Cartilage markers and their association with cartilage loss on magnetic resonance imaging in knee osteoarthritis: the Boston Osteoarthritis Knee Study

Osteoarthritis (OA) is characterized by the degeneration of articular cartilage. This results from a direct attack on matrix molecules, resulting in their cleavage, damage to these molecules, and their loss. It is also accompanied by a response of the tissue to this damage which involves enhanced matrix synthesis and turnover. The most direct evidence of pathology is cartilage degradation. A secondary and more indirect indication is cartilage matrix synthesis. The amount of synthesis in relation to degradation may prove of great importance in determining disease progression [1].

The ability to use biochemical markers to predict disease progression and identify patients most likely to progress is a top priority in the future management of OA. Ultimately, it would enable much more rapid assessment of structure-modifying therapies in clinical trials. It may also allow the identification of patients at highest risk of progression, allowing the efficient testing of new treatments. Biochemical markers of OA progression represent a surrogate for structural change which may have advantages over existing methods of measuring structure. Therapeutic development in OA is constrained by the slow progress of structural changes using standard imaging techniques. The development and validation of biochemical markers may accelerate the pace of therapeutic development.

Some recent work on type II collagen has suggested that assays for type II collagen degradation, when used in combination or with markers of collagen synthesis, can distinguish populations with knee OA which exhibit progression of joint damage from non-progressors. The ratio of the type II collagen crosslinking C-telopeptide (CTX-II) to the amino propeptide of type IIA collagen [2] or the ratio of two collagenase-generated cleavage epitopes in the helical region (C1,2C to C2C) [3] each can make this distinction. The results from each of these studies need to be confirmed. But, clearly, these two independent studies point to differences in collagen turnover as being suggestive of disease progression and provide encouragement for future work in this area. Preliminary plain radiographic studies suggest that cartilage oligomeric matrix protein (COMP) may be a useful prognostic marker of disease progression in knee [4–6] and hip [7] OA.

The overarching aim of this investigation was to conduct a study within an existing longitudinal dataset of knee OA with serial knee magnetic resonance imaging (MRI) to evaluate and validate promising biochemical markers, markers that have been reported in either cross-sectional or longitudinal studies to be related to OA or its progression. MRI of the knee has the advantage of covering the whole joint in one examination, meaning that the cartilage defects in the joint can be visualized directly, regardless of their location [8]. Direct visualization of cartilage defects enhances the ability to detect cartilage loss that can be missed using joint space narrowing from plain radiographs [8, 9].

More specifically, we assessed the baseline levels of cartilage degradation, synthesis, and turnover products using collagenase-generated C1,2C, and C2C; Col II C-telopeptide (CTX-II assay); C-propeptide of type II collagen; aggrecan 846 epitope; and COMP in a sample of knees with known knee cartilage loss and controls. Our prior hypotheses were that increased levels of cartilage degradation products would be predictive of cartilage loss and that imbalance of cartilage synthesis and degradation would be predictive of cartilage loss.

Increased COMP levels predict subsequent cartilage loss on MRI, but the association is modest (area under the curve = 0.60). The other biochemical markers of cartilage synthesis, turnover, and degradation do not facilitate prediction of cartilage loss.

Articular cartilage is a multiphasic material with at least two major phases: a fluid phase composed of water and electrolytes and a solid phase composed of chondrocytes together with matrix molecules that include collagen and proteoglycans. The predominant type of collagen is type II, which is found predominantly in cartilage (some also in the vitreous of the eye). It forms the basic fibrillar structure of the extracellular matrix which imparts its tensile strength.

As articular cartilage degenerates in OA, chondrocytes upregulate their biosynthetic activities, including type II collagen, as if to compensate for this damage. Only after secretion, as the molecules reach the extracellular space, are the non-helical domains at the end (the amino-terminal type II and carboxy-terminal type II procollagen propeptides [PIINP and PIICP, respectively]) cleaved from the helical domain.

The C-propeptide content and release from the cartilage are directly correlated with collagen synthesis [24]. In OA, a variant form of type II collagen is produced in which the N-propeptide contains an additional gene product of exon 2 [25]. This is type IIA collagen and it represents a form found in development but apparently not in healthy adult cartilage. The remainder of the collagen is called type IIB and lacks the exon 2 product. By immunoassay, it is possible to detect both type IIA and IIB collagen with the use of the C-propeptide assay [24].

It is hoped that the availability of assays to measure degradative, synthetic, and turnover products of cartilage matrix metabolism in body fluids offers opportunities to try and monitor cartilage turnover in vivo. Type II collagen is degraded by proteolytic enzymes secreted by chondrocytes and synoviocytes. The cleavage of the type II collagen triple helix by collagenases results in the generation of neoepitopes at cleavage sites. Since the initial cleavage that generates the neoepitope is followed by subsequent cleavage of the alpha chain, there is release of the epitope from the tissue [26]. Thus, an increase in these cleavage products can be detected in vivo by immunoassays with antibodies that recognize cleavage epitopes called COL2-3/4C_longmono (also known as C2C) and specific for type II collagen [14], and collagenase cleavage epitopes called COL2-3/4C_short or C1,2C, which detect cleavages of both type II and I collagens [15], have been generated and are markedly elevated in experimental OA [27]. These epitopes are generated in OA articular cartilage as shown by Billinghurst and colleagues [15]. However, our investigation suggests that they are not detectably elevated in patients at risk of cartilage loss, and thus their role in predicting OA progression in human MRI studies is questionable based on our data. The inability to detect an increase may be related to the number of joints involved in OA (joint load) compared with RA [19, 28]. Alternatively, it may be because, in serum, we are unable to distinguish increases in pathology from normal turnover.

Some recent work on type II collagen has suggested that assays for type II collagen degradation, when used alone or in combination or with markers of collagen synthesis, can distinguish populations with knee OA which exhibit progression of joint damage from non-progressors [2, 29–31]. The ratio of the type II collagen crosslinking C-telopeptide (CTX-II) to the amino propeptide of type IIA collagen [2] or the ratio of two collagenase-generated cleavage epitopes in the helical region (C1,2C to C2C) [3] each can potentially make this distinction. In one of these clinical studies, progression was identified by the increase in type II collagen cleavage products compared to a decrease in the propeptide marker of synthesis [2]. That study had a number of major limitations, including the control subjects having no radiographic evaluation, incomplete evaluation of OA subjects, and only 12 months of follow-up. However, the greater the distinction between increased degradation and decreased synthesis, the more progression was observed. In another unpublished study of progression [3], the relative amount of primary and secondary cleavage correlated with progression. Both C2C and C1,2C epitopes contain the cleavage site generated by collagenase. The greater the amount of the shorter epitope (C1,2C) containing the cleavage site, relative to the longer epitope (C2C), the greater the progression. This suggests that there is a difference in proteolysis linked to progression that leads to increased generation of the shorter C1,2C epitope. Our investigation did not corroborate these findings, potentially due to differences in study design.

In addition to type II collagen, the second main component of the extracellular matrix of articular cartilage is aggrecan, which, like collagen type II, is almost specific to this tissue. Aggrecan is a proteoglycan composed of a core protein to which GAG chains are covalently attached. The compressive stiffness of articular cartilage is a product of the hydration and swelling of aggrecan, embedded as macromolecular aggregates within the collagen fibrillar network. The monoclonal antibody CS 846 prepared for aggrecan reveals the presence of the largest apparently intact molecules that predominate in fetal cartilages but that are almost absent from healthy adult cartilage [32]. In OA, these larger molecules reappear and increase in amount with increased synthesis of this molecule [17] in synovial fluid and serum [33, 34]. The 846 monoclonal antibody usually recognizes the epitope on the largest molecules and likely signifies the presence of more recently synthesized molecules [17]. We are unaware of previous human clinical studies that have evaluated this promising biomarker and its relation to knee OA progression. Based upon our investigation, its role in predicting progression in knee OA may be limited, although its early increase has been observed in serum in experimental dog OA [35].

Cartilage oligomeric protein is a pentameric protein of the thrombospondin family which can bind type I, II, and IX collagens [36]. It is synthesized by chondrocytes, synovial cells, and other cells of the skeleton. Its synthesis is increased in chondrocytes and in synovial cells when activated by proinflammatory cytokines [37]. Preliminary plain radiographic studies suggested that COMP may be a useful prognostic marker of disease progression in knee [4–6] and hip [7] OA, and longitudinal analysis of COMP may predict episodic or phasic progression of OA [38]. We were able to corroborate these findings in a knee MRI study, suggesting that this marker may be a useful means of identifying progressors, albeit the estimate was modest.

Some further limitations of this work, some of which are generic to the application of biomarkers, warrant mentioning. Age-related increases are commonly seen in biochemical markers and these may produce variation in both biomarker levels and cartilage loss [39]. Efforts were made to adjust for age in analyses. The BOKS study assessed the local structural changes in the knees only and in only one knee (not both). It may be that other studies that investigate the total body burden of OA, including other joint areas, may be able to detect an association in patients with symptomatic OA. Another potential explanation for our null findings in patients with symptomatic OA is that we have insufficient power. It is further possible that other biomarkers that we did not measure could have a relation to MRI cartilage loss such as matrix metalloproteinase-3, high sensitive C-reactive protein, hyaluronan, CTX-I, keratan sulphate, matrillin-1, and cartilage intermediate layer protein. In addition, longitudinal analyses of biomarker change would be interesting to explore.

Another potential explanation for the lack of association found relates to limitations with the endpoint, namely cartilage loss on MRI. This was measured semiquantitatively with inherent potential observer bias and possible measurement error. Nonetheless, a number of studies have found plausible biologic associations with cartilage loss on MRI in this dataset, including relations to alignment [40], bone marrow lesions [41], and meniscal abnormalities [42]. Thus, any lack of ability to detect a strong association between biomarkers and MRI is unlikely to be a result of limitations in the MRI variable.

With the exception of COMP, if changes in cartilage turnover in patients with symptomatic knee OA are associated with cartilage loss, they do not appear to affect systemic biomarker levels. Where there are other markers such as alignment and bone marrow lesions that are potent predictors of progression, we would not advocate one-time measurement of biochemical markers to predict MRI progression in patients with symptomatic knee OA, with the possible exception of COMP.