Covalently-Bound Heparin Makes Collagen Thromboresistant
http://www.100md.com
《动脉硬化血栓血管生物学》
From the Department of Biochemistry (J.F.W.K., T.L.) and the Cardiovascular Research Institute (S.J.H.W., T.L.), Maastricht University, 6200 MD Maastricht, The Netherlands; and the Medtronic Bakken Research Center B.V. (A.D., M.V., M.H.), 6229 GW Maastricht, The Netherlands.
Correspondence to Dr. T. Lindhout, Department of Biochemistry, Maastricht University, PO Box 616, 6200 MD Maastricht, The Netherlands. E-mail t.lindhout@bioch.unimaas.nl
Abstract
Objective— Blood compatibility of artificial surfaces depends on their immunogenic and thrombogenic properties. Collagen’s weak antigenicity makes it an attractive candidate for stent coatings or fabrication of vascular grafts. However, the thrombogenic nature of collagen limits its application. We examined whether heparinization can make collagen more thromboresistant.
Methods and Results— Collagen was heparinized by crosslinking collagen with extensively periodate oxidized heparin and/or by covalently bonding of mildly periodate oxidized heparin. Both ways of heparinization have no effect on platelet adhesion and could not abolish induction of platelet procoagulant activity. However, thrombin generation was completely prevented under static and flow conditions. The functionality of immobilized heparin was confirmed by specific uptake of antithrombin, 13.5±4.7 pmol/cm2 and 1.95±0.21 pmol/cm2 for mildly and heavily periodated heparin, respectively.
Conclusions— These results indicate that immobilization of heparin on collagen, even as a crosslinker, is a very effective way to prevent surface thrombus formation. These data encourage the application of heparinized collagen as stent-graft material in animal and eventually human studies.
Key Words: collagen ? heparin ? thrombogenicity ? thrombosis ? blood flow
Introduction
Cardiovascular disease, including vascular stenosis, is still the leading cause of death in Western society. Obstructive atherosclerotic disease, causing angina pectoris or even myocardial infarction, is currently treated by the implantation of a stent or through bypass surgery. Unfortunately, attempts to implant vascular grafts with a small diameter are not successful because of thrombotic and inflammatory reactions.1–6 Thus, to enhance blood compatibility, a surface has to be designed that is both anti-immunogenic and thromboresistant. Collagen has been widely used in medical applications, including skin replacement, bone substitutes, and artificial valves.7,8 Recently, a number of investigators attempted to use modified collagen as a vascular graft material.9–12 Potential advantages of the natural biological polymer collagen are its weak antigenicity and high tensile strength, which can resist high arterial blood pressures. Furthermore, collagen is a suitable substrate for endothelial cell growth in vitro,11 which makes its application as an artificial vessel even more attractive. Nevertheless, the prothrombotic properties of collagen13 are a major drawback in its applicability in blood contacting devices.
To make collagen more thromboresistant, Wissink et al14 coupled the sulfated polysaccharide, heparin, to collagen. An in vitro assay showed that the immobilized heparin was functionally active as it accelerated the thrombin-antithrombin (AT) reaction. In addition, this group demonstrated that crosslinking of collagen had an adverse effect on the antigenicity and degradation rate of collagen, but stimulated endothelial cell adhesion and proliferation.11,12 The suitability of collagen as a vascular graft material was also emphasized in a rabbit study, which established that collagen-grafts were completely endothelialized after one month of implantation.10
This study was undertaken to get a better understanding of the precise antithrombotic functions of immobilized heparin. Three types of heparinized collagen (heparin-crosslinked collagen, heparinized heparin-crosslinked collagen, and heparinized EDC/NHS-crosslinked collagen) were evaluated for their thrombogenicity by measuring platelet adhesion, platelet activation, and thrombin generation.
Methods
Preparation of Collagen and Heparinized Collagen Sheets
Four different types of type I collagen (Syntacoll) sheets were prepared; namely, noncrosslinked collagen (NC), heparin-crosslinked collagen (HC), heparinized EDC/NHS-crosslinked collagen (HEC), and heparinized heparin crosslinked collagen (HHC). For a detailed description, see http://atvb.ahajournals.org.
Heparin Assay
In all (static) experiments collagen and heparinized collagen sheets, mounted in the wells of a 48-well microtiter plate (Corning), were repeatedly rinsed in Hepes buffer (5 mmol/L Hepes, 136 mmol/L NaCl, 2.7 mmol/L KCl, 2 mmol/L MgCl2, 1 mg/mL BSA, pH 7.45) to remove unbound heparin. The amount of heparin in the rinse-buffer was measured by adding a 30 μL aliquot into 112.5 μL AT (0.66 μmol/L). After an incubation of 3 minutes at 37°C, 7.5 μL bovine factor Xa (1 μmol/L) was added. At timed intervals, 25-μL samples were transferred into a cuvette containing 425 μL Hepes buffer containing 20 mmol/L EDTA and 50 μL chromogenic substrate S2765 (2 mmol/L, Chromogenix). The amount of heparin was calculated from the rate of disappearance of the factor Xa activity. A reference curve was constructed from heparin (190 U/mg; Diosynth) that was used to prepare heparinized collagen.
Platelet Adhesion and Procoagulant Activity
Venous blood was obtained from volunteers who denied taking any medication in the 2 weeks before sampling. Washed platelets were prepared as described.15 Wells were blocked with 20 mg/mL BSA in Hepes buffer. A 1-mL suspension of washed platelets (3x108 platelets) was then added to collagen and heparinized collagen films in a 48-well microtiter plate, and nonbound platelets were removed after an incubation of 45 minutes at room temperature under stationary conditions. For scanning electron microscopy (SEM) analysis, collagen sheets were treated as described.16 SEM images were analyzed for platelet surface coverage using ImagePro software (Media Cybernetics). Collagen adherent platelets were incubated with 3 mmol/L CaCl2 in Hepes buffer for 15 minutes. This was followed by incubation with 1 μg/mL Oregon Green-labeled annexin V (NeXins) and 3 mmol/L CaCl2 in Hepes buffer. Nonbound annexin V was removed with Hepes buffer containing 3 mmol/L CaCl2, and platelet bound annexin V was eluted in Hepes buffer containing 20 mmol/L EDTA. The amount of recovered annexin V was measured in a spectrofluorometer (SLM) with excitation wavelength 485 nm and emission wavelength 535 nm. A reference curve was constructed with known amounts of Oregon Green-labeled annexin V.
Determination of Specific Antithrombin Binding to Heparinized Collagen
Specific binding of AT to heparinized collagen was determined as described.17 See http://atvb.ahajournals.org for further information.
Thrombin Generation
Collagen films with adherent platelets were exposed to 300 μL of citrated PFP or AT depleted plasma (Biopool) containing 250 μmol/L of the fluorogenic thrombin substrate Z-GGR-AMC (Bachem). Coagulation was started by the addition of 20 mmol/L of CaCl2. Fluorescence tracings as a result of thrombin generation were recorded with a temperature controlled microplate fluorometer (SPECTRAmax GEMINI XS, Molecular Devices) at 37°C with ex=368 nm and em=460 nm. Floating point averaged, first derivative traces were constructed to obtain thrombin generation, as described by Hemker et al.18
Thrombin generation at collagen surfaces in flowing whole blood was measured as follows: Small pieces of the collagen sheets were cut into pieces of 15 mm length and 10 mm width. The sheets were prewetted in deionized water and folded around a stent of 15 mm length and 3.0 mm diameter (beStent BravaTM, Medtronic) with a few mm overlap. After drying at room temperature, 1.5-mm silicone rings were fixed at both ends. Next, the stents were expanded in silicone tubing with inner diameter of 1.5 mm and rinsed overnight with Hepes buffer at a flow rate of 500 μL/min to remove nonbound heparin. Flowing citrated whole blood (27 μL/min) was mixed with flowing (3 μL/min) Hepes buffer containing 200 mmol/L CaCl2, and perfused through the stent-containing silicone tube as described.19 Timed samples (30 μL) were taken at the outlet and mixed with 420 μL Hepes buffer containing 20 mmol/L EDTA and 200 μmol/L S2238 (Chromogenix). The diluted blood sample was stored on ice for the duration of the flow experiment and subsequently centrifuged at 10,000 rpm for 20 seconds. The supernatant was assayed for thrombin activity as described.20
Statistical Analysis
The data were expressed as mean±SD. To determine the statistical significance of differences, probability values were obtained with a nonparametric test for two (Mann–Whitney Test) or more (Kruskal–Wallis Test) independent variables.
Results
Effect of Heparinization on Platelet Adhesion
Prior to experimentation, collagen films were extensively rinsed to remove any nonbound heparin. Heparin in the effluent was assayed as described. The collagen films were then incubated with washed platelets. Platelet adhesion was assessed from SEM micrographs (Figure 1A, middle). It is clearly seen that all adherent platelets formed pseudopods (Figure 1A, bottom). No significant differences in platelet adhesion to NC, HC, HEC, and HHC were observed, indicating that heparinization of collagen does not affect platelet adhesion (Figure 1B).
Figure 1. Adhesion of platelets to (heparinized) collagen films. A, SEM images of noncrosslinked collagen (NC), collagen crosslinked with heparin (HC), collagen crosslinked with carbodiimide and additional heparin treatment (HEC), and collagen crosslinked with heparin and additional heparin treatment (HHC) before (top) and after platelet adhesion (middle and bottom). B, platelet surface coverage as a percentage of total area. Shown are mean±SD of at least 10 different fields.
Effect of Heparinization on Collagen-Induced Platelet Procoagulant Activity
Collagen-adherent platelets were examined for their loss of plasma membrane phospholipid asymmetry, utilizing the property of annexin V to bind only to cell membranes containing anionic phospholipids such as phosphatidylserine.21 The collagen surfaces, except HHC, did not bind annexin V. No significant difference could be observed in the amount of annexin V bound to platelets on NC and HC (P=0.42, Mann–Whitney test), indicating that multi-point coupled heparin, when used as a crosslinker of collagen, does not prevent the collagen-induced activation of adherent platelets (Table 1). However, annexin V binding was significantly reduced for HEC and HHC surfaces with adherent platelets (P0.01, Kruskal–Wallis test). It is apparent that immobilization of heparin, using mildly periodate oxidized heparin, partially inhibits collagen-induced exposure of anionic phospholipids in the outer leaflet of the platelet plasma membrane.
Annexin V Binding to Collagen-Adherent Platelets
Antithrombin Binding
The AT binding capacities of the different collagen surfaces are shown in Figure 2. The NC surface bound hardly any AT (0.09±0.08 pmol/cm2). When the collagen surfaces contained heparin, either as crosslinker (HC) or immobilized on crosslinked collagen (HHC or HEC), significant amounts of AT bound to these surfaces. HC bound 1.95±0.21 pmol/cm2 AT, which is about 25-fold more than NC. Interestingly, heparinization of HC greatly increased the AT binding capacity (23.8±4.1 pmol/cm2). It is very likely that the mild periodate oxidation procedure used to immobilize heparin to the HC preserves the binding sites for AT to a larger extent than the modification of the crosslinker heparin. This notion is supported by the finding that HEC also binds relatively high amounts of AT (13.5±4.7 pmol/cm2).
Figure 2. Binding of antithrombin to immobilized heparin. NC, HC, HEC, and HHC were incubated with AT (5 μmol/L) for one hour. Unbound AT was removed and bound AT was eluted with heparin and quantified (see http://atvb.ahajournals.org for more details). Data are mean±SD of three independent experiments.
Thrombin Generation at Collagen Surfaces under Static Conditions
Collagen surfaces with adherent platelets were exposed to recalcified citrated platelet-free plasma, and thrombin generation was continuously monitored as described. It was first established that under the conditions of the assay, collagen does not generate thrombin in the absence of adherent platelets. Thus, under the conditions of the experiment, thrombin generation requires the presence of (activated) platelets (data not shown). When NC surfaces with adherent platelets were exposed to PFP, thrombin activity became detectable after a lag phase of about 12 minutes, and a peak level of thrombin was reached after about 20 minutes (Figure 3, left). In contrast, when the same experiment was performed with HC, HEC, or HHC, thrombin generation was not detected. To examine the contribution of heparin cofactor II, heparinized collagen surfaces with adherent platelets were exposed to AT-depleted plasma. In contrast with normal PFP, significant amounts of thrombin were generated on heparinized collagen films. Maximal levels of thrombin were reached after approximately 20 minutes for HC and 25 minutes for HEC and HHC surfaces (Figure 3, right). The sustained levels of thrombin activity in the case of AT-depleted plasma indicate that heparin cofactor II hardly contributes to the inactivation of thrombin in plasma. Consequently, heparinized-collagen surfaces do not inhibit thrombin generation in AT-depleted plasma.
Figure 3. Thrombin generation under static conditions. Collagen and heparinized collagen films with adherent platelets were exposed to recalcified platelet-free plasma (A1-D1) or AT deficient plasma (A2-D2) containing the fluorogenic thrombin substrate Z-GGR-AMC. Thrombin formation on NC (A), HC (B), HEC (C), and HHC (D) are shown as a function of time.
Thrombin Generation at Collagen Surfaces under Flow Conditions
Stents were wrapped in a collagen sheet, inserted in silicone tubes (inner diameter 1.5 mm), expanded, and perfused with recalcified whole blood. Figure 4 shows the thrombin concentration measured at the outlet of the flow system as a function of the perfusion time. In the control, an empty silicone sleeve with (Figure 4, ) and without (Figure 4, ) an inserted stent, traces of thrombin appeared 16 and 18 minutes after the start of the perfusion, respectively. Noncrosslinked collagen dramatically shortened the lag phase in thrombin generation and increased the maximal thrombin concentration (Figure 4, ?). The blood perfusion had to be stopped after 40 minutes because of massive clot formation in the silicone tube, resulting in an irregular flow. Neither thrombin generation, nor blood clot formation, were detectable in silicon tubes with stents that were wrapped in collagen that had been cross-linked with heparin or heparinized after cross-linking with heparin or EDC/NHS. These findings suggest that also with flowing whole blood, multi-point attached heparin as a cross-linker by virtue of its AT binding capacity can prevent the initiation and propagation of collagen-induced thrombin generation at the surface of adherent platelets. These findings suggest that multi-point attached heparin, as a cross-linker by virtue of its AT binding capacity, can also prevent the initiation and propagation of collagen-induced thrombin generation at the surface of adherent platelets with flowing whole blood.
Figure 4. Thrombin generation in flowing blood in contact with collagen coated stents. Empty silicone sleeve (), bare stent (), and stent wrapped in a sheet of NC (?), HC (), HEC (), or HHC () were perfused with recalcified whole blood at a flow rate of 30 μL/min. Depicted is the thrombin concentration at the outlet of the flow reactor as a function of perfusion time. Arrows indicate time of extensive clot formation and occlusion of the flow reactor tubing.
Discussion
Different heparinized collagen preparations, namely collagen crosslinked with extensively periodated heparin (HC), HC with immobilized mildly periodated heparin (HHC), and collagen crosslinked with EDC/NHS and treated with mildly periodated heparin-crosslink (HEC), were evaluated for their thrombogenic properties. We wished to reveal to what extent heparin as a collagen crosslinker or covalently bound to the collagen surface is able to reduce collagen’s thrombogenicity. Our results demonstrate that numerous platelets adhered to collagen films, and that these platelets supported the process of thrombin generation at the collagen surface. Heparinization of collagen could not abolish the binding and collagen-induced activation of platelets. However, in contrast to nontreated collagen surfaces, no thrombin generation could be detected on heparinized surfaces.
Platelet adhesion, determined as the percentage of surface coverage, was the same for NC, HC, HEC, and HHC. This result clearly indicates that heparinization of collagen (either as a crosslinker or immobilized on crosslinked collagen) does not prevent, or even reduce, platelet adhesion. Moreover, platelets that adhere to heparinized collagen become activated, resulting in pseudopod formation and the appearance of phosphatidylserine in the outer leaflet of the plasma membrane. This result appears to contrast with findings of other investigators, who demonstrated that heparin coatings significantly decreased platelet adhesion.22,23 However, our observation that platelet adhesion on collagen is not inhibited by the presence of immobilized heparin is in line with a previously reported study that demonstrated that heparin immobilization on carbodiimide crosslinked collagen even slightly increased the number of adherent platelets.14 Contradictory findings in the effects of immobilized heparin on platelet adhesion might be related to the type of surface to which heparin is bound.
Collagen-induced platelet procoagulant activity is mediated by the interaction of collagen with its platelet receptor, GPVI.24 It is apparent that crosslinking collagen with heparin did not inhibit the loss in phospholipid membrane asymmetry, as monitored by annexin-V binding. In contrast, collagen-induced platelet procoagulant activity was significantly reduced when collagen films were treated with mildly periodate oxidized heparin, implying that under these circumstances, when the polysaccharide chain of heparin is almost fully intact, immobilized heparin interferes with platelet activation by collagen. Whether immobilized heparin affects the interaction of collagen with its primary platelet receptor GPVI or with a secondary one (P2Y2 and/or P2Y12) awaits further experimentation. Because these experiments were performed with adherent platelets that were extensively washed, heparin-accelerated inactivation of traces of thrombin is highly unlikely. Interference of heparin with platelet activation by collagen was previously demonstrated under both static and flow conditions.25–27 One study showed that under static conditions or at low shear rate, immobilized heparin did not affect platelet deposition to collagen, but strongly inhibited the platelet release reaction.26 Another study in rats revealed that under flow conditions (low and high shear rate), heparin released by stimulated mast cells did not block 2?1-dependent platelet adhesion, but attenuated subsequent platelet activation as well as fibrinogen binding to platelets.27 Together, these results support the notion that mildly periodate oxidized heparin may interfere with collagen-induced platelet activation (ie, release reaction and exposure of negatively charged phospholipids in outer leaflet of the plasma membrane).
Exposure of heparinized collagen films with adherent (activated) platelets to recalcified platelet-free plasma demonstrated that, independent of the method of heparinization, thrombin formation is abolished completely during the time of the experiment (60 minutes). We surmise that the amount of AT (±2 pmol/cm2) that binds to collagen films crosslinked with highly periodated heparin is already sufficient to prevent thrombin generation at these surfaces. These results extend earlier findings from our group illustrating that AT bound to immobilized heparin is probably the most effective way to prevent thrombus formation at an artificial surface.19,28,29 To confirm that the thromboresistance of the heparinized collagen surfaces was indeed AT-dependent, thrombin generation was measured with AT-depleted plasma. While thrombin generation on heparinized collagen with adhering platelets was strongly inhibited when exposed to normal plasma, significant thrombin generation was observed when the same experiment was performed with AT-depleted plasma, suggesting that the thromboresistance of heparinized collagen surfaces is fully dependent on the presence of AT. The physiological relevance of our findings was established in a perfusion setup, where stents coated with heparinized collagen films were exposed to flowing whole blood. While an explosive thrombin formation was observed in blood that passes the nonheparinized collagen-coated stents, stents coated with heparinized collagen appeared to be fully thromboresistant. We postulate that immobilization of heparin effectively catalyzes the thrombin–antithrombin reaction, whereby levels of surface-located thrombin remain below the threshold values that are needed to exert its positive feedback reactions, such as activation of the cofactors factors V and VIII and generation of procoagulant cell membranes, that regulate thrombin formation. The in vitro finding that heparinized collagen films are highly thromboresistant encourages further research on the application of collagen in medical devices in contact with blood.
References
Link J, Feyerabend B, Grabener M, Linstedt U, Brossmann J, Thomsen H, Heller M. Dacron-covered stent-grafts for the percutaneous treatment of carotid aneurysms: effectiveness and biocompatibility–experimental study in swine. Radiology. 1996; 200: 397–401.
Holmes DR, Camrud AR, Jorgenson MA, Edwards WD, Schwartz RS. Polymeric stenting in the porcine coronary artery model: differential outcome of exogenous fibrin sleeves versus polyurethane-coated stents. J Am Coll Cardiol. 1994; 24: 525–531.
Ratcliffe A. Tissue engineering of vascular grafts. Matrix Biol. 2000; 19: 353–357.
Nelson SR, deSouza NM, Allison DJ. Endovascular stents and stent-grafts: is heparin coating desirable? Cardiovasc Intervent Radiol. 2000; 23: 252–255.
Welt FGP, Rogers C. Inflammation and restenosis in the stent era. Arterioscler Thromb Vasc Biol. 2002; 22: 1769–1776.
Ross R. Atherosclerosis is an inflammatory disease. Am Heart J. 1999; 138: S419–S420.
Lee CH, Singla A, Lee Y. Biomedical applications of collagen. Int J Pharm. 2001; 221: 1–22.
Culav EM, Clark CH, Merrilees MJ. Connective tissues: matrix composition and its relevance to physical therapy. Phys Ther. 1999; 79: 308–319.
Lambert AW, Fox AD, Williams DJ, Horrocks M, Budd JS. Experience with heparin-bonded collagen-coated grafts for infrainguinal bypass. Cardiovasc Surg. 1999; 7: 491–494.
Cloft HJ, Kallmes DF, Lin HB, Li ST, Marx WF, Hudson SB, Helm GA, Lopes MB, McGraw JK, Dion JE, Jensen ME. Bovine type I collagen as an endovascular stent-graft material: Biocompatibility study in rabbits. Radiology. 2000; 214: 557–562.
Wissink MJB, Beernink R, Scharenborg NM, Poot AA, Engbers GHM, Beugeling T, van Aken WG, Feijen J. Endothelial cell seeding of (heparinized) collagen matrices: effects of bFGF pre-loading on proliferation (after low density seeding) and pro-coagulant factors. J Control Release. 2000; 67: 141–155.
van Wachem PB, Plantinga JA, Wissink MJB, Beernink R, Poot AA, Engbers GHM, Beugeling T, van Aken WG, Feijen J, van Luyn MJA. In vivo biocompatibility of carbodiimide-crosslinked collagen matrices: effects of crosslink density, heparin immobilization, and bFGF loading. J Biomed Mater Res. 2001; 55: 368–378.
Heemskerk JWM, Siljander PRM, Bevers EM, Farndale RW, Lindhout T. Receptors and signalling mechanisms in the procoagulant response of platelets. Platelets. 2000; 11: 301–306.
Wissink MJ, Beernink R, Pieper JS, Poot AA, Engbers GH, Beugeling T, van Aken WG, Feijen J. Immobilization of heparin to EDC/NHS-crosslinked collagen. Characterization and in vitro evaluation. Biomaterials. 2001; 22: 151–163.
Heemskerk JW, Vuist WM, Feijge MA, Reutelingsperger CP, Lindhout T. Collagen but not fibrinogen surfaces induce bleb formation, exposure of phosphatidylserine, and procoagulant activity of adherent platelets: evidence for regulation by protein tyrosine kinase-dependent Ca2+ responses. Blood. 1997; 90: 2615–2625.
Briedé J, Wielders S, Heemskerk J, Baruch D, Hemker H, Lindhout T. Von Willebrand factor stimulates thrombin-induced exposure of procoagulant phospholipids on the surface of fibrin-adherent platelets. J Thromb Haemost. 2003; 1: 559–565.
Blezer R, Cahalan L, Cahalan PT, Lindhout T. Heparin coating of tantalum coronary stents reduces surface thrombin generation but not factor IXa generation. Blood Coagul Fibrinolysis. 1998; 9: 435–440.
Hemker HC, Giesen PL, Ramjee M, Wagenvoord R, Beguin S. The thrombogram: monitoring thrombin generation in platelet-rich plasma. Thromb Haemost. 2000; 83: 589–591.
Keuren JFW, Wielders SJH, Willems GM, Morra M, Lindhout T. Fibrinogen adsorption, platelet adhesion and thrombin generation at heparinized surfaces exposed to flowing blood. Thromb Haemost. 2002; 87: 742–747.
Schoen P, Lindhout T, Willems G, Hemker HC. Continuous flow and the prothrombinase-catalyzed activation of prothrombin. Thromb Haemost. 1990; 64: 542–547.
Stuart MC, Bevers EM, Comfurius P, Zwaal RF, Reutelingsperger CP, Frederik PM. Ultrastructural detection of surface exposed phosphatidylserine on activated blood platelets. Thromb Haemost. 1995; 74: 1145–1151.
Niimi Y, Ishiguro Y, Nakata Y, Goto T, Morita S, Yamane S. Platelet adhesion to heparin coated oxygenator fibers under in vitro static conditions: impact of temperature. ASAIO J. 2001; 47: 361–364.
Bickel C, Rupprecht HJ, Darius H, Binz C, Hauroder B, Krummenauer F, Meyer J. Substantial reduction of platelet adhesion by heparin-coated stents. J Interv Cardiol. 2001; 14: 407–413.
Nieswandt B, Watson SP. Platelet-collagen interaction: is GPVI the central receptor? Blood. 2003; 102: 449–461.
Griffin MJ, Rinder HM, Smith BR, Tracey JB, Kriz NS, Li CK, Rinder CS. The effects of heparin, protamine, and heparin/protamine reversal on platelet function under conditions of arterial shear stress. Anesth Analg. 2001; 93: 20–27.
Kauhanen P, Kovanen PT, Lassila R. Coimmobilized native macromolecular heparin proteoglycans strongly inhibit platelet-collagen interactions in flowing blood. Arterioscler Thromb Vasc Biol. 2000; 20: E113–E119.
Lassila R, Lindstedt K, Kovanen PT. Native macromolecular heparin proteoglycans exocytosed from stimulated rat serosal mast cells strongly inhibit platelet-collagen interactions. Arterioscler Thromb Vasc Biol. 1997; 17: 3578–3587.
Keuren JFW, Wielders SJH, Willems GM, Morra M, Cahalan L, Cahalan P, Lindhout T. Thrombogenicity of polysaccharide-coated surfaces. Biomaterials. 2003; 24: 1917–1924.
Blezer R, Fouache B, Willems GM, Lindhout T. Activation of blood coagulation at heparin-coated surfaces. J Biomed Mater Res. 1997; 37: 108–113.(Jeffrey F.W. Keuren; Simo)
Correspondence to Dr. T. Lindhout, Department of Biochemistry, Maastricht University, PO Box 616, 6200 MD Maastricht, The Netherlands. E-mail t.lindhout@bioch.unimaas.nl
Abstract
Objective— Blood compatibility of artificial surfaces depends on their immunogenic and thrombogenic properties. Collagen’s weak antigenicity makes it an attractive candidate for stent coatings or fabrication of vascular grafts. However, the thrombogenic nature of collagen limits its application. We examined whether heparinization can make collagen more thromboresistant.
Methods and Results— Collagen was heparinized by crosslinking collagen with extensively periodate oxidized heparin and/or by covalently bonding of mildly periodate oxidized heparin. Both ways of heparinization have no effect on platelet adhesion and could not abolish induction of platelet procoagulant activity. However, thrombin generation was completely prevented under static and flow conditions. The functionality of immobilized heparin was confirmed by specific uptake of antithrombin, 13.5±4.7 pmol/cm2 and 1.95±0.21 pmol/cm2 for mildly and heavily periodated heparin, respectively.
Conclusions— These results indicate that immobilization of heparin on collagen, even as a crosslinker, is a very effective way to prevent surface thrombus formation. These data encourage the application of heparinized collagen as stent-graft material in animal and eventually human studies.
Key Words: collagen ? heparin ? thrombogenicity ? thrombosis ? blood flow
Introduction
Cardiovascular disease, including vascular stenosis, is still the leading cause of death in Western society. Obstructive atherosclerotic disease, causing angina pectoris or even myocardial infarction, is currently treated by the implantation of a stent or through bypass surgery. Unfortunately, attempts to implant vascular grafts with a small diameter are not successful because of thrombotic and inflammatory reactions.1–6 Thus, to enhance blood compatibility, a surface has to be designed that is both anti-immunogenic and thromboresistant. Collagen has been widely used in medical applications, including skin replacement, bone substitutes, and artificial valves.7,8 Recently, a number of investigators attempted to use modified collagen as a vascular graft material.9–12 Potential advantages of the natural biological polymer collagen are its weak antigenicity and high tensile strength, which can resist high arterial blood pressures. Furthermore, collagen is a suitable substrate for endothelial cell growth in vitro,11 which makes its application as an artificial vessel even more attractive. Nevertheless, the prothrombotic properties of collagen13 are a major drawback in its applicability in blood contacting devices.
To make collagen more thromboresistant, Wissink et al14 coupled the sulfated polysaccharide, heparin, to collagen. An in vitro assay showed that the immobilized heparin was functionally active as it accelerated the thrombin-antithrombin (AT) reaction. In addition, this group demonstrated that crosslinking of collagen had an adverse effect on the antigenicity and degradation rate of collagen, but stimulated endothelial cell adhesion and proliferation.11,12 The suitability of collagen as a vascular graft material was also emphasized in a rabbit study, which established that collagen-grafts were completely endothelialized after one month of implantation.10
This study was undertaken to get a better understanding of the precise antithrombotic functions of immobilized heparin. Three types of heparinized collagen (heparin-crosslinked collagen, heparinized heparin-crosslinked collagen, and heparinized EDC/NHS-crosslinked collagen) were evaluated for their thrombogenicity by measuring platelet adhesion, platelet activation, and thrombin generation.
Methods
Preparation of Collagen and Heparinized Collagen Sheets
Four different types of type I collagen (Syntacoll) sheets were prepared; namely, noncrosslinked collagen (NC), heparin-crosslinked collagen (HC), heparinized EDC/NHS-crosslinked collagen (HEC), and heparinized heparin crosslinked collagen (HHC). For a detailed description, see http://atvb.ahajournals.org.
Heparin Assay
In all (static) experiments collagen and heparinized collagen sheets, mounted in the wells of a 48-well microtiter plate (Corning), were repeatedly rinsed in Hepes buffer (5 mmol/L Hepes, 136 mmol/L NaCl, 2.7 mmol/L KCl, 2 mmol/L MgCl2, 1 mg/mL BSA, pH 7.45) to remove unbound heparin. The amount of heparin in the rinse-buffer was measured by adding a 30 μL aliquot into 112.5 μL AT (0.66 μmol/L). After an incubation of 3 minutes at 37°C, 7.5 μL bovine factor Xa (1 μmol/L) was added. At timed intervals, 25-μL samples were transferred into a cuvette containing 425 μL Hepes buffer containing 20 mmol/L EDTA and 50 μL chromogenic substrate S2765 (2 mmol/L, Chromogenix). The amount of heparin was calculated from the rate of disappearance of the factor Xa activity. A reference curve was constructed from heparin (190 U/mg; Diosynth) that was used to prepare heparinized collagen.
Platelet Adhesion and Procoagulant Activity
Venous blood was obtained from volunteers who denied taking any medication in the 2 weeks before sampling. Washed platelets were prepared as described.15 Wells were blocked with 20 mg/mL BSA in Hepes buffer. A 1-mL suspension of washed platelets (3x108 platelets) was then added to collagen and heparinized collagen films in a 48-well microtiter plate, and nonbound platelets were removed after an incubation of 45 minutes at room temperature under stationary conditions. For scanning electron microscopy (SEM) analysis, collagen sheets were treated as described.16 SEM images were analyzed for platelet surface coverage using ImagePro software (Media Cybernetics). Collagen adherent platelets were incubated with 3 mmol/L CaCl2 in Hepes buffer for 15 minutes. This was followed by incubation with 1 μg/mL Oregon Green-labeled annexin V (NeXins) and 3 mmol/L CaCl2 in Hepes buffer. Nonbound annexin V was removed with Hepes buffer containing 3 mmol/L CaCl2, and platelet bound annexin V was eluted in Hepes buffer containing 20 mmol/L EDTA. The amount of recovered annexin V was measured in a spectrofluorometer (SLM) with excitation wavelength 485 nm and emission wavelength 535 nm. A reference curve was constructed with known amounts of Oregon Green-labeled annexin V.
Determination of Specific Antithrombin Binding to Heparinized Collagen
Specific binding of AT to heparinized collagen was determined as described.17 See http://atvb.ahajournals.org for further information.
Thrombin Generation
Collagen films with adherent platelets were exposed to 300 μL of citrated PFP or AT depleted plasma (Biopool) containing 250 μmol/L of the fluorogenic thrombin substrate Z-GGR-AMC (Bachem). Coagulation was started by the addition of 20 mmol/L of CaCl2. Fluorescence tracings as a result of thrombin generation were recorded with a temperature controlled microplate fluorometer (SPECTRAmax GEMINI XS, Molecular Devices) at 37°C with ex=368 nm and em=460 nm. Floating point averaged, first derivative traces were constructed to obtain thrombin generation, as described by Hemker et al.18
Thrombin generation at collagen surfaces in flowing whole blood was measured as follows: Small pieces of the collagen sheets were cut into pieces of 15 mm length and 10 mm width. The sheets were prewetted in deionized water and folded around a stent of 15 mm length and 3.0 mm diameter (beStent BravaTM, Medtronic) with a few mm overlap. After drying at room temperature, 1.5-mm silicone rings were fixed at both ends. Next, the stents were expanded in silicone tubing with inner diameter of 1.5 mm and rinsed overnight with Hepes buffer at a flow rate of 500 μL/min to remove nonbound heparin. Flowing citrated whole blood (27 μL/min) was mixed with flowing (3 μL/min) Hepes buffer containing 200 mmol/L CaCl2, and perfused through the stent-containing silicone tube as described.19 Timed samples (30 μL) were taken at the outlet and mixed with 420 μL Hepes buffer containing 20 mmol/L EDTA and 200 μmol/L S2238 (Chromogenix). The diluted blood sample was stored on ice for the duration of the flow experiment and subsequently centrifuged at 10,000 rpm for 20 seconds. The supernatant was assayed for thrombin activity as described.20
Statistical Analysis
The data were expressed as mean±SD. To determine the statistical significance of differences, probability values were obtained with a nonparametric test for two (Mann–Whitney Test) or more (Kruskal–Wallis Test) independent variables.
Results
Effect of Heparinization on Platelet Adhesion
Prior to experimentation, collagen films were extensively rinsed to remove any nonbound heparin. Heparin in the effluent was assayed as described. The collagen films were then incubated with washed platelets. Platelet adhesion was assessed from SEM micrographs (Figure 1A, middle). It is clearly seen that all adherent platelets formed pseudopods (Figure 1A, bottom). No significant differences in platelet adhesion to NC, HC, HEC, and HHC were observed, indicating that heparinization of collagen does not affect platelet adhesion (Figure 1B).
Figure 1. Adhesion of platelets to (heparinized) collagen films. A, SEM images of noncrosslinked collagen (NC), collagen crosslinked with heparin (HC), collagen crosslinked with carbodiimide and additional heparin treatment (HEC), and collagen crosslinked with heparin and additional heparin treatment (HHC) before (top) and after platelet adhesion (middle and bottom). B, platelet surface coverage as a percentage of total area. Shown are mean±SD of at least 10 different fields.
Effect of Heparinization on Collagen-Induced Platelet Procoagulant Activity
Collagen-adherent platelets were examined for their loss of plasma membrane phospholipid asymmetry, utilizing the property of annexin V to bind only to cell membranes containing anionic phospholipids such as phosphatidylserine.21 The collagen surfaces, except HHC, did not bind annexin V. No significant difference could be observed in the amount of annexin V bound to platelets on NC and HC (P=0.42, Mann–Whitney test), indicating that multi-point coupled heparin, when used as a crosslinker of collagen, does not prevent the collagen-induced activation of adherent platelets (Table 1). However, annexin V binding was significantly reduced for HEC and HHC surfaces with adherent platelets (P0.01, Kruskal–Wallis test). It is apparent that immobilization of heparin, using mildly periodate oxidized heparin, partially inhibits collagen-induced exposure of anionic phospholipids in the outer leaflet of the platelet plasma membrane.
Annexin V Binding to Collagen-Adherent Platelets
Antithrombin Binding
The AT binding capacities of the different collagen surfaces are shown in Figure 2. The NC surface bound hardly any AT (0.09±0.08 pmol/cm2). When the collagen surfaces contained heparin, either as crosslinker (HC) or immobilized on crosslinked collagen (HHC or HEC), significant amounts of AT bound to these surfaces. HC bound 1.95±0.21 pmol/cm2 AT, which is about 25-fold more than NC. Interestingly, heparinization of HC greatly increased the AT binding capacity (23.8±4.1 pmol/cm2). It is very likely that the mild periodate oxidation procedure used to immobilize heparin to the HC preserves the binding sites for AT to a larger extent than the modification of the crosslinker heparin. This notion is supported by the finding that HEC also binds relatively high amounts of AT (13.5±4.7 pmol/cm2).
Figure 2. Binding of antithrombin to immobilized heparin. NC, HC, HEC, and HHC were incubated with AT (5 μmol/L) for one hour. Unbound AT was removed and bound AT was eluted with heparin and quantified (see http://atvb.ahajournals.org for more details). Data are mean±SD of three independent experiments.
Thrombin Generation at Collagen Surfaces under Static Conditions
Collagen surfaces with adherent platelets were exposed to recalcified citrated platelet-free plasma, and thrombin generation was continuously monitored as described. It was first established that under the conditions of the assay, collagen does not generate thrombin in the absence of adherent platelets. Thus, under the conditions of the experiment, thrombin generation requires the presence of (activated) platelets (data not shown). When NC surfaces with adherent platelets were exposed to PFP, thrombin activity became detectable after a lag phase of about 12 minutes, and a peak level of thrombin was reached after about 20 minutes (Figure 3, left). In contrast, when the same experiment was performed with HC, HEC, or HHC, thrombin generation was not detected. To examine the contribution of heparin cofactor II, heparinized collagen surfaces with adherent platelets were exposed to AT-depleted plasma. In contrast with normal PFP, significant amounts of thrombin were generated on heparinized collagen films. Maximal levels of thrombin were reached after approximately 20 minutes for HC and 25 minutes for HEC and HHC surfaces (Figure 3, right). The sustained levels of thrombin activity in the case of AT-depleted plasma indicate that heparin cofactor II hardly contributes to the inactivation of thrombin in plasma. Consequently, heparinized-collagen surfaces do not inhibit thrombin generation in AT-depleted plasma.
Figure 3. Thrombin generation under static conditions. Collagen and heparinized collagen films with adherent platelets were exposed to recalcified platelet-free plasma (A1-D1) or AT deficient plasma (A2-D2) containing the fluorogenic thrombin substrate Z-GGR-AMC. Thrombin formation on NC (A), HC (B), HEC (C), and HHC (D) are shown as a function of time.
Thrombin Generation at Collagen Surfaces under Flow Conditions
Stents were wrapped in a collagen sheet, inserted in silicone tubes (inner diameter 1.5 mm), expanded, and perfused with recalcified whole blood. Figure 4 shows the thrombin concentration measured at the outlet of the flow system as a function of the perfusion time. In the control, an empty silicone sleeve with (Figure 4, ) and without (Figure 4, ) an inserted stent, traces of thrombin appeared 16 and 18 minutes after the start of the perfusion, respectively. Noncrosslinked collagen dramatically shortened the lag phase in thrombin generation and increased the maximal thrombin concentration (Figure 4, ?). The blood perfusion had to be stopped after 40 minutes because of massive clot formation in the silicone tube, resulting in an irregular flow. Neither thrombin generation, nor blood clot formation, were detectable in silicon tubes with stents that were wrapped in collagen that had been cross-linked with heparin or heparinized after cross-linking with heparin or EDC/NHS. These findings suggest that also with flowing whole blood, multi-point attached heparin as a cross-linker by virtue of its AT binding capacity can prevent the initiation and propagation of collagen-induced thrombin generation at the surface of adherent platelets. These findings suggest that multi-point attached heparin, as a cross-linker by virtue of its AT binding capacity, can also prevent the initiation and propagation of collagen-induced thrombin generation at the surface of adherent platelets with flowing whole blood.
Figure 4. Thrombin generation in flowing blood in contact with collagen coated stents. Empty silicone sleeve (), bare stent (), and stent wrapped in a sheet of NC (?), HC (), HEC (), or HHC () were perfused with recalcified whole blood at a flow rate of 30 μL/min. Depicted is the thrombin concentration at the outlet of the flow reactor as a function of perfusion time. Arrows indicate time of extensive clot formation and occlusion of the flow reactor tubing.
Discussion
Different heparinized collagen preparations, namely collagen crosslinked with extensively periodated heparin (HC), HC with immobilized mildly periodated heparin (HHC), and collagen crosslinked with EDC/NHS and treated with mildly periodated heparin-crosslink (HEC), were evaluated for their thrombogenic properties. We wished to reveal to what extent heparin as a collagen crosslinker or covalently bound to the collagen surface is able to reduce collagen’s thrombogenicity. Our results demonstrate that numerous platelets adhered to collagen films, and that these platelets supported the process of thrombin generation at the collagen surface. Heparinization of collagen could not abolish the binding and collagen-induced activation of platelets. However, in contrast to nontreated collagen surfaces, no thrombin generation could be detected on heparinized surfaces.
Platelet adhesion, determined as the percentage of surface coverage, was the same for NC, HC, HEC, and HHC. This result clearly indicates that heparinization of collagen (either as a crosslinker or immobilized on crosslinked collagen) does not prevent, or even reduce, platelet adhesion. Moreover, platelets that adhere to heparinized collagen become activated, resulting in pseudopod formation and the appearance of phosphatidylserine in the outer leaflet of the plasma membrane. This result appears to contrast with findings of other investigators, who demonstrated that heparin coatings significantly decreased platelet adhesion.22,23 However, our observation that platelet adhesion on collagen is not inhibited by the presence of immobilized heparin is in line with a previously reported study that demonstrated that heparin immobilization on carbodiimide crosslinked collagen even slightly increased the number of adherent platelets.14 Contradictory findings in the effects of immobilized heparin on platelet adhesion might be related to the type of surface to which heparin is bound.
Collagen-induced platelet procoagulant activity is mediated by the interaction of collagen with its platelet receptor, GPVI.24 It is apparent that crosslinking collagen with heparin did not inhibit the loss in phospholipid membrane asymmetry, as monitored by annexin-V binding. In contrast, collagen-induced platelet procoagulant activity was significantly reduced when collagen films were treated with mildly periodate oxidized heparin, implying that under these circumstances, when the polysaccharide chain of heparin is almost fully intact, immobilized heparin interferes with platelet activation by collagen. Whether immobilized heparin affects the interaction of collagen with its primary platelet receptor GPVI or with a secondary one (P2Y2 and/or P2Y12) awaits further experimentation. Because these experiments were performed with adherent platelets that were extensively washed, heparin-accelerated inactivation of traces of thrombin is highly unlikely. Interference of heparin with platelet activation by collagen was previously demonstrated under both static and flow conditions.25–27 One study showed that under static conditions or at low shear rate, immobilized heparin did not affect platelet deposition to collagen, but strongly inhibited the platelet release reaction.26 Another study in rats revealed that under flow conditions (low and high shear rate), heparin released by stimulated mast cells did not block 2?1-dependent platelet adhesion, but attenuated subsequent platelet activation as well as fibrinogen binding to platelets.27 Together, these results support the notion that mildly periodate oxidized heparin may interfere with collagen-induced platelet activation (ie, release reaction and exposure of negatively charged phospholipids in outer leaflet of the plasma membrane).
Exposure of heparinized collagen films with adherent (activated) platelets to recalcified platelet-free plasma demonstrated that, independent of the method of heparinization, thrombin formation is abolished completely during the time of the experiment (60 minutes). We surmise that the amount of AT (±2 pmol/cm2) that binds to collagen films crosslinked with highly periodated heparin is already sufficient to prevent thrombin generation at these surfaces. These results extend earlier findings from our group illustrating that AT bound to immobilized heparin is probably the most effective way to prevent thrombus formation at an artificial surface.19,28,29 To confirm that the thromboresistance of the heparinized collagen surfaces was indeed AT-dependent, thrombin generation was measured with AT-depleted plasma. While thrombin generation on heparinized collagen with adhering platelets was strongly inhibited when exposed to normal plasma, significant thrombin generation was observed when the same experiment was performed with AT-depleted plasma, suggesting that the thromboresistance of heparinized collagen surfaces is fully dependent on the presence of AT. The physiological relevance of our findings was established in a perfusion setup, where stents coated with heparinized collagen films were exposed to flowing whole blood. While an explosive thrombin formation was observed in blood that passes the nonheparinized collagen-coated stents, stents coated with heparinized collagen appeared to be fully thromboresistant. We postulate that immobilization of heparin effectively catalyzes the thrombin–antithrombin reaction, whereby levels of surface-located thrombin remain below the threshold values that are needed to exert its positive feedback reactions, such as activation of the cofactors factors V and VIII and generation of procoagulant cell membranes, that regulate thrombin formation. The in vitro finding that heparinized collagen films are highly thromboresistant encourages further research on the application of collagen in medical devices in contact with blood.
References
Link J, Feyerabend B, Grabener M, Linstedt U, Brossmann J, Thomsen H, Heller M. Dacron-covered stent-grafts for the percutaneous treatment of carotid aneurysms: effectiveness and biocompatibility–experimental study in swine. Radiology. 1996; 200: 397–401.
Holmes DR, Camrud AR, Jorgenson MA, Edwards WD, Schwartz RS. Polymeric stenting in the porcine coronary artery model: differential outcome of exogenous fibrin sleeves versus polyurethane-coated stents. J Am Coll Cardiol. 1994; 24: 525–531.
Ratcliffe A. Tissue engineering of vascular grafts. Matrix Biol. 2000; 19: 353–357.
Nelson SR, deSouza NM, Allison DJ. Endovascular stents and stent-grafts: is heparin coating desirable? Cardiovasc Intervent Radiol. 2000; 23: 252–255.
Welt FGP, Rogers C. Inflammation and restenosis in the stent era. Arterioscler Thromb Vasc Biol. 2002; 22: 1769–1776.
Ross R. Atherosclerosis is an inflammatory disease. Am Heart J. 1999; 138: S419–S420.
Lee CH, Singla A, Lee Y. Biomedical applications of collagen. Int J Pharm. 2001; 221: 1–22.
Culav EM, Clark CH, Merrilees MJ. Connective tissues: matrix composition and its relevance to physical therapy. Phys Ther. 1999; 79: 308–319.
Lambert AW, Fox AD, Williams DJ, Horrocks M, Budd JS. Experience with heparin-bonded collagen-coated grafts for infrainguinal bypass. Cardiovasc Surg. 1999; 7: 491–494.
Cloft HJ, Kallmes DF, Lin HB, Li ST, Marx WF, Hudson SB, Helm GA, Lopes MB, McGraw JK, Dion JE, Jensen ME. Bovine type I collagen as an endovascular stent-graft material: Biocompatibility study in rabbits. Radiology. 2000; 214: 557–562.
Wissink MJB, Beernink R, Scharenborg NM, Poot AA, Engbers GHM, Beugeling T, van Aken WG, Feijen J. Endothelial cell seeding of (heparinized) collagen matrices: effects of bFGF pre-loading on proliferation (after low density seeding) and pro-coagulant factors. J Control Release. 2000; 67: 141–155.
van Wachem PB, Plantinga JA, Wissink MJB, Beernink R, Poot AA, Engbers GHM, Beugeling T, van Aken WG, Feijen J, van Luyn MJA. In vivo biocompatibility of carbodiimide-crosslinked collagen matrices: effects of crosslink density, heparin immobilization, and bFGF loading. J Biomed Mater Res. 2001; 55: 368–378.
Heemskerk JWM, Siljander PRM, Bevers EM, Farndale RW, Lindhout T. Receptors and signalling mechanisms in the procoagulant response of platelets. Platelets. 2000; 11: 301–306.
Wissink MJ, Beernink R, Pieper JS, Poot AA, Engbers GH, Beugeling T, van Aken WG, Feijen J. Immobilization of heparin to EDC/NHS-crosslinked collagen. Characterization and in vitro evaluation. Biomaterials. 2001; 22: 151–163.
Heemskerk JW, Vuist WM, Feijge MA, Reutelingsperger CP, Lindhout T. Collagen but not fibrinogen surfaces induce bleb formation, exposure of phosphatidylserine, and procoagulant activity of adherent platelets: evidence for regulation by protein tyrosine kinase-dependent Ca2+ responses. Blood. 1997; 90: 2615–2625.
Briedé J, Wielders S, Heemskerk J, Baruch D, Hemker H, Lindhout T. Von Willebrand factor stimulates thrombin-induced exposure of procoagulant phospholipids on the surface of fibrin-adherent platelets. J Thromb Haemost. 2003; 1: 559–565.
Blezer R, Cahalan L, Cahalan PT, Lindhout T. Heparin coating of tantalum coronary stents reduces surface thrombin generation but not factor IXa generation. Blood Coagul Fibrinolysis. 1998; 9: 435–440.
Hemker HC, Giesen PL, Ramjee M, Wagenvoord R, Beguin S. The thrombogram: monitoring thrombin generation in platelet-rich plasma. Thromb Haemost. 2000; 83: 589–591.
Keuren JFW, Wielders SJH, Willems GM, Morra M, Lindhout T. Fibrinogen adsorption, platelet adhesion and thrombin generation at heparinized surfaces exposed to flowing blood. Thromb Haemost. 2002; 87: 742–747.
Schoen P, Lindhout T, Willems G, Hemker HC. Continuous flow and the prothrombinase-catalyzed activation of prothrombin. Thromb Haemost. 1990; 64: 542–547.
Stuart MC, Bevers EM, Comfurius P, Zwaal RF, Reutelingsperger CP, Frederik PM. Ultrastructural detection of surface exposed phosphatidylserine on activated blood platelets. Thromb Haemost. 1995; 74: 1145–1151.
Niimi Y, Ishiguro Y, Nakata Y, Goto T, Morita S, Yamane S. Platelet adhesion to heparin coated oxygenator fibers under in vitro static conditions: impact of temperature. ASAIO J. 2001; 47: 361–364.
Bickel C, Rupprecht HJ, Darius H, Binz C, Hauroder B, Krummenauer F, Meyer J. Substantial reduction of platelet adhesion by heparin-coated stents. J Interv Cardiol. 2001; 14: 407–413.
Nieswandt B, Watson SP. Platelet-collagen interaction: is GPVI the central receptor? Blood. 2003; 102: 449–461.
Griffin MJ, Rinder HM, Smith BR, Tracey JB, Kriz NS, Li CK, Rinder CS. The effects of heparin, protamine, and heparin/protamine reversal on platelet function under conditions of arterial shear stress. Anesth Analg. 2001; 93: 20–27.
Kauhanen P, Kovanen PT, Lassila R. Coimmobilized native macromolecular heparin proteoglycans strongly inhibit platelet-collagen interactions in flowing blood. Arterioscler Thromb Vasc Biol. 2000; 20: E113–E119.
Lassila R, Lindstedt K, Kovanen PT. Native macromolecular heparin proteoglycans exocytosed from stimulated rat serosal mast cells strongly inhibit platelet-collagen interactions. Arterioscler Thromb Vasc Biol. 1997; 17: 3578–3587.
Keuren JFW, Wielders SJH, Willems GM, Morra M, Cahalan L, Cahalan P, Lindhout T. Thrombogenicity of polysaccharide-coated surfaces. Biomaterials. 2003; 24: 1917–1924.
Blezer R, Fouache B, Willems GM, Lindhout T. Activation of blood coagulation at heparin-coated surfaces. J Biomed Mater Res. 1997; 37: 108–113.(Jeffrey F.W. Keuren; Simo)