Search

A number of recent reports have described murine monoclonal antibodies that react specifically with the complex formed by human platelet membrane glycoproteins (GP) IIb and IIIa. We show that the IgG L, a previously described human alloantibody isolated from a polytransfused thrombasthenia patient, has similar properties. When used in non- precipitating amounts in crossed immunoelectrophoresis (CIE), 125I-IgG L bound strongly to the IIb-IIIa complex. However, after dissociation of the complex with EDTA, only a weak binding to GP IIb and no binding to GP IIIa was detected. In further studies, increased amounts of IgG L were interacted with 125I-labeled membrane glycoproteins in (a) CIE and (b) classical indirect immunoprecipitation experiments. Although the antibody was able to quantitatively precipitate the IIb-IIIa complex from Triton X-100-soluble extracts of platelet membranes, no precipitation of GP IIb or GP IIIa was observed after divalent cation chelation. Addition of EDTA to immunoprecipitates containing GP IIb- IIIa resulted in dissociation and partial release of both glycoproteins. The interaction of the IgG L with electrophoretically separated GP IIb and GP IIIa was studied using a Western blot procedure in the presence of Ca2+, Mg2+, or EDTA. The presence of divalent cations did not increase the reactivity of the antibody with the individual glycoproteins. Overall, our results show that acquired antibodies to IIb-IIIa, such as the IgG L, may predominantly react with complex-dependent determinants.

We prepared a heterohybrid cell line that secretes a human IgM monoclonal autoantibody that recognizes an antigen found on thrombin- activated or stored platelets. The surface expression of the epitope recognized by this autoantibody, 5E5, increases with time as platelets age in vitro, suggesting that it may represent a senescence or activation-specific antigen. 5E5 binds to the purified platelet membrane glycoprotein (GP) IIb-IIIa complex in an enzyme-linked immunosorbent assay (ELISA). In an immunoblot technique, 5E5 binds to a protein with an apparent mol wt of 95,000, which is identical to that of GPIIIa under nonreduced conditions. In crossed immunoelectrophoresis (CIE), the predominant antigen recognized by 5E5 is contained in the GPIIb-IIIa precipitin arc. An additional precipitin arc recognized by 5E5 is often observed only on gels derived from lysates of platelets stored under blood bank conditions for greater than 3 days. These findings illustrate the usefulness of human monoclonal antibodies for the identification of membrane neoantigens expressed as a result of platelet activation or revealed as platelets age in vitro.

The gray platelet syndrome is a rare inherited platelet disorder characterized by the absence of alp ha-granules as observed by electron microscopy. Analysis of the glycoprotein composition of the platelets of 2 such patients by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) revealed decreased or absent staining for carbohydrate of several high molecular weight glycoproteins. The major periodic acid Schiff (PAS) staining membrane glycoproteins were normally detected and were normally labeled with 125I during lactoperoxidase-catalyzed iodination. Analysis of the protein composition of gray platelets by single or two- dimensional SDS-PAGE followed by Coomassie blue staining revealed an apparent absence of GP Ig (thrombospondin), markedly reduced platelet fibrinogen and albumin concentrations, and severely reduced levels of 2 low molecular weight polypeptides exhibiting identical rates of migration on SDS-PAGE as platelet factor 4 and beta-thromboglobulin. SDS-PAGE profiles similar to those of the gray platelets were observed with normal human platelets that had undergone the release reaction induced by thrombin. Analysis of gray platelet proteins by crossed immunoelectrophoresis using a rabbit anti-human platelet antibody preparation and rocket immunoelectrophoresis using monospecific antisera confirmed the above findings and showed additional severe deficiencies of factor VIIIR:Ag and cold-insoluble globulin. In contrast, factor XIII (subunit A), a cytoplasmic protein, was normally detected. Our studies provide further evidence that circulating gray platelets specifically lack, or have markedly decreased concentrations of, the alpha-granule proteins.

Washed human platelets were labeled with 125I by the lactoperoxidase- catalyzed method and solubilized in 1% Triton X-100. The soluble proteins were analyzed by crossed-immunoelectrophoresis in 1% agarose, employing a rabbit antibody raised against whole human platelets. Analysis of autoradiograms developed from dried agarose gels led to the establishment of a normal reference pattern that was consistent for platelets obtained from more than 50 normal individuals. Six platelet membrane glycoprotein antigens contained in four distinguishable precipitates were identified. Each identification was based on direct sodium dodecyl sulfate-polyacrylamide gel electrophoretic analysis of 125I-antigens contained in individually excised precipitates. These platelet antigens include major membrane glycoproteins previously designated la, lb, lla, llb, llla, and lllb. Glycoproteins llb and llla were shown to be contained in a single immunoprecipitate, while glycoproteins la and lla were routinely detected in a single different immunoprecipitate. Analysis of soluble proteins from platelets of five patients with Glanzmann's thrombasthenia demonstrated either a complete absence or a marked reduction of only one radiolabeled precipitate, that containing membrane glycoproteins llb and llla. Platelet samples from two patients with Bernard-Soulier syndrome were devoid of a different precipitate, that containing membrane glycoprotein lb.

Triton X-100 soluble proteins from 125I-labeled human platelets were studied by crossed immunoelectrophoresis employing a multispecific rabbit antibody raised against whole normal platelets. Emphasis was placed upon an analysis of immunoprecipitates containing 125I-labeled major membrane glycoproteins, and in particular, a prominent immunoprecipitate containing a glycoprotein antigen (s) previously designated as protein 16. SDS-polyacrylamide gel electrophoresis of protein 16 precipitated by a monospecific alloantibody. IgG L . . . , confirmed the presence of both glycoproteins IIb and IIIa. 125I-IgG L . . . , at concentration below that capable of precipitating protein 16 by itself, bound specifically to the precipitate containing protein 16 produced by the multispecific rabbit antibody. No other precipitates formed by the rabbit antibody contained either glycoprotein IIb or IIIa. When platelet proteins, incubated with optimum concentrations of ethylenediamine tetraacetic acid (EDTA) or ethyleneglycol bis (B- aminoethylether) NN1-tetraacetic acid (EGTA), were electrophoresed against the rabbit antibody, previously unobserved immunoprecipitates that contained either free glycoprotein IIb or free glycoprotein IIIa were detected. Upon readdition of excess Ca++, but not Mg++, to the same protein samples, a single immunoprecipitate containing both glycoproteins was once again observed. It is thus demonstrated that glycoproteins IIb and IIIa can form Ca++-dependent complexes (protein 16) in Triton X-100 extracts of normal platelets. The potential significance of the reversible association of these glycoproteins to normal platelet function is discussed.

The sequence beta (3)203–228 is involved, in a yet undetermined manner, in alpha IIb beta 3 function. We now show that murine monoclonal antibody (MoAb) AP6, specific for beta (3)211–221, binds to alpha IIb beta 3 on adenosine diphosphate (ADP)-activated platelets only when the receptor is occupied by intact fibrinogen. The ligand-induced binding- site reported by AP6 is unique in that it is not expressed following occupancy by either RGD peptides or the gamma-chain carboxy-terminal dodecapeptide. Binding of AP6 to platelets coincides temporally with the binding of the MoAb 9F9, specific for a receptor-induced binding site on fibrinogen. Thus, AP6 reports the binding of fibrinogen to the recognition pocket of alpha IIb beta 3. Its binding to thrombin- stimulated washed platelets correlates with secretion as determined using an MoAb to P-selectin. When ultrathin sections of nonactivated platelets were examined by immunogold staining and electron microscopy, AP6 identified a pool of alpha IIb beta 3 colocalizing with P-selectin and suggesting the presence of alpha IIb beta 3-ligand complexes in the alpha-granule membrane. There was little binding of AP6 to surface alpha IIb beta 3 of unstimulated platelets. After ADP-induced activation, AP6 was abundantly distributed over the entire platelet surface, including pseudopods, but only when fibrinogen was present in the medium. ADP had little effect on AP6 reactivity within platelets. This contrasted with washed platelets and thrombin, where extensive AP6 binding was observed within internal membrane pools as early as 10 to 15 seconds after stimulation. Surface labeling with AP6 followed slower kinetics. Flow cytometry on Triton X-100 permeabilized fixed platelets confirmed AP6 binding to alpha IIb beta 3 within the platelet. Thus, our results provide evidence of (1) a pool of alpha-granule alpha IIb beta 3 occupied by ligand in nonactivated platelets; (2) thrombin- induced activation of alpha IIb beta 3 within the platelet, and (3) thrombin-induced mobilization of ligand-bound alpha IIb beta 3 to the surface.

Certain monoclonal antibodies (MoAbs) specific for platelet membrane glycoproteins are known to be capable of activating platelets, and it is generally thought that platelets from normal subjects are equally susceptible to stimulation by such MoAbs. We found that platelets from 20 normal donors varied significantly in their sensitivity to three IgG1 murine MoAbs specific for membrane glycoproteins CD9, GPIV (CD36), and the GPIIb/IIIa complex (CD41), respectively. The response of platelets to these MoAbs was blocked by prior addition of MoAb IV.3 specific for the Fc gamma RII receptor, indicating that activation was Fc receptor mediated. Platelets that responded poorly to these MoAbs failed to bind the MoAb 41H.16, specific for the “responder” form of Fc gamma RII, but platelets that responded well reacted with this MoAb. The average number of Fc gamma RII receptors on platelets from “responders” and “non-responders” was approximately the same. However, the number of Fc gamma RII receptors expressed influenced sensitivity of a subgroup of “responder” platelets to the anti-CD41 MoAb. These platelets were judged on the basis of MoAb binding studies to be heterozygous for the two alleles of Fc gamma RIIA. In contrast to their varying sensitivity to IgG1 MoAbs, members of the platelet panel responded equally well to 50H.19, an IgG2a MoAb specific for CD9, and these responses could not be blocked by MoAb IV.3 in the presence of plasma. This appears to be because of dual actions of 50H.19 on platelets: one FcR-dependent and the other complement-dependent. Our findings confirm previous reports that certain IgG1 MoAbs activate platelets through binding of their Fc domains to Fc gamma RII receptors and demonstrate that this response is influenced both by Fc gamma RII phenotype and (in the case of the anti-CD41 MoAb) by the number of Fc gamma RII receptors expressed. The failure of nonresponding platelets to bind detectable amounts of MoAb 41H.16, which is thought to recognize all Fc gamma RII receptors except for one allele of the Fc gamma RIIA gene, is consistent with the possibility that Fc gamma RIIA gene products, but not Fc gamma RIIB or Fc gamma RIIC gene products, are expressed on platelets.

We have characterized a murine IgG monoclonal antibody, OP-G2, specific for platelet glycoprotein (GP) IIb-IIIa (alpha IIb beta 3). OP-G2 Fab fragments inhibit fibrinogen-mediated platelet aggregation and competitively inhibit adenosine diphosphate-induced binding of 125I- fibrinogen to washed platelets. OP-G2 binding to GPIIb-IIIa is specifically inhibited by RGD-containing peptides but not the fibrinogen gamma-chain carboxy-terminal peptide, and OP-G2 Fab fragments, like RGD-containing peptides, alter the conformation of GPIIb-IIIa resulting in the expression of a ligand-induced binding site (LIBS) recognized by PMI-1. OP-G2 fails to bind to the recombinant Cam variant of GPIIb-IIIa (alpha III beta 3Cam) wherein an Asp119 to Tyr119 substitution in GPIIIa abrogates the ability to recognize RGD. These data indicate that OP-G2 recognizes an epitope at or in very close proximity to the RGD recognition site of GPIIb-IIIa and that, in every aspect tested, OP-G2 behaves like a macromolecular RGD ligand. Interestingly, two-color flow cytometry shows that OP-G2 IgG can bind to nonactivated platelets. Quantitative binding assays indicate that nonactivated platelets bind approximately 50,000 125I-OP-G2 molecules/platelet. Furthermore, the affinity of OP-G2 for platelets activated with thrombin is roughly fivefold higher (nonactivated, kd = 24.8 nmol/L; activated, kd = 4.9 nmol/L). These results suggest that the RGD recognition site of GPIIb-IIIa is available to macromolecules that contain RGD even on nonactivated platelets, provided that the affinity of the ligand is adequate.

We have compared the binding of affinity-purified anti-PlA1 IgG from seven nonrelated donors with chimeric integrin subunit beta 3 molecules expressed in the baculovirus-Spodoptera frugiperda insect cell system. Beta 3 chimeras were engineered to include segments of antigenic human beta 3 sequences spliced to intervening segments of nonantigenic Xenopus beta 3 sequence. Our results clearly show that antibodies from all seven donors will bind to nondenatured molecules containing the antigenic human beta 3 Cys26-Cys38 loop only when it is presented in a correct orientation that must be maintained by noncontiguous human sequences. Key downstream sequences are located within the region beta 3(288–490), flanking either side of the putative long-range disulfide at Cys435. Although our results confirm unambiguously that the Leu/Pro polymorphism at position 33 in human beta 3 is necessary for the expression of PlA epitopes, they also indicate that this polymorphic sequence alone is not sufficient. The requirement for additional human beta 3 sequence transcends the need to maintain a correct orientation within the Cys26-Cys38 loop itself, because the murine monoclonal antibody SZ21, which recognizes the sequence beta 3(28–35) contained within the Cys26-Cys38 loop, binds to all chimeras containing this loop, even if the same chimeras are not recognized by anti-PlA1. Our results indicate that additional noncontiguous residues encompassed by the sequence 288–490 either directly contribute to the composition of the PlA1 epitope or, more likely, maintain the Cys26-Cys38 loop in a proper orientation with respect to the remainder of the beta 3 molecule and thereby maintain proper antigenic presentation of the sequences in that loop.

We have used two murine monoclonal antibodies, each directed against one component of the human platelet membrane glycoprotein (GP) IIb-IIIa complex, to examine further the molecular requirements for fibrinogen binding to the platelet surface and subsequent platelet-platelet cohesion (aggregation). Although neither AP3, which is directed against GPIIIa, nor Tab, which is specific for GPIIb, were individually able to inhibit adenosine diphosphate (ADP)-induced fibrinogen binding, platelet aggregation, or secretion, the combination of AP3 and Tab completely abolished platelet aggregation and the release reaction. Unexpectedly, this synergistic inhibition of platelet-platelet cohesion occurred in the presence of apparently normal fibrinogen binding. Both the number of fibrinogen molecules bound and the dissociation constant for fibrinogen binding remained essentially unchanged in the presence of these two antibodies. Inhibition of aggregation was dependent upon the divalency of both AP3 and Tab because substitution of Fab fragments of either antibody for the intact IgG resulted in a complete restoration of both aggregation and secretion. In contrast to ADP induction, thrombin-activated platelets neither aggregated nor bound fibrinogen in the presence of AP3 plus Tab but were fully capable of secretion, which illustrated the multiple mechanisms by which the platelet surface can respond to different agonists. These data demonstrate that fibrinogen binding to the platelet surface alone is not sufficient to support platelet-platelet cohesion and that an additional post-fibrinogen-binding event(s) that is inhibitable by these two monoclonal antibodies may be required.

In agreement with previous studies, we observed that incubation of washed human platelets with EDTA at 37 degrees C for short periods caused an irreversible loss of their aggregation response to adenosine diphosphate and markedly diminished their capacity to bind fibrinogen. AP-2 is a monoclonal antibody that reacts with a determinant specific to the glycoprotein (GP) IIb-IIIa complex. We now report that in a direct binding assay, the number of sites for AP-2 on platelets incubated with EDTA at 37 degrees C fell to approximately 30% of those present on control platelets. This effect of EDTA was not observed at room temperature. Analysis of the treated platelets by sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed normal amounts of GP IIb and GP IIIa. However, studies using crossed immunoelectrophoresis with 125I-AP-2, 125I-Tab (anti-GP IIb), or 125I- AP-3 (anti-GP IIIa) in intermediate gels showed that at 37 degrees C, EDTA was inducing an irreversible change in GP IIb-IIIa complexes. A reduction in size and probable dissociation of the GP IIb-IIIa precipitate was accompanied by the appearance of precipitates having the characteristics of those given by free GP IIb and free GP IIIa and the location of a major new cathodal precipitate, which bound Tab and AP-3 but not AP-2. Membrane modifications associated with the loss of antigenic determinants on GP IIb-IIIa may explain EDTA-induced loss of platelet aggregability at 37 degrees C.

A murine monoclonal antibody specific for glycoprotein (GP)IIIa was prepared by immunization with a GPIIb- and GPIIIa-enriched Triton X-114 extract of platelet membranes. This antibody, designated AP-3, was shown by indirect immunoprecipitation to react solely with GPIIIa derived from either P1A1-positive or -negative individuals. The epitope on GPIIIa recognized by AP-3 is expressed on dissociated GPIIIa as well as on Ca+2-dependent complexes of GPIIb and GPIIIa, as shown by crossed immunoelectrophoresis in the presence or absence of EDTA. A previously described monoclonal antibody specific for the GPIIb/IIIa complex (AP- 2) inhibited platelet aggregation induced by ADP, thrombin, collagen, or arachidonic acid (Pidard et al, J Biol Chem 258:12582–12586, 1983). In contrast, AP-3 had no effect on aggregation induced by any of these reagents, a finding similar to that previously reported for the GPIIb- specific monoclonal antibody, Tab (McEver et al, J Clin Invest 66:1311- 1318, 1980). At saturation, 40,200 AP-3 molecules were bound per platelet, a value similar to that obtained for AP-2 or Tab. Thus, data derived using AP-3 indicate that significant amounts of free GPIIIa are not present, thereby supporting the hypothesis that GPIIb and GPIIIa exist complexed in a 1:1 stoichiometry in the plasma membrane of intact, nonactivated platelets.

In human platelet lysates prepared by addition of nonionic detergent (Triton X-100) or by sonication, the multimer composition and electrophoretic mobility of platelet von Willebrand factor (vWF) were consistently modified under conditions that would favor activation of the endogenous calcium-activated, sulfhydryl-dependent neutral protease (CAP). By sodium dodecylsulfate-agarose gel electrophoresis, native platelet vWF contained some multimers that were larger than those characteristic of plasma vWF. Modified platelet vWF contained a multimer population equivalent to or smaller than that of plasma vWF plus an additional fast-migrating band. In crossed immunoelectrophoresis (CIE), modified platelet vWF was characterized by a more anodic distribution and the appearance of a distinct, cross- reactive, anodic component previously designated VIIIR:Ag fragment. In the presence of calcium, radiolabeled purified plasma vWF was also degraded by the protease in question, with a decrease in the apparent molecular weight of the reduced monomer from 230,000 to 205,000. The VIIR:Ag fragment isolated from the same degraded plasma vWF by preparative CIE was shown to be composed of an identical mol wt 205,000 subunit. Because cleavage of plasma or platelet vWF was inhibited by prior addition of leupeptin, EDTA, ethylene glycol bis (beta-aminoethyl ether)-N, N, N', N'-tetraacetic acid (EGTA), or N-ethylmaleimide (agents known to inhibit platelet CAP) but was unaffected by numerous other protease inhibitors, including soybean trypsin inhibitor, benzamidine, hirudin, phenylmethylsulfonyl fluoride, aprotonin, or epsilon-aminocaproic acid (none of which inhibits platelet CAP), we conclude that proteolysis of vWF observed in this study is a direct effect of CAP and is not mediated by way of secondary proteases.

An electroimmunoassay was applied to the quantitation of platelet- associated IgG (PAIgG). Protein solubilized by Triton X100 from washed platelets was electrophoresed at pH 5.0 in agarose gels containing carbamoylated rabbit anti-human IgG (pI approximately equal to 5.0). Because the rabbit antibody is immobilized under these conditions, while PAIgG is freely mobile, rocket precipitates were produced, the heights of which were directly proportional to the amount of antigen (PAIgG) present. By this method, PAIgG for normal individuals was found to be 4.3 +/- 1.7 fg/platelet (mean +/- 2 SD; n = 35). Increased PAIgG levels (direct assay) were found in 27 of 29 patients with a diagnosis of clinically active idiopathic thrombocytopenic purpura (ITP), ranging from 10.5 to 101.5 fg/platelet. Moderately elevated PAIgG was found in 8 of 10 patients in an early stage of recovery from ITP (range 7.5–9.5 fg/platelet) and in 3 of 6 patients with apparent nonimmune thrombocytopenia (range 14.5–24.0 fg/platelet). Electroimmunoassay for PAIgG can be performed on patients with platelet counts as low as 2000/microliters, yields results in less than 24 hr, is highly reproducible, and appears to provide a useful tool for the evaluation of patients with immunologically mediated thrombocytopenia.

Platelets from patients with Montreal platelet syndrome (MPS) consistently display a defect in the mechanisms that regulate platelet size during shape change and undergo spontaneous aggregation and stir- induced microaggregate formation. We now provide data that the surface glycoprotein composition of MPS platelets is indistinguishable from that of normal platelets. However, a defect in calcium-activated neutral proteinase (calpain) was detected in MPS platelets. The specific activity of calpain in the cytosolic fraction of platelets from four MPS patients was found to be only 30% of that in platelets from normal control donors (n = 18, P less than .001). Additionally, platelets from MPS patients (n = 3) contained only 50% (P less than .001) of the calpain I catalytic subunit antigen found in platelets from normal control donors (n = 9). Platelets from the asymptomatic father/grandfather of the MPS patients had normal amounts of both total calpain proteolytic activity and calpain I catalytic subunit antigen. This represents the first report of a defect in calpain in human cells. The abnormally low calpain activity in MPS platelets may account for the platelet defects characteristic of this disorder.

Baka is a platelet alloantigen whose putative allele, Bakb, has not been identified previously. By using a serum, “Har,” obtained from a patient with posttransfusion purpura, we describe the platelet alloantigen Bakb. The Har serum reacted with an NP-40-extractable platelet membrane protein of 142 kd with mobility similar to platelet glycoprotein IIb alpha. We found that the antigen recognized by the Har serum is inherited in an autosomal dominant mode with an apparent gene frequency of .39. Chi-square analysis of observed and expected phenotype frequencies indicated that serum Har recognizes Bakb, the anticipated allele of Baka. Our findings provide new evidence for polymorphism of glycoprotein IIb and for the association of posttransfusion purpura with alloimmunization to determinants on this glycoprotein.

An antibody (DIL) from a patient with idiopathic thrombocytopenic purpura (ITP) was shown to have autospecificity on the basis of reactions with autologous platelets that were identical to those obtained with platelets from normal subjects. DIL antibody also reacted strongly in an immunofluorescence test with platelets from a patient with Glanzmann's thrombasthenia, but failed to react with platelets from a patient with the Bernard-Soulier syndrome who was known to be deficient in glycoprotein Ib (GPIb). Purified GPIb and control platelets, but not Bernard-Soulier platelets, inhibited the lytic activity of DIL. Using the GPIb-specific monoclonal antibody AP1 and one-dimensional rocket electrophoresis into gels containing rabbit antihuman platelet membrane antibody, it was shown that staphylococcal protein A-Sepharose beads coated with DIL antibody selectively remove GPIb from solubilized platelet preparations. By crossed immunoelectrophoresis it was found that DIL recognizes a determinant on GPIb on the membrane side of the cleavage site of the platelet calcium- activated protease (calpain). These studies provide direct evidence for binding of a platelet autoantibody to a determinant on GPIb relatively close to the site of insertion of this protein into the platelet membrane.

Conditions were developed in which 80% to 90% of platelet fibrinogen could be routinely purified in nondegraded form from the fluid phase of platelet suspensions stimulated with the calcium ionophore, A23187, in the presence of calcium, leupeptin, and prostaglandin E1. Fibrinogen was separated from other released proteins by chromatography on diethylaminoethanol (DEAE)-cellulose using a continuous pH and ionic strength gradient. Purified platelet fibrinogen, greater than 98% homogeneous by immunoelectrophoresis and sodium-dodecyl sulfate- polyacrylamide gel electrophoresis (SDS-PAGE), consisted of intact A alpha, B beta and gamma A chains, but not gamma' chains, and was 95% to 96% clottable. Platelet fibrinogen was shown to compete for the binding of radiolabeled plasma fibrinogen to ADP-activated platelets in a manner identical to that of unlabeled plasma fibrinogen itself. Also, at equivalent protein concentrations, platelet and plasma fibrinogens supported platelet aggregation to an equivalent extent. Based upon these results, we conclude that there is no significant difference between platelet and plasma fibrinogen with respect to their size, their clottability, their affinity for the activated platelet fibrinogen receptor, or their capacity to support subsequent platelet aggregation.

Analysis of human platelet membrane proteins by crossed immunoelectrophoresis (CIE) in the presence of Triton X-100 (TX-100) has previously shown that glycoproteins (GP) IIb and IIIa are located in a single immunoprecipitate, band 16.2 To investigate whether IIb and IIIa are associated in a complex, we have analyzed TX-100-solubilized 125I-labeled membrane proteins by density gradient ultracentrifugation using 10%-40% sucrose gradients containing the nonionic detergent. studies were performed using soluble proteins derived from membranes isolated in the presence or absence of EDTA. Analysis of gradient fractions by SDS-polyacrylamide gel electrophoresis showed that in the absence of divalent cation chelation, GP IIb and IIIa penetrated well into the gradient (fractions 15–17). Analysis of fractions 15–17 by CIE revealed the presence of band 16. In contrast, when the membrane proteins were incubated with EDTA prior to or after TX-100 solubilization, IIb and IIIa remained near the top of the gradient (fractions 8–11) and gave separate immunoprecipitates during CIE. Incubation of washed platelet lysates with leupeptin, an inhibitor of the Ca2+-dependent protease of human platelets, had no effect on the shape of the band 16 immunoprecipitate. Thus, for the first time, direct evidence has been obtained that GP IIb and IIIa may form a divalent cation-mediated complex. Calibration of the sedimentation profiles using proteins of known molecular weight suggests that the complex is of limited size. Indirect evidence suggests that the complex is a heterodimer.

To evaluate the possibility that changes in lipid composition might be related to the functional lesion that develops when platelets are stored as concentrates for several days, we measured lipid constituents of platelets in freshly prepared concentrates and in concentrates stored for 72 hr at 4 degrees C or at 20 degrees C under standard blood banking conditions. At 20 degrees C, but not at 4 degrees C, platelets lost about 15% of total cholesterol and 7%--11% of total phospholipid. The distribution of individual phospholipids remained unchanged. This was also true of the fatty acid distribution in total phospholipids and in individual phospholipids except for a statistically significant reduction of linoleic acid (18:2) and an increase in oleic acid (18:1) in phosphatidyl inositol (PI). Platelets collected in citrate-phosphate- dextrose (CPD) anticoagulant did not differ significantly in lipid composition from those collected in acid-citrate-dextrose (ACD) anticoagulant during the period of observation. These findings do not provide a basis to suggest that functional abnormalities developing in stored platelets are related to changes in lipid composition.

We have characterized a 120-Kd antigen that frequently reacts with serum antibodies from patients with immune thrombocytopenia or normal subjects. Immunoblots made after two-dimensional nonreduced/reduced sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) or two-dimensional isoelectric focusing/SDS-PAGE demonstrated that this 120-Kd protein has the same molecular weight under nonreduced or reduced conditions, is not a surface protein, and has an isoelectric point (pl) of 6.4 to 6.5. From these data, one likely candidate is the intracellular platelet protein, vinculin. Monoclonal antivinculin antibody reacts with this 120-Kd protein, and purified human platelet vinculin is bound by antibodies that recognize the 120-Kd protein. Therefore, we conclude that this 120-Kd protein is identical to vinculin. Data obtained from a sensitive enzyme-linked immunosorbent assay demonstrate the presence of naturally occurring antivinculin antibodies in many normal sera. However, the incidence of antivinculin antibodies in patient sera (67%; 55 of 82 sera) is significantly (P less than .01) higher than that in normal sera (40%; 32 of 80 sera), and there is a significant difference (P less than .05) between the mean levels of antivinculin antibodies in patient and normal sera. Whereas the levels of these antibodies in patient and normal sera overlap, 2 of 82 sera from patients with thrombocytopenia express unusually high levels of such antibodies. The pathologic significance of these antibodies remains to be determined.

One proposed ligand binding site on platelet integrin alpha IIb beta 3 is the region of the beta 3 subunit encompassing amino acids 211–221. However, we recently showed that synthetic peptides corresponding to amino acids 211–221 inhibit fibrinogen binding to alpha IIb beta 3 by binding to alpha IIb beta 3 and not to fibrinogen. In this study, we show that AP6, a monoclonal antibody (MoAb) directed against amino acids 214–221 of beta 3, bound to immobilized active alpha IIb beta 3 but did not inhibit fibrinogen binding to the complex. We then determined whether nonfunctional alpha IIb beta 3 on platelets with a beta 3 Arg-214-->Trp mutation (Strasbourg I variant of Glanzmann's thrombasthenia or GTV) could be induced to aggregate after treatment with dithiothreitol (DTT). DTT has been shown to expose the fibrinogen receptor on normal platelets. DTT treatment of GTV platelets did result in the formation of the fibrinogen binding site as indicated by the binding of pI-55, an MoAb that only binds to the activated form of alpha IIb beta 3. Furthermore, DTT-treated GTV platelets aggregated in the presence of fibrinogen and divalent cations. This aggregation was inhibited by EDTA, RGDS, and the selective alpha IIb beta 3 antagonist, Ro 43–5054. These data show that Arg-214 of beta 3 is not required for fibrinogen binding or for platelet aggregation. However, this amino acid appears to be critical for the formation and for the maintenance of the correct tertiary structure of the fibrinogen binding site on alpha IIb beta 3.

The relative antithrombotic effectiveness of targeting glycoprotein (GP) Ib-dependent versus GPIIb-IIIa-dependent platelet interactions has been determined in baboons by measuring thrombus formation after infusing comparable antihemostatic doses of anti-von Willebrand factor (vWF) monoclonal antibody (MoAb) BB3-BD5, anti-GPIb MoAb AP1, and anti- GPIIb-IIIa MoAb LJ-CP8 under conditions of arterial and venous flow (shear rates of 750 to 1,000 seconds-1 and 100 seconds-1, respectively). Thrombus formation was quantified as 111In-platelet deposition and 125I-fibrin accumulation on segments of collagen-coated tubing interposed in chronic exteriorized arteriovenous (AV) shunts for 40 minutes. In vitro, anti-vWF MoAb BB3 BD5 (IgG) and anti-GPIb MoAb AP1 [IgG or F(ab)2 fragments] inhibited ristocetin-induced platelet aggregation (IC50 50 nmol/L and 1 mumol/L, respectively), but neither of these MoAbs blocked platelet aggregation induced by adenosine diphosphate (ADP) (P > .5). Conversely, anti-GPIIb-IIIa MoAb LJ-CP8 inhibited platelet aggregation induced by ADP (IC50 1 mumol/L, but failed to block ristocetin-induced platelet aggregation (P > .5). In vivo, the intravenous infusion of anti-vWF MoAb BB3 BD5 or anti-GPIIb- IIIa MoAb LJ-CP8 into baboons at doses that abolished corresponding agonist-induced aggregation ex vivo (bolus injections of 0.5 mg/kg and 10 mg/kg, respectively) prolonged template bleeding times from baseline values of 4.0 +/- 0.3 minutes to > 27 +/- 4 minutes, and to > 26 +/- 4 minutes, respectively (P 3 .001 in both cases), without affecting the peripheral platelet count (P > .5). However, injection of anti-GPIb MoAb AP1 [10 mg/kg as IgG or 1 mg/kg as F(ab)2 fragments] produced immediate irreversible thrombocytopenia (< 40,000 platelets/microL). Anti-GPIIb-IIIa MoAb LJ-CP8 abolished platelet deposition and fibrin accumulation on collagen segments under both arterial and venous flow conditions (P < .01 in all cases), whereas MoAb BB3 BD5 produced minimal inhibition of platelet deposition and no decrease in fibrin accumulation at arterial shear rates and undetectable antithrombotic outcomes at low shear. Thus, inhibiting GPIIb-IIIa-dependent platelet recruitment abrogates both thrombus formation and platelet hemostatic function at both venous and arterial shear rates. By contrast, interfering with GPIb-vWF-dependent platelet interactions abolishes platelet hemostatic function without producing corresponding antithrombotic effects.

The activity and surface antigenicity of alpha 2 beta 1 on platelets from 27 normal subjects were found to vary significantly. A fourfold range of surface antigen correlates with a 20-fold variation in the ability of nonactivated, washed platelets to adhere to type I collagen and a fivefold variation in the adhesion of platelets to type III collagen. These differences in surface receptor are reflected in significant variation in the lag time required for type I collagen- induced platelet aggregation in platelet-rich plasma. Among the same individuals, no difference was observed in surface levels or activities of two other platelet integrins, the fibronectin receptor alpha 5 beta 1 and the fibrinogen receptor alpha IIb beta 3. In all cases studied, we observed complimentary differences in the incorporation of 125I into surface alpha 2 beta 1, in quantity of surface alpha 2 beta 1 antigens, and in alpha 2 beta 1 collagen receptor activity. Despite variations in these parameters, there was no difference in the electrophoretic mobility or isoelectric point of either integrin subunit among the individuals studied. The wide range of activity and antigenicity of this platelet collagen receptor may result from polymorphism(s) in the alpha 2 beta 1 genes, or the activity of alpha 2 beta 1 may be variably regulated by another gene product. The heterogeneity of platelet alpha 2 beta 1 that we describe in this report certainly explains previous discrepancies concerning the contributions of this integrin to platelet adhesion to collagens. Most importantly, differences in surface collagen receptor activity may correlate with a long-term risk toward thrombosis, impaired hemostasis, and/or cardiovascular disease.