If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Immunohematology Diagnostics, Sanquin, Amsterdam, the NetherlandsDepartment of Immunohematology and Blood Transfusion, Leiden University Medical Center, Leiden, the NetherlandsCenter for Clinical Transfusion Research, Sanquin Research, Leiden, the Netherlands
Human platelet antibody (HPA) detection is necessary for the diagnosis and therapeutic decisions for refractoriness to platelet transfusions, post transfusion purpura and fetal and neonatal alloimmune thrombocytopenia. In the last four to five decades many new developments, both in knowledge and methods, have increased the quality of platelet serology. However, the quest for the optimal antibody detection method(s) encountered, sometimes unexpected, difficulties. In this review the various aspects concerning platelet antibody test methods and detection of platelet antibodies both for the diagnostic and screening setting are discussed.
1.1 Phase I and II platelet antibody detection assays
Early methods (phase I) for platelet antibody detection such as platelet aggregation tests, serotonin release and complement fixation techniques made use of the functional properties of blood platelets for indirect evidence for the presence of antibodies [
]. In the 1970s the availability of radiolabelled or immune fluorescence labelled anti-human immunoglobulins of the IgG, IgM or IgA class led to the development of more sensitive and specific methods (phase II) e.g. radio immune assay (RIA), platelet immunofluorescence test (PIFT) and mixed passive haemagglutination assay (MPHA) [
]. With these assays, detection of antibody binding on donor platelets incubated with serum from the patient (indirect method) or directly on patient platelets (direct method) became possible. However, major drawbacks of these whole platelet assays are the co-expression of human leucocyte antigen (HLA)-class-I antigens and Fc-gamma-receptor (FcγR)IIa on platelets causing difficulties to distinguish HLA antibodies from human platelet antigen (HPA) antibodies and non-specific binding of IgG via the Fc-part to the Fc-receptor. In order to circumvent the HLA problem incubation of platelets in chloroquine was used. Chloroquine removes the HLA antigens from the platelet surface and prevents HLA antibodies from binding while the human platelet antigens carrying glycoproteins remain intact [
The first assays aimed at detecting antibodies using isolated glycoproteins (GPs), thus bypassing the HLA and Fc-receptor ‘problem’, applied sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) [
Platelet plasma membrane glycoproteins. Evidence for the presence of nonequivalent disulfide bonds using nonreduced-reduced two-dimensional gel electrophoresis.
In the second half of the 1970′s solubilization of the platelet membrane and extraction of the membrane proteins, retaining their antigenicity, with non-ionic detergents was described [
Characterization of human platelet proteins solubilized with Triton X-lOO and examined by crossed immunoelectrophoresis. Reference patterns of extracts from whole platelets and isolated membranes.
]. This, together with the availability of GP-specific monoclonal antibodies (moab) made the development of (phase III) glycoprotein-specific platelet antibody detection methods possible.
The first moab-based GP-specific immunoassay for the detection of platelet antibodies against GPIIb/IIIa and GPIb/IX was described by Woods and McMillan [
]. In this assay platelets were solubilized with Triton-X100 and GP’s in the lysate were captured on GP-specific mouse-anti-human moab coated microtiterplate wells. After incubation with patient plasma radiolabeled anti-human-IgG was added to detect free circulating platelet autoantibodies. Also polystyrene beads coated with anti-human-IgG were incubated with lysate from antibody-sensitized platelets (i.e. platelets from ITP patients directly or from healthy donors after incubation with serum), after which radiolabeled GP-specific mouse-anti-human moabs were added to detect GP-associated autoantibodies [
]. Shortly thereafter the antigen-capture ELISA (ACE) and monoclonal antibody immobilization of platelet antigens (MAIPA, see Fig. 1) assay were introduced by Furihata et al. [
] In the ACE, isolated glycoproteins from HPA typed donorplatelets were captured on a solid support (polystyrene bead or microtiter plate well), pre-coated with mouse-anti-human GP-specific moab’s and incubated with the patient’s serum. In MAIPA assays, HPA typed donor platelets are incubated with mouse-anti-human monoclonal antibodies directed against specific platelet glycoproteins and the patient’s serum. The moab-GP-antibody complexes are then isolated from the platelets by solubilization and subsequently captured on microtiterplate wells coated with goat-anti-mouse moab’s. In both assays labelled anti-human-Ig conjugate was used to visualise antibodies binding to the GP. The MAIPA assay bypassed the problems with non-specific reactions by antibody binding to the plastic of the microtiterplate and with false negative results due to antigenicity loss as a result of conformational changes of the GP’s following platelet lysis. Although this method proved to be a major step forward in platelet serology, some problems with non-specific reactions due to unexpected naturally occurring anti-mouse antibodies in human sera were seen [
] introduced the modified antigen-capture ELISA (MACE, see Fig. 1) to prevent this problem and Kiefel, in his overview of the applications in immunohematology suggested to simply overcome this problem by first incubate donor platelets with human plasma, followed by a washing step and only then incubate with the mouse monoclonal antibodies [
]. Both assays proved to be more sensitive and specific for the detection of platelet alloantibodies (see ‘2. HPA antibody detection’ for more details), both for serum tested with a panel of HPA typed donor platelets and for cross-matching with paternal platelets which is used for diagnosing fetal and neonatal alloimmune thrombocytopenia (FNAIT), and were implemented in most platelet serology laboratories in the world [
Sensitivity of the platelet immunofluorescence test (PIFT) and the MAIPA assay for the detection of platelet-reactive alloantibodies: a report on two U.K. National Platelet Workshop exercises.
A panel of platelets carrying different HPA typed glycoproteins are incubated with human (maternal or patient) serum. If the serum contains anti-HPA-1a antibodies, they will bind to HPA-1a positive platelets. After adding a glycoprotein-specific mouse-anti-human monoclonal antibody (moab, in this figure anti-GPIIb/IIIa), the platelets are solubilized. The human antibody-GP-mouse moab complex is then captured in a microtiterplate well (or on a bead) using goat-anti-mouse moab’s. The presence of the human antibody (anti-HPA-1a in this figure) is visualized using a color reaction with horseradish peroxidase.
The same procedure is used for MACE, but solubilization is performed after incubation of platelets with the human serum and the human antibody-GP complex is captured with mouse-anti-human moab’s coated on the microtiterplate or bead.
In the 1990s, beads coated with sheep or goat-anti-mouse moab serving as solid phase target for the specific mouse-anti-human-platelet glycoprotein complexes were used in modified MAIPA assays [
]. The main advantage of these beads assays was the short test time of approximately 3.5 h compared to eight hours for the MAIPA or later 5 h for the modified rapid MAIPA [
A modified rapid monoclonal antibody-specific immobilization of platelet antigen assay for the detection of human platelet antigen (HPA) antibodies: a multicentre evaluation.
1.3 Methods for the simultaneous detection of multiple HPA antibodies
The detection of antibodies on different GPs simultaneously was first introduced by Joutsi et al. 1997 using different fluorescent labels for the GP-specific moab’s in a MAIPA setting and later by using polystyrene beads with varying fluorescence intensity [
Comparison of the direct platelet immunofluorescence test (direct PIFT) with a modified direct monoclonal antibody-specific immobilization of platelet antigens (direct MAIPA) in detection of platelet-associated IgG.
]. A ‘beads MAIPA’ called SASPA, using rat-anti-mouse moab, with different IgG subtype specificities to prevent cross-capture of the GP-specific moab’s, coated beads was introduced by Nguyen et al. in 2004 [
]. Furthermore, an HPA antibody detection method using gel ID cards (gel antigen specific assay, GASA), containing anti-human-IgG and microbeads on which initially GPs were captured by using GP-specific moab’s but later by using biotin-streptavidin coupling has also been described (Meyer (2006) and Bakchoul (2008)) [
However, because the (modified) MAIPA assay is time-consuming and technically demanding, many (non-reference) laboratories use more convenient solid-phase ELISA techniques with GP’s from HPA typed donors already coated on microtiterplates or Luminex beads (Fig. 2), which became commercially available in the decades after the introduction of the MAIPA [
Luminex beads with different colors carrying different HPA typed GP structures, comparable with a platelet antibody identification panel, are used. In this figure the human serum containing anti-HPA-1a antibodies is added to a mixture of beads. The HPA-1a antibodies will bind to the HPA-1a positive beads. After incubation with a fluorescence labelled anti-human IgG, lasers measuring the bead color and the presence of fluorescence (mean fluorescence intensity) are used for the identification of the antibody.
Although not available in all platelet serology laboratories, the use of a surface plasmon resonance (SPR) method for the detection of HPA antibodies seems promising [
Low-avidity anti-HPA-1a alloantibodies are capable of antigen-positive platelet destruction in the NOD/SCID mouse model of alloimmune thrombocytopenia.
]. Bakchoul et al. (2011) successfully tested Isolated IgG fractions from maternal sera in flow cells with purified HPA-1 typed GPIIb/IIIa from platelet donors [
Low-avidity anti-HPA-1a alloantibodies are capable of antigen-positive platelet destruction in the NOD/SCID mouse model of alloimmune thrombocytopenia.
]. Sixty-eight anti-HPA-1a antibodies could be detected with MAIPA and 75 with SPR in maternal samples from 83 suspected FNAIT cases with HPA-1a negative mothers. Also Peterson et al. (2012) showed a higher sensitivity using SPR for 61 suspected FNAIT cases with HPA-1a negative mothers, but without detectable antibodies in MACE [
1.5 HPA-typed cell lines and glycoprotein contstructs
The above-mentioned test methods are all based on the use of HPA typed donor platelets. The availability of certain typed donors can be a problem as the frequency of some HPA’s is extremely low. To overcome this problem, Hayashi et al. established an HPA typed cell line panel as an alternative source of platelet antigens. For this they co-transduced HPA typed wild-type β3 and αIIb in a human erythroleukaemia K562 cell line. After isolation of the GPs expressed on these cells, they were successfully used in the MAIPA [
]. Furthermore, Skaik and colleagues reported the large-scale production of the whole extracellular part of recombinant soluble β3-integrins in HEK cells displaying HPA-1a or HPA-1b epitopes, coupled to a His-tag for use in a single-antigen magnetic bead assay (SAMBA) [
Development of a single-antigen magnetic bead assay (SAMBA) for the sensitive detection of HPA-1a alloantibodies using tag-engineered recombinant soluble β3 integrin.
In addition to the low availability of certain typed donors, conformational changes of GPs after isolation and incomplete solubilisation can cause incorrect results. To prevent these problems the use of recombinant glycoprotein constructs, as described by Stafford et al. in 2008 [
A multicenter validation of recombinant β3 integrin-coupled beads to detect human platelet antigen-1 alloantibodies in 498 cases of fetomaternal alloimmune thrombocytopenia.
] can be a solution. They first tested several different β3 constructs and showed that recombinant calmodulin-tagged β3 integrin fragments ΔβA-Leu33 (rHPA-1a) and ΔβA-Pro33 (rHPA-1b) could be used in an ELISA method and a Luminex bead-based assay for the detection of both type I and II anti-HPA-1 antibodies (for details regarding type I and type II antibodies see ‘2. HPA antibody detection’) [
A multicenter validation of recombinant β3 integrin-coupled beads to detect human platelet antigen-1 alloantibodies in 498 cases of fetomaternal alloimmune thrombocytopenia.
Until now 41 human platelet antigens are known, of which 12 are part of bi-allelic systems, with varying allele frequencies in different populations (for references and further information see http://www.versiti.org/HPA).
Several aspects need to be considered for platelet alloantibody detection. To value the antibody detection results it is necessary to consider the clinical symptoms, the ethnic origin of the patient and the HPA pheno- or genotyping results. The intact platelet antibody detection assays are generally less sensitive and specific than glycoprotein-specific assays. This is partly due to the low and/or varying expression of certain GPs (e.g. GPIa/IIa carrying HPA-5 and CD109 carrying HPA-15) [
]. On the other hand, disadvantages of GP-specific assays are: 1) GP isolation prior to incubation with the patient’s serum can result in antigen loss and non-specific binding of antibodies due to damage or conformational change [
Studies on the binding of immunoglobulins and immune complexes to the surface of human platelets: IgG molecules react with platelet Fc receptors with the CH3 domain.
Involvement of the cysteine-rich domain of glycoprotein IIIa in the expression of the human platelet alloantigen, PlA1: evidence for heterogeneity in the humoral response.
] 2) non-specific binding of antibodies to the microtiterplate wells or beads used in the GP-specific assays can cause false positive results 3) in general glycoprotein-specific assays only test for antibodies of the IgG class 4) the mouse-anti-human moab can block human antibody binding [
] 5) of course, only antibodies directed against antigens located on the investigated glycoproteins will be detected in glycoprotein-specific assays. Most commercial assays test for the a and b antigens within the HPA-1, -3 and -4 systems (GPIIb/IIIa), as well as the HPA-2 (GPIb/IX) and HPA-5 (GPIa/IIa) systems. The commercial Luminex beads assay PAKLx also includes a GPIV specific bead. However, none of them includes CD109 (HPA-15), which might be due to technical difficulties in coating this glycosylphosphatidylinositol (GPI)-linked cell surface antigen [
], we provided some clarity about the value of various techniques for the detection of platelet antibodies. Using only the PIFT, 53 of 73 (73 %) HPA antibodies were detected. Laboratories using both the PIFT and a glycoprotein-specific solid phase ELISA (SPhE) or the PIFT and the MAIPA detected 35 of 35 (100 %) platelet reactive antibodies. The specificities of SPhE and MAIPA were respectively, 88 % and 97 %. False positive reactions in the SPhE, often with inert blood group AB serum samples from never transfused, male donors, were not HPA-specific but occurred with one or more than one GPs.
Furthermore, the results of the Dutch EQA for platelet serology showed considerable inter-laboratory variation. The percentages of correct results ranged from 57 to 86 percent for the detection of platelet-specific antibodies in laboratories in which only the PIFT was used and from 70 to 100 percent correct results in laboratories where the MAIPA was applied. This inter-laboratory variation is, although in lesser extent, also seen in the results of the quality assurance exercises of the ISBT International Platelet Immunology Working Party and emphasizes the need for standardization and well-trained and experienced technicians even for performing very standardized ELISA techniques [
Most platelet serology reference laboratories nowadays use one or several glycoprotein-specific antibody detection methods. The majority (> 95 %) use MAIPA, whether or not in combination with commercial or in-house modified ELISA assays or Luminex based beads assays [
], showed four (6 %) false-negative results in PAKLx in a total of 67 HPA or GP-reactive antibodies containing sera and eight (4.1 %) false-positive GP reactions in a total of 194 sera compared with the MAIPA assay. The sensitivity for the detection of anti-HPA-1a antibodies was comparable with MAIPA. Comparable results were seen in a PAKLx validation study by Cooper et al. [
], but testing more HPA-3 specific antibody samples, they noticed a limited sensitivity (80 %) for detecting anti-HPA-3 antibodies.
2.2 HPA-1a antibody detection
The HPA-1 system consists of the antigens HPA-1a and -1b, encoded by a single-nucleotide polymorphism (SNP rs5918) Leu33Pro in the ITGB3 gene (among Caucasians these antigens are expressed by 97.6 % and 28 %, allele frequencies 0.846 and 0.154, respectively). These antigens are located on the β3 integrin, where β3 may be combined with α2b (α2bβ3, GPIIb/IIIa, CD41/CD61) on platelets or with αv (αvβ3, CD51/CD61) on different cell types such as endothelial cells, osteoclasts and trophoblasts [
]. Immunization against HPA-1a is strongly correlated with the presence of the HLA-DRB3*01:01 allele, for which approximately 28 % of the population is positive [
HPA-1a antibodies can cause refractoriness for platelet transfusions (PTR), FNAIT and post-transfusion purpura (PTP). In the Caucasian population, approximately 1:450 random women are immunized against HPA-1a during pregnancy, which leads to neonatal thrombocytopenia in about 1: 1000 pregnancies, with a high risk of intracranial bleeding which occurs in approximately one in 10,000 pregnancies [
]. The risk for an HPA-1a negative, HLA-DRB3*01:01 positive woman, to become immunized after giving birth to an HPA-1a positive child has been calculated at 12.7 %, (95 % confidence interval: 8.6 %–16.8 %) [
In the exceptional cases (approximately 0.5 %) that HPA-1a negative/HLA-DRB3*01:01 negative women get immunized against HPA-1a, the antibodies are often weak reactive or non-detectable [
It is shown that anti-HPA-1a can differ in the way they interact with β3. They can either (type I) bind to a very restricted epitope at or near the Leu33/Pro33 polymorphic amino acid residue located on the β3 plexin/semaphorin/integrin (PSI) domain, or (type II) they do also require residues in the hybrid and epidermal growth factor (EGF) domains [
Involvement of the cysteine-rich domain of glycoprotein IIIa in the expression of the human platelet alloantigen, PlA1: evidence for heterogeneity in the humoral response.
]. These differences might be one of the mechanisms explaining the sometimes very weak reactive or undetectable anti-HPA-1a antibodies in FNAIT cases with severe neonatal thrombocytopenia [
Low-avidity anti-HPA-1a alloantibodies are capable of antigen-positive platelet destruction in the NOD/SCID mouse model of alloimmune thrombocytopenia.
]. It is possible that conformational changes resulting from manipulating the GPs during the test procedure can prevent type II antibodies from binding.
2.3 Anti-HPA-1a antibody detection in a diagnostic setting versus a screening setting
It must be emphasized that post-natal anti-HPA-1a detection in cases of suspected FNAIT because of neonatal thrombocytopenia requires a different antibody detection approach than anti-HPA-1a screening in pregnancy. Neonatal thrombocytopenia combined with maternal-fetal HPA-1a incompatibility but without detectable antibodies is a reason for more extensive diagnostic serologic tests (see below). For large scale screening this is not feasible. For HPA-1a negative mothers, an antibody detection assay with maximal sensitivity and specificity at the right time point(s) during pregnancy is needed to identify the cases at risk and to start treatment if necessary. Therefore, some remarks now follow about the sensitivity and specificity of HPA-1a antibody detection both in a diagnostic setting and in screening programs.
In a total of 1683 post-natal FNAIT serology requests received in our laboratory (unpublished data), HPA antibody detection showed a total of 326 HPA antibodies: anti-HPA-1a (n = 235), anti-HPA-1a + 3a (n = 2), anti-HPA-1a + 5b (n = 9), anti-HPA-1b (n = 5), anti-HPA-3a (n = 5), anti-HPA-5a (n = 7), anti-HPA-5b (n = 53), anti-HPA-6bw (n = 1), anti-HPA-11bw (n = 1), anti-HPA-15a (n = 3), anti-HPA-15b (n = 2). 271 mothers (16 %) were typed HPA-1a negative, leaving 25 (i.e. 271–246) cases with HPA-1a positive thrombocytopenic neonates but no detectable antibodies. This number is not increased compared to the random population, indicating that the HPA-1a incompatibility combined with the neonatal thrombocytopenia might be merely coincidence. We test 120 μL maternal serum in the routine (rapid) MAIPA assay but in cases with HPA-1a discrepancies without detectable HPA-1a antibodies we extent the investigation with double serum incubation with typed donor platelets, use a different mouse-anti-human GPIIb/IIIa (Y2.51) monoclonal antibody in MAIPA, use a Luminex beads based assay and/or if still no antibodies are detected we ask for a new sample drawn 1–2 weeks post-partum. If still negative we do advice the treating physician to request HPA-1a antibody testing in the next pregnancy. In about 4–5 % of cases, where the mother is HPA-1a negative, we detect HPA-1a antibodies only after more extensive testing. This percentage was significantly higher when we tested 120 μL instead of 40 μL in the MAIPA. Our experience and the results of the various international workshops have made it clear that to obtain good results with the MAIPA, many different details must be taken into account, including the amount of serum being tested, the incubation times, the correct GP-specific moab’s, etc. In addition, the technicians experience is of great importance.
In the pre-natal screening study by Williamson et al. (1998) [
], including 24,417 women, 387 of 618 HPA-1a negative women were examined for antibodies against HPA-1a in MAIPA. Antibodies were detected in 46 women (i.e. 1:332 random women and 12 % of HPA-1a negative women) of which seven (15 %) only postnatally. Thrombocytopenia in cord blood samples was found in 12 of 237 HPA-1a negative women with no detectable HPA-1a antibodies in the initial MAIPA nor after supplementary testing. Remarkable was that all 12 women were HLA-DRB3*01:01 negative, which as indicated above can be a reason for very weak reactive or undetectable HPA-1a antibodies. In the Norwegian screening study [
] including 100,448 women, 1990 of 2111 HPA-1a negative women were examined for antibodies against HPA-1a in MAIPA. Antibodies were detected in 210 women (i.e. 1:450 random women and 10.6 % of HPA-1a negative women) of which 39 (19 %) only postnatally. The majority of these women were HLA-DRB3*01:01 positive [
]. Neonatal platelet counts for these 39 cases were not obtained and therefore it is not clear whether these cases were all post-natal immunizations or whether the antibodies were too weak reactive during pregnancy and only became detectable after the antibody titre was boosted due to fetal-maternal platelets transfer at birth.
] extended the standard FNAIT diagnostic serological investigation with MACE, using SPR, for 61 of 677 (9 %) cases with HPA-1a negative mothers and HPA-1a positive fathers without detectable antibodies. In 18 of these 61 cases HPA-1a antibodies were detected with SPR. For 13 of these infants clinical information could be obtained. Six children suffered from thrombocytopenia (mean 32 × 109/L, range 8−61 × 109/L) of whom five had petechial haemorrhages in the skin or mucosal membranes of suspected FNAIT. In our standard work-up for suspected FNAIT, we encounter difficulties in detecting 4–5 % of the HPA-1a antibodies with MAIPA (unpublished data).
We are currently running a 27th week pregnancy HPA-1a antibody screening program [
HIP-Study (HPA-screening In Pregnancy): Protocol of a Nationwide, Prospective and Observational Study to Assess Incidence and Natural History of Fetal/Neonatal Alloimmune Thrombocytopenia and Identifying Pregnancies at Risk.
], after which we perform an HPA antibody screening for the HPA-1a negative women. For this antibody screening we use a Luminex beads HPA antibody detection assay (PAKLx) [
]. This assay is very easy to perform and only uses 10 μL serum, which is very welcome because of the limited available material. Although too early for data on the sensitivity and specificity of this assay in this screening protocol, we already notice some difficulties in assessing weakly reactive HPA-1a positive GPIIb/IIIa carrying beads. The PAKLx assay contains six GPIIb/IIIa carrying beads, three HPA-1a homozygous, two HPA-1b homozygous and one HPA-1ab heterozygous. As a result of slightly varying back ground mean fluorescence intensity values, the reactions with the six HPA-1a and -1b carrying beads can mimic a very weak HPA-1a antibody pattern. This means that for a high number of weak reactive samples further testing is necessary to distinguish HPA-1a specific reactions from insignificant back ground reactions. This will postpone results and increase costs and therefore is not acceptable in large scale screening programs.
These data show that, for screening purposes, it can be valuable to search for easily standardized methods for HPA-1a antibody detection with higher sensitivity and specificity. It is important to take into account the possibility of being able to detect the different types of weak antibodies, without compromising specificity. To investigate these possibilities it will be worthwhile to validate and explore the possibilities with the above described recombinant β3 constructs or recombinant soluble β3 integrins on beads or in SPR [
Low-avidity anti-HPA-1a alloantibodies are capable of antigen-positive platelet destruction in the NOD/SCID mouse model of alloimmune thrombocytopenia.
Development of a single-antigen magnetic bead assay (SAMBA) for the sensitive detection of HPA-1a alloantibodies using tag-engineered recombinant soluble β3 integrin.
A multicenter validation of recombinant β3 integrin-coupled beads to detect human platelet antigen-1 alloantibodies in 498 cases of fetomaternal alloimmune thrombocytopenia.
], there was a significant correlation (p = 0.004) between the occurrence of severe neonatal thrombocytopenia (platelet count < 50 × 109/L) and a third trimester anti-HPA-1a antibody titre ≥ 1:32 in MAIPA (PPV 75 %, NPV 88 %). Based on the HPA-1a antibody titre and a standard curve created with serial dilutions of a standard high concentration anti-HPA-1a, a MAIPA quantification method was introduced in the Norwegian anti-HPA-1a screening program [
HIP-Study (HPA-screening In Pregnancy): Protocol of a Nationwide, Prospective and Observational Study to Assess Incidence and Natural History of Fetal/Neonatal Alloimmune Thrombocytopenia and Identifying Pregnancies at Risk.
] showing a PPV of 54 % and a NPV of 95 % (p = 0.001) for the prediction of severe thrombocytopenia (<50 × 109/L) when using a cut-off level of 3 AU/ml either at 22 or at 34 weeks gestation [
International Collaboration for Transfusion Medicine Guidelines. Maternal HPA-1a antibody level and its role in predicting the severity of Fetal/Neonatal Alloimmune Thrombocytopenia: a systematic review.
]. Quantitative MAIPA results generated by different reference laboratories in an international workshop showed comparable results for most laboratories, especially when using a standardized protocol with the same GP-specific moab and HPA-1a reference standard [
SSC of the ISTH. Interlaboratory workshop on anti-HPA-1a alloantibody quantification with the mAb-specific immobilization of platelet antigen technique.
] showed that variations in the fucosylation at Asn297 in the Fc-tail of anti-HPA-1a-specific IgG1 from FNAIT patients affected the binding affinity to FcγRIIIa and FcγRIIIb. HPA-1a antibodies with a low amount of Fc fucose showed enhanced phagocytosis of platelets using FcγRIIIb(+) polymorphonuclear cells or FcγRIIIa(+) monocytes as effector cells and the degree of anti-HPA-1a fucosylation correlated positively with the neonatal platelet counts in FNAIT, and negatively to the clinical disease severity. Further studies are needed to see if the degree of fucose of the Fc part of anti-HPA-1a in combination with the MAIPA quantification results will improve prediction of the severity of the neonatal thrombocytopenia.
Building on the findings of van Gils et al. (2008) [
] showed that in addition to the platelet specific type I and II HPA-1a antibodies, there is also a type III antibody, directed against HPA-1a located on the αVβ3 complex. A strong correlation was demonstrated between the presence of these antibodies in the maternal circulation and the occurrence of fetal intracranial bleeding, indicating that not only thrombocytopenia but also damage to the endothelial vessel wall can be an important underlying mechanism for the occurrence of fetal/neonatal intracranial haemorrhage.
2.5 Non HPA-1a antibodies
In contrast to the HPA-1a antibodies, little is known about the sensitivity and specificity for the detection of other HPA antibodies and about the existence of low avidity antibodies against other HPA’s.
Antibodies against HPA-1b are only occasionally involved in FNAIT but can be detected in sera from patients suffering from PTR. Antibodies against HPA-2a and -2b, a single-nucleotide polymorphism located on GPIbα (rs6065 T145 M encoded by the GPIBA gene, allele frequencies in Caucasian population 0.92 and 0.08, respectively), in the GP complex Ibα/Ibβ/IX/V are also very rarely involved in FNAIT. However, HPA-2b antibodies are, next to anti-HPA-1b, regularly seen in PTR and are often a combination of antibodies of the IgG and IgM class. Results from the workshops of the 16th to 19th ISBT International Platelet Immunology Working Party (unpublished data, i.e. mainly MAIPA results) show a high sensitivity (± 93 %) for the detection of HPA-1b antibodies and a varying sensitivity (75–100 %) for the detection of HPA-2b antibodies. Immunization against HPA-2a is very uncommon (we only detected this antibody once in a 15-years period). In our laboratory, we use MAIPA with anti-human-IgG and if necessary –IgM antibodies) and a commercial Luminex beads anti-HPA detection method (anti-human-IgG) in combination with PIFT (with both anti-human-IgG and -IgM) for the detection of these antibodies.
Antibodies against HPA-3a and -3b, located on α2 in the α2β3 complex (rs5911 I843S encoded by the ITGA2B gene, allele frequencies in Caucasian population are approximately 0.6 and 0.4, respectively) can be involved in FNAIT, RPT and PTP. Already for decades it is known that HPA-3 antibodies can be difficult to detect with GP-specific antibody detection assays, as was also shown in the 15th and 19th international society of blood transfusion platelet immunology workshops [
]. A changed glycoprotein conformation and glycoprotein de-glycosylation caused by the platelet solubilization steps and storage are thought to be underlying causes for the low and varying anti-HPA-3 detection [
]. Furthermore, because HPA-3 is localized close to the platelet membrane, there can be damage because of the solubilization process even in glycoprotein-specific methods with pre-solubilization serum incubation [
]. It is therefore recommended to (also) use intact platelet assays for the detection of HPA-3 antibodies. Antibodies against HPA-3a are detected in approximately 2 % of FNAIT cases [
], which is in accordance with the experience from our laboratory (see above). Alloantibodies against HPA-3b are very rare. We only detected (in PIFT not in MAIPA) these antibodies twice in a 30 years period, both in FNAIT cases. Very recent Zhang et al. showed promising results for the detection of HPA-3 (and HPA-9) antibodies using CD41+/CD42b+ human megakaryocytes differentiated from bioengineered induced pluripotent stem cell lines (iPSCs). This might increase the percentage of detected HPA-3 antibodies [
HPA-4a and -b are located on β3 in the α2β3 complex (rs5917 R143Q encoded by the ITGB3 gene, among Caucasians these antigens are expressed by nearly 100 % and <0.1 %, respectively; among Asians the frequencies are 99.9 % and depending on the population 0.9–1.7 %, respectively). In the Asian population, immunization against HPA-4b is regularly seen. Ohto et al. (2004) [
] detected 223 HPA antibodies in serum samples from 24,630 pregnant (most Japanese) women, i.e. anti-HPA-5b (n = 168), anti-HPA-4b (n = 49), anti-HPA-5a (n = 3), anti-Naka (n = 2), anti-HPA-4b + 5b (n = 3). Two of the neonates born to mothers having anti-HPA-4b showed generalized purpura. No severe neonatal bleeding was seen.
We do not routinely screen for anti-HPA-4 antibodies, but we do perform a cross-match between maternal serum and paternal platelets in MAIPA. The commercial PAKLx kit contains HPA-4 typed beads including one HPA-4b positive bead.
In our series with a total of 326 FNAIT cases with detectable HPA antibodies, anti-HPA-5 antibodies were detected in 69 cases of which seven anti-HPA-5a and 62 anti-HPA-5b (nine in combination with anti-HPA-1a antibodies). HPA-5 is located on GPIa of the GPIa/IIa complex (Very Late Antigen-2, α2β1, CD49b/CD29, rs10471371 E505 K encoded by the ITGA2 gene, allele frequencies in Caucasian population 0.92 and 0.08, respectively), of which approximately 1000–2500 copies are present per platelet [
]. Therefore, the number of HPA-5a or -5b copies per platelet in the per definition HPA-5 heterozygous neonates suffering from FNAIT is approximately 500–1250 copies per platelet. This low expression, compared with 40.000 copies of HPA-1a in heterozygous neonates, may be an important reason for the often only mildly decreased or normal platelet counts in children born of women who are HPA-5b immunized. This low expression is also the reason why anti-HPA-5 antibodies are often undetectable in whole platelet HPA antibody detection assays and can best be detected with the more sensitive GP-specific methods [
]. The antibody detection results in the 16th-19th workshops of the ISBT International Platelet Immunology Working Party showed excellent results (95–100 %) for the detection of both HPA-5a and -5b antibodies in MAIPA or MACE.
Also the platelet expression of HPA-15 located on CD109 (rs10455097 S682Y, HPA-15a and -15b allele frequencies in Caucasian population 0.5 and 0.5, respectively) is low, i.e. approximately 2000 copies per platelet, and shows inter-individual variability [
]. However, because of the low CD109 expression, fetal/neonatal platelet destruction due to maternal HPA-15 antibodies might not always cause severe fetal/neonatal thrombocytopenia and no diagnostics will be requested leaving the antibodies undetected. The importance of anti-HPA-15 antibodies in PTR patients might also be overestimated as these antibodies are mostly found in combination with anti-HLA antibodies. For an optimal detection of HPA-15 antibodies it is necessary to use homozygous HPA-15a or -15b typed donor platelet suspensions from donors with a known high expression of CD109 [
The wide spread class B scavenger receptor CD36 is known as platelet GPIV (NAKa) on platelets A lack of GPIV on platelets (known as type II CD36 deficiency) is seen in approximately 0.1 % of Caucasians, 1.3 % in Chinese, 2.4 % African-Americans, 4.8 % Afro-Caribbean’s, 8 % sub-Saharan Africans and 5–10 % Japanese [
Quantitation of CD36 (platelet glycoprotein IV) expression on platelets and monocytes by flow cytometry: application to the study of Plasmodium falciparum malaria.
]. Immunization against GPIV, also known as NAKa can occur in a small percentage of these individuals (approximately 0.5 % of Asians and Africans) who also lack CD36 on other cell types (type I CD36 deficiency), causing FNAIT or PTR. Antibodies against NAKa can best be detected with GP-specific antibody detection assays, but false negative results can occur because of GP-specific mouse-anti-human moab’s blocking the human antibody binding [
]. We noticed this in a sample reactive in PAKLx but not in MAIPA with moab P58 (data not shown).
3. Conclusion
Considerable progress has been made in the last four to five decades in platelet antibody detection. The availability of labeled anti-human-Ig antibodies, platelet membrane solubilization, GP-specific moab’s, Luminex beads and recombinant GP constructs in combination with HPA (geno)typing possibilities have changed platelet serology. However, there still is not one method which can detect all possible clinical important antibodies. This can be partly assigned to the specifics of the methods but also because of the, for certain antibodies demonstrated, variation in antigen recognition. Depending on the clinical situation different approaches and multiple assays are necessary. It remains clear that antibody detection should be done in laboratories with sufficient knowledge and experience and that also for those (reference) laboratories standardization of test methods is necessary. Finally, in addition to the detection and quantification of antibodies, various aspects of HPA antibodies, e.g. the ‘αvβ3 HPA-1a’ specificity and the IgG Fc-tail glycosylation patterns, have been shown to play an important role in the occurrence of endothelial damage [
Platelet plasma membrane glycoproteins. Evidence for the presence of nonequivalent disulfide bonds using nonreduced-reduced two-dimensional gel electrophoresis.
Characterization of human platelet proteins solubilized with Triton X-lOO and examined by crossed immunoelectrophoresis. Reference patterns of extracts from whole platelets and isolated membranes.
Sensitivity of the platelet immunofluorescence test (PIFT) and the MAIPA assay for the detection of platelet-reactive alloantibodies: a report on two U.K. National Platelet Workshop exercises.
A modified rapid monoclonal antibody-specific immobilization of platelet antigen assay for the detection of human platelet antigen (HPA) antibodies: a multicentre evaluation.
Comparison of the direct platelet immunofluorescence test (direct PIFT) with a modified direct monoclonal antibody-specific immobilization of platelet antigens (direct MAIPA) in detection of platelet-associated IgG.
Low-avidity anti-HPA-1a alloantibodies are capable of antigen-positive platelet destruction in the NOD/SCID mouse model of alloimmune thrombocytopenia.
Development of a single-antigen magnetic bead assay (SAMBA) for the sensitive detection of HPA-1a alloantibodies using tag-engineered recombinant soluble β3 integrin.
A multicenter validation of recombinant β3 integrin-coupled beads to detect human platelet antigen-1 alloantibodies in 498 cases of fetomaternal alloimmune thrombocytopenia.
Studies on the binding of immunoglobulins and immune complexes to the surface of human platelets: IgG molecules react with platelet Fc receptors with the CH3 domain.
Involvement of the cysteine-rich domain of glycoprotein IIIa in the expression of the human platelet alloantigen, PlA1: evidence for heterogeneity in the humoral response.
HIP-Study (HPA-screening In Pregnancy): Protocol of a Nationwide, Prospective and Observational Study to Assess Incidence and Natural History of Fetal/Neonatal Alloimmune Thrombocytopenia and Identifying Pregnancies at Risk.
International Collaboration for Transfusion Medicine Guidelines. Maternal HPA-1a antibody level and its role in predicting the severity of Fetal/Neonatal Alloimmune Thrombocytopenia: a systematic review.
SSC of the ISTH. Interlaboratory workshop on anti-HPA-1a alloantibody quantification with the mAb-specific immobilization of platelet antigen technique.
Quantitation of CD36 (platelet glycoprotein IV) expression on platelets and monocytes by flow cytometry: application to the study of Plasmodium falciparum malaria.
30278-2&id=gr1.jpg){kind=link}
30278-2&id=gr2.jpg){kind=link}