Serological identification of fast progressors of structural damage with rheumatoid arthritis

Rheumatoid arthritis (RA) is a chronic autoimmune disease characterized by inflammation and massive tissue destruction in multiple synovial joints. Aggressive treatment for patients whose disease is rapidly progressing (progressors) may preserve joint function. However, it is very difficult using standard technologies to identify progressors. Identification of progressors is also important as most anti-inflammatory treatments are associated with side effects emphasizing that only those who are in most need of aggressive treatment (that is, progressors) should be appropriately treated, which may improve the benefit-risk ratio of these therapies.

In RA, the periarticular bone and cartilage undergo major destructive changes, whereas the altered turnover of the synovial membrane results in an expansion of the synovial membrane. The extracellular matrix (ECM) is the most prominent component of these tissues and the most abundant ECM protein is type I collagen. However, type II collagen is the most abundant protein in cartilage and the most investigated collagen with arthritides. Destruction of the ECM is mediated by enzymatic cleavage by different proteases, though predominantly by matrix metalloproteinases (MMPs). MMPs have been shown to be highly up-regulated with RA [1‐3] and the tissue destruction with specific MMPs elevated with disease is a central player in inflammatory disease, such as RA [3, 4]. The action of MMPs on ECM results in the release of small protein 'fingerprints' into the circulation, which may be used as markers of tissue destruction [5, 6]. It seems reasonable to hypothesize that connective tissue destruction mediated by MMPs could be associated with disease progression in RA.

Type I collagen can be degraded by several proteinases and in the process yields various products, namely CTX-I, ICTP, PINP and C1M (Figure 1). CTX-I is a measurement of Cathepsin K-mediated destruction of type I collagen and is the standard measure for bone resorption [7]. ICTP is a triple cross-linked carboxyterminal telopeptide of type I collagen generated by MMPs and destroyed by Cathepsin K and, therefore, not related to bone type I collagen [7, 8]. Lastly, PINP is widely used as a measure of type I collagen formation in bone [9‐12].

We recently developed and validated an enzyme-linked immunosorbent assay (ELISA) using a monoclonal antibody detecting a degradation fragment of the helical domain of type I collagen mediated by MMP-cleavage and destroyed by Cathepsin K (C1M) [13]. This ELISA does not recognize intact type I collagen, other collagens, or the products of other proteinase-mediated collagen degradation. C1M is a product of MMP-2, -9 and -13 cleavage of type I collagen, but not MMP-3, -8 or Cathepsin K cleavage. Therefore, C1M is not a biomarker of hard tissue (that is, bone), but a biomarker of soft tissue destruction, due to the destruction of the fragment by Cathepsin K. Therefore, it represents another pool of type I collagen degradation than PINP, CTX-I and ICTP, namely, exclusively MMP-mediated connective tissue destruction [13].

The aim of the present study was to investigate whether MMP-mediated type I collagen destruction, measured by serum levels of C1M, could be a novel biochemical marker for identification of progressors in RA. We investigated whether baseline levels of serum C1M were correlated with structural status and progression of RA, by measuring change in joint space narrowing, total Sharp score and erosion, in the placebo group of the biomarker study of the LITHE study [14]. We, furthermore, investigated whether serum C1M levels were altered following treatment of RA with an anti-IL-6R monoclonal antibody, tocilizumab (TCZ), in combination with methotrexate (MTX) in the biomarker study in the LITHE study [14, 15]. Lastly, we investigated if the change in C1M at four weeks was related to change in one year radiographic measures and if including the standard biomarker (C-reactive protein, CRP) of RA increased the predictive value of C1M.

RA is a complex disease in which progression is very heterogeneous and non-continuous, in which flares may be important drivers of severe progression. Selecting the patients who will benefit from a specific treatment would be an advantage in inflammatory arthritis diseases in particular, as most, if not all, anti-inflammatory treatments are associated with severe, but manageable, side effects. In this study we investigated if a novel assay detecting a protein fingerprint of type I collagen degradation could be a novel biochemical marker for identification of progressors in RA. We also investigated the change in serum C1M levels following treatment with TCZ and the predictive value of C1M in combination with CRP. We found that 1) C1M levels were associated with disease progression, 2) C1M levels were correlated with disease activity at baseline, 3) baseline C1M was a significant predictor of structural progression, 4) TCZ dose-dependently inhibited the level of C1M as early as two weeks after initiating treatment, and 5) change in C1M at 16 weeks after treatment initiation was related to the level of radiographic changes over one year. The current findings contribute to the literature and the scientific field by demonstrating that quantifying specific tissue destruction may aid in selecting the RA patients in most need of aggressive treatment.

Serological biochemical markers are receiving increased attention in drug development and for guiding evidence-based patient management decisions [18]. A range of traditional markers may provide valuable in managing RA, such as CRP, ESR, and the interleukins-1 and -6. However, these markers are systemic biomarkers. A newer class of molecular biochemical markers, also assessed by simple ELISA technologies, provides independent and additional information reflecting structural changes within the tissue [5]. These markers - protein fragments - are referred to as protein fingerprints, because they are end results of specific pathological events, proteolytic degradation of, that is, the joint structure [5, 6]. The protein fingerprint degradation markers of type I, II and III collagen have been studied and all three have been shown to be descriptive of disease progression and response to anti-inflammatory therapy [19]. In this study, we further show that C1M is highly correlated to CRP levels (Table 2), indicating that type I collagen degradation in RA is related to inflammation. Therefore, exploring C1M in RA will indicate the state of the specific inflammatory structural changes and not the systemic state of the inflammation. Due to the relationship between C1M and CRP, we have not investigated the difference between the biomarkers in predicting progressors of disease.

Interpreting biochemical markers found in serum is associated with many limitations, as several different tissues at different rates may produce, and thus contribute to the total pool of, the molecular marker. Type I collagen is abundant in many tissues throughout the body [5]. Bone in particular may be a prime candidate contributing to the serological C1M pool as it is naturally being constantly renewed. However, C1M has been shown not to be correlated with bone metastases, which have high and very aggressive bone resorption, and, furthermore, not to recognize MMP-3, -8 and Cathepsin K-cleaved type I collagen [13]. Thus, C1M is a biochemical marker with increased accuracy for connective tissue turnover. Other biomarkers of type I collagen degradation have previously been described in the literature in relation to RA. ICTP, a C-terminal degradation fragment of the mature type I collagen, has been shown to be related to the severity of RA [20, 21] as well as disease activity [22]. It has been shown to be useful for assessing the effect of potentially disease modifying therapies [23]. ICTP in synovial fluid and serum has been shown to be correlated to radiographic measures of joint destruction [22, 24]. However, RA disease duration has shown to be a cofounder in ICTP levels [25]. ICTP has also been described to be correlated to disease activity in reactive arthritis [26]. Another marker of type I collagen, CTX-I, a marker of Cathepsin K-mediated type I collagen degradation reflecting bone resorption, has previously been described to predict disease progression, albeit only in the worst quartile [27, 28]. Furthermore, it has been shown in a study of early RA patients that CTX-I at baseline, adjusted for cofounders, was significantly associated with radiographic progression in the RA MRI erosion score and classical radiographs over one year [20]. A drawback of this study, however, was that there was a considerable overlap in CTX-I between progressors and non-progressors. The same group of investigators has shown that the relationship between elevated serum levels of CTX-I and subsequent joint destruction still existed at 10-year follow-up. However, the relationship was weaker [29].

In the assessment of C1M as a marker of structural progression in RA, we have utilized radiographic changes (JSN, mTSS and erosion score) after 24 and 52 weeks. The significance level of the correlation between C1M and radiographic scores was strongest at 52 weeks, especially with mTSS. This is due to the low sensitivity of the radiographic scores to changes in joints over a short period [30]. The low sensitivity led to high variation in scores at 24 weeks. The variation decreased when assessing the radiographic scores at 52 weeks and hence increasing the statistical power. JSN is a visualization of the space between the joints and a reduction in JSN indicates loss of cartilage and meniscal damage, hence only a part on the JSN could is associated with type I collagen, which partly could explain the lack of correlation.

A limitation of this study is the effect of concomitant intake of non-steroidal anti-inflammatory drugs and corticosteroid. These medications could have an influence on the level of C1M, but this contribution cannot be studied in this setting. Especially, prednisolone intake influences the level of inflammation and joint damage [31, 32]. However, patients already take stable doses of prednisolone and/or other drugs at the time they enter the study, hence the effect of the drugs might be obliterated in the baseline measurements of C1M. Moreover, the study population of this investigation has long disease duration at baseline and profound joint damage, hence the study population is in an advanced disease stage. It is therefore not a disease population in which a biomarker of progression is needed. A study population with shorter disease duration is more ideal for this type of study. However, the disease progression of these patients is not as profound and an even larger study population would be needed. Our present study aims at identifying progressors of disease by the biomarker by which this study population is adequate, due to the rapid progression.

We demonstrated that a simple serum biochemical marker measurement is able to predict fast progressors in RA and we also illustrated that C1M could be used as an efficacy marker. It is critical to effectively treat tissue destructive diseases, such as RA, early to suppress inflammation and prevent destruction of the joint.