Thomas W. Chow

Antibody inhibitors to von Willebrand factor metalloproteinase and increased binding of von Willebrand factor to platelets in ticlopidine-associated thrombotic thrombocytopenic purpura.

Ticlopidine, a potent antiplatelet agent used to maintain patency after coronary artery stenting and to prevent strokes in high-risk persons (1), has been associated with thrombotic thrombocytopenic purpura (TTP) (2-5). Thrombotic thrombocytopenic purpura, first described by Moschcowitz (6), is characterized by extensive platelet thrombi in the arterioles and capillaries. Abnormalities in von Willebrand factor multimers, including the presence of unusually large multimers and disappearance of the large multimers found in normal plasma, have been detected in many cases of the disease (7, 8). Furthermore, von Willebrand factor is abundant in the thrombi of patients with TTP (9), and flow cytometric studies have demonstrated that the factor is bound to platelets in the circulation of these patients during the most thrombocytopenic phase of the disease (10). The von Willebrand factor, a glycoprotein critical in mediating platelet deposition at sites of vessel injury, is synthesized and secreted by endothelial cells as a disulfide-linked polymer composed of a 2050amino acid monomer (11). Upon release into the circulation, it is cleaved by a plasma metalloproteinase in a shear-dependent manner (11) at the peptide bond between tyrosine-842 and methionine-843 (12). This cleavage decreases the size of the von Willebrand factor polymer, generates a series of multimers found in normal plasma, and produces dimers of 176-kD and 140-kD fragments (11). In the absence of the proteinase, large and unusually large von Willebrand factor multimers accumulate in the plasma. When unfolded by shear stress (13), these multimers exhibit an increased capacity to support platelet aggregation (14). Indeed, a deficiency of the proteinase has been reported in idiopathic TTP (15, 16). We investigated whether von Willebrand factor is involved in ticlopidine-associated TTP. Methods Patients Seven consecutive patients who developed TTP after initiation of ticlopidine therapy and were treated at the participating institutions from 1 January 1996 to 31 December 1998 were investigated. The criteria for the diagnosis of TTP were those described elsewhere (10, 16). We also determined proteinase activity in 17 controls: 7 consecutive, unselected patients without thrombocytopenia (age range, 62 to 81 years; 5 men and 2 women) who donated blood samples at routine follow-up examinations after 3 to 5 weeks of ticlopidine therapy prescribed for cardiac stents, and 10 randomly selected hospitalized patients not taking ticlopidine. Blood samples were obtained by venipuncture or at the time of plasmapheresis. The investigational protocol was approved by the institutional review boards of the participating centers. von Willebrand Factor Studies Platelet-bound von Willebrand factor, von Willebrand factor multimers, von Willebrand factorcleaving metalloproteinase activity, and the inhibitory activity of IgG to the von Willebrand factorcleaving metalloproteinase were measured as described elsewhere (10, 16). The von Willebrand factor bound to single platelets in EDTA-anticoagulated whole-blood samples was quantified by flow cytometry. Proteinase activity was expressed as a percentage of that in the pooled normal plasma control. Results The initial clinical and laboratory findings of the patients are summarized in the Table. The duration of ticlopidine therapy before diagnosis of TTP ranged from 2 to 7 weeks (median, 3 weeks). None of the patients had a history of autoimmune disorders, and none were receiving penicillins, antineoplastic chemotherapy, or oral contraceptives before onset of the disease. In all patients, remission occurred after ticlopidine therapy was discontinued and daily plasma exchange was instituted. The median number of plasma exchanges received by the patients was 10 (range, 5 to 30). None of the patients had relapse after plasma exchange was discontinued. Table. Clinical and Laboratory Findings von Willebrand Factor Binding to Single Platelets Binding of von Willebrand factor to platelets was studied in patients 1 to 3; the test was not available for the other 4 patients. Platelet-bound von Willebrand factor was 7.5, 4.5, and 4.5 arbitrary units, respectively (normal value<2.1 arbitrary units) in the initial blood samples; these values returned to normal when patients received plasma exchange and platelet counts increased. von Willebrand Factor Multimers In all seven patients, the largest multimers, which are found in normal plasma, were missing in the initial plasma samples. For patient 1, von Willebrand factor multimeric patterns in three of seven subsequent plasma samples were abnormal; one sample lacked the largest von Willebrand factor multimers, and two contained unusually large multimers. The initial plasma sample (obtained on day 1) for patient 3 was missing the largest multimers. The next two samples (obtained on days 2 and 4) contained unusually large multimers. The multimers were normal in the subsequent two samples (obtained on days 8 and 9). For patients 1, 2, and 6, plasma samples obtained upon remission were available for investigation; all samples showed normal multimeric patterns. von Willebrand Factor Metalloproteinase Activity For patient 1, only plasma samples obtained on days 4 to 6 after admission (when his platelet counts were 77 109/L, 70 109/L, and 76 109/L) were available for the study. These samples contained 28%, 17%, and 14%, respectively, of the proteinase activity found in plasma from normal controls. Proteinase activity was 0% in the initial plasma samples of patients 2 to 4 and 7%, 12%, and 4%, respectively, in patients 5 to 7. However, the plasma samples of these three patients were obtained from the plasmapheresis bags in amounts of 200 to 250 mL during the initial plasma exchanges. For patient 2, the protease activity increased to 100% on day 3, when the platelet count was 260 109/L. For patient 3, protease activity increased to 6%, 10%, 81%, and 77%, respectively, on days 2, 4, 8, and 9, when platelet counts were 25 109/L, 130 109/L, 280 109/L, and 325 108/L. Plasma proteinase levels in patient 5 increased to 23% and 55% on days 4 and 6, respectively, when platelet counts were 140 109/L and 180 109/L. A plasma sample obtained from patient 6 on day 5, when the platelet count had increased to 277 109/L, contained 94% proteinase activity. The mean (SD) plasma proteinase activity in the 7 controls receiving ticlopidine for 3 to 5 weeks was 114% 36%, which did not differ from the activity in the 10 randomly selected controls who were not treated with ticlopidine (97% 17%). In a previous study (16), 74 randomly selected patients without TTP had proteinase activity of 103% 23%. Inhibitors of von Willebrand Factor Proteinase To explore the causes of proteinase deficiency, the initial plasma sample of patient 2 was mixed with normal control plasma after treatment at 56 C for 30 minutes. The von Willebrand factor metalloproteinase activity in the mixture was suppressed to 23% of the activity found in a control mixture of heated normal pooled plasma and normal control plasma. Plasma samples from patients 3 to 7 were sufficient in volume for studies to determine whether their immunoglobulins inhibited the proteinase. The IgG isolated from patient 3 on day 1 exhibited a concentration-dependent inhibition of proteinase activity in normal control plasma. The concentration of the IgG molecules required to inhibit 50% of the protease activity in the mixture (IC50) was 2.2 mg/mL. The IgG isolated from the same patient on day 9 was not inhibitory. The IC50 of the IgG isolated from initial plasma samples of patients 4 to 7 was 5.5, 2.2, 4.4, and 2.2 mg/mL, respectively. In tests comparing susceptibility to inhibition, von Willebrand factor metalloproteinase in plasma from the normal controls and that in the plasma samples from controls who received ticlopidine were equally sensitive to inhibition by IgG isolated from patients with ticlopidine-associated TTP (data not shown). The inhibitory activity of IgG was abolished when it was incubated with antibodies to IgG Fab (data not shown). Discussion In two series of single-episode and intermittent idiopathic TTP (15, 16), inhibition of plasma von Willebrand factor proteinase by IgG autoantibodies was found to be characteristic. In support of a role of von Willebrand factor proteinase deficiency in the pathogenesis of platelet thrombosis, the deficiency was not observed in persons who did not have the disease. Furthermore, shear stress was found to increase the capacity of von Willebrand factor to support platelet aggregation (16). We now describe seven patients with ticlopidine-associated TTP who also had severely decreased levels of von Willebrand factor proteinase 2 to 7 weeks after initiation of ticlopidine therapy. The durations of ticlopidine therapy before the diagnosis of TTP are similar to the 2 to 12 weeks observed in 98 cases of TTP in a recently described series (17). The deficiency in our patients resolved after ticlopidine therapy was discontinued and plasmapheresis was instituted. The deficiency was not observed in randomly selected patients who had been receiving ticlopidine for similar durations but did not develop the disease. The absence or severe reduction of von Willebrand factor metalloproteinase was accompanied by binding of von Willebrand factor to platelets. Concurrently, the large von Willebrand factor multimers were missing. The level of von Willebrand factor proteinase activity required to prevent or decrease binding of von Willebrand factor to platelets and thrombosis was low (approximately 10% to 15%). Thus, even a slight increase in the proteinase activity was sufficient to suppress the values of platelet-bound von Willebrand factor. At this low level of proteinase activity, von Willebrand factor proteolysis remained defective. This explained the presence of unusually large von Willebrand factor multimers in the plasma. The decrease in von Willeb

Increased von Willebrand Factor Binding to Platelets in Single Episode and Recurrent Types of Thrombotic Thrombocytopenic Purpura

Extensive microvascular platelet aggregation is characteristic of thrombotic thrombocytopenic purpura (TTP). Previous studies have indicated that abnormalities of von Willebrand factor (vWf) are often present in TTP patient plasma. There has not been previously any direct evidence linking these abnormalities to the process of intravascular platelet aggregation in TTP. We used flow cytometry to analyze the binding of vWf to single platelets, and the presence of platelet aggregates, in the blood of 4 children with chronic relapsing (CR) TTP and 5 adults with single episode or recurrent TTP. vWf on the single platelets of CRTTP patients at all time points studied was significantly increased compared to controls, and was increased further as platelet counts decreased to levels below 40,000/μl. The single episode and recurrent adult TTP patients had platelet aggregates in the blood, as well as increased vWf on single platelets, before therapy commenced and thereafter until recovery was in process. In the one unresponsive single episode TTP patient, vWf on single platelets remained elevated, and platelet aggregates persisted, until her death. The platelet α‐granular protein, P‐selectin, was not increased on the single platelets of most TTP blood samples, suggesting that it is vWf from plasma (rather than from α‐granules) that attaches to platelet surfaces in association with platelet aggregation. These results suggest that vWf‐platelet interactions are involved in the platelet clumping process that characterizes TTP. Am. J. Hematol. 57:293–302, 1998.

Chimeric 7E3 Fab (ReoPro) decreases detectable CD11b on neutrophils from patients undergoing coronary angioplasty.

OBJECTIVES The purpose of this study was to monitor the effects of chimeric 7E3 Fab (ReoPro) on leukocyte and platelet activation and interaction during coronary angioplasty. BACKGROUND Increased expression of CD11b on monocytes and neutrophils promotes their adhesion to endothelial cells, extracellular matrix and smooth muscle cells. Thrombin-activated platelets adhere via P-selectin to monocytes and neutrophils. These cell interactions may affect the outcome of coronary angioplasty. METHODS During coronary angioplasty, venous blood was obtained for flow cytometric detection of leukocyte CD11b; platelet CD41a, CD61a and CD62P; the percentage of leukocytes with adherent platelets and the intensity of bound platelet fluorescence. RESULTS Leukocyte CD11b expression increased after angioplasty in control patients (neutrophils 171+/-25 to 255+/-31 mean fluorescence intensity [MFI, mean+/-SEM], n=25, p < 0.0001; monocytes 200+/-40 to 248+/-36 MFI, n=17, p < 0.05) and decreased in the patients selected to receive chimeric 7E3 Fab (neutrophils 146+/-30 to 82+/-22 MFI, n=25, p < 0.0001; monocytes 256+/- 53 to 160+/-38 MFI, n= 17, p < 0.05). Neutrophil CD11b decreased after in vitro incubation of whole blood with chimeric 7E3 Fab (n=5, p=0.01), but fMLP-induced increases in CD11b were not prevented. The CD11b expression was unchanged and increased with fMLP stimulation after in vitro incubation of isolated neutrophils with chimeric 7E3 Fab. Direct-labeled chimeric 7E3 Fab was not detected bound to neutrophils in whole blood or isolated cells using flow cytometric techniques. Adhesion of isolated neutrophils to protein-coated glass was not prevented by in vitro incubation with chimeric 7E3 Fab. Platelet activation increased after angioplasty in control patients (CD62P 8.9+/-0.8 to 12.3+/-1.2 MFI, n=25, p < 0.05; CD41a 382+/-25 to 454+/-26 MFI, n=25, p < 0.05, CD61a 436+/-52 to 529+/-58 MFI, n=11, p < 0.05); it did not increase in the patients selected to receive chimeric 7E3 Fab (CD62P 13.2+/-1.0 to 9.0+/-0.9 MFI, n=25, p < 0.05; CD61a 398+/-32 to 410+/-38 MFI, n=7, p=NS). Leukocytes with adherent platelets tended to increase in the control group of patients and decrease after the procedure in patients selected to receive chimeric 7E3 Fab; individual and procedure-related variability were marked. CONCLUSIONS Despite standard aspirin and heparin therapy, leukocyte and platelet activation with platelet adherence to leukocytes occurs after coronary angioplasty. Although chimeric 7E3 Fab does not bind to leukocytes directly, it influences CD11b expression in whole blood. Modulation of platelet and leukocyte activation and interaction by chimeric 7E3 Fab may contribute to an improved outcome after coronary angioplasty.

Thrombotic thrombocytopenic Purpura: Understanding a Disease No Longer Rare

Serial studies of plasma samples from patients during episodes of thrombotic thrombocytopenic purpura (TTP) have often shown either the presence of unusually large (UL) von Willebrand factor (vWf) multimers or, alternatively, absence of the largest plasma vWf forms. The presence of ULvWf multimers in TTP patient plasma may reflect impaired processing of the ULvWf forms released from endothelial cells. The disappearance of ULvWf and large vWf multimers in some TTP patient plasma samples during acute TTP episodes may be predominantly because these ULvWf forms, along with the largest vWf multimers, bind to platelets and cause aggregation. Serial flow cytometry studies of EDTA-whole blood samples from patients with initial episode, intermittent, and chronic relapsing types of TTP confirm that vWf is the likely aggregating agent, perhaps in association with fluid shear stress. The amount of vWf bound to single platelets has been found to be significantly increased during TTP relapses relative to remission periods in patients with all types of TTP. A substance in normal platelet-poor plasma and the cryoprecipitate-depleted fraction of normal plasma (cryosupernatant) is capable in vitro of reversibly reducing the size of ULvWf multimeric forms released by endothelial cells into the somewhat smaller vWf multimers ordinarily in circulation. This activity has characteristics of a limited disulfide bond reductase. The process of ULvWf breakdown may be made irreversible by the tandem proteolysis, catalyzed by a vWf metalloproteinase, of partially reduced vWf multimers. Several patients with chronic relapsing TTP have decreased or absent plasma vWf metalloproteinase activity, apparently on a congenital basis. Adult initial episode and intermittent TTP patients have been found to have vWf metalloproteinase activity inhibited by an autoantibody during, but not after, TTP epidsodes.

Shear Stress-Induced Binding of von Willebrand Factor to Platelets

Shear stress-induced platelet aggregation requires von Willebrand factor (vWF), platelet glycoprotein (GP) Ib, GPIIb-IIIa, Ca2+, and adenosine diphosphate (ADP). Recent reports using vWF labeled with either 125I or fluorescein isothiocyanate (FITC) have demonstrated that in shear-fields, vWF binds to both GPIb and GPIIb-IIIa. The sequence of the vWF finding to the two platelet receptors has not been precisely determined in these reports. In this study, a flow cytometry technique using a primary anti-vWF antibody and a secondary FITC IgG antibody was used to measure shear stress-induced vWF binding to platelets. Washed normal platelets suspended at 50,000/microliters with purified large vWF multimers were exposed to laminar shear stresses of 15 to 120 dynes/cm2 for 30 sec. At this low platelet count, little or no aggregation occurred in the shear fields. A significant increase in post-shear vWF-positive platelets was consistently observed. Experiments with platelets from normal and severe von Willebrands disease (vWD) (which lack plasma and platelet alpha-granule vWF) demonstrated that exogenous vWF predominately contributed to the platelet-vWF binding. Blockade of platelet GPIb with the monoclonal anti-GPIb antibody, 6D1, completely inhibited shear stress-induced platelet-vWF attachment. In contrast, blockade of GPIIb-IIIa with monoclonal anti-GPIIb-IIIa antibodies, 10E5, or c7E3, or with the GPIIb-IIIa-blocking tetrapeptide, RGDS had little or no inhibitory effect on platelet-vWF binding. These data demonstrate that the binding of vWF to GPIb is likely to be the initial shear-induced platelet-ligand binding event.

Increased von Willebrand factor (vWf) binding to platelets associated with impaired vWf breakdown in thrombotic thrombocytopenic purpura.

Thrombotic thrombocytopenic purpura (TTP) is a disorder of systemic platelet aggregation. Evidence has accumulated that the aggregating agonist in TTP of all types is likely to be von Willebrand factor (vWf), especially unusually large vWf multimers derived from endothelial cells. Recent evidence indicates that a metalloproteinase involved in vWf breakdown is produced in inadequate amounts in children with chronic relapsing TTP. Chronic relapsing TTP is, therefore, likely to be a congenital enzyme deficiency. In adults with single episode or intermittent types of TTP, the vWf metalloproteinase is inhibited by autoantibodies that are present either transiently or intermittently in patient blood. Single episode and intermittent types of TTP in adults are likely to be short‐term or recurrent autoimmune processes, respectively. J. Clin. Apheresis 13:126–132, 1998.

Thrombin receptor activating peptide (SFLLRN) potentiates shear-induced platelet microvesiculation

Shear-induced activation of platelets plays a major role in vascular thrombosis. Shear stress tends to increase both platelet aggregation and procoagulant activity. One mechanism for increased procoagulant activity is promotion of the transbilayer movement of anionic phospholipids from the inner to the outer leaflet of the platelet membrane bilayer. This is accompanied by vesiculation of the platelet membrane, resulting in the formation of procoagulant membrane particles called microvesicles. In this study we have examined the effect of various platelet agonists on shear-induced platelet microvesiculation and the development of platelet procoagulant activity. Normal citrated whole blood was subjected to laminar shear rate up to 12,500 sec(-1) (shear stress approximately 375 dyne/cm2) in a cone-and-plate viscometer, and the formation of platelet microvesicles was measured by flow cytometry under different conditions. Elevated levels of shear stress induced significant microvesiculation. We investigated the effects of adenosine diphosphate, epinephrine, thromboxane A2 analog, collagen, and thrombin receptor activation peptide (SFLLRN) on shear-induced platelet microvesiculation. The thrombin peptide significantly increased shear-induced microvesicle formation. In contrast, under similar conditions, the other agonists had no significant effect on shear-induced microvesiculation. These studies suggest that thrombin formed in the vicinity of primary hemostatic plugs in areas of elevated shear stress may have a major role in the propagation of thrombi by potentiating shear-induced platelet microvesiculation.

Importance of plasma fibronectin in determining PFP and PRP clot mechanical properties

Plasma fibronectin has been shown to be incorporated into fibrin clots. The serum concentration of fibronectin is 20-50% lower than the plasma concentration (1). Grinnell and Feld, using an indirect immunofluorescence analysis on blood clots prepared on plastic strata, demonstrated that the fibrin-platelet meshwork was covered with a uniform coating of fibronectin (2). Fibronectin binds to fibrin by either non-covalent attachment or covalent crosslinking. The covalent crosslinking of fibronectin to the fibrin alpha-chain is via an epsilon-(gamma-glutamyl)lysine linkage, mediated by Factor XIIIa (3). Each fibronectin molecule has two such binding sites. Thus it appears that fibronectin may play a role in blood coagulation. It has been suggested that the mechanical properties of fibrin clots may be enhanced by the presence of fibronectin crosslinking (2). Kamykowski, et al. (4) have shown that the shear modulus of a ligated clot formed from purified fibrinogen can be either increased or decreased if fibronectin is present, depending on the ionic strength and the pH during network formation. Although fibronectin is associated with the network structure in plasma clots, the mechanical role of this fibronectin has not been established. In this study we examine quantitatively the effect of plasma fibronectin on the dynamic rigidity moduli of fibrin clots formed from platelet free plasma (PFP), as well as from platelet rich plasma (PRP).

In vitro comparison of blood pump induced platelet microaggregates between a centrifugal and roller pump during cardiopulmonary bypass

Platelets are consumed during cardiopulmonary bypass (CPB) and mechanical ventricular assistance, at least partly as a result of the formation of platelet microaggregates in the blood pump. There is no commonly accepted method currently available to detect platelet microaggregates during the use of CPB or left ventricular assist devices (LVAD). The purpose of this study was to develop a flow cytometric method for the quantification of platelet microaggregates generated in blood pumps, and to evaluate the effect of cellular fragments from hemolyzed erythrocytes on the perioperative assessment of platelet counts during CPB. Method Fresh human anticoagulated blood (1IU heparin /mL, activated clotting time 250 ± 24 sec.) was circulated for 120 minutes in an artificial circulatory system, containing either a centrifugal pump (CP) or roller pump (RP). Whole blood was used to quantify platelet consumption and to detect circulating platelet microaggregates in a flow cytometer. Platelet consumption was additionally analyzed using an automated “Coulter” blood cell counter. Hemolysis was analyzed by measurement of plasma free hemoglobin (fHb), as well a by flow cytometric detection of red blood cell (RBC) fragments. Results Flow cytometric analysis demonstrated significantly more circulating platelet aggregates and platelet consumption in the RP than in the CP (p<0.01). Quantification of RBC fragments and plasma free hemoglobin (fHb) levels also indicated significantly increased hemolysis in the RP than in the CP (p<0.01). In contrast, the Coulter count data indicated less platelet consumption in the PP compared to the CP. Conclusion Fragments from hemolyzed erythrocytes have the same size distribution as intact platelets and the number of RBC fragments correlates with the extent of pump-induced hemolysis during CPB. Our data suggest that assessment of platelets by “Coulter counting” cannot distinguish platelets from RBC fragments and may underestimate platelet consumption in the presence of hemolysis during CPB. We conclude that flow cytometry is more accurate in the perioperative assessment of platelet count and platelet aggregation during CPB and LVAD support.

Direct detection of red blood cell fragments: a new flow cytometric method to evaluate hemolysis in blood pumps.

Pump induced hemolysis is presently evaluated by measuring plasma free hemoglobin (fHb). However, this method has disadvantages because quantification of fHb depends on hematocrit (HCT) and hemoglobin (Hb) levels. The aim of this work was to devise a hemoglobin independent method, capable of quantifying cell trauma directly by measuring the number of red blood cell (RBC) fragments. Whole blood flow cytometry was used to quantify circulating RBC fragments derived from a roller pump (Sarns, Inc. Model 2 M 6,002) and a centrifugal pump (Gyro C1E3, Kyocera Corp.). The pumps were tested in a mock circuit for 2 hr (5 L/min flow against 100 mm Hg pressure head). Red blood cell fragments were quantified by a phycoerythrin (PE) labeled glycophorin A antibody specific for erythrocytes. Red blood cell fragments were smaller than the intact RBC population and overlapped in size with the platelet population (based on forward- and side-light scattering measurements). For the roller pump, the values for RBC fragments increased from 1,090 ± 260/&mgr;l at 0 min to 14,880 ± 5,900/&mgr;l after 120 min. In contrast, using the centrifugal pump, there was little increase in RBC fragments (from 730 ± 270/&mgr;l at 0 min to 1,400 ± 840/&mgr;l after 120 min). Flow cytometry can be used for the rapid, sensitive, hemoglobin independent evaluation of pump induced RBC trauma.