAAGACGCTTCCA-3′ and reverse 5′-AGGTAGTCCTGAACAGTTTTTGTATTTTTGTC-3′; MMP-13, forward 5′-TTGCAGAGCGCTACCTGAGA-3′ and reverse 5′-TCATGGAGCTTGCTGCATTC-3′; MMP-10, forward 5′-AATGAGGTACAAGCAGGTTATCCAAAGTCTTCCAAT-3′ and reverse 5′-ACAGCTGCATCAATTTTCCTTATGCCTACTGTTGCTGTG-3′; MMP-14, forward 5′-GCCTGCGTCCATCAACACT-3′ and reverse 5′-AACACCCAATGCTTGTCTCCTTT-3′; pentraxin-3 (PTX-3), forward 5′-AGGACCCCACGCCGT-3′ and reverse 5′-CTTCGCCAGGCTTTCC-3′; activin A, forward 5′-CCGAGTCAGGAACAGCCAG-3′ and reverse 5′-ACTTTGGTCCTGGTCCTGGTCCTGTTG-3′; and IL-8, forward 5′-CTCTGTGTGAAGGTGCAGTTTTG-3′ and reverse 5′-GACAGAGCTCTCTTCCATCAGAAAG-3′.

Total RNA (1 μg) was reverse transcribed in a 20-μl reaction using 2 μg of random hexamers and Superscript II reverse transcriptase in accordance with the manufacturer's instructions. Relative quantification of gene expression was performed using the Lightcycler (Roche Diagnostics, Lewes, UK) or a 7900HT PCR system (Applied Biosystems). PCRs were performed in triplicate using 5 mM MgCl2, 2 μl HotStart SYBR Green Master Mix, and 0.5–1 μM of each primer in a 20-μl reaction. Thermocycler conditions comprised an initial activation step at 95°C for 10 minutes. This was followed by a 3-step program consisting of 95°C for 20 seconds, 60°C for 5 seconds, and 72°C for 10 seconds for 45 cycles. A 1-step melt-curve analysis was also performed at the end of the run, to ensure the crossover values obtained were due to the amplification of a specific product. The GAPDH gene was used as an endogenous control, to normalize for differences in the amount of total RNA in each sample. A 2-step program was performed to amplify GAPDH, 95°C for 20 seconds and 60°C for 20 seconds for 45 cycles, using 2 μl of HotStart hydridization Master Mix (Roche), 5 mM MgCl2, and 1 μl primer/probe mix (Applied Biosystems) in 20-μl reactions.

Arthritis model and immunohistochemistry.

An established model of inflammatory arthritis was used, in which replication-deficient adenovirus was engineered to express murine IL-1 and murine OSM at 1 × 106 plaque-forming units/joint/vector, as described previously (9, 10). At 7 days after injection of the adenovirus vectors, joints were dissected, fixed overnight in 7% formaldehyde in PBS (pH 7.4), decalcified in 10% EDTA in PBS for 10 days, and wax-embedded. Formalin-fixed paraffin sections (5 μm) were deparaffinized and rehydrated in decreasing concentrations of ethanol (99%, 95%, 70%, and 50%, in deionized water), for 3 minutes each. Antigen retrieval was performed by placing sections in 10 mM sodium citrate, pH 6.0, for 2 hours at room temperature. Sections were then placed in 3% H2O2 in Tris buffered saline (TBS) for 15 minutes. Thereafter, serial sections were blocked with 1.5% normal rabbit serum in TBS for 30 minutes and then incubated for 90 minutes at room temperature with various polyclonal primary (goat) antibodies: anti-human activin A, anti-human PTX-3, and anti-mouse KC (all at 5 μg/ml). Normal goat IgG (5 μg/ml) was used as an isotype-matched control antibody. Sections were subsequently washed twice in TBS for 5 minutes and then incubated with biotinylated secondary antibody (rabbit anti-goat IgG, diluted 50-fold in TBS according to the Vectastain kit instructions) in 1.5% rabbit serum in TBS for 30 minutes, followed by incubation with avidin–biotin complex for 30 minutes, using Vectastain kit 6105 in accordance with the manufacturer's instructions (Vector). Sections were then washed twice for 5 minutes in TBS.

Protein signals were developed using diaminobenzidine tetrahydrochloride (Dako, Ely, UK), following the manufacturer's protocol. Sections were counterstained with hematoxylin for 5 seconds and washed extensively in water. Images of stained sections were captured using a 3-CCD color video camera (JVC, Tokyo, Japan) and displayed on a computer monitor.

RESULTS

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

Variable gene expression by chondrocyte populations in response to IL-1 plus OSM.

Preliminary experiments using Panorama cytokine arrays assessed the gene expression induced by IL-1 plus OSM in 3 different populations of primary human articular chondrocytes. A degree of variability was seen in these preparations, especially for genes that were expressed at relatively low levels; in fact, definitive assignment of a synergistic induction was difficult. This variability in response may reflect differences in the basal expression levels of genes in primary chondrocytes isolated from different patients (20). We therefore analyzed the RNA preparations using real-time PCR for genes that have already been shown to be synergistically up-regulated by IL-1 plus OSM (11).

The results from analyses of 4 different human articular chondrocyte populations revealed a reproducible and synergistic induction of MMP-1, and although, as predicted, the relative levels varied, the magnitude of induction with IL-1 plus OSM relative to that with IL-1 alone was similar among all cell populations (Figure 1). These findings differed slightly from the data obtained by Panorama array in that the array did not reveal a synergistic response in all of the populations, probably due to a lack of sensitivity as compared with the results produced by real-time PCR. Despite these limitations, the Panorama cytokine array indicated that MMPs 1, 3, 8, 10, 12, and 13, as well as ADAM-10, were up-regulated in a synergistic or additive manner. Furthermore, a variety of cytokines and chemokines were also up-regulated, including monocyte chemoattractant protein 1 (MCP-1), epithelial neutrophil–activating peptide 78 (ENA-78), pre–B cell colony-enhancing factor (PBEF), and IL-8. The acute-phase protein PTX-3 was also shown to be up-regulated (Table 1 and data not shown).

thumbnail image

Figure 1. Variation in interleukin-1 (IL-1) plus oncostatin M (OSM)–induced matrix metalloproteinase 1 (MMP-1) expression levels in different human chondrocyte populations. Primary human articular chondrocytes were stimulated with medium alone (control), IL-1 (0.02 ng/ml), OSM (10 ng/ml), or IL-1 + OSM for 24 hours. Total RNA from 4 separate chondrocyte populations was isolated and subjected to real-time polymerase chain reaction for MMP-1 (AD). The data are presented relative to GAPDH. Values for the fold induction with IL-1 + OSM relative to IL-1 alone are shown above the bar for each population; the overall mean ± SEM induction of MMP-1 was 2.6 ± 0.18.

Download figure to PowerPoint

Table 1. Genes synergistically induced in chondrocytes following stimulation with IL-1 plus OSM*
GeneSignal intensitySignal log ratio vs. controlFold change vs. control
ControlIL-1OSMIL-1 + OSMIL-1OSMIL-1 + OSMIL-1OSMIL-1 + OSM
  • *

    RNA extracted from SW1353 chondrocytes (15 μg) was labeled and hybridized to the U133 microarrays. The majority of synergistically induced genes, as selected by a specific algorithm (that mimicked the profile in Figure 1), are presented. Arbitrary values for signal intensities are given for RNA isolated from control, interleukin-1 (IL-1) (0.2 ng/ml), oncostatin M (OSM) (10 ng/ml), and IL-1 plus OSM–treated cells. Gene expression was classified as present (P) at P < 0.04, marginal (M) at P = 0.04–0.06, or absent (A); signals are presented to the nearest whole number. Signal log ratios (presented to 1 decimal place) and the fold change were calculated using Microarray Suite software. MMP-1 = matrix metalloproteinase 1; SCCA-2 = squamous cell carcinoma antigen; MCP-1 = monocyte chemoattractant protein 1; LIF = leukemia inhibitory factor; OSMβR = OSM β receptor; ENA-78 = epithelial neutrophil–activating peptide 78; PBEF = pre–B cell colony-enhancing factor; PTX-3 = pentraxin 3; SOD-3 = superoxide dismutase (extracellular form).

  • These genes were also expressed on the Panorama array, and in all cases were at least additively, if not synergistically, induced by IL-1 plus OSM.

Proteases and inhibitors          
 MMP-122A1,875P28P4,217P6.10.37.367.61.3157.6
 MMP-3376P5,862P318P8,910P3.40.24.410.21.121.1
 MMP-1030A99P14A177A1.2−0.62.32.3−0.74.8
 MMP-1257A1,015P37A2,447P3.6−0.34.912.5−0.830.7
 MMP-133A280P18P1,548P5.92.08.458.94.1342.5
 MMP-1422A28A97A248P0.01.83.21.03.58.9
 Antileukopeptidase112P175P68A400P0.4−0.72.41.1−0.66.9
 SCCA-24A20A525P1,125P2.25.87.24.455.7147.0
 C1r274P1,217P509P1,912P2.21.22.94.62.37.5
Chemokines, cytokines, receptors, and signal transduction          
 IL-82A896P2A1,972P7.7−0.29.0207.9−0.9512.0
 IL-1β24A150A48A560P2.90.84.27.21.718.4
 MCP-13A175P43A285P5.03.96.332.014.478.8
 MCP-343A369P89P734P3.81.67.313.93.028.8
 IL-653A251P43A1,486P1.7−0.14.33.3−0.920.4
 LIF12A125M9A226P3.1−0.14.04.3−0.910.2
 OSMβR25P71P155P253P1.22.73.32.46.59.2
 ENA-789A10A6A217P0.6−0.24.41.5−0.320.4
 PBEF305P928P604P2,303P1.61.03.03.02.08.1
 Activin A58A90P17A125A0.7−2.31.51.8−0.72.3
 Jak 2 kinase28A88P98P371P1.41.53.22.72.89.1
Extracellular proteins          
 Decorin variant A14P64P23P126P1.80.13.32.31.218.4
 Decorin variant C98P255P99P826P1.70.03.53.31.011.3
 Fibronectin130P140P397P709P0.111.62.71.02.96.3
 Serum amyloid A25A289P9A2,288P5.150.98.985.01.1362.0
 Calcium binding protein A97A45A18A339P1.511.4338.92.82.645.3
 Calcium binding protein A833A63A19A604P0.91−0.24.01.91.716.0
 PTX-318A123P24P407P2.590.75.16.01.633.6
 Chitinase-3–like 279A215P71A1,750P1.56−0.33.92.91.014.9
 Chitinase-3–like 1243P579P1,666P2,320P1.212.73.22.46.49.2
 SOD-324P232P25A404P3.45−0.25.724.81.541.6

Due to the inherent variability of different human articular chondrocyte preparations, we used the chondrocyte cell line SW1353 for further genome-wide profiling, which allowed us to maximize the detection of genes regulated by IL-1 plus OSM. This also allowed the reproducible expansion of the large quantities of cells required to ensure sufficient isolated RNA for microarray experiments.

Synergistic induction of multiple genes in IL-1 plus OSM–stimulated chondrocytes.

Microarray analysis of RNA from IL-1 plus OSM–stimulated SW1353 chondrocytes confirmed the findings obtained with the Panorama cytokine arrays on primary human articular chondrocytes, and identified a number of genes that were synergistically regulated by this cytokine combination (Table 1). A specific clustering algorithm that mimicked the profile seen in Figure 1 was used to derive this data set. Many genes were up-regulated by IL-1 plus OSM. Each of these genes appeared to have different roles, such as a contribution to cartilage catabolism (MMPs 1, 3, 10, 12, and 13), inflammation (IL-1, IL-6, and IL-8), and cartilage repair (activin A, PTX-3, decorin, and fibronectin).

Differences in the temporal synergistic regulation of MMPs by IL-1 plus OSM in chondrocytes.

Real-time PCR using RNA from IL-1 plus OSM–treated SW1353 chondrocytes confirmed the synergistic induction of MMPs 1, 3, 10, and 13 (Figure 2A). Moreover, synergy was also observed in the induction of MMP-8, despite a lack of signal on the Affymetrix array. A maximal induction was seen at 8 hours for MMPs 3, 10, and 13, whereas MMP-8 was maximally induced at a much later time point, as previously observed (11); this late induction of MMP-8 is the most probable explanation for the lack of signal at 24 hours. A synergistic induction of MMP-14 was also observed, although this varied with respect to time (Figure 2A). In general, the magnitude of induction observed on the Affymetrix microarrays (Figure 2B) correlated well with the real-time PCR data at the same 24-hour time point (Figure 2A).

thumbnail image

Figure 2. Matrix metalloproteinase (MMP) gene-expression levels in interleukin-1 (IL-1) plus oncostatin M (OSM)–stimulated SW1353 chondrocytes. A, SW1353 cells were treated for the indicated times with medium alone (control) (diagonally hatched bars), IL-1 (0.2 ng/ml) (open bars), OSM (10 ng/ml) (horizontally hatched bars), or IL-1 + OSM (solid bars). Total RNA was isolated and analyzed by real-time polymerase chain reaction for the indicated MMPs. The data in A are presented relative to GAPDH. B, The fold induction (compared with control) of various MMPs was determined from the RNA isolated at 24 hours and hybridized to the Affymetrix U133 microarrays; no signal was detected for MMP-8.

Download figure to PowerPoint

Up-regulation of PTX-3, activin A, and IL-8 expression by IL-1 plus OSM in a murine model of arthritis.

Previous studies have shown that the combination of IL-1 and OSM is a potent inducer of MMP expression in T/C28a4 chondrocytes. We also demonstrated, in an in vivo model, that overexpression of IL-1 and OSM in murine knee joints results in marked morphologic changes, including synovial hyperplasia, pannus formation, and cartilage and bone erosions characteristic of RA (9, 21), and these changes correlate with increased MMP expression. Furthermore, Northern blot analysis of primary human articular chondrocytes showed that IL-1 plus OSM up-regulates MMPs 1, 3, 8, and 13 (21).

To assess whether the proteins encoded by synergistically induced genes were also up-regulated in vivo, we used our established murine model of arthritis in which IL-1 and OSM are overexpressed, and compared the results with the microarray and real-time PCR data. Three genes that exhibited a moderate, high, and very high level of induction by IL-1 plus OSM were selected (Table 1). These were activin A, the long pentraxin, PTX-3, and IL-8, respectively. Studies have identified expression of activin A in chondrocytes (22), and this has been shown to suppress IL-6 activity, thereby dampening inflammatory responses. Although the array data indicated that activin A was moderately up-regulated following stimulation with IL-1 plus OSM for 24 hours, the in vivo data showed marked staining for activin A that was localized to the chondrocytes (Figure 3A). In SW1353 cells, high levels of activin A expression occurred at 4 hours (Figure 3B), and again the magnitude of induction obtained from the microarrays was similar to that obtained using real-time PCR (Figure 3C). High levels of activin A were also synergistically induced by IL-1 plus OSM in primary human chondrocytes (Figure 3D).

thumbnail image

Figure 3. Up-regulation of activin A expression in vivo and in vitro by interkeukin-1 (IL-1) plus oncostatin M (OSM). A, Murine joints were treated with vector alone (control), vector encoding murine IL-1, vector encoding OSM, or vectors encoding both IL-1 and OSM. Subsequently, joints were wax-embedded and 5-μm sections were prepared. Activin A was detected using a goat anti-human activin A antibody. B, SW1353 cells were treated for 4–24 hours with medium alone (control), IL-1 (0.2 ng/ml), and/or OSM (10 ng/ml), and activin A expression levels were analyzed by real-time polymerase chain reaction (PCR). The fold expression induced by each treatment is shown relative to control. C, Total RNA (15 μg) from SW1353 cells, treated for 24 hours as described in B, was labeled and hybridized to U133 microarrays. The fold change in expression induced by each treatment is shown relative to control. D, Primary human articular chondrocytes were stimulated with medium alone (control), IL-1 (0.02 ng/ml), OSM (10 ng/ml), or IL-1 + OSM for 24 hours, and activin A expression was assessed using real-time PCR. The fold change in expression induced by each treatment is shown relative to control. Bars show the mean and SEM. ∗ = P < 0.05 versus either IL-1 or OSM.

Download figure to PowerPoint

PTX-3 expression in RA synovium has been reported previously, but this is the first report of its expression by chondrocytes. In the present study, samples were prepared from murine joints treated with vector alone or with vector overexpressing IL-1, OSM, or IL-1 plus OSM. The joints were subsequently wax-embedded and sections were cut and subjected to immunohistochemistry in order to detect PTX-3. Weak staining for PTX-3 was observed in sections from the control, IL-1–treated, and OSM-treated cartilage, whereas there was more marked PTX-3 staining in the IL-1 plus OSM–treated cartilage (Figure 4A). Real-time PCR using RNA from IL-1 plus OSM–treated SW1353 chondrocytes demonstrated a maximal synergistic induction of PTX-3 at 4 hours (Figure 4B). The induction observed on the Affymetrix microarrays at 24 hours was concordant with the corresponding real-time data (Figure 4C). In addition, PTX-3 was also synergistically up-regulated in primary chondrocytes (Figure 4D).

thumbnail image

Figure 4. Up-regulation of pentraxin 3 (PTX-3) expression in vivo and in vitro by IL-1 plus OSM. A, Murine joint sections were prepared as described in Figure 3, and PTX-3 expression was detected using a goat anti-human PTX-3 polyclonal antibody. B, SW1353 cells were treated for 4–24 hours with medium alone (control), IL-1 (0.2 ng/ml), and/or OSM (10 ng/ml), and PTX-3 expression levels were analyzed by real-time PCR. The fold expression induced by each treatment is shown relative to control. C, SW1353 cells were treated as described in B for 24 hours, and labeled RNA (15 μg) was hybridized to the U133 microarrays. The fold change in expression induced by each treatment is shown relative to control. D, Primary human articular chondrocytes were stimulated with medium alone (control), IL-1 (0.02 ng/ml), OSM (10 ng/ml), or IL-1 + OSM for 24 hours, and PTX-3 expression was assessed using real-time PCR. The fold change in expression induced by each treatment is shown relative to control. Bars show the mean and SEM. See Figure 3 for other definitions.

Download figure to PowerPoint

KC is a chemokine that is found in mice and is functionally equivalent to IL-8 (23). This was up-regulated by IL-1 plus OSM in the murine model, especially at the articular surface (Figure 5A). Furthermore, IL-8 was markedly up-regulated in SW1353 cells, with maximal induction occurring 24 hours poststimulation (Figure 5B). Again, the 24-hour data from the microarrays and real-time analysis were very similar (compare Figures 5C and 5B). IL-8 expression was also induced by the cytokine combination in primary human chondrocytes (Figure 5D).

thumbnail image

Figure 5. Up-regulation of IL-8/KC expression in vivo and in vitro by IL-1 plus OSM. A, Murine joint sections were prepared as described in Figure 3, and KC (the IL-8 homolog in mice) was detected using a goat anti-mouse KC antibody. B, SW1353 cells were treated for 4–24 hours with medium alone (control), IL-1 (0.2 ng/ml), and/or OSM (10 ng/ml), and IL-8 expression levels were analyzed by real-time PCR. The fold change in expression induced by each treatment is shown relative to control. C, SW1353 cells were treated as described in B for 24 hours, and total RNA (15 μg) was labeled and hybridized to the microarrays. The fold change in expression induced by each treatment is shown relative to control. D, Primary human articular chondrocytes were stimulated with medium alone (control), IL-1 (0.02 ng/ml), OSM (10 ng/ml), or IL-1 + OSM for 24 hours, and IL-8 expression was assessed using real-time PCR. The fold change in expression induced by each treatment is shown relative to control. Bars show the mean and SEM. See Figure 3 for other definitions.

Download figure to PowerPoint

DISCUSSION

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

Although considerable data exist on the nature of the genes that contribute to pathologic cartilage destruction, such as MMPs (see ref.1 and references therein), less data are available on the repair responses that are invoked following a proinflammatory stimulus. The aim of the current study was to identify genes up-regulated by IL-1 plus OSM that may contribute to such a repair mechanism. We have shown that the combination of IL-1 and OSM up-regulates many of the MMPs known to play a key role in cartilage degradation (1), as well as some for which a defined role has yet to be demonstrated, including MMPs 10 and 12.

The data show that MMP-10 is synergistically up-regulated in chondrocytes by IL-1 plus OSM. Other studies have shown that MMP-10 can degrade aggrecan, link protein, and fibronectin, and that it activates proMMP-1 and proMMP-8 (1). MMP-12 (macrophage elastase) is also present in RA synovium (24) and can degrade elastin, fibronectin, and laminin in addition to cleaving and activating pro–tumor necrosis factor α (TNFα), proMMP-2, and proMMP-3 (25). MMP-12 also cleaves urokinase-type plasminogen activator receptor (uPAR), resulting in its inhibition (26). This cell-surface receptor localizes and enhances uPA activity, which converts plasminogen to plasmin; this has been implicated in procollagenase activation and cartilage collagenolysis (6). The array data also identified other genes that are likely to be involved in cartilage degradation, such as C1r. Expression of complement component in chondrocytes has been previously reported (27), and C1r activates C1s, which can degrade type II collagen and decorin (28). Serum amyloid A2 (SAA2) was also induced. SAA expression has been reported in RA synovial tissue (29), and SAA proteins induce MMP production in synovial fibroblasts (29), thereby enhancing ECM breakdown.

Various cytokines, chemokines, and their receptors involved in inflammatory processes were up-regulated. These include MCPs 1 and 3, IL-6, leukemia inhibitory factor (LIF), and the OSM-specific receptor OSMβR. IL-6, LIF, and OSM all belong to the glycoprotein-130–binding cytokine family (30). IL-6 and OSM are produced in RA synovium and can act synergistically with IL-1 and TNFα in the presence of their soluble receptors (7, 31) to promote cartilage breakdown (6, 9, 10, 21). An increase in the expression of OSMβR may lead to prolonged activation of OSM-mediated signaling pathways (30), and one component that facilitates such signaling, Jak-2 kinase, was also notably induced. Combined with the marked induction of IL-1β, this could result in an exacerbation of IL-1 plus OSM–induced effects within cartilage by the resident chondrocytes. This indicates that cartilage may be a much more active player in RA pathogenesis than previously thought.

The chemokines IL-8 and ENA-78 were synergistically up-regulated. We confirmed that KC, a murine equivalent to IL-8 (23), was up-regulated by IL-1 plus OSM in a murine model of arthritis. KC expression was primarily localized to the articular surfaces of the cartilage, concomitant with its role as a chemoattractant inducing the migration of neutrophils from the synovium toward the cartilage. IL-8 and ENA-78 are potent inducers of angiogenesis (32), which is a marked feature in our arthritis model (9, 10). IL-8 also contributes to the pathologic changes observed in arthritis through p38 MAPK pathway activation (33), which can lead to hypertrophic differentiation, alteration in collagen subtype expression, and cartilage calcification. PBEF was also up-regulated, and this cytokine perpetuates inflammation since it stimulates IL-6 and IL-8 expression (34). Cartilage may therefore inadvertently contribute to joint inflammation, since the inflammatory process is presumably initiated as a repair response to the original procatabolic stimulus.

Stimulation with IL-1 plus OSM significantly induced the calcium binding proteins S100 A8 and S100 A9. These proteins have been localized to RA synovial tissue, in particular, the synovium–pannus junction (35). S100 A8 and S100 A9 activate endothelium, promoting further recruitment of inflammatory cells into the synovium (36) and thus perpetuating inflammation. PTX-3 expression has been reported in RA synovium (37), but this study is the first to demonstrate PTX-3 expression by chondrocytes. Several functions have been attributed to PTX-3, including C1q binding, complement activation (38), and inhibition of angiogenesis through its interaction with fibroblast growth factor 2 (39).

A variety of other genes that represent a repair response mechanism were also up-regulated by IL-1 plus OSM. The serine protease inhibitors antileukopeptidase and squamous cell carcinoma antigen were induced. Antileukopeptidase prevents cartilage and bone erosion in anti–type II collagen antibody–induced arthritis (40). Activin A is a member of the transforming growth factor β (TGFβ) superfamily, mediating its affects through Smad transcription factors. The Array data, as well as immunolocalization, showed that activin A is significantly up-regulated by IL-1 plus OSM in chondrocytes. Previous studies have shown that activin A is expressed in RA synovial tissue and can induce the proliferation of fibroblast-like synoviocytes in culture (41). In osteoarthritic cartilage, activin A exhibits anabolic properties, inducing expression of tissue inhibitor of metalloproteinases 1 and increasing expression of type II collagen and proteoglycan synthesis in chondrocytes (42, 43). Our findings also indicate that activin A expression appears to be prolonged, since its expression was relatively unaltered regardless of the adenovirus titer used (data not shown). In cartilage, activin A expression may be a repair response, since it appears to have a protective role by preventing MMP-mediated cartilage degradation and promoting ECM-component synthesis. However, we have shown that, unlike TGFβ (44), activin A fails to prevent IL-1 plus OSM–mediated cartilage collagenolysis in an in vitro model of cartilage breakdown (Hartland S and Rowan AD: unpublished results).

Another markedly induced gene following IL-1 plus OSM stimulation was YKL-40 (chitinase-3–like protein 1), a protein that is reported to be present in degenerate articular cartilage and in inflamed, hyperplastic synovium (45, 46). Recent studies have shown that purified YKL-40 promotes connective tissue growth, provides a signal through kinase-mediated signaling pathways (47), and inhibits fibroblast responses to IL-1 through its effects on these pathways, resulting in a reduction in MMP-1, MMP-3, and IL-8 production (48). These data support the concept of a protective repair response elicited by chondrocyte-derived YKL-40. Indeed, it is known that 33% of conditioned medium from stimulated chondrocytes can exhibit YKL-40 (49), and our observations support this concept (Catterall JB, et al: unpublished results). Expression of the structural components decorin and fibronectin was increased by IL-1 plus OSM, again indicating a repair response aimed at synthesizing new ECM components. Another protein that may have a protective role is superoxide dismutase, which is an antioxidant that removes superoxide anions that have been implicated in hyaluronic acid and cartilage ECM damage. The role of this protein is supported by evidence showing that a genetic deficiency in superoxide dismutase results in an enhancement of collagen-induced arthritis in mice (50).

We have provided further evidence of marked induction of MMPs in chondrocytes following stimulation with IL-1 plus OSM, confirming their undoubted importance in the degradation of the cartilage ECM. We have also provided evidence that chondrocytes are capable of expressing a variety of factors following stimulation, some of which are protective. It would appear that these reparative mechanisms, initiated by chondrocytes, are ultimately overwhelmed by the continued inflammatory stimuli that predominate in cartilage catabolism. Our data suggest that blockade of such proinflammatory stimuli may inadvertently suppress potential repair mechanisms as well as catabolic processes, and such interventions therefore need to be fully evaluated. Moreover, our data indicate that cartilage may be an active player in the disease process, especially in RA, which is often viewed as a synovium-driven disease. Thus, microarray analyses have highlighted many genes that may play a role in prevention of cartilage breakdown and mechanisms of the repair response. These genes could be exploited for therapeutic intervention in the future.

Acknowledgements

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

We thank Dr. Keith Ray and Prof. John Heath for providing some of the reagents. In addition, we are indebted to Dr. Chris Morris for allowing the use of his immunohistochemistry facilities, and Arthur Oakley for help with the microscopy.

REFERENCES

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