(Received for publication, August 31, 1996)
From the Cancer Research Unit, Faculty of Medicine
and Health Sciences, The University of Newcastle, Callaghan, New
South Wales 2308, Australia and the § School of Chemical
Sciences, Swinburne University of Technology,
Hawthorn, Victoria 3122, Australia
Thrombospondin, a major secretory product of the
-granules of activated platelets, is a large trimeric glycoprotein
that plays an important role in platelet aggregation. On resting
platelets, thrombospondin binds to a single receptor in a
cation-independent manner, but upon platelet activation it binds at
least two further, distinct receptors that are both dependent upon
divalent cations. Each of these receptors on the platelet surface binds
to different regions of the thrombospondin molecule, and such binding
may be responsible for the multifunctional role of thrombospondin in aggregation. We show here that a peptide from the carboxyl terminus of
thrombospondin, RFYVVMWK, directly and specifically induces the
activation and aggregation of washed human platelets from different
donors at concentrations of 5-25 µM. At lower
concentrations the peptide synergizes with suboptimal concentrations of
ADP to induce aggregation. Peptide affinity chromatography and
immunoprecipitation with a monoclonal antibody were used to identify
the receptor for the carboxyl-terminal peptide as the
integrin-associated protein. The integrin-associated protein remained
bound to the RFYVVMWK-containing peptide column when washed with a
scrambled peptide in the presence of 5 mM EDTA, indicating
a divalent cation-independent association. It is suggested that
integrin-associated protein is the primary receptor for thrombospondin
on the surface of resting platelets and is implicated in potentiating
the platelet aggregation response.
Platelet aggregation and clot formation in vivo are
initiated when platelets are activated by soluble physiological
activators such as thrombin, or directly by binding to components of
the subendothelial matrix, such as von Willebrand factor or
fibronectin. This event activates the platelets in a process that
induces platelet shape change and secretion of the contents of the
platelet dense granules and -granules. It also induces a
conformational change in the integrin
IIb
3 (GPIIbIIIa),
enabling this cell surface molecule to function as a receptor for
fibrinogen. Platelet aggregation proceeds when fibrinogen, present both
in the plasma and secreted from the
-granules, is bound by the
IIb
3 receptor in a process that cross-links the platelets
(reviewed in Ref. 1).
It has been recognized for some time that thrombospondin 1 (TSP)1 plays an essential accessory role in
platelet activation and aggregation (2, 3). The concentration of TSP in
plasma is very low (4, 5), but it is the most abundant protein
component of platelet -granules, representing some 25% of the total
protein secreted upon platelet activation, after which a proportion of TSP becomes bound to the surface of the activated platelet (6) and also
becomes incorporated into the fibrin clot (7). TSP has also been
reported to promote platelet aggregation of both non-stimulated
platelets and platelets stimulated with thrombin or ADP (8). In
addition, some anti-TSP antibodies can inhibit platelet aggregation (2,
9, 10). Surface-bound TSP also interacts with fibrinogen bound to the
surface of the platelet via
IIb
3 (11). Leung (2) has suggested
that TSP serves to stabilize fibrinogen binding to the activated
platelet surface reinforcing the strength of interplatelet interactions
and thereby determining the size and reversibility of platelet
aggregates; this in turn serves to regulate clot formation.
The TSP molecule is large and complex and can be depicted as comprising
a series of modules, each exhibiting discrete functions. Proteolytic
digestion of TSP, together with the expression of individual modules as
recombinant proteins or synthetic peptides, has enabled the
identification of several cell-binding domains within the TSP molecule
(reviewed in Refs. 3 and 12). Platelets exhibit at least four known
putative TSP receptors: a proteoglycan that binds predominantly to the
heparin-binding domain in the amino-terminal region of TSP (10); CD36,
which binds to a sequence CSVTCG present in the type 1 repeats (13);
the integrins IIb
3 (GPIIbIIIa) and
v
3, which can bind the
RGD sequence contained within the last of the type III repeats (14,
15); and an unknown receptor that binds to the carboxyl terminus of TSP
(9). In addition, TSP can bind platelet sulfatides (16) and to
fibrinogen (11), interactions which may modulate TSP binding to
platelets. This plethora of receptors for TSP on platelets has
confounded studies seeking to identify the role of individual
receptors, and it is apparent that there exists some redundancy
in the system; for example, platelets deficient in CD36 expression bind
TSP normally (17), as do platelets deficient in
IIb
3 and
fibrinogen binding (18, 19).
Gartner and Dockter (20) found that TSP bound to platelet (and erythrocyte) membranes in the absence of divalent cations, apparently contradicting the work of Phillips et al. (6) who had demonstrated that this association was cation dependent. These findings were resolved in a detailed study by Wolff et al. (21) showing that resting platelets displayed a single class of low capacity, high affinity binding sites for TSP, binding to which was cation independent, whereas thrombin-activated platelets bound TSP with high capacity and lower affinity in a cation-dependent manner. The identity of the receptor for TSP on resting platelets is not known, but it is unlikely to be an integrin or CD36 since binding to these receptors is cation dependent (13, 22). Recently, Kosfeld and Frazier (23) identified peptides from the carboxyl terminus of TSP that were active in cell adhesion, and the same group used cross-linking of iodinated peptides to the surface of K562 erythroleukemia cells to identify a putative cell surface receptor for this region of TSP (24). Significantly, like cell attachment to the peptides, binding of labeled peptide to this protein was cation independent and was stimulated with EDTA (24). Very recently, Gao et al. (25) demonstrated that cells transfected with integrin-associated protein (IAP;CD47) were able to bind the carboxyl-terminal TSP peptide, RFYVVMWK, even in the presence of EDTA to chelate cations. IAP is a ubiquitously expressed protein originally identified and characterized by Brown and colleagues (26, 27, 28). It is expressed on erythrocytes that lack integrins (29), but on most cell types, including platelets, IAP is physically and functionally associated with integrins (26, 27, 28, 29, 30). Antibodies against IAP can inhibit the calcium influx induced in endothelial cells adhering to fibronectin or vitronectin (28) and can induce an oxidative burst in neutrophils (31). Further data to suggest that IAP is a signaling receptor were provided by Tsao and Mousa (32) in showing that the TSP-derived peptide RFYVVMWK triggered a calcium influx in IRM-90 fibroblasts and that this increase was partially blocked by a mAb to IAP.
In this report we demonstrate that the RFYVVMWK peptide stimulates the
aggregation of resting, washed human platelets. An mAb to IAP
synergizes in this aggregation response. Peptide affinity chromatography of platelet lysate reveals a protein that binds the
peptide in the presence of EDTA, and immunoprecipitation shows this
protein to be IAP in association with the integrin receptors IIb
3
and possibly
2
1. It is suggested that IAP is the TSP receptor on
resting platelets and that binding of exogenous TSP by unstimulated
cells could serve to recruit platelets into the developing clot.
Prostaglandin E1, HEPES, leupeptin, phenylmethylsulfonyl fluoride, iodoacetamide, soybean trypsin inhibitor, Triton X-100, and biotinamidocaproate N-hydroxysuccinimide ester (biotin-CNHS) were purchased from Sigma. Acrylamide was from ICN (Costa Mesa, CA). Na125I was purchased from Australian Radioisotopes (Sydney, New South Wales). 32Pi was from Bresatec. Streptavidin-biotin-horseradish peroxidase complexes were purchased from Amersham (Buckinghamshire, United Kingdom). ECL reagents were obtained from DuPont NEN. Iodogen was purchased from Pierce. ADP was from Chrono-Log (Havertown, PA). Thrombin (bovine) was purchased from Armour Pharmaceutical Company (Kankakee, IL). Nitrocellulose was purchased from Sartorious (Goettingen, Germany). Sepharose-2B was purchased from Amrad Pharmacia Biotech (Uppsala, Sweden). Rabbit anti-mouse IgG was from Dako (Carpinteria, CA). SDS-PAGE molecular mass standards, horseradish peroxidase-conjugated anti-species IgG, and Affi-Gel-501 were purchased from Bio-Rad. All other chemicals were purchased from either Ajax Chemicals (Auburn, New South Wales) or BDH (Kilsyth, Victoria) and were of the highest quality available.
Monoclonal AntibodiesMonoclonal antibodies to GPIb-IX
(AK2), 2
1 (AK7), and CD36 (IE8) have been described previously
(33, 34); anti-
v
3 (AP-3) and anti-CD4 (OKT-4) were obtained from
the American Type Culture Collection; and the anti-integrin-associated
protein monoclonal antibody B6H12 (26) was a kind gift of Dr. Eric J. Brown (Washington University School of Medicine, St. Louis, MO).
Peptides were assembled on an Applied Biosystems 431A peptide synthesizer using highly optimized tert-butyloxycarbonyl solid phase synthesis chemistry based on manual in situ neutralization chemistry cycles as described by Schnolzer and Kent (35). Crude peptides were analyzed and purified to >95% by semi-preparative reversed-phase high performance liquid chromatography before final characterization and assessment of purity using capillary electrophoresis. The peptides synthesized were carboxyl-terminal TSP peptide, RFYVVMWK; RFYVVMWKQVTQSYWDTNKKC, which was coupled to Affi-Gel-501 for affinity chromatography studies; scrambled carboxyl-terminal TSP peptide, VFRWKYVM; and heparin-binding domain TSP peptide ELTGAARKGSGRRLVKGPD. RGDS was purchased from Auspep (Parkville, Victoria); KCSVTCG and KRFYVVMWKK were purchased from Chiron Mimotopes (Clayton, Victoria). The carboxyl-terminal peptide RFYVVMWK, its scrambled counterpart, the lysine extended version, and the heparin-binding domain peptide were each solubilized into 40 µl of Me2SO, and the volume made up to 200 µl with water to give final stock concentrations of 5 mg/ml. All other peptides were solubilized into water to the same final concentration.
Platelet IsolationFresh human platelets from consenting adult donors were isolated from 3.13% sodium citrate anticoagulated forearm venous blood (1:9 v/v) collected through a 19 G needle after discarding the initial 10 ml of the transfusion. Whole blood samples were immediately spun at 120 × g for 10 min to obtain platelet-rich plasma. Prostaglandin E1 was added to 1 µM, and platelets were gently pelleted by further centrifugation at 120 × g for 10 min. Finally, platelets were carefully resuspended using a plastic pasteur pipette from the soft pellet into buffer A (138 mM NaCl, 2.9 mM KCl, 0.5 mM MgCl2, 12 mM NaHCO3, 0.3 mM NaH2PO4, 5.5 mM glucose, 10 mM HEPES, pH 7.4).
Platelet AggregationPlatelet aggregation was monitored on a dual-channel platelet aggregometer (Chrono-Log, Havertown, PA). 450- or 250-µl platelet suspensions (1-3 × 108/ml) were treated with thrombin, ADP, or a range of monoclonal antibodies or challenged with defined peptides. Appropriate vehicle blanks were processed in parallel. The platelet aggregation response was monitored under conditions of constant stirring at 1000 rpm at 37 °C.
Isolation of Platelet MembranesPlatelet-rich plasma was isolated as described above from fresh whole blood samples collected into acid citrate dextrose anticoagulant containing fenwals. The pH of the plasma was adjusted to 6.6 with 1 M citric acid as required. Platelets were recovered by centrifugation as described above. Pelleted platelets were washed twice with CGS buffer (120 mM NaCl, 13 mM trisodium citrate, 30 mM glucose, pH 7.0) followed by two washes with HES buffer (10 mM HEPES, pH 7.4, 5 mM EDTA, 150 mM NaCl). Washed platelets were resuspended into HES buffer and layered onto a 0-40% glycerol (in HES buffer) gradient. Samples were spun at 1500 × g for 40 min and then 5900 × g for 10 min at room temperature. The platelet pellet was resuspended into HES buffer containing 10 µg/ml leupeptin, 10 µg/ml soybean trypsin inhibitor, 2 mM phenylmethylsulfonyl fluoride, and 20 mM iodoacetamide. The suspension was sonicated by 4 × 15-s pulses at maximal power using a 6-mm probe (MSE Soniprep 150). The sonicated suspension was centrifuged at 7900 × g for 30 min at 4 °C, and platelet membranes were recovered from the supernatant by further centrifugation at 100,000 × g for 60 min and again at 4 °C. Pelleted membranes were resuspended into HES buffer containing the protease inhibitors detailed above and maintained at 4 °C.
Affinity Column Chromatography of Platelet Membrane SuspensionsFollowing resuspension into HES buffer, platelet membrane preparations were homogenized by 15 strokes in a tight fitting Wheaton siliconized glass homogenizer. This and subsequent procedures were performed at 4 °C. Membrane suspensions were then applied to an RFYVVMWKQVTQSYWDTNKKC-Affi-Gel 501 column equilibrated with HES buffer containing protease inhibitors (all solutions used for affinity chromatography contained protease inhibitors unless stated otherwise). The homogenate was allowed to run into the gel matrix and stand within the column bed for 30 min before washing the column bed with five column volumes of fresh HES buffer. After washing, one column volume of 5 mg/ml of scrambled peptide (VFRWKYVM) solubilized in a minimal volume of Me2SO and made up to volume with HES buffer was run into the column bed. Again, the solution was kept within the column for 30 min before allowing the column volume to void with any eluted protein(s). The column was re-washed with five column volumes of HES buffer. To elute specific surface receptor(s) for the TSP carboxyl-terminal peptide, a 5 mg/ml solution of RFYVVMWK (prepared in the same manner as the scrambled version) was applied to the column and processed as described for the randomly scrambled peptide. In one experiment protease inhibitors were omitted from all column solutions to exaggerate any proteolytic products.
Iodination of Eluted ProteinsIodogen reagent
(1,3,4,6-tetrachloro-3,6
-diphenylglycouril), 5.2 µg, was
solubilized in CHCl3 and coated onto the surface of a glass
tube by evaporation of the vehicle under a stream of N2.
Sample protein from the peptide column was added to the tube followed
by 0.1 mCi of Na125I. The sample was incubated on ice for 5 min, and the reaction was stopped by removal of the sample from the
reaction tube. Samples were made to 1% Triton X-100 and stood on ice
for 60 min prior to immunoprecipitation.
Platelets were recovered as described above and resuspended at 1 × 107 platelets/ml into borate buffer (10 mM sodium borate, pH 8.8, 150 mM NaCl). Biotin-CNHS was added to 50 µg/ml, and the sample was allowed to stand at room temperature for 15 min. The reaction was stopped by the addition of NH4Cl to 10 mM. Labeled cells were washed in buffer B (50 mM Tris-HCl, pH 7.4, 25 mM KCl, 5 mM MgCl2, 1 mM EGTA). Platelets were lysed in 1% Triton X-100 lysis buffer (1% Triton X-100, 150 mM NaCl, 10 mM Tris-HCl, pH 7.4) containing protease inhibitors as described above. Samples stood at 4 °C for 60 min. Insoluble materials were removed from lysates by centrifugation at 15,000 × g for 15 min.
ImmunoprecipitationSoluble platelet lysates and detergent-treated column eluates were precleared with rabbit anti-mouse IgG coupled to Sepharose-4B beads for 2 h at 4 °C. Appropriate antibodies, either coupled directly to Sepharose beads or as free antibody (indirect), were added to the precleared lysates. Sepharose-4B rabbit anti-mouse IgG was simultaneously added to indirect samples. All samples were allowed to mix end-over-end for at least 2 h at 4 °C. Following immunoprecipitation, antigen-bead conjugates were washed with two consecutive cycles of radioimmunoprecipitation assay buffers 1 (25 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Triton X-100, 0.2% sodium deoxycholate, 0.05% SDS) and 2 (25 mM Tris-HCl, pH 7.4, 500 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate). SDS-PAGE sample buffer (62 mM Tris-HCl, pH 6.8, 2.3% SDS, 10% glycerol) was added to the washed antigen-bead samples, and the samples stood at 100 °C for 5 min. Precipitated proteins were analyzed by SDS-PAGE. To detect biotinylated proteins, separated proteins were electrophoretically transferred to nitrocellulose membranes. Streptavidin-biotin-horseradish peroxidase conjugates were used to probe the membranes and protein (s) visualized using ECL.
In Vivo 32P Labeling and Peptide Challenge of PlateletsPlatelets were isolated as described above and
resuspended following recovery from platelet-rich plasma into buffer C
(10 mM HEPES, pH 7.4, 150 mM NaCl, 1 mM EDTA). 32Pi (0.5 mCi) was added
to the suspension, and the cells were allowed to stand at 37 °C for
90 min. Platelets were recovered by centrifugation at 120 × g for 10 min and gently resuspended into buffer A at 3 × 108 platelets/ml. Samples were allowed to stand at 37 °C
for 30 min to recover from these manipulations. Labeled platelets were
treated with either 1 unit/ml thrombin, 45 µM RFYVVMWK
peptide, 45 µM scrambled peptide, or left untreated.
Aggregation with thrombin and RFYVVMWK was allowed to proceed with
mixing until aggregation was maximal as determined by aggregometer
trace. The scrambled sample was treated for the same time period as
used for the unscrambled TSP peptide. All samples were lysed with an
equal volume of ice-cold Triton X-100 lysis buffer (2% Triton X-100,
150 mM NaCl, 20 mM Tris-HCl, pH 7.4, 2 mM EDTA) plus protease inhibitors as described before.
Insoluble material was removed by centrifugation at 15000 × g for
15 min at 4 °C. Equivalent protein concentrations from the soluble
lysate of each sample were analyzed by SDS-PAGE and gels
autoradiographed overnight at 70 °C.
Over a range of concentrations the peptide
RFYVVMWK induced platelet aggregation; the lowest concentration to
induce such aggregation in several experiments was found to be between
11 and 25 µM with peak aggregation commonly occurring at
45 µM concentration, although this varied with platelets
from different donors (data not shown). This peptide did not appear to
be completely soluble after addition of water to the Me2SO
carrier (see "Experimental Procedures"). We therefore prepared a
more water soluble version of the peptide by the addition of charged
(lysine) residues to each end of the peptide. In aggregation assays
this modified peptide (KRFYVVMWKK) gave almost identical results,
except that the lower aggregation threshold was 5 µM
(Fig. 1A). Therefore, in subsequent experiments described below, the RFYVVMWK peptide was used. Aggregation was rapid after addition of either peptide, and the aggregates formed
were large as seen both microscopically (data not shown) and from the
amplitude of the trace (Fig. 1A). A scrambled version of the
peptide (VFRWKYVM) did not cause any platelet aggregation, and it did
not influence aggregation induced by the subsequent addition of the
RFYVVMWK peptide (Fig. 1B). To demonstrate that the peptide
was inducing activation and aggregation of the platelets rather than
agglutination, platelets were loaded with 32Pi
and treated with scrambled peptide (VFRWKYVM) or stimulated with
RFYVVMWK peptide or with thrombin. The platelet lysate was examined for
the appearance of phosphoproteins as a measure of activation (36). As
expected, the RFYVVMWK peptide and thrombin treatments induced the
appearance of a phosphoprotein band at 47 kDa, which was not seen in
platelets treated with the scrambled peptide (Fig. 1C). This
band is likely to be pleckstrin, which is the major protein
phosphorylated upon platelet activation (37).
TSP carboxyl-terminal peptide (K)RFYVVMWK(K) induces platelet aggregation and activation. A, dose-response aggregation of resting platelets induced by the TSP carboxyl-terminal peptide (K)RFYVVMWK(K). Resting, washed human blood platelets were stimulated with a range of concentrations of peptide (5-200 µM), and aggregation was monitored optically using a Chrono-Log aggregometer as described under "Experimental Procedures." Addition of peptide is indicated by winged arrowhead. The data represent a typical response derived from at least three individual donors. B, absence of effect of the randomly scrambled version of the active peptide RFYVVMWK upon resting platelets. Resting platelets were stimulated with either TSP carboxyl-terminal peptide (RFYVVMWK, 45 µM) (trace i) or with scrambled peptide (VFRWKYVM, 45 µM) (trace iii) or pre-treated with scrambled peptide (45 µM) prior to stimulation with TSP carboxyl-terminal peptide (45 µM) (trace ii). Aggregation was monitored as in A. Winged arrowheads indicate addition of carboxyl-terminal peptide; triangular arrowheads indicate addition of scrambled peptide. C, activation of platelets by TSP carboxyl-terminal peptide RFYVVMWK demonstrated by the phosphorylation of p47 protein. Resting platelets were loaded with 32Pi and then stimulated with 45 µM RFYVVMWK, 45 µM VFRWKYVM (scrambled), 1.0 unit/ml thrombin, or left untreated (control). Solubilized lysates from each treatment regime were separated by SDS-PAGE, and the dried gel was autoradiographed to visualize labeled proteins. Phosphorylated p47 protein is arrowed. Molecular mass markers are shown at left in kDa.
Effects of Cell Binding Peptides from Other Domains of Thrombospondin upon RFYVVMWK-mediated Platelet Aggregation
Gartner et al. (10) reported that a
polyclonal antibody to the heparin-binding domain of thrombospondin
inhibited platelet aggregation induced by ADP, collagen, or thrombin. A
peptide from this region of thrombospondin that binds to the cellular
proteoglycan receptor in endothelial cells was recently identified
(38). In addition, Tuszynski and colleagues (39) reported that the CSVTCG peptide from the type 1 repeats of thrombospondin, which binds
CD36 (13), inhibited ADPinduced platelet aggregation in a
dose-dependent manner. We therefore tested these peptides
for their effect upon RFYVVMWK-induced aggregation of platelets.
Neither the heparin-binding domain peptide (ELTGAARKGSGRRLVKGPD) nor
the CSVTCG peptide themselves induced any platelet aggregation when tested over a range of concentrations (data not shown). The
heparin-binding domain peptide however inhibited the extent of platelet
aggregation induced by RFYVVMWK in a dose-dependent manner
(Fig. 2A), although it did not alter the
shape change or initial peptide-induced aggregation (see below). In
contrast, pre-treatment of the platelets with CSVTCG peptide prior
to stimulation with RFYVVMWK peptide had no effect on the aggregation
response (Fig. 2B).
Effect of cell-binding peptides from TSP on the platelet aggregation response induced by TSP carboxyl-terminal peptide RFYVVMWK. A, effect of heparin-binding domain peptide (ELTGAARKGSGRRLVKGPD) on the aggregation of resting platelets induced by the TSP carboxyl-terminal peptide RFYVVMWK. Resting washed platelets were stimulated with carboxyl-terminal peptide alone (trace i, 45 µM) or exposed to heparin-binding domain peptide (trace ii, 100 µM; trace iii, 200 µM) for 60 s prior to stimulation with carboxyl-terminal peptide (45 µM). Aggregation was monitored optically using a Chrono-Log aggregometer as described under "Experimental Procedures." Winged arrowheads indicate addition of RFYVVMWK; triangular arrowheads indicate addition of heparin-binding domain peptide. The data represent a typical response derived from at least three individual donors. B, effect of CD36-binding peptide KCSVTCG on platelet aggregation induced by RFYVVMWK. Resting washed platelets were stimulated with KCSVTCG alone (100 µM) (trace ii) or KCSVTCG (100 µM) 60 s prior to addition of RFYVVMWK (45 µM) (trace i). Aggregation response was monitored as described in A. Winged arrowheads indicate addition of RFYVVMWK; triangular arrowheads indicate addition of KCSVTCG peptide. C, effect of RGDS peptide on the aggregation response of resting platelets induced by the TSP carboxyl-terminal peptide RFYVVMWK. Resting washed platelets were stimulated with carboxyl-terminal peptide alone (trace i, 45 µM) or treated with increasing concentrations of RGDS peptide (trace ii, 0.3 mM; trace iii, 0.6 mM; trace iv, 1.2 mM) 60 s prior to stimulation with RFYVVMWK peptide (45 µM). Aggregation was monitored as in A. Winged arrowheads indicate addition of RFYVVMWK peptide; triangular arrowheads indicate addition of RGDS. The data represent a typical response derived from three individual donors.
Contained within the last type III repeat of thrombospondin is the RGD
motif that serves as a recognition sequence for a number of integrins
including IIb
3 (GPIIbIIIa) of platelets and is well characterized
as inhibiting platelet aggregation induced by a number of agonists
(13). As expected, upon testing in the RFYVVMWK-induced platelet
aggregation assay, the RGDS peptide inhibited aggregation in a
dose-dependent manner (Fig. 2C). This result
indicates direct involvement of the
IIb
3 integrin in the
peptide-induced aggregation response.
The decrease in the extent of
aggregation caused by the heparin-binding domain peptide (Fig.
2A) is unlikely to be caused by any antagonistic effect on
the RFYVVMWK peptide but is most likely the result of inhibition of the
platelet cross-linking caused by secreted endogenous platelet
thrombospondin that results in stabilization of formed platelet
aggregates as propounded by Leung (2). Taken together, these data may
account for reports demonstrating that a monoclonal antibody against
the carboxyl region of thrombospondin (9), and also a polyclonal
antibody to the amino-terminal heparin-binding domain (10) of
thrombospondin, can both inhibit platelet aggregation. From our data
based on the peptide from the carboxyl-terminal region of
thrombospondin, it appears that this region is involved in an
activation step rather than in the thrombospondin-mediated
stabilization of platelet aggregation (2). Since the plasma levels of
TSP are low (4, 5), the major source of TSP accessible to resting
platelets is likely to be that secreted from the -granule store of
adjacent, activated platelets or other cells (40, 41). Such secretion from activated platelets would be accompanied by the release of ADP,
itself a weak platelet agonist (42). If the carboxyl terminus of TSP
indeed plays a role in initiating platelet aggregation, it might
therefore be expected to synergize with ADP. To test this concept,
platelets were treated with suboptimal doses of the weak agonist ADP or
with suboptimal concentrations of the RFYVVMWK peptide, or with both
together. As shown in Fig. 3A, the peptide
and ADP synergized to induce rapid and complete platelet aggregation.
No aggregation was seen with suboptimal concentrations of ADP together
with the scrambled peptide (Fig. 3A).
Synergistic enhancement of platelet
aggregation caused by combining suboptimal concentrations of TSP
carboxyl-terminal peptide RFYVVMWK and ADP or anti-IAP mAb B6H12.
A, suboptimal doses of ADP synergize with suboptimal
concentrations of carboxyl-terminal peptide RFYVVMWK. Resting washed
platelets were treated with either a non-aggregating concentration of
carboxyl-terminal peptide alone (trace i, 20 µM), scrambled carboxyl-terminal peptide alone
(trace ii, 45 µM), treated with suboptimal ADP
(20 µM) prior to addition of scrambled peptide (45 µM) (trace iii), or treated with suboptimal ADP (20 µM) followed by suboptimal RFYVVMWK peptide (20 µM) (trace iv). In each case aggregation
response was monitiored as described in Fig. 1. Winged
arrowheads indicate addition of carboxyl-terminal peptide
RFYVVMWK, triangular arrowheads indicate addition of
scrambled peptide, and asterisked arrowheads indicate
addition of ADP. These data are representative of at least three
separate assays for each combination, and the non-aggregating
concentration of carboxyl-terminal peptide used for platelets from
different donors varied between 10 and 20 µM.
B, anti-IAP mAb B6H12 synergizes with suboptimal doses of
RFYVVMWK peptide. Resting washed platelets were exposed to a suboptimal
dose of TSP carboxyl-terminal peptide, RFYVVMWK (15 µM)
(trace i), mAb anti-2
1 (10 µg/ml, AK7) (trace
ii), anti-
2
1 (10 µg/ml) prior to 15 µM
RFYVVMWK peptide (trace iii), anti-IAP (10 µg/ml, B6H12)
(trace iv), or suboptimal RFYVVMWK peptide (15 µM) followed by anti-IAP (B6H12) (10 µg/ml)
(trace v). Aggregation response was monitored as described
in Fig. 1. Winged arrowheads indicate addition of RFYVVMWK
peptide; triangular arrowheads indicate addition of antibody. These data are
representative of at least three experiments with different platelets
for each combination.
RFYVVMWK Peptide Binds to IAP from Platelets
Gao et
al. (25) have recently identified IAP as a cellular receptor for
the carboxyl-terminal domain of thrombospondin. Since Brown and
associates (26) have produced antibodies to IAP that can influence
cellular function, we tested one of these antibodies for its effect on
platelet aggregation. The antibody, B6H12, has been reported as a
functional antibody (for example, endothelial cell migration is blocked
(26, 27, 28, 29, 30, 31)). When tested for its effects upon platelets, this antibody
induced direct platelet aggregation of the platelets from some subjects
(Fig. 3B) but not others. Monoclonal antibodies directed
against other platelet glycoproteins that are able to induce
aggregation (anti-CD9 (43), anti-CD36 (19), and anti-PTA1 (34)) all
require involvement of the Fc receptor (44). Because of a polymorphism
of the human Fc receptor, FcRII, only platelets from some
individuals (about 50%) can bind the Fc of mouse monoclonal antibodies, and only platelets from these "responder" subjects are
activated by the antibodies; "non-responder" platelets bind antibody only through antigen recognition and are not induced to
aggregate (45). We therefore tested platelets from a number of subjects
known to be "responders" or "non-responders" to the anti-PTA1
antibody (34, 44). The B6H12 antibody to IAP aggregated platelets only
from the responder group suggesting that, as with other activating
antibodies, Fc receptor engagement is required for this response (data
not shown). Also in common with other platelet-activating antibodies
(19, 43, 34), B6H12 antibody-induced platelet aggregation was preceded
by a prolonged lag phase (Fig. 3B), the length of which
correlated with the concentration of antibody used (data not shown).
This lag phase was greatly shortened when the platelets were treated
with suboptimal concentrations of ADP (data not shown). In addition,
the B6H12 antibody also synergized with suboptimal concentrations of
the RFYVVMWK peptide to induce platelet aggregation (Fig.
3B).
These results suggest that IAP may function as a platelet receptor for
this carboxyl-terminal peptide from thrombospondin. To confirm that
platelets express IAP on their surface, immunoprecipitation experiments
were carried out from surface-labeled platelets. As shown in Fig.
4A, a band in the position of IAP was
immunoprecipitated specifically with the B6H12 antibody. IAP on the
cell surface labels very poorly (26), and without carrying out binding
analysis it cannot be stated whether or not the faint nature of this
band is the result of low copy numbers of IAP receptors on the cell surface. We then sought to determine whether the RFYVVMWK peptide bound
IAP. Gao and Frazier (24) had reported that in a cross-linking study
the RFYVVMWK peptide bound to a 52-kDa protein when cross-linked to the
surface of K562 cells, but binding was substantially reduced after the
cells had been lysed in Triton X-100-containing buffers. We attempted
to affinity purify RFYVVMWK-binding proteins using detergent-lysed
platelet membranes and obtained no specific binding proteins (data not
shown). This result substantiates the suggestion of Gao and Frazier
(24) that the receptor for this peptide requires a specific membrane
conformation to bind its ligand and that this is disrupted by Triton
X-100. We therefore prepared disrupted platelet membranes in the
absence of detergent (see "Experimental Procedures"). These were
passed over an affinity column of an RFYVVMWK-containing peptide, which
was then extensively washed with buffer containing EDTA, and bound
materials subsequently eluted with the scrambled peptide (VFRWKYVM) in
the continued presence of EDTA followed by specific elution with
RFYVVMWK peptide. Material eluted at each step was concentrated and
analyzed by SDS-PAGE and silver staining (Fig. 4B). The only
bands seen from the specific eluate were proteins of ~45 and 35 kDa
(Fig. 4B). In different experiments the proportions of the
45- and 35-kDa proteins varied, and in the absence of protease
inhibitors the 35-kDa protein became more prominent (Fig.
4B). These results are in good agreement with the RFYVVMWK
peptide cross-linking data of Gao and Frazier (24) who suggest that the
35-kDa band represented a major degradation product of the 52-kDa
protein from K562 cells. Next, we sought to identify the 45-kDa band by immunoprecipitation. The peptide affinity chromatography experiments were repeated, and the proteins eluted by the scrambled peptide and the
specific peptide were detergent treated to disrupt the membranes and
then labeled with 125I. The labeled proteins were then
immunoprecipitated with antibodies to the platelet membrane receptors:
IAP, the integrin 3 subunit of
IIb
3, the
subunit of the
integrin
2
1, the Ib subunit of GPIb-IX, CD36, or with an antibody
irrelevant to platelets (CD4). IAP protein was seen to be precipitated
from the specific eluate but not from the scrambled eluate (Fig.
4C). This same band appeared also to be coprecipitated by
the antibody against
3, an expected result since Brown and
colleagues have shown previously that IAP from platelets associates
with the integrin
3 subunit (26) and also that IAP on K562 cells
cross-linked to the RFYVVMWK peptide is coprecipitated with antibodies
to
3 (24). The IAP band was not precipitated with antibodies to CD36
or CD4, but a trace was seen in the GPIb-IX precipitant. A band
apparently co-incident with IAP was also seen in the
2
1
precipitant (Fig. 4C). Brown et al. (26) have
reported that IAP associates with integrins other than
3, and it may
be that in platelets this receptor preferentially associates with
2
1. However, the relatively strong labeling of this band in the
anti-
2
1 immunoprecipitant compared with the IAP immunoprecipitant
together with its slightly less diffuse appearance does not allow this
conclusion without further analysis.
IAP of platelets binds to the
carboxyl-terminal peptide RFYVVMWK. A, integrin-associated
protein is expressed on the surface of resting platelets. Proteins on
the surface of resting platelets were labeled with biotin, the
platelets were lysed, and immunoprecipitation was performed on the
soluble lysate with antibodies to IAP (B6H12), 3 (AP-3), CD36 (IE8),
and a platelet irrelevent antibody (OKT-4). Precipitated proteins were
separated by SDS-PAGE, transferred to nitrocellulose, and biotinylated
protein(s) detected by ECL. Molecular mass markers in kDa are shown at
left. B, silver stain of SDS-PAGE analysis of
protein(s) from platelet membrane suspensions bound to an
RFYVVMWK-containing peptide affinity column in the presence of EDTA.
Resting platelet membrane protein(s) eluted in EDTA-containing
buffer from the peptide affinity column in the absence of protease
inhibitors were analyzed by SDS-PAGE and silver staining.
Homogenate, total material loaded to column; Scrambled (VFRWKYVM), proteins eluted by the scrambled
peptide (5 mg/ml); column wash, post-scrambled elution;
RFYVVMWK, proteins eluted by the TSP carboxyl-terminal
peptide (5 mg/ml). Proteins eluted by the carboxyl-terminal peptide are
arrowed at right. Molecular weight markers are
shown at left in kDa. C, immunoprecipitation identifies integrin-associated protein as binding the TSP
carboxyl-terminal peptide RFYVVMWK. Proteins eluted by the scrambled
and specific peptides with buffer containing EDTA and protease
inhibitors were iodinated and immunoprecipitated with a range of
antibodies to platelet surface proteins: IAP (B6H12),
3 (AP-3),
2
1 (AK-7), GPIb-IX (AK-2), CD4 (OKT-4), and CD36 (IE8).
Precipitated material was analyzed by SDS-PAGE, and labeled protein was
detected by autoradiography. Molecular mass markers are shown at
left in kDa.
Further work will be required to place these results in a physiological
context. The platelets used in our study were washed since the peptide
did not induce aggregation of platelets contained in plasma.
Preliminary studies have shown that this is the result of the peptide
binding to an unidentified component of plasma (also present in murine
ascites fluid) rather than the status of the platelets; by every
criterion measured the washed platelets used for the aggregation
studies were resting.2 As such, the data
suggest that the IAP receptor is functionally expressed on the surface
of resting platelets. Extrapolating from this, it is likely that IAP is
the cation-independent receptor for TSP on resting platelets identified
by Wolff et al. (21). These authors carried out TSP binding
studies to demonstrate that resting platelets bound 3100 ± 1000 TSP molecules/platelet with a Kd of 50 nM. This binding was not disrupted by the presence of 5 mM EDTA, and the same group went on to establish that TSP
binding to resting platelets was not influenced by IIb
3 (GPIIbIIIa) or fibrinogen binding (18). Our affinity chromatography was
carried out in 5 mM EDTA, confirming that the binding of
RFYVVMWK peptide to IAP is not inhibited by the presence of EDTA (24), and it will be instructive to measure the numbers of IAP molecules expressed by resting platelets; although Brown and associates (26) have
shown that platelets bind less anti-IAP antibody than erythrocytes,
which, in a later study (44), were demonstrated to display some 10,000 antibody binding sites per erythrocyte. The relatively high
concentrations of peptide required to induce platelet activation is
unlikely to be a measure of the true binding affinity of TSP for IAP
since the peptide is removed from its natural conformation and the
trimeric nature of TSP could alter the avidity of binding. Based
primarily upon blocking studies with anti-TSP and anti-receptor
antibodies, it has been recognized for some time that TSP plays an
important role in platelet aggregation (2, 3, 8, 9, 10). While there has
been some disagreement about the relative importance of the different
receptors for TSP identified on platelets, it is generally conceded
that TSP is involved in determining the size and reversibility of
platelet aggregates (2). As evidenced by TSP binding to thrombasthenic platelets, TSP binding in the absence of a functional
IIb
3
receptor is insufficient to promote aggregation (18). However, the
intriguing experiments described by Aiken et al. (19)
suggest that TSP alone may support the aggregation of stimulated normal
platelets in the virtual absence of platelet binding to other adhesive
proteins, including fibrinogen. Our results with a carboxyl-terminal
peptide from TSP substantiate and extend the observations made by
Tuszynski et al. (8), that TSP is also able to potentiate
the reversible first phase of ADP-induced platelet aggregation and to
drive the platelets through to the irreversible second phase. This may
be of significant biological consequence. Platelets binding to
immobilized ligands such as fibronectin or collagen spread out and
become activated in the absence of soluble agonists (45). These ligands would be well expressed upon damage to endothelial cells; in this event
TSP secreted by the substrate-bound, activated platelets could play a
substantive role in recruiting resting platelets into the forming
clot.
We thank the Red Cross Blood Transfusion Service (Newcastle) and Hampson Pathology for provision of blood samples, the numerous individuals who generously donated the blood that was used for this study, and Dr. Eric J. Brown for providing the anti-IAP antibody and for critically reviewing the manuscript.