A Point Mutation Thr₇₉₉Met on the ₂ Integrin Leads to the Formation of New Human Platelet Alloantigen Sit^a and Affects Collagen-Induced Aggregation

IMMUNIZATION AGAINST human platelet antigens (HPA) is the central event in the pathophysiology of neonatal alloimmune thrombocytopenia (NAIT) and in posttransfusion purpura (PTP) and is also of importance in some patients refractory to platelet transfusions.1 

Currently, 5 HPA systems, HPA-1 (Pl^A, Zw), -2 (Ko, Sib), -3 (Bak, Lek), -4 (Pen, Yuk), and -5 (Br, Zav), are officially recognized. In addition, a number of low frequency alloantigens, HPA-6bW (Ca^a, Tu^a), -7bW (Mo^a), -8bW (Sr^a), -9bW (Max^a), -10bW (La^a), Gro^a, and Iy^a, have also been reported. Meanwhile, the alloantigenic determinants of all of these HPAs could be localized on platelet membrane glycoprotein (GP) Ia, GPIb_α, GPIb_β, GPIIb, and GPIIIa. All of these have been found to result from point mutations in the encoding genes, which lead to single amino acid substitutions.2-13 

GPIIb/IIIa, also known as the integrin α_IIbβ₃,14 which serves as fibrinogen receptor on platelets, carries the majority of these HPAs.

Another integrin, α₂β₁ (GPIa/IIa), which mediates platelet adhesion to both fibrillar and nonfibrillar collagen,15,16 is known to bear the alloantigenic determinant of the HPA-5.17 Recently, we could demonstrate that the HPA-5a, -5b phenotypes are associated with a single base G₁₆₄₈A substitution of the GPIa gene leading to an aminoacid Glu₅₀₅Lys dimorphism.6Alloimmunization against HPA-5 has been associated with NAIT,18 PTP,19 platelet transfusion refractoriness,20 and thrombocytopenia due to passively transferred platelet antibodies after transfusion of blood products21 or bone marrow transplantation.22 

In addition, the important functional role of GPIa in primary hemostasis is stressed by the observation of congenital and acquired bleeding tendency in patients with deficient GPIa and with autoantibodies. In vitro studies showed impaired reactivity of platelets with collagen in all these cases.23-26 

Current evidence suggests that adhesion and activation of platelets by collagen are mediated through distinct receptors. Activation of platelets by collagen is believed to occur via a 2-site, 2-step model in which GPIa/IIa provides an initial interaction (adhesion) that brings a second site in the collagen molecule into the activitory receptor GPVI.27 

Recently, Kunicki et al28 were able to identify clusters of linked silent polymorphisms within the coding sequence of the α₂ gene that which are responsible for the variation in α₂β₁ receptor density on the platelet surface. Three allelic differences could now be identified by 4 dimorphisms at positions 807, 837, 873, and 1648. Allele 1 (T₈₀₇/T₈₃₇/A₈₇₃/G₁₆₄₈) is associated with increased levels of α₂β₁, whereas allele 2 (C₈₀₇/T₈₃₇/G₈₇₃/G₁₆₄₈) and allele 3 (C₈₀₇/C₈₃₇/G₈₇₃/A₁₆₄₈) are associated with lower levels of α₂β₁.29 Carriers of the allele 1 express high levels of GPIa/IIa, whereas individuals who carry alleles 2 and 3 exhibit lower expression of the platelet integrin. Furthermore, Kritzig et al29 could demonstrate that the density of GPIa/IIa on the platelet surface correlates with the rate of platelet attachment in whole blood to type I collagen. A clinical impact of these findings was demonstrated by the association of the GPIa C₈₀₇T gene polymorphism with nonfatal myocardial infarction and stroke in younger individuals.30 31 

In this report, we describe another case of alloimmunization against a new genetic variant of GPIa that was responsible for a case of severe NAIT, and we describe the biochemical, molecular, biological, and functional properties of platelets carrying this antigen.

A healthy 19-year-old white mother (Sit) gave birth to her first child after a normal pregnancy. The male child (2,150 g) was delivered at term with an Apgar score of 8. At birth, a severe thrombocytopenia was observed (31 × 10⁹/L) and the platelet count decreased continually below measurable value. Clinical and laboratory examinations showed no other abnormalities. Petechiae were observed at the age of 4 days and a tentative diagnosis of NAIT was made. Because maternal platelets were not available, the infant was transfused with platelets from a healthy donor prepared by platelet apheresis with good increment.The platelet count increased to 158 × 10⁹/L. In addition, the child was treated with high doses of intravenous IgG from day 3 to 5 of life and with prednisolone from day 2 to 10 of life. After a moderate decrease of the platelet count to 67 × 10⁹/L, the child was discharged with a normal platelet count on day 18. Ultrasound examination showed no signs of cerebral hemorrhage.

Blood samples of the mother, the father, and the child were referred to us 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. Normal donor platelets for the characterization of platelet antibodies were selected from a large number of donors with known human platelet alloantigens (HPA 1-5) and ABO blood group antigens.

Phenotyping of human platelets was performed using the glycoprotein-specific immunoassay (monoclonal antibody-specific immobilization of platelet antigens [MAIPA]), as previously described.32 

Antibodies against Br^a and Br^b alloantigens were obtained from mothers of children with NAIT and from polytransfused patients, respectively.17 Monoclonal antibody (MoAb) Gi5 and MoAb Gi9 specific for GPIIb/IIIa complex (CD41) and GPIa (CD49b), respectively, were produced and characterized in our laboratory.6 MoAb SAM1 directed against the GPIc (CD49e) subunit and MoAb B1G6 specific for β₂ microglobulin of HLA class I molecule were purchased from Immunotech (Hamburg, Germany). MoAb FMC 25 against the GPIX (CD42a) subunit of GPIb/IX/V complex was a generous gift of Dr H. Zola (Adelaide, Australia).

Platelets and Chinese hamster ovary (CHO) stable transfectants were surface-labeled with 5 mmol/L NHS-LC-Biotin (Paesel, Frankfurt, Germany) and were immunoprecipitated as previously described.9 

Total platelet RNA was isolated from 20 mL EDTA anticoagulated blood of phenotyped donors by a modification of the guanidium isothiocyanate method as previously described.4 Platelet RNA was resuspended in 100 μL diethyl pyrocarbonate (DEPC)-treated H₂O and stored at −70°C. Genomic DNA was obtained from peripheral blood leukocytes derived from 3 mL EDTA anticoagulated blood after removing platelet-rich plasma. The DNA was extracted by the salting out procedure as described by Miller et al,33dissolved in 300 μL TE buffer (10 mmol/L Tris, 1 mmol/L EDTA, pH 8.0), and stored at 4°C.

To amplify the entire coding region of GPIa by polymerase chain reaction (PCR), 8 overlapping sets of primers were constructed based on the published cDNA sequence of GPIa.34 For cDNA amplification of a region encompassing nucleotides 2304-2858, forward primer P48 (2212-GAAAGGTGCC TGCAGAAG-2229), reverse primer P40 (2858-TCAGCATCATACAGGAGAGG-2839), and nested primer P18 (2304-CTCTTTGGATTTGCGTGTGGACATC-2328) were used. In addition, primer pair P61 (2414-GTGGTGAGGATGGACTTTGC-2433), P42 (2572-CAATTCCAGTGTTGTATGCAC-2552) and primer pair P6 (2551-AGTGCATA CAACACTGGAAT-2570), P7 (2699-TTTAAAGCAGGGTAGCCTAC-2680) were used to amplify GPIa gene from genomic DNA.

One hundred microliters of platelet RNA were heated to 68°C for 10 minutes and quickly cooled on ice water. The first-strand cDNA was synthesized using 10 μmol/L oligo dT, 40 U RNAse inhibitor (Boehringer Mannheim, Mannheim, Germany), 2 mmol/L of each dNTP (Pharmacia, Freiburg, Germany), 500 U Moloney's murine leukemia virus (MMuLV) reverse transcriptase, and 5× enzyme buffer (GIBCO BRL, Eggenstein, Germany) for 40 minutes at 45°C in a total volume of 30 μL and was stopped by chilling to 0°C. Aliquots of 20 μL of cDNA were preheated to 90°C for 5 minutes and were chilled to 0°C. After digestion with 2 U of RNAseH enzyme (GIBCO BRL) at 37°C for 20 minutes, cDNA was stored at −20°C. Five microliters of cDNA were diluted with 10× PCR buffer, 0.3 μmol/L of each primer (P48, P40), 175 μmol/L dNTP, and 2.5 U Taq Gold polymerase (Perkin Elmer, Norwalk, CT) in a total volume of 50 μL. Amplification was performed on DNA thermal cycler 480 (Applied Biosystem, Weiterstadt, Germany) for 20 cycles. Each cycle consisted of denaturation at 93°C for 1 minute, annealing at 53°C for 1 minute, and extension at 72°C for 2 minutes. In the final cycle, the samples were kept at a temperature of 72°C for 10 minutes and then chilled to 4°C. Two-microliter aliquots of PCR products were reamplified for 30 cycles using nested primer P18 and reverse primer P40 under identical PCR conditions. PCR products were analyzed on 1.6% SeaKemGTG agarose gel (Biozym, Hamel, Germany) containing ethidium bromide.

Two hundred nanograms of genomic DNA were amplified using primer pairs P61, P42 and P6, P7 as described above. Thirty-two cycles of denaturation at 94°C for 1 minute, annealing at 58°C for 1 minute, and extension at 72°C for 2 minutes were performed.

Amplified DNA was purified on SeaKemGTG gel (Biozym) by QIAquick (Qiagen, Düsseldorf, Germany). Purified DNA was trimmed using Klenow DNA polymerase (Biolabs, Schwalbach, Germany) and were subcloned into the EcoRV cloning site of the plasmid vector pGEM-5Zf (Promega Biotech, Madison, WI) and then transformed into the DH5α high efficiency competent Escherichia coli (GIBCO BRL). Plasmid DNA was prepared for nucleotide sequencing analysis by QIAprep (Qiagen). The inserts from 8 positive clones were sequenced using Taq-FS Dye-Terminator Cycle Sequencing Kit according to the manufacturer's protocol and were analyzed on ABI PRISM Genetic Analyzer 310 (Applied Biosystem). SP6 and T7 primers that hybridize to vector sequence and flank the cloned insert were used for sequencing.

The DNA fragment of interest was gel purified from 50 μL PCR reaction and directly sequenced as described above. PCR primers were used as sequencing primers.

For genotyping of Sit^a alloantigen, 3 to 5 μg genomic DNA was amplified using forward primer P61 (see above) and an intronic reverse primer P83 5′-TACCGGTAGGGAGAATGATGC-2602-3 under the following PCR conditions: 30 cycles at 94°C for 1 minute, 58°C for 1 minute, and 72°C for 2 minutes, followed by final extension at 72°C for 10 minutes. Aliquots of 3 μL of PCR products were digested in a thermal cycler with 2 U Mae III endonuclease (Boehringer Mannheim) for 6 hours at 55°C, respectively. Restriction fragments were analyzed on 1.6% agarose gel using Tris borate buffer system (Biozyme). Genotyping for HPA-5 was performed by RFLP using Mnl I endonuclease as previously described.35 

Genotyping of the C₈₀₇T dimorphism was performed using PCR-SSP as recently described.30 

A full-length human GPIa cDNA (A₁₆₄₈AG/AC₂₅₃₁A) in the mammalian expression vector pMPSV,36which encodes Lys₅₀₅Thr₇₉₉ GPIa isoform, was kindly provided by Dr E. Klein (Department of Dermatology, University of Würzburg, Würzburg, Germany). An allele-specific recombinant form G₁₆₄₈AG/AC₂₅₃₁A encoding the Glu₅₀₅Thr₇₉₉ GPIa isoform was produced by cartridge mutagenesis. A 1,435-bp cDNA fragment spanning nucleotides 1486 to 2921 of platelet GPIa mRNA derived from a Br^bhomozygous donor was amplified by PCR.6 After digestion with Bgl II and BstEII endonucleases (Biolabs), the 1,374-bp fragments containing the polymorphic base G₁₆₄₈were then shuttled into the pMPSV expression vector containing the GPIa insert (A₁₆₄₈AG/AC₂₅₃₁A), which had been digested with the same enzymes. To exclude contamination of undigested plasmid, the pMPSV expression vector was linearized with Swa I endonuclease (Biolabs) before ligation. After subcloning in E coli, the resulting plasmid constructs were amplified in PCR (bases 1486-1900) and were screened for A₁₆₄₈G dimorphism by RFLP analysis using Mnl I endonuclease. Specific mutations CA→TG at positions 2531 and 2532 were introduced into the Lys₅₀₅Thr₇₉₉ and Glu₅₀₅Thr₇₉₉ constructs by site-directed mutagenesis using QuickChange Mutagenesis Kit (Strategene, Heidelberg, Germany).

For PCR amplification, 2-nucleotides (underlined) mismatched forward primer 5′-GTTAACATTTTCAGTAATGCTGAAAAATAAAAGGG-3′ and reverse primer 5′-CCCTTTTATTTTTCAGCATTACTGAAAATGTTAACC-3′ from base 2513 to 2548 of GPIa cDNA were constructed. After denaturation for 30 seconds at 95°C, aliquots of 20 ng plasmid were amplified for 12 cycles (denaturation for 30 seconds at 95°C, annealing for 60 seconds at 55°C, and extension for 17 minutes at 68°C). PCR products were digested with Dpn I endonuclease for 1 hour at 37°C and transformed into DH5α high efficiency competent E coli. Plasmid DNA from positive clones were amplified by PCR using primers P18 and P40. PCR products were subjected to RFLP analysis with Mae III as described above. For subsequent transfection studies, all GPIa allele-specific constructs were validated by nucleotide sequence analysis.

CHO (American Type Tissue Collection, Rockville, MD) cells were grown in RPMI 1640 medium supplemented with 10% fetal calf serum (FCS; Seromed, Berlin, Germany), 1% sodium pyruvate, 1% glutamine, and 1% penicilline/streptomycine (GIBCO BRL; complete medium) and were transfected with allele-specific GPIa constructs by the use of the reagent Lipofectin (GIBCO BRL). In brief, 6 μg of the GPIa construct was mixed with 25 μL Lipofectin in 2 mL OptiMEM Medium (GIBCO BRL) and then added to a subconfluent 10-cm plate of CHO cells for 12 hours. For antibiotic selection, cotransfection with 6 μg RSV neo plasmid (Invitrogen, Leek, Netherlands) containing Neomycin resistent gene was performed. Nine milliliters of complete medium was added and the incubation was continued for 48 hours. After splitting, CHO transfectant was then selected with Genicitin (G418; final concentration, 1 mg/mL; GIBCO BRL) for approximately 2 weeks. Positive clones expressing GPIa/IIa were enriched by adhesion of 3 × 10⁶ stable transfectant onto 3.6-cm petri dishes coated with collagen type I (5 μg/mL of 0.1 mol/L NaHCO₃). After subcloning, the surface GPIa/IIa expression was analyzed by flow cytometry (see below). Stable cell lines were grown in complete medium supplemented with 200 μg/mL of G418.

Stable transfectants were harvested with trypsin-EDTA (GIBCO BRL), washed in phosphate-buffered saline (PBS; GIBCO BRL), and fixed with 1% paraformaldehyde for 3 minutes. After three washes, 450 μL (10⁶/μL) of cell suspensions were incubated with 50 μL MoAb of Gi14 (20 μg/mL) or 20 to 30 μL human sera (anti-Br^a, anti-Br^b, and anti-Sit^a) for 30 minutes at room temperature. Labeled cells were then washed twice, stained with 500 μL fluorescein isothiocyanate-conjugated rabbit antimouse IgG or antihuman IgG (dilution 1:40; Dako, Hamburg, Germany), and analyzed by flow cytometry (Ortho Diagnostic, Neckargemünd, Germany).

Adhesion of platelets to immobilized collagen was investigated as recently described.37 Platelets were isolated from ACD-anticoagulated blood by differential centrifugation and washed with AMPL buffer (140 mmol/L NaCl, 5 mmol/L KCl, 1 mmol/L MgCl₂, 10 mmol/L glucose, 10 mmol/L sodium pyruvate, 5 mmol/L sodium malate, 10 mmol/L HEPES, and 0.3 mmol/L albumin, pH 7.4) containing 0.02 U/mL apyrase and 5 U/mL hirudin (Paesel & Lorey, Frankfurt, Germany). Aliquots of 1 × 10⁹ washed platelets in PBS (137 mmol/L NaCl, 2 mmol/L KCl, 8 mmol/L Na₂HPO₄, and 1 mmol/L KH₂PO₄, pH 7.4) were labeled with 16 μL 75 mmol/L NHS-LC-Biotin (Paesel) at room temperature for 15 minutes. Labeled platelets were washed twice with 5 mL AMPL buffer and resuspended in PBS at a final concentration of 3 × 10⁸/mL. For adhesion, microtiter wells were coated overnight with collagen type I, III, or V (50 μg/mL; Sigma, Dreieich, Germany) or bovine serum albumin (BSA; 10 mg/mL; Sigma), washed 3 times with 200 μL PBS, and blocked with 200 μL 1% BSA in PBS for 1 hour at 37°C. Aliquots of 100 μL of biotinylated platelets were added in triplicate to wells coated either with BSA or collagen and were permitted to adhere at 37°C for 1 hour. To evaluate the functional effect of anti-Sit^a alloantibodies, platelets were mixed with either 40 μL affinity-purified AB-sera or of anti-Sit^a for 30 minutes in a 5% CO₂atmosphere. Nonadherent cells were removed by gentle absorption onto absorptive pads and by washing of the wells 3 times with 150 μL PBS. Bound platelets were detected by the addition of 100 μL alkaline phosphatase-labeled streptavidin (500 μg/mL, dilution 1:1,500; Dianova) and incubation at 37°C for 1.5 hours. After a further 3 washes with PBS and one wash with diethanolamine buffer, pH 9.8, 100 μL of p-nitro-phenylphosphate (1 mg/mL; Sigma) was added. The reaction was stopped after 5 minutes at room temperature by the addition of 100 μL of 1.0 mol/L NaOH and was read in an enzyme-linked immunosorbent assay (ELISA) plate reader (OD₄₀₅, Titertex Multiskan; Pharmacia).

Platelet-rich plasma (PRP) of ACD anticoagulated blood from Sit^a-phenotyped individuals was ajusted to 3 × 10⁵ platelets/μL by dilution with autologous plasma. To aliquots of 180 μL PRP, 20 μL of various dilutions of ADP (5, 10, and 25 μmol/L; Sigma) or collagen (2.5, 5, and 10 μg/mL; Nycomed, Munich, Germany) were added and the change in the optical density was monitored using APACT aggregometer (Labor Timer, Ahrensburg, Germany) with continuous stirring at 37°C. In some instances, PRP was incubated with anti-Sit^a alloantibodies as described above. After stimulation with 10 μg/mL collagen, platelet aggregation was recorded as described above.

When serum from the mother Sit was tested against autologous, paternal, or donor panel platelets using a platelet adhesion immunofluorescence test, a positive reaction was only observed with paternal platelets (data not shown). An IgG antibody from serum Sit^a reacted strongly with GPIa/IIa immobilized by MoAb Gi9 from paternal platelets only, but not with epitopes on GPIc/IIa, GPIIb/IIIa, GIb/IX, and HLA class I (Fig 1). Platelets from 11 typed panel donors, including those of Br^a and Br^bhomozygous individuals, did not react either. These results indicate that the Sit^a serum recognizes a new low-frequency platelet alloantigen residing on GPIa/IIa complex. Because the Sit^aalloantigenic determinants could be immobilized by MoAb directed against GPIa/IIa only, but not by MoAb specific for GPIc/IIa complex, Sit^a epitopes are probably localized on the GPIa subunit.

In a population of 400 unselected donors, GPIa/IIa of 1 individual (Dre) carried the Sit^a antigen. Figure 2 shows the pedigrees of the family Dre and the index family Sit. Phenotyping analysis of both families demonstrated that Sit^a antigen is inherited as an autosomal dominant trait. In addition, all Sit^a (+) individuals were also phenotyped for Br^a and Br^b(Table 1). Three Br^b homozygous Sit^a (+) individuals (B.II.2, II.4, and III.3) were found (Fig 2B) in family Dre, indicating that Sit^a alloantigen is inherited with the Br^b allele.

To confirm the localization of Sit^a alloantigen on GPIa/IIa, immunoprecipitation analysis of biotin-labeled glycoproteins derived from Sit^a (+) and Sit^a (−) Br^a/Br^b heterozygous individuals was performed. As shown in Fig 3, anti-Sit^a(lanes 1 and 4) only precipitated GPIa/IIa complex from Sit^a (+) platelets, but not from Sit^a (−) platelets. In the control experiments, anti-Br^a (lanes 2 and 5) and anti-Br^b (lanes 3 and 6) precipitated GPIa/IIa complex from both Br^a/Br^b heterozygous individuals.

To analyze the nucleotide sequence encoding for GPIa, platelet mRNA was sequentially amplified by reverse transcription using 8 sets of primers, subcloned, and sequenced. Nucleotide sequence analysis of the 554 bp encompassing nucleotides 2304-2858 derived from a Sit^a (+) individual showed 2-base mutations C₂₅₃₁A_{2532 →}T₂₅₃₁G₂₅₃₂(Fig 4) in 10 of 12 subcloned examined, consistent with the hypothesis that all Sit^a (+) individuals found to date are heterozygous for this character. These mutations predicted Thr₇₉₉(AC₂₅₃₁A₂₅₃₂) in the Sit^a (−) and Met₇₉₉ (AT₂₅₃₁G₂₅₃₂) in the Sit^a (+) phenotype. Analysis of the other 7 regions of GPIa mRNA showed no other nonconservative nucleotide differences.

To establish genomic DNA typing, we first elucidated the exon-intron boundaries surrounding the Sit polymorphic base by PCR. When PCR with genomic DNA was performed using exonic primer pairs P61, P42 and P6, P7, fragments of 580 bp and nearly 1,400 bp were obtained. Nucleotide sequence analysis of both PCR products identified 421-bp and 1.2-kb introns (Fig 5) with conserved donor and acceptor splice junctions, which are classified as phase 2 and 0, respectively. These results demonstrate that these 2 introns enclose a 142-bp (bases 2478-2619) exon encoding 47 amino acids. Based on these intron sequences, reverse primer P83 downstream of the 142-bp exon was constructed. After 35 cycles of amplification using primer P61 as forward primer, the expected 630-bp product was obtained from genomic DNA derived from peripheral blood lymphocytes (Fig 6). The PCR product was digested withMae III endonuclease, which cleaves 5′-GTAAC-3′ (wild-type) but not 5′-GTAAT-3′ sequences (mutant). Sit^a-negative individuals could be clearly distinguished from Sit^a heterozygous individuals by the absence of the uncut 630-bp band. The results of genotyping by PCR-RFLP of 400 Sit^a-phenotyped donors were consistent with the serological phenotyping.

Because different triplets theoretically could code the wild-type Thr₇₉₉ (ACA, ACC, ACG, and ACT), we analyzed the members of the families Sit and Dre (Table 1) by direct nucleotide sequencing. Two different codons, ACA and ACG (Fig 4), both of which encode the wild-type Thr₇₉₉ form of GPIa, were found among Sit^a (−) homozygous and Sit^a (+) heterozygous individuals (Table 1). To determine the distribution of the codon usage for Thr₇₉₉, we analyzed 22 unselected blood donors by DNA sequencing. Nineteen individuals presented with codons ACG/ACG, 2 with ACG/ACA, and 1 with ACA/ACA. Our results suggest that the codon usage for Thr₇₉₉ is 90.9% ACG and 9.1% ACA. Other conceivable codons (ACC and ACT) were not found in our population so far. It is of note that the Thr₇₉₉ ACA allele is uniquely recognized by the restriction enzyme TspRI and therefore represents a potentially useful genetic marker for RFLP (data not shown).

Thus, the mutation C₂₅₃₁T in Sit^a (+) individuals most probably occurred in the high-frequeny codon ACG (Thr) to ATG (Met) due to a 1-point mutational event.

To analyze the linkage associations between C₂₅₃₁T with A₂₅₃₂G, A₁₆₄₈G, and C₈₀₇T polymorphisms, we genotyped all members of the families Sit and Dre for A₁₆₄₈G, C₈₀₇T, and A₂₅₃₂T dimorphisms by PCR-RFLP, PCR-SSP, and direct nucleotide sequencing analysis, respectively. Their genotypes are summarized in Table 1.

In accordance with our previous phenotyping results by MAIPA assay, we could confirm that the Sit^a allele segregated together with the Br^b allele. Four Sit^a (+) individuals (A.II.1, B.II.2, B.II.4, and B.III.3) are homozygous for the G₁₆₄₈ allele.

Furthermore, we observed that all Sit^a (+) individuals were homozygous for the C₈₀₇, indicating the linkage association between the Sit^a allele with the C₈₀₇ allele.

Finally, we found that the G₁₆₄₈ haplotype exclusively segregated with the G₂₅₃₂ allele. All Br^bhomozygous family members are homozygous for the G₂₅₃₂allele. In contrast, the A₁₆₄₈ haplotype most probably segregates with the A₂₅₃₂ allele, because all Br^a heterozygous individuals were found to be heterozygous G/A at position 2532.

These observations indicate a linkage disequilibrium between polymorphisms at positions 807, 1648, 2531, and 2532. Under the consideration of the linkage associations between the polymorphisms at bases 807/837/873,28 we could now define 4 α₂gene alleles by 6 dimorphisms at positions 807/837/873, 1648, 2531, and 2532 (Table 2).

To examine the involvement of amino acid 799 in the formation of the Sit^a antigenic determinants, we transfected allele-specific expression vectors encoding for the Lys₅₀₅Thr₇₉₉, Glu₅₀₅Thr₇₉₉, and Glu₅₀₅Met₇₉₉ GPIa into CHO cells. Stable transfectants expressing GPIa recombinant proteins were surface-labeled with biotin and analyzed by immunoprecipitation. As shown in Fig 7, anti-Sit^a (lanes 2) precipitated exclusively the Met₇₉₉ GPIa isoform, but not both Thr₇₉₉ isoforms having either glutamic acid or lysine at position 505. The associated band (110 kD) most probably represents endogenous hamster β₁ integrin subunit, which is coprecipitated with the human GPIa subunit (Mr 150 kD). In the control experiment, anti-Br^b (lanes 1) reacted with Thr₇₉₉ and Met₇₉₉ GPIa isoforms, which have the amino acid glutamic acid at position 505. In contrast, anti-Br^a (lane 3) only recognized the Lys₅₀₅GPIa isoform.

These findings demonstrate that amino acid Met₇₉₉ directly controls the expression of the Sit^a epitopes, whereas residue 505 (lysine or glutamic acid) is responsible for the formation of Br^a and Br^b alloantigenic determinants, respectively.

To determine the possible effects of Thr₇₉₉Met mutation on platelet function, we studied platelet ad- hesion and platelet aggregation of Sit^a-phenotyped individuals.

The results of platelet adhesion to immobilized collagen type I of Sit^a phenotyped platelets were summarized in Table 3. The platelet adhesion response to high collagen (50 μg/mL) and low collagen concentration (1 μg/mL) of Sit^a (+) platelets was indistinguishable from Sit^a (−) platelets.

Furthermore, we analyzed the effect of anti-Sit^aalloantibodies on platelet adhesion. In comparison to the control experiment with MoAb Gi9 specific for a functional epitope of the GPIa/IIa complex, anti-Sit^a did not inhibit platelet adhesion to type I collagen. Similar results were obtained with type III and V collagens (data not shown).

However, in standard aggregation assay (Fig8), the platelet aggregation response to low collagen concentration (2.5 μg/mL) in the Sit^a (+) individual (B.III.3, see Fig2) was diminished in comparison to the Sit^a (−) individual (B.III.2). In contrast, the platelet aggregation of the Sit^a (+) individual after stimulation with high concentration of collagen (10 μg/mL) was indistinguishable from the Sit^a (−) individual. The addition of other agonists (ADP and ristocetin) to platelets from Sit^a (+) individuals resulted in completely normal aggregation response (data not shown).

Similar results were observed with 3 other Sit^a (+) individuals (B.II.1, B.II.2, and B.II.4; data not shown). Because GPIa of both individuals (B.III.3 v B.III.2) only differs at nucleotide 2531 (T or C), but carries identical nucleotide sequences at polymorphic positions 807 (homozygous C) and 1648 (homozygous G), our results suggest that only the mutation Thr→Met at position 799 affects the function of GPIa/IIa as a collagen receptor. The differences in response to collagen thus cannot be explained by possible differences in GPIa expression density related to the base C₈₀₇T polymorphism.

When platelet aggregation by collagen was evaluated in the presence of purified anti-Sit^a, no inhibition was found (data not shown).

We report here the identification and characterization of a new low-frequency platelet alloantigen, Sit^a, in whites that was involved in a case of severe NAIT. A study in 400 unrelated individuals identified another Sit^a (+) individual, showing that this low frequency alloantigen is not restricted to a single family. Studies with immunochemical techniques allowed us to localize the Sit^a alloantigenic determinant on GPIa. Examination of the nucleotide sequence derived from a Sit^a (+) father showed a mutation comprising 2 bases CA→TG at positions 2531 and 2532, respectively, in a heterozygous state in the GPIa mRNA transcripts.

An analysis of the exon-intron structure allowed us to define the polymorphic site on an exon with a length of 142 bp. A comparison with other α subunits of integrin showed that this exon corresponds to exon 20 of the α_x and α_Mgenes.38 39 All 3 exons have identical length (142 bp) and are flanked by introns with identical phases (2 and 0).

Interestingly, we also discovered an additional silent polymorphism at base 2532 (G or A) adjacent to the polymorphic base 2531. By direct nucleotide sequencing of genomic DNA derived from Sit^a(−) individuals, we could define 2 different codons, ACG and ACA, that encode the wild-type Thr₇₉₉ form of GPIa.

Because the Sit^a alloantigen is a low-frequency antigen, it seems to represent a recent mutational event. According to the current observations that most polymorphisms responsible for platelet alloantigens are point mutations of the wild-type allele,40the Sit^a alloantigen most probably occurred due to the single base mutation C→T at position 2531 of the high-frequency wild-type allele ACG (90.9%) rather than the low-frequency ACA allele (9.1%).

Recently, Kritzig et al29 could define 3 allelic differences by 4 dimorphisms at positions 807, 837, 873, and 1648. Our data indicate that the A₂₅₃₂ allele may be linked to the C₈₀₇A₁₆₄₈ allele, whereas the G₂₅₃₂allele can be associated either with the C₈₀₇G₁₆₄₈ or T₈₀₇G₁₆₄₈allele. These observations are in accordance with the frequencies of C₈₀₇A₁₆₄₈ and A₂₅₃₂ genotypes in our population (10% v 9.1%). However, analysis of A₁₆₄₈ and G₁₆₄₈ genotyped individuals among larger populations is necessary to establish this association. Interestingly, the GPIa cDNA derived from fibroblast cell library also represented the rare C₈₀₇A₁₆₄₈A₂₅₃₂allelic form.34 

The mutation C₂₅₃₁T results in a substitution of the polar uncharged amino acid threonine into a hydrophobic nonpolar residue methionine at position 799 on mature GPIa polypeptide. Stable expression of recombinant allele-specific GPIa in CHO cells led to the confirmation that the single amino acid substitution Thr₇₉₉Met was sufficient to induce Sit^a epitope formation recognized by the Sit^a antibody present in the mother's serum. As expected from our previous observations,6 we could demonstrate in this study that the residue 505 (Lys or Glu) specifically controls the formation of Br^a and Br^b alloantigenic determinants, respectively.

Functional studies showed no influence of the Sit-polymorphism on platelet adhesion to immobilized collagen. In contrast, after stimulation with a low concentration of collagen, the platelet aggregation response of Sit^a (+) platelets was reduced as compared with the response of platelets from Sit^a (−) individuals.

Until now, the region comprising the amino acid 799 has so far not been implicated in the adhesion to collagen, which was mapped to the I-domain of the GPIa subunit.41,42 Recently, several lines of evidence suggest that activation of platelets occurs via a 2-site, 2-step model in which the initial interaction occurs via the GPIa/IIa, allowing binding to GPVI leading to activation.27 It may be speculated that the mutation Thr→Met at position 799 affects the activation of platelets, possibly through interaction between GPIa/IIa and GPVI.

Although the aggregation response to collagen in Sit^a (+) carriers was diminished, none of these individuals had signs of an altered hemostasis.

In NAIT, alloantibodies against the Br^a epitopes on GPIa often induces moderate thrombocytopenia in affected children.18 It is interesting to note that NAIT due to anti-Sit^a was associated with severe thrombocytopenia. It is conceivable that the additional functional defect of GPIa/IIa complex resulted in the pronounced clinical course of this child. An additional influence of anti-Sit^a alloantibodies on platelet adhesion and aggregation by collagen could not be observed.

In conclusion, we have identified that a single point mutation Thr₇₉₉Met adjacent to a hot spot in the GPIa gene is responsible for the formation of a new low-frequency platelet alloantigen, which we termed Sit^a. In contrast to other known platelet alloantigens, this point mutation appears to affect platelet function.

The authors thank Dr H. Schachinger, who kindly referred this NAIT case to us. We are grateful to Micaela Boehringer and Monika Kummel for their technical assistance. Our gratitude is also extended to the families concerned for their cooperation in this study.