In this paper, we describe a method to assess endothelial von Willebrand factor release and the subsequent platelet capture under fluid shear stress in response to inflammatory stimuli using an in vitro flow chamber system.
Von Willebrand factor (VWF) is a multimeric glycoprotein coagulation factor that mediates platelet adhesion and aggregation at sites of endothelial damage and that carries factor VIII in the circulation. VWF is synthesized by endothelial cells and is either released constitutively into the plasma or is stored in specialized organelles, called Weibel-Palade bodies (WPBs), for on-demand release in response to hemostatic challenge. Procoagulant and proinflammatory stimuli can rapidly induce WPB exocytosis and VWF release. The majority of VWF released by endothelial cells circulates in the plasma; however, a proportion of VWF is anchored to the endothelial cell surface. Under conditions of physiological shear, endothelial-anchored VWF can bind to platelets, forming a VWF-platelet string that may represent the nidus of thrombus formation. A flow chamber system can be used to visually observe the release of VWF from endothelial cells and the subsequent platelet capture in a manner that is reproducible and relevant to the pathophysiology of VWF-mediated thrombus formation. Using this methodology, endothelial cells are cultured in a flow chamber and are subsequently stimulated with secretagogues to induce WPB exocytosis. Washed platelets are then perfused over the activated endothelium. The platelets are activated and subsequently bind to elongated VWF strings in the direction of fluid flow. Using extracellular histones as a procoagulant and proinflammatory stimulus, we observed increased VWF-platelet string formation on histone-treated endothelial cells compared to untreated endothelial cells. This protocol describes a quantitative, visual, and real-time assessment of the activation of VWF-platelet interactions in models of thrombosis and hemostasis.
Thrombosis is a leading cause of mortality worldwide1 and can develop in response todysregulated platelet activation and thrombin generation in both veinsand arteries. Plasma levels of VWF are a key regulator of blood coagulation, whereby low levels (<50%) result in the bleeding disorder known as von Willebrand disease (VWD)2 and high levels (>150%) are associated with an increased risk of venous3 and arterial4 thrombosis.
VWF is a multimeric glycoprotein synthesized by megakaryocytes and endothelial cells and stored in platelet α-granules and WPBs, respectively. Upon hemostatic challenge, VWF can be released from endothelial WPBs to tether circulating platelets to activated endothelial cells5 or exposed collagen on the vessel wall6. Anchoring of VWF to endothelial cells has been shown to be mediated by P-selectin7 and integrin αvβ38. The subsequent release of platelet α-granule stores can further increase localized VWF concentrations to stabilize platelet-platelet interactions for platelet plug formation, the scaffold needed for the propagation of the coagulation cascade and fibrin deposition. The platelet-binding activity of VWF is regulated by its multimeric structure, with high-molecular weight multimers possessing greater hemostatic activity9,10. In circulation, VWF also acts as a carrier for the coagulation factor VIII.
Fluid shear stress is an essential regulator of VWF physiology. In the absence of shear stress, VWF exists in a globular form, concealing binding domains for platelet glycoprotein Ib adhesion11. When shear stress is present, the cleavage site for a metalloprotease, A disintegrin and metalloprotease with thrombospondin motif (ADAMTS13), is exposed. ADAMTS13 cleaves naked and platelet-decorated VWF strings to regulate multimer size, thereby reducing its hemostatic activity12.
VWF is an acute-phase protein, and numerous stimuli, including hypoxia13, infection14, and proinflammatory cytokines, have been shown to mediate VWF release from endothelial cells. Similar to other inflammatory agents, extracellular histones have also been shown to induce systemic VWF release in mice15,16 and the activation of platelets in vitro17,18,19. This was shown to be dependent upon histone subtype, as differences in lysine and arginine content may influence function15. Our study aims to establish a flow chamber model to investigate the influence of lysine-rich (HK) and arginine-rich (HR) histone subtypes and secretagogues on endothelial VWF release and real-time platelet capture, potential early events in inflammation-induced thrombosis.
This flow chamber methodology recapitulates in vivo interactions between subendothelial collagen, endothelial cells, VWF, and platelets in an in vitro system that is visual, reproducible, and quantifiable. It allows for the real-time assessment of all aspects of the pathway that regulates VWF-platelet interactions, including WPB secretion, platelet activation, and VWF proteolysis. Studies of VWF under controlled shear stress conditions have been used to evaluate VWD mutations that impair VWF release and platelet-binding function20, WPB physiology21, and VWF cleavage by ADAMTS135. We use this methodology to quantify VWF-platelet string formation as a consequence of an inflammatory stimulus: extracellular histones.
These studies were approved by the Research Ethics Board of Queen's University, Canada.
1. Endothelial Cell Stimulation
2. VWF Quantification by Enzyme-linked Immunosorbent Assay (ELISA)
3. Solid-phase Histone-VWF Binding Assay
4. Seeding Endothelial Cells onto Flow Chambers
5. Isolating Platelets from Human Whole Blood
6. Flow Apparatus Assembly
7. Microscope Settings
8. VWF-platelet String Formation
9. VWF-platelet String Quantification
NOTE: The continuous flow of PBS is required to maintain the VWF-platelet string elongation for image analysis. Cessation of flow will result in the VWF becoming globular and will impair VWF-platelet string quantification. It is essential to top up each reservoir with PBS during image capture. HBSS can replace PBS if cell shrinking or detachment is observed during this step.
To directly assess the effect of histones on VWF release from endothelial cells, we exposed confluent BOECs to serum-free medium containing PMA (positive control), UH, HR, and HK for 2 h. We showed that HK induced a 2-fold increase in VWF protein (VWF:Ag) in the medium of treated endothelial cells (Figure 1). Interestingly, when BOECs were stimulated with UH and HR, there was less VWF:Ag detected in the medium than in the untreated condition. We hypothesized that because histones can bind directly to VWF30, they may differentially interfere with VWF:Ag detection in the culture medium. To evaluate histone binding to VWF, we used a solid-phase binding assay. We showed that UH and HR bound more strongly to VWF than HK (Figure 2), supporting the detection interference demonstrated in Figure 1.
Extracellular histones have been shown to activate platelets in vitro19, and our data suggest that histones mediate VWF release from endothelial cells. To characterize the effect of histones on VWF-platelet interactions, we used a flow chamber model of VWF-platelet string formation in which labelled platelets were perfused on stimulated endothelial cells. Representative images were captured post-flow, and the number of VWF-platelet strings were quantified.
We observed that, following the initiation of flow conditions, untreated cells had very few VWF-platelet strings (<1 string per review field), while histamine and PMA-treated cells had an average of 2.10 and 3.10 strings, respectively. Significantly more VWF-platelet strings were formed when cells were treated with HK (3.52 strings, P = 0.019), HR (6.00 strings, P = 0.004), and UH (5.51 strings, P = 0.012). A graphical representation of the VWF-platelet string count is shown in Figure 3G. Real-time VWF-platelet string formation in response to UH compared to untreated cells is shown in Videos 1A and B.
These studies, along with investigations by other groups, show that histones28 and other proinflammatory agents such as histamine5 and cytokines31 stimulate VWF secretion from endothelial cells, mediating subsequent platelet capture in vitro. We demonstrated that the static stimulation of endothelial cells was not sufficient to elucidate VWF release due to the capability of histones to bind and sequester VWF. The design of a flow chamber model of VWF release and platelet adhesion was therefore required to more completely characterize and model the in vivo interactions between the endothelium, VWF, and platelets.
Figure 1: Lysine-rich Histones Induces VWF Release from Cultured Endothelial Cells. BOECs were stimulated with unfractionated histone (UH), lysine-rich (HK), and arginine-rich (HR) histone fractions. HK stimulated VWF release into the culture medium. UH and HR treatment decreased the detection of VWF:Ag in the endothelial cell medium relative to the untreated (Untr) control. N ≥3 individual experiments were performed. The data are shown as the mean values ± SE. *P <0.05, **P <0.01, and ***P <0.001 indicate the significance relative to the untreated condition. This figure has been modified from Michels et al15. Please click here to view a larger version of this figure.
Figure 2: VWF-histone Binding Interferes with VWF Detection In Vitro. Plasma-derived VWF binds more strongly to HR and UH in a dose-dependent fashion compared to lysine-rich (HK) histones in a solid-phase binding assay. The data are shown as the mean values ± SE. N≥3 individual experiments were performed. This figure has been modified from Michels et al15. Please click here to view a larger version of this figure.
Figure 3: Histones Mediate VWF-platelet String Formation. Endothelial cells seeded on flow chamber slides were treated with unfractionated histone (UH), lysine-rich (HK), or arginine-rich (HR) histone fractions or histamine/PMA (positive controls) and were perfused at a shear stress of 4.45 dyn/cm2 for 10 min with washed, fluorescently labeled platelets. Representative images were obtained post-platelet flow from the untreated control (A) and cells treated with histamine (B), Phorbol 12-myristate 13-acetate (PMA) (C), HK (D), HR (E), and UH (F). PMA treatment resulted in an average of 3.1 strings per field of view, comparable to histamine. VWF-platelet strings were quantified from 10 consecutive review fields and averaged for n≥3 individual experiments (G). The data are shown as the mean values ± SE. *P <0.05 and **P <0.01 indicate significance relative to the untreated condition. This figure has been modified from Michels et al15. Please click here to view a larger version of this figure.
Supplemental Videos. VWF-platelet String Formation in Response to Unfractionated Histone (2) Compared to No Treatment (1): This video has been modified from Michels et al15. Please click here to download these videos.
While the physiological relevance of VWF-platelet strings remains controversial due to their rapid dissolution in the presence of the VWF-cleaving protease ADAMTS13, they serve as a quantifiable in vitro model of platelet recruitment by VWF to a site at which a thrombus might form in the presence of localized increases in histone levels5. Moreover, in pathologies lacking ADAMTS13 activity-such as thrombotic thrombocytopenic purpura (TTP)-or in inflammatory microenvironments-where ADAMTS13 activity is inhibited-VWF-platelet strings can be observed in vivo32 and may contribute to platelet microaggregation and leukocyte extravasation, linking the hemostatic and proinflammatory functions of VWF33. Finally, it has been hypothesized that VWF-platelet string formation might be an initiating event in venous and/or arterial thrombosis, forming the nidus around which a pathological thrombus is built.
VWF-platelet string formation in the flow chamber system can be used to study WPB exocytosis, ADAMTS13 cleavage, and the platelet binding capacity of VWF. There are several critical steps to successfully executing these experiments. However, depending on the application of the model, modifications can be made to the protocol.
Endothelial cells from specific vascular beds are heterogeneous in their response to mechanical and biochemical signals34. It may therefore be useful to select a cell type suitable for the application of this methodology. BOECs were used in the current study because of their proliferative potential; maintenance of phenotype across several (10+) passages; and robust VWF expression, facilitating the quantification of VWF exocytosis. Other investigators have used human umbilical vein endothelial cells (HUVECs)5,8,28 and non-endothelial cells (human embryonic kidney 293 cells) transfected with VWF cDNAs expressing VWD mutations35. It has been demonstrated that VWF expression varies significantly between tissue and vessel physiology36. Cultured endothelial cells arising from various micro- and macrovasculature have observable differences in VWF protein basal secretion37, and there is evidence that calcium channel differences between endothelial cell subtypes influence WPB exocytosis in response to secretagogues38.
Histones were used in this study as a novel secretagogue. They were compared to the positive control, histamine, a frequently described physiological stimulant of VWF release. A VWF:Ag ELISA was used to verify VWF release from endothelial cells under static conditions. There are two intracellular mechanisms that trigger the release of VWF: the elevation of intracellular calcium39 or pf cyclic AMP levels40. Differential activation of these pathways can be induced by secretagogues such as thrombin/histamine or epinephrine/vasopressin (or, relevant to VWD, desmopressin), respectively, and it may be interesting to compare these two pathways using this flow chamber model of VWF-platelet string formation. In the context of thrombo-inflammatory disease, further elucidation of inflammatory biomarkers, such as cytokines41 or damage/pathogen-associated molecular patterns (DAMPS/PAMPS), and their roles in VWF-platelet string formation is encouraged.
Fluid shear stress is important when globular VWF unfolds to expose the A1 domain binding site for glycoprotein Ib on platelets11. It has been demonstrated that a minimum fluid shear stress of approximately 0.73 dyn/cm2 is required for platelet binding to VWF strings42. Furthermore, these strings form on endothelial cells at relatively low shear stresses (2.5 dyn/cm2) but can be dislodged at shear stresses of 20 dyn/cm2 or above8. In the current investigation, we selected a venous shear stress (4.45 dyn/cm2) that allowed for interactions between the endothelium, VWF, and platelets while conserving the perfusion buffer for a 10 min experiment.
Using this flow chamber system to evaluate ADAMTS13 cleavage of VWF-platelet strings requires an increase in the fluidic shear stresses over those used in this experiment. Maximal ADAMTS13 cleavage of VWF-platelet strings occurs between 10 and 30 dyn/cm2 12, a range that encompasses high venous to moderate arterial shear stress43. Modifying the shear rate in this system may be useful for comparing VWF-platelet string formation and dissolution, contributing factors in the development of venous, arterial, or microvascular thrombosis.
Additional modifications of this technique can be made to the perfusion material and the flow chamber. Secretagogues can be administered under flow conditions to resemble in vivo environments. Priming platelets with agonists to evaluate their reactivity or including leukocyte populations are extensions of this technique. Furthermore, including therapeutic agents targeting VWF multimer length, platelet binding capacity, or platelet reactivity may elucidate important mechanisms that regulate the initiation of thrombus development. Other investigators have used either in-house44 or purchased glass coverslip flow chambers45, as opposed to the polymer-based slides used in this study, and seem to obtain comparable results.
A critical consideration in the VWF-platelet string protocol is the preparation of washed platelets. It is not recommended to use whole blood in this protocol due to the presence of ADAMTS13. Agents preventing platelet activation, such as prostacyclin or apyrase, should be avoided, as this may later adversely influence platelet adhesion to VWF. Ensuring that the buffers are brought to room temperature prior to platelet isolation and using careful pipette techniques (i.e., avoiding bubbles and harsh shear stresses) are key to maintaining the platelets in their inactive state. Furthermore, labeling platelets with a mitochondrial stain, such as DIOC6 or rhodamine 6G, is preferred over interfering with extracellular domains that may be important for VWF binding. However, as DIOC6 is a membrane-permeable dye, it may also be taken up in small quantities by the endothelial cells over the course of the experiment. In the absence of a fluorescent microscope, other studies have observed VWF-platelet strings using bright-field microscopy46.
In conclusion, this flow chamber methodology of VWF-platelet string formation can provide mechanistic, real-time data to support in vivo observations in the pathogenesis of thrombosis. As exposure to shear strongly influences the pathophysiological role that VWF plays in thrombo-inflammatory disease, the flow chamber model is a vital investigative tool for any lab interested in understanding endothelial cell-platelet interactions in this context.
The authors have nothing to disclose.
Alison Michels is a recipient of a Frederick Banting and Charles Best Canada Graduate Scholarship from the Canadian Institutes of Health Research (CIHR). Laura L. Swystun is the recipient a CIHR fellowship. David Lillicrap is the recipient of a Canada Research Chair in Molecular Hemostasis. This study was funded in part by a CIHR operating grant (MOP-97849).
Calf-thymus unfractionated histones (UH) | Worthington Biochemical | HLY | Reconstituted in serum-reduced media (5 mg/mL) |
Calf-thymus lysine-rich histones (HK) | Sigma-Aldrich | H5505 | Reconstituted in serum-reduced media (5 mg/mL) |
Calf-thymus arginine-rich histones (HR) | Sigma-Aldrich | H4830 | Reconstituted in serum-reduced media (5 mg/mL) |
Phorbol 12-myristate 13-acetate (PMA) | Sigma-Aldrich | P8139 | Reconstituted in DMSO (20 mM) |
Histamine | Sigma-Aldrich | H7125-1G | Reconstituted in water (50 mg/mL) |
3,3' Dihexyloxacarbocyanine Iodide (DiOC6) | Invitrogen | D273 | Reconstituted in methanol (20 mM) |
Rabbit Anti-VWF Coating Antibody | DAKO | A0082 | For VWF ELISA |
Rabbit Anti-VWF Detection Antibody, HRP conjugated | DAKO | P0026 | For VWF ELISA and histone-VWF binding assay |
Nunc MaxiSorp flat-bottom 96-well microplates | eBioscience | 44-2404-21 | For histone-VWF binding assay |
Immulon 4 HBX Flat Bottom Microtiter 96-Well Plates | Thermo Scientific | 3855 | For VWF ELISA |
Humate-P | CSL Behring | N/A | Plasma-derived human von Willebrand factor/factor VIII complex |
Normal Reference Plasma | Precision BioLogic | CCNRP-05 | For VWF ELISA standard curve |
O-Phenylenediamine dihydrochloride (OPD) reagent | Sigma-Aldrich | P8287 | Equivalent product available through ThermoFisher Scientific (Catalogue Number: 34006) |
EGM-2 BulletKit | Lonza | CC-3162 | For culturing and initial seeding of BOEC |
Hank's Balanced Salt Solution (HBSS) | ThermoFisher Scientific | 14025092 | |
Rat-tail Collagen Type 1 | Corning | 354236 | |
Gibco Opti-MEM I Reduced Serum Media | ThermoFisher Scientific | 31985070 | For endothelial cell stimulations |
METAMORPH Microscopy Automation and Image Analysis Software | Molecular Devices | N/A | |
BD Vacutainer Blood Collection Tubes, No Additive | BD Biosciences | 366703 | |
µ-Slide III 0.1 (flow chambers) | Ibidi | This product has been discontinued. We suggest using µ-Slide VI 0.1 (#80661) or 0.4 (# 80601) and recalculating flow rate and platelet volume needed to maintain a shear stress of 4.45 dyn/cm2 | |
Silicone Tubing 1.6 mm ID: 5 m, sterilized | Ibidi | 10842 | |
Luer Lock Connector Female: natural Polypropylene, sterilized | Ibidi | 10825 | |
Elbow Luer Connector Male: white Polypropylene, sterilized | Ibidi | 10802 | |
Blunted 18G Needle | BD Biosciences | 305180 | |
20 mL syringes | BD Biosciences | 302830 | |
Syringe Pump | New Era Pump Systems Inc. NE-1600 Multi-PhaserTM | N/A | |
Quorum WaveFX- 4X1 spinning disk microscope | Quorum Technologies | N/A | |
Image Processing Software | ImageJ | N/A |