Ultrastructural localization of human platelet thrombospondin, fibrinogen, fibronectin, and von Willebrand factor in frozen thin section

We have investigated the localization of thrombospondin (TSP), fibrinogen, fibronectin, and von Willebrand factor in human platelets by transmission electron microscopy of antibody-stained ultrathin frozen sections. In negatively stained thin sections, alpha granules were identified on the basis of their smooth, roughly spherical shape, size, single limiting electron-lucent 100 A membrane, and frequent presence of electron-dense nucleoid. In contrast, mitochondria exhibited characteristic double membranes and cristae. Sections were separately stained with affinity-purified polyclonal antibodies to these proteins as well as with three monoclonal anti-TSP antibodies. Antibody specificity was documented in radioimmunoassays, by immunofluorescent cross-blocking, and by staining of bands of appropriate mobility in Western blots of whole platelets. Bound antibody was visualized using a 5-nm colloidal gold-avidin conjugate. In resting cells, staining of virtually all alpha granules was observed for all four proteins. In contrast, consistent staining was absent from other organelles, including plasma membranes, mitochondria, and vacuolar structures that may represent the open canalicular system.     

Arg-Gly-Asp-dependent occupancy of GPIIb/IIIa by applaggin: evidence for internalization and cycling of a platelet integrin

Using indirect immunofluorescence microscopy we examined the distribution and cycling of GPIIb/IIIa after binding to applaggin, a high-affinity Arg-Gly-Asp (RGD)--containing ligand. Resting, unfixed platelets were incubated with applaggin for 30 minutes at 37 degrees C, and bound applaggin was detected by an affinity-purified rabbit anti- applaggin antibody. Examination of intact cells showed a rim pattern for applaggin, consistent with its binding to the platelet surface. Staining of Triton X-100--permeabilized cells showed an intracellular pool of applaggin. Competition of applaggin binding by either AP-2, an anti-GPIIb/IIIa monoclonal antibody (MoAb) that blocks fibrinogen binding, or the synthetic peptide RGDW eliminated both surface and intracellular staining, indicating that applaggin is binding to GPIIb/IIIa in an RGD-dependent manner. Inhibition of platelet activation by PGE1 and theophylline had no effect on the observed staining patterns, indicating that cellular activation is not required for surface binding and subsequent internalization. To evaluate whether occupancy of functional binding sites on GPIIb/IIIa is required for internalization, we used mAb15, an anti-GPIIIa antibody that neither blocks fibrinogen binding nor induces the expression of ligand-induced binding sites on GPIIb/IIIa. In these studies mAb15 was internalized in a manner analogous to both AP-2 and applaggin, showing that occupancy of the RGD binding site is not required to initiate receptor internalization. To estimate the size of the newly internalized pool of applaggin, 125I-applaggin--binding studies were performed. Displacement of bound 125I-applaggin by excess unlabeled applaggin or EDTA showed that at least 17% of bound applaggin was nondisplaceable when binding was performed under conditions permitting membrane flow and internalization. These data indicate that GPIIb/IIIa is internalized in unstimulated platelets independent of cellular activation or occupancy of the functional binding site(s) of GPIIb/IIIa by RGD-containing ligands. Thus, internalization of GPIIb/IIIa may represent a mechanism by which the surface expression of this adhesion receptor is regulated.     

Ultrastructural localization of coagulation factor V in human platelets

The distribution and transport in platelets of human coagulation Factor V was investigated by immunofluorescent and immunoelectron microscopy. In resting intact platelets, little surface staining was observed by immunofluorescence. In permeable resting cells, punctate staining similar to that reported for fibrinogen (Fbg), thrombospondin (TSP), fibronectin (Fn), von Willebrand factor (VWF), B-thromboglobulin (BTG), and platelet Factor 4 (PF4) was observed. Double label immunofluorescent staining for Fbg and Factor V demonstrated colocalization, suggesting their presence in the same intracellular structure. Thrombin stimulation induced the appearance of larger (approximately 0.5 mu) immunofluorescent masses of these proteins which exactly colocalized. Thus, at the light level, Factor V and Fbg are localized in the same structure in resting and thrombin-stimulated cells. On the ultrastructural level, an alpha granule localization for Fbg has previously been established. We have extended our immunofluorescent observations regarding the localization of Factor V in human platelets by use of transmission electron microscopy of antibody-stained ultrathin frozen sections. In resting cells, staining of virtually all alpha granules was observed for Factor V. In contrast, consistent staining was absent from other organelles including plasma membranes, mitochondria, and vacuolar structures which may represent the open canalicular system. These data thus establish at the ultrastructural level an alpha granule localization of human coagulation Factor V.     

Localization of internal pools of membrane glycoproteins involved in platelet adhesive responses.

To investigate the existence of intracellular pools of membrane glycoproteins involved in platelet adhesive reactions, the authors have studied the distribution of glycoprotein (GP) Ib and IIb/IIIa by immunofluorescence and immunoelectron microscopy. Studies on whole cells and frozen thick sections revealed a rim pattern of fluorescence for GPIb and GPIIb/IIIa consistent with a surface distribution. In addition, extensive staining occupying the entire cell interior was observed for anti-GPIIb/IIIa, whereas anti-GPIb revealed staining of large intracellular structures that contained no stainable fibrinogen. On the ultrastructural level, the extracellular face of the plasma membrane and the intraluminal face of vacuolar structures were stained with both anti-GPIb and anti-GPIIb/IIIa. Additionally, GPIIb/IIIa antigen was localized to alpha-granule membranes. To determine whether alpha-granule GPIIb/IIIa could be transported to the cell surface, the authors employed a calcium-dependent monoclonal anti-GPIIb/IIIa antibody. Incubation of platelets with EGTA at 37 C abolished staining of plasma membrane and vacuolar but not alpha-granule GPIIb/IIIa. Recalcification of these cells failed to restore the epitope; however, thrombin treatment of recalcified cells reconstituted surface staining with a concurrent loss of internal staining. These data suggest that GPIIb/IIIa is present in alpha-granule membranes and may be transported to the cell surface in response to thrombin treatment. In addition, both GPIb and GPIIb/IIIa antigens are present in intracellular membrane-bounded vacuolar structures which are closed to antibody probes in fixed cells. Redistribution of these internal pools of adhesive protein "receptors" may participate in the regulation of platelet adhesive properties.     

Internalization of bound fibrinogen modulates platelet aggregation

In agonist-stimulated platelets, the integrin alpha IIb beta 3 (glycoprotein IIb-IIIa) is converted from an inactive to an active fibrinogen receptor, thereby mediating platelet aggregation. With time after agonist addition, at least two events occur: fibrinogen becomes irreversibly bound to the platelet and, when stirring is delayed, platelets lose the ability to aggregate despite the presence of maximally bound fibrinogen. Because we previously identified an actively internalized pool of alpha IIb, beta 3 in platelets, we explored the possibility that both of these events might result from the internalization of fibrinogen bound to active alpha IIb beta 3. Under conditions of irreversible fibrinogen binding, fluorescence microscopy showed that biotinylated fibrinogen is rapidly internalized by activated platelets to a surface-inaccessible, intracellular pool. Flow cytometric analysis showed that the observed loss in accessibility to extracellular probes immediately precedes a loss in ability to the platelets to aggregate. Moreover, prevention of irreversible fibrinogen binding results in a prevention of internalization and a retention of aggregation capacity. Thus, the internalization of fibrinogen from the activated platelet surface appears to contribute not only to the irreversible phase of fibrinogen binding, but also to the downregulation of platelet adhesiveness. Fibrinogen internalization is therefore likely to represent a fundamental regulatory mechanism that modulates platelet function.     

A platelet alpha granule membrane protein that is associated with the plasma membrane after activation. Characterization and subcellular localization of platelet activation-dependent granule-external membrane protein.

We have identified and purified a platelet integral membrane protein (140,000 mol wt), using the KC4 monoclonal antibody specific for activated platelets, that is internal in resting platelets but exposed on activated platelets (Hsu-Lin S.-C., C.L. Berman, B.C. Furie, D. August, and B. Furie, 1984, J. Biol. Chem. 259: 9121-9126.). The expression of the protein on the platelet surface is secretion-dependent. This protein has been named platelet activation-dependent granule-external membrane (PADGEM) protein. PADGEM protein is distinct from the surface glycoproteins of resting platelets, but identical to the S12 antigen, GMP-140. Using immunofluorescent staining, resting platelets failed to stain for PADGEM protein with the KC4 antibody, but after permeabilization showed a punctate staining of the cell interior. Thrombin-stimulated intact platelets stained with a peripheral rim pattern thus demonstrating the translocation of PADGEM protein from an internal location to the cell surface. PADGEM protein expression on the platelet surface at varying thrombin concentrations correlated with alpha granule release, as measured by the secretion of platelet factor 4. Further evidence for an alpha granule localization of PADGEM protein was provided by nitrogen cavitation of resting platelets followed by metrizamide density gradient centrifugation; PADGEM protein codistributed with platelet factor 4. Using immunoelectron microscopy, the protein was localized to the alpha granule in frozen ultrathin sections of resting platelets labeled using rabbit anti-PADGEM protein antibodies, whereas in thrombin-activated platelets, the plasma membrane was labeled. These studies indicate that PADGEM protein is a component of the alpha granule membrane of resting platelets and is incorporated into the plasma membrane upon activation and secretion.     

ADP induced depolarization of human platelet membrane potential

The transmembrane potential of human blood platelets suspended in plasma was investigated by studying the distribution of a radiolabeled permeant ion [14C] thiocyanate. The membrane potential of resting platelets was found to be -54.50 mV +/- 9.23 S.D. with a range of -39 to -76 mV (n = 27). The possibility that platelet activation alters membrane potential or that changes in membrane potential serve as an activation trigger was investigated. Stimulation by ADP (10 microM) resulted in a significant (p less than 0.05) depolarization of the membrane potential. Preincubation with 6 mM EGTA failed to inhibit ADP-induced depolarization even though EGTA effectively prevented primary and secondary aggregation but not shape change. Preincubation with PGE1 inhibited shape change, aggregation, and the ADP-induced depolarization. No significant change in membrane potential was observed following stimulation by epinephrine (50 microM). These results suggest that the initial interaction of ADP and its receptor may involve an inward positive current which can be determined by thiocyanate distribution.     

Plasma membrane GPIIb/IIIa. Evidence for a cycling receptor pool.

The author used immunofluorescence and digital image processing to investigate the dynamic distribution of GPIIb/IIIa in living platelets. Resting cells were incubated with AP-2, a complex-specific, monoclonal, anti-GPIIb/IIIa antibody. Examination of intact cells demonstrated a rim pattern for GPIIb/IIIa consistent with a surface localization. Permeabilization revealed a time-dependent increase in the labeling of apparent intracellular vacuoles. This pattern is distinct from the "patch-cap" pattern observed when unfixed platelets were incubated with fluoresceinated concanavalin A. Additionally, labeling of this vacuolar pool of GPIIb/IIIa was inhibited by treatment with 2% sodium azide or by incubation at 4 degrees C. Identical staining patterns were obtained with Fab fragments of AP-2. Ultrastructural examination confirmed the presence of labeled intracellular vacuolar structures. Parallel studies performed with AP-1, a monoclonal anti-GPIb antibody, failed to demonstrate internalization of GPIb. Finally, thrombin stimulation of resting platelets, which had been preincubated with AP-2, resulted in the clearing of this newly internalized pool of GPIIb/IIIa; presumably via translocation to the surface. These data suggest the presence of an actively cycling pool of GPIIb/IIIa that has not been described previously. The dynamic distribution of this pool may be important in the regulation of platelet adhesiveness.     

Immunolocalization of beta 1 integrins in platelets and platelet- derived microvesicles

Human platelets contain several adhesion receptors belonging to the integrin superfamily. At least three beta 1 integrins are present on platelets and have been shown to mediate platelet adhesion to collagen, fibronectin, and laminin. To study the cellular localization of the beta 1 integrins in platelets, we produced a polyclonal antibody by immunization of goat 172 with purified beta 1 subunit from HPB-ALL cells. Antibody 172 (Ab172) specifically immunoblotted a 135-Kd protein in a lysate of whole platelets. The reactivity of Ab172 with platelet membrane proteins was further determined by immunoprecipitation of lysates of surface-radioiodinated platelets. Ab172 immunoprecipitates, resolved by nonreducing/reducing two-dimensional sodium dodecyl sulfate- polyacrylamide gel electrophoresis consisted of three labeled proteins with migrational properties of platelet glycoprotein (GP)Ia, GPIc and GPIIa. Neither GPIIb/IIIa nor the vitronectin receptor were immunoprecipitated by Ab172, confirming a lack of cross-reactivity with the beta 3 integrins in platelets. Immunofluorescence studies using Ab172 were performed to investigate the cellular distribution of beta 1 integrins in platelets. Fluorescent labeling of intact cells demonstrated the presence of beta 1 antigen on the surface of resting cells. Permeabilization of platelets with Triton X-100 showed the presence of an intracellular pool of beta 1 antigen. Double-label experiments using Ab172 and AP-2 (anti-GPIIb/IIIa) showed identical labeling patterns, suggesting a similar subcellular distribution for these integrins. Following thrombin stimulation, permeabilized cells showed a centralized clearing of both beta 1 antigen and GPIIb/IIIa as well as an intensification of surface labeling for beta 1 antigen. These findings suggest the translocation of intracellular beta 1 antigen to the platelet surface as a result of thrombin stimulation. Because platelet-derived microvesicles have been reported to contain GPIIb/IIIa, we investigated the possible distribution of beta 1 integrins in these structures. Microvesicles, produced as a result of platelet activation, were labeled with Ab172, suggesting the distribution of beta 1 integrins in these structures as well as in intact cells.     

Identification and quantitation of protein S in human platelets

Gel filtered human platelets contaminated with less than 0.02% of plasma protein S contained 490 ng of protein S antigen per 3 X 10(8) platelets, equivalent to 2.5% of protein S in whole blood. Three patients with heterozygous plasma protein S deficiency, a congenital disorder associated with venous thrombotic disease, had platelet protein S antigen levels that were 40% of the mean platelet level in ten normal volunteers. In immunoblotting analysis, platelet protein S was indistinguishable from plasma protein S. Thrombin stimulation of platelets caused release of 63% of total protein S antigen and this release was abolished when platelets were preincubated with metabolic inhibitors. Thrombin effected limited proteolysis of platelet protein S and this reaction was inhibited by calcium ions. Immunofluorescent staining of platelets using protein S antibodies demonstrated that protein S colocalized with fibrinogen, an established alpha-granule protein. Thus, human platelets contain protein S in alpha granules that can be released by thrombin stimulation. The released protein S may bind to stimulated platelets and thereby promote and localize the anticoagulant activity of activated protein C on the platelet surface.