I M M U N O H E M AT O L O G Y A new platelet alloantigen, Swia, located on glycoprotein Ia identified in a family with fetal and neonatal alloimmune thrombocytopenia _3038 1745..1754 Hartmut Kroll, Korinna Feldmann, Claudia Zwingel, Jochen Hoch, Rainer Bald, Gregor Bein, Behnaz Bayat, and Sentot Santoso BACKGROUND: Fetal and neonatal alloimmune thrombocytopenia (FNAIT) is a bleeding disorder caused by transplacental passage of maternal antibodies to fetuses whose platelets (PLTs) express the corresponding human PLT antigen (HPA). STUDY DESIGNS AND METHODS: We observed a fetus with FNAIT who died from a severe intracranial hemorrhage. Analysis of maternal serum in antigen capture assay with paternal PLTs showed reactivity with PLT glycoprotein (GP)IIb/IIIa (aIIbb3) and GPIa/IIa (a2b1 integrin), indicating the presence of anti-HPA-1a and an additional alloantibody against GPIa (termed anti-Swia). RESULTS: By immunochemical studies, the localization of the Swia antigen on GPIa/IIa could be confirmed. Analysis of paternal GPIa full-length cDNA showed a single-nucleotide substitution C3347T in Exon 28 resulting in a Thr1087Met amino acid substitution. Testing of family members by polymerase chain reaction-restriction fragment length polymorphism using MslI endonuclease showed perfect correlation with phenotyping. Extended family and population studies showed that 4 of 10 members of the paternal family but none of 500 unrelated blood donors were Swia carriers. Expression studies on allele-specific transfected Chinese hamster ovary (CHO) cells confirmed that the single-amino-acid substitution Thr1087Met was responsible for the formation of the Swia epitope. Adhesion of CHO cells expressing the Swia alloantigen to immobilized collagens was not impaired compared to the wild-type control and was not inhibited by anti-Swia alloantibodies. CONCLUSION: In this study we defined a new PLT alloantigen Swia that was involved in a case of additional immunization against HPA-1a. Our observations demonstrate that combinations of PLT-specific alloantibodies may comprise low-frequency alloantigens. F etal and neonatal alloimmune thrombocytopenia (FNAIT) is a severe bleeding disorder of the fetus and the newborn, which is caused by destruction of platelets (PLTs) by maternal alloantibodies during the pregnancy and after birth. The alloantibodies are directed against fetal PLT-specific antigens that are inherited from the father.1 The most common human PLT antigens (HPAs) responsible for the maternal immunization in the Caucasian populations are HPA-1 (approx. 70%), HPA-5 (approx. 20%), and HPA-2, -3, and -15 (approx. 10%).2-4 An increasing number of private or rare HPAs associated with FNAIT have been reported during the past two decades. This discovery was enabled with the introduction of a maternopaternal crossmatch strategy, that is, assessment of maternal serum with paternal PLTs, in a glycoprotein (GP)-specific immunoassay, for example, monoclonal antibody-specific immobilization of PLT ABBREVIATIONS: CHO = Chinese hamster ovary; FNAIT = fetal and neonatal alloimmune thrombocytopenia; GP = glycoprotein; HPA = human platelet antigen; MAIPA = monoclonal antibody–specific immobilization of PLT antigens; PRP = platelet-rich plasma. From the Institute for Transfusion Medicine Dessau, Red Cross Blood Transfusion Service NSTOB, Dessau; the Institute for Clinical Immunology and Transfusion Medicine, Justus Liebig University, Giessen; the Institute for Experimental Hematology and Transfusion Medicine, Bonn; and the Department for Prenatal Medicine, Clinics for Gynaecology and Obstetrics, Leverkusen, Germany. Address reprint requests to: Hartmut Kroll, MD, Institute for Transfusion Medicine Dessau, Red Cross Blood Transfusion Service NSTOB, 06847 Dessau, Germany; e-mail: hartmut.kroll@ bsd-nstob.de. Received for publication September 5, 2010; revision received November 28, 2010, and accepted November 29, 2010. doi: 10.1111/j.1537-2995.2010.03038.x TRANSFUSION 2011;51:1745-1754. Volume 51, August 2011 TRANSFUSION 1745 KROLL ET AL. antigens (MAIPA) assay.5 Meanwhile, 15 low-frequency HPAs (HPA-6bw through -14bw and -16bw through -21bw) have been assigned. Most of them reside on the GPIIb/IIIa (n = 12), one (HPA-12bw) on GPIbb, and two (HPA-13bw, -18bw) on GPIa (for comprehensive information see http://www.nibsc.ac.uk/ipd/hpa/). The GPIa/IIa complex or a2b1 integrin is expressed on numerous different cell types including PLTs and functions as a receptor for a number of matrix (collagens, laminin) and nonmatrix ligands (collectin/C1q, decorin, E-cadherin, and viruses).6 Four allelic forms of the GPIa gene have been described so far. Alleles 1 and 2 encode for HPA-5a, whereas the less frequent Allele 3 for HPA-5b.7,8 The transcript of the low-frequency Allele 4 carries HPA13bw.9 Recently, a further new rare GPIa allele encoding the HPA-18bw has been described.10 Alleles 1 and 2 are distinguished by a nonsense point mutation C to T at position 759 within the coding sequence of the GPIa gene. Carriers of the T759 allele express high levels of GPIa/IIa, whereas individuals carrying the C759 allele exhibit lower expression.11 Carriers of the T759 allele showed a higher risk for myocardial infarction.8 Interestingly, a linkage disequilibrium between the C759T dimorphism and the point mutations G1600A, G2235T, and C2483T associated with HPA-5, HPA-18bw, and -13bw systems has been observed.9,10 Recently, a new rare alloantigen Hita on GPIIIa, associated with FNAIT, has been reported.12 The Hita alloantigen is interesting because it represents the third immunogenic allele of the HPA-7 system and was found in 0.15% of the Japanese population.12,13 Third alleles of the HPA-1 and HPA-5 systems have already been described; however, alloantibody formations against these third allelic isoforms were not observed so far.14,15 In this study, we describe a case of the fatal FNAIT caused by hydrocephalus as a consequence of massive intracranial bleeding. In the maternal serum an alloantibody directed against a new alloantigen residing on PLT GPIa, Swia, was identified in addition to an anti-HPA-1a. CASE REPORT The first child of family Swi was born after an uneventful pregnancy from healthy parents with normal PLT counts (mother 338 ¥ 109/L, father 287 ¥ 109/L). The newborn did not show any signs of hemorrhagic diathesis; therefore, no serologic analysis was done. During the second pregnancy an ultrasound examination revealed normal findings at 29 weeks’ gestation. At 35 weeks’ gestation the mother realized that she was missing fetal movements. Upon sonography no cardiac action and a massive hydrocephalus were observed. A cranitomy showed approximately 300 mL coagula. Pathologic examination postpartum demonstrated that the fatal hydrocephalus was the consequence of massive intracranial hemorrhage. Analysis for PLT antibodies showed anti-HPA-1a and additional anti1746 TRANSFUSION Volume 51, August 2011 Swia. One year later the mother became pregnant again. At 24 weeks’; gestation fetal blood sampling revealed a PLT count of 7 ¥ 109/L. The pregnancy was managed by 12 weekly intrauterine PLT transfusions. Despite good increments the pretransfusion PLT count nadir during fetal blood samplings was 12 ¥ 109/L, indicating a strong clinical relevance of the maternal antibody. At 36 weeks’ gestation the child was delivered by cesarean section. Immediate PLT transfusion normalized the PLT count to 320 ¥ 109/L. The PLT count declined to 92 ¥ 109/L on Day 7 and then returned to normal values. No hemorrhagic diathesis was observed and the child was discharged in good health. MATERIALS AND METHODS Antibodies Antibodies against HPA-5a and -5b were obtained from mothers of children with FNAIT and from multiply transfused patients, respectively. Monoclonal antibodies (MoAbs) Gi5 and Gi9 against GPIIb/IIIa and GPIa/IIa complex, respectively, were produced and characterized in our laboratory.7 MoAb FMC25 against GPIb/IX complex was purchased from AbD Serotec (Oxford, UK). MoAbs against the b2-microglobulin subunit of HLA Class I antigen (Clone B1G6), a5 integrin subunit (Clone SAM-1), and b1 integrin subunit (Clone A1A5) were purchased from Immuntech (Hamburg, Germany). Characterization of PLT alloantibodies by antigen capture assay PLTs from the father and known HPA-phenotyped healthy blood donors were isolated from ethylenediaminetetraacetate (EDTA)-anticoagulated blood by differential centrifugation and stored at 4°C in isotonic saline containing 0.1% NaN3. Antibody detection was performed using MAIPA assay using a panel of MoAbs (see above), as previously described.16 Immunoprecipitation PLTs and stable transfected Chinese hamster ovary (CHO) cells were surface labeled with 5 mmol/L NHS-LC-Biotin (Pierce, Rockford, IL), and surface GPs were precipitated as previously described.17 A total of 100 to 300 mL of labeled cell lysates were incubated with 50 mL of serum or MoAb (20 mg/mL) overnight at 4°C in the presence of 100 mL of protein G beads (Pierce). After washings with immunoprecipitation buffer (50 mmol Tris, 150 mmol NaCl, 1% Triton X-100), bound proteins were eluted by adding sodium dodecyl sulfate (SDS) buffer for 5 minutes at 100°C. Eluates were analyzed on 7.5% SDSpolyacrylamide gel electrophoresis (PAGE) under reduc- PLT ALLOANTIGEN Swia ing conditions. Separated proteins were transferred onto nitrocellulose membranes and developed with peroxidase-labeled streptavidin and a chemiluminescence system (ECL, Amersham Biosciences, Freiburg, Germany). In some experiments, labeled PLT lysates were cleared three times with MoAb Gi5 (anti-GPIIb/IIIa complex) before immunoprecipitation with human serum. Sequence analysis of PCR products Amplification products from cDNA and from genomic DNA were separated by gel electrophoresis (SeaKemGTG, Biozym, Hess. Oldendorf, Germany), purified by a DNA purification kit (QIAquick, Qiagen, Hilden, Germany), sequenced with PCR primers using cycle sequencing technology (Taq-FS dye terminator, Applied Biosystems, Darmstadt, Germany) and analyzed on a genetic analyzer (ABI PRISM, Model 310, Applied Biosystems). Amplification of PLT RNA Total RNA was isolated from washed PLTs derived from 100 mL of EDTA-anticoagulated blood by the use of a kit for total RNA isolation (Roti Quick Kit, Roth, Karlsruhe, Germany) as recommended by the manufacturer. To amplify the entire coding region of GPIa by polymerase chain reaction (PCR), eight overlapping sets of primers were constructed based on the published cDNA sequence (NM_002203.3). For cDNA amplification of a region encompassing Nucleotides 3098 through 3772, forward primer V49 (2997-3077), reverse primer R5 (3772-3689), and nested primer V65 (3098-3120) were used. Aliquots of 30 mL of PLT RNA were reverse transcribed as recommended by the manufacturer (Ready-To-Go kit, GE Health Care, Munich, Germany). Five microliters of cDNA was diluted with 10¥ PCR buffer, 1.25 mL of each primer (V49, R5; 5 mmol/L), 2.5 mL of dNTP (1.25 mmol/L each nucleotide), and 1.25 U of Taq Gold polymerase (Perkin Elmer, Vaterstetten, Germany) in a total volume of 25 mL. Amplification was performed on a DNA thermal cycler (Model 480, Applied Biosystems, Weiterstadt, Germany). Conditions were as follows: initial denaturation at 96°C for 10 minutes; 25 cycles consisting of 96°C for 1 minute, 56°C for 1 minute, 72°C for 2 minutes; and final extension 72°C for 10 minutes. One microliter of the PCR products was amplified for 32 cycles using nested primer V65 and reverse primer R5 under identical PCR conditions. PCR products were analyzed on 1.6% agarose gel electrophoresis (Biozyme, Hamel, Germany) containing ethidium bromide. Amplification of genomic DNA Genomic DNA was extracted from EDTA-anticoagulated blood with a blood and tissue kit (DNeasy, Qiagen, Duesseldorf, Germany). One-hundred nanograms of genomic DNA was amplified using forward primer (nt 52382.547565) and reverse primer (nt 52383.067-047), 2.5 mL of dNTP (1.25 mmol/L each nucleotide), and 1.25 U Taq Gold polymerase (Perkin Elmer) in a total volume of 50 mL. Thirty-two cycles of denaturation at 95°C for 1 minute, annealing at 58°C for 1 minute, and extension at 72°C for 1 minute were performed. Genotyping of Swia alloantigen by restriction fragment length polymorphism Two microliters of genomic DNA (100 ng/mL) was amplified using intronic forward (Exons 28-108, 5′GCTCTTGTCACATTCCACCAGG-3′) and reverse primers (Exons 28 + 207, 5′-GGCATGGAGGAAGGCAAGTGTA-3′) using the following conditions: 30 cycles of denaturation at 93°C for 1 minute, annealing at 57°C for 1 minute, extension at 72°C for 1 minute, and final extension at 72°C for 10 minutes. Aliquots of 8 mL of PCR products were digested with 5 U of MslI endonuclease (New England Biolabs, Frankfurt am Main, Germany) for 4 hours at 37°C. Restriction fragments were analyzed on 1.6% agarose gel using Tris borate buffer system (Biozyme). Genotyping for HPA-1, -2, -3, -5, and -13bw and the GPIa C759T dimorphism was performed as previously described.8,18 Construction of GPIa allelic expression vector A full-length GPIa cDNA in the mammalian vector pMPSV encoding for GPIa Glu505Met1087 isoform (HPA-5a, Swia) was produced by site-directed mutagenesis using a commercially available kit (QuickChange, Stratagene, Heidelberg, Germany) as previously described.9 For PCR amplification, site-directed mutagenesis primers encompassing Nucleotides 3330 and 3364 of GPIa cDNA were constructed. After denaturation at 95°C for 30 seconds, aliquots of 20 ng plasmid were amplified for 12 cycles (denaturation at 95°C for 30 sec, annealing at 55°C for 60 sec, extension at 65°C for 12 min). PCR products were digested with Dpn endonuclease for 1 hour at 37°C and transformed into DH5a high-efficiency competent Escherichia coli bacteria (Invitrogen, Darmstadt, Germany). Plasmid DNA from positive clones was verified by nucleotide sequencing as described above. Stable transfection of Swia alloantigen in CHO cells CHO (American Type Tissue Collection, Rockville, MD) cells were grown and were transfected with allele-specific GPIa constructs encoding for Thr1087 or Met1087 isoform Volume 51, August 2011 TRANSFUSION 1747 KROLL ET AL. as previously described.9 Stable expressing cells were selected with Genicitin (G418, final concentration 1 mg/ mL; Gibco BRL, Eggenstein, Germany) and were enriched by adhesion onto petri dishes coated with collagen Type I. Subsequently, the expression of recombinant GPIa/IIa complex on the cell surface was measured by flow cytometry as previously described.9 Adhesion of stable transfected CHO cells onto collagen One milliliter of transfected CHO cells (5 ¥ 106 cells) in a-MEM (PAA, Marburg, Germany) was labeled with 35 mL of 2′,7′-bis-(2-carboxyethyl)-5-(and -6)carboxyfluorescein (Molecular Probes, Invitrogen) for 30 minutes at 37°C. Labeled cells were then washed twice with 5 mL of a-MEM medium and adjusted to a concentration of 1 ¥ 106/mL. Aliquots of 300-mL cell suspensions were treated with 150 mL of AB serum, 150 mL of anti-Swia, or 30 mL of MoAb Gi9 (30 mg) for 30 minutes at 37°C. For adhesion test, microtiter wells were coated overnight with collagen Type I (50 mg/mL; Sigma, Dreieich, Germany), bovine serum albumin (10 mg/mL; Sigma), or MoAb Gi14 (10 mg/mL); washed three times with 200 mL phosphatebuffered saline (PBS); and blocked with 200 mL of 1% bovine serum albumin (BSA) in PBS for 1 hour at 37°C. Aliquots of labeled cells were added in triplicate to wells coated with either BSA or collagen and were permitted to adhere at 37°C for 30 minutes. Nonadherent cells were removed by gentle aspiration and by washing of the wells two times. Bound cells were measured on a fluorescence microplate reader (Flx-800, Biotek, Neufahrn, Germany). PLT aggregation PLT-rich plasma (PRP) obtained from ACD-anticoagulated blood from Swia-positive family members and Swianegative individuals was adjusted to 3 ¥ 1011 PLTs/L by dilution with autologous plasma. To aliquots of 180 mL of PRP, 20 mL of collagen (Nycomed, Munich, Germany) in different concentrations was added (2.5, 5, 10 mg/mL) and light transmission was monitored using an aggregometer (APACT, Labor Timer, Ahrensburg, Germany) during continuous stirring at 37°C. For some experiments, PRP was incubated with anti-Swia alloantibody. RESULTS Serologic identification of anti-Swia When serum from the mother Swi was tested against paternal PLTs in the MAIPA assay, strong reactions were observed with immobilized GPIIb/IIIa and GPIa/IIa (Fig. 1A), but not with immobilized GPIb/IX and HLA Class I antigens. Analysis with a panel of PLTs with known 1748 TRANSFUSION Volume 51, August 2011 HPA-1, -3, -5, and -13bw phenotypes showed that maternal serum contained a strong anti-HPA-1a, but was not reactive with HPA-1b, -3a, -3b, -5a, -5b, and -13bw. Further evaluation in the MAIPA assay using capture MoAbs against different b1 integrins showed positive reactions when anti-a2b1 (Clone Gi9) and anti-b1 (Clone A1A5) were used but negative results with anti-a5b1 (Clone SAM-1; Fig. 1B). Taken together, these observations indicated that the maternal serum contained an anti-HPA-1a and an additional alloantibody against a new PLT alloantigen located on PLT GPIa (a2 subunit). Within the family of the father three further Swia-positive individuals were identified by MAIPA assay. A population study revealed no Swia-positive individual in a cohort of 500 healthy blood donors. Immunochemical investigations To confirm the localization of Swia alloantigen on the GPIa/IIa complex, immunoprecipitation analysis of biotin-labeled GPs from paternal PLTs was performed. As shown in Fig. 2A, maternal serum strongly precipitated the GPIIb/IIIa complex. In the control experiments, antiHPA-1a and anti-HPA-5a precipitated GPIIb/IIIa and GPIa/IIa complex, respectively. Maternal serum pulled down GPIa as a faint band and the expected GPIIa was superimposed by the presence of a strong GPIIbb band. However, after complete preclearing of GPIIb/IIIa from PLT lysates, precipitation with anti-Swia resulted in considerable detection of GPIa/IIa complex (Fig. 2B). Genetic analysis To analyze the nucleotide sequence encoding GPIa, paternal PLT mRNA was sequentially amplified by PCR after reverse transcription using eight sets of primers. Nucleotide sequence analysis of the 674-bp encompassing nucleotides 3098 through 3772 showed one nucleotide substitution C>T at Position 3347 (Fig. 3). Analysis of the other seven regions showed no other nonconservative nucleotide differences (data not shown). This result was confirmed by direct nucleotide sequencing of genomic DNA derived from other Swia-positive individuals (data not shown). This mutation predicted Thr1087 in Swianegative and Met1087 in Swia-positive individuals. The C3347T substitution is located at the end of Exon 28, two bases upstream from the exon-intron boundary, and creates a restriction site for MslI endonuclease (Fig. 4). Based on this knowledge PCR-restriction fragment length polymorphism (RFLP) strategy was developed for the genotyping of the Swia alloantigen. Swia-positive individuals were characterized by the presence of 212- and 150-bp restriction fragments, and in contrast Swia-negative individuals by a 362-bp band. Figure 4 shows the genotyping of the index family Swi. Five Swia-heterozygous individuals PLT ALLOANTIGEN Swia HPA father 1aa3aa A) IIb/IIIa 1aa3bb 1bb3ab father 5aa Ia/IIa 5bb 13ab father Ib/IX father HLA class I Gi9 (a2b 1) B) SAM1 (a5b 1) A1A5 (b 1) 0 2 1 3 Optical Density (492nm) Fig. 1. The reactivity of anti-Swia in antigen capture assay MAIPA assay. (A) Paternal PLTs and HPA-phenotyped PLTs (as indicated) were incubated with maternal serum and MoAbs against GPIIb/IIIa, GPIa/IIa, Ib/IX, or HLA Class I. (B) Paternal PLTs were incubated with anti-Swia ( ) or anti-HPA-5a ( ) and MoAbs against a2b1 (GPIa/IIa), a5b1 (GPIc/IIa), or b1 (GPIIa) integrin. After lysis, antigen-antibody complexes were immobilized on a microtiter plate. Bound human antibodies were measured by enzyme-linked immunosorbent assay using enzyme-labeled anti-human IgG (see Materials and Methods). A) GPIa GPIIa, GPIIbb B) GPI GPIa GPIIa GPIIIa 1 2 3 4 1 2 Fig. 2. Immunoprecipitation analysis of anti-Swia. (A) Paternal PLTs were surface labeled with biotin and lysed. Labeled cell lysates were precipitated with AB serum (Lane 1), maternal serum (Lane 2), anti-HPA-1a (Lane 3), or anti-HPA-5a (Lane 4). Immunoprecipitates were run on 7.5% SDS-PAGE under reducing conditions, transferred onto nitrocellulose membrane, and visualized by a streptavidin-chemiluminescence system. (B) Labeled cell lysates were cleared with MoAb Gi5 against GPIIb/IIIa and precipitated subsequently with anti-HPA-1a (Lane 1) or maternal serum (Lane 2) and analyzed as described above. were found within three generations; all of them were HPA-5aa homozygous, indicating that the Swia alloantigen is inherited with the HPA-5a allele. The Swia PCR-RFLP typing results matched perfectly with the serologic analy- sis in the MAIPA assay. No Swia-positive individual was identified upon genotyping of 100 unrelated blood donors. Furthermore, PCR and nucleotide sequencing analysis of five Swia-positive individuals (Fig. 4) were performed Volume 51, August 2011 TRANSFUSION 1749 KROLL ET AL. (A) (B) cDNA DNA Exon 29 Exon 28 Exon 28 Intron 5‘ 3‘ t C3347T a c g a g t g G C T A T T G T C A C C3347T Val1086 Thr1087 Val1086 Thr1085 Met Ile1088 Thr1087 Met Fig. 3. Nucleotide sequencing analysis of amplified GPIa cDNA and genomic DNA. (A) Eight overlapping fragments of GPIa cDNA derived from paternal PLTs were amplified by PCR. The analysis of PCR product encompassing Nucleotides 3098 through 3772 was presented. The C>T base exchange of the wild-type to the mutant GPIa at Position 3347 (arrow) results in Thr1087 (ACG)>Met1087 (ATG). (B) Genomic DNA from a Swia-positive individual was amplified by PCR using an intronic primer pair flanking Exon 28. The exon-intron boundary is presented and the point mutation C>T at Position 3347 is indicated by arrow. aa aa to identify polymorphism at Position 759. One Swiapositive individual (grandmother) was identified as TT759 homozygous, indicating that the T3347 allele arises from the less frequent T759 allele (Fig. 5). Previous studies demonstrated that this allele is associated with higher-density expression of GPIa/IIa complex on PLT surface. Indeed, Swia-positive individuals expressed GPIa/IIa in high amounts (3315 molecules/PLT, data not shown). Anti-Swia aa aa aa bb ab ab ab 362 bp 212 bp 50 bp 150 43 bp Msl I 5‘ 43 bp Msl I 150 bp 212 bp 3‘ Swia + 5‘ 43 bp F: Ex28-108 362 bp 405 bp 3‘ Swia R: Ex28+207 Fig. 4. Pedigree of the index family Swi and PCR-RFLP genotyping of the Swia alloantigen. (䊉, ) Swia-positive individuals; (䊊, 䊐) Swia-negative individuals. The HPA-1 genotypes (aa, ab, bb) are given below the symbols. The arrow indicates the mother who was immunized against Swia and HPA-1a. Positions of the restriction sites for MslI endonuclease (arrows) and the length of restriction fragments for Swia (+) and (-) are shown. Genomic DNA was amplified using intronic forward primer (F = Exon 28-108 nucleotides) and reverse primer (R = Exon 28 + 207 nucleotides) as indicated. PCR products were digested with MslI endonuclease and were analyzed on 1.6% agarose gel electrophoresis stained with ethidium bromide. The lanes refer to the individuals in the pedigree above. Molecular weight marker was run in parallel (left). 1750 TRANSFUSION Volume 51, August 2011 Analysis of recombinant GPIa allelic isoforms To confirm the responsibility of Amino Acid 1087 for the formation of the Swia epitope, allele-specific constructs encoding for Glu505Thr1087 (wild-type) and Glu505Met1087 (mutant) were transfected into CHO cells. Stable transfected cells expressing allelic forms of GPIa/IIa were surface labeled and analyzed by immunoprecipitation (Fig. 6). In the control experiment, anti-HPA-5a precipitated GPIa/IIa complex from wild-type as well as from mutant isoforms. In contrast, anti-Swia only recognized the Met1087 GPIa isoform. Similar results were obtained when allele-specific CHO cells were analyzed with antiSwia in the MAIPA assay (data not shown). Effect of the Thr1087Met polymorphism and anti-Swia alloantibody on cell function To determine possible effects of the Thr1087Met substitution on PLT function, aggregation studies with Swiaphenotyped individuals were performed. The PLT aggregation response to high (50 mg/mL) as well as to low collagen concentrations (1 mg/mL) of Swia-positive PLTs was not different from Swia-negative PLTs. In addition, no HPA-5a (Allele 1; 0.376) T 759 1600 2235 2483 3347 PLT ALLOANTIGEN Swia C A G C C C G G C C G C C G C G HPA-5b (Allele 3; 0.076) HPA-5a (Allele 2; 0 0.529) 529) C T G HPA-5a, -13bw (Allele 4; <0.01) HPA-5a, HPA-18bw (Allele 5; <0.01) HPA-5a, Swia (Allele 6; <0.01) T G T T C C G G C T low expression high expression Fig. 5. GPIa alleles associated with alloantigenic determinants and their frequencies in the Caucasian population. Arrows indicate the point mutations that have occurred during evolution. Gene frequency within the Caucasian population is in parentheses. Nucleotide numbering according to the Nomenclature System for Human Gene Mutations. The first Nucleotide A of the start codon is designated as nucleotide +1.19 (A) (B) WT Mut Cell Numbers C s Mut + Ctrl Mut + Ctrl Mut + Gi14 GPIa GPIIa WT + Gi14 Fluorescence intensity 1 2 3 1 2 3 Fig. 6. Analysis of CHO cells transfected with GPIa allelic constructs. (A) Flow cytometry analysis of stable transfected cells expressing wild-type (WT) and mutant (Mut) GPIa isoforms with isotype control (Ctrl) and with MoAb Gi14 against GPIa/IIa complex. (B) Immunoprecipitation analysis of stable transfected cells expressing wild-type (WT) and mutant (Mut) GPIa isoforms with AB serum (Lanes 1), anti-HPA-5a (Lanes 2), and anti-Swia (Lanes 3). Immunoprecipitates were run on 7.5% SDS-PAGE under reducing conditions, transferred onto nitrocellulose membrane, and visualized using a streptavidin-chemiluminescence system. inhibition of PLT aggregation by anti-Swia was observed (data not shown). These results could be confirmed by analysis of adhesion of allele-specific transfected CHO cells on immobilized collagen (Fig. 7). Similar adhesion capacity was detected with wild-type and mutant transfectants, and this cell adhesion could not be inhibited by anti-Swia. In the control experiment, MoAb Gi9 specific for a functional epitope on the GPIa/IIa complex significantly blocked the adhesion of both cell lines onto collagen. DISCUSSION In this study, we report on the characterization of a new rare PLT alloantigen residing on PLT GPIa (integrin a2 subunit), which we termed Swia according to the name of the index family. Together with an anti-HPA-1a, anti-Swia was found in serum of a mother, who lost her baby due to massive intracranial hemorrhage. Serologic analysis of the maternal serum with paternal PLTs and a panel of HPA-typed PLTs in the MAIPA Volume 51, August 2011 TRANSFUSION 1751 KROLL ET AL. Cell A Adhesion ((Collagen/G Gi14) 1.2 1 0.8 0.6 0.4 0.2 0 BSA Col 5 Col 2.5 - - - Col 2.5 AB Col 2.5 Swi Col 2.5 Gi9 Fig. 7. Adhesion of CHO cells expressing wild-type ( ) and mutant ( ) GPIa isoforms on collagen matrix. Fluorescencelabeled cells before and after treatment with antibodies were allowed to adhere to microtiter wells coated either with BSA, MoAb Gi14, or collagen Type I. After being washed bound cells were measured on fluorescence microtiter reader. Binding was calculated as ratio between the cells bound to collagen and to MoAb Gi14 (n = 3). assay documented the presence of anti-HPA-1a. In addition, strong reactivity was observed with paternal PLTs when GPIa/IIa was captured in the MAIPA assay. These results were confirmed by immunoprecipitation analysis of paternal PLTs. Nucleotide sequencing analysis of GPIa transcripts from paternal PLTs showed a single-nucleotide substitution C>T at Position 3347 in a heterozygous state, which resulted in a Thr>Met mutation at Position 1087 (numbering according to the international nomenclature system).19 This mutation is encoded by the last triplet (ACG or ATG) of Exon 28 and creates a restriction site for MslI endonuclease. PCR-RFLP analysis demonstrated that the C3347T substitution correlated with the serologic phenotype of all Swia-positive and Swia-negative family members as well as Swia-negative unrelated blood donors. Immunoprecipitation analysis of heterologous transfected mammalian cells confirmed that the single-aminoacid substitution Thr1087Met (polar uncharged into hydrophobic nonpolar amino acid residue) is sufficient to induce the formation of the epitope(s) recognized by anti-Swia. The Thr1087Met dimorphism is located in the C-terminal region, on the calf-2 domain of a2 integrin, only 15 amino acids away from the transmembrane region. The integrin a2b1 interacts with collagen via the I-domain, which is located on the N-terminal region of this molecule.20 As expected from the localization of the mutation within the molecule, the Thr1087Met mutation did not show any effect on the adhesion of transfected cells expressing Swia antigen onto immobilized collagen or on collagen-induced PLT aggregation, and anti-Swia did 1752 TRANSFUSION Volume 51, August 2011 not interfere with cell adhesion or aggregation. These findings are comparable with the capacities of the other HPAs located on GPIa to influence the adhesion to collagen.9,10 Meanwhile, five allelic forms of GPIa carrying four HPAs have been described (see Fig. 5). This report adds another low-frequency variant of GPIa. Our previous studies showed that the most frequent HPA-5a is encoded by Alleles 1 (gene frequency, 0.394) and 2 (gene frequency, 0.529), whereas the less frequent HPA-5b is encoded by Allele 3 (gene frequency, 0.076).9 The rare alloantigen, HPA-13bw (Sita), arose most probably from the most frequent Allele 2. Interestingly, both new antigens HPA-18bw (Caba) and Swia appear to originate from the less frequent Allele 1.10 Alleles 1 and 2 differ by a nonsense mutation at Position 759 (T or C). Kunicki and coworkers11,21 have demonstrated that carriers of the T759 allele express a higher density of GPIa/IIa molecules on the PLT surface in comparison to individuals carrying the C759 allele. Accordingly, we found in PLTs of Swia-positive individuals, who were heterozygous for CT759, higher GPIa/IIa surface density in comparison to CC759 individuals. However, it has not been determined whether having a higher expression of GPIa/IIa predisposes an individual to a greater risk of alloimmunization against alloantigenic epitopes on the GP. So far antibodies against HPA-5b and HPA-13bw, which are expressed on lowdensity alleles (C759) have been observed more frequently in FNAIT than antibodies against antigens on highdensity alleles (HPA-5a or HPA-18bw).4,22 Most of the immunogenic polymorphisms relevant in FNAIT are located on GPIIb/IIIa. This study together with the recent description of HPA-18bw indicates that GPIa is the second most relevant target for alloantibodies on the PLT surface.10 Overall, the clinical consequences of immunizations against epitopes on GPIa/IIa seem to be less pronounced than in antibodies against GPIIb/IIIa, indicated by a lower frequency of intracranial bleeding and an often milder thrombocytopenia.9,10,23 In our case, the isolated clinical relevance of anti-Swia could not be estimated since maternal serum additionally contained a strong anti-HPA-1a and the affected children (two and three) did not carry the Swia antigen. Because the only Swia-positive child in the index family was unaffected, there is no clinical evidence that anti-Swia can cause thrombocytopenia. To estimate the potential of anti-Swia to eliminate PLTs from the circulation in vivo mouse models could reveal further information. In a recent large cohort study on antibodies in FNAIT, low-frequency HPAs were estimated as not significant.24 However, intensive analysis of maternal sera with paternal PLTs in experienced laboratories revealed a constantly growing number of low-frequency PLT antigens as well as first examples of repeated immunizations against already PLT ALLOANTIGEN Swia known HPAs.10,22,25,26 In addition, our report demonstrates that not only combinations of antibodies against the more frequently involved HPAs occur, for example, HPA-1a and HPA-5b,4 but also combinations of anti-HPA-1a with antibodies against a low-frequency HPA must be expected. One may speculate that careful screening of maternal sera with a suitable panel of MoAbs will lead to the identification of more of such cases. platelet alloantigen Sita and affects collagen-induced aggregation. Blood 1999;94:4103-11. 10. Bertrand G, Jallu V, Saillant D, Kervran D, Martageix C, Kaplan C. The new platelet alloantigen Caba: a single point mutation Gln716His on the a2 integrin. Transfusion 2009;49: 2076-83. 11. Kritzik M, Savage B, Nugent D, Santoso S, Ruggeri ZM, Kunicki TJ. Nucleotide polymorphisms in the a2 gene define multiple alleles that are associated with differences in platelet a2b1 density. Blood 1998;92:2382-8. ACKNOWLEDGMENTS This study is part of the doctoral thesis of KF. The authors thank 12. Koh Y, Taniue A, Ishii H, Matsuyama N, Amakishi E, Hayashi T, Furata RA, Fukumori Y, Hirayama F, Yoshimura Olga Eva and Silke Werth for their technical assistance. Our grati- K, Nagamine T, Tamai S, Nakano S. Neonatal alloimmune tude is extended to the family Swi for their cooperation in this thrombocytopenia caused by an antibody specific for a study. newly identified allele of human platelet antigen-7. Transfusion 2010;50:1276-84. CONFLICT OF INTEREST The authors declare that they have no conflicts of interest relevant to the manuscript submitted to TRANSFUSION. REFERENCES 1. Kaplan C. Foetal and neonatal alloimmune thrombocytopenia. Orphanet J Rare Dis 2006;I:39. 2. Mueller-Eckhardt C, Kiefel V, Grubert A, Kroll H, Weisheit M, Schmidt S, Mueller-Eckhardt G, Santoso S. 348 cases of suspected neonatal alloimmune thrombocytopenia. Lancet 1989;I:363-6. 3. Davoren A, Curtis BR, Aster RH, McFarland JG. Human platelet antigen-specific alloantibodies implicated in 1162 cases of neonatal alloimmune thrombocytopenia. Transfusion 2004;44:1220-5. 4. Ertel K, Al-Tawil M, Santoso S, Kroll H. Relevance of the HPA-15 (Gov) polymorphism on CD109 in alloimmune thrombocytopenic syndromes. Transfusion 2005;45:366-73. 5. Kroll H, Kiefel V, Santoso S, Mueller-Eckhardt C. Sra, a private platelet antigen on glycoprotein IIIa associated with neonatal alloimmune thrombocytopenia. Blood 1990; 76:2296-302. 6. Zutter MM, Edelson BT. The a2b1 integrin: a novel collectin/C1q receptor. Immunobiology 2007;212:343-53. 13. Kuijpers RW, Simsek S, Faber NM, Goldschmeding R, van Wermerkerken RK, von dem Borne AE. Single point mutation in human glycoprotein IIIa is associated with a new platelet-specific alloantigen (Mo) involved in neonatal alloimmune thrombocytopenia. Blood 1993;81:70-6. 14. Santoso S, Kroll H, Andrei-Selmer CL, Socher I, Rankin A, Kretzschmar E, Watkins NA, Ouwehand WH. A naturally occuring Leu33Val mutation in b3-integrin impairs the HPA-1a epitope: the third allele of HPA-1. Transfusion 2006;46:790-9. 15. Peterson JA, Gitter ML, Pietz B, Bellissimo DB, Curtis BR, Aster RH. A third allele of the HPA-5 (Br) platelet alloantigen system identified in investigating a case of neonatal thrombocytopenia. Transfusion 2010;50:1855-6. 16. Kiefel V. The MAIPA assay and its applications in immunohaematology. Transfus Med 1992;2:181-8. 17. Santoso S, Kalb R, Kroll H, Walka M, Kiefel V, MuellerEckhardt C, Newman PJ. A point mutation leads to an unpaired cysteine residue and a molecular weight polymorphism of a functional platelet beta3 integrin subunit. J Biol Chem 1994;269:8439-44. 18. Carl B, Kroll H, Bux J, Bein G, Santoso S. B-lymphoblastoid cell lines as source of reference DNA for genotyping of human platelet and neutrophil antigens. Transfusion 2000; 40:62-8. 19. Antonarakis SE. Recommendations for a nomenclature system for human gene mutations. Nomenclature Working 7. Santoso S, Kalb R, Kiefel V, Mueller-Eckhardt C, Newman PJ. The human platelet alloantigens Bra and Brb are associated with a single amino acid polymorphism on glycopro- Group. Hum Mutat 1998;11:1-3. 20. Emsley J, King SL, Bergelson JM, Liddington RC. Crystal structure of the I domain from integrin a2b1. J Biol Chem 1997;272:28512-7. tein Ia (integrin a2 subunit). J Clin Invest 1993;92:2427-32. 8. Santoso S, Kunicki TJ, Kroll H, Haberbosch W, Gardemann A. Association of the platelet glycoprotein Ia C807T gene polymorphism with myocardial infarction in younger 21. Kunicki TJ, Kritzik M, Annis DS, Nugent DJ. Hereditary variation in platelet integrin a2b1 density is associated with two silent polymorphisms in the a2 gene coding sequence. Blood 1997;89:1939-43. patients. Blood 1999;93:2449-53. 9. Santoso S, Amrhein J, Hofmann H, Sachs U, Walka M, Kroll H, Kiefel V. A mutational hot spot Thr799Met on the a2 integrin subunit leads to the formation of the new human 22. Kroll H, Yates J, Santoso S. Immunization against a lowfrequency human platelet alloantigen in fetal alloimmune thrombocytopenia is not a single event: characterization by the combined use of reference DNA and novel alleleVolume 51, August 2011 TRANSFUSION 1753 KROLL ET AL. specific cell lines expressing recombinant antigens. Transfusion 2005;45:353-8. 23. Kaplan C, Morel-Kopp MC, Kroll H, Kiefel V, Schlegel N, Chesnel N, Mueller-Eckhardt C. HPA-5b (Bra) neonatal alloimmune thrombocytopenia: clinical and immunological analysis of 39 cases. Br J Haematol 1991;78:425-9. 24. Ghevaert C, Rankin A, Huiskes E, Porcelijn L, Javela K, Kekomäki R, Bakchoul T, Santoso S, Nutland S, Smyth DJ, Smith GA, McBride S, Watkins NA, Ouwehand WH. Alloantibodies against low-frequency human platelet antigens do not account for a significant proportion of cases of fetoma- 1754 TRANSFUSION Volume 51, August 2011 ternal alloimmune thrombocytopenia: evidence from 1054 cases. Transfusion 2009;49:2084-9. 25. Peterson JA, Balthazr SM, Curtis BR, McFarland JG, Aster RH. Maternal alloimmunization against the rare plateletspecific antigen HPA-9b (Maxa) is an important cause of neonatal alloimmune thrombocytopenia. Transfusion 2005;45:1487-95. 26. Peterson JA, Gitter ML, Kanack A, Curtis BR, McFarland JG, Bougie D, Aster RH. New low-frequency platelet glycoprotein polymorphisms associated with neonatal alloimmune thrombocytopenia. Transfusion 2010;50:324-33.