A functional platelet fibrinogen receptor with a deletion in the cysteine-rich repeat region of the β₃ integrin: the Oe^a alloantigen in neonatal alloimmune thrombocytopenia

The platelet membrane glycoprotein (GP) IIb-IIIa complex, also referred to as α_IIbβ₃, belongs to the integrin family of heterodimeric adhesion receptors implicated in cell-cell and cell-matrix interactions.1 As receptor for ligands containing the Arg-Gly-Asp sequence such as fibrinogen and von Willebrand factor, GPIIb-IIIa plays an essential role in hemostasis by mediating platelet aggregation and platelet spreading on vascular matrices.2 GPIIb-IIIa primarily functions as a responsive adhesion receptor serving as a bidirectional conduit for inside-out and outside-in signaling processes across the plasma membrane. Whereas intracellular signals modulate the ligand-binding function of GPIIb-IIIa (inside-out), signals generated by ligation and clustering of GPIIb-IIIa regulate the extent of platelet aggregation and spreading (outside-in).3 

The mature GPIIIa subunit is a disulfide bond-rich single polypeptide consisting of 762 amino acids encompassing a large extracellular domain, a single transmembrane domain, and a cytoplasmic tail. The extracellular domain carries an I-domain–like ligand-binding region (residues 110-294) and a cysteine-rich repeat region (423-622) that contains 31 of 56 cysteine residues. All cysteine residues are located in highly conserved regions of the molecule and play an important role in preserving the 3-dimensional structure, because pertubation or absence of one or more of these residues could affect stability or ligand-binding function.4,5 Recently, Yan and Smith demonstrated that a selected group of the cysteines located within the extracellular cysteine-rich domain remains unpaired. The redox status of these cysteine residues can directly influence the activation state of GPIIb-IIIa.6 

Several mutations of GPIIIa receptor lead to the loss of either expression or function of the GPIIb-IIIa and are therefore responsible for Glanzmann thrombasthenia, an inherited bleeding disorder characterized by the failure of platelet aggregation.7 

Glycoprotein IIIa represents one of the most polymorphic molecules on the platelet surface. So far, 7 human platelet alloantigen systems (HPA-1, -4, -6, -7, -8, -10, and -11) could be characterized on this receptor. The HPA systems are induced by single point mutations of the high frequency GPIIIa Leu₃₃ isoform (HPA-1a; Pl^A1, Zw^a), leading to the generation of 6 less frequently occurring GPIIIa alleles.8 

Glycoprotein IIIa complementary DNA (cDNA) of human and other species showed a remarkably high degree of homology. Alignment analysis showed that the deduced amino acid sequence of canine and mouse corresponds to the high-frequency human GPIIIa isoform.9 

Platelet alloantigens can elicit the production of alloantibodies and lead to neonatal alloimmune thrombocytopenia (NAIT), posttransfusion purpura, and platelet transfusion refractoriness.10 NAIT is induced by maternal immunization against a fetal HPA and subsequent transplacental transfer of the maternal antibody into the fetal circulation. In the white population about 75% of the NAIT cases are caused by immunization against HPA-1a (Pl^A1, Zw^a) carried on the GPIIIa Leu₃₃ isoform.11 

In recent years, several groups reported that polymorphism of GPIIIa is not only immunologically relevant but may also contribute as a risk factor in the development of coronary heart disease.12,13Individuals carrying the HPA-1b allele seem to have increased risk for adverse cardiovascular events. At present, however, there is only limited evidence that integrin signaling may be affected by GPIIb-IIIa polymorphisms.14,15 

In this study, we describe the biochemical, molecular, and functional characterization of a new immunogenic variant of GPIIIa, which is involved in NAIT.

The 27-year-old mother (Ms Oe) had a healthy child after her first pregnancy. The second pregnancy ended with an abortion. After her third uneventful pregnancy she gave birth to a male infant (3540 g, Apgar score 1/5/10). The neonate exhibited cutaneous signs of a thrombocytopenic hemorrhagic diathesis: multiple hematoma and petechiae on head, back, and arms. Six hours postpartum the platelet count was 36 000/μL and further declined to a nadir of 29 000/μL a few hours later. Initially, septicemia was suspected and antibiotic therapy was initiated, but no other signs of a severe infection were detectable. During treatment with intravenous IgG over 2 days the platelet count increased to 63 000/μL. For 10 further days the patient received corticosteroids. Two weeks later the platelet count had increased to normal values and the newborn could be dismissed in good health.

Blood samples from the mother, the father, and the child were referred to our laboratory shortly after delivery because of suspected NAIT. Platelets were isolated from EDTA-anticoagulated blood by differential centrifugation and stored at 4°C in isotonic saline containing 0.1% NaN₃. For the characterization of platelet alloantibodies, platelets were selected from a large number of donors with known HPAs and ABO blood group antigens.

All samples were obtained with informed consent and with the approval of the National Blood Service ethics review board.

Phenotyping of HPAs was performed by the monoclonal antibody–specific immobilization of platelet alloantigens (MAIPA assay) as previously described.16 

Monoclonal antibodies (mAbs) Gi5 and Gi9 specific for GPIIb-IIIa and GPIa-IIa, respectively, were produced in our laboratory.17 The mAb FMC25 directed against GPIX subunit of GPIb-IX-V complex was provided by Dr Heddy Zola (Adelaide, Australia).18 The mAb AP3 specific for GPIIIa and mAb D3 against a ligand-induced binding site (LIBS) on GPIIIa were kindly provided by Dr Peter Newman (Milwaukee, WI) and Dr Lisa Jennings (Memphis, TN), respectively.19,20 High-performance liquid chromatography–purified RGDW and RGEW peptides were purchased from Bachem (Heidelberg, Germany). The mAbs 77 and PY20 directed against pp125^FAK and phosphotyrosine, respectively, as well as mAb PAC-1 against activated GPIIb-IIIa complex were purchased from Becton Dickinson (Heidelberg, Germany).

Platelets were isolated from acid-citrate-dextrose anticoagulated blood by differential centrifugation. Platelets (4 × 10⁶) were lysed in 2% sodium dodecyl sulfate (SDS), 30 mM N-ethylmaleinimide, 100 mM phenylmethylsulfonyl fluoride (PMSF) for 5 minutes at 100°C. The lysate was separated in a 7.5% SDS–polyacrylamide gel electrophoresis (SDS-PAGE) and blotted on a nitrocellulose membrane. Then, 200 μL eluate prepared from either maternal (anti-Oe^a) or control sera (anti–HPA-1a, AB serum) was incubated with the membrane for 90 minutes at room temperature. An alkaline phosphatase–labeled rabbit anti–human IgG (Dako, Hamburg, Germany) was added and bound antibodies were visualized with 5-bromo-4-chloro-3-indolyl phosphate (BCIP) substrate solution.

Platelets and Chinese hamster ovary (CHO) stable transfectants were surface labeled with 5 mmol/L NHS-LC biotin (Paesel, Frankfurt, Germany) as previously described.21 Then, 5 × 10⁸ labeled platelets in 500 μL phosphate-buffered saline (PBS) were digested with 500 μL chymotrypsin (1 mg/mL, Sigma, Taufkirchen, Germany) at 37°C overnight and the reaction was terminated by adding 100 μL PMSF (5 mg/mL). After washings labeled platelets were lysed in 1 mL solubilization puffer (25 mM Tris, 10 mM EDTA, 100 mM NaCl, 1% Triton X-100, 2 mM PMSF, 1 mM leupeptin, and 2 mM N-ethylmaleimide) for immunoprecipitation.

Total platelet RNA was isolated from 20 mL EDTA anticoagulated blood as previously described.21 Genomic DNA was obtained from leukocyte 3 mL EDTA anticoagulated blood using QIAquick-Kit (Qiagen, Hilden, Germany).

Platelet RNA (100 μL) was reverse transcribed using 10 μM oligodT (Boehringer Mannheim, Mannheim, Germany). To amplify the entire coding region of GPIIIa by polymerase chain reaction (PCR), 4 overlapping sets of primers (nucleotides 56-698, 633-1341, 1116-1799, 1666-2415) were constructed based on the published cDNA.22For cDNA amplification of the fourth region forward primer no. 1 (1503-CCAGTGTGAGTGCTCAGAGGA-1524), reverse primer no. 2 (2415-GGCTGATAATGATCTGAGGATGACTG-2389), and nested primer no. 3 (1666-GTGACGACTTCTCCTGTGTCCGCTACAAGG-1695) were used. Then 5 μL cDNA was diluted with 10 times PCR buffer, 0.25 μM of primer no. 1 and no. 2, 175 μM of each dNTP, and 2.5 U TaqGold polymerase (Applied Biosystems, Weiterstadt, Germany) in a total volume of 50 μL. Amplification was performed on DNA thermal cycler 480 (Applied Biosystems) for 15 cycles. Each cycle consisted of denaturation at 96°C for 90 seconds, annealing at 52°C for 60 seconds, and extension at 72°C for 120 seconds. In the final cycle, the samples were kept at 72°C for 10 minutes and then chilled for 4°C. An aliquot of 1 μL diluted PCR products (1:100; 1:1000) was reamplified using primer no. 1 and no. 3 for 30 cycles under the same conditions with annealing temperature of 58°C. PCR products were analyzed on 1.6% SeaKem GTG agarose gel (Biozyme, Hameln, Germany). After purification by QIAquick Kit, PCR products were subcloned into theEcoRV cloning site of pGEM-5Zf (Promega Biotech, Madison, WI). Plasmid DNA from 16 positive clones was sequenced using Taq-FS Dye Terminator Cycle Sequencing Kit and were analyzed on ABI PRISM Genetic Analyzer 310 (Applied Biosystems).

To analyze the entire exon 10 of GPIIIa,23 5 μg genomic DNA was amplified using intronic sense primer no. 4 (5′-cagcgggtccaccttcct-3′) and antisense primer no. 5 (5′-cctgcctcccggctctct-3′) for 30 cycles. Each cycle consisted of denaturation at 95°C for 4 minutes, annealing at 61°C for 30 seconds, and extension at 72°C for 60 seconds. After purification by QIAquick Kit, 60 ng amplified DNA was directly sequenced as above.

Genotyping of HPA-1 antigen and Leu₄₀Arg dimorphism from human genomic DNA was performed by restriction fragment length polymorphism (PCR-RFLP) analysis as previously described.24,25 PCR products were digested withMspI or ScrFI endonuclease and were analyzed on 2.2% agarose gel.

DNA typing of Oe^a alloantigen was performed by sequence-specific PCR (PCR-SSP). In brief, 5 μg genomic DNA was added to 50 μL reaction mixture containing 10 mM Tris, 50 mM KCl, 2.75 mM MgCl_2, 0.2 mM of each dNTP, 0.4 μM each of sense (1753-ACTGGACCGGCTACTACTGCAA-1774) and sequence-specific antisense primer (ctccagactccacactcacTTC/A-1931), 0.2 μM each of human growth factor (HGH) I (5′-CAGTGCCTTCCCAACCATTCCCTTA-3′) and HGH II (5′-ATCCACTCACGGATTTCTGTTGT GTTTC-3′), and 2.5 UTaqGold polymerase. After initial denaturation at 96°C for 10 minutes, amplification was performed in a DNA thermocycler (Hybaid, Heidelberg, Germany) for 35 cycles (denaturation at 94°C for 50 seconds, annealing 55°C for 30 seconds, and extension 72°C for 15 seconds). The PCR products were analyzed on 2.0% agarose gel electrophoresis using molecular marker V (Boehringer Mannheim) as DNA standard.

DNA from nonhuman primates (prosimian, new world monkeys, old world monkeys, and Hominoidea) were kindly provided by Dr C. Roos (Institute for Genetics, University of Munich, Germany). For the genotyping of GPIIIa gene polymorphism DNA was amplified and analyzed by direct nucleotide sequencing as described above.

Full-length human wild-type GPIIIa and GPIIb cDNAs in the mammalian expression vector pcDNA3/Neo, which encode Leu₃₃GPIIIa isoform and Ser₈₄₃ GPIIb isoform, respectively, were kindly provided by Dr P. Newman (Blood Research Institute, Milwaukee, WI). Allele-specific recombinant GPIIIa isoforms Pro₃₃, Leu₃₃ΔLys₆₁₁, and Pro₃₃ΔLys₆₁₁ were produced from the wild-type GPIIIa cDNA by site-directed mutagenesis using Quick-Change Mutagenesis Kit (Strategene, Heidelberg, Germany).

For the construction of the expression vector encoding for Pro₃₃ isoform PCR was performed using one mismatched (underlined) forward primer (5′-CTGATGAGGCCCTGCCTCCGGGCTCACCTCGCTGTG-3′) and reverse primer (5′-CACAGCGAGGTGAGCCCGGAGGCAGGGCCTCATCAG-3′) from base 178-213 of GPIIIa cDNA as previously described.21 Both Leu₃₃ and Pro₃₃ GPIIIa constructs were then deleted for 3 bases (nucleotides 1929-1931) by site-directed mutagenesis using forward primer (5′-CCAAGATGCCTG CACCTTTAAAGAATGTGTGGAGTG-3′) and reverse primer (5′-CACTCCAC ACATTCTTTAAAGGTGCAGGCATCTGG-3′) encompassing nucleotides 1910-1948 as described above. Full-length GPIIb in pcDNA3/Neo plasmid was digested with XbaI and EcoRI endonucleases (Biolabs, Frankfurt, Germany) and was shuttled into the pcDNA3.1/Zeo vector, which had been digested with the same enzymes.

After subcloning in Escherichia colibacteria, all constructs were validated by nucleotide sequence analysis for subsequent transfection studies.

The CHO cells were transfected with allele-specific GPIIIa and GPIIb expression vectors by the use of the reagent Lipofectin (Gibco, Karlsruhe, Germany) as previously described.21 In brief, 3 μg of each plasmid was mixed with 25 μL lipofectin in 2 mL OptiMEM Medium (Gibco) and then added to a subconfluent 10-cm plate of CHO cells (8 × 10⁵ cells) for 12 hours. Medium (9 mL) was then added and the incubation was continued for 48 hours. After splitting, transfectants were selected with Geneticin (800 μg/mL, Gibco) and Zeocin (500 μg/mL, Invitrogen, Groningen, The Netherlands) for approximately 2 weeks. After subcloning, surface GPIIb-IIIa was analyzed by flow cytometry using mAbs AP3 and Gi5 specific for GPIIb and GPIIb-IIIa complex, respectively. To obtain a homogenous cell population, stable cell lines were cloned 3 times and different clones (3 cell lines for each transfection) were maintained in the same selection media.

Stable transfectants were analyzed by flow cytometry as previously described.21 Cell suspensions of 200 μL (8 × 10⁵ cells) were incubated with 20 μL mAb Gi5 (20 μg/mL), washed, and labeled with 40 μL fluorescein isothiocyanate (FITC)–conjugated rabbit anti–mouse IgG (1:80 dilution; Dako). Cell were resuspended in 500 μL PBS containing 0.5% bovine serum albumin (BSA) and 0.1% NaN₃ and were analyzed by flow cytometry (FACS Calibur, Becton Dickinson). Binding of the LIBS mAb D3 in the absence of RGDW or RGEW peptide was assessed as described.26 Cells were incubated with 1 mM peptides for 5 minutes at room temperature prior to labeling with 20 μL mAb D3 (0.02 mg/mL).

PAC-1 binding was analyzed as described by Lyman et al.27Stable transfectant (2 × 10⁷) in 1 mL Tyrode buffer (137 mM NaCl, 2.8 mM KCl, 12 mM NaHCO₃, 10 mM Hepes, pH 7.4) was treated with 10 mM dithiothreitol (DTT) or buffer for 20 minutes at room temperature. Cells were then washed once and resuspended in 1 mL Tyrode buffer supplemented with 3.5 mg/mL BSA (TB). Aliquots of 100 μL untreated and DTT-treated transfectants in TB were stained with 20 μL FITC-labeled mAb PAC-1 (1 μg/mL) in the presence of 10 μL 10 mM MgCl₂ and 1 mM CaCl₂ for 30 minutes at room temperature. Cells were washed once and resuspended in 500 μL TB containing MgCl₂ and CaCl₂ for FACS analysis.

Adhesion of transfected CHO cells on fibrinogen-coated wells was performed as described by Wang and Newman.26Microtiter wells were coated with 100 μL PBS containing different concentrations of fibrinogen (Calbiochem, Schwalbach, Germany), 10 μg/mL mAb Gi5, or 1% BSA at 4°C overnight and blocked with 1% BSA in PBS at room temperature before use. Stable transfectants were labeled with 2 μM calcein-am (Mobitec, Göttingen, Germany) at 37°C for 30 minutes, washed, and resuspended in serum-free media. An aliquot of 100 μL cell suspension (2 × 10⁵ cells) was added to each well and cells were allowed to adhere for 1 hour at 37°C. The wells were then washed 3 times with media containing 1% BSA. Finally, 200 μL wash media was added to each well and the plates were read in a microtiter plate fluorescence reader (Spectrafluor Plus, Tecan, Crailsheim, Germany) at excitation and emission wavelengths of 485 and 515 nm, respectively.

Tissue culture plates (100 mm) were coated overnight at 4°C with either fibrinogen (Calbiochem) or BSA (10 mg/mL; Serva, Heidelberg, Germany) and blocked with 1% BSA for 2 hours at room temperature. Transfectants in serum-free media were added and incubated at 37°C for 60 minutes. Adherent cells were lysed in RIPA buffer containing 1% Triton X-100, 150 mM NaCl, 10 mM Tris, 1 mM EDTA, 1 mM Na₃VO₄, 0.5% Nonidet P40, 1% sodium desoxycholate, 2 mM PMSF, and 10 μg/mL each aprotinin and leupeptin (Sigma). After centrifugation at 15 000g at 4°C for 30 minutes, protein concentration of the cell lysate was determined using the BCA protein assay (Perbio, Bonn, Germany). Proteins (500 μg) were incubated with 5 μL rabbit polyclonal anti-pp125^FAK(Becton Dickinson) overnight and immunoprecipitated as previously described.21 Immunoprecipitates were transferred onto PVDF membrane and the tyrosine phosphorylation state was then detected with mAbs 77 (250 μg/mL) and PY20 (1 mg/mL) specific for pp125^FAK and phosphotyrosine, respectively (dilutions 1:1000). Prior to incubation with anti-pp125^FAK blots probed with antiphosphotyrosine were stripped for 30 minutes at 50°C in 62.5 mM Tris buffer, pH 6.7, containing 2% SDS and 100 mM β-mercaptoethanol. Labeled proteins were then visualized using peroxidase-labeled rabbit anti–mouse IgG (dilution 1:10 000; Dianova, Hamburg, Germany) and chemiluminescence substrate (ECL Plus; Amersham Pharmacia, Freiburg, Germany). The signals were scanned using Correl PhotoPaint software and the densitometric quantitation was performed using Kodax 1D Image. Phosphorylation of pp125^FAK was determined from densitometric scans of pTyr mean intensity divided by pp125^FAK mean intensity.

When maternal serum was tested in the MAIPA assay using mAb Gi5 specific for GPIIb/IIIa complex, a positive reaction was obtained with platelets from the father and grandfather, but not with autologous platelets and with platelets from unselected donors. In addition, the Oe^a serum did not react with platelets carrying known low-frequency antigens on the GPIIb or GPIIIa subunits. Analysis of maternal serum with mAb Gi9 and FMC25 against GPIa-IIa and GPIb-IX complex, respectively, yielded negative results (data not shown). In a population study, none of 600 unrelated blood donors was found to carry the Oe^a alloantigen (phenotype frequency < 0.0017).

Figure 1 shows the pedigree of the family Oe^a. This segregation pattern demonstrated that the Oe^a antigen is inherited as a codominant allele. Interestingly, the Oe^a (+) grandfather was homozygous for the HPA-1b character.

A rare Leu₄₀Arg variant of the GPIIIa gene associated with the HPA-1b allelic isoform was found in the human gene pool.24 Because this mutation could theoretically be linked to the Oe^a alloantigen, we did HPA-1 genotyping of Oe^a (+) individuals by RFLP using MspI. As shown in Figure 2, HPA-1b individuals carrying the Arg₄₀ (lane 5) could be clearly distinguished from HPA-1b/1b homozygous healthy blood donor (lane 4) and HPA1b, Oe^a (+) individuals (lanes 2 and 3) by the presence of a 151–base pair (bp) instead of 171-bp restriction fragment. This finding demonstrates that the Arg₄₀ is not associated with the Oe^a alloantigen.

To further localize the Oe^a alloantigenic determinants, immunoblotting analysis was performed with platelet lysates derived from an Oe^a (+) individual (Figure3). Under nonreduced conditions, Oe^a alloantibodies recognize the GPIIIa subunit. However, the reactivity of this serum with GPIIIa was abolished after treatment with β-mercaptoethanol (data not shown).

In addition, we performed immunoprecipitation analysis with chymotrypsin-treated platelets. Labeled platelets of an Oe^a(+) individual were partially digested with chymotrypsin and were then subjected for immunoprecipitation. As shown in Figure4, anti–HPA-1a alloantibody recognized 72- and 66-kd proteolytic fragments of GPIIIa (lane 1), whereas Oe^a and HPA-4a alloantibodies (lanes 2 and 3) failed to react with both fragments. In contrast, undigested GPIIIa could be precipitated with all alloantibody specificities. These results indicated that the Oe^a alloantigenic determinant is located on the GPIIIa loop, which is cleaved after chymotrypsin treatment or resides on the remnant membrane- bound GPIIIa, and is destroyed by chymotrypsin treatment.

To elucidate the molecular basis underlying the Oe^aalloantigens, we amplified the entire coding region of GPIIIa cDNA derived from an Oe^a (+) HPA-1b homozygous individual. PCR products were then subcloned and sequenced. Figure5A shows the sequence analysis of 2 GPIIIa regions (surrounding nucleotide 1925 and nucleotide 191). In the wild-type GPIIIa allele, the amino acids Phe₆₁₀Lys₆₁₁Lys₆₁₂ are encoded by the nucleotides sequences TTT AAG AAA. In contrast, the mutant GPIIIa allele (5 of 16 clones) shows a deletion of AAG triplet encoding Lys₆₁₁. Furthermore, both Oe^a (+) and (−) HPA-1b homozygous individuals carry the nucleotide C at position 196, which encodes Pro₃₃ GPIIIa isoform and is responsible for the formation of the HPA-1b epitopes. These results are in accord with our serologic findings showing that Oe^a alloantigen is inherited with the HPA-1b allele (Figure 1).

To analyze the genomic DNA, we performed direct nucleotide sequencing analysis from Oe^a-phenotyped individuals (Figure5B). In comparison to Oe^a (−), Oe^aheterozygous individuals are striking by a shift of 3 bases due to the deletion of AAG triplet. Using this technique, we sequenced 2 Oe^a (+) individuals and 20 Oe^a (−) blood donors and the results were in full accordance with our phenotyping results.

Figure 6 illustrates the location of ΔLys₆₁₁ in GPIIIa cDNA and in genomic DNA. The AAG deletion encodes for the last amino acid on exon 10, which creates a recognition site for DraI restriction enzyme in Oe^a carriers. Thus, typing by PCR-RFLP technique could be performed with cDNA (data not shown). However, this technique could not be adapted for genomic DNA typing, because the introduction of the intronic nucleotide G in the mutant allele abolished the recognition site of DraI. Therefore, we developed the PCR-SSP approach (Figure 6). Both allele-specific primers amplified the 197- and 200-bp fragments of Oe^a heterozygous individuals. In contrast, DNA derived from Oe^a (−) blood donors could only be amplified with the wild-type primer. As internal control, amplification of the 439-bp fragment of the HGH gene was used. The genotypes of 22 individuals determined by PCR-SSP were identical with the phenotypes detected by MAIPA assay.

To demonstrate whether this deletion is directly responsible for the formation of Oe^a epitopes, we constructed allele-specific expression vectors. Site-directed mutagenesis was performed with the wild-type GPIIIa HPA-1a construct to introduce C-T mutation at position 33. In both HPA-1a and -1b constructs (Leu₃₃ and Pro₃₃) the codon at positions 1929-1931 (Leu₃₃ΔLys₆₁₁ and Pro₃₃ΔLys₆₁₁) was deleted. All 4 GPIIIa allele-specific constructs were cotransfected with GPIIb expression vector into CHO cells. Stable transfectants expressing allele-specific GPIIb-IIIa complex were established and labeled with biotin for immunoprecipitation analysis (Figure 7A).

As shown in the right panel, the Pro₃₃Lys₆₁₁ as well as the Pro₃₃ΔLys₆₁₁ GPIIIa isoforms could be precipitated by anti–HPA-1b alloantibodies (lanes 3 and 7). In contrast, anti-Oe^a alloantibodies did not recognize the Pro₃₃Lys₆₁₁ (lane 4), but only reacted with the Pro₃₃ΔLys₆₁₁ GPIIIa variant (lane 8). In the control experiments anti–HPA-1a (lanes 2 and 6) and AB serum from a healthy blood donor (lanes 1 and 5) did not precipitate any protein. In addition, immunoprecipitation analysis of the Leu₃₃ GPIIIa isoforms (Figure 7A, left panel) showed that anti-Oe^a was also able to precipitate the Leu₃₃ΔLys₆₁₁isoform (lane 8). These results clearly demonstrate (1) that the deletion of Lys₆₁₁ leads to the creation of Oe^aepitopes and (2) this deletion has no influence on the formation of HPA-1a as well as HPA-1b alloantigenic determinants.

To analyze the influence of ΔLys₆₁₁ on the structure of GPIIIa we compared the migration of GPIIIa subunits of Pro₃₃Lys₆₁₁ and Pro₃₃ΔLys₆₁₁ mutant under nonreduced and reduced conditions (Figure 7B). Under nonreduced conditions, the Lys₆₁₁ isoform (lane 1) precipitated with mAb Gi5 migrated slightly slower than ΔLys₆₁₁ isoform (lane 3). In the control experiment, mixture of both isoforms migrated as doublet (lane 2). Under reduced conditions, however, both GPIIIa, Lys₆₁₁and ΔLys₆₁₁ isoforms migrated with the same mobility. These results indicate that the molecular size polymorphism observed in the mutant GPIIIa is dependent on the disulfide bond structure of GPIIIa molecule. This is in accord with our immunoblotting experiment showing the sensitivity of Oe^a epitopes for disulfide reduction.

To study the adhesion properties of Oe^a allelic isoform, clones of stable transfectants expressing high and similar amounts Pro₃₃Lys₆₁₁ or Pro₃₃ΔLys₆₁₁ form of GPIIb-IIIa were selected. Flow cytometry analysis using mab Gi5 specific for GPIIb-IIIa complex revealed equivalent levels of receptor expression on both cell lines (Figure 8A). To examine whether the Oe^a allelic form of GPIIIa could undergo conformational changes for ligand binding, we compared the binding of anti-LIBS mab D3 to both Lys₆₁₁ or ΔLys₆₁₁ transfectants in the presence or absence of RGDW peptide. Equivalent expressions of D3 epitopes were observed in both cell lines in the presence of RGDW peptide (Figure 8B). In the control experiment, no binding of mAb D3 could be detected in the presence of RGEW peptide.

To further determine if the ligand binding domains of the mutant GPIIb-IIIa was functionally intact we measured binding of the ligand mimetic mAb PAC-1 to DTT-treated “activated” transfectants. As shown in Figure 8C, ΔLys₆₁₁ transfectants bound PAC-1 antibody in an equivalent activation-dependent manner as wild-type cells. This indicated that the ligand-binding domains were functional.

Furthermore, we compared the adhesion ability of both transfectants to immobilized fibrinogen. To exclude the influence of different expression level on the evaluation of cell adhesion, adhesion to immobilized fibrinogen was normalized to cell binding on Gi5. As shown in Figure 9A, Lys₆₁₁ and ΔLys₆₁₁ transfected cells showed an equivalent adhesion to immobilized fibrinogen. These results demonstrated that the ΔLys₆₁₁ form of GPIIb-IIIa is able to undergo conformational change and binds fibrinogen in a manner similar to Lys₆₁₁.

To examine whether ΔLys₆₁₁ affects the outside-in signaling of GPIIb-IIIa, tyrosine phosphorylation of focal adhesion kinase pp125^FAK was measured after adhesion of GPIIb-IIIa–transfected cells to fibrinogen (Figure 9B). Blots were incubated subsequently with antiphosphotyrosine and anti-pp125^FAK. Equivalent phosphorylation of pp125^FAK was observed in Pro₃₃ and Pro₃₃ ΔLys₆₁₁ transfectants. Thus, the deletion of Lys, which is responsible for the formation of the Oe^a alloantigenic determinant, affects neither the adhesive properties of GPIIb-IIIa nor the outside-in signaling.

To better understand the evolutionary relationship of GPIIIa alleles, we analyzed DNA derived from 13 nonhuman primates representing 4 different species ranging from prosimian (Varecia variegata), new world monkeys (Callimico goeldii, Saimiri sciureus), old world monkeys (Cercopithecus griseoviridis, Macaca mulatta, Macaca nemestrinus, Macaca fascicularis, Mandrilus sphinx), and Hominoidea (Hylobates syndactylus, Pongo abelii, Gorilla gorilla, Pan paniscus, Pan troglodytes). Amplified DNA encompassing the polymorphic regions 196, 217, 526, 1317, 1564, which encode Leu₃₃Pro, Leu₄₀Arg, Gln₁₄₃Arg, Pro₄₀₇Ala and Arg₄₈₉Gln, respectively, were sequenced directly using PCR primers. All nonhuman primates revealed the amino acid residues Leu₃₃, Leu₄₀, Arg₁₄₃, Pro₄₀₇, and Arg₄₈₉ (Figure 10A). This allelic isoform, referred to as GP3A*01 carrying the HPA-1a epitopes is high frequent distributed among humans.8 These data indicate that GP3A*01 represents the ancestral allele. Seven point mutations of this ancestral allele were found (Figure 10B). The first mutation led to formation of GP3A*02, which encodes GPIIIa Pro₃₃ form carrying HPA-1b epitopes. A further mutational event (ΔLys₆₁₁) occurred in this allele leading to the formation of Oe^a alloantigenic determinant.

In this study, we present data on a new immunogenic GPIIIa variant, the Oe^a alloantigen, responsible for a typical case of NAIT in a white family. Analysis by the use of antigen capture assay, immunoprecipitation, and immunoblot allowed us to localize the Oe^a alloantigenic determinant on the platelet GPIIIa subunit. Nucleotide sequence analysis of GPIIIa cDNA in the Oe^a (+) family members (father and grandfather of the affected child) showed a deletion of one codon (AAG at positions 1929-1931). This deleted triplet encodes for the last amino acid Lys₆₁₁ on exon 10. Our findings could be confirmed by direct nucleotide sequencing analysis of genomic DNA. Furthermore, full accordance between Oe^a phenotyping and genotyping was shown by allele-specific PCR analysis.

Stable expression of recombinant allele-specific GPIIIa in mammalian cells allowed confirmation that the single deletion of Lys₆₁₁ was sufficient to induce the Oe^adeterminant. In addition, we could demonstrate that this deletion did not impair the epitopes of HPA-1.

For the formation of the Oe^a alloantigenic determinants, 3-dimensional structures of GPIIIa appeared to be required, because reduction of disulfide bonds and chymotrypsin treatment abolished the presentation of Oe^a epitopes.

Treatment of platelets with chymotrypsin led to digestion of GPIIIa giving rise to 3 major fragments that represent amino acids 1-100, 101-321, and 322-762. The 66-kd fragment is composed by the amino terminal part (amino acids 1-100) and the carboxy terminal portion (amino acids 322-762), which are disulfide linked and anchored in the platelet membrane.28,29 Although the ΔLys₆₁₁responsible for Oe^a epitopes is located in this part of the GPIIIa molecule, anti-Oe^a failed to react with the 66-kd fragment. This result indicated that the cleaved fragment (amino acids 101-321) took part in the formation of Oe^a epitopes.

Interestingly, the ΔLys₆₁₁ GPIIIa isoform was found to migrate slightly faster than the wild-type GPIIIa isoform under nonreducing conditions. After disulfide reduction, however, both isoforms migrated with the same mobility, indicating that ΔLys₆₁₁ altered the 3-dimensional structure of GPIIIa by impairment of disulfide bonds.

Four main structural domains in GPIIIa could be assigned; the N-terminal cysteine-rich domain (residues 1-62), the fibrinogen-binding domain (residues 101-422), the cysteine-rich proteinase-resistant core (residues 423-622), which is bound to the N-terminal domain by a single disulfide bond (Cys₅-Cys₄₃₅), and the C-terminal domain, comprising an extracellular subdomain (residues 623-692), a transmembrane (residues 693-721), and cytoplasmic subdomains (residues 722-762). Within the cysteine-rich proteinase-resistant core, 4 cysteine-rich repeats have been assigned.30 The ΔLys₆₁₁ responsible for the formation of Oe^a alloantigenic determinants is located in the fourth cysteine-rich repeat encompassing cysteine residues 601, 604, 608, 614, 617, and 687, which are known to be paired among themselves. One could speculate that the ΔLys₆₁₁ may impair the adjacent disulfide bond(s). Further investigations of the mutant ΔLys₆₁₁ using proteomic approach will help to ascertain the GPIIIa disulfide pattern in the cysteine repeat region.

Another molecular weight polymorphism associated with a human platelet alloantigen has been observed in HPA-8 system.31 In HPA-8b individuals we found an Arg₆₃₆Cys mutation in the cysteine-rich repeat, which led to the presence of a free sulfhydryl group in the mutant GPIIIa isoform. This mutation resulted in different glycosylation of GPIIIa, which appeared as molecular weight polymorphism.

Recently, several observations suggested that the cysteine-rich repeat of GPIIIa (β3 integrin) plays an important role in the regulation of α_IIbβ₃ functions. Sequence alignment of β₃ integrins indicates that about 90% of noncysteine residues in the cysteine-rich repeats are conserved between mouse, rodent, canine, and human. In contrast, these noncysteine residues are poorly conserved among β₁, β₂, and β₃ integrins.9,32 Wippler et al33 demonstrated that recombinant α_IIbβ₃ lacking the cysteine repeats of β₃ showed a high-affinity binding to fibrinogen. Recently, Kashiwagi et al34 demonstrated that a point mutation (Thr₅₆₂Asn) in the extracellular cysteine-rich repeat region is directly involved in integrin structural changes during inside-out signaling. This mutation resulted in a constitutive activation of α_IIbβ₃ and α_Vβ₃ integrins leading to spontaneous binding of soluble fibrinogen.

To study the influence of ΔLys₆₁₁ on GPIIb-IIIa receptor function, we examined ligand binding and postligand binding events of GPIIb-IIIa in CHO cells. Using LIBS antibody we could demonstrate that the ΔLys₆₁₁ GPIIIa isoform was not constitutively active and was capable of undergoing conformational changes on ligand binding like the wild-type GPIIIa form. In addition, the ΔLys₆₁₁GPIIIa could mediate downstream signaling events such as phosphorylation of pp125^FAK. These data suggest that the Lys₆₁₁ deletion has no effect on integrin function.

Ten different genetic variants of the GPIIIa molecule have been identified (Figure 10). The GP3A*01 allele encoding for the GPIIIa Leu₃₃ isoform, which carries the HPA-1a alloantigenic determinants, is the most frequent allele with a gene frequency of nearly 85% within the white population. This allele differs from the second most common form of GP3A*02 (HPA-1b, frequency ∼15%) by a single amino acid substitution (Leu₃₃Pro). Other alleles of GPIIIa (GP3A*03, *04…) are much less frequently represented and appear to have arisen from the high frequency allele GP3A*01 by independent point mutations, a scenario for the evolution of human platelet alloantigens that has been proposed by Newman and Nathalie.35 From the higher frequency of GP3A*01 in human and sequence comparison with rodents and Xenopus, the authors hypothesized that this variant comprises the ancestralGPIIIa gene. Unfortunately, the evolutionary relationships of the HPA-1 alloantigen system could not be directly examined, because these species carry neither leucine nor proline at position 33.32,36 Recently, Lipscomb et al9 reported that the deduced amino acid sequence of canine GPIIIa has leucine at position 33 and therefore corresponds to the high frequency human GP3A*01. In this study, we could demonstrate that GPIIIa of nonhuman primates also carry Leu₃₃ supporting the model for the GPIIIa allelic generation as described above.

Recently, we have reported a Leu₄₀Arg dimorphism on GPIIIa that was associated with HPA-1b, but so far no corresponding alloantibody has been detected.24 Interestingly, we observed that the Oe^a alloantigen also segregated with the HPA-1b allele.

By the genotyping analysis we could exclude a relation between Oe^a and the Leu₄₀Arg dimorphism. Therefore, 2 independent mutations of the HPA-1b allele have occurred. The proportional distribution of rare mutations among the more frequent GP3A*01 and *02 alleles (7 versus 2) suggests that mutational events occurred randomly during evolution and no further selection took place (Figure 10).

Point mutations or a deletion (in this study) responsible for the formation of human platelet alloantigens on GPIIIa were found in different domains of the molecule: HPA-1 and HPA-10 in the amino terminal domain, HPA-4 in the ligand-binding domain, HPA-6 and Oe^a in the cysteine-rich repeat domain, and HPA-8 and HPA-11 in the extracellular C-terminal domain. To date, comparative functional analysis has been performed for the HPA-1 and HPA-4. Whereas the HPA-4 polymorphism does not seem to have an influence on GPIIb-IIIa function, some data on the functional relevance of different HPA-1 alleles have been reported.26,14 From these and our results, one could speculate that the alloantigenic polymorphism of GPIIIa does not have a major influence on integrin function.

The authors thank Dr K. Adam, Children Hospital, Weiden, Germany, who kindly referred this NAIT case to us. We are grateful to M. Ernst-Schlegel, M. Boehringer, S. Werth, and A. Wittchen for their technical assistance. We are grateful to Dr P. Newman for his support. Our gratitude is also extended to the families concerned for their cooperation in this study. This work is part of the doctoral theses of A. Rachman and B. Carl.

In agreement with the policy of ISBT/ISTH working party on nomenclature of platelet alloantigens, materials have been sent to another expert laboratory for confirmation. The existence of the new human platelet alloantigen Oe^a could be confirmed by Dr B. Curtis and Dr R. H. Aster (Blood Research Institute, Milwaukee, WI) by analyzing maternal serum, paternal platelets, and DNA.