Decreases in serum levels of S100A8/9 (calprotectin) correlate with improvements in total swollen joint count in patients with recent-onset rheumatoid arthritis

Rheumatoid arthritis (RA) is a chronic inflammatory autoimmune disease characterised by synovitis and joint destruction in which the infiltration of inflammatory cells, the activation of synovial fibroblasts and the production of a wide range of inflammatory mediators play significant roles [1, 2]. However, the exact pathological processes involved in the initiation of RA remain incompletely understood. Very early RA is suggested to represent an immunopathologically distinct phase of the disease in which a "window of opportunity" for early drug intervention with the potential to prevent joint damage may exist [3]. Recent studies have shown that the development of established RA in patients in the early stages of the disease can be predicted by using clinical and serological measures [4‐6]. Therefore, a better understanding of the pathological mechanisms and biomarkers during this early phase would be an important way to determine possible new therapeutic targets and to tailor therapy to ensure optimal treatment for individual patients.

S100 calcium-binding proteins are multifunctional proteins that are implicated in the regulation of a variety of cellular activities [7]. The most familiar S100 proteins, myeloid-related proteins S100A8/9 (calprotectin) and S100A12 (calgranulin C), have recently been proposed as "alarmins," which are the endogenous molecules that signal the early phase of tissue and cell damage [8]. The S100 proteins are expressed predominantly by neutrophils, monocytes and activated macrophages, and increased S100 levels have been demonstrated in several inflammatory diseases [9]. S100A8/9 and S100A12 are increased locally at sites of inflammation as well as in the circulation of patients with RA [10‐13]. Moreover, a tight correlation between S100 proteins and laboratory and clinical markers of disease activity has been demonstrated in patients with different arthritides [13‐16]. In addition, S100A8/9 and S100A12 were shown to be decreased locally in synovial tissue as well as in the blood in response to different anti-inflammatory therapies, including TNFα inhibitors, and they were upregulated weeks before relapse became clinically apparent in patients with previously well-controlled disease [16‐19]. S100A8/9 was associated with measures of joint damage in one cross-sectional study [20]. More importantly, longitudinal data demonstrated that S100A8/9 was a good prognostic biomarker for long-term radiographic joint progression in patients with established RA [21].

However, S100 proteins have not yet been studied in treatment-naïve RA patients. Therefore, we explored the following: (1) the levels of S100 proteins in patients with recent-onset RA, (2) the effect of conventional treatment on the levels of serum S100 proteins, (3) the association between S100 proteins and disease activity and (4) a potential role of S100 proteins as surrogate predictive markers in a short-term longitudinal study.

Table 1 shows the baseline characteristics of the patients and healthy controls. Overall, three patients with recent-onset RA had erosions at baseline. Prior to treatment, 22 patients had highly active disease (mean DAS28 > 5.1), 19 patients had moderate disease activity (mean 3.2 < DAS28 ≤ 5.1) and 2 patients had low disease activity (mean 2.6 ≤ DAS28 < 3.2). Initially, DMARD treatment was started in 42 patients. Thirty-five patients received methotrexate (mean dose at month 3: 14.86 mg/week; range: 7.5 to 20 mg/week), six received sulphasalazine, one received leflunomide and thirty-eight received GCs (prednisone or equivalent; initial mean daily dose: 8.9 mg/day; range: 2.5 to 20 mg/day). After three months of treatment, a significant reduction in disease activity was observed (mean DAS28: 5.3 ± 1.5 to 2.8 ± 1.3; mean SJC: 10.0 ± 9.2 to 1.7 ± 3.6; mean CRP: 16.2 ± 19.8 to 4.2 ± 5.8; P < 0.0001 for all comparisons). Thirty-nine patients (90.7%) achieved good or moderate improvement, and eighteen (41.9%) reached remission according to the EULAR response criteria [24].

To the best of our knowledge, this study is the first to show elevated serum levels of S100 proteins in patients with recent-onset DMARD/GC-naïve RA, the association between S100 proteins and disease activity and normalisation of S100A8/9 in patients who achieve remission after conventional treatment. Furthermore, we have demonstrated that decreases in S100A8/9 levels are associated with clinical improvement in the number of affected joints.

S100 proteins were previously found to be upregulated in the inflamed synovial tissue, synovial fluid and blood of patients with established RA [10‐13]. However, in this study, we have shown for the first time significantly increased serum levels of both S100A8/9 and S100A12 proteins in patients with recent-onset RA who had not yet been exposed to conventional treatment. This is consistent with previous reports showing elevated circulating S100 proteins in patients with previously established disease [10‐13]. However, compared to previous studies [19, 25], the levels of S100A8/9 measured in our study were higher, which can be explained by the use of various antibodies to detect different epitopes in different ELISAs. In our commercial assay, an antibody detecting an epitope which is dependent on S100A8/S100A9 heterocomplex formation and calcium binding was used. Nevertheless, this did not affect the comparative study in general. It has previously been demonstrated that the S100A8/9 and S100A12 proteins show decreased levels in the peripheral circulation in response to different anti-inflammatory therapies [16, 17]. Consistent with such findings, we have shown in the present study that the conventional treatment of patients with DMARD/GC-naïve early-stage RA reduces serum levels of S100 proteins to the normal levels found in healthy individuals. As expected, patients with very active disease at baseline had higher levels of S100A8/9 and S100A12 proteins that decreased more significantly as the disease improved than did patients with moderate disease activity. Furthermore, an association between both serum S100 proteins in patients with recent-onset RA is in agreement with previous reports on established RA [26, 27], which have indicated that S100A8/9 and S100A12 proteins may be coregulated early in the disease process.

In agreement with previous reports on established RA [10‐13, 15, 20, 21], we found an association of both S100A8/9 and S100A12 proteins with laboratory and clinical markers of disease activity in patients with recent-onset RA. Interestingly, this association was more pronounced for S100A8/9 than for S100A12. Higher baseline levels of S100A8/9 predicted decreased CRP levels and improvements in the total number of swollen joints over time. Moreover, decreases in S100A8/9 levels were directly related to clinical and laboratory improvements over time. Multivariate regression analysis revealed that baseline S100A8/9 values were not more predictive of disease activity than traditional biomarkers such as CRP and that they were only marginally acceptable as predictors of baseline disease activity. However, both CRP and S100A8/9 levels were found to be weak predictors for the baseline number of swollen joints. Interestingly, however, changes in the levels of S100A8/9, but not CRP, were associated with changes in the total number of swollen joints over time. It is evident that S100 proteins are extensively produced by activated immune cells of the synovial membrane and synovial fluid in affected joints and pass into the blood circulation [12, 13]. Our data support the hypothesis that S100A8/9 protein represents a suitable marker that provides important information about the extent of local inflammation in affected joints, as shown by the strong associations between the decrease in S100A8/9 levels and improvements in swollen joint counts. Although researchers in some studies have found an association between S100 proteins and autoantibodies in patients with RA [15, 20, 25], we have not confirmed these data, which may be explained by a stronger association with disease activity than with the autoimmune response in the early phase of recent-onset RA. S100A8/9 protein has recently been demonstrated to predict 10-year radiographic progression in patients with established RA [21]. Although our study was not sufficiently long to assess the effect of changes in S100A8/9 levels on the radiographic progression of early RA, it can be suggested that a decrease in serum levels of S100A8/9 over time, which is associated with improvement in the number of affected joints, might be associated with inhibition of further structural joint damage. This remains to be determined in future studies.

In summary, our data show elevated serum levels of S100 proteins at the onset of RA and a normalisation of S100A8/9 levels in patients who achieved remission shortly after the initiation of conventional treatment. Furthermore, decreases in S100A8/9 rather than CRP levels were associated with improvements in the total number of swollen joints over time. Further studies are needed to determine whether S100A8/9 levels may have a predictive value for further structural damage in patients with recent-onset RA.