The β3 subunit of the integrin αIIbβ3 regulates αIIb-mediated outside-in signaling

Integrins are αβ heterodimeric receptors required for numerous essential functions in metazoan cells.¹  The megakaryocyte- and platelet-specific integrin αIIbβ3 is essential for normal hemostasis.²  Platelet adhesion, aggregation, and bidirectional signaling are mediated by αIIbβ3.³  Inappropriate activation of αIIbβ3 contributes significantly to cardiovascular disease,²  the leading cause of death in the Western world. Consequently, understanding the regulation of αIIbβ3 function has enormous potential for facilitating design of antiplatelet therapeutics to reduce the risk of thrombotic events. Additionally, acquisition of this understanding has far-reaching significance for cell biology because basic concepts explaining regulation of αIIbβ3 function typically are relevant to related integrins.^4-6 

Many integrins, including αIIbβ3, are unable to bind their ligands or signal in their low-affinity, or resting, state.¹  Transformation from the resting state to the active or high-affinity state typically results from integrin-mediated inside-out signaling initiated indirectly by activation of other receptors.¹  This transformation induced by inside-out signaling (cytoplasmic interactions) may affect the equilibrium between the 2 states by trapping the receptor in the active state. The active, oligomerized receptors can initiate and propagate integrin-mediated outside-in signaling. The αIIbβ3-mediated outside-in signaling elicits thromboxane A₂ (TxA₂) production and adenosine triphosphate (ATP) secretion that amplifies, propagates, and thereby perpetuates the signaling initiated by physiologic agonists.⁷  The clinical importance of TxA₂ production is evident from the efficacy of aspirin as an antithrombotic agent.⁸  Despite important advances in our understanding of the regulation of integrin inside-out signaling, little is known about the regulation of αIIbβ3-mediated outside-in signaling^1,9  even though that signaling elicits the production of clinically important autocrines.⁷ 

The mechanisms underlying the transformation of the integrin to the high-affinity state is controversial.^1,10-13  But the generally accepted model of activation is described as a switchblade-like transformation from the bent, compact, highly subunit-interacting resting state into an extended, less interacting, more open active conformation.^14-17 

Retention of the integrin in the bent, compact resting state apparently is controlled by interaction between the membrane proximal, highly conserved regions of the cytoplasmic domains of the α and β subunits.^18-21  Disruption of this interaction by mutation results in the constitutive activation of the affected αIIbβ3 heterodimers expressed in CHO and 293T cells.^19,22,23  Agonist-induced physiologic disruption of this interaction appears to be caused by the binding of talin⁵  or other proteins^1,9  to the cytoplasmic domain of β3.

Interactions between the cytoplasmic domains of αIIbβ3 also affect αIIbβ3-mediated inside-out signaling. For example, a myristoylated peptide corresponding to the αIIb cytoplasmic domain (residues K989-E1008)¹³  and a palmitoylated peptide (myristoylation²⁴  and palmitoylation²⁵  each enable the derivatized peptides to enter the cell^13,25,26 ) corresponding to αIIb K994-D1003²⁷  prevented or greatly diminished αIIbβ3 activation induced by inside-out αIIbβ3 signaling in response to adenosine diphosphate (ADP) and epinephrine. Presumably, binding of the myristoylated or palmitoylated αIIb peptides to the cytoplasmic domain of β3 prevented β3 inside-out signaling that would have activated αIIbβ3 under normal conditions.^13,27  Thus, the interaction of membrane distal regions of the cytoplasmic domains of αIIb and β3 appear to regulate αIIbβ3-mediated inside-out signaling.

Given the regulation of β3 inside-out signaling by the cytoplasmic domains of αIIbβ3, we hypothesized that these interactions might also regulate αIIb-mediated outside-in signaling. Here we show that this hypothesis is correct and that the membrane distal region of the cytoplasmic domain of β3 negatively regulates αIIb-mediated outside-in signaling.

Apyrase, prostaglandin E₁ (PGE₁), the peptide RGDS, protein G, and o-phenylenediamine dihydrochloride (OPD) peroxidase substrate were from Sigma-Aldrich (St Louis, MO). Purified normal human αIIbβ3 was from Enzyme Research Laboratories (South Bend, IN). The soluble extracellular domain of human αIIbβ3 lacking the transmembrane and cytoplasmic domains was a generous gift from Dr Timothy A. Springer (Harvard Medical School, Cambridge, MA). This construct has not been described in the literature, but it was made using the same technique described for the synthesis of the soluble extracellular fragment of αVβ3.¹⁶  The complete extracellular domains of the 2 subunits were fused at each C terminus to peptides that form a disulfide-linked α-helical coiled coil.¹⁶  Donkey anti–mouse antibody was from Jackson ImmunoResearch Laboratories (West Grove, PA). The monoclonal antibodies (mAbs) PT25-2,²⁸  AP3,²⁹  and 7E3³⁰  were generous gifts from Dr Makoto Handa (Keio University, Tokyo, Japan), Dr Peter J. Newman (The Blood Center of Southeastern Wisconsin, Milwaukee), and Dr Barry Coller (Rockefeller University, New York, NY), respectively. D3 was prepared as previously described.³¹  Peptides and palmitoylated peptides were synthesized and characterized by analytical high-performance liquid chromatography and time-of-flight matrix-assisted laser desorption ionization mass spectroscopy and purified, if necessary, by the Hartwell Center for Bioinformation and Biotechnology (St Jude Children's Research Hospital, Memphis, TN). Fluorescein isothiocyanate (FITC) derivatization of peptides was performed as described,³²  and the derivatized peptides were purified by using Waters Sep-Pak Vac 6cc C18 column (Waters, Milford, MA). Care was taken to keep the pH neutral during the FITC derivatization process to minimize the chance of FITC-facilitated hydrolysis of the derivatized amino-terminal residues (Edman degradation).

The patient has a variant form of Glanzmann thrombasthenia designated here as VGTΔ724.⁴  The patient is a young man with a life-long history of enhanced bruising, mucosal bleeding, and petechiae. He has a normal platelet count but a prolonged bleeding time. The patient contains 2 distinct mutant alleles for β3; an allele that contains a transition (CGA [R] to TGA [nonsense]) mutation that results in a truncated β3 lacking all but the 8 membrane proximal cytoplasmic domain residues. The other allele sustains a deletion of β3 T1181 resulting in a frame shift and a nonsense codon corresponding to amino acid 642. So, the abnormal platelets express only αIIbβ3 complexes containing the truncated β3 subunit, which are otherwise structurally normal. These complexes are expressed at about 40% of the normal level. As a consequence of the truncation, the platelets do not aggregate to physiologic agonists because the truncated β3 subunit cannot mediate inside-out signaling that activates the altered αIIbβ3 complex.

After informed consent was obtained, blood was collected from healthy donors and the variant thrombasthenic patient (with approval from the Cincinnati Children's Hospital Medical Center Institutional Review Board) into empty syringes and then transferred to polypropylene centrifuge tubes containing 100 μl/mL Whites anticoagulant (2.94% sodium citrate, 136 mM glucose, pH 6.4) 0.1 μg/mL PGE₁, and 1 U/mL apyrase. Platelet-rich plasma (PRP) was prepared by differential centrifugation. Washed platelets were prepared by differential centrifugation (1100g for 10 minutes) of the PRP containing 5 mM EDTA (ethylenediaminetetraacetic acid). Platelets were resuspended into modified Tyrode solution (12 mM NaHCO₃, 138 mM NaCl, 5.5 mM glucose, 2.9 mM KCl, 2 mM MgCl₂, 0.42 mM NaH₂PO₄, 10 mM HEPES [N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid], pH 7.4). Aggregation was measured in a lumi-aggregometer (Chrono-Log, Havertown, PA) using washed platelets (300 μL) adjusted to approximately 10⁶ platelets/μL.

ATP secretion was measured using the CHRONO-LUME reagent (Chrono-Log) as described.⁷  ATP secretion was used as a surrogate measure of ADP release, so the terms are used interchangeably in the text.

TxB₂, a stable metabolite of TxA₂, was measured as described.⁷  TxB₂ was measured to indirectly estimate TxA₂ production, so the terms are used interchangeably in the text.

Platelet factor 4 (PF4) was released from the α-granules by the activated platelets. After washed platelets were stimulated, supernatant fractions were collected and assayed using an Asserachrom PF4 quantitative enzyme-linked immunosorbent assay (ELISA) kit (Diagnostica Stago, Parsippany, NJ), as described.³³ 

Laser scanning confocal microscopy was used to visualize FITC-labeled palmitoylated peptide and FITC-labeled nonpalmitoylated peptide-treated, washed normal platelets. For this purpose, the platelets (300 μL) in Tyrode solution (containing 5 mmol/L EDTA) were incubated with 10 μmol/L final concentration of either the FITC-labeled palmitoylated peptide RKEFAKFEEER or the FITC-labeled nonpalmitoylated peptide RKEFAKFEEER, at 37°C with stirring at 1000 rpm in a Chrono-Log aggregometer for 5 minutes. Then the platelets were centrifuged at 1100g for 5 minutes, and each pellet was suspended in 300 μL Tyrode solution. These platelets were fixed by incubation in an equal volume of fixative solution (2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer containing 2% sucrose, pH 7.3) for 10 minutes at 37°C. The fixed platelets were washed 3 times with phosphate-buffered saline (PBS), and placed on plus glass slides. Successive z-axis serial sections (optical slices) of the fixed platelets were obtained using a Nikon C1 laser scanning confocal microscope (Nikon, Tokyo, Japan) with a × 60 Plan Apo 1.2 NA water objective. Representative images corresponding to the ventral surface, the interior of platelets, and the dorsal surface were selected from a optical-sectioning series of each platelet containing 30 images. Sections were taken every 0.3 μm starting from ventral surface. The color images were transformed into black-and- white images through Photoshop (Adobe System, Mountain View, CA).

The ability of a peptide corresponding to residues R724-R734 of the membrane distal cytoplasmic domain of β3 to bind to the cytoplasmic domain of αIIb was tested using a microtiter well binding assay.³³  For this purpose, 200 μL of 3% bovine serum albumin (BSA) in PBS, or a 5-mmol/L solution of RKEFAKFEEER (β3, R724-R734), its scrambled control EAERKFERKFE, ARAKWDTANN (β3, A735-N744) its scrambled control NNWTAAARKD, KVGFFKRNRPPLEED (αIIb, K989-D1003) and its scrambled control FPFVGNKDKRLEREP, and LSARLAF³³  or its scrambled control FRALASL³³  were incubated in carbonate coating buffer (1.59 g/L Na₂CO₃ and 2.93g/L NaHCO₃, pH 9.6) in the wells of XENOBIND Covalent Binding Microwell Plates (Xenopore, Saddle Brook, NJ) at 37°C for 2 hours to immobilize the peptides. After incubation, the wells were washed 3 times with 300 μL PBS (pH 7.5, with 0.1% Tween 20). Then the wells were postcoated with albumin by using 300 μL PBS containing 3% BSA for 2 hours at 37°C. The postcoated wells were washed 3 times with 300 μL PBS. Then, 200 μLof an 8-μg/mL solution of purified αIIbβ3 or of the extracellular aspect of αIIbβ3 (truncated form of αIIbβ3 lacking the transmembrane and cytoplasmic domains of the receptor¹⁶ ) were incubated with or without peptides (10 μM) in the wells at 37°C for 2 hours. After washing the wells 6 times with PBS, they were incubated for 2 hours with 200 μL of a 1-μg/mL solution of AP3 in PBS containing 3% BSA at 37°C for 2 hours. After incubation, the wells were washed 6 times with PBS, then the wells were incubated with a final concentration of 0.2 μg/mL horseradish peroxidase–conjugated donkey anti–mouse antibody (Jackson ImmunoResearch Laboratories, West Grove, PA) in PBS containing 3% BSA at 37°C for 1 hour. After washing 6 times with PBS, the plates were developed according to the manufacturer's instructions and absorbance was read at 490 nm.

Means were compared using the Student t test.

The hypothesis that the membrane distal region of the cytoplasmic domain of β3 negatively regulates αIIb-mediated outside-in signaling was tested using human platelets that express a truncation of the membrane distal region of the cytoplasmic domain of β3. These platelets (VGTΔ724) have normal external and transmembrane domains of αIIbβ3 but lack all of the β3 cytoplasmic domain other than the 8 membrane proximal residues because of a truncation at residue 724.⁴  These platelets fail to aggregate in response to physiologic agonists or spread on fibrinogen.⁴  Although the abnormal integrin complexes are not constitutively active and cannot be activated by αIIbβ3-mediated inside-out signaling in response to thrombin and other agonists, they can be activated and induced to bind ligand by the ligand-induced binding site (LIBS)–specific mAbs D3 and LIBS6.⁴  The lack of activation of αIIbβ3 in these platelets in response to physiologic agonists apparently results from the inability of the truncated β3 cytoplasmic domain to mediate the inside-out signaling required to activate αIIbβ3 in response to those agonists.^4,5,18,34 

The VGTΔ724 platelets that lack the 39 membrane distal β3 cytoplasmic domain residues⁴  were treated with the LIBS-specific mAbs D3³¹  and PT25-2²⁸  to test our hypothesis. These antibodies cause αIIbβ3 activation but do not elicit TxA₂ production or ATP secretion by normal platelets.^28,31  Normal and VGTΔ724 platelets were treated independently with D3 and PT25-2 in the presence of fibrinogen with stirring. Normal and VGTΔ724 platelets aggregated in the presence of fibrinogen in response to D3 and PT25-2 (Figure 1). In contrast to normal platelets, which neither underwent shape change nor produced TxA₂ or secreted the contents of their storage granules, the VGTΔ724 platelets changed shape, produced TxA₂, and secreted ATP and PF4 (Figure 1). The D3-induced shape change was confirmed using scanning electron microscopy (not shown). These results support the view that the cytoplasmic domain of β3 negatively regulates αIIb-mediated outside-in signaling.

Blockade of fibrinogen binding by the peptide RGDS³⁵  or the anti-αIIbβ3 mAb 7E3³⁰  (both agents block fibrinogen binding) in the presence of D3 eliminated shape change, aggregation, and TxA₂ production and ATP secretion (Figure 2). Also, D3 or fibrinogen alone did not cause TxA₂ production or ATP secretion (not shown). These results demonstrate that the D3-induced signaling measured here requires ligand cross-linking of the receptors. This is evident because RGDS treatment of platelets enhances D3 binding,³¹  but unlike fibrinogen, RGDS cannot cross-link the receptors.

Furthermore, platelet aggregation appears to be required for the D3 plus fibrinogen-induced signaling characterized here because treatment of the VGTΔ724 platelets with D3 and fibrinogen in the absence of stirring did not elicit a substantial level of TxA₂ production or ATP secretion (Figure 2). This conclusion is supported by the observation that a 2-fold dilution of platelets resulted in approximately a 4-fold decrease of TxA₂ production and about a 3-fold decrease of ATP secretion (the dilution experiments were repeated 3 times), rather than the 2-fold decrease in TxA₂ production and ATP secretion that should have occurred if aggregation was not required for signaling (Figure 2). However, our data do not exclude the occurrence of a low level of signaling in the VGTΔ724 platelets that is elicited by ligand cross-linking of the receptors but is enhanced exponentially by aggregation.

The signaling that occurred in response to D3 or PT25-2 plus fibrinogen presumably might have resulted from either αIIbβ3 outside-in signaling or Fc receptor–mediated signaling in response to clustered antibodies. Accordingly, the role of the Fc receptor (FcγRIIA) in D3 plus fibrinogen-induced signaling by VGTΔ724 platelets was evaluated by the following experiments. Normal and VGTΔ724 platelets were treated with D3 and a donkey anti–mouse IgG1κ immunoglobulin (D3 is an IgG1κ immunoglobulin) to cause Fc-dependent TxA₂ production and ATP secretion.³⁶  This treatment of the platelets was repeated in the presence of protein G, which binds to the Fc domain of IgG³⁷  and thereby prevents Fc receptor–dependent signaling. Cross-linking of D3 with the anti-IgG1κ immunoglobulin caused shape change of and TxA₂ production and ATP secretion by both normal and VGTΔ724 platelets, but these responses were eliminated by the protein G. In contrast, the signaling caused by D3 plus fibrinogen treatment of the VGTΔ724 platelets was not affected by protein G (Figure 3). These results demonstrate that the signaling induced by D3 and presumably PT25-2 was not Fc receptor (FcγRIIA) dependent because it was not blocked by protein G and therefore support the view that the LIBS-specific antibody-induced TxA₂ production and ATP secretion resulted from αIIb3-mediated outside-in signaling.

The results presented support the view that the membrane distal region of the β3 cytoplasmic domain negatively regulates αIIb outside-in signaling. If the mutual interactions between the membrane distal regions of the cytoplasmic domains of αIIbβ3 regulate both αIIb and β3 outside-in signaling, a palmitoylated peptide (palmitoylation of the peptide enables it to enter the platelets^25,26 ) corresponding to an appropriate region of the β3 cytoplasmic domain missing in the VGTΔ724 platelets might be expected to inhibit the αIIb-mediated outside-in signaling induced by D3 plus fibrinogen. Accordingly we tested 2 peptides, one containing β3 sequence RKEFAKFEEER, corresponding to residues R724-R734, the other containing the β3 sequence ARAKWDTANN, corresponding to residues A735-N744. Residues R724-R734 correspond to the N-terminal 11 amino acids of the β3 cytoplasmic domain segment missing in the VGTΔ724 platelets (Figure 4). Residues A735-N744 correspond to the C-terminal 10 amino acids contiguous with RKEFAKFEEER. The β3 segment represented by these peptides was selected as a potential regulator of αIIb signaling because it is part of the membrane distal region of the β3 cytoplasmic domain missing in VGTΔ724 platelets, and because this region does not participate in maintaining αIIbβ3 complexes in the resting configuration^18-23  as is evident from the fact that the αIIbβ3 expressed on VGTΔ724 platelets is in the resting configuration.⁴ 

The results of these tests support the view that a region of the membrane distal region of β3 can negatively regulate αIIb outside-in signaling. A palmitoylated derivative of the peptide R724-R734 (pRKEFAKFEEER), but not the nonpalmitoylated peptide RKEFAKFEEER, or a scrambled control peptide containing the same amino acid composition, but a different sequence (pEAERKFERKFE), or pARAKWDTANN inhibited D3 plus fibrinogen-induced shape change or TxA₂ production and ATP secretion by VGTΔ724 platelets (Figure 4). These results demonstrate that a peptide corresponding to a specific region of the β3 cytoplasmic domain missing in the VGTΔ724 platelets can inhibit D3 plus fibrinogen-induced signaling in those platelets. The conclusion that the inhibition of the αIIb-mediated signaling results from pRKEFAKFEEER acting inside the platelets is supported by the results of a laser confocal microscopy study (Figure 5). In this study, visualization of platelets treated with either FITC-derivatized palmitoylated or nonpalmitoylated RKEFAKFEEER demonstrates that palmitoylation renders the FITC-conjugated peptide platelet permeable (Figure 5).

A hypothetical explanation for the inhibition of the αIIb outside-in signaling in the VGTΔ724 platelets demonstrated in Figure 4 is that it results from the direct interaction of the membrane distal cytoplasmic domains of β3 and αIIb. As part of a test of this hypothesis, the binding of the membrane distal cytoplasmic domain of β3 to αIIbβ3 was tested using a microtiter well binding assay.³²  The β3 peptide RKEFAKFEEER or control peptides were immobilized via their amino termini to the wells of XENOBIND plates. Intact αIIbβ3 complexes or complexes of truncated αIIbβ3 corresponding to the extracellular aspect of the receptor¹⁶  were tested for the ability to bind to the immobilized peptides. Intact αIIbβ3 bound to RKEFAKFEEER, LSARLAF,³³  an agonist peptide that binds to αIIbβ3 and causes outside-in signaling,³⁸  but not to the β3 peptide ARAKWDTANN or to the scrambled control peptides EAERKFERKFE, NNWTAARKD, and FRALSAL (scrambled versions of RKEFAKFEEER, ARAKWDTANN, and LSARLAF, respectively; Figure 6). In contrast, complexes of truncated αIIbβ3 bound only to LSARLAF, not to RKEFAKFEEER. These results demonstrate that RKEFAKFEEER binds to a cytoplasmic or transmembrane domain of αIIbβ3 but not to the extracellular aspect of αIIbβ3.

The binding assay was also used to demonstrate that the β3 peptide RKEFAKFEEER binds to the cytoplasmic domain of αIIb. This was done by incubating an αIIb peptide with the αIIbβ3 during the binding assay. The objective was to learn if the αIIb peptide could prevent the binding of αIIbβ3 to the β3 peptide. The αIIb peptide KVGFFKRNRPPLEED, which corresponds to αIIb cytoplasmic domain residues K989-D1003, was tested in this assay. Binding of αIIbβ3 to RKEFAKFEEER was inhibited by KVGFFKRNRPPLEED, but not by the corresponding scrambled control peptide FPFVGNKDKRLEREP. Also, αIIbβ3 bound to the immobilized αIIb peptide KVGFFKRNRPPLEED, but not to the immobilized scrambled control peptide FPFVGNKDKRLEREP (Figure 6). Truncated αIIbβ3 did not bind to the αIIb peptide. In the reciprocal experiment, the β3 peptide RKEFAKFEEER, but not the scrambled version EAERKFERKFE, inhibited the binding of αIIbβ3 to KVGFFKRNRPPLEED. Assuming the absence of allosteric effects, these results demonstrate that a membrane distal region of the cytoplasmic domain of β3, RKEFAKFEEER, or a sequence therein appears to be able to bind to the cytoplasmic domain of αIIb.

The results demonstrate that the β3 sequence RKEFAKFEEER can negatively regulate αIIb-mediated outside-in signaling in VGTΔ724 platelets, but they do not address the issue of the relevance of that regulation to αIIbβ3 outside-in signaling initiated by a physiologic agonist in normal platelets. That issue was resolved by evaluating the ability of the palmitoylated β3 peptide pRKEFAKFEEER to inhibit αIIbβ3-mediated outside-in signaling induced in normal platelets by γ-thrombin. Washed control platelets were treated with a level of γ-thrombin that caused secretion-dependent aggregation. Under those conditions, most of the TxA₂ production and ATP secretion that occurred in response to γ-thrombin was dependent on aggregation and the aggregation was dependent on aggregation-driven secretion (Figure 7). This relationship is evident from the ability of the antihuman αIIbβ3 mAb 7E3 to inhibit the aggregation of (Figure 7A), and about 80% of the TxA₂ production and ATP secretion by the γ-thrombin–stimulated platelets (Figure 7B). Therefore, the majority of the TxA₂ production and ATP secretion that occurred in response to this level of γ-thrombin was dependent on αIIbβ3 outside-in signaling.

The results of this investigation were unambiguous; pRKEFAKFEEER, but not the palmitoylated scrambled control peptide pEAERKFERKFE, the nonpalmitoylated β3 peptide RKEFAKFEEER, or the palmitoylated β3 peptide pARAKWDTANN inhibited aggregation (Figure 7A), TxA₂ production, or ATP secretion (Figure 7B).

As a control to demonstrate that the β3 peptide was not eliminating inside-out signaling, exogenous fibrinogen was added to the γ-thrombin, pRKEFAKFEEER-treated platelets. The expectation was that if pRKEFAKFEEER eliminated inside-out signaling, exogenous fibrinogen would not mediate aggregation. In contrast, if pRKEFAKFEEER did not eliminate the inside-out signaling, the exogenous fibrinogen should support a low level of platelet aggregation similar to that supported by D3 plus fibrinogen. Furthermore, the aggregation facilitated by the exogenous fibrinogen should not stimulate TxA₂ production and ATP secretion over and above the level stimulated by γ-thrombin in the presence of the mAb 7E3 because the outside-in signaling would be inhibited by pRKEFAKFEEER. In accordance with these expectations, aggregation of γ-thrombin–stimulated platelets was not prevented by pRKEFAKFEEER in the presence of fibrinogen, but was barely discernible in the absence of fibrinogen (Figure 7A). Also, TxA₂ and ATP secretion were not enhanced by aggregation mediated by exogenous fibrinogen in the presence of pRKEFAKFEEER (Figure 7B). Therefore, it is clear that αIIbβ3 was activated by treatment with a low level of γ-thrombin, but an insufficient level of fibrinogen was secreted to support easily discernible aggregation in the absence of outside-in signaling. These results demonstrate that the regulation of integrin-mediated signaling documented here is not unique to the VGTΔ724 platelets or to αIIbβ3-mediated signaling induced by D3 or PT25-2 and fibrinogen (Figure 7).

The results presented here extend our understanding of one aspect of the regulation of αIIbβ3-mediated bidirectional signaling by the cytoplasmic domains of αII and β3. Previous work had established that inside-out signaling requires the membrane distal region of β3^4,5,18,34  and that the membrane distal region of αIIb negatively regulates inside-out signaling that appears to be mediated by β3.^13,27  However, little had been established about the regulation of αIIb outside-in signaling.^1,9  Our results provide the new insight that the cytoplasmic domain of β3 can negatively regulate αIIbβ3-mediated outside-in signaling in normal platelets stimulated with a physiologic agonist.

The data presented in Figures 1 and 2 demonstrate that treatment of VGTΔ724, but not normal platelets, with the LIBS-specific mAbs D3 or PT25-2 in the presence of fibrinogen caused shape change, TxA₂ production, and ATP secretion. This difference in behavior occurred even though both types of platelets aggregated in response to the antibodies in the presence of fibrinogen. Furthermore, cross-linking of αIIbβ3 by the ligand apparently is necessary for the signaling induced by treatment with the antibodies because although the binding of the peptide RGDS to αIIbβ3 is enhanced by D3 treatment of the platelets,³¹  it did not cause shape change, TxA₂ production, or ATP secretion by VGTΔ724 platelets. However although necessary, cross-linking by ligand apparently is not sufficient to enable D3 to elicit the signaling described here because treatment of VGTΔ724 platelets with D3 plus fibrinogen in the absence of stirring did not elicit substantial signaling. The apparent requirement for aggregation for this signaling was confirmed by a dilution experiment (Figure 2). So, both receptor cross-linking and platelet aggregation appear to be required to elicit the signaling characterized here in the VGTΔ724 platelets.

In accordance with our hypothesis that the membrane distal region of the cytoplasmic domain of β3 negatively regulates αIIb outside-in signaling, a palmitoylated peptide corresponding to residues R724-R734 of β3 inhibited D3 plus fibrinogen-induced signaling by VGTΔ724 platelets (Figure 4). The specificity of inhibition by pRKEFAKFEEER was demonstrated by the inability of pARAKWDTANN, a palmitoylated peptide corresponding to the contiguous 10 C-terminal β3 residues A735-N744, to inhibit D3 plus fibrinogen-induced signaling by VGTΔ724 platelets or to bind to αIIbβ3.

Importantly, the negative regulation of αIIb outside-in signaling demonstrated by the data presented here is not limited to signaling by VGTΔ724 platelets. The broad physiologic significance of this regulation is demonstrated by the data in Figure 7; as with VGTΔ724 platelets, pRKEFAKFEEER, but not the control peptides, inhibited αIIbβ3 outside-in signaling induced in normal platelets by γ-thrombin. So, the negative regulation of αIIb signaling by the membrane distal region of the β3 cytoplasmic domain apparently is a normal aspect of the regulation of αIIbβ3-mediated bidirectional signaling in platelets.

The conclusion that pRKEFAKFEEER is working inside the platelets rather than externally is supported by the inability of nonpalmitoylated RKEFAKFEEER to inhibit either the D3 plus fibrinogen or the γ-thrombin–induced αIIb outside-in signaling (Figures 4 and 7) and by our demonstration that FITC-labeled pRKEFAKFEEER, but not FITC-labeled nonpalmitoylated RKEFAKFEEER, could be visualized by confocal microscopy in platelets treated independently with these peptides (Figure 5).

The results of microtiter well binding assays also support the model that pRKEFAKFEEER inhibits αIIb signaling by binding to the cytoplasmic domain of αIIb rather than by interacting with the extracellular aspect of αIIbβ3. Those assays demonstrated that the β3 peptide RKEFAKFEEER apparently can bind to the cytoplasmic domain of αIIb but not the extracellular aspect of αIIbβ3 (Figure 6). The results of the binding experiment in combination with the signaling behavior of the VGTΔ724 platelets are consistent with the model that the membrane distal region of the cytoplasmic domain of β3 can regulate αIIb outside-in signaling by directly binding to the cytoplasmic aspect of αIIb.

A number of unanswered questions were raised by our observations. For example, the inability of D3 or PT25-2 plus fibrinogen-mediated aggregation to elicit the outside-in signaling characterized here in normal platelets remains unexplained. That behavior contrasts with the ability of αIIbβ3-mediated outside-in signaling in normal platelets stimulated with a low level of γ-thrombin to enhance TxA₂ production and granule secretion (Figure 7). Even more perplexing is the observation that although receptor activation and aggregation are required for D3 plus fibrinogen to elicit signaling in the VGTΔ724 platelets, they are not necessarily sufficient to elicit signaling in those platelets. This is evident because treatment of those platelets with the reducing agent dithiothreitol plus fibrinogen³⁹  resulted in platelet aggregation, but did not elicit the signaling documented here (not shown). The reason for this unexpected behavior is not known. Further work is required to resolve these issues.

In summary and conclusion, the data presented here demonstrate that a region within the membrane distal cytoplasmic domain of β3 can bind to the cytoplasmic domain of αIIb and in that manner negatively regulate αIIb-mediated outside-in signaling in platelets.

We thank Dr Makota Handa,²⁸  Peter J. Newman,²⁹  and Barry S. Coller³⁰  for the generous gifts of PT25-2, AP3, and 7E3, respectively; Drs Timothy Springer, Junichi Takagi,¹⁶  and Aideen Mulligen for the truncated version of αIIbβ3; Sharon Frase of the University of Memphis Integrated Microscopy Center for her assistance with the microscopy; and Ann Becker for her expert technical assistance.