Pro-Resolving Mediators in Rotator Cuff Disease: How Is the Bursa Involved?

Abstract

So far, tendon regeneration has mainly been analyzed independent from its adjacent tissues. However, the subacromial bursa in particular appears to influence the local inflammatory milieu in the shoulder. The resolution of local inflammation in the shoulder tissues is essential for tendon regeneration, and specialized pro-resolving mediators (SPMs) play a key role in regulating the resolution of inflammation. Here, we aimed to understand the influence of the bursa on disease-associated processes in neighboring tendon healing. Bursa tissue and bursa-derived cells from patients with intact, moderate and severe rotator cuff disease were investigated for the presence of pro-resolving and inflammatory mediators, as well as their effect on tenocytes and sensitivity to mechanical loading by altering SPM signaling mediators in bursa cells. SPM signal mediators were present in the bursae and altered depending on the severity of rotator cuff disease. SPMs were particularly released from the bursal tissue of patients with rotator cuff disease, and the addition of bursa-released factors to IL-1β-challenged tenocytes improved tenocyte characteristics. In addition, mechanical loading modulated pro-resolving processes in bursa cells. In particular, pathological high loading (8% strain) increased the expression and secretion of SPM signaling mediators. Overall, this study confirms the importance of bursae in regulating inflammatory processes in adjacent rotator cuff tendons.

Keywords: mechanical stress/loading; pro-resolving mediators; resolution of inflammation; rotator cuff disease; subacromial bursa.

Figure 1

Study design to investigate the role of the subacromial bursa in pro-resolving processes at the human rotator cuff. The yellow-to-red bursa tissue indicates the different disease stages of the rotator cuff. The image was created with graphics from Servier Medical Arts. SPM: specialized pro-resolving mediators.

Figure 2

Gene expression of pro- and anti-inflammatory pathway targets in bursae of intact rotator cuff (n = 14), moderate disease (n = 12) or severe disease (n = 18) measured using real-time PCR. (A) Heatmap derived from normalized gene expression values by collapsing the individual values as median of each group. Rows (genes) were centered, and unit variance scaling was applied to rows. Rows were clustered using maximum distance and average linkage. (B) Scatter plot of selected genes (IRF-1, IL6, IL8, IL1β) with significant differences between groups (highlighted with a red box in (A)). Gene expression of target genes was normalized to the three reference genes Cyclophilin A (Ppia), Hypoxanthin-Phosphoribosyl-Transferase (HPRT) and 18s and calculated using an efficiency corrected equation. Dunn’s Multiple Comparison test was conducted, with * p < 0.05, ** p < 0.01 and *** p < 0.001 indicating significant differences. ALOX5: Arachidonat-5-Lipoxygenase; IRF1: interferon regulatory factor 1; IL-8: Interleukin 8; IL-1β: Interleukin 1β; FGL2: fibrinogen-like protein 2; CISH: cytokine inducible SH2-containing protein; IL-6: Interleukin 6; APOL3: Apolipoprotein L 3; VAMP5: vesicle-associated membrane protein 5; SRRM2: serine/arginine repetitive matrix 2; CD1D: Cluster of Differentiation 1D; IL-10: Interleukin 10; IDO1: indoleamine 2,3-dioxygenase 1; FPR2: formyl peptide receptor 2; ChemR23: Chemerin Receptor 23; PTX3: Pentraxin 3; CD206: Cluster of Differentiation 206; CXCL11: C-X-C motif chemokine ligand 11; MCP-1: monocyte chemoattractant protein 1; WARS: tryptophanyl-tRNA synthetase; TNF-α: tumor necrosis factor α; ANXA1: Annexin A1; TGM2: Transglutaminase 2; GPR18: G-protein-coupled receptor 18; CCL20: C-C motif chemokine ligand 20; CD163: Cluster of Differentiation 163; FPR1: formyl peptide receptor 2.

Figure 3

Analysis of SPM receptors ChemR23 and FPR2 in histological sections: (A) Exemplary immunofluorescence image of a bursa sample from the severe disease group showing single staining channels for Dapi, ChemR23, FPR2 and CD45, as well as channel overlap (merged), and the respective result after image analysis (Makro). Scale bars represent 50 µm. (B) Quantification of immunofluorescence staining of total ChemR23, FPR2 and CD45 (single- + double- + triple-positive staining area) in bursae of intact controls (n = 8), moderate disease (n = 7) and severe disease (n = 8) using a self-designed image analysis tool. (C) Distribution of immunofluorescence staining area of single-positive ChemR23, FPR2 and CD45 signal; double-positive ChemR23/FPR2, ChemR23/CD45 and FPR2/CD45 signal; or triple-positive ChemR23/FPR2/CD45 signal. The staining areas are given in % to the total positive staining area.

Figure 4

Quantification of cells isolated from bursa tissue using flow cytometric analysis. (A) CD45+ immune cells (T cells (CD3+), macrophages (CD68+), NK cells (CD56+)) and (B) CD45− non-immune cells (endothelial cells (CD31+), fibroblasts (CD90+)) in all investigated bursae (n = 20) and distribution of (C) T cells, (D) endothelial cells, (E) macrophages and (F) fibroblasts between intact (n = 7–8), moderate (n = 5) and severe (n = 6–7) disease given as percentage to CD45+ or CD45− populations. (G–I) Percentages of (G) ChemR23+, (H) FPR2+ and (I) ChemR23+FPR2+ cells among the main analyzed immune and non-immune cell populations. (J–L) Distribution of (J) ChemR23+ macrophages, (K) ChemR23+FPR2+ macrophages and (L) CD80+ macrophages between intact, moderate and severe disease. Statistics: Dunn’s Multiple Comparison test and Mann–Whitney U Test given as median with interquartile range. * p < 0.05, ** p < 0.01, *** p < 0.001; a dashed line indicates a trend (p < 0.1).

Figure 5

Release of pro-resolving and inflammatory mediators in tissue culture of bursae of patients with intact rotator cuff (n = 16), moderate disease (n = 16) or severe disease (n = 29) measured using ELISA and Luminex multiplex assay. (A) Heatmap derived from normalized protein concentrations by collapsing the individual values as median of each group. Rows (proteins) were centered, and unit variance scaling was applied to rows. Rows were clustered using maximum distance and average linkage. (B) Scatter plots of release of selected pro-resolving mediators (highlighted with a red box in (A)) showing the protein concentrations of ANXA1, LXA4, RvD1 and RvD2 as medians with interquartile ranges. Dunn’s Multiple Comparison test was conducted, with p < 0.1 indicating trend for differences with dashed line.

Figure 6

Influence of bursa-released factors on tenocytes: (A) Experimental design, (B) protein concentrations measured using ELISA or Luminex^® Assay in high ANX, low ANX, high RvD and low RvD groups (each n = 6). (C–M) Results of either unstimulated tenocytes as negative control (NC) or tenocytes pre-stimulated with 3 ng/mL IL-1β 24 h before adding bursa supernatants with high or low ANX or RvD concentrations. Values were normalized to the IL-1β stimulation group (set to 1). (C) Cell viability measured using Presto Blue^® assay, (D) migration analyzed using scratch assay, (E) Col I secretion measured using ELISA. (F–M) Gene expression of target genes (Col1A1, Col3A1, ANXA1, ChemR23, GPR18, IL-6, IL-1β) normalized to the housekeeping gene HPRT, calculated using an efficiency corrected equation and given as fold to the IL-1β stimulation group. Statistics: Dunn’s Multiple Comparison test, * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001, dashed lines indicate trends (p < 0.1).

Figure 7

Role of mechanical loading on SPM signaling mediators in bursa cells (n = 6). (A) Exemplary results shown as histograms of surface marker expression of unstimulated bursa cells and after mechanical stimulation with 2% and 8% strains. In addition, 100 ng/mL LPS was used as positive control to show the general immune responsiveness of the bursa cells. FMO (fluorescence-minus-one) controls were included for proper gating. (B) Gene expression of target genes (ANXA1, FPR2, ChemR23, GPR18, IL-6, IL-1β, Col1A1, MMP-1, MMP-2, TIMP-1) normalized to the reference gene HPRT, calculated using an efficiency corrected equation and given as fold to the unstimulated control. (C) Secretion of ANXA1 and RvD1 normalized to unstimulated control (set to 1). Statistics: Dunn’s Multiple Comparison test, * p < 0.05, ** p < 0.01; a dashed line indicates trends (p < 0.1).