Synovial fluid concentration of gremlin1 but not frizzled-related protein is increased in patients after acute knee injury

Abstract

Purpose: Degenerative joint diseases are characterized by a disturbance in joint homeostasis, which is poorly understood at the molecular level. Previously, we have identified the secreted WNT antagonists dickkopf 1 homolog (Xenopus laevis) (DKK1) and frizzled-related protein (FRZB) and the BMP antagonists Gremlin1 (GREM1) as regulators of cartilage homeostasis by preventing hypertrophic differentiation of articular chondrocytes, consistent with their role in articular cartilage homeostasis. In this study we tested the hypothesis that health status of the joint is reflected by the expression levels of FRZB and GREM1 in synovial fluid. To test this hypothesis, we performed a cross sectional analysis of synovial fluid samples collected from patients with acute and chronic knee injury and compared these with synovial fluid samples from knee healthy controls. Methods: Synovial fluid was aspirated (Lund University, Sweden) from patients with acute (n =106, mean age 30 years [range 17-57], 0-77 days from injury) and chronic (n = 50, mean age 38 years [18-70] 1-37 years after injury) knee injuries, and from knee healthy subjects (n = 8, mean age 28 years [17-85]). Synovial fluid concentrations of FRZB and GREM1 were determined using ELISA (FRZB, RD GREM1, Bio-connect diagnostics). Concentration of cartilage biomarkers in synovial fluid; ARGS-aggrecan (in-house), cartilage oligomeric matrix protein (COMP; BioVendor) and collagen epitope C2C (IBEX), and proinflammatory cytokines (IL-1β, IL-6, IL-8 and TNF-α; Meso Scale Discovery) were analyzed by immunoassays. Mann-Whitney tests for group comparisons and Spearman’s rank test (rS) for correlation analysis was used. Results: Synovial fluid concentration of FRZB in the acute injury group was no different from the level in the reference group (p = 0.99), but was 2-fold (based on median concentrations) higher than the level of the chronic injury group (p = 0.023) (Fig. 1a). Synovial fluid concentrations of GREM1 in the acute and chronic injury groups were both higher than the level of the reference group, 2-fold (p = 0.003) and 1.8-fold (p = 0.031), respectively (Fig. 1b); also, in these patients the GREM1 concentration was negatively correlated to time after injury (rS = -0.224, p = 0.005). The FRZB and GREM1 concentrations of the acute and chronic injury groups were not influenced by gender (p = 0.550, p = 0.825) or by age (rS =-0.146, p = 0.068; rS = -0.137, p = 0.089), although, the FRZB concentration was significantly decreased in knee healthy subjects with age (rS= -0.833, p = 0.005). In the acute and chronic injury groups, there was no correlation between FRZB and GREM1 concentrations in synovial fluid (rS = 0.04, p = 0.585). Also, for a portion of the acute injury group (n = 52) synovial fluid FRZB concentration were correlated with COMP levels (rS = 0.284, p = 0.041) but not with other cartilage markers (C2C p = 0.55, ARGS-aggrecan p = 0.311) nor with any cytokine levels (p> 0.05 for all cytokines); the synovial fluid GREM1 concentration did not correlated with any cartilage marker nor with cytokine levels (p > 0.05 for all markers). Conclusions: Compared to chronic injured knees, acute injured knees have higher synovial fluid concentrations of FRZB and GREM1. Their levels were not influenced by gender or age, indicating that the differences of FRZB and GREM1 concentrations between acute and chronic injured knee joints depend on differences in disease conditions. Persistent synovial fluid levels of FRZB in acute injured knee joints might be a way for cells trying to protect the cartilage from degradation by inhibiting the Wnt signaling pathway. Similarly, high synovial fluid levels of GREM1 in injured knee joints (acute and chronic) suggests of an effort to maintain cartilage homeostasis by inhibiting BMP signaling.