Shear Stress–Induced Detachment of Blood Platelets From Various Surfaces
Abstract Platelet accumulation is the result of platelet adhesion and detachment. This study describes platelet detachment from fibronectin, laminin, fibrinogen, von Willebrand Factor (vWF), endothelial cell matrix (ECM), and collagen type III. Platelets adhered after 5 minutes’ perfusion of anticoagulated whole blood at different shear rates were subjected to a brief flush of 1 minute with HEPES-buffered saline at varying shear stress. Platelets adhering to fibronectin and laminin were most easily detached. Fibrinogen and vWF had an intermediate position. Almost no detachment occurred from ECM and collagen type III. Dendritic platelets were removed more easily than spread platelets. When the shear rate at which adhesion had occurred was raised, platelet detachment decreased strongly. When the time period between adhesion and detachment was increased, platelet detachment also decreased. From these results, we conclude that detachment is determined initially by the shear rate at which platelets adhere, then by the time they are allowed to settle, then by the nature of the surface, and then by the degree of spreading. The shear optimum for a given adhesive protein is not determined by the detachment.
- Received November 25, 1996.
- Accepted April 21, 1997.
Platelet adhesion needs to be studied under flow conditions to mimic the physiological situation. It is usually quantified by evaluating the accumulation of platelets on a surface after a given time. Platelet accumulation is the result of platelet adhesion and platelet detachment. Real-time studies using fluorescently labeled platelets have shown that a considerable degree of detachment may occur under some conditions.1 This detachment may vary depending on the adhesive protein to which platelets stick and on the shear rate used for the study.
The effect of the shear rate on platelet adhesion has been studied for most adhesive proteins.2,3 No shear optimum was found for collagen type III, ECM, and surface-coated vWF: Increased shear rate leads to increased adhesion up to 2000 s−1, and adhesion remains unchanged above this shear rate.3 Fibronectin showed an optimum at 300 s−1,4 laminin at 800 s−1,5 and thrombospondin at 1500 s−1.6 Adhesion to fibrinogen/fibrin showed a broad shear optimum between 500 and 1000 s−1, with a gradual decrease toward 50% of the optimum at 2000 s−1.7 It is currently unclear whether this shear optimum is determined to a larger extent by adhesion or detachment of platelets. In the current study, we have measured the detachment of platelets from various adhesive proteins and from ECM to shed some light on this question.
C17, a monoclonal antibody directed against GPIIIa inhibiting ADP-induced platelet aggregation,8 was a kind gift from Dr A.E. von dem Borne (Central Laboratory Blood Transfusion Service, Amsterdam, The Netherlands). Human collagen type III from placental origin was obtained from Sigma Chemical Co. LMWH (Fragmin) was from Kabi Pharmacia. Mouse laminin (A1, B1 ,B2) was from GIBCO. Human fibrinogen was from IMCO. Synthetic dRGDW peptide was kindly provided by Dr J. Bouchaudon (Rhône-Poulenc-Rorer, Vitry, France). Fibronectin was isolated from human plasma by affinity chromatography on gelatin-Sepharose as previously described.9 vWF was purified from cryoprecipitates by gel filtration on Sepharose 4B (Pharmacia AB) as described earlier.9,10 vWF in the void volume was stored at 4°C and used within 2 weeks after purification.
Coating and Spraying of the Coverslips
Glass coverslips (Menzel Gläser 18×18 mm) were cleaned overnight by a chromium trioxide solution and rinsed with distilled water before spraying. Monomeric collagen type III was solubilized in 50 mmol/L acetic acid and spray-coated with a density of 30 μg/cm2 on glass coverslips with a retouching airbrush (Badger model 100, Badger Brush Co).11 After the spraying procedure, the collagen surface was incubated for 1 hour with 1% human albumin in PBS (10 mmol/L phosphate buffer, pH 7.4, .15 mol/L NaCl) to block further protein binding.
Human endothelial cells derived from umbilical veins were isolated and cultured as described before.12,13 For the experiments described, endothelial cells were cultured on glass coverslips, previously coated with gelatin. To isolate the extracellular matrix, cells were grown to confluence and exposed to 0.1 mol/L NH4OH for 3 to 5 minutes at room temperature. The cell layer was completely removed by this procedure, leaving the extracellular matrix intact.14 The extracellular matrix was washed three times with PBS before use.
Platelet adhesion to adsorbed adhesive proteins (vWF, fibronectin, laminin, fibrinogen) was studied by using square glass microcapillaries (Microcell, The Wilmad Glass Company Inc). Coating was performed by incubating the microcapillaries with 25 μL of purified vWF (10 μg/mL), fibronectin (300 μg/mL), laminin (100 μg/mL), and fibrinogen (100 μg/mL) in PBS for 1 hour, followed by a 15-minute incubation with 1% human albumin solution in PBS to block further protein binding. The concentration chosen for the coating of the capillaries is the concentration of the different proteins that gives optimal platelet adhesion. After coating, microcapillaries were kept in PBS until perfusion. In separate experiments it was shown that the coated proteins remain bound to the surfaces during the perfusion experiments.
Perfusion studies over ECM and collagen type III were carried out in a specially devised small parallel plate perfusion chamber with well-defined rheological characteristics devised to accommodate a glass coverslip.15 Whole blood obtained by venipuncture from healthy volunteer donors was anticoagulated with 0.1 vol 200 U/mL LMWH(LMWH-blood) or 0.1 vol 110 mmol/L trisodium citrate (citrated blood). Citrated blood was use for platelet adhesion studies to vWF, fibronectin, fibrinogen, and endothelial cell extracellular matrix. LMWH-blood was used for platelet adhesion to laminin and collagen because for these surfaces the presence of cations is necessary for optimal adhesion. Whole blood was prewarmed at 37°C for 5 minutes and then drawn from a container through the chamber for 5 minutes at different wall shear rates with the use of a syringe and an automated syringe pump (Harvard Apparatus). Some of the coverslips were then taken from the chamber for evaluation as described below. The other coverslips were subjected to a brief perfusion at variable shear rates with 10 mmol/L HEPES buffer, pH 7.4, containing 0.15 mol/L NaCl (HBS) during 1 minute (viscosity HEPES buffer 0.75±0.01 cpoise). These coverslips were then taken from the chamber, rinsed in HBS, fixed in 0.5% glutaraldehyde/PBS, dehydrated in methanol, and stained with May-Grünwald–Giemsa as previously described.16 Platelet adhesion was evaluated with a light microscope, and the coverage was measured with an image analyzer (AMS 40 to 10). A homogeneous coverage of platelets was always observed. Fixation, staining, and evaluation for the square microcapillaries was identical. The microscope was focused on the upper glass wall of the capillary. Adhesion was expressed as the percentage of the surface covered with platelets. The effect of the applied shear stress during the second perfusion was presented as attached platelets remaining after the second perfusion. This effect was expressed as percentage of the platelet coverage obtained after the first perfusion (not subjected to the 1-minute perfusion with HBS). In some experiments, spread and dendritic platelets were counted. For this purpose, five duplicate fields at a magnification of 400× were imaged and printed, and the total number of spread and dendritic cells was counted. Dendritic and spread cells were recognized by eye. The mean area of a dendritic cell was 8.2±2.2 μ2 and the mean area of a spread cell 42.1±11.3 μ2, indicating that clear distinction had been achieved.
Perfusions were also performed with glass microcapillaries coated by adsorption with vWF, fibronectin, laminin, and fibrinogen as described. This procedure was followed to reach very high shear rates. Perfusions were performed with a single passage, drawing blood from a container through the tube at various wall shear rates as described above. Perfusions were directly followed by a 1-minute flush of HBS at varying shear stress as described above.
Scanning Electron Microscopy
For scanning electron microscopy, coverslips with platelets fixed for 1 hour with 2% glutaraldehyde/PBS at room temperature were postfixed in 1% osmium tetroxide for 1 hour, dehydrated in a graded series of ethanol, and dried by the critical point procedure, using CO2 as transitional fluid. The samples were sputter-coated with a thin layer of gold and viewed in a scanning electron microscope (Cam Scan S2).
Platelet Detachment From Isolated Adhesive Proteins
Platelets were allowed to adhere to glass coverslips coated with the adhesive proteins fibronectin, laminin, fibrinogen, and vWF for 5 minutes at a shear rate of 300 s−1. The percentage coverage and morphology of the adhered platelets varied dependent on the respective adhesive protein, as described previously.4,5,7 For fibronectin, surface coverage was 22.0±0.5% and consisted of mostly dendritic with some spread platelets. For laminin, surface coverage was 37.7±15.5% and consisted of dendritic and spread platelets with still prominent pseudopods. For fibrinogen, adhesion surface coverage was 49.4±2.9% and consisted of a mixture of dendritic platelets, with little sign of spreading and many thin pancake-like fully spread platelets. For vWF, surface coverage was 49.8±10.1% and consisted of a mixture of dendritic and thin fully spread platelets. Platelets were then detached from the surface by a 1-minute flush of HBS. The results of these detachment studies are shown in Fig 1⇓. The remaining adhesion to fibronectin after a 1-minute flush with HBS at different shear stress is linearly related to the logarithm of the shear stress with a half maximum of approximately 3 dyne/cm2. The remaining adhesion to laminin was linearly related to the logarithm of the shear stress as well, with a half maximum at 2.6 dyne/cm2. In the case of fibrinogen, a linear relationship was found between the surface coverage and the shear stress, with a half-maximal shear stress of 26 dyne/cm2. For vWF, a linear relationship was found between the surface coverage and the logarithm of the shear stress, with a half-maximal shear stress of 26 dyne/cm2. A finding characteristic of vWF was that not all platelets were completely removed, even at the highest shear rate that could be reached.
Morphology of the Remaining Platelets
Intuitively, one does expect that platelets that are more spread out will be less easily removed at high shear rates. At very high shear rates, damage to adherent cells may occur, which may lead to spuriously low results. We therefore quantified the morphology of adherent platelets on various surfaces after the detachment experiment and compared it with the morphology before the experiment. The results are shown in Fig 2A⇓ through 2D. For all four surfaces, dendritic platelets detached more quickly than spread platelets. This finding was most pronounced for fibronectin (Fig 2B⇓). The detachment of spread platelets tended to be linear with the shear stress. Of note is the presence of some dendritic platelets still present on vWF at high shear stress. Spread platelets remaining on fibrinogen and vWF at the highest shear stress regularly contained a central hole. This hole is shown in more detail in the scanning electron micrograph (Fig 3⇓). Immunofluorescence staining of fibrinogen demonstrated that these holes showed exactly the same staining as the surface surrounding the attached platelets, suggesting that part of the platelet had completely disappeared and that the underlying surface coat had become accessible to the fluorescent antibody (data not shown).
Platelet Detachment From the ECM
Platelet adhesion studies to the ECM were performed at 100 and 300 s−1, giving a surface coverage of 16.5±3.4% and 34.8±1.6%, respectively (Table 1⇓). Detachment studies were performed in the single-pass small perfusion chamber, in which a maximum shear stress of 30 dyne/cm2 could be attained. A flush with HBS lasting 1 minute at this shear stress caused a small but significant decrease in surface coverage to 87.5±0.9% of the control value when platelet adhesion was caused by flowing blood at a wall shear rate of 300 s−1. A remaining attachment of 57.8±4.7% of control was found when the adhesion had occurred at 100 s−1. When platelet spreading at 300 s−1 was prevented by using a high concentration of dRGDW (100 μmol/L), platelet adhesion was decreased to 12.1±0.9%, but the subsequent platelet detachment was increased, so that 56.6±6.0% of control remained attached to the ECM (Table 2⇓).
Platelet Detachment From Collagen Type III
Platelets adhere to collagen type III at 100 and 300 s−1, with formation of aggregates. Surface coverage was 4.6±0.2% and 15.4±0.5%, respectively. A brief flush with HBS for 1 minute at 30 dyne/cm2 had no effect at all at 300 s−1 but caused a significant reduction to 59.7±4.1% of the control when adhesion had occurred at 100 s−1 (Table 1⇑). When platelet aggregation and platelet spreading were prevented at 300 s−1 with use of dRGDW, platelet adhesion increased considerably, as has been published before.17 A perfusion with HBS during 1 minute at 30 dyne/cm2 resulted in only a slight nonsignificant decrease toward 87.6±11.3% of control (Table 2⇑).
Effect of the Shear Rate at Which Platelets Have Adhered on the Subsequent Detachment
The studies on platelet detachment from the ECM and collagen type III showed the importance of the initial shear rate at which platelets had attached for the subsequent detachment at high shear stress. We have studied this detachment in detail also for fibronectin, laminin, fibrinogen, and vWF. The results are summarized in Fig 4⇓, which shows the results on ECM and collagen type III (Table 1⇑) as well, for comparison. The percentage of platelets that remains attached after a 1-minute flush at a shear stress of 2.2 (fibronectin and laminin) or 30 dyne/cm2 (fibrinogen, vWF, ECM, and collagen type III) increased sharply in the range between 100 and 600 s−1. For ECM and collagen type III, almost no detachment was observed by 300 s−1. Fibronectin was studied only up to a shear rate of 300 s−1 because the adhesion decreases strongly above this shear rate.4
Effect of the Time Period Between the Two Perfusions on Detachment
Dendritic cells were removed more easily than spread cells. In time, platelets showed an increased tendency to spread. Therefore, we studied the influence of the perfusion time of the first run on the detachment of platelets. The results are summarized in Fig 5⇓, which shows that after 3 minutes’ perfusion, 65% of the platelets detached in the second flush. After 5 minutes, the percentage of detached platelets decreased to 56%, while after 10 minutes, only 19% of the platelets were detached with the second flush.
Platelet adhesion to the injured vessel wall is the first step in the formation of a hemostatic plug or a thrombus. The adhesion process is usually studied in perfusion chambers in which blood is circulated over an adhesive surface during a given length of time.18,19 The platelet adhesion is then evaluated by morphometry or by use of a fluorescent or radioactive label. Direct observations in real time with the use of fluorescent platelets have shown that adhesion is resultant of a complex series of events consisting of rolling, attachment, formation of pseudopods, spreading, and detachment.1 Studies in which the shear rate was varied showed a strong effect of these variations on platelet adhesion. This shear rate effect is the overall result of increased transport, increased spreading, platelet activation, and increased detachment. Although the overall process of platelet adhesion to various surfaces has been studied extensively, only one study has reported a detailed investigation of platelet detachment under shear stress.17 In that study, platelets were allowed to settle during 12 minutes from platelet-rich plasma onto fibrinogen-coated glass. Detachment was studied in a special tapered perfusion chamber after a 5-minute flush with PBS. This study showed a linear relationship between the shear stress and detachment, with half-maximal detachment at 35 dyne/cm2. The detachment appeared to depend on platelet shape, with complete detachment of round platelets, 50% detachment of dendritic platelets, and almost no detachment of fully spread platelets.
Relation Between Shear Stress and Detachment
We have studied the detachment of platelets that adhered under flow from a series of adhesive surfaces and adhesive proteins. We found, on the basis of a shear rate for adhesion of 300 s−1, that there was a hierarchy, with ECM and collagen type III as the surfaces to which platelets were attached most firmly, von Willebrand factor and fibrinogen as a second group with firm attachment, and fibronectin and laminin as the proteins with the least firm attachment. The results that we observed on platelets adhering to fibrinogen after 5 minutes’ flow at 300 s−1 were in the same range as the results of Jen et al17 after 12 minutes’ adhesion from platelet-rich plasma under static conditions. Detachment was linear, with the shear stress with half-maximal detachment at 26 dyne/cm2. The results on fibronectin, laminin, and von Willebrand factor were somewhat different. Detachment was linear with the logarithm of the shear stress. A likely explanation may be that platelets are spread to a larger degree on fibrinogen. As seen in Fig 2A⇑, at least 40% of the platelets are fully spread out on fibrinogen, which corresponds to a large proportion of the surface coverage. Spread platelets are removed in a linear relation to the shear stress, and this dominated the overall picture in the case of fibrinogen to a larger extent than for the other adhesive proteins.
Relation Between Morphology and Detachment
The platelets that remained attached to fibrinogen and vWF were fully spread, and high stress (50 dyne/cm2) caused mechanical damage, with a central hole where part of the platelet had been torn away. This may indicate that the central part of a spread platelet is less strongly attached to the surface than the periphery. Also, on other surfaces, spread platelets remained attached relatively better than dendritic platelets. Jen et al17 suggested on the basis of their results on fibrinogen that the degree of spreading is the main determinant of platelet detachment. Our current results confirm this observation when a single shear rate is selected for adhesion and when a single surface is studied. This does not mean, however, that detachment is completely determined by the degree of spreading. Closer study of Fig 2⇑ shows that at a shear rate at which all dendritic platelets have disappeared from fibronectin, 50% remain still attached to fibrinogen or vWF. Also, perfusion studies depicted in Fig 4⇑ indicate that dendritic platelets may remain fully attached when they have adhered at high shear rates, since up to 30% of the adhering platelets is dendritic and no detachment occurs even at the highest shear stress.
Detachment as Determinant of Shear Resistance
As mentioned in the introduction, adhesion to various proteins shows a clear-cut hierarchy regarding the shear rate. Fibronectin has a shear optimum at 300 s−1 and laminin at 800 s−1.4,5 For von Willebrand factor, ECM, and collagen type III, no shear optimum was observed.3 Adhesion to fibrin(ogen) showed a broad shear optimum between 500 and 1000 s−1, with a gradual decrease toward 50% of the optimum at 2000 s−1.7 Comparison of the shear rate dependence of adhesion with the detachment data observed at a shear rate of 300 s−1 shows considerable discrepancy. Laminin and fibronectin show similar detachment, but their shear optima are different. vWF and fibrinogen have approximately similar detachment but a different shear optimum for adhesion. This observation suggests that the adhesion optimum is not primarily determined by the detachment.
Detachment From Complex Matrices
The detachment from ECM and collagen type III was minimal when adhesion occurred at 300 s−1. This is most likely due to the fact that several different adhesive interactions are involved in both of these cases. Adhesion to collagen type III is dependent on vWF and on direct interactions between collagen type III and platelet receptors. Platelet adhesion to collagen type III leads to thrombus formation. It is striking that a shear stress of 30 dyne/cm2 has an effect neither on adhesion nor on thrombus formation once thrombi had formed during 5 minutes’ perfusion at 300 s−1.
Platelet spreading on ECM is known to be mediated by GPIIb-IIIa.20 Adhesion in the presence of the GPIIb-IIIa blocking peptide dRGDW gave a decrease in platelet adhesion and a pronounced increase in detachment. On collagen type III, a surface on which GPIIb-IIIa interaction and spreading are not important, detachment was not affected by dRGDW, although thrombus formation was completely abolished, with a rise in platelet adhesion as the consequence of the higher platelet concentration in the marginal layer, as has been reported before.21
In conclusion, platelet detachment is determined by the shear rate at which adhesion occurred, the adhesive surface, and the degree of spreading, in that order of importance. Complex matrices are better able to retain platelets on their surface than single proteins. Adhesion is a complex process determined only to a limited extent by detachment of platelets that have adhered as dendritic or spread platelets.
Selected Abbreviations and Acronyms
|ECM||=||endothelial cell matrix|
|vWF||=||von Willebrand factor|
The monoclonal antibody C17 was a kind gift from Dr A.E. von dem Borne (Central Laboratory Blood Transfusion Service, Amsterdam, The Netherlands) and the synthetic dRGDW peptide was provided by Dr J. Bouchaudon (Rhône-Poulenc-Rorer, Vitry, France).
Wu YP, van Breugel HHFI, Lankhof H, Wise RJ, Handin RI, de Groot PG, Sixma JJ. Platelet adhesion to multimeric and dimeric von Willebrand factor and to collagen type III preincubated with von Willebrand factor. Arterioscler Thromb Vasc Biol. 1996;16:611–620.
Beumer S, IJsseldijk MJW, De Groot PG, Sixma JJ. Platelet adhesion to fibronectin in flow: dependence on surface concentration and shear rate, role of membrane GP IIb/IIIa, and VLA-5 and inhibition by heparin. Blood. 1994;84:3724–3733.
Hindriks GA, IJsseldijk MJW, Sonnenberg A, Sixma JJ, De Groot PG. Platelet adhesion to laminin: role of Ca2+ and Mg2+ ions, shear rate, and platelet membrane glycoproteins. Blood. 1992;79:928–935.
Agbanyo FR, Sixma JJ, De Groot PG, Languino LR, Plow EF. Thrombospondin-platelet interactions: role of divalent cations, wall shear rate and platelet membrane glycoproteins. J Clin Invest. 1993;92:288–296.
Hantgan RR, Hindriks G, Taylor R, Sixma JJ, De Groot PG. Glycoprotein Ib, von Willebrand factor, and glycoprotein IIb/IIIa are all involved in platelet adhesion to fibrin in flowing whole blood. Blood. 1990;76:345–353.
Houdijk WPM, Sakariassen KS, Nievelstein PFEM, Sixma JJ. Role of factor VIII-von Willebrand factor and fibronectin in the interaction of platelets in flowing blood with monomeric and fibrillar collagen types I and III. J Clin Invest. 1985;75:531–540.
Jaffe EA, Nachman RL, Becker CG, Minick CR. Culture of human endothelial cells derived from umbilical veins: identification by morphologic and immunologic criteria. J Clin Invest. 1973;52:2745–2756.
van Zanten GH, Saelman EUM, Schut-Heese KM, Wu Y-P, Slootweg PJ, Nieuwenhuis HK. Platelet adhesion to collagen type IV under flow conditions. Blood. 1996;85:3862–3871.
Nievelstein PFEM, d’Alessio PA, Sixma JJ. Fibronectin in platelet adhesion to human collagen types I and III: use of nonfibrillar and fibrillar collagen in flowing blood studies. Arteriosclerosis. 1988;8:200–206.
Jen CJ, Li HM, Wang JS, Chen HI, Usami S. Flow-induced detachment of adherent platelets from fibrinogen-coated surface. Am J Physiol. 1996;270:H160–H166.
Saelman EUM, Hese KM, Nieuwenhuis HK, Uzan A, Cavero I, Marguerie G, Sixma JJ, de Groot PG. Aggregate formation is more strongly inhibited at higher shear rates by dRGDW, a synthetic RGD-containing peptide. Arterioscler Thromb. 1993;13:1164–1170.