MODIFICATION OF PLATELET ACTION BY FIBRINOGEN- COATED PARTICLES BACKGROUND OF THE INVENTION Platelets play a critical role in the maintenance of hemostasis in humans and other animals. An early event in the control of bleeding is the activation of platelets at a site of vascular damage. This activation has several consequences, for example, secretion of chemicals (such as ADP and thromboxin) that cause activation of additional platelets, and binding of the protein fibrinogen to the surface of activated platelets. The bound fibrinogen mediates aggregation and cross-linking of platelets at the wound site (i. e., formation of the platelet"plug"). In addition, the bound fibrinogen is modified by proteolytic action of thrombin to form fibrin monomers, which rapidly polymerize to form a reticulum of fibrin threads. A platelet fibrin clot develops when additional activated platelets (as well as other cells) are caught in, and bind, the fibrin reticulum. These in vivo processes can be modeled in vitro. For example, platelets can be activated in vitro by the addition of ADP in the presence or absence of fibrinogen. Similarly, binding by activated platelets of "polymerizing fibrin"is the in vitro analogue of the association of platelets and the fibrin reticulum in vivo.

Binding of fibrinogen and fibrin to platelets occurs via the allbß3 receptor.

This binding results in changes in platelet metabolism, particularly in levels of polyphosphoinositides. Consistent with their different roles in clot formation, the metabolic changes caused by binding of polymerizing fibrin to platelets are different from those caused by fibrinogen binding (Vickers et al., 1990, Platelets 1 : 199; Vickers et al., 1986, Biochem.

J. 237: 327). In either case, however, the changes are likely involved in signaling pathways that link the binding of the ligands and internal changes in the platelets, particularly remodeling of the actin cytoskeleton.

SUMMARY OF THE INVENTION Because of the critical role of clotting in hemostasis, compositions that affect platelet binding to fibrin and/or fibrinogen (e. g., through effects on allbß3 receptor conformation) may have therapeutic utility. In addition, an understanding of the interaction via the allbß3 receptor of platelets to immobilized fibrinogen or similar polypeptides provides a means for making agents with therapeutic utility, such as artificial platelets. The present disclosure provides these and other agents and methods.

The present invention is related to the discovery that binding of fibrinogen- coated protein microspheres to the fibrin (ogen) receptor allbß3 on platelets causes an "outside-in"signal that decreases phosphatidylinositol 4,5-biphosphate (PIP2) levels, similar to changes caused by the binding of polymerizing fibrin to Mj, , and different from the effects of fibrinogen binding.

In one aspect, the invention provides a method for affecting phosphoinositide metabolism in platelets, comprising exposing the platelets to fibrinogen-coated microspheres (FCMs).

In another aspect, the invention provides compositions useful for treating conditions, such as thrombocytopenia, that are characterized by a deficiency or dysfunction of platelets. In one embodiment, these compositions are microspheres (also referred to as microparticles) having a size range of from about 5 to about 5000 nanometers and comprising on their surface a compound that binds the (Xjjbpg receptor of activated human platelets, wherein incubation of the particles and activated platelets results in an decrease in platelet PIP2 when compared to activated human platelets incubated with similar particles not comprising the compound. The microparticles can be of natural or artificial materials but typically are made from a protein, such as human serum albumin. The protein may be cross-linked. The microspheres are coated with a peptide, polypeptide, or nonproteinaceous molecule capable of interacting with the platelet a"bß3 receptor, which may be cross-linked to the basic microsphere (e. g., a cross-linked human serum albumin microsphere). For example, the microspheres may be coated with human fibrinogen, or with a variant or mutant (e. g., a deletion mutant) of human fibrinogen or a fibrinogen chain (i. e., other than naturally occurring human fibrinogen).

The invention also provides a method of reducing bleeding time in a patient comprising administering a therapeutically effective dose of the coated microparticles.

In still another aspect, the invention provide an assay for identifying an antithrombocytopenic or putatively antithrombocytopenic microparticle, comprising coating a protein microparticle with a test compound, incubating the coated microparticle with activated human platelets, detecting phosphatidylinositol 4,5-biphosphate (PIP2) and phosphatidylinositol 4-phosphate (PIP) levels in the platelets, and correlating a decrease in PIP2 levels with the ability of the coated microparticle to act as an antithrombocytopenic microparticle.

BRIEF DESCRIPTION OF THE FIGURES Figure 1 shows changes in the labeling of PIP2 and PIP in platelets exposed to polymerizing fibrin (compared with fibrinogen) or TS1 microspheres (compared to CS control spheres).

Figure 2 shows labeling of PIP2 and phosphatidic acid in chymotrypsin- treated rabbit platelets incubated with fibrinogen, polymerizing fibrin, control spheres (CS), or TS1 microspheres for 2 minutes with stirring at 200 rpm in an aggregometer. The labeling in the control samples (Fbg for Fgb/Fgn or CS for CS/TS1) was expressed as 100%.

Figure 3 shows ADP-stimulated changes in the labeling of PIP2 and PIP in platelets prelabelled with [32P] phosphate, showing the effects of previous exposure in vivo of the platelets to saline (Control), control spheres (CS) or Thrombospheres- (TS). Results are from three experiments with triplicate samples for CS and TS, duplicate for saline.

Statistical significance indicates the significance of the ADP-induced change in labeling.

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to the interaction of fibrinogen-coated protein microspheres with platelets, specifically the platelet allbß3 receptor. In one aspect, the present invention is related to the discovery that fibrinogen-coated microspheres (FCM) interact with the (XllbP3 receptor and cause changes in platelet metabolism, e. g., a decrease in PIP2 levels, that are comparable to changes caused by the binding of polymerizing fibrin but different from changes caused by soluble fibrinogen (i. e., fibrinogen not bound to a microsphere). Without wishing to be bound by any particular mechanism, these receptor mediated changes resulting from FCM-platelet interactions are believed to:

(i) modulate the signals which originally activated the platelet (i. e., signals that changed the conformation of the acid 3 receptor); and (ii) modify the response of platelets to stimuli (e. g., including stimuli which occur subsequent to platelet interaction with FCMs, such as stimuli occurring after FCMs are cleared from blood in vivo). Examples of stimuli include the binding to platelets of fibrinogen, immobilized fibrin and fibrin.

By way of background, the confirmation of the allbP3 receptor is believed to be affected both by the binding of ligands to the external domain and by intracellular mechanisms (e. g., reflected in changes in phosphinositide levels). For example, fibrinogen binding to the a"bß3 receptor usually requires internal changes in the platelet (i. e., platelet activation) that cause a conformational change in the external portion of a"bß3. However, conformational changes in allbP3 in the absence of activation are also sufficient to expose the fibrinogen binding site. For example, binding of anti-LIBS (anti-ligand-induced binding site) antibodies (which bind with high affinity to ligand-occupied a"bß3 receptors and with lower affinity to unoccupied a"bß3), can cause a conformational change that permits binding of fibrinogen.

Conformational changes in the a"bß3 receptor may permit interactions with components of the intracellular signal transduction mechanisms, stimulating internal signals which result in modification of the cytoskeleton, association of contractile elements with the individual receptors and clustering of the receptors mediated by the contractile apparatus. A 7 state model for all3, based on structural and functional criteria, has been proposed (see, Plow and Ginsberg, Prog. Hemost. Thromb. 9: 117,1989).

Thus, binding of ligands to the a"bß3 receptor of platelets results in changes in the internal state of the platelet and, conversely, changes in the internal state of the platelet (e. g., a decrease in PIP2) affect the properties of the receptor. These properties include conformation, affinity for fibrinogen, and other changes which occur subsequently to fibrinogen binding in vivo, such as clot retraction subsequent to fibrin binding or spreading.

The discoveries disclosed herein have particular relevance to the therapeutic use of fibrinogen-coated microspheres (FCM) including fibrinogen-coated protein microspheres, e. g., for treatment of thrombocytopenia in humans. It has been discovered that fibrinogen coated human serum albumin microspheres, exemplified by ThrombospheresTM-brand microspheres such as TS1 and TS3 microspheres described infra,

have the ability to support hemostatic function in vivo under conditions of low platelet counts (experimental thrombocytopenia). The administration of Thrombospheres shortens bleeding times in thrombocytopenic rabbits; interestingly, this effect persists after the protein microspheres have been cleared from the circulation (see commonly assigned provisional patent application U. S. S. N. 60/048685, attorney docket no. 016197-001400, entitled"Fibrinogen Coated Microparticles"which was filed on June 5,1997, and which is incorporated herein by reference in its entirety and for all purposes). This phenomenon may be explained, in part, by the discovery, disclosed herein and described in detail in the Examples, that interaction of FCMs with platelets activates a signaling pathway from the ajjbps receptor to enzymes involved in phosphoinositide interconversion. Moreover, it has been discovered that this pathway is activated by the presentation on FCMs of a high density of OGIIbt'3 binding sites (i. e., fibrinogen molecules). This understanding of FCM/platelet interaction provides the basis for the compositions of the present invention.

Thus, in one aspect, the present invention provides a method for affecting phosphoinositide metabolism in platelets by exposing the platelets to FCMs.

In a second aspect, the invention provides FCM, especially FCMs comprising deletion mutants of fibrinogen wherein the deletions have been selected on the basis of their a"bß3-mediated effects on polyphosphoinositide metabolism. In particular, FCMs coated with fibrinogen deletion mutants lacking the y chain C-terminal sequence will reduce bleeding time to the same extent as FCMs coated with full-length fibrinogen, but will be even more resistant to causing platelet aggregation.

In a third aspect, the invention provides an assay for identifying compositions that, when bound to microspheres, e. g., protein microspheres, will interact with platelets to produce an antithrombocytopenic effect. Specifically, activated platelets are incubated with a microsphere coated with the test compound and the levels of phosphatidylinositol 4,5- biphosphate (PIP2) and phosphatidylinositol 4-phosphate (PIP) assayed by any standard method (see Examples, infra). Compositions that cause a significant decrease in PIP2 levels (as described in the Examples, infra, for TS1) are determined to be putative antithrombocytopenic agents. A significant decrease is a decrease of at least about 3 %, more often about 4% or about 5 %, most often at least about 6%.

In a related aspect, the invention provides a method of identifying an antithrombocytopenic microsphere by incubating a plurality (usually at least 106-109

microspheres) of microspheres with activated platelets (e. g., activated human platelets), and detecting phosphatidylinositol 4,5-biphosphate (PIP2) levels in the platelets. A decrease in PIP2 levels indicates that the microsphere is useful as an antithrombocytopenic microsphere.

Usually the decrease is a decrease of at least about 3 %, more often about 4% or about 5 %, most often at least about 6% compared to controls.

In another aspect, the invention provides a method of reducing bleeding time in an animal, such as a human patient, e. g., a human suffering from thrombocytopenia, by administering a therapeutically effective dose of microspheres. According to the invention, the microspheres comprise on their surface a compound that binds the a"bß3 receptor of activated platelets, wherein incubation of the particles with activated platelets results in an decrease in platelet phosphatidylinositol 4,5-biphosphate (PIP2) levels when compared to activated platelets incubated with similar particles not comprising the compound. Typically the decrease is a decrease of at least about 3 %, more often about 4 % or about 5 %, most often at least about 6%.

The terms"therapeutically effective dose"or"pharmacologically effective amount"are well recognized phrases and refer to that amount of an agent effective to produce the intended pharmacological result. Thus, a therapeutically effective amount is an amount sufficient to ameliorate the symptoms of the disease being treated, e. g., thrombocytopenia.

Administration can be via any accepted systemic or local but usually will be by intravenous injection or infusion. The actual dose of microspheres administered will depend on the disease condition being treated, the health of the patient, and other factors. It is expected that a dose will comprise between about 101 and 1012 microspheres per kg, more often between about 109 and 10"microspheres per kg. The amount actually administered will be dependent upon the individual to which treatment is to be applied, and will preferably be an optimized amount such that the desired effect is achieved without significant side-effects. The determination of a therapeutically effective dose is well within the capability of those skilled in the art. Usually the microspheres will be administered as a suspension, for example an aqueous suspension comprising an excipient (e. g., 0. 1M arginine, 5 mM citrate, 0.5 mM EDTA, 1% lactose, 1% maltose and 0.1 % Tween-80@).

The microspheres of the invention are similarly useful for forming an aggregate inside blood vessels only at the site of a wound due to the action of thrombin by

administering a therapeutically effective dose of microspheres comprising on their surface a compound that binds theallbP3 receptor of activated human platelets.

I. Microspheres As used herein the terms"microsphere"and"microparticle"are used interchangeably and refer to particles having a size range of from about 5 to about 5000 nanometers (diameter), with a mean size of about 800 to about 1500 nm preferred and about 1200 nm most preferred. For therapeutic uses it is preferred that the preparation be substantially free of large particles or aggregates of particles (e. g., larger than 25,10 or 7 uM in diameter).

The microspheres of the invention may be composed of any biocompatible material, but usually a made primarily from protein, e. g., human serum albumin. In a preferred embodiment, the albumin is cross-linked (e. g., using glutaraldehyde or another aldehyde).

In preferred embodiments, the microspheres of the invention are made by the method of Yen, e. g., as described in U. S. Patent 5,069,936, PCT Publication WO 96/39128, and coassigned patent application U. S. S. N. 60/048685, (described supra), each of which is incorporated by reference in their entirety and for all purposes. In a most preferred embodiment the spheres are made according to the following protocol (describing the production of"TS3"spheres); except that"fibrinogen variant"may be substituted with an equivalent amount of another peptide, polypeptide, etc.

1) HSA, 25% USP is first diluted with normal saline (0.9% sodium chloride solution USP), and Sotradecol (3 % USP) is added to a final concentration containing 15 % HSA (w/v) and 0.002% (v/v) of Sotradecol (this mixture hereafter referred to as sHSA) The volumes of each ingredient subsequently added are multiples or fractions of the initial volume of Sotradecol containing-15% HSA used, which is defined as"one volume".

Sotradecol is sodium tetradecyl sulfate.

2) 1.0 volume of the above mixture (sHSA) is mixed quickly with 1.8 volume of 70% ethanol (with the remainder volume 30% injection-grade water), at which time turbidity is immediately observed; 3) 0.11 volume of glutaraldehyde is mixed into the turbid suspension immediately;

4) Thereafter, 1.45 volumes of normal saline is added; 5) The bulk suspension is then placed in a cold room (about 4°C) and stirred slowly (40 10 rpm) in the final bulk container by placing the unopened container on a moving platform until the step for filling into glass bottles or filtration was ready to proceed.

6) Large (> 7 Itm diameter) particles and aggregates are removed from the sphere suspension, e. g., by filtration (e. g., using a 3 micron or a 5 micron filtering system such as a Membrex Pacesetter Pilot Filter System 400cm2 (cat. # pspilot) fitted with a 3/mi (cat. # 3039-003) or 5/mi (cat. # 3039-005) SteelPore 400cm2 Cartridge, Membrex Inc., 155 Route 46 West, Fairfield, NJ 07004) or a similar filtering system). The filtrate is then concentrated and residual soluble HSA removed using a sterile 0.2 micron Asahi hollow fiber cartridge dialysis system (Plasmaflo AP-05HL, Asahi Medical Co.) or the equivalent.

Large particles can also be removed by other means, such as centrifugation.

8) Glutaraldehyde (1.25 %) is added to a final concentration of 0.05%, immediately followed by addition of a 1 mg/ml protein solution (e. g., a fibrinogen variant or other polypeptide) with mixing to result in a final concentration of 0.33 mg/ml of the "coating"protein (e. g., fibrinogen variant).

9) A 5-fold concentrated solution of excipients is added (1 part per 4 part of adjusted suspension volume). In a preferred embodiment the final concentration of excipient is as follows: arginine (2.1%), maltose (2.0%), lactose (2.0%), citrate (0.0105%), EDTA (0.0186%), Tween 80 (0.01 %), adjusted to pH 6.4 with sodium hydroxide solution.

Thereafter the preparation is lyophilized 10) The microparticles are resuspended in water. After no solid particles remain visible to the eye, the product is assayed or used.

It will be appreciated that numerous variations in concentration, choice of aldehyde cross-linkers, incubation times, etc. can be introduced into the protocol above without materially changing the resulting microspheres. For example, different excipients may be used, centrifugation may be substituted for the filtration step, etc.

II. Fibrinogen (e. g.. Deletion Mutants ! As noted supra the microparticles may be coated with a variety of compounds such that the coated microparticle will bind to the llbß3 receptor of activated platelets or the equivalent receptor on other cells. Suitable compounds include peptides, polypeptides and other molecules which can be identified on the basis of structural or functional similarity to fibrinogen or by using the assay disclosed herein. Peptides and polypeptides can be isolated from natural sources or produced by recombinant or synthetic means well known in the art (see, e. g., Sambrook et al., Molecular Cloning-A Laboratory Manual, 2nd Ed., Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1989, and Stewart and Young, Solid Phase Peptide Synthesis, 2d. ed., Pierce Chemical Co., 1984, which is incorporated by reference herein).

Preferred compounds are variants and mutants (e. g., deletion mutants) of fibrinogen. Fibrinogen is a large (340 kDa) protein consisting of two sets of Aa, Bß and y chains (see, e. g., Rooney et al. JBiol. Chem 271: 8553,1996, and Weisel et al. Science 230: 1388). Expression vectors encoding the Aa, Bp and y chains have been described (e. g., Rooney et al., supra ; Binnie et al., Biochemistry 32: 107,1993; Lord et al., Biochemistry 35: 2342, each of which is incorporated by reference herein in its entirety and for all purposes). Mutants and deletions of fibrinogen and fibrinogen chains can be made by one of skill using routine molecular biological techniques. See, e. g., Sambrook et al., supra, and Ausubel et al., Current Protocols in Molecular Biology, Greene Publishing and Wiley-Interscience, New York (1995) each of which is incorporated by reference herein in its entirety and for all purposes. Thus, typically, the receptor-binding compound is human fibrinogen, or a fragment or variant of a human fibrinogen polypeptide. In each case, the fragment or variant will be capable of specifically interacting (e. g., binding) to the llbß3 receptor. Thus, in some embodiments, the compound is naturally occurring (e. g., isolated) human fibrinogen, or a recombinant protein or polypeptide having the amino acid sequence of naturally occurring fibrinogen. Alternatively, a receptor binding fragment of fibrinogen (e. g., a fragment that is less than the full-length molecule). Usually a polypeptide fragment is at least about 6 residues in length, more often at least about 10,12,15,25,50, or 100 residues in length up to full-length.

Of particular interest are microspheres coated with recombinant fibrinogen lacking the C-terminal sequence of the y-chain. Fibrinogen mutants of this type have been

reported to support platelet retraction of clots, but not platelet aggregation (Rooney et al., J Biol. Chem 271: 8553,1996, which is incorporated by reference). Thus, microspheres coated with this (and similar) fibrinogen deletion polypeptides will have the same effect (antithrombocytopenic) effect as microspheres (e. g., Thrombospheres) coated with full- length fibrinogen as to reduced bleeding time, but may have even less likelihood of causing platelet aggregation.

III. Assay As is described in detail in the Examples, platelet-ligand interactions can be characterized by measuring the effect of the interaction on phosphoinositide metabolism.

The interaction of Thrombosphere-brand FCMs shown to have therapeutic effects (e. g., reduction in bleeding time in thrombocytopenic rabbits) with platelets has now been shown to result in a specific effect (decrease in PIP2). Accordingly, the present invention provides a method for identifying a ligand, such as a fibrinogen-deletion mutant or other polypeptide, that, when coated onto a microparticle, e. g., a protein microsphere, has similar or improved therapeutic effects. As used herein, a microparticle is deemed to be antithrombocytopenic when, if administered to a experimental thrombocytopenic rabbit according to, e. g., Blajchman and Lee, 1997, Transfusion Med. Reviews 11: 95-105, reduces bleeding time by at least about 200% or about 300% at 5,10 or 24 hours post-administration compared to animals to which the agent is not administered. Other suitable assay conditions (e. g., reduction in blood loss) are described in detail in commonly assigned provisional patent application U. S. S. N. 60/048685, described supra.

Compounds that can be tested in the assay include, but are not limited to, fibrinogen deletion mutants and other peptides and polypeptides such as described supra.

EXAMPLES Materials and Methods A. Fibrinogen-coated albumin microspheres Fibrinogen-coated albumin microspheres (TS1) were prepared as described in U. S.

Patent No. 5,069,936, and in commonly assigned provisional patent application USSN 60/048,685. TS1 microspheres were produced as follows:

1) HSA, 25% USP was first diluted with normal saline (0.9% sodium chloride solution USP), and Sotradecol (3 % USP) was added to a final concentration containing 15 % HSA (w/v) and 0.002% (v/v) of Sotradecol (this mixture hereafter referred to as sHSA) [the volumes of each ingredient subsequently added were multiples or fractions of the initial volume of Sotradecol containing-15% HSA used, which was defined as"one volume"]; 2) 1.0 volume of the above mixture (sHSA) was mixed quickly with 1.8 volume of 70% ethanol (with the remainder volume 30% injection-grade water), at which time turbidity was immediately observed; 3) 0.11 volume of 1.25 % glutaraldehyde (diluted from a 25 % glutaraldehyde stock with normal saline) was mixed into the turbid suspension immediately; 4) A 1.45 volume of fibrinogen solution (1.0 mg/ml, diluted from a stock of 10 mg/ml dissolved with water) was mixed in within 10 minutes to coat the spheres in the presence of glutaraldehyde.

5) The TS1 was then lyophilized.

6) To resuspend, normal saline was injected into the vials with a syringe and needle for reconstitution into suspensions. After no solid particles remain visible to the eye, samples were used as described infra.

Control spheres (CS) are albumin microspheres produced in the same manner as TS1 but not coated with fibrinogen. CS microspheres are produced as described above except, that normal saline was added in place of the fibrinogen solution at step (4).

B. Platelets Rabbit platelets were obtained from New Zealand White rabbits as described in Kinlough-Rathbone et al."Platelet Aggregation,"in Harker, (ed.) METHODS IN HEMATOLOGY, MEASUREMENT OF PLATELET FUNCTION, Edinburgh, U. K., Churchill, Livingstone 64 1983 (based on original methods of Mustard et al., Brit. J. Haematol, 22: 193,1972).

Chymotrypsin treatment of platelets was carried out by adding chymotrypsin to platelet suspensions to a final concentration of 10 U/ml, and incubating for 30 min, 22°C with 10 zM PGEl present, as described in Greenberg et al., Blood 54: 753,1979. The effect of pretreatment of human or rabbit platelets with chymotrypsin on their responses to human fibrinogen and aggregating agents.

C. Phosphoinositide Assays Phosphoinositide assays were carried out using washed rabbit platelets prelabelled with 32P and pretreated with chymotrypsin (10 U/ml, 30 min, 22°C with 10 yM PGE, present) were suspended in Tyrode-albumin solution (Kinlough-Rathbone et al., supra) with CA" (2mM) and apyrase (Molna et al., 1961, Arch. Biochem. Biophys. 93: 353- 363). TS were incubated with chymotrypsin-treated (CT) platelets and the phosphoinositides quantified by extraction of platelet suspensions, thin layer chromatography and counting of incorporated 32p. Methods for extraction, fractionation and quantitation both of chemical amount and radioisotopic labeling of phosphoinositides have been published (Vickers et al., Anal. Biochem. 224: 449,1995).

D. Polymerizing fibrin Polymerizing fibrin was produced by the action of the snake venom enzyme batroxobin on fibrinogen (0.4 mg/ml) which removes fibrinopeptide A converting fibrinogen to fibrin which spontaneously polymerizes EXAMPLE 1 PLATELET AGGREGATION Platelet aggregation in the presence of TS was measured using a Coulter counter to follow decreases in particle numbers, due to aggregation of platelets in twos or threes. Activation of platelets with ADP resulted, as expected, in a decrease in particle count (Table 1). Addition of increasing amounts of Thrombospheres caused increasingly larger decreases in the particle count for ADP-activated platelets. In contrast, control spheres had no effect on particle count for ADP-activated platelets.

TABLE 1 Decreases in Platelet Count Upon Activation of Platelets with 10, uM ADP in the Absence and Presence of Fibrinogen-Coated Microspheres Microspheres Added'Platelet Count (Platelets x 10-8/mol) (number/ml final Control2 ADP Change concentration) None (Tyrode's 8.22 6.16-0.206 Solution) Control Spheres 1.5x 109 7.63 5.50-0.213 Thrombospheres 1.5 x 10'10.09 6.23-0.386 1.5 x 108 9.64 5.64-0.400 1.5 x 109 8. 73 3.80-0.493 1. Platelets stirred at 200 rpm. 2. Control platelet count approximately 1 x 109 platelets/ml.

Although the number of Thrombospheres in the two lower concentration samples is considerably less than the number of platelets, the decrease in particle count is consistent with a substantial portion of the platelets being in aggregates.

Since it was apparent that more platelets were involved in aggregates with the highest concentration of Thrombospheres, that concentration (1.5 x 109/ml) of Thrombospheres and Control Spheres was used for examination of the effect on phosphoinositide metabolism.

EXAMPLE 2 PHOSPHOINSOSITIDE CHANGES IN ADP-STIMULATED PLATELETS INDUCED BY INCUBATION WITH MICROSPHERES The effects of Thrombospheres on phosphatidylinositol 4,5-biphosphate (PIP2) and phosphatidylinositol 4-phosphate (PIP) levels was examined in unstimulated (Tyrode's solution added) and ADP-stimulated platelets. Platelets were prelabelled with [32P] phosphate and phosphoinositides quantified by determination of labeling with [32p phosphate as previously described (chemical amount was determined by phosphate assay but data are not shown since they parallel labeling results). See, Vickers, J. D., 1995, Anal.

Biochem. 224: 449. ADP-stimulated- (10 jus) or unstimulated-platelets were exposed to polymerizing fibrin, fibrinogen, Thrombospheres, or control spheres for 1 minute while stirring in an aggregometer cuvette at 200 rpm, and then extracted to recover phospholipids.

The effect of polymerizing fibrin was compared to fibrinogen and the effect of Thrombospheres was compared to control spheres (Figure 1 ;"Fibrin"= the labeling induced by polymerizing fibrin minus the background labeling induced by fibrinogen, and "Thrombospheres"= the labeling induced by Thrombospheres minus the background labeling induced by control spheres). Fibrinogen was used as a control for polymerizing fibrin since in earlier studies of ADP stimulated platelets polymerizing fibrin caused a change beyond the fibrinogen-dependent change and in studies of fibrinogen agglutination of CT platelets (human published, rabbit not published) was found to not cause a change in PIP2. See, Vickers et al., 1990, Platelets 1: 199 and Vickers et al., 1994, Ann. N. Y. Acad.

Sci. 714: 287. Statistical significance (denoted by an asterisk) indicates that the treatment (polymerizing fibrin or Thrombospheres) changed the level of PIP2 or PIP significantly compared to the corresponding control treatment.

Neither polymerizing fibrin nor Thrombospheres caused statistically significant changes in PIP2 or PIP compared to fibrinogen or control spheres, respectively, in unstimulated platelets (Figure 1). The observed decrease, though not statistically significant, suggests that polymerizing fibrin may interact to some extent with unactivated platelets causing a weak signal. That is, the interaction may be sufficient to cause a small conformational change in CCIlbp3 which results in internal changes in the platelets, but the change is below a threshold necessary to cause the large changes in PIP2 seen with the activated platelets.

In ADP-stimulated platelets, polymerizing fibrin caused a large decrease in PIP2 compared with fibrinogen (Figure 1, upper panel, left side). This result is consistent with previous studies (Vickers et al., 1986, Biochem. J. 237: 327; Vickers et al., 1990, Platelets 1: 199; Vickers et al., 1994, Ann. N. Y. Acad. Sci. 714: 287)), Similarly, Thrombospheres caused a significant decrease in PIP2 in comparison with control spheres, when binding was activated by ADP-stimulation, although the magnitude was less than that caused by polymerizing fibrin (Figure 1, upper panel, right side). These results suggest that Thrombospheres act, at least in part, through a mechanism that is similar to the binding of polymerizing fibrin.

Consistent with previous observations, the large decrease in PIP2 caused by polymerizing fibrin was paralleled by an increase in PIP (Figure 1, lower panel). (The absolute magnitudes of the changes cannot be compared since PIP2 contains three phosphates and PIP only two.) Thrombospheres also caused an increase in PIP in activated platelets (Figure 1, lower panel), although it was not statistically significant.

An interesting result was the failure of ADP to cause a decrease in PIP2 in the presence of control spheres (data not shown). Rabbit platelets do not require added fibrinogen to aggregate and the decrease in PIP2 after ADP stimulation has been found to occur whether or not fibrinogen is added. A possible explanation is that the control spheres act as a physical impediment to platelet interactions required for the decrease in PIP2.

Taken together these results indicate that Thrombospheres cause a signal in platelets similar to that caused by polymerizing fibrin, although it is weaker.

The indication that polymerizing fibrin may cause a weak signal even with unactivated platelets opens the possibility that transient low affinity interaction (receptor in a low affinity (inactivated) state) may cause changes (e. g., to receptor conformation and in PIP2 levels) which affect subsequent responses. An example of a subsequent response is the activation response (possibly including aggregation, adhesion as well as the internal biochemical changes including the decrease in PIP2) upon stimulation by agonist.

EXAMPLE 3 PHOSPOINOSITIDE CHANGES DUE TO AGGLUTINATION OF CHYMOTRYPSIN- TREATED PLATELETS BY FIBRINOGEN. POLYMERIZING FIBRIN. AND THROMBOSPHERES For these studies platelets were prelabelled with [32P] phosphate and phosphoinositides quantified by determination of chemical amount (phosphate assay, data not shown since they paralleled labeling results) and labeling with [32P] phosphate incorporation. After labeling, platelets were treated with chymotrypsin to cause a limited digestion to modify the a"bß3 receptor, followed by washing to remove chymotrypsin.

Chymotrypsin treatment has been shown previously to remove a 4 kDa peptide from allb (Pidard et al., 1991, Eur. J. Biochem. 200: 437-447) which results in exposure of the binding site for fibrinogen and fibrin. Chymotrypsin treatment renders the fibrin (ogen) receptor a"bß3, active and permits study of signaling without the complications of agonist-

induced signals involved in receptor activation (Vickers et al., 1994, Ann. N. Y. Acad. Sci.

714: 287).

In control samples, incubation of Thrombospheres with untreated, unstimulated platelets had no effect on the phosphoinositides. Incubation of chymotrypsin- treated (CT) platelets with fibrinogen caused agglutination of the platelets but had no effect on phosphoinositide levels (data not shown). This result is consistent with previous findings using human platelets (Vickers et al., 1994, Ann. N. Y. Acad. Sci. 714: 287) and rabbit platelets (not shown). In contrast, incubation for 3 min. with polymerizing fibrin resulted in a decrease (26 9.7%) in labeling of PIP2 (p < 0. 05, n=6) (Figure 2) and a parallel increase in PIP (not shown). Incubation of CT platelets with a low concentration of Thrombospheres (1.2 x 108/mol) had no detectable effect on the levels of the phosphoinositides (not shown), but incubation for 3 min. with higher concentrations of Thrombospheres [i. e., 0.75 x 109/ml (not shown) or 1.5 x 109/ml (Figure 2)], resulted in a significant (13 4%) decrease in labeling of PIP2 with [32p] phosphate compared to CT platelets incubated with control spheres without bound fibrinogen (p < 0. 05, n=8).

Thrombospheres at the higher concentrations did not significantly affect PIP. The results in Figure 2 were obtained by labeling of PIP2 and phosphatidic acid (PAP in [32P] phosphate- labeled chymotrypsin-treated rabbit platelets incubated with fibrinogen (Fbg), polymerizing fibrin (Fbn), control spheres (CS) or Thrombospheres (TS) for 2 minutes with stirring at 200 rpm in an aggregometer. The labeling in the control samples (Fbg for Fgb/Fgn or CS for CS/TS) were expressed as 100%.

To confirm results from previous studies that phospholipase C activation was not involved in the decrease in PIP2, changes in phosphatidic acid were examined.

Phosphatidic acid increases up to 20 fold in platelets in which phospholipase C is activated by thrombin stimulation. Polymerizing fibrin caused only a small change and Thrombospheres did not affect the labeling of phosphatidic acid (Figure 2).

Thus, by using an experimental system in which changes in the phosphoinositides resulting from ADP-induced activation of the platelets can be distinguished from effects caused by ligand (e. g., fibrin or fibrinogen) binding, Thrombospheres appear to cause changes that are similar to those caused by polymerizing fibrin and different from fibrinogen.

EXAMPLE 4 EFFECTS OF IN VIVO EXPOSURE OF PLATELETS TO THROMBOSPHERES ON EX VIVO ACTIVATION OF PLATELETS BY ADP For this study, rabbits were injected with saline (control), control spheres and Thrombospheres (three animals/experimental treatment) and 40 hours later the rabbits were killed and platelets collected. After washing and labeling as described supra, platelets were aggregated by stimulation with ADP and phospholipids assessed. No consistent differences were observed in platelet aggregation related to the experimental treatment. As shown in Figure 3, in three experiments platelets exposed to Thrombospheres in vivo showed a significant decrease in PIP2 upon stimulation with ADP for 1 min, whereas platelets from control animals (exposed to saline or control spheres) showed smaller decreases which were not sufficiently consistent across the three animals to be statistically significant. The results with the platelets from rabbits injected with saline or control spheres is consistent with experience over the years that a large number of experiments (5-7) is required to demonstrate a significant decrease in PIP2 in ADP-stimulated platelets. The contrasting statistically very significant ADP-stimulated decrease in PIP2 in the platelets from rabbits injected with Thrombospheres, emphasizes that there has been a substantial change in the platelets to yield this result. The magnitude of the decrease which is about 10% is consistent with previously reported decreases in PIP2.

The difference between the changes in PIP2 contrasts with the similarity of the increases in PIP in response to ADP stimulation. In a previously published study of changes in phosphoinositides in ADP-stimulated platelets, it has been demonstrated that the decrease in PIP2 is dependent upon binding of fibrinogen to the platelets, whereas the increase in PIP is independent of fibrinogen binding and thus only dependent upon initial activation of the platelets by ADP (Vickers et al., 1993, Eur. J. Biochem. 216: 231). Thus the effect of prior exposure of the platelets to Thrombospheres in vivo appears to have specifically affected the pathway involved in the decrease in PIP2 in ADP-stimulated platelets, which involves allb, B3 and the enzymes of PIP2 metabolism, and not other pathways involved in ADP-stimulated platelet activation including that responsible for the increase in PIP (which is unknown).

Although the foregoing invention has been described in detail for purposes of clarity of understanding, it will be obvious that certain modifications may be practiced within the scope of the appended claims. All publications and patent documents cited above are hereby incorporated by reference in their entirety for all purposes to the same extent as if each were so individually denoted.