Location of the quinine-dependent epitope on platelet glycoprotein IIIa.

Quinine-dependent (Q) IgG antibodies (Q.Ab) in drug-induced immune thrombocytopenia are heterogeneous and bind to different platelet surface glycoproteins (GP), namely GPIb, IX, IIb, IIIa and an unidentified 57-kDa membrane proteins. Although both the Q-dependent epitope on GPIIIa and the P1A1 antigen require intact disulphide bonds for their expression, they are distinct because Q.Ab bind to GPIIIa lacking P1A1. Epitopes for both antigens were examined on Western blots of either intact washed human platelets or purified GPIIIa. When intact platelets were digested with trypsin and washed and solubilised prior to electrophoresis, membrane-associated fragments of GPIIIa of 78 kDa were found to be reactive with both antibodies. In addition, 60- and 68-kDa fragments bound anti-P1A1 but not Q.Ab. Similar digestion with chymotrypsin produced only 60-kDa fragments containing both epitopes. Digestion of purified GPIIIa with chymotrypsin produced 60-kDa peptides reactive with Q.Ab and anti-P1A1 in immunoblotting studies. Similar digestion with elastase produced 58-kDa fragments also containing the epitopes for both antibodies. Longer digestion times or sequential digestion with different enzymes did not reveal extra fragments. However, immunoprecipitation of trypsin-digested 125I-labelled GPIIIa with affinity-purified Q.Ab produced a 17-kDa fragment containing the Q-dependent epitope.

Expression of the drug-dependent antigen for quinine-dependent antiplatelet antibodies on GP IIIa but not that on GPIb, IIb or IX on human endothelial cells.

In quinine- and quinidine-induced thrombocytopenic purpura IgG antibodies are known to react in a drug-dependent manner with different combinations of surface glycoproteins (GP) Ib, IIb, IIIa and IX. Because endothelial cells share a number of properties of platelets, including the presence of GP IIIa and GPIb-like proteins, we have compared these two cell types for their quinine-dependent IgG binding abilities. By immunoblotting of endothelial cells, quinine-dependent IgG binding from four patient sera was observed only to a 93 kDa component corresponding to GP IIIa. Strong IgG binding independent of the drug was found at 170-180 kDa. Thus endothelial cells express the GP IIIa quinine-dependent epitope on platelet GP IIIa, but not those on other platelet glycoproteins.

Platelet aggregating IgG antibody to platelet surface glycoproteins associated with thrombosis and thrombocytopenia.

Previously described platelet-aggregating antibodies associated with thrombosis and thrombocytopenia required heparin for their in vivo and in vitro expression. We have observed a patient with thrombosis who became thrombocytopenic during heparin treatment, but who suffered further thrombotic events and continued thrombocytopenia for 3 months after heparin withdrawal. The patient's plasma contained a potent platelet aggregating factor reactive with both his own and normal platelets in the absence of heparin. It also caused [14C]serotonin secretion from labelled platelets from normal donors and patients with either Glanzmann's thrombasthenia or Bernard-Soulier syndrome. This factor was an IgG and was neutralized by antibody specific for IgG lambda light chains. While the patient was thrombocytopenic an IgG paraprotein with lambda light chains was detected by isoelectrofocussing. After corticosteroid treatment it disappeared and the patient recovered. The active, but not the recovery serum contained IgG which immunoprecipitated a glycoprotein with characteristics of Glycoprotein IV from platelets labelled with Na[3H]BH4/periodate. Thus platelet-aggregating IgG antibodies with direct specificity for platelet surface glycoproteins may be associated with thrombosis/thrombocytopenia.

Naturally occurring human IgG antibodies to intracellular and cytoskeletal components of human platelets.

Immunoblotting of platelets that have been subjected to SDS-PAGE has revealed that sera from normal individuals contain IgG which binds to many platelet components. This binding was seen with autologous and heterologous platelets using serum of males and of nulliparous females who had not received blood transfusions. Although binding patterns of different sera were not identical, almost all sera caused IgG binding to platelet components of 87-90 kD, 140 kD (identified as vinculin) and 220-240 kD (tentatively identified as talin and actin-binding protein). Purified IgG showed the same binding pattern as whole serum and F(ab')2 fragments retained their ability to bind to many components. The titre of IgG binding in serum was 1:50-1600 while that of alloantibodies to the PlA1 antigen was 1:3200. IgG binding components were not secreted when platelets were stimulated and were rarely associated with isolated membranes, but were located either in platelet cytoplasm or cytoskeletons. IgG binding was decreased by absorbing sera with lysed platelets or isolated cytoskeletons, but only slightly with intact platelets. Microaffinity purification of IgG which formed a major band on immunoblots showed that it was antibody with specificity for vinculin or its degradation products. These findings suggest that normal sera contain naturally occurring IgG antibodies with specificity for intracellular platelet antigens and that in some cases their titre approaches that of antibodies of pathological significance.

Arterial thrombosis associated with immune thrombocytopenia: presence of a platelet aggregating IgG synergistic with thrombin and adrenalin.

We report the case of a 50-year-old lady who presented with arterial thrombosis in the setting of thrombocytopenia. Investigations confirmed the diagnosis of idiopathic thrombocytopenic purpura. A spontaneous platelet aggregating factor (SPAF) was isolated from the immunoglobulin fraction of the patient's plasma. The isolated IgG irreversibly aggregated platelet-rich plasma and washed platelets, an effect abolished by pretreating the platelets with aspirin. The activity of the IgG was greatly enhanced by subaggregatory concentrations of thrombin and adrenalin and was localized to the F(ab')2 of the molecule. Plasmapheresis in combination with anti-platelet therapy resulted in an increase in the patient's platelet count, reduced platelet aggregating activity of plasma and significant clinical improvement. We suggest that the presence of this platelet aggregating IgG contributed to the development of thrombosis in our patient and postulate that a similar factor may explain the paradox of thrombosis observed in a select group of thrombocytopenic patients.

Differing susceptibilities of platelets from normal individuals and patients with von Willebrand's disease to attack by quinine- and quinidine-dependent antiplatelet antibodies.

Reactivity of quinine- and quinidine-dependent antiplatelet antibodies has been compared in platelet-rich-plasma (PRP) from normal donors and from patients with von Willebrand's disease (vWd). One quinine-dependent antibody (Q.Ab) caused platelet aggregation and [14C] serotonin release with only 7 of 12 normal donors, while another Q.Ab and a quinidine-dependent antibody (Qd.Ab) caused aggregation and release with all 12. Drug-dependent IgG binding and PF 3 availability induced by the antibodies were, however, comparable in all donors. Differences in responsiveness were associated with platelets and not plasma. vWd platelets showed normal drug-dependent IgG binding, but decreased aggregation and serotonin release to most drug-dependent antibodies. Responsiveness was not restored by purified vWf:Ag, but, in one case, was corrected by normal plasma or cryoprecipitate. Drug-dependent binding of the Q.Ab which caused variable responsiveness in normals was to the same platelet antigens (GPIb and GPIIIa) in both normal and vWd platelets and did not require plasma components. Reduced PF 3 availability was seen with some antibodies in some vWd patients. Plasma from two of these patients inhibited aggregation of normal platelets to Q.Ab and one of these inhibited aggregation to ADP. Antiplatelet antibodies were detected in these two plasmas by ELISA. Thus some Q.Ab produce different responses with platelets from different donors. In vWd, reduced responsiveness to Q.Ab and Qd.Ab may result from production of inhibitory antiplatelet antibodies.

Heterogeneity of drug-dependent platelet antigens and their antibodies in quinine- and quinidine-induced thrombocytopenia: involvement of glycoproteins Ib, IIb, IIIa, and IX.

The molecular nature of platelet receptors for quinine- and quinidine-dependent antiplatelet antibodies (Q.Ab and Qd.Ab) was studied by immunoblotting. One Q.Ab caused quinine-dependent IgG binding to platelet proteins with molecular weights (mol wts) of 174 Kd and 93 Kd and another to only a 93-Kd protein. A third Q.Ab caused binding to 174-, 140-, 93-, and 57-Kd proteins, while a fourth Q.Ab and a Qd.Ab caused IgG binding to 174- and 18-Kd proteins. Using platelets from patients with Glanzmann's thrombasthenia or Bernard Soulier syndrome and purified GPIIIa, these proteins were shown to be GPIb, GPIIb, GPIIIa, GPIX, and an unidentified 57-Kd protein missing in Bernard Soulier syndrome. Binding to the 93-Kd protein was independent of the PIA1 antigen. Absorption of one Q.Ab with Glanzmann's thrombasthenia platelets revealed different populations of antibodies with different specificities within the one patient. Thus Q.Ab and Qd.Ab are heterogeneous and may be directed toward different epitopes on major platelet glycoproteins.

Thrombin-independent activation of platelet factor XIII by endogenous platelet acid protease.

Platelets contain factor XIII, an A subunit zymogen form of transglutaminase (TGase), that is activated by thrombin. In addition a thrombin-independent TGase (A#) was observed. A# was formed in platelet preparations lysed at acid pH, and its generation inhibited by protease inhibitors and alkaline pH. When maximal A# activity was generated in acidified lysates no further TGase activity could be induced by subsequent treatment with thrombin. Both FXIII zymogen and A# copurified as for FXIII, from either alkaline or from acidified platelet lysates respectively, by the conventional procedure. The pH optima, Km's for NN dimethyl casein, molecular weights, heat lability of active forms, requirements for calcium and reducing agents, and immunological characteristics of both TGases were the same. Studies with inhibitor substrates suggested that a thrombin-like cathepsin C or carboxypeptidase was responsible for A# formation. Purified FXIII zymogen could be activated directly by cathepsin C. Thus, the predominant, and probably only, TGase of platelets is factor XIII, which may be activated either by thrombin or by endogenous platelet acid protease(s).

Coagulation activities of plasma microparticles.

An evaluation of the effect of plasma microparticles (MP) on in vitro coagulation has been undertaken using platelet rich (PRP), platelet poor (PPP) and platelet free (PFP) plasmas prepared by differential centrifugation. MP provide coagulant material which shortens the activated partial thromboplastin time (APTT) and dilute simplastin time (DSTT) which is different from that contributed by commercial phospholipid preparations. The amount of platelet factor three (PF3) available in plasma is directly correlated with the centrifugal force used in its preparation and is present in large amounts in the MP pellet remaining after preparation of PFP. Factor VIII (F.VIII:C) and von Willebrand factor (vWf) were associated with the MP fraction but could be separated from MP on sucrose gradients. The effect of MP on the APTT was independent of the F.VIII:C/vWf and was not solely due to their PF3 content. Plasma prepared for routine coagulation assays contains MP which contribute to the APTT and DSTT and should be considered in their assessment. High speed centrifugation of plasma reduces the F.VIII:C/vWf:Ag/RCoF levels and this may contribute to losses of these proteins during preparative procedures utilising high speed centrifugation.

Platelet-associated immunoglobulins G, A and M are secreted during platelet activation: normal levels but defective secretion in grey platelet syndrome.

Platelet-associated (PA) IgG is known to be released from normal human platelets when they are stimulated by aggregating agents. We have studied whether PA-IgA and PA-IgM are also secreted during platelet activation or during blood collection and processing and whether their levels are related to those in serum. Processing of platelets from normal donors in the presence of secretion inhibitors prostaglandin E1 (PGE1) and theophylline increased levels of both surface and total PA-immunoglobulins (Ig) in intact and lysed platelets respectively, with increases being significant for surface PA-IgA and PA-IgM and total PA-IgM. About 50% of total PA-IgM, 40% PA-IgA and 20% PA-IgG was detectable on intact platelets. All three PA-Ig and PA-albumin were secreted in response to thrombin and this release was inhibited by PGE1. The platelet:serum ratio of each Ig and albumin were similar. In grey platelets deficient in alpha-granules, PA-Ig and PA-albumin levels were raised per platelet but when increased platelet size was taken into account PA-Ig were normal and PA-albumin just below normal. Although thrombin caused release of most of the small amounts of beta-thromboglobulin present, PA-Ig and PA-albumin were not released. This suggests that PA-Ig and albumin from plasma may enter a pool of secretory proteins in normal platelets, but in grey platelets they remain in some other site.

Effects of nicotine on platelet function.

In non-smokers, only high concentrations of nicotine (10 mM) caused platelet aggregation in platelet-rich plasma and release of 5-hydroxytryptamine (5-HT). Both responses to ADP and 5-HT were enhanced at 1 and 10 mM nicotine while they were inhibited to collagen, ristocetin, adrenaline and arachidonic acid. Lower concentrations of nicotine had no effect with any agent other than 5-HT, with which a variable enhancement of 5-HT-induced aggregation was observed. Uptake of 14C-5-HT was inhibited by nicotine at 100 microM or higher while platelet factor 3 availability was unaffected. Thus it is unlikely that direct effects of nicotine on platelets are responsible for smoking-related changes in platelet reactivity.

Shortening of bleeding time by 1-deamino-8-arginine vasopressin (DDAVP) in the absence of platelet von Willebrand factor in Gray platelet syndrome.

The Gray platelet syndrome is a rare disorder characterised by the absence of platelet alpha-granules and their contents. We describe a new patient and the effects of infusions of 1-deamino-8-arginine vasopressin (DDAVP). The patient had a prolonged skin bleeding time and his platelets had reduced numbers of alpha-granules, increased vacuolation and reduced retention on glass beads. Platelet von Willebrand factor antigen (vWf:Ag) was undetectable and levels of platelet fibrinogen, beta-thromboglobulin, platelet factor 4 and thrombospondin were reduced. All tests of plasma coagulation factors were normal, including Factor VIII (F.VIII:C), vWf:Ag, ristocetin cofactor (R:CoF) and botrocetin cofactor. Platelet ATP, ADP, platelet albumin, surface membrane glycoproteins and 14C-serotonin uptake were also normal. Infusions of DDAVP increased plasma F.VIII:C, vWf:Ag and R:CoF and shortened the bleeding time on two occasions. This suggests that DDAVP shortens the bleeding time by releasing vWf:Ag and/or other proteins from cellular storage sites other than the platelet.

Platelets in idiopathic thrombocytopenic purpura are increased in size but are of normal density.

To determine whether platelet size and volume are related to one another or to platelet age, subpopulations of platelets from patients with idiopathic thrombocytopenic purpura (ITP) have been produced on the basis of density using Percoll gradients. The density distribution of platelets from patients with ITP and from patients with other forms of thrombocytopenia (thought to be nonimmune in nature) was the same as in normal controls. However, the platelets in each density subpopulation from ITP patients were increased in size. beta-thromboglobulin (beta TG) content of platelets from each patient group and the normals increased with density and tended to be higher in ITP than in normal controls. beta TG concentration per unit platelet volume and its level in plasma were similar in ITP patients and in normal controls. This suggests that the apparently normal density of ITP platelets was not a result of degranulation of large, dense platelets. Thus platelet size and density are independently determined and the increased size of platelets in immune thrombocytopenia may be the result of abnormalities in their production.

Analysis of immunoglobulins that bind to platelets from serum of patients with immune thrombocytopenia: molecular weight distribution.

The nature of platelet- bindable immunoglobulins (PB-Ig) in serum has been investigated. PB-IgG, -A and -M were measured by an ELISA using platelets coated on microtitre plates. This assay detected alloantibodies at high serum dilutions. In 32 patients with idiopathic thrombocytopenic purpura (ITP) or systemic lupus erythematosus (SLE) raised levels of at least one PB-Ig class were found in 18. To distinguish binding due to immune complexes, the molecular weight of PB-IgG was studied by gel filtration on Sepharose 4B. In sera from patients with ITP and SLE, PB-IgG with Mr of primarily 150 Kd was observed, compatible with monomeric IgG antiplatelet antibodies. Levels of PB-IgG in serum were not related to total serum IgG. In sera from the patients with SLE and some with ITP (most of whom had several of the features of SLE), PB-IgG with Mr of 200 Kd - greater than 1000 Kd was seen. In heat-aggregated preparations of normal IgG, PB-IgG with Mr up to 1000 Kd was also found. Rabbit IgG was able to block PB-IgG in fractions of high molecular weight in purified normal IgG, heat-aggregated normal IgG and in patient serum, but had no effect on the 150 Kd peak. In whole serum from patients who had high molecular weight PB-IgG, the inhibitory effects of rabbit IgG were much less than in isolated high molecular weight column fractions. Thus although the majority of PB-IgG is monomeric antiplatelet antibody, some PB-IgG with higher molecular weight, characteristic of immune complexes, occurs in sera of some patients with autoimmune thrombocytopenia and it makes a small contribution to PB-IgG levels measured in whole serum.

Different platelet specificities of heparin-dependent platelet aggregating factors in heparin-associated immune thrombocytopenia.

Delayed onset heparin-associated thrombocytopenia (HAT) is thought to be a result of formation of antiplatelet antibodies which cause platelet aggregation in the presence of heparin. Platelet aggregation in response to serum from patients with HAT has been studied in platelet-rich plasma (PRP) from a panel of normal blood donors. Heparin-dependent aggregation with any HAT serum occurred in PRP from only some donors. PRP from the non-responding donors did, however, aggregate in the presence of heparin with other HAT sera. The same patterns of aggregation or lack of response to HAT sera were seen in washed platelet suspensions. Heparin (0.06-2 U/ml) did not cause aggregation in the presence of normal serum with PRP from these donors. However, in PRP from four of the 17 individuals studied, heparin (0.25-1 U/ml) alone caused rapid platelet aggregation and some HAT sera heated at 56 degrees C caused platelet aggregation without added heparin. Sub-aggregating concentrations of adrenaline could replace heparin in promoting aggregation by heated HAT sera in PRP of the other donors. HAT IgG showed the same platelet specificities as the serum in causing either heparin- or adrenaline-dependent aggregation. Thus in HAT, antibodies are directed towards different platelet antigens which are expressed differently in different individuals. Platelet activation by heparin and adrenaline either exposes these antigens or causes aggregation of antibody-coated platelets.

Liberation of surface and internal platelet-associated IgG (PA-IgG) during platelet activation.

The relationship between platelet-associated IgG (PA-IgG) of intact platelets to that of lysed platelets has been studied using a competitive ELISA. In platelets from 20 normal individuals mean surface PA-IgG constituted 35% of mean total PA-IgG and showed a significant linear relationship with total PA-IgG. In platelets from 61 patients with various forms of thrombocytopenia including that with immune causes, 38 showed a similar proportion of PA-IgG on the surface after storage at 2 degrees C, as seen in normal platelets, while in 23, levels of surface and total PA-IgG were equal. Normal platelets activated with thrombin or calcium ionophore A23187, also had levels of surface PA-IgG close to total PA-IgG. Supernatants of platelet suspensions activated with aggregating agents contained increased amounts of IgG and when the platelets were washed, both total and surface PA-IgG were decreased. Liberation of IgG from platelets was slower than that of [14C]serotonin but was decreased by release reaction inhibitors. Platelets treated at 2 degrees C with soluble heat-aggregated IgG, which binds to the Fc gamma receptor, showed increased surface PA-IgG, but, after incubation at 37 degrees C, although [14C]serotonin was released, PA-IgG levels were no longer increased. Thus since platelet activation, including that mediated by IgG binding to the Fc gamma receptor, causes PA-IgG release, levels of both surface-located and total PA-IgG may be affected by platelet activation, either in vivo, or in vitro during sample preparation and assay.

Relationship of platelet-associated immunoglobulin G and platelet protein to platelet size and density in normal individuals and patients with thrombocytopenia.

Total platelet-associated immunoglobulin G (PA-IgG) and platelet volume and protein content in normal individuals and in patients with idiopathic thrombocytopenic purpura (ITP) and other thrombocytopenias of presumed nonimmune origin have been measured on platelets separated into subpopulations on the basis of density on continuous polyvinylpyrrolidone (Percoll) gradients. In all subjects PA-IgG per platelet was primarily found in the lightest platelets at levels up to sevenfold greater than in the heavier platelets. PA-IgG level per platelet was raised in light platelets in 29% of patients with thrombocytopenia and in heavy platelets in 60%. In almost half of these instances the PA-IgG level fell to within the normal range when considered in relation to either platelet volume or protein content. PA-IgG levels of patients with untreated ITP did not differ significantly from those with treated ITP or thrombocytopenia of other causes. Mean platelet volume and protein content of the total platelet population of all subjects showed significant linear correlation (P less than 0.01). Thus PA-IgG of both controls and patients with thrombocytopenia of all causes is preferentially located in the lightest platelets, but increases in PA-IgG in immune thrombocytopenias occur more frequently in the heavier platelets. These findings suggest that part of the process of IgG accumulation by platelets is the same in normal individuals as in patients with thrombocytopenia.

Platelets deficient in glycoprotein I have normal Fc receptor expression.

Platelet glycoprotein I (GPI) is known to be required for the interaction of platelets with ristocetin and factor VIII:von Willebrand factor (VIII:vWf). However, its role as Fc receptor is not clear. Some studies have shown that enzymatic removal of GPI destroys the ability of platelets to react with VIII:vWf but not their ability to bind Ig G (IgG). Others have shown that IgG immune complexes which block the Fc receptor also inhibit VIII:vWf interaction with platelets. This subject has been re-examined by testing the ability of platelets with reduced amounts of GPI to aggregate and undergo the release reaction in response to stimuli which act at the platelet Fc receptor. Platelets from two patients with Bernard-Soulier syndrome, congenitally deficient in GPI, both aggregated and released 14C-serotonin normally when exposed to latex particles coated with IgG. Levels of GPI were decreased experimentally in normal platelets by treating them with chymotrypsin. Platelets treated in this manner did not aggregate or release [14C]serotonin in response to ristocetin-VIII:vWf. They did, however, both aggregate and release when incubated with heat-aggregated IgG, antigen-antibody complexes or latex particles coated with IgG. Thus the presence of GPI is not a prerequisite for platelet stimulation via the Fc receptor.