Sphingosine-1-phosphate protects endothelial glycocalyx by inhibiting syndecan-1 shedding

Abstract

Endothelial cells (ECs) are covered by a surface glycocalyx layer that forms part of the barrier and mechanosensing functions of the blood-tissue interface. Removal of albumin in bathing media induces collapse or shedding of the glycocalyx. The electrostatic interaction between arginine residues on albumin, and negatively charged glycosaminoglycans (GAGs) in the glycocalyx have been hypothesized to stabilize the glycocalyx structure. Because albumin is one of the primary carriers of the phospholipid sphingosine-1-phosphate (S1P), we evaluated the alternate hypothesis that S1P, acting via S1P1 receptors, plays the primary role in stabilizing the endothelial glycocalyx. Using confocal microscopy on rat fat-pad ECs, we demonstrated that heparan sulfate (HS), chondroitin sulfate (CS), and ectodomain of syndecan-1 were shed from the endothelial cell surface after removal of plasma protein but were retained in the presence of S1P at concentrations of >100 nM. S1P1 receptor antagonism abolished the protection of the glycocalyx by S1P and plasma proteins. S1P reduced GAGs released after removal of plasma protein. The mechanism of protection from loss of glycocalyx components by S1P-dependent pathways was shown to be suppression of metalloproteinase (MMP) activity. General inhibition of MMPs protected against loss of CS and syndecan-1. Specific inhibition of MMP-9 and MMP-13 protected against CS loss. We conclude that S1P plays a critical role in protecting the glycocalyx via S1P1 and inhibits the protease activity-dependent shedding of CS, HS, and the syndecan-1 ectodomain. Our results provide new insight into the role for S1P in protecting the glycocalyx and maintaining vascular homeostasis.

Keywords: MMP; S1P; albumin; endothelial cells; glycocalyx.

Fig. 1.

Plasma protein protects the endothelial glycocalyx: the potential role of S1P. Immunofluorescence images (left) and coverage analysis (right) of heparan sulfate (HS; A) and chondroitin sulfate (CS; B) in the presence of p-DMEM (DMEM with 5% FBS and 0.5% BSA), DMEM with 2% FBS and 0.2% BSA, and f-DMEM (DMEM without FBS or BSA) for 2 h. C: the S1P₁ receptor antagonist W146 abolished the protection of the glycocalyx (HS and CS) by plasma protein. Note that a 10-min pretreatment with f-DMEM containing W146, which has no effect on coverage, is necessary for the reduction of CS in the p-DMEM after 2 h. Without the pretreatment, the decrease of CS coverage takes longer. For example, we observed a similar level of CS coverage in p-DMEM with W146 after 12 h. Scale bar: 20 μm. Significant difference (n = 3): *P < 0.05; **P < 0.01.

Fig. 2.

Effects of FBS, BSA, and essential fatty acid-free BSA on retention of the glycocalyx. Immunofluorescence images (A) and coverage analysis (B) show the effects of FBS, BSA, and essential fatty acid-free BSA on CS. The average S1P content was 100, 87, and 57 nM in 5% FBS, 0.5% BSA, and 0.5% fatty acid-free BSA, respectively, and 187 nM in p-DMEM (5% FBS plus 0.5% BSA). DMEM, which contains no serum or BSA, was used for preparing the experimental media. The data suggest that S1P is the active component in p-DMEM that protects the glycocalyx. **Significant difference (P < 0.01; n = 3).

Fig. 3.

S1P preserves the phosphorylation of S1P1 after removal of plasma proteins. Cells were treated with p-DMEM, 1 μM S1P, and f-DMEM for 2 h. Then, anti-S1P1 (phospho T236) was used to label the levels of activated S1P1. Immunofluorescence images (A) and the relative fluorescence intensity of phosphorylated-S1P1 (p-S1P1) on these maximum-intensity projections (B) were obtained. Phosphorylation was strong in the presence of p-DMEM and 1 μM S1P but not f-DMEM. **Significant difference vs. p-DMEM (P < 0.01; n = 3).

Fig. 4.

Effects of S1P and S1P₁ receptor antagonist W146 on sGAGs (HS and CS). Coverage analysis of HS (A and B) and CS (C and D) on rat fat-pad endothelial cells (RFPECs), which were treated with p-DMEM or f-DMEM in the presence of S1P (A and C), or treated with p-DMEM or f-DMEM with or without S1P in the presence of W146 (B and D) for 2 h. Cells were treated with p-DMEM or f-DMEM to serve as controls. W146 abolishes the protective roles of p-DMEM and 1 μM S1P. This is also observed in human coronary artery ECs (passage 6, Cell Applications), indicating the protein effect is not a cell-specific phenomenon (data not shown). The dose-response relationships of S1P with the retention of HS (E) and CS (F) on RFPECs. Significant difference: **P < 0.01 (for A–D); **P < 0.01 vs. 0 μM S1P (for E and F); ##P < 0.01 vs. 0.1 μM S1P (n = 3).

Fig. 5.

The sGAG levels in media. The sGAG (HS and CS) released into media in the presence of p-DMEM and f-DMEM with or without S1P for 2 h was determined using a Blyscan sGAG assay. p-DMEM and f-DMEM plus S1P protect against sGAG release into the media. *Significant difference (P < 0.05; n = 6).

Fig. 6.

Effects of plasma protein and S1P on syndecan-1 and glypican-1. A: the core-domain structure of glypican-1 and syndecan-1. GPI, glycosylphosphatidylinositol; TM, transmembrane domain; C1 and C2, conserved regions 1 and 2, respectively; V, variable region. The solid and broken lines indicate the HS and CS attachment sites, respectively. *Antibody mark regions of H-95 (glypican-1), H-174 (syndecan-1), and N-18 (ectodomain of syndecan-1). Immunofluorescence images (left) and coverage analysis (right) of glypican-1 (B), syndecan-1 (C), and syndecan-1 ectodomain (D) in the presence of p-DMEM and f-DMEM for 2 h. E: the syndecan-1 ectodomain (N-18) in media was analyzed by Western blotting. Control, fresh medium; Treat, the medium collected after treating with cells for 2 h. Only the syndecan-1 ectodomain is shed from the surface to the media. Scale bar: 20 μm. **Signficant difference (P < 0.01; n = 3).

Fig. 7.

Role of MMPs and MMP activity. f-DMEM induced shedding of CS (A), and syndecan-1 ectodomain (B) was abolished by the general MMP inhibitor GM6001 after 2 h (n = 3). GM6001 NC served as negative control. C: f-DMEM-induced activation of MMPs was suppressed by S1P (2 h) (n = 4). Data shows the involvement of MMPs in CS shedding. **Significant difference (P < 0.01).

Fig. 8.

The separate and concurrent effects of specific MMP inhibitors on CS. Before staining of CS, cells were treated with p-DMEM, 400 nM MMP-9 inhibitor alone, 1 μM MMP-13 inhibitor alone, and f-DMEM for 2 h (A and C), or were treated with p-DMEM, 400 nM MMP-9 inhibitor and 1 μM MMP-13 inhibitor together, and f-DMEM for 2 h (B and D). MMP-9 plus MMP-13 can account for all CS shedding. Scale bar: 20 μm. **Significant difference vs. p-DMEM control (P < 0.01; n = 3).

Fig. 9.

Summary of the role of plasma protein (S1P) in glycocalyx protection. Starting from the left: serum albumin carries S1P, which can activate the S1P₁ receptor. The activation of S1P₁ inhibits the activity of MMPs and abolishes MMP activity-dependent syndecan-1 ectodomain shedding. Starting from the right: the S1P- S1P₁ receptor agonist activity can be blocked by W146. After the bound S1P on the cell surface is cleared by removal of plasma protein (or S1P), the inhibition of MMP activity (MMP-9 and -13) is attenuated (broken line), and this induces the shedding of syndecan-1 by cleaving its ectodomain.