Dr Alan T.Nurden UMR 5533 CNRS, Hôpital Cardiologique, 33604 Pessac, France.
Glanzmann's thrombasthenia (GT) results from a qualitative or quantitative defect of GPIIb–IIIa complexes (integrin αIIbβ3), the fibrinogen receptor on platelets. This integrin plays a critical role in platelet aggregation. In this report we describe the molecular abnormalities of a patient with clinical and laboratory findings typical of type I Glanzmann's thrombasthenia. SDS-PAGE with Western blotting revealed an absence of GPIIb but small amounts of normally migrating GPIIIa in his platelets. A non-radioactive PCR-SSCP procedure and direct sequence analysis of PCR-amplified DNA fragments showed the patient to be a compound heterozygote for mutations in the GPIIb gene. A single point mutation (G to A) at nucleotide 1064 of the cDNA derived from the mother's allele led to a Glu324 to Lys amino acid substitution in GPIIb. It was responsible for a MscI restriction site in exon 12 of the GPIIb gene. This amino acid substitution changes the electric charge between the second and third Ca++-binding domains of GPIIb. The second mutation was inherited from his father and is in exon 18 of the GPIIb gene. It was a T → C base transition at position 1787 of GPIIb cDNA and results in a Ile565 to Thr substitution. The two GPIIb mutations identified in this study will provide new information on GPIIb–IIIa structure and biosynthesis.
The autosomal recessive bleeding disorder Glanzmann's thrombasthenia (GT) is characterized by an absence of platelet aggregation linked with a quantitative or qualitative defect of the fibrinogen receptor on the platelet surface: the GPIIb–IIIa complex (αIIbβ3) ( George et al, 1984 , 1990; French, 1998). Type I patients either lack GPIIb–IIIa or have only trace amounts (<5% of normal levels), whereas type II platelets express reduced, but more easily detectable, levels of this complex. Type III, the so-called ‘variant’ forms of thrombasthenia, are characterized by having half to normal levels of a dysfunctional form of GPIIb–IIIa on the platelet surface. Platelets from GT patients either fail to bind adhesive protein ligands or bind them in insufficient amounts to allow platelet aggregation, and the defect extends to all physiological agonists. The structure and organization of the GPIIb and GPIIIa genes is now known ( Poncz et al, 1987 ; Heidenreich et al, 1990 ; Zimrin et al, 1990 ; Wilhide et al, 1997 ). These advances have enabled the detection of molecular abnormalities giving rise to Glanzmann's thrombasthenia and, over the last few years, more than 30 different genetic defects have been described that result in the thrombasthenic phenotype. Deletions, insertions, splice defects and point mutations have been identified in both the GPIIb and GPIIIa genes (see Bray, 1994; French, 1998). The majority of these lesions give rise to single amino acid substitutions, small deletions or truncations. Analysis of these molecular defects has led to an improved understanding of the structural requirements necessary for GPIIb–IIIa subunit biogenesis, as well as receptor maturation, cell surface expression, and ligand binding function.
We now report the results of our studies on a Swiss patient with type I Glanzmann's thrombasthenia whose platelets had no detectable GPIIb and a low content of GPIIIa as determined by SDS-PAGE and Western blotting. Using a PCR-SSCP procedure and direct sequence analysis of PCR-amplified DNA fragments, we have examined all exons and splice sites of the GPIIb and GPIIIa genes and identified the patient as a compound heterozygote for two apparently rare mutations in the GPIIb gene.
MATERIALS AND METHODS
The patient is a 5-year-old son of non-consaguinous Swiss parents. They have no other children. Diagnosis was based on the following criteria: (i) frequent mucocutaneous bleeding with transfusion being required shortly after birth; (ii) an absence of platelet aggregation in response to ADP, epinephrine, thrombin and collagen; (iii) normal platelet agglutination with ristocetin; and (iv) < 2% of the normal platelet content of GPIIb–IIIa as determined by flow cytometry using the monoclonal antibodies 10E5 ( Coller et al, 1983 ) and 7E3 ( Coller et al, 1986a ) in standard procedures. Blood samples for the studies described below were obtained from the patient, his parents and five other members of his family (see Fig 1). None of the other studied family members have experienced abnormal bleeding. Controls were volunteers from hospital staff. Informed consent was obtained.
Platelet isolation and Western blot analysis
Washed platelets were prepared from acid–citrate–dextrose anticoagulated blood and samples prepared for SDS–polyacrylamide gel electrophoresis as previously described ( Jallu et al, 1994 ). Solubilizing buffer was 10 m M Tris-HCl, pH 7.2, 130 m M NaCl, 5 m M EDTA, 1% w/v SDS, 5 m MN-ethylmaleimide and contained the protease inhibitors PMSF, leupeptin and benzamidine, platelets were suspended at 5 or 2 × 109 platelets/ml. Platelet proteins (5 μg) from the patient, family members, and controls, were applied to 7–12% gradient acrylamide gels and the proteins electrophoretically transferred to nitrocellulose membrane. Membranes were saturated with buffer (pH 8.3) containing 5% (w/v) nonfat milk, 154 m M NaCl, 0.05% (v/v) Tween 20 and 20 m M Tris-base, and the presence of GPIIb and GPIIIa was evaluated by Western blotting performed using monoclonal antibodies (mAbs) to GPIIb (SZ22, Immunotech, Marseille, France) or GPIIIa (XIIF9, our laboratory) at 0.5 and 0.1 μg/ml respectively. Bound murine mAb was detected using peroxidase-labelled anti-mouse IgG (Jackson ImmunoResearch, West Grove, Pa., U.S.A.) and a chemiluminescence procedure (Amersham-France, Les Ulis, France) ( Jallu et al, 1994 ).
Isolation of genomic DNA
Genomic DNA was isolated using a QIAamp blood kit (QIAGEN Corporation, Chatsworth, Calif., U.S.A.) according to the manufacturer's instructions from whole blood (9 vol) taken into EDTA anticoagulant (1 vol). Samples (10 μl) were analysed on 0.7% (w/v) agarose dissolved in electrophoresis buffer containing 50 m M Tris base, 50 m M boric acid, 1 m M EDTA, pH 8.3 (0.5 × TBE buffer). Ethidium bromide (1 μg/ml) was included in the gel. To the samples were added a 1:5 vol of loading buffer composed of 10% (w/v) Ficoll 400, 0.25% (w/v) bromophenol blue, 0.25% (w/v) xylene cyanol FF, 0.4% (w/v) Orange G, 50 m M EDTA, 10 m M Tris-HCl, pH 7.5. Migration was for 1 h at 50 V. The band corresponding to purified DNA was detected using UV light and imaged on PolaroidTM film (Poly Labo, Strasbourg, France). In order to assess the size of the DNA, a sample (1 μg) of a size marker ‘Lambda EcoRI-Hind III’ (Promega Corporation, Madison, Wis., U.S.A.) was electrophoresed in parallel.
27 pairs of oligonucleotide primers were used to amplify the promoter and 30 exons of the GPIIb gene, exons 5 + 6, 8 + 9, 10 + 11 and 15 + 16 being coamplified. All oligonucleotides hybridized in intronic sequences enabling amplification of the entire exon and a small part of flanking introns including splice sites. The design of the oligonucleotides was based on known GPIIb genomic sequences ( Heidenreich et al, 1990 ). The oligonucleotides used to amplify the fragment containing exon 12 of the GPIIb gene were 5′-TCCAGTCCCATGTAACCACT-3′ (forward primer, located in intron 11) and 5′-CTCTGCAGCAAGTAGGGCT-3′ (reverse primer, located in intron 12); for exon 18 of the GPIIb gene were 5′-GGTAGCAAGATGGCCTGACT-3′ (forward primer, located in intron 17) and 5′-TGCACCTCCCTGGCCTGT-3′ (reverse primer, located in intron 18). The sizes of the amplified products were 280 bp and 196 bp respectively. Oligonucleotides were synthesized by Dr J.-C. Gandar (Institut de Biochimie et Génétique Cellulaire, Bordeaux, France). They were subjected to two cycles of precipitation in a mixture of absolute ethanol and 3.3 M sodium acetate (2.5 vol:0.1vol) prior to use. Each PCR amplification was performed for 40 cycles using 1 U Taq polymerase (Promega) in a DNA programmable thermal cycler (Perkin-Elmer, Norwalk, Conn., U.S.A.) under the following conditions: denaturation, 98°C for 15 s; annealing, 50°C or 58°C for 30 s; and extension, 72°C for 30 s. An additional elongation at 72°C for 3 min in 50 μl buffer containing 1.5 m M MgCl2 followed the final cycle. Yields of PCR products were determined from ethidium bromide-stained 1.5% agarose gels after electrophoresis. Migration was compared to that of standard DNA size markers ‘100 bp DNA marker’ (Promega). Analysis of the exons of the GPIIIa gene was performed as described by Jin et al (1993 ).
SSCP analysis was performed according to published procedures ( Jin et al, 1993 ; Peyruchaud et al, 1995a , b). An aliquot (3 μl) of each PCR product was mixed with an equal volume of denaturating solution (95% v/v formamide, 10 m M NaOH, 0.05% w/v bromophenol blue, 0.05% w/v xylene cyanol) and cooled on ice to facilitate the formation of stable secondary single-strand structures and to avoid double-stranded renaturation. Polyacrylamide gel electrophoresis was performed using the minigels of the Pharmacia PhastSystem (Pharmacia Biotech, Saint-Quentin-Yvelines, France). The optimal conditions for the analysis of amplification products containing exon 12 of the GPIIb gene were 270 Vh at 15°C using a 12.5% polyacrylamide gel, and for exon 18 of this gene they were 200 Vh at 4°C using a 12.5% polyacrylamide gel. Gels were stained with silver nitrate according to the PhastSystem development technique file No. 210 from Pharmacia Biotech.
Amplified DNA fragments containing exons 12 or 18 of the GPIIb gene were purified by ‘PCR preps DNA purification resin’ and WizardTM minicolumns according to the conditions recommended by the manufacturer (Promega). They were then directly sequenced using a fmolTM DNA Sequencing System Kit (Promega) in a DNA programmable thermal cycler (Perkin-Elmer) according to the manufacturer's instructions and using a primer 5′ end-labelled with γ-33P-ATP (Isotopchim, Ganagobie-Peyruis, France). The DNA fragments were sequenced respectively using a forward primer and a reverse primer. The sequences of these primers are given above. After the reaction mixture (6 μl) was overlaid with 30 μl mineral oil, the thermal cycler was preheated to 95°C for 2 min to prevent nonspecifically annealed primers from being extended. The cycling programme was 95°C for 30 s (denaturation), 42°C for 30 s (annealing), and 70°C for 1 min (extension) for 30 cycles. All sequencing products were size-fractionated on a 6% w/v polyacrylamide sequencing gel prepared in TBE buffer containing 7 M urea, and electrophoresed at 60 W at room temperature. The samples were denaturated for 2 min at 80°C prior to electrophoresis and 1.5 μl deposited in each well. The gels were dried and subjected to autoradiography for 24 h.
PCR amplification products containing exon 12 of the GPIIb gene were purified by ‘PCR preps DNA purification resin’ and WizardTM minicolumns (see above) and then digested with 6 U MscI restriction enzyme (New England Biolabs, Mass., U.S.A.) for 2 h at 37°C in a 40 μl reaction mixture. The MscI digests were separated by electrophoresis on a 10% (w/v) polyacrylamide gel in TBE buffer for 16 h, and visualized by ethidium bromide staining. The bands were visualized using UV light and imaged on PolaroidTM film (Poly Labo).
Platelets from the patient, his parents and other family members were first examined by immunoblotting to see if trace amounts of GPIIb and/or GPIIIa remained (Fig 1). No GPIIb was detected in his platelets. In contrast, the patient's platelets contained residual GPIIIa with relatively normal migration. The platelets from his mother (no. 1), his maternal uncle (no. 7) and his maternal grandmother (no. 3) also contained decreased amounts of GPIIb compared to the control. This suggested that they were potential carriers of a thrombasthenic trait carried by GPIIb. The patient's results confirmed the diagnosis of type I disease and led us to first examine the GPIIb gene for potential molecular genetic defects in this family using PCR-SSCP analysis.
PCR-SSCP, direct sequencing and ASRA analysis
Each PCR reaction was controlled by electrophoresis of the amplified fragments on 1.5% agarose gels. For the patient, each of the amplified products was of the correct size, confirming the absence of large homozygous deletions or insertions for this patient (data not shown). SSCP analysis revealed migration changes respective to the control for amplified DNA fragments corresponding to exons 12, 18, 21, 22, 26 and 30 of the GPIIb gene (exon numbering according to Heidenreich et al, 1990 ), the patterns of the other fragments were identical to control samples. The differences for the fragments incorporating exon 26 and exon 21 were due to known polymorphisms linked to the HPA-3 alloantigen system (exon 26) and to a 9 bp intronic deletion linked to the HPA-3b determinant (intron 21) (see Peyruchaud et al, 1995b ). The patient was found to be heterozygote for HPA-3a/HPA-3b. The differences for the fragments containing exon 22 and exon 30 corresponded to two new silent polymorphisms which are also linked to the HPA-3 system ( Ruan et al, 1998 ). In contrast, the two altered patterns seen for fragments containing exon 12 and exon 18 did not correspond to any known polymorphisms of the GPIIb gene.
The agarose gel shown in Fig 2A contains the PCR products obtained by amplifying exon 12 from genomic DNA of a control donor (C), the patient's mother (lane 1), the patient and his father (lane 2). As expected, a single fragment of the correct size was obtained. Fig 2B shows the SSCP pattern obtained from the products shown in Fig 2A. The SSCP pattern for the control (lane C) and the patient's father (lane 2) shows two bands labelled (a) and (b). The pattern for the patient and his mother (lane 1) shows not only bands a and b after SSCP, but also two additional bands (a′) and (b′). Both subjects were therefore presumed to be heterozygous for a genetic alteration in this exon. This was confirmed by direct sequencing of PCR products from the patient and a normal control which revealed that the patient's DNA contained a heterozygous point mutation (G to A) at nucleotide 1064 of the cDNA leading to a Glu324 to Lys amino acid substitution (Fig 2C). This mutation is located between the second and third Ca++ binding domains of GPIIb. It should be noted that the nomenclature used here does not take into account the leader sequences and signal peptides for the GPIIb and GPIIIa genes (see French, 1998). When these are taken into account, the alternative numbering for the amino acid mutations described becomes Glu355 → Lys (Glu324 → Lys) and Ile596 → Thr (Ile565 → Thr) respectively.
As this mutation gives rise to a MscI restriction site in exon 12 of the GPIIb gene, ASRA analysis was performed for each of the family members available for study. As illustrated in Fig 3, the mutation results in the appearance of a 184 bp fragment and a 96 bp fragment in place of a 280 bp fragment seen for normal alleles. The ASRA pattern for the patient, his mother (lane 1), his maternal uncle (lane 7) and his maternal grandmother (lane 3) present a combination of both of the normal and the mutant profiles. They were therefore heterozygous for this mutation. The other family members gave a normal profile. These results confirm that the patient had inherited a mutated maternal allele.
Fig 4 illustrates the analysis of exon 18 of the GPIIb gene. The fragments containing exon 18 were amplified by PCR for a normal control, the patient, his father and his mother. As expected, a single 196 bp fragment was obtained (Fig 4A). The SSCP patterns for the control (lane C) and the patient's mother (lane 1) show a major band labelled (a). Fine inspection of the patterns for the patient and his father (lane 2) shows two closely migrating bands (a) and (b) in this position (Fig 4B). This prompted us to sequence amplified genomic DNA fragments containing exon 18, and the results showed the patient to be heterozygous for a T to C transition at position 1787 of GPIIb cDNA. This mutation leads to an Ile565 to Thr substitution. The analysis of the SSCP pattern for the other family members showed that the patient's father and his paternal grandmother (no. 5 in Fig 1) were also heterozygotes for the second mutation (data not shown). As no restriction site was created, we sequenced the amplified fragments for exon 18 for his paternal grandparents. Only the grandmother was heterozygous for the Ile565 to Thr substitution.
Finally, PCR-SSCP analysis of the exons coding for GPIIIa confirmed the absence of abnormalities in the GPIIIa gene (results not illustrated).
Our aim was to determine the molecular abnormalities responsible for the phenotype of a boy from a Swiss family whose clinical laboratory findings were consistent with type I Glanzmann's thrombasthenia. The patient's platelets lacked detectable GPIIb by Western blotting and contained only small amounts of GPIIIa of normal size. The limit of detection of GPIIb under the conditions used is about 1% of normal values. This initial finding led us to examine each of the 30 exons and each of the exon–intron boundaries of the GPIIb gene by PCR-SSCP analysis. Combined with direct cycle sequencing of PCR-amplified genomic DNA, we found that the patient was probably a compound heterozygote for Glanzmann's thrombasthenia, inheriting one mutant allele from the GPIIb gene of each parent. The maternal mutant allele located in exon 12 was a single point mutation (G to A) at nucleotide 1064 of cDNA leading to a Glu324 → Lys amino acid substitution in the GPIIb subunit. This mutation resulted in the presence of a MscI restriction site and ASRA defined the inheritance of this mutation in the patient's immediate family and furthermore confirmed the results of the PCR-SSCP analysis. In fact, this mutation is not new to us, having been found in a homozygous state in an Algerian woman with type I Glanzmann's thrombasthenia ( Bourre et al, 1995 ). The location of the same defect in unrelated patients of Swiss and Algerian nationalities implies that the mutations occurred independently. When investigating the Algerian mutation, Bourre et al (1995 ) performed site-directed mutagenesis and showed that GPIIb–IIIa containing GPIIb Lys324 failed to reach the surface of Cos-7 cells when transiently expressed with wild-type GPIIIa, thereby confirming the link between this mutation and the type I Glanzmann's thrombasthenia phenotype.
This mutation in the maternal allele of our patient is located between the second and third Ca2+-binding domains of GPIIb (see Ginsberg et al, 1995 ). The replacement of a negatively charged Glu324 by a positively charged Lys will influence the electric charge within this region of GPIIb. Significantly, Glu324 is well conserved within the α-subunits of integrins although it is replaced by an Arg in the α-subunits of the β2 integrins (LFA-1, Mac-1 and p153/95). This conservation implies that Glu324 has an important structural or functional role. Other thrombasthenic patients have been described that harbour single amino acid substitutions located within or in proximity to the four putative Ca2+-binding domains of GPIIb. Two point mutations (Gly242 to Asp within the first Ca2+-binding domain ( Poncz et al, 1994) and Gly418 to Asp flanking the fourth Ca2+-binding domain ( Wilcox et al, 1994 ) introduced an additional negatively charged residue proximal to aspartate residues and in each case resulted in type I GT. The third point mutation, an Arg327 to His located between the second and third Ca2+-binding domains ( Wilcox et al, 1995 ), preserved an existing positive charge and gave rise to type II GT with the level of GPIIb–IIIa complexes surface-expressed in platelets estimated to be of the order of 7%. A small deletion of two amino acids (Val425 and Asp426) located at the proximal end of the fourth Ca2+-binding domain in GPIIb resulted in the loss of a charged residue ( Basani et al, 1996 ). In order to confirm the pathogenicity of these mutations, and to study the biological role of the Ca2+-binding domains of GPIIb during the synthesis of the GPIIb–IIIa complex, the above authors used site-directed mutagenesis to introduce the cited mutations into the cDNA for wild-type GPIIb, and then coexpressed the mutants with wild-type GPIIIa in eukaryotic cells (such as Cos-1, Cos-7 or CHO cells). The results showed that the induced mutations did not impair pro-GPIIb synthesis or pro-GPIIb–IIIa complex assembly, but altered the conformation of the immature complex which was not transported from the endoplasmic reticulum (ER) to the Golgi complex where the pro-GPIIb was normally cleaved into heavy and light chains. Overall, it appears that the four Ca2+-binding domains are critical for maintaining the correct conformation of nascent GPIIb and for normal intracellular trafficking and maturation of the GPIIb–IIIa complex.
The Glu324 → Lys mutation was detected in a heterozygous state in three other members of the mother's family, and although each of them showed a lower than normal expression of GPIIb in their platelets, Western blotting clearly distinguished them from the patient. Thus this mutation could not alone explain the type I phenotype of the patient. A second candidate mutation was inherited from the paternal mutant allele, an Ile565 → Thr amino acid substitution located in the extracellular domain of GPIIb. Although Western blotting did not reveal noticeably decreased GPIIb in the platelets of other family members heterozygous for this mutation on the father's side, this could be a reflection of variations in the levels of GPIIb and GPIIIa within the normal population (see Coller et al, 1986 b). In this regard, it should be noted that the only member of this family to be studied with two normal GPIIb alleles (no. 6, Fig 1) possessed the highest levels of GPIIb of the subjects tested (see Fig 1). Site-directed mutagenesis followed by expression of the mutated GPIIb with wild-type GPIIIa in heterologous cells will be required to provide absolute proof for the pathological nature of this mutation which differs from all known polymorphisms of GPIIb. However, the recent detection of the Ile565 → Thr mutation in another unrelated doubly heterozygous patient with Glanzmann's thrombasthenia is a strong indication that this mutation is involved in the expression of the Glanzmann's phenotype ( French et al, 1997 ). The closest reported mutation to these in the GPIIb gene is a nonsense mutation at amino acid Arg553 (Arg584 taking into account the signal peptide) coded for in exon 17 ( Gu et al, 1993 ). However, this latter mutation gives rise to a stop codon and a truncated protein. The T → C change, which gives rise to the Ile565 → Thr substitution, is the only mutation yet to be described in exon 18.
Recently, Springer (1997) has shown that the N-terminal domain (representing about 440 amino acids) of integrin α subunits contains sequence repeats involving FG (phenylalanyl-glycyl) and GAP (glycyl-alanyl-prolyl) consensus sequences. These repeats are predicted to fold into a β-propeller domain with, for GPIIb, seven four-stranded β-sheets. In terms of this model, the mutated Glu324 would fall within the fifth β-sheet. Modifications of this domain close to a Ca2+-binding site would be expected to modify the folding within the β-sheet and destabilize the capacity of GPIIb to associate with GPIIIa. In contrast, the Ile565 → Thr mutation, although on the GPIIb heavy chain and extracellular, is to be found within the central torus of the β-propeller. Its location in two unrelated Glanzmann's patients points to an as yet uncharacterized role of this domain in GPIIb–IIIa complex formation.
In conclusion, the patient appeared to be a compound heterozygote as a result of point mutations in two different exons of the GPIIb gene. As reviewed by French (1998), compound or double heterozygotes are a common form of Glanzmann's thrombasthenia in patients where consanguinity can be excluded and where the patient does not belong to a minority ethnic group with a known high risk for a particular Glanzmann's genotype. The two mutations described here will help further define the structure/function relationship involved in GPIIb–IIIa biosynthesis and expression.
This work was supported by funding from the CNRS, Université de Bordeaux II, the Conseil Régional d'Aquitaine and the Ministère de l'Enseignement Supérieur et de la Recherche (ACC-SV no. 9). O. Peyruchaud was a recipient of a postdoctoral fellowship from the Société Française d'Hématologie. J. Ruan received a doctoral grant from the Sanofi Association for Thrombosis Research. The participation of Dr J. Beer in some of the studies leading to the initial characterization of this Glanzmann's thrombasthenia patient in Bern is gratefully acknowledged.