Background
Articular cartilage is a dynamic tissue in synovial joints that can withstand substantial loads within physiological levels [
1,
2]. Aggrecan, the major proteoglycan of cartilage, is substituted with negatively charged sulfated glycosaminoglycans (sGAGs) and forms large aggregates by binding to hyaluronan [
3]. The negative charge density induces a high osmotic swelling pressure within cartilage, restrained by the collagen fibril network [
3]. These osmotic and electrostatic repulsive interactions provide more than 50% of cartilage’s equilibrium compressive modulus. In addition, the closely spaced sGAG chains resist fluid flow caused by dynamic compression quantified at both the molecular level [
4] and the tissue level [
5]. The resulting intratissue pressurization caused by dynamic compression greatly increases cartilage’s dynamic modulus at high loading rates. Proteolytic cleavage of aggrecan is an important feature in osteoarthritis (OA) [
6‐
8]. Aggrecanase-1 and aggrecanase-2 (a disintegrin and metalloproteinase with thrombospondin motifs 4 and 5 [ADAMTS4 and ADAMTS5, respectively]) and the matrix metalloproteinases (MMP-1, MMP-3, MMP-8, MMP-9, and MMP-13) are believed to be the most important enzymes responsible for the proteolytic degradation of aggrecan [
9].
A severe joint injury leads to an increased risk of developing posttraumatic osteoarthritis (PTOA) [
10]. PTOA pathogenesis is multifactorial, and many risk factors associated with developing nontraumatic OA, such as age, obesity, and genetic variation, may also apply to PTOA [
10]. With regard to knee OA, the risk is related to the integrity of the menisci; to associated compressive injuries to ligaments, cartilage, and bone; and to joint synovitis [
11‐
14]. In the acute phase of injury, impact forces applied over the knee joint in combination with hemarthrosis lead to activation and recruitment of immune cells to the knee joint, producing an inflammatory and procatabolic joint environment. Several reports have demonstrated rapid increases in pro- and anti-inflammatory cytokines, proteases, and proteolytic activity after a severe knee injury [
15‐
22]. Cells of the synovium and fibrous joint capsule (referred to hereafter as the
synovium + joint capsule [SJC]), comprising resident inflammatory cells, synovial cells, and other cells such as endothelial cells and blood leukocytes, contribute to joint inflammation by producing proinflammatory mediators and proteases such as aggrecanases and MMPs [
23,
24]. Matrix molecule fragments and cell debris released into the joint after trauma lead to activation of various cells in the cartilage and the SJC, further increasing the proinflammatory response [
24]. In the long term, imbalance between pro- and anticatabolic activities may lead to progressive cartilage extracellular matrix degradation and development of OA [
15,
19,
25‐
29].
In the present in vitro study, we investigated differences in the preferred enzymatic cleavage of aggrecan in bovine cartilage explants cocultured with or without SJC, cartilage exposed to mechanical trauma, or cartilage exposed to exogenous tumor necrosis factor-α (TNF-α). We hypothesized that in this coculture system, there is cross talk between traumatized cartilage and the SJC through cell- and matrix-derived factors, leading to increased aggrecanase and MMP cleavage of the cartilage aggrecan. To quantify aggrecanase and/or MMP activity toward aggrecan in these different conditions, we used a set of well-characterized antibodies directed at different cleavage sites in the aggrecan molecule.
Discussion
Proteolysis of human aggrecan showing the fragmentology of aggrecan has been described previously [
8,
9,
34,
38], but details such as those in this study, showing aggrecan fragments present in cartilage and released into medium, have not been described to date, to our knowledge. Our findings suggest that coincubation of cartilage with SJC leads to increased proteolytic activity of both MMPs and aggrecanases against aggrecan. These quantitative findings were supported by the qualitative analysis of aggrecan fragments in medium, showing that conditions where the SJC was present were associated with additional enzymatically cleaved aggrecan fragment types released into medium and that proteolytic cleavage of aggrecan started earlier in time in these conditions. Mechanical injury of cartilage alone was associated with a small increase in the amount of FFGV fragments released into the medium, suggesting increased MMP activity toward aggrecan compared with the uninjured control.
Whereas high aggrecanase activity against the IGD cleavage site, as seen by Western blotting, was observed in cartilage stimulated by mechanical injury + TNF-α, no MMP activity against the IGD was observed in this condition during the 6 days of culture. This novel finding suggests that the activated proteolytic pathways are different between the TNF-α-stimulated cartilage and when cartilage is cultured alone, mechanically traumatized, or coincubated with SJC.
Activation of inflammatory pathways may be involved in the pathogenesis of OA and has been shown to associate with pain and to be prognostic of OA progression [
14,
39]. As this study and others have shown, the synovial cells and/or other cells of the fibrous joint capsule may induce aggrecan cleavage by producing proteases themselves and could, by producing inflammatory mediators, induce chondrocytes to produce aggrecanases and MMPs [
23,
26]. In the sequelae of severe joint injury, the activation of Toll-like receptors (TLRs) by matrix fragments may induce a vicious cycle by downstream activation of inflammation pathways [
40,
41]. Aggrecan fragments may also contribute to continued activation of inflammatory pathways, and a 32-mer aggrecan peptide was recently shown to have antianabolic, procatabolic, and proinflammatory bioactivity mediated through TLR2 and nuclear factor-κB [
42]. Importantly, with regard to the findings of the present study, cells other than synoviocytes (leukocytes, macrophages, fibroblasts, and endothelial cells) reside in the matrix underlying the synovial lining and could contribute to increased protease activity in the cartilage explant in vitro system [
43,
44].
Our results extend previous findings suggesting that cross talk between cartilage and SJC could increase aggrecanase activity in the cartilage [
23,
45] by showing that coincubation of mechanically injured cartilage with SJC increases both MMP and aggrecanase activity. We also found that the SJC produces active aggrecanases and MMPs. Our study further extends previous investigations by showing that the coincubation of cartilage with SJC induces not only aggrecan cleavage in the IGD (generating ARGS-CS2 fragments) [
45,
46] but also proteolytic cleavage between CS2 and G3 (generating ARLE-G3 fragments), and leads to a higher degree of cleavage in the CS2 region (generating G1-KEEE fragments). This is important, considering that aggrecan degradation in this explant study, and in other in vitro studies [
38,
47], starts with cleavage in the CS region followed by cleavage in the IGD.
Interestingly, the findings of the present study indicate that mechanical injury to cartilage alone, previously shown to induce increased MMP and aggrecanase gene expression (250-fold and 40-fold, respectively) [
25], did not increase aggrecanase-mediated digestion of aggrecan. This novel observation suggests that additional soluble factors from the SJC are essential to increasing aggrecanase degradative activity toward aggrecan in the cartilage. We also found evidence of more MMP activity in mechanically injured cartilage with or without coincubation with SJC compared with cartilage cultured alone. These findings are consistent with previous studies which have shown that the combination of mechanical injury of cartilage plus incubation with inflammatory cytokines led to the release of significantly greater amounts of sGAG and specific mass spectrometry-detected matrix fragments (e.g., aggrecan, collagen III, cartilage oligomeric matrix protein) from the cartilage than cytokine treatment alone [
27,
48]. Taken together, the findings of the present study are consistent with the hypothesis that mechanical injury alone does not induce or alter the mechanisms underlying aggrecan proteolysis but can affect the rate of release of the resulting fragments from the tissue via altered transport/diffusion. Also, the increased MMP activity observed in the mechanically injured cartilage coincubated with SJC could be related to increased access to the cartilage for proinflammatory molecules and proteases released from the SJC [
27].
In contrast, when mechanically injured cartilage was cultured in the presence of TNF-α, only aggrecanase and no MMP activity was detected. These findings are interesting and could imply that different intracellular pathways for inducing aggrecan degradation are activated, depending on if cartilage is stimulated by SJC or by TNF-α. Such differences may be crucial when interpreting results from in vitro studies and when assessing the physiological relevance of any given model.
These findings are furthermore consistent with previous in vitro and in vivo investigations showing that high concentrations of catabolic cytokines such as TNF-α lead to early (8 h to 2 days) aggrecanase activity [
49,
50]. In agreement with Madsen et al. [
50], who used a similar bovine explant system, we did not detect MMP-generated FFGV fragments in the medium or in the cartilage during the 6 days of cartilage explant culture in the presence of TNF-α. In all other conditions, FFGV-aggrecan was detected as early as 1–2 days after incubation started. More complex systems, such as in vivo lipopolysaccharide-induced joint inflammation or acute knee injury, have implicated increased MMP activity toward type II collagen within 1 day of insult [
15,
20,
51,
52]. However, whether MMP activity toward aggrecan is increased in these conditions is not well understood [
53]. Studies designed to identify which factors released from the SJC increase aggrecanase and MMP activity are warranted. If these factors could be singled out, then they could represent novel targets for treatment of PTOA or joint injury.
There are several limitations to the present investigation. Owing to the small study sample size (three animals), our findings need to be repeated. Of note, the excision of SJC before culture traumatizes the SJC tissue, and resident cells may respond to this injury with an inflammatory phenotype. This is also applicable to the cartilage. We must also keep in mind that very young bovine cartilage was used in the experiment, and we do not know if the pattern of proteolysis is similar in older bovine cartilage or in adult human cartilage, even though the aggrecan fragment pattern observed in the bovine medium resembles the pattern seen in synovial fluid from patients with OA (Additional file
1: Figure S1). However, using this young bovine system of very tightly controlled age, we have previously shown that animal-to-animal variations in cartilage behavior are typically no larger than specimen-to-specimen variations within a single animal, giving us confidence that the results presented are meaningful. A strength of the present study is that we were able to follow a course of events using different experimental conditions, using well-characterized bovine aggrecan fragments rather than observing momentary time frames of the joint milieu, as is the case in analysis of joint fluid after, for example, a joint injury.
Acknowledgements
We thank Jonas Ranstam (Lund University) for help with the statistical analysis, Michael W. Lark (Trevena, King of Prussia, PA, USA) for the kind gift of the LGQR and KEEE antibodies and MMP-3, Takehiko Nakamura (Seikagaku, Tokyo, Japan) for the kind gift of keratanase and keratanase II, Sanjay Kumar and Michael Pratta (GlaxoSmithKline, Collegeville, PA, USA) for the kind gift of ADAMTS-4 and ARGS (OA-1) and AGEG antibodies, Peter Roughly (Shriners Hospital, Montreal, QC, Canada) for the kind gift of ARLE antibody, and John Sandy (Rush University, Chicago, IL, USA) for the kind gift of G1 antibody.