Osteoarthritis with its accompanying cartilage degeneration is a common consequence of aging [20
]. Many components of the tissue matrix undergo aging changes, including the major proteoglycan, aggrecan [21
], the various collagenous components [22
] and the chondrocytes themselves [23
]. Excessive proteolysis has long been recognized as a major contributory factor to cartilage matrix degradation in OA. A variety of proteases have been implicated, ranging from lysosomal cysteine proteases, such as cathepsin B, which can cleave aggrecan and non-helical telopeptide sites in Col II, to MMPs, such as MMP-3 (stromelysin), aggrecanases and collagenases [24
]. Collagen degradation is considered to be a key process in articular cartilage degeneration in OA [2
]. Collagenases are involved in normal extracellular turnover of collagen but their up-regulation is associated with the generation of inflammatory mediators [25
], which are believed to play a central role in the pathology of OA [26
]. Enhanced Col II degradation by collagenases has been detected in OA articular cartilages [2
] together with increased denaturation of Col II that results from intrahelical cleavage of Col II [17
]. It has long been considered that MMPs with the capacity to generate the classic collagenase cleavage site are the principal proteases capable of cleaving triple helical collagen. Recent studies have provided evidence that collagen degradation in health and disease may also involve the action of cathepsin K [4
], a cysteine protease that is up-regulated in OA chondrocytes [3
] and is capable of degrading native triple helical collagens, including Col II [4
], as well as other matrix molecules of the hyaline cartilage, such as aggrecan [28
The ability of cathepsin K to cleave native Col II at sites in the triple helix distinct from that of the primary collagenase cleavage site [4
], renders it unique among other tissue proteases. The present study, which used immunoperoxidase localization to identify and compare the distribution of Col II cleavage neoepitopes generated by these different proteases in healthy and OA cartilages, was prefaced by analytical immunoassays of these articular cartilages with the antibodies C2C and C2K, which revealed a significant increase in the generation of these two neoepitopes in aging and more so in OA cartilage [4
]. Using immunolocalization in the present study, we were able to demonstrate the more widespread distribution of Col II cleavage in arthritic cartilage and to identify the sites of more intense proteolytic activity within the tissue. Moreover, we showed that sites of cleavage of Col II by collagenases are usually co-localized with sites where the cathepsin K cleavage neoepitope is also found, suggesting that different proteases are active in sites of Col II cleavage in both normal and arthritic cartilages.
Our results for the collagenase neoepitope analyses complement our earlier data [18
] and demonstrate that in the majority of cases similar results were obtained with both C1,2C and C2C antibodies. In the cases where differences were observed with more staining for C1,2C, this may result from secondary cleavages in the collagen α chain (Figure ), which would only be recognized by the C1,2C antibody. Our results further indicate that Col II degradation usually starts around chondrocytes, in the upper region of cartilage, at and close to the articular surface, with progressive involvement of the mid and deep zones in aging and OA. A "stepwise" activation of chondrocytes by cytokines initiated in the superficial cartilage, slowly extending through chondrocyte activation into the deeper zones may account for the clear demarcation between the severely degenerate extracellular matrix and the more normal underlying cartilage. This "creeping substitution" of good with bad appears to be the manner whereby cartilage damage extends ever deeper into the cartilage with time. However, which cytokines stimulate cathepsin K expression in chondrocytes remains to be determined.
In most specimens that were used for the study, a similar distribution was observed for both collagenase and cathepsin K-generated cleavage of Col II. The sites of collagen degradation by collagenases are known to correlate with the distribution of the collagenases MMP-1 and MMP-13 in both healthy and OA cartilages [18
]. On the other hand, cathepsin K positive cells have been identified in all zones of degraded cartilage and there is evidence that cathepsin K may be released into the extracellular matrix, as indicated by immunohistochemical detections of the enzyme in the territorial matrix around cells [3
]. Recently, in a study of equine OA, use of the same antiserum to the neoepitope generated by cathepsin K revealed that sites of extracellular collagen cleavage were increased in degenerate cartilage and co-localized with cathepsin K staining [29
]. These and our present and recent findings [4
] together indicate that cleavage of Col II as detected in this study can involve both collagenases and cathepsin K that originate from chondrocytes.
Concomitant with aging in human cartilage, the accumulation of advanced glycation end products has been shown to decrease the susceptibility of this tissue to proteolytic degradation [30
]. Our immunohistochemical findings are consistent with this progressive crosslinking of the cartilage matrix, which limits the ability of cleavage products to be released from the tissue. This appears to account for the increased staining intensity of the collagen neoepitopes with age and the requirement for extensive chymotrypsin pretreatment of the tissue for immunoassay quantitation. In addition, despite a major effort, we have been unable to biochemically characterize the collagen peptides bearing the cathepsin K neoepitope. The product of in vitro
collagen cleavage was detectable by immunoassay but was not demonstrable by Western blotting even when high percentage polyacrylamide gels were used.
The present study mainly provides information regarding the distribution of collagen degradation within the tissue. Making quantitative estimates to compare the amount of epitopes generated by each of the enzymes is not possible since the antibody dilutions were adjusted to give the strongest specific antigen staining with the lowest non-specific background. However, our recent quantitative analyses also revealed a significant correlation between cleavage generated by collagenases and that which can be generated by cathepsin K [4
]. The present study indicates that Col II cleaving activity of the kind generated by cathepsin K is usually evident in the same sites as collagenase activity in both normal and diseased cartilages. Both studies thus revealed that neoepitopes are generated by both collagenases and cathepsin K in the same sites and increase with age and more so with OA and that these events are correlated in some way. That this non-collagenase cleavage is produced by cathepsin K is supported by our recent observation that the generation of the cathepsin K neoepitope in OA cartilage from some patients in culture can be inhibited by a specific non-toxic inactivator of this cysteine protease [4
]. In patients where inhibition was not observed, it appears that cathepsin K was not active in that period of time. This does not preclude the enzyme from being active at other times in these patients. But it also raises the issue of whether proteases other than cathepsin K and collagenases, the latter having not been shown by us to generate this cathepsin K cleavage neoepitope, may also produce this neoepitope. In fact, there have been several studies which have identified other cleavage sites in Col II distinct from those described here with no evidence for a single protease being involved in their generation [32
]. In a recent proteomic study [35
], products of human articular cartilage digested in vitro
with a series of MMPs and aggrecanases were characterized. While a large number of Col II cleavage products were identified, none were compatible with cathepsin K activity. Thus, only the collagenase generated neoepitope would appear to be specific for this protease [2
Together these results suggest that whereas collagenases may be the dominant mediator of Col II, cleavage in a majority of patients, cathepsin K and other proteases may also play a significant role. In both normal aging and in cartilage degeneration accompanying osteoarthritis, the degradative processes mediating collagen cleavage by these enzymes appear to be similar.