A Val193Met mutation in GPIIIa results in a GPIIb/IIIa receptor with a constitutively high affinity for a small ligand

Authors


Prof. Desmond Fitzgerald, Department of Clinical Pharmacology, Royal College of Surgeons in Ireland, 123 St. Stephen's Green, Dublin 2, Ireland. E-mail: dfitzgerald@rcsi.i.e.

Abstract

We have identified a patient designated as (GTa) with Glanzmann's Thrombasthenia (GT) diagnosed on the basis of a prolonged bleeding time and failure of the patient's platelets to aggregate. The number of glycoprotein (GP)IIb/IIIa receptors on the platelet surface was 37% of normal and those receptors displayed a defect in soluble fibrinogen binding. Nevertheless, GTa platelets showed increased adhesion to solid-phase fibrinogen and binding affinity for the RGD-mimetic 3H-SC52012, a non-peptide GPIIb/IIIa antagonist. Dithiothreitol (DTT) and ADP enhanced the affinity for [3H]-SC52012 in normal platelets, but had little effect in GTa platelets. These findings suggested that GTa platelets were locked in an altered affinity state. Genetic analysis showed that GTa was a compound heterozygote for the GPIIIa gene. One allele showed a deletion at the 3′ end of exon 3 resulting in a premature stop codon. The second GPIIIa allele had a G to A transition at nucleotide 577, resulting in a Val193Met substitution. HEK 293T cells transfected with mutant GPIIb/IIIaV193M bound [3H]-SC52012 with a higher affinity than wild-type GPIIb/IIIa, and this was not increased by DTT. The mutant receptor distinguishes between platelet adhesion and aggregation, and demonstrates the phenotype that may be expected when platelet aggregation alone is inhibited.

The αIIbβ3 receptor [glycoprotein (GP) IIb/IIIa] is one of the integrin family of adhesion receptors and is present in high density on the platelet surface (≈ 50 000 molecules per platelet) (Wagner et al, 1996; Quinn et al, 1999). GPIIb/IIIa is the receptor for fibrinogen and other adhesive molecules including Von Willebrand factor, vitronectin and fibronectin (Pytela et al, 1986; Plow & Ginsberg, 1989; Hantgan et al, 1990; Smith et al, 1990). Under resting conditions, GPIIb/IIIa has a low affinity for soluble fibrinogen, although it recognizes solid-phase fibrinogen (Kieffer et al, 1991; Savage & Ruggeri, 1991; Savage et al, 1992; Haimovich et al, 1993). However, upon activation of platelets by an agonist (e.g. ADP, thrombin, adrenaline), the affinity for soluble fibrinogen increases markedly (O'Toole et al, 1990; Sims et al, 1991; Du et al, 1993). The mechanism underlying receptor activation is poorly understood. However, it appears to involve a change in receptor conformation as new epitopes appear that are detected by monoclonal antibodies (Kunicki et al, 1996). This conformational change is normally induced by intracellular signals generated in response to platelet agonists. However, the activation state can be mimicked by treating GPIIb/IIIa with the reducing agent dithiotreitol (DTT) (Zucker & Masiello, 1984; Zucker & Mauss, 1986; Peerschke, 1995). The receptor undergoes further conformational changes upon ligand binding that are also detected by monoclonal antibodies (e.g. D3GP3) and are called Ligand Induced Binding Sites or LIBS (Kouns et al, 1990). LIBS have been linked to post-occupancy intracellular signals, called ‘outside-in’ signalling (Honda et al, 1998). Anti-LIBS antibodies such as D3GP3 force an active conformation in GPIIb/IIIa, induce a low level of platelet aggregation and enhance platelet adhesion (Frelinger et al, 1991). Thus, it appears that GPIIb/IIIa assumes multiple conformations that are important for function.

Inherited qualitative and/or quantitative abnormalities of the GPIIb/IIIa receptor complex give rise to Glanzmanns Thrombasthenia (GT), which is a recessive haemorrhagic disorder. Mutations in either GPIIb or GPIIIa may give rise to the GT phenotype. Studies of such naturally occurring mutations are invaluable in deciphering the structure–function relationship of GPIIb/IIIa. In this paper, we describe a patient with GT arising from different mutations on each allele of the GPIIIa gene. One mutation gives rise to an unstable mRNA transcript that is rapidly degraded and results in a reduced number of receptors expressed on the platelet surface. The second mutation is a G to A transition at nucleotide 577, resulting in a Val193Met substitution. This mutation gives rise to a receptor that is expressed but fails to support platelet aggregation despite exhibiting enhanced affinity for solid-phase fibrinogen and the RGD-mimetic 3HSC52012, a GPIIb/IIIa antagonist.

Patients and methods

Patient information GTa, a 37-year-old female Irish Caucasian, presented initially with excessive bleeding after a normal delivery and a history of excessive bruising. There was no family history of any bleeding disorder. The Simplate bleeding time on two occasions was 16 and 17 min respectively (normal < 10min). Platelet aggregation to ADP, thrombin receptor activating peptide (TRAP), thrombin and U46619 was < 10% and platelet count was 161 × 109/l. Clot retraction was 50% (normal 48–64%) and ristocetin-induced agglutination was normal.

Materials Reagents were all of analytical grade. Dithiothreitol (DTT), adenosine diphosphate, collagen, arachidonic acid and Sepharose 2B-300 were obtained from Sigma, Aldrich, Ireland). Thrombin was obtained from Diagnostic Reagents, Oxon, UK. U44619 was obtained from Cayman Chemical, Ann Arbor, MI, USA. TRAP was a generous gift from Dr Patrick Harriot, Queens University, Belfast, UK. The RGD mimetic [3H]-SC52012 was a kind gift from Dr Nancy Nicholoson, Searle, Skokie, IL, USA. Oregon green fibrinogen was obtained from Molecular Probes, Oregon, USA.

Antibodies Anti-P-selectin and anti-GPIIb/IIIa (CD41) antibodies were obtained from Immunotech, Marseilles, France. The LIBS antibody anti-D3GP3 was a kind gift from Dr Lisa Jennings, University of Tennessee, Memphis, TN, USA. Purified platelet activation marker (PAC-1) and fluorescein isothiocyanate (FITC)-labelled secondary antibodies were obtained from Becton Dickinson, Oxford, UK. The platelet GPIIb/IIIa receptor occupancy kit (anti-GPIIIa Mab L18) was obtained from Biocytex, Marseilles, France.

Platelet isolation Venous whole blood was drawn into 0·1 volume 3·8% sodium citrate. Platelet-rich plasma (PRP) was obtained by centrifugation of the whole blood at 180 g for 12 min at room temperature. After removal of PRP, platelet-poor plasma (PPP) was obtained by centrifuging the remaining blood at 1000 g for 10 min at room temperature.

Gel-filtered platelets were obtained as described (Mendelsohn et al, 1991). Venous whole blood was collected into 0·15 Vol. ACD (38 mmol/l citric acid anhydrous, 75 mmol/l sodium citrate, 124 mmol/l dextrose) and centrifuged to obtain PRP. PRP was acidified to pH 6·5 with ACD and PGE1 (1µmol/l) was added. Platelets were pelleted through the plasma by centrifugation at 1000 g for 12 min at room temperature. Supernatant was removed and platelets were re-suspended in buffer A (130 mmol/l NaCl, 10 mmol/l trisodium citrate, 9 mmol/l NaHCO3, 6 mmol/l dextrose, 0·9 mmol/l MgCl2, 0·81 mmol/l KH2PO4, 10 mmol/l Tris pH 7·35). Re-suspended platelets were applied to a Sepharose 2B-300 column equilibrated with buffer A and filtered. Calcium chloride at a final concentration of 1·8 mmol/l was added to the filtered platelets.

Platelet aggregation Platelet aggregation was monitored in an aggregometer (PAP-4, BioData, Horsham, PA, USA). PRP was aliquoted into siliconized glass tubes in a final volume of 50μl, and stirred at 1100 r.p.m. at 37°C. A baseline for PRP aggregations was obtained using PPP. Platelets were activated upon addition of agonists (0·1 U/ml thrombin, 2 µmol/l U46619, 2 µmol/l TRAP, 2 µmol/l ADP, 5 µmol/l adrenaline, 0·2 mg/ml collagen). Platelet aggregation was monitored for 4 min.

Platelet shape change Shape-change experiments were performed as described previously (Takahara et al, 1990). Briefly, the 0% and 100% limits of the platelet aggregometer calibration curve were set using undiluted PRP and PPP respectively. PRP (240 μl) was diluted in an equal volume of a modified Tyrode's buffer containing 5 mmol/l EDTA. After the addition of 2 µmol/l U46619, no aggregation was observed because of the presence of EDTA, but shape change registered as a transient decrease in light transmission.

Platelet adhesion Gel-filtered platelets were prepared as described above. Platelet adhesion to fibrinogen was studied in 96-well polystyrene plates (Sarstedt, Wexford, Ireland). The platelet adhesion protocol was as previously described (Haimovich et al, 1993). Briefly, 50μl aliquots of platelets, diluted to 3 × 108/ml in buffer A, were dispensed into 96-well plates that had been coated with fibrinogen (50μl of a 20µg/ml solution) for 1 h at room temperature and post-coated with bovine serum albumin (BSA) (1 mg/ml). Platelets were allowed to adhere for 60 min at room temperature. Plates were washed extensively with ice-cold buffer A. Adhered platelets were quantified using a crystal violet protein stain and compared with a standard curve to estimate total adhesion. Parallel adhesion assays were performed in 96-well plates coated only with BSA as an estimate of non-specific binding. Data are expressed as absorbance of crystal violet at 590 nm.

Oregon green fibrinogen binding A 1:10 dilution of PRP was made in buffer A supplemented with 1·8 mmol/l CaCl2. Aliquots (100μl) were incubated with 20µg/ml Oregon green fibrinogen in the presence and absence of 20µmol/l ADP for 20min at room temperature in the dark. Volumes were increased to 1 ml in buffer A and analysed using flow cytometry.

Radio-ligand binding studies Ligand binding to control and activated platelets was determined using the radiolabelled RGD mimetic [3H]-SC52012 (specific activity of 37GBq/mmol). Binding assays were performed using a filtration technique to separate free from bound ligand. Gel-filtered platelets were diluted to a concentration of 0·5 × 108/ml in buffer A. Platelets were activated with either ADP (20 µmol/l or DTT (20 mmol/l) for 5 min at 37°C before the addition of ligand. Aliquots (100μl) of control or activated platelets were incubated with increasing concentrations of [3H]-SC52012 from 1 to 500 nmol/l, in triplicate in the presence or absence of excess unlabelled SC-52012 (10 µmol/l). Incubations were performed in a final volume of 250μl at room temperature for 30 min. Free ligand was separated from bound ligand by filtration through a Whatman GFB filter in a cell harvester (Brandel M-24RP, Gaithersburg, MD, USA). Filters were rapidly (10 s) washed three times with 4 ml of ice-cold 10 mmol/l Tris pH 7·4. Filter discs were counted in 5 ml of Ecoscint in a beta-liquid scintillation counter (LKB Wallac 1214, Wallac Oy, Finland). Specific binding was determined by subtracting counts obtained in the presence of excess unlabelled ligand from the equivalent counts obtained in the absence of competitor. Data were analysed and plotted using Deltagraph Pro® 3·0.4.

Single-point [3H]-SC52012 binding studies were performed on transfected Human Embryonic Kidney (HEK)293T cells in a similar manner. Briefly, transfected HEK293T cells (9 × 104) were incubated in the presence and absence 20 mmol/l DTT (Sigma) for 5 min at 37°C. Control and DTT-treated cells were incubated with 50 nmol/l [3H]-SC52012, in triplicate for 30 min at room temperature. One set of cells contained excess unlabelled ligand at a final concentration of 10 µmol/l to estimate non-specific binding and the results were corrected for expression efficiency.

Flow cytometry analysis A 1:10 dilution of PRP was made in phosphate-buffered saline (PBS) containing 1% BSA and samples were treated with agonists or 5 mmol/l EDTA as indicated. Aliquots (50μl) of each sample were added to 20µg/ml of the indicated primary antibody and incubated at room temperature for 20 min. Each sample was washed with 1 ml of PBS containing 1% BSA, and platelets were pelleted at 10 000 g for 3 min at room temperature. The supernatant was removed and the platelets were resuspended in 100μl of an FITC-labelled secondary antibody, and incubated at room temperature for a further 20 min. As a negative control, samples were incubated with secondary antibody alone to determine background FITC non-specific binding. Each sample was washed with 1 ml of PBS containing 1% BSA, and pelleted a second time. Finally, platelets were re-suspended in 1 ml of PBS containing 1% BSA and 1% formaldehyde and analysed on a FACS flow cytometer. GPIIb/IIIa surface expression on transfected HEK293T cells were also measured using flow cytometry.

Purification of platelet RNA and leucocyte DNA PRP was prepared from ACD anticoagulated blood (60 ml) by centrifugation at 280 g for 20 min. PRP was centrifuged at 130 g for 20 min to remove peripheral white blood cells before pelleting the platelets at 2500 g. Total RNA was prepared from the platelet pellet using 2 ml of Tri-Reagent (Sigma, St. Louis, MO, USA), according to the manufacturer's instructions. High-molecular-weight DNA was prepared from peripheral white blood cells by a salting-out procedure (Miller et al, 1988).

Genetic analysis of the patients β3 subunit from genomic DNA and platelet mRNA Total RNA isolated from Tri-Reagent was reverse transcribed into cDNA, and this was used as template for the polymerase chain reaction (PCR). Briefly, RNA samples of 1µg were denatured at 65°C for 10 min. Random hexamers 100 ng, 20 units of RNase inhibitor (Boehringer, Mannheim, Germany), 1 × reverse transcription buffer, DTT 0·1 mol/l, deoxyribonucleoside triphosphates 25 µmol/l and 10 units of Moloney murine leukaemia virus reverse transcriptase (Promega, Madison, WI, USA) were added and incubated at 37°C overnight in a reaction volume of 20 μl. The reverse transcription (RT) reaction was stopped by heating to 95°C for 5 min. A 1:10 volume of the generated cDNA reaction was used in subsequent amplification reactions. Oligonucleotide primers for GPIIIa-specific amplification were synthesized by Sigma-Genosys Biotech, Cambridge, UK. The GPIIIa-specific primers for RT-PCR and conditions were as described previously (Grimaldi et al, 1996). Either high fidelity Pfu Turbo DNA polymerase (Stratagene, La Jolla, CA, USA) or AccuTaq (Sigma, Aldrich, Ireland) enzyme was used in the PCR amplification. Samples were analysed on a 0·8% ethidium-stained agarose gel. Negative controls were carried out in parallel with the GPIIIa-specific PCRs. PCR amplification of the patients genomic DNA was carried out using the following primer pairs; exon 3 sense 5′-AAC TAA CAT CTT CGT GCC TTC CA-3′ and exon 3 antisense 5′-GCG TCT GGA GGA GGG ACT TAC-3′, exon 4 sense 5′-GAA GAC CAC CTG CTT GCC CAT G-3′ and exon 4 antisense 5′-GAC TGT AGC CTG CAT GAT GGC-3′. These PCRs were carried out using AccuTaq (Sigma Aldrich, Ireland) and the parameters were initial denaturation at 94°C for 1 min followed by 30 cycles of 98°C for 20 s, 60°C for 30 s, 68°C for 30 s. A final elongation step of 68°C for 7 min was used before rapid cooling to 4°C. The overlapping PCR fragments generated by the RT-PCR of the patient's cDNA were subcloned into the pCR-Blunt vector (Invitrogen, The Netherlands). Plasmid DNA (six individual colonies for each fragment) was prepared for sequencing using Wizard Plus SV Mini-prep DNA Purification System (Promega). Cloned inserts were sequenced using BigDye Terminator cycle sequencing chemistry (Perkin-Elmer Applied Biosystems, Warrington, UK). Sequencing of cloned inserts was carried out using M13 forward and reverse sequencing primers (Genosys Biotech., Cambridge, UK). Mutation verification was carried out using direct PCR sequencing on the products generated from genomic DNA according to the manufacturer's protocol. PCR products used for direct sequencing were purified and concentrated using Microcon 100 microconcentrators (Amicon, Beverly, MA, USA). Sequence analysis was performed on an ABI Prism 310 Genetic Analyser using capillary electrophoresis. All sequencing reactions were carried out using the BigDye Terminator cycle sequencing chemistry as it is optimal for heterozygote calling of mutations.

Site-directed mutagenesis of β3 CDNA A single nucleotide substitution (G577A) corresponding to that found in the patients β3 subunit was generated in the GPIIIa cDNA construct in a mammalian cell expression vector, pcDNA3·1 (Invitrogen). The mutation was constructed using the Quick Change Site-Directed Mutagenesis Kit (Stratagene) according to manufacturer's instructions. The mutagenic primers were as follows; sense primer 5′-GGC TAC AAA CAC ATG CTG ACG CTA ACT GAC C−3′ and antisense primer 5′-GGT CAG TTA GCG TCA GCA TGT GTT TGT AGC C−3′. Modifications to the protocol were as follows; 200 ng of GPIIIa cDNA was subjected to 18 cycles of PCR and the resultant PCR product digested for 2 h by incubation with DpnI at 37°C. Next, 2μl mutant cDNA constructs were transformed into XL1-Blue supercompetent cells (Stratagene). Plasmid DNA was isolated and subjected to automated sequencing to verify introduction of the mutation.

Cell culture and transfection HEK 293T cells were transiently co-transfected with wild-type pcDNA3-GPIIb and pcDNA3-GPIIIa or with wild-type pcDNA3-GPIIb and mutant pcDNA3-GPIIIaVal193Met. Unligated pcDNA-3 was used as a mock control with all transfections. Briefly, GPIIb and GPIIIa cDNA constructs (4·5µg of each) were mixed with 120µg LipofectAmine reagent (Gibco, Life Technologies, Paisley, UK) in Optimem (1·6 ml) for 15 min in the dark. The transfection mixture was added to HEK 293T cells and incubated at 37°C for 5 h. Media were removed and replaced with minimal essential medium (MEM; Gibco, Life Technologies) containing 20% FBS. Cells were harvested after a 60 h incubation at 37°C, 5% CO2. GPIIb/IIIa surface expression was measured using flow cytometry. Cells (9 × 104) were incubated with 10μl of CD41a (Gibco, Life Technologies), a monoclonal antibody specific for the GPIIb/IIIa receptor complex. The samples were analysed with a FACSCalibur flow cytometer (Becton Dickinson, Mountain View, CA, USA) using cellquest software.

Results

Platelet aggregation and GPIIb/IIIa receptor density

Platelets isolated from GTa failed to aggregate to ADP (2 µmol/l, 20 µmol/l), TRAP (2 µmol/l), thrombin (0·1 U/ml), the thromboxane mimetic, U46619 (2 µmo/l) or collagen (10µg/ml) (Fig 1A). However, platelet-shape change (Fig 1C, inset) and ristocetin-induced platelet agglutination (not shown) were normal. Similarly, secretion from dense granules as measured by P-selectin expression on the platelet surface after agonist stimulation was normal (Fig 1C). Quantification of platelet surface GPIIb/IIIa using the monoclonal antibody CD41 revealed levels that were 37% of normal (Fig 1D) and this was confirmed using L18 (data not shown).

Figure 1.

Patient platelets failed to aggregate to ADP 2 µmol/l and TRAP 2 µmol/l (A) compared with control, normal platelets (B). Shape change (inset C) was measured as the change in percentage light absorption in response to agonists in the presence of 5 mmol/l EDTA. The magnitude of the response to 2 μmol/l U46619 is similar for control (upper trace) and patient (lower trace) platelets. The secretary response (C) was measured in PRP while stirring in the presence or absence of TRAP (10 µmol/l) for 3 min at 37°C. The samples were incubated with an antibody to CD62P and analysed using flow cytometry. The data are expressed as mean ±S.E.M. of two experiments performed in duplicate. Quantification of GPIIb/IIIa on the surface of patient platelets with CD41a antibody showed a reduction in GTa of > 50% (D). The data are expressed as mean ±S.E.M. of four independent experiments.

Oregon green fibrinogen binding

GTa platelets showed markedly reduced binding of soluble fibrinogen upon treatment with ADP when compared with controls (Fig 2). Mean fluorescence of the control platelets, as a consequence of Oregon green fibrinogen binding increased when treated with ADP, whereas there was little increase in GTa platelets. GTa platelets also showed decreased binding of PAC-1, a monoclonal antibody that binds to the active form of GPIIb/IIIa. The mean fluorescence of control platelets incubated with PAC-1 increased from 1·45 to 23·77 when stimulated with ADP, whereas GTa platelets mean fluorescence increased from 1·33 to 2·4.

Figure 2.

Binding of soluble fibrinogen to control and patient platelets. The mean fluorescence of control platelets incubated with Oregon green fibrinogen increased when stimulated with ADP, while remaining unchanged in patient platelets.

Radioligand binding to platelets

GPIIb/IIIa protein expression on GTa platelets was also quantified using equilibrium binding studies with a tritiated ligand, [3H]-SC52012, a specific GPIIb/IIIa antagonist. In normal platelets, the Kd for [3H]-SC52012 was 108 nmol/l and the Bmax or number of binding sites was 38 860 binding sites per platelet (Fig 3A). In GTa platelets, the affinity for [3H]-SC52012 was markedly increased with a Kd of 23 nmol/l and the Bmax was reduced at 26 387 sites per platelet (Fig 3b). The affinity of [3H]-SC52012 binding to normal washed platelets was increased after activation with either ADP 20 µmol/l or DTT 20 mmol/l (for 5 min at 37°C), the Kd falling to 31·3 nmol/l and 11·5 nmol/l respectively. Thus, the affinity of GTa platelets at rest was equivalent to normal platelets activated with ADP. In GTa platelets, ADP and DTT reduced the affinity constant further to 10·3 nmol/l and 16 nmol/l respectively. Treatment of platelets with ADP had no effect on the number of binding sites detected by [3H]-SC52012 in either normal or GTa platelets. In normal donors, treatment with DTT increased the Bmax, whereas in GTa platelets, Bmax decreased.

Figure 3.

Ligand binding studies: PRP was incubated with PBS, ADP (10 µmol/l) or DTT (20 mmol/l). After 5 min at 37°C, increasing concentrations of the ligand [3H]-SC52012 ranging from 1 to 500 nmol/l, were added and equilibrium binding of the ligand assayed after 30 min. Control data represent the mean of three experiments, whereas the patient platelet data represent the specific binding per sample of one experiment performed in triplicate. (A) Control platelets; (B) patient platelets.

Platelet adhesion studies

Despite the lower platelet density of GPIIb/IIIa, GTa platelets showed an equal or enhanced adhesion response to immobilized fibrinogen (Fig 4). In normal platelets, direct activation of the platelet GPIIb/IIIa with D3GP3 enhanced platelet adhesion (Mondoro et al, 1996). In contrast, adhesion of GTa platelets was slightly reduced by pre-treatment with D3GP3 (Fig 4). Similarly, ADP-induced activation enhanced the adhesion of normal, but not GTa, platelets (data not shown).

Figure 4.

Adhesion of platelets to immobilized fibrinogen: Gel-filtered platelets were diluted to 3 × 108/ml and plated on fibrinogen-coated polystyrene tissue culture plates for 60 min at room temperature in the presence and absence of the monoclonal antibody D3GP3 (10 µg/ml) or EDTA (5 mmol/l) as indicated. The data represent absorbance at 590 nm and are expressed as mean ±S.E.M. of three experiments performed in triplicate.

Genetic studies

Total RNA isolated from GTa platelets was used as template for RT-PCR. Seven primer pairs were used to amplify over-lapping regions of the entire GPIIIa-coding sequence, using cDNA as template (Miller et al, 1988). The PCR-amplified products from the patient were cloned and sequenced. Only one mutation was identified, and this was a G to A transition at nucleotide 577, resulting in a Val193Met substitution in the GPIIIa subunit. The presence of this mutation was confirmed using sequence analysis of genomic DNA. The G to A substitution in exon 4 was verified, and the patient was found to be heterozygous for this mutation. Sequencing of exon 3 revealed in 50% of clones an adenosine deletion at the 3′ end of exon 3, resulting in a premature stop codon at the beginning of exon 4. Direct PCR sequencing of this exon showed that the mutation was heterozygous. This deletion was not detectable on sequencing of the cDNA, presumably, as it results in an unstable mRNA transcript caused by premature loss of ribosomal protection. Thus, genetic analysis of the patient identified her as being a compound heterozygote with two distinct mutations, one in exon 3 and the other in exon 4 of the GPIIIa-coding sequence.

Transient expression of mutant GPIIb/IIIa receptor complexes in HEK293T cells

Mutations resulting in a premature stop codon at the extreme 5′ portion of a gene leads to a message that is not translated into protein as the mRNA transcript is unstable. Thus, the mutation in exon 3 may be responsible for the quantitative defect of GPIIb/IIIa on the platelet surface. On the other hand, the mutation in exon 4 of GPIIIa is likely to be expressed, and may be responsible for the functional defect. In order to determine if the Val193Met substitution was responsible for Glanzmann's thrombasthenia in this patient, a GPIIIa cDNA mutant construct was generated with a G to A transition at nucleotide 577 (GPIIIaVal193Met). The GPIIIaVal193Met construct was transiently co-transfected with wild-type GPIIb into HEK 293T cells to generate GPIIb/IIIa-GT. Mock transfected cells showed no expression of GPIIb/IIIa as was expected. Fluorescence analysis showed equal expression of wild-type and mutant receptors (Fig 5). At a low concentration of [3H]-SC52012, GPIIb/IIIa-GT transfected cells bound higher levels of the radioligand than HEK cells expressing the wild-type receptor (Fig 5). This is consistent with GTa platelets having a higher ligand affinity. Moreover, although [3H]-SC52012 binding increased in cells expressing wild-type GPIIb/IIIa exposed to DTT, treatment of GPIIb/IIIa-GT expressing cells with DTT did not give rise to an increase in [3H]-SC52012 binding.

Figure 5.

3H-SC52012B binding to HEK 293 cells transiently transfected with wild-type GPIIb/IIIa and GPIIb/IIIa-GT in the absence (closed bars) or presence (open bars) of 20 mmol/l DTT. The data represent the mean ±S.E.M. of nine separate experiments performed in duplicate. The inset shows the FACScan analysis of HEK cells showing equivalent binding of CD41a antibody to cells expressing either isoform.

Discussion

Mutations in either the GPIIb or GPIIIa responsible for GT lead to defects in mRNA splicing (Burk et al, 1991; Newman et al, 1991; Iwamoto et al, 1994; Jin et al, 1996), mRNA stability (Kato et al, 1992), divalent cation binding (Poncz et al, 1994; Basani et al, 1996), subunit association (Bajt et al, 1992; Lanza et al, 1992; Tao et al, 2000), intracellular trafficking (Wilcox et al, 1994), ligand binding (Loftus et al, 1990; Ward et al, 2000) and integrin-mediated signal transduction (Wang et al, 1997). In the Strasbourg variant of GT (GPIIIa Arg214Tyr), the receptor retained the ability to bind RGD ligands but did not recognize fibrinogen (Lanza et al, 1992). We describe a novel mutation in the GPIIIa gene, giving rise to a phenotype in which the receptor recognizes solid-phase fibrinogen and a small ligand antagonist, but does not facilitate platelet aggregation. This is caused by an inability to recognize the soluble form of fibrinogen after platelet activation with an agonist. Normal platelets adhere to solid-phase fibrinogen without the need for prior activation, presumably because GPIIb/IIIa recognizes a change in the conformation of fibrinogen when plated (Parise et al, 1993). Thus, despite the reduced expression of GPIIb/IIIa and the failure to aggregate normally, the patient has a partially functioning receptor capable of recognizing a natural ligand.

The adhesion of GTa platelets to fibrinogen was enhanced compared with normal platelets in the absence of any activation despite lower expression. Previous activation of the receptor with D3GP3 increased the adhesion of normal platelets, but not GTa platelets. These findings suggested that the receptor was already in a high-affinity state for some ligands. Consistent with this hypothesis, the binding affinity of [3H]-SC52012 (a GPIIb/IIIa antagonist) was increased in GTa platelets, and was equivalent to normal platelets activated with ADP. Previous studies have demonstrated a marked increase in GPIIb/IIIa affinity for small ligands and fibrinogen upon exposure to DTT, presumably because of the reduction of disulphide bonds. Although all of the cysteine pairings have not been mapped, DTT-induced receptor activation is mimicked by disrupting the disulphide linkage between cysteine residues 5 and 435 in GPIIIa (Liu et al, 1997). Treatment of GTa platelets with DTT induced only a minor shift in receptor affinity for [3H]-SC52012 and failed to increase the number of binding sites, as in normal platelets. Similarly, ADP induced only a small increase in the affinity for [3H]-SC52012. Thus, the GPIIb/IIIa expressed on GTa platelets appears to be in a constitutively high-affinity state for certain ligands. The mechanism underlying the increase in ligand binding capacity by DTT in normal platelets is not understood, but may be caused by conformational changes resulting in the exposure of further binding sites as recent studies in our laboratory have shown a potential role for thiol bonds in the mechanism of GPIIb/IIIa activation (O'Neill et al, 2000).

Genetic studies focused on the GPIIIa as defects in this subunit often give rise to a defective receptor, whereas mutations in GPIIb generally effect receptor expression (Tao et al, 2000). Indeed, two mutations in GPIIIa were found. First, there was a deletion in the extreme 5′ portion of the GPIIIa-coding region that results in a frame shift and premature stop codon at the 5′ intron C-exon 4 splice donor site, MET180OPA. The unstable mRNA that arises from such a defect is inefficiently protected by ribosomes and rapidly degrades. Consequently, this mutation was not detected at the mRNA level. This mutation would be expected to reduce the expression a GPIIb/IIIa complex on the platelet surface. However, the patient's platelets did express a GPIIb/IIIa receptor complex albeit at a lower than normal density. This implied that there was a second mutation underlying the qualitative defect. Screening of the entire GPIIIa gene at both the DNA and RNA (cDNA) level identified a heterozygous G to A transition at nucleotide 577 in the second GPIIIa allele, resulting in a Val193Met substitution. This amino acid substitution is a subtle change as both valine and methionine are small hydrophobic residues, valine with an aliphatic side chain and methionine containing a sulphur atom in a thioether linkage.

To confirm that the Val193Met substitution was responsible for the functional defect, this mutation was co-expressed with wild-type GPIIb in HEK 293T cells. Flow cytometric studies showed expression of the mutant receptor was similar to that of the wild-type receptor. However, the expressed mutant receptor exhibited an increased affinity for [3H]-SC52012, and this was not enhanced by DTT. In contrast, the reducing agent increased radioligand binding to the expressed wild-type receptor. Thus, the expressed mutant GPIIb/IIIa mimicked the platelet studies showing a receptor with an enhanced affinity for a small ligand that was insensitive to DTT.

The Val193Met substitution lies adjacent to the calcium-binding MIDAS-like domain of GPIIIa (residues 118–131), an area that appears to influence receptor affinity (D'Souza et al, 1994). Mutations in this region do not affect the expression of GPIIb/IIIa on the cell surface, but inhibit the interaction of the receptor complex with ligands (Bajt & Loftus, 1994). In an adjacent region, a naturally occurring mutation (GPIIIa Arg214Gln) inhibits the ability of platelets to aggregate and the resulting GPIIb/IIIa is defective in binding multiple ligands (Bajt et al, 1992). Peptides derived from this loop (211–221), in particular RNRDA (214–218), inhibit fibrinogen binding to purified GPIIb/IIIa. These peptides interact with the receptor, not the ligand, suggesting that the region may play a role in the folding necessary to assume a high-affinity conformation. Indeed the region overlaps with Asp217 and Glu220, which contribute to cation binding. Interestingly, mutation of Arg214 to a Trp prevents fibrinogen binding although the receptor can still interact with an RGD peptide, consistent with the GT patient's phenotype (Lanza et al, 1992: Djaffar & Rosa, 1993). It is worth noting also that a Leu262Pro mutation results in a receptor capable of binding to fibrin but not to fibrinogen (Ward et al, 2000). Thus, this region of GPIIIa appears to be involved in regulating the accessibility of fibrinogen to GPIIb/IIIa after platelet activation.

The discrimination between platelet adhesion and aggregation is a major goal in the development of GPIIb/IIIa antagonists on the basis that preservation of adhesion may translate into a lower bleeding risk. The assumption is that platelet aggregation plays an important role in pathological thrombosis, whereas adhesion is important for haemostasis. Indeed, GTa had an intermediate bleeding time, whereas the available GPIIb/IIIa antagonists which inhibit platelet adhesion in addition to aggregation often prolong the bleeding time to > 30 min. Despite the modest prolongation of the bleeding time, the patient exhibited a bleeding diathesis, with easy bruising and serious blood loss during delivery. Thus, the findings suggest that platelet aggregation plays a role in haemostasis independently of adhesion. The results also suggest that clot retraction (normal in GTa), is dependent largely on adhesion to fibrinogen, as clot retraction is abolished when the receptor is absent.

In conclusion, the GPIIIa(Val193Met) mutation results in the expression of a GPIIb/IIIa receptor that exhibits a high affinity for solid-phase fibrinogen and an RGD mimetic, but does not support platelet aggregation. This inability to aggregate is caused by a defect in the binding of soluble fibrinogen. Despite the normal adhesion, the patient had a bleeding diathesis, evidence of a functional role for platelet aggregation (independent of platelet adhesion) in haemostasis.

Acknowledgments

The GPIIb and GPIIIa cDNA expression constructs were a generous gift from P. J. Newman. We would like to thank the Blood Research Institute, Milwaukee, WI, USA, for their invaluable technical advice and the patient for her time and patience. This work was supported by grants from the Higher Education Authority of Ireland (D.J.F.), the Wellcome Trust (N.M., D.J.F.), the Irish Heart Foundation (R.M.) and Enterprise Ireland (J.F.). R.M. was a Health Research Board of Ireland Research Fellow during the performance of this work.

Ancillary