Complement lectin pathway components MBL and MASP-1 promote haemostasis upon vessel injury in a microvascular bleeding model

Abstract

Background: Complement lectin pathway components, in particular mannan-binding lectin (MBL) and MBL-associated serine proteases (MASPs) have been shown to interact with coagulation factors and contribute to clot formation. Here we investigated the role of MBL and MASP-1 in the haemostatic response following mechanical vessel injury in a human microfluidic bleeding model.

Methods: We studied haemostasis in a microvascular bleeding model in the presence of human endothelial cells and human whole blood under flow conditions. We monitored incorporation of proteins into the clot with fluorescently labelled antibodies and studied their effects on clot formation, platelet activation, and bleeding time with specific inhibitors. Platelet activation was also studied by flow cytometry.

Results: Upon vessel injury, MBL accumulated at the injury site in a well-defined wall-like structure. MBL showed partial colocalisation with fibrin, and strong colocalisation with von Willebrand factor and (activated) platelets. Flow cytometry ruled out direct binding of MBL to platelets, but confirmed a PAR4- and thrombin-dependent platelet-activating function of MASP-1. Inhibiting MBL during haemostasis reduced platelet activation, while inhibiting MASP-1 reduced platelet activation, fibrin deposition and prolonged bleeding time.

Conclusion: We show in a microvascular human bleeding model that MBL and MASP-1 have important roles in the haemostatic response triggered by mechanical vessel injury: MBL recognises the injury site, while MASP-1 increases fibrin formation, platelet activation and shortens bleeding time. While the complement lectin pathway may be harmful in the context of pathological thrombosis, it appears to be beneficial during the physiological coagulation response by supporting the crucial haemostatic system.

Keywords: MBL-associated serine protease-1 (MASP-1); complement system; haemostasis; mannan-binding lectin (MBL); microfluidics.

Figure 1

Microfluidic bleeding model before and after injury. From top left to bottom right: Brightfield and CellMask-stained images of the uninjured and injured vessel. The collagen pouch adjacent to the vessel is used to introduce the vessel injury. The blood then flows into the collagen container. Blood flow is indicated with orange arrows. The bracket indicates the injury site. Scale bar: 50 µm.

Figure 2

MBL and fibrin accumulate upon vessel injury. Microfluidic bleeding model 20 min after injury. From top left to bottom right: Brightfield, anti-MBL signal (green), AlexaFluor647-conjugated fibrin signal (cyan), and merged signals. MBL accumulates at the injury site in a wall-like structure but shows only partial colocalisation with fibrin. Scale bar: 50 µm.

Figure 3

MBL colocalises with activated platelets. Microfluidic bleeding model 60 min after injury. From top left to bottom right: Brightfield, anti-MBL signal (green), anti-CD62P signal (purple), and merged signals. CD62P-expressing activated platelets and MBL showed a complete colocalisation. Scale bar: 50 µm.

Figure 4

MBL colocalises with von Willebrand factor. Microfluidic bleeding model 45 min after injury. From top left to bottom right: Brightfield, anti-MBL signal (green), anti-von Willebrand factor signal (yellow), and merged signals. MBL and von Willebrand factor showed strong colocalisation. Scale bar: 50 µm.

Figure 5

MASP-1 and platelet activation. Whole blood supplemented with corn trypsin inhibitor and acetylsalicylic acid was re-calcified and incubated with rMASP-1cf (10 μg/ml and 50 μg/ml) and fixed at different time points. The entire platelet population was identified with an anti-CD41 antibody conjugated with APC. The activated population was identified with an anti-CD62P antibody conjugated with BV421. (A) Addition of rMASP-1cf increased the percentage of platelets expressing CD62P (shown as mean with SD) in a time- and dose-dependent manner. Statistical analysis: The Shapiro-Wilk test confirmed normal distribution, and p-values for differences between groups were determined with an ANOVA test (ns not significant, * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001). Number of experiments: Control: n=8; 10 μg/ml MASP-1: n=3; 50 μg/ml MASP-1: n=6). (B) The effect of 50 μg/ml rMASP-1cf was tested in the presence of PAR4 inhibitor BMS986120 and/or hirudin. The samples were fixed after 15min. Addition of the PAR4 inhibitor significantly reduced the effect of rMASP-1cf, while addition of hirudin canceled it out completely. Statistical analysis: The Shapiro-Wilk test confirmed normal distribution, and p-values for differences between groups were determined with an ANOVA test (ns not significant, * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001). Number of experiments: Control: n=8; MASP-1 only: n=6; MASP-1+ BMS: n=3; MASP-1+ Hirudin: n=4; MASP-1+ BMS + Hirudin: n=4.

Figure 6

Effects of inhibiting MBL-binding in the bleeding model. To investigate the effects of the MBL inhibitory antibody 3F8, the signal intensity relative to the first time point was monitored for MBL (A), MASP-1 (B), activated platelets (CD62P) (C) and fibrin (D). The area under the curve was calculated using GraphPad Prism (E). Experiments using the inhibitory 3F8 antibody showed a reduced increase of the MBL, MASP-1 and CD62P signal in comparison to the non-inhibitory control antibody 1C10, but no clear difference in fibrin signals was detected. Each experiment was performed three times, data are shown as mean with SD, and p-values for differences between groups were determined with a t-test (ns not significant, * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001).

Figure 7

Effects of MASP-1-inhibition in the bleeding model. To investigate the effects of the MASP-1 inhibitor SGMI-1, the signal intensity relative to the first time point was monitored for activated platelets (CD62P) (A) and fibrin (B). A) Both concentrations of SGMI (20.4 μg/ml; n=4 and 40.8ug/ml; n=2) resulted in a reduced increase of the CD62P signal in comparison to the control (n=6). B) Fibrin deposition was reduced in the presence of SGMI-1 (n=3). The area under the curve (C) was measured using GraphPad Prism. (D) SGMI-1 prolonged the bleeding time (n=3). Data are shown as mean with SD, and p-values for differences between groups were determined with ANOVA or t-test (* p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001).