Myeloperoxidase as a novel diagnostic biomarker in equine infectious joint disease

Introduction: Equine joint infection is a life-threatening disorder. A presumptive diagnosis is commonly based on synovial fluid total and differential white blood cell counts together with total protein levels and is confirmed with a positive bacterial culture. However, cytological results are not always conclusive and up to 45% of the cultures may remain negative, depending on the method used. Efforts to define novel biomarkers for the discrimination of infectious and non-infectious joint disease, such as the synovial fluid D-dimer and serum amyloid A concentration and the neutrophil viability, were recently reported (Jacobsen et al., 2006; Wauters et al., 2010; Ribera et al., 2011). In the current study, myeloperoxidase (MPO), an anti-bacterial enzyme of the neutrophils involved in the production of reactive oxygen species, was evaluated as a candidate biomarker. Detection of MPO has rapidly gained importance in the diagnosis of equine gastro-intestinal, pulmonary and orthopaedic disease (Fietz et al., 2008). We therefore aimed to investigate the synovial fluid levels of both the total (= sum of enzymatically active and inactive fractions) MPO and enzymatically active MPO fraction in patients by immunological detection. Materials and methods: Four patient groups were included: (1) 29 healthy horses (29 joints), (2) 24 horses with OCD-caused synovitis (24 joints), (3) 16 horses with traumatic non-infectious synovitis (16 joints) and (4) 17 horses with culture confirmed infectious synovitis (19 joints). Diseased joint samples were obtained from client-owned horses, while healthy joints were obtained from slaughterhouse horses. Samples were collected in EDTA tubes after standard surgical preparation of the synoviocentesis sites. The total nucleated and differential cell count was performed with an automated blood counter prior to cell fraction isolation. Synovial fluid viscosity was therefore decreased with hyaluronidase before collecting the supernatants by 10 min centrifugation at 500 g at 4°C and storage at - 80°C prior to analysis. Total MPO was measured with a commercial horse-specific sandwich ELISA, while the active MPO fraction was monitored by the SIEFED (specific immunological extraction followed by enzymatic detection) technique, both according to the manufacturers’ recommendations but with adaptation of the dilution factor (Franck et al., 2005 and 2006). The ELISA and SIEFED methods resulted in skewed data that were evaluated through non-parametric statistical analysis (Kruskal-Wallis with Post-Hoc Dunn’s multiple comparison test). Results: Statistically significant differences were observed between patient groups for both total and active MPO (p < 0.0001). As illustrated in Figure 1 and 2, samples of infected joints contained significantly more MPO protein compared to samples from healthy horses (p < 0.001), OCD horses (p < 0.001) and horses diagnosed with severe traumatic non-infectious synovitis (p < 0.01). The same was true for the active MPO concentration but SIEFED determination also showed a significantly higher MPO value in traumatic non-infectious synovitis samples compared to the healthy control group (p < 0.01). Routine synovial fluid analysis results are shown in Table 1 for each patient group. No correlation was observed between the synovial fluid MPO concentrations and the total and differential nucleated cell counts and only a weak correlation could be observed with the synovial fluid total protein levels. Discussion: Synovial fluid total and active MPO levels are demonstrated to be higher in infected compared to either non-infected traumatic or OCD-related or healthy samples. Interestingly, the MPO concentration was not correlated with the synovial fluid cell counts or differentiation and therefore we suggest MPO might be a valuable biomarker to include in the diagnostic tool kit for infectious arthritis, particularly in these cases where synovial fluid cytological results are not conclusive for infectious synovitis. Both techniques (cytology and MPO determination) could therefore be used complementary to each other to predict more accurately whether a sample is infected. A larger clinical study should however still be undertaken to determine an MPO cut-off value that can be used in clinical circumstances. Results obtained for ELISA and SIEFED were comparable and thus only one of both techniques should be adapted for clinical use. Both techniques allow quasi immediate results, in contrast to bacterial culture, and are easy to perform. The potential of a customized low-cost one-sample ELISA for quick and easy reading (as in e.g. the “snap-foal” test) seems to favour determination of total MPO. Nevertheless, further research should first elucidate whether the level of active MPO could be correlated with other clinically interesting parameters such as joint damage and long-term outcome, as this would favour commercialisation of SIEFED detection compared to ELISA