F.A. van de Loo

Effects of murine recombinant interleukin 1 on intact homologous articular cartilage: a quantitative and autoradiographic study.

Murine recombinant interleukin 1 (IL1) was tested for its ability to affect intact murine articular cartilage. IL1 caused enhanced proteoglycan degradation and severe inhibition of chondrocyte synthetic function at a concentration of 3 U/ml (100 pg/ml). Inhibition of proteoglycan synthesis appeared to be delayed in onset but occurred consistently after 24 hours. Pulse chase experiments made it clear that proteoglycan degradation and inhibition of proteoglycan synthesis are two distinct actions of IL1. No indications were obtained for selective degradation of either newly synthesised or processed proteoglycan. Moreover, chondrocyte synthetic activity appeared to be inhibited uniformly throughout the cartilage matrix, i.e., no evidence was found for selective suppression of cells in certain regions. IL1 uptake measurement in the cartilage, using [125I]IL1, yielded a partition coefficient far below 1, and autoradiography demonstrated a faint but even distribution within the cartilage matrix. The coordinated induction of enhanced breakdown of proteoglycan and inhibition of proteoglycan synthesis, with such low concentrations of IL1 reaching the chondrocytes, underlines the impressive destructive potential of IL1.

A novel Saa3 -promoter reporter distinguishes inflammatory subtypes in experimental arthritis and human synovial fibroblasts

Objective To evaluate the applicability of a lentiviral (LV) serum amyloid A3 (Saa3)-promoter luciferase (Luc) reporter for assessing inflammation in experimental arthritis, synovial fibroblasts (SF) from osteoarthritis (OA) and rheumatoid arthritis (RA) patients. Methods In mice, synovium was transduced in vivo by cholesterol optimised LV, and two flares of acute joint inflammation were induced by injection of streptococcal cell wall (SCW) material into the knee-joint cavity. The time course of synovial inflammation was assessed using ex vivo luciferase assays, and histology. Uptake of 99mtechnetium (Tc) was used to assess oedema. SF (n=12) of RA and OA patients were stratified by hierarchical clustering of whole genome expression profiles. Relative Saa3-promoter responses were determined in cytokine- or toll-like receptor (TLR)-stimulated SF subgroups. Results In vivo, the Saa3-promoter reporter activity was strongly upregulated at 1 and 2 days after the first and second SCW challenge. The Saa3-promoter activities during acute inflammation correlated with Tc uptake measurements but were more sensitive and able to respond to the ongoing synovitis in the chronic phase of SCW arthritis. Molecular stratification defined two inflammatory SF subtypes, unrelated to disease classification. Relative Saa3-promoter responses to interleukin 1β, tumour necrosis factor α and TLR4 agonist were significantly increased in OA/RA SF with a high compared to a low inflammatory profile subtype. Serum stimulation of the Saa3-promoter reporter cell-line could distinguish between healthy and RA patients. Conclusion The Saa3-promoter reporter demonstrates a robust and feasible tool for assessing the course and severity of experimental arthritis and for distinguishing molecularly distinct inflammatory SF subtypes from a heterogeneous patient population.

Higher efficacy of anti-IL-6/IL-21 combination therapy compared to monotherapy in the induction phase of Th17-driven experimental arthritis

Th17 cells and their cytokines are linked to the pathogenesis of rheumatoid arthritis, a chronic autoimmune disease characterized by joint inflammation. Th17 development is initiated by combined signaling of TGF-β and IL-6 or IL-21, and can be reduced in the absence of either IL-6 or IL-21. The aim of this study was to assess whether combinatorial IL-6/IL-21 blockade would more potently inhibit Th17 development, and be more efficacious in treating arthritis than targeting either cytokine. We assessed in vitro Th17 differentiation efficacy in the absence of IL-6 and/or IL-21. To investigate in vivo effects of IL-6/IL-21 blockade on Th17 and arthritis development, antigen-induced arthritis (AIA) was induced in IL-6-/- x IL-21R-/- mice. The therapeutic potential of this combined blocking strategy was assessed by treating mice with collagen-induced arthritis (CIA) with anti-IL-6R antibodies and soluble (s)IL-21R.Fc. We demonstrated that combined IL-6/IL-21 blocking synergistically reduced in vitro Th17 differentiation. In mice with AIA, absence of IL-6 and IL-21 signaling more strongly reduced Th17 levels and resulted in stronger suppression of arthritis than the absence of either cytokine. Additionally, anti-IL-6/anti-IL-21 treatment of CIA mice during the arthritis induction phase reduced disease development more potent than IL-6 or IL-21 inhibition alone, as effective as anti-TNF treatment. Collectively, these results suggest dual IL-6/IL-21 inhibition may be a more efficacious therapeutic strategy compared to single cytokine blockade to suppress arthritis development.

FRI0039 Synovial Macrophages Promote TGF-Beta Signaling but Protect against Influx of S100a8/s100a9-Producing Cells Following Intra-Articular Injections of Oxidized Low-Density Lipoproteins

Background In previous studies we found that synovial macrophages regulate joint pathology during experimental osteoarthritis (OA) and, more recently, we found that high systemic levels of low-density lipoproteins (LDL) aggravate joint pathology during experimental OA with synovitis. LDL in inflamed synovium is oxidized and taken-up by macrophages, leading to an activated macrophage phenotype. Objectives In this study, we investigate whether injection of oxidized LDL directly into a murine knee joint induces joint pathology and elucidate the role of synovial macrophages in that process. Methods Synovium was obtained from end-stage OA patients and stained for apolipoprotein B (APOB). Murine knee joints were injected five consecutive days with oxLDL, LDL, or an equal volume of vehicle (PBS). This procedure was repeated in mice depleted of synovial macrophages by intra-articular injection of clodronate liposomes seven days prior to the consecutive injections. Joint pathology was investigated by immunohistochemistry and flow cytometry (FCM) analysis, and RNA expression and protein production by synovium were determined using RT-PCR and luminex, respectively. Aggrecanase activity was measured using NITEGE-staining and active TGF-β was measured using a functional CAGA-luciferase assay. Data are depicted as mean ± standard deviation. Results Synovial macrophages and fibroblasts of end-stage OA patients showed extensive accumulation of APOB, the main protein present in LDL and oxLDL. Multiple injections of oxLDL in murine knee joints significantly increased TGF-β activity in synovial wash-outs by 28% [from 16562 ± 2326 relative light units (RLU) to 21151 ± 3823 RLU; P<0.05], but did not induce catabolic or inflammatory processes. In contrast, repeated injections of oxLDL in macrophage-depleted knee joints led to a 3.1 fold increase of synovial thickening, compared with injection of PBS (P<0.01), while LDL injections did not alter synovial thickening. Protein and RNA levels of chemokines CCL2 and CCL3 were significantly upregulated in macrophage-depleted joints after oxLDL injections (6.7 fold and 4.6 fold, respectively; P<0.01). Furthermore, FCM-analysis revealed increased presence of monocytes (from 1422 ± 1105 to 3029 ± 1644 cells) and neutrophils (from 4014 ± 3511 to 12708 ± 7829 cells) in the synovium of macrophage-depleted joints after injection of oxLDL (P<0.05), which was confirmed by immunohistochemical staining. Also protein levels of S100A8/A9, markers for inflammation, were significantly increased in synovial wash-outs of oxLDL-injected joints compared with LDL (fold increase 5.6; P<0.05) or PBS (fold increase 8.3; P<0.01) injection, as was NITEGE expression (fold increase 1.92; P<0.05). Interestingly, no raise in active TGF-β was measured in these macrophage-depleted joints. Conclusions Synovial macrophages promote anabolic processes after intra-articular oxLDL injections. In absence of synovial macrophages, however, oxLDL induces production of pro-inflammatory mediators and aggrecanase activity, in combination with increased influx of monocytes and neutrophils. Disclosure of Interest None declared

A5.3 Elevated levels of BMP2 compensate for loss of TGF-BETA on proteoglycan level in articular cartilage during experimental osteoarthritis

Purpose We demonstrated earlier that in aging articular cartilage TGF-beta signalling via Smad2/3 is drastically reduced and loss of Smad2/3signalling predisposed cartilage for OA. We showed that TGF-beta inhibition reduced proteoglycan content in articular cartilage. In contrast, during OA we found elevated levels of BMP2 surrounding cartilage lesions. It is unclear what is the effect of this BMP2 presence on articular cartilage. Therefore, we investigated whether elevated BMP-2 could counteract the loss of TGF-beta signalling during OA. Methods We made a unique inducible transgenic mouse expressing human BMP2 under control of the Col2a1 promoter only when exposed to doxycycline (Col2a1-rtTA-BMP2). Functionality was tested by measuring human BMP2 mRNA on Q-PCR from articular cartilage, spleen and liver 72 hours after exposure to doxycycline food or standard diet. We induced OA in Col2a1-rtTA-BMP2 by destabilization of the medial meniscus (DMM-model) while treating them with doxycycline- versus standard diet. To study the effect of loss of TGF-beta activity during OA, we additionally intra-articularly injected an adenovirus overexpressing TGF-beta inhibitor LAP (Ad-LAP). Four weeks after DMM-induction knee joints were isolated for histology. OA was scored based on cartilage damage (adapted OARSI score, scale 0-30). Proteoglycan content was measured with digital image analysis in Safranin O stained articular cartilage of the medial tibia. Results Col2a1-rtTA-BMP2 mice with doxycycline food clearly had elevated expression of hBMP2 mRNA in articular cartilage, but not in spleen and liver thereby confirming functionality of the transgenic animals. Doxycycline exposure in Col2a1-rtTA-BMP2 up to 8 weeks did not result in any detectible alterations in healthy articular cartilage. When OA was induced OA score clearly increased (average of all DMM groups 16.9 versus 2.5 in non-DMM groups), but this was not significantly affected by elevated chondrocyte-specific BMP2. TGF-beta inhibition during DMM significantly reduced the proteoglycan content by 18% compared to DMM alone. BMP2 overexpression only did not affect the proteoglycan content during DMM. Nevertheless, the proteoglycan depletion by inhibition of TGF-beta during DMM was significantly and nearly completely counteracted by elevated chondrocyte-specific BMP2. Conclusions Our data show that in healthy articular cartilage and in OA cartilage in young animals elevated levels of BMP2 did not have any detectible effects on its own. However, when TGF-beta signalling was lost, a phenomenon occurring in aged individuals, this resulted in decreased levels of PG content in articular cartilage during OA. In this setting, elevated levels of BMP2 could compensate this loss of PG. Therefore the elevated levels of BMP2 near OA lesions could be a reparative response of the articular cartilage. Especially with ageing, when TGF-beta signalling is drastically reduced this compensatory mechanism could be of great importance as an attempt to restore damaged articular cartilage.

AB0077 RA-Patients Show Differential Responses in A Novel Cell-Based Promoter Reporter Assay

Background Disease activity in rheumatoid arthritis (RA) patients is commonly determined by calculating the DAS28 score, incorporating the number of swollen and tender joints, along with a measurement of the acute phase response (APR). Because the joint count relies on a personal assessment, many effort is being put into finding biomarkers that can improve an objective disease assessment. However, no single biomarker has been found that can sufficiently reflect disease activity. A novel approach for a bioassay can be through disease-inducible promoter-reporters, consisting of a promoter sensitive to RA-related disease factors that can drive the expression of a reporter gene. After incubation with patient serum, the signal from the reporter assay is a direct measurement of the disease activity. Objectives This study aims to develop promoter-reporter constructs that are sensitive to serum from RA patients for the development of a novel bioassay. Methods Microarrays on joint tissue of 20 RA patients and 7 controls were analysed to find upregulated genes during active disease. The upregulation was validated by qPCR in Pam3Cys stimulated human THP-1 monocytes. The promoters of the upregulated genes were isolated from human cDNA and cloned into a lentiviral vector containing the firefly luciferase reporter gene. THP-1 cells were transduced with the lentiviral constructs and stimulated for 6 hours with pro-inflammatory stimuli, 64 RA patient sera and 34 control sera to determine the inducibility of the promoters. Results The microarray analysis resulted in a list of 8 genes that were upregulated ≥4-fold in RA synovium. The increased expression was confirmed by qPCR. To mimic the inflammation in RA patients in vitro, THP-1 monocytes were stimulated by Pam3Cys. 6/8 genes were upregulated, showing the validity of the selected genes and the applicability of the THP-1 cells. After transduction of THP-1 cells with the promoter reporter constructs, Pam3Cys stimulation resulted in a 15.5-fold upregulation of the CXCL10 promoter reporter and a 26.5-fold upregulation of the IL-8 promoter reporter, indicating that the isolated promoter sequences contain the important regulatory binding sites. In addition, a promoter that contained 4 repeated binding sites for NF-κB was induced 68.5-fold by Pam3Cys. Next, the transduced THP-1 cells were stimulated with human sera. The promoter constructs with CXCL10, IL-8 and 4x NF-κB could distinguish between RA and healthy serum (P=0.017, P=0.036 and P=0.023 respectively). Finally, we compared the “assay-positive” RA sera that induced a signal of at least 1x standard deviation above the healthy control average in at least one reporter construct and found that these sera had a significantly higher erythrocyte sedimentation rate (P=0.0022), an acute phase response marker often incorporated in the DAS28, compared to “assay-negative” RA sera. Furthermore, the differences between the binding sites present on the promoters and the differences in RA sera that were tested positive between the promoters indicates that the different reporters might reflect different disease processes. Conclusions Promoter-reporter constructs can respond to RA-related disease factors in the serum of RA patients, indicating their potential for the development of a novel disease characterizing assay. Disclosure of Interest None declared DOI 10.1136/annrheumdis-2014-eular.2895

A10.26 Synovial Wnt and WISP1 Expression Induces Expression of Cartilage-Degrading Metalloproteinases in the Synovium

Background Many osteoarthritis (OA) patients show synovial involvement. We found strong upregulation of Wnts 2b and 16 and the downstream protein WISP1 in knee joints in experimental OA. The role of the synovium in the induction of OA pathology under the influence of Wnt signalling is unclear. Here we investigated the potential of Wnt signalling to increase the expression of cartilage-degrading enzymes in the synovium. Methods Pathway analysis of microarray data from the synovium of a collagenase-induced OA mouse model was done using DAVID software. In vivo synovial overexpression of genes from the canonical Wnt signalling pathway was achieved by intra-articular injection of adenoviral vectors. Joint pathology was assessed by histology. Gene expression was analysed by qPCR. Human OA synovial specimens were collected from joint replacement surgery and stimulated or used for outgrowth of fibroblasts. Results Pathway analysis showed that the Wnt signalling pathway was enriched in the synovium during experimental OA. To determine the effects of Wnt signalling on synovial tissue, we stimulated human OA synovial specimen with Wnt3a or WISP1, which resulted in increased expression of MMP3, MMP9 and MMP13, whereas expression of TIMP1 and 3 was not altered. Next, we investigated which cell-type in the synovium may cause the increased MMP expression. Stimulation of synovial OA fibroblasts with Wnt3a increased the expression of MMP3 and MMP13, whereas WISP1 led to increased MMP3 expression. Stimulation of THP-1 cells with Wnt3a resulted in highly increased MMP3, MMP9 and MMP13 expression. WISP1 stimulation led to increased expression of MMP3. TIMP1 and 3 expression was not altered. Synovial overexpression of Wnt8a, Wnt16 and WISP1 in murine knee joints with use of adenoviral vectors led to cartilage damage in vivo. 7 days after overexpression, we found a significant induction of OA pathology at the medial margin of the medial tibial plateau, a preferential site for damage in experimental OA. Lesions were found in 92% (n = 12) of the knee joints after Wnt8a overexpression compared to 17% (N = 12) for the control virus and 80% (N = 5) for Wnt16 overexpression, but only 20% (N = 5) for the control virus. In addition, overexpression of WISP1 led to significantly increased cartilage damage after 7 days. Conclusions Canonical Wnts produced in the synovium may play an important role in OA pathology by increasing the expression of cartilage-degrading MMPs in synovial tissue, where fibroblasts and macrophages showed comparable patterns of MMP induction. Synovium-specific overexpression of Wnt signalling members induces cartilage damage.

AB0070 Wnt and wisp1 expression in the synovium induces cartilage damage by skewing of tgf-beta signaling via the canonical wnt signaling pathway

Background Many osteoarthritis (OA) patients show significant synovial involvement. Recently, we found strong upregulation of canonical Wnts 2b and 16 and WISP1, a downstream protein, in the synovium in two murine OA models. Wnt signaling has been implicated in OA through activation of the β-catenin pathway. In addition, TGF-β signaling is critical in cartilage maintenance. TGF-β signals via both ALK5 and ALK1 and downstream via Smad 2/3 and Smad 1/5/8 respectively. Objectives Investigate the potential of canonical Wnts to skew TGF-β signaling from the protective Smad 2/3 pathway towards the Smad 1/5/8 pathway, which can induce chondrocyte hypertrophy. Methods Pathway analysis of microarray data from the synovium of a collagenase-induced OA (CIOA) mouse model was done using DAVID software. In vivo synovial overexpression of genes from the canonical Wnt signaling pathway was achieved by intra-articular injection of adenoviral vectors. Gene expression in human chondrocytes was analyzed by qPCR. Detection of Smad 2/3 and Smad 1/5/8 phosphorylation was done by Western blot analysis. Functional TGF-β signaling via ALK5 was measured using the luciferase reporter construct CAGA-Luc. Results Pathway analysis using DAVID, showed that both Wnt and TGF-β signaling were enriched in the synovium of mice with CIOA. Because of their small size, Wnts and WISP1 can migrate into the cartilage and possibly alter the chondrocyte phenotype. To determine if synovial overexpression of canonical Wnts leads to Wnt signaling in the cartilage, we injected adenoviral vectors for Wnt8a and Wnt16 into murine knee joints. These vectors cannot target the cartilage, due to their size. Synovial overexpression of Wnt8a and Wnt16 led to β-catenin accumulation in chondrocytes, a tell-tale sign of canonical Wnt signaling. In vitro overexpression of canonical Wnts and WISP1 in human chondrocytes led to significantly increased collagen X and decreased collagen II expression, suggesting a loss of chondrocyte phenotype. Moreover, pre-incubation with Wnt3a and/or WISP1 resulted in decreased TGF-β-induced phosphorylation of Smad 2/3, whereas phosphorylation of Smad 1/5/8 was increased in both murine and human chondrocytes. This implies a shift towards dominant TGF-β signaling via the hypertrophy-inducing ALK1 pathway. On a functional level, pre-incubation with Wnt3a and/or WISP1 led to decreased CAGA-Luc reporter construct activity, confirming decreased ALK5 signaling. Moreover, the expression of the anti-hypertrophic factor Sox9 was decreased after pre-incubation with Wnt3a and WISP1. In order to investigate whether Wnt3a skews the TGF-β signaling via the canonical signaling pathway, we pre-incubated chondrocytes with Wnt3a and/or WISP1 in the presence of DKK-1, a selective inhibitor of canonical Wnt signaling. Compared to the groups without DKK-1, we found increased Smad 2/3 phosphorylation and decreased phosphorylation of Smad 1/5/8 after Wnt3a stimulation. Conclusions Wnts produced in the synovium may play an important role in OA pathology by changing the chondrocyte phenotype, probably through modulation of the important TGF-β signaling pathway via the canonical Wnt signaling pathway. This points towards Wnt/WISP1 expression in the synovium as an interesting target for OA therapy. Disclosure of Interest None Declared

218 CANONICAL WNT EXPRESSION IN THE SYNOVIUM INDUCES OA-LIKE CARTILAGE DEGENERATION

INTRODUCTION: In osteoarthritis (OA), cartilage damage is one of the main pathological features. There is, however, a significant involvement of the synovium in a large proportion of OA-patients. The mechanism through which the synovium contributes to OA pathology is not yet known. In cartilage, it is proposed that developmental processes important in embryonic development are reactivated. Associations have been identified of the occurrence of OA with polymorphisms of genes from the wnt/β-catenin pathway, a pathway that is involved in normal cartilage development. In the wnt/β-catenin pathway soluble wntproteins are secreted by cells and bind to Frizzled (Fzd) receptors, which sets into motion a signaling cascade leading to intra-cellular β-catenin accumulation and the transcription of a plethora of genes, like several proteases. The levels of β-catenin in the chondrocytes are crucial for the stability of chondrocyte phenotype. The aim of the present study is to investigate the contribution of the synovium to OA pathology via wnt signaling. METHODS: To demonstrate the presence of canonical wnt signaling in murine models for OA, murine knee joints isolated at several time point after induction of collagenase induced OA (CIOA), surgically induced OA and STR/ort mice were stained for the presence of β-catenin. CIOA was generated by intra-articular injection of collagenase, which induces joint instability. Surgical OA was induced by destabilization of the medial meniscus, by transecting the medial meniscotibial ligament. A longitudinal expression analysis was performed in 2 of the models, one with clear synovial involvement, CIOA, and the spontaneous OA model in STR/ort mice, which shows less synovial involvement. Synovial expression of components of the wnt signaling pathway was determined at several time points. From these results, targets were selected for the generation of adenoviral vectors to overexpress specific genes. To study in vivo effects, adenoviral vectors were injected intra-articularly in murine knee joints and joints were isolated and processed for histological examination at day 1, 3, and 7 after injection. To study the effect of these genes on chondrocytes, human chondrocytes were isolated from cartilage that was obtained from joint replacement surgery. One day after isolation, these cells were transfected with the adenoviral vectors, and incubated for 7 and 14 days. Hereafter Q-PCR was performed at the mRNA that was obtained from these cultures, to detect expression of aggrecan, collagens type I, II and X and several MMPs. Cartilage and synovial specimen were obtained at the time of joint replacement surgery after a signed informed consent was obtained. All animal experiments were approved by the institutional review board conforming to the local laws and regulations. RESULTS SECTION: In all experimental murine models for OA, β-catenin stained stronger in the cartilage compared to naïve mice, especially in the superficial cartilage layer (Figure 1). In the deeper, calcified layer cartilage stained positive in naïve mice as well. Also in synovium, β-catenin staining increased when OA developed (not shown). This indicates that during experimental OA, canonical wnt signaling occurs in the cartilage and synovium. Strong upregulation of the canonical wnts wnt16 (up to 256-fold) and wnt2b (up to 90-fold) was found in both models, although regulation was stronger in CIOA. Expression in the synovium was clearly higher compared to the cartilage. In the cartilage, no clear upregulation of canonical wnt-proteins was found. Clear intracellular accumulation of βcatenin was found in both synovium and cartilage, which indicates the activation of wnt/β-catenin in both tissues. This suggests that wntproteins that are expressed in the synovium, diffuse to the cartilage and induce wnt-signaling in chondrocytes. Wnt-1 induced signaling protein (WISP-1), a protein downstream canonical wnt signaling, was highly expressed in the synovium as well, again indicating activation of this pathway in the synovium. To determine whether canonical wnt expression in the synovium has the potency to cause cartilage damage, canonical wnt8a was overexpressed specifically in the synovium by intra-articular injection of an adenoviral vector. At day 1 and 3, no significant differences were observed in the cartilage from wnt8 overexpressing knee joints compared to joints transfected with control virus. Remarkably, at day 7, a strong induction of cartilage pathology was observed at the medial margin of the medial tibial plateau (Figure 2), a preferential site for the start of cartilage damage in our models. This shows that expression of canonical wnt in the synovium causes cartilage degeneration. In addition, overexpression of WISP-1, a mediator that is induced by canonical wnts induces MMPand aggrecanase mediated cartilage damage, already 4 days after transfection. Due to their size, wnt proteins and WISP-1 can reach the chondrocytes in the cartilage matrix and may alter the chondrocyte phenotype. This was underlined by increased β-catenin staining in cartilage of wnt8a overexpressing mice. Adenoviral overexpression of wnt8, wnt16 and WISP1 in human primary chondrocytes led to a significant increase within 14 days of Collagen type I, and a significant decrease of Collagen type II, suggesting loss of the articular chondrocyte phenotype. DISCUSSION: Canonical wnt expression and subsequent WISP-1 production is increased in the synovium during experimental OA. This synovial expression may lead to the degradation of cartilage as soon as 7 days after transfection, possibly by inducing changes in the articular chondrocyte phenotype. This identifies synovial wnt expression as a potential target for OA therapy. Future studies should focus on the inhibition of the canonical wnt16 and WISP-1, to underline its efficacy as a target for OA therapy. Paper No. 6 • 56th Annual Meeting of the Orthopaedic Research Society

A15 ELEVATED EXTRACELLULAR MATRIX PRODUCTION AND GAG DEGRADATION UPON BMP-2 STIMULATION POINT TOWARDS A ROLE FOR BMP-2 IN CARTILAGE REMODELING

Conclusions: We found that N-domain and full-length TIMP-3 are similarly effective in most in vitro and cell-based assays, although full-length was more potent against MMP-13. The Ndomain was more effective in blocking IL-1-induced GAG release from bovine nasal cartilage explants. Both forms, but most dramatically the full-length protein, inhibited cartilage degradation, osteophyte growth and bone lesions in a rat meniscal tear model of osteoarthritis. The greater efficacy seen with the full-length form could be due to its greater potency against MMP-13.