a) Update and quality control issues on laboratory testing for ITP.  - Kiefel V.        

Dr. Kiefel reviewed principles and practical aspects of the laboratory testing for platelet autoantibodies in patients with autoimmune thrombocytopenic purpura (AITP). Issues raised included: should testing be performed (i.e., what is their diagnostic value)? What are the relevant target platelet antigens? Should efforts be made to detect non-IgG antibodies (i.e., IgA, IgM). Which assays should be set up (PAIgG vs GP-PAIgG vs testing platelet eluates vs testing free serum/plasma platelet antibody)? How do commercial assays perform?

In general, the diagnostic specificity of classic "PAIgG" assays for AITP was said to be very poor, and essentially of no diagnostic value. Discrepancies in the amounts of PAIgG quantified (ranging from 0.2 to 400 fg/platelet) among various assays and research laboratories also exist. In part, this may result from IgG binding to platelets in non-immune fashion (e.g., IgG uptake into platelet alpha granules, IgG complexes bound to platelet Fc receptors, "nonspecific" binding of IgG to platelet membranes) as well as "specific" binding of autoantibodies against GP target(s). Certain assays may detect PAIgG in these different platelet compartments differently. In contrast, the presence of GP-specific autoantibodies (e.g., detected using direct MAIPA) is much more specific (80-95%) for AITP, although the sensitivity is lower (about 40-50%). Another approach is to obtain an eluate from platelets, and then react the eluate against platelets: only platelet autoantibodies react under these circumstances, and this results in sensitivity and specificity for AITP similar to that of the direct MAIPA. A comparison of direct vs indirect testing for GP-specific autoantibodies showed that direct assays are far more likely to detect autoantibodies than indirect assays, which are positive only in 3-12% of patients. Other rare, but occasionally relevant target autoantibodies include: GPV, Ia/IIa, VI, GMP140, and CD-9.

The laboratory also needs to review the peripheral blood film to rule out disorders such as pseudothrombocytopenia, platelet satellitism, May-Hegglin anomaly, Sebastian syndrome, etc. A commercial GP-specific assay was assessed by Dr. Kiefer but was found to react poorly with many platelet autoantibodies (this assay is adequate for detecting alloantibodies).

Some technical issues were raised:

1. sufficient positive and negative samples should be tested to determine the upper range of normal, which ideally should be set at 0.15-0.2 OD;
2. select the right monoclonal antibodies; and
3. be aware of GPIb/IX/V proteolytic degradation, which can cause autoantibodies to be falsely non-reactive. Regarding counseling of patients, it is important to point out that since the test is moderately sensitive, but highly specific, a negative test does not rule out AITP, though a positive test is strong evidence for AITP.

Some conclusions were offered: (a) obtain sufficient blood samples for testing; (b) IgG antibody testing is sufficient; (c) GPs IIb/IIIa, Ib/IX, and V are the most relevant target antigens; (d) degradation of GPIb/IX/V complexes is a relevant issue; and (e) commercial assays should be assessed by experienced labortories with well-defined reagents.

b) Clinical approach to the diagnosis of AITP. - Bussel J.      
    
Currently, the diagnosis of AITP is "clinical." This is based upon the low sensitivity (50%) of even of the most advanced assays, and also the lack of availability of advanced platelet antibody testing in many medical centers. The author poses: can a diagnostic algorithm for AITP be developed? Drew Provan has developed an algorithm in an ITP registry developed by the EHA Working Party. "Confirmation" of diagnosis would be based upon results of laboratory testing as well as response to treatment (if any) or observation of the natural history of the thrombocytopenia. In children, the diagnosis is based upon a low platelet count, otherwise unremarkable CBC (Complete Blood Count), absence of hepatosplenomegaly, absence of lymphadenopathy, and no evidence of associated diseases (CLL, TTP, drug-induced thrombocytopenia, MDS, etc.). If available, other lab tests that might be useful to collect include: aPTT (screen for antiphospholipid antibodies), HIV, immunoelectrophoresis, thyroid stimulating hormone, direct antiglobulin test, lactate dehydrogenase, reticulocyte count, urea, creatinine, bone marrow examination (including special studies such as cytogenetics). Also important would be "clinical evolution" assessed at 3-6 mo. intervals, with focus on serial blood counts and bleeding complications. "Gold standards" would include chronicity of stable thrombocytopenia (i.e., no evolution to another disease), response to treatment, e.g., IVIG or anti-D (especially if repeated response); response to corticosteroids and splenectomy is considered less specific. It is suggested that at least 100 patients would be required to establish a preliminary algorithm, with additional patients being used to refine the algorithm. Funding would be required for all parts of the registry. Choice of laboratory and /or cytogenetic studies are not necessarily straightforward. Potential advantages of developing this algorithm could include: more precise diagnosis and better management of AITP; identification of subtypes of AITP, better discrimination regarding published studies (by using established diagnostic criteria according to a validated algorithm). Potential disadvantages: this might be too complex a topic (too many variables), too many patients required, too much missing data, etc., might be relevant issues.

A small working party will be formed to formulate diagnostic criteria and develop an diagnostic algorithm for AITP, and to report to the subcommittee in 2004.

a) Guidelines for diagnosis and management of fetal/neonatal alloimmune thrombocytopenia (FNAIT). Murphy M, Kaplan C, Bussel J, Kroll H. (Presented by H Kroll and C Kaplan)

i) Dr Kroll presented on laboratory issues. The last guidelines were developed in 1991 by the Neonatal Hemostasis Subcommittee of ISTH. However, significant developments in the last 12 years in laboratory detection of alloantibodies and antenatal management have occurred. A proposal for revised guidelines arose from a joint initiative of the ISTH/ISBT in collaboration with a European Working Group, which met in 2002 in Belgirate, Italy. The draft guidelines include the following issues.

1. When to suspect fetal AIT: fetus with abnormal cerebral ultrasound, hydrocephalius, porencephalic cyst, unexplained fetal anemia/hydrops, unexplained in utero death in second or third trimester, recurrent late miscarriage, modified fetal activity, unexplained incidentally-detected fetal thrombocytopenia; and known maternal immunization.
2. When to suspect neonatal AIT: clinically apparent bleeding; sibling or family history of neonatal bleeding or thrombocytopenia; maternal autoimmune thrombocytopenia. Laboratory workup should be performed when the neonatal platelet count is <50 X109/L or any unexplained thrombocytopenia.
3. Laboratory testing should be performed in a reference laboratory, including maternal platelet antibody testing, parental platelet typing for HPA-1,3,5 alloantigens (depending on ethnic background), with further testing depending on results of initial investigations. Antibody detection tests include those using intact platelets (PSIFT, PAIFT, ELISA); glycoprotein-specific assays (MAIPA, MACE); binding assay with chloroquine-treated platelets; platelet phenotyping (using above methods), or platelet genotyping (various methods described). Maternal antibody testing should always include cross-match against paternal platelets, using global and GP-specific assays, against at least GPIIb/IIIa, Ia/IIa, IbIX, ?CD109, ?CD36), as well as against a typed donor panel; testing for maternal platelet autoantibodies should also be performed; also, it was recommended to test for HLA class I antibodies and for ABO incompatibility, although the clinical significance of anti-HLA and/or anti-ABO antibodies remains unclear. Technical issues include epitope-specific problems, e.g., HPA-2 alloantigens are protease-sensitive; HPA-3 alloantigen destruction occurs during storage; CD109 is unstable; thus, fresh platelets should be used. Also, careful selection of monoclonal antibodies is important because of the potential for antibody competition. Diagnostic interpretation of certain reaction patterns was also discussed. Difficult remaining issues include: undetermined significance of HLA and ABO antibodies; antibodies can be undetectable even when they are strongly suspected to be present (HPA alloantigen incompatibility in a fetus/neonate with clinically-suspected FNAIT), new (as yet undiscovered) alloantigens.

ii) Dr. Kaplan presented on management issues. Most cases of FNAIT are unexpected (i.e., no family history). It was emphasized that the main goal of treatment is to transfuse compatible platelets (whether from mother or a phenotyped donor) if the platelet count is <30 x 109/L or bleeding is present. IVIg is not considered a substitute for platelet transfusions. Discordant results using random platelets have been reported. If the platelet count is >30x 109/L, close monitoring is required. Usually, subsequent pregnancies are more severely affected (notwithstanding data from Scandinavian prospective studies), and so current management is aimed at preventing intracranial hemorrhage during pregnancy and delivery (although the optimal means to achieve this remains debatable). Management options include:

1. weekly maternal injections of high-dose IVIG (with or without corticosteroids);
2. repeated intrauterine platelet transfusions (performed at experienced centre); and
3. fetal genotyping should be performed if the father is heterozygous for the implicated alloantigen, to determine if the fetus might be unaffected and not at risk. Fetal blood sampling may be required for this. In general, the risks of repeated fetal blood sampling/transfusion suggest that maternally-directed therapy with IVIG might be the first line of treatment, with maternal therapy stratified based upon the family history. Treatment issues include the uncertain effect of IVIG (no randomized data, plus the possibility of improving natural history with subsequent pregnancies).

Antenatal screening issues were also discussed. The rationale for screening includes the (retrospective) data that 50% of affected infants are first-born, with 10% death rate, and 20% neurologic sequelae (intracranial hemorrhage), and the observation that fatal hemorrhage can occur in utero. Prospective studies indicate that there is a 2/1000 chance of anti-HPA-1a immunization during first pregnancy, with the prevalence of neonatal thrombocytopenia linked with anti-HPA-1a at about 1/1000. Overall, the risk of FNAIT appears to be about 1/800 pregnancies. Currently, systematic screening for FNAIT has not been performed since 1996 (despite promising pilot studies). This is a public health issue. However, practical issues include the difficulty in predicting which alloimmunized pregnant women would have affected offspring, and what the optimal therapy would be. Advantages and disadvantages of maternal versus neonatal screening were discussed. Neonatal screening was more cost-effective in one study. In conclusion, it was stated that it is difficult to justify routine screening in the absence of well-established policies on antenatal risk assessment and appropriate treatment.

It is proposed that one manuscript on guidelines for the diagnosis and another for the management of FNAIT will be submitted for review and approval as official SSC publications.

b) Recombinant platelet proteins and EBV-transformed B-cell lines for the improvement of platelet antibody and antigen detection.-  Kroll H, Santoso S. (Presented by Dr. Kroll)
    
Reference material and techniques are required for alloantigen genotyping. For antibody characterization, the following are desirable: recombinant antigens (transient/stable transfectants and purified proteins), synthetic peptides, recombinant antibodies, and reference techniques. The platelet-specific alloantigens were summarized (9 HPA-designated polymorphisms on GPIIIa, 2 HPA-designated polymorphisms on GPIIb, 2 HPA-designated polymorphisms on GPIa, 1 HPA-designated polymorphism each on GPIb-alpha and GPIb-beta, 1 HPA-designated polymorphism on CD109, and 1 as yet non-designated alloantigen on IIb or IIIa known as Va. In many cases, availability of reference material for testing is hampered by the rarity of platelets or DNA from individuals of interest. Thus, HPA alleles of have been prepared using EBV-transformed cell lines, including the alleles of HPA-1a/b, -2a/b, -3a/b, -4a/b, -5a/b, -6b, -7b, -8b, 9b, 11b, 12b, and 13b. There is now availability of EBV-transformed B-cell lines from most of the low frequency alleles (missing HPA-10 and -16). A genotyping workshop evaluating proficiency testing using the transformed cell lines showed errors in typing were made in 6 of 24 labs. Currently, most of the HPA alleles are availability in the repository laboratory (Giessen). The generation of transfected mammalian cell lines expressing platelet antigen isoforms was summarized: 1. Place platelet GP cDNA in an expression vector; 2. Perform site-directed mutagenesis to obtain the rare allele of interest to generate the allele-specific cDNA construct; 3. Perform transfection in a mammalian cell line, resulting in surface expression of the alloantigen of interest (e.g., use CHO cells). Examples of use of this technique in identifying platelet alloantibodies were shown. This technique can be applied to obtain the 10 "private" alloantigens. The speaker also discussed a mutation (Csy345Ala) on the beta-3 chain of GPIIb/IIIa that leads to loss of a long-range disulfide bond. Interestingly, these leads to isoforms of GPIIIa that exhibit different reactivity among anti-HPA-1a alloantisera, with so-called type II anti-HPA-1a requiring an intact 3-dimensional protein structure (i.e., Cys345 is required). Type II anti-HPA-1a represents a minority of patients with anti-1a alloantibodies (NAT and PTP). In summary, cell lines expressing recombinant platelet GPs are a useful tool for alloantibody detection. Standard assays such as MAIPA can be used with these transfected cells. A panel of rare recombinant alloantigens can be made available for investigations of certain patients. Affinity-purified recombinant GPs can be used; certain modified/truncated proteins and peptides that are recognized by only a subpopulation of antibodies can be studied.

c)  Human platelet alloantigens (HPAs) - Santoso S, Ouwehand W.

i) Nomenclature of  Human Platelet Alloantigens.- Santoso S.     In 1990, the HPA nomenclature system was introduced (von dem Borne). However, the molecular genetic basis of the platelet alloantigens it was suggested by P. Newman. There was a need for supplementary nomenclature describing the various alleles. The use of both molecular-based and HPA nomenclature was accepted by the ISBT in 1995. Given the increasing attention to platelet alloantigens both by Transfusion Medicine and Thrombosis & Hemostasis societies, a joint consensus committee was approved by both ISBT and ISTH. The first meeting occurred at SSC of the ISTH (Paris, 2001) consisting of a Joint Platelet Nomenclature Committee (JPNC), with representatives from both societies. Uniform HPA nomenclature and guidelines for HPA inclusion were recommended, along with recommended membership of reference and repository laboratories. At the next SSC meeting (Boston 2002), the proposal of uniform nomenclature and guidelines was presented and accepted. Then, at the ISBT meeting (Vancouver 2002), the proposal was also discussed and accepted by that society. The final document (SOP) will be verified by the JPNC and published in the official journals of both societies, and in a website. Representatives of the HPA Nomenclature Committee include: P Metcalfe (UK), N A Watkins (UK), W H Ouwehand (UK), C Kaplan (France), P Newman (USA), R Kekomaki (Finland), M de Haas (Netherlands), R H Aster (USA), Y Shibata (Japan), J Smith (Canada), V Kiefel (Germany) and S Santoso (Germany). The SOP deals with the following: (1) Membership of the JPNC; (2) Responsibilities of the JPNC; (3) Roles of the reference and repository laboratories; (4) HPA nomenclature–terminology and rules; and (5) Rules and organization for assignment of new HPA. There are ten voting and two non-voting members of the JPNC, with three executive positions: chairman (C Kaplan), co-chairman (S Santoso), and secretary (W Ouwehand). Reference laboratories are located in six countries (USA, Canada, Netherlands, Germany, France, UK) and repository laboratories are included in Germany and the UK. Some examples of HPA nomenclature were shown, such as the designation of the high-frequency allele of Gov (Govb) designated as HPA-15a.

A report on nomenclature of human platelet alloantigens will be submitted for review and approval by SSC Council as an official SSC publication.

ii) Single nucleotide polymorphisms (SNPs) in blood cell genes: a model for more to come. - Ouwehand W.
    
The human genome consists of 3 billion base-pairs (about 28,000 genes), with the difference between humans (same gender) of only 0.1%. Most differences are caused by SNPs, small deletions/insertions, or gene amplification/deletion. Examples of mutations relevant to transfusion medicine were described (ABO, RhD, Lu): of these three, the first and last are examples of SNPs leading to significant phenotypic consequence. Modeling studies of the platelet alloantigens reveals mutations typically to be on loops. For example, the HPA-2 alloantigens are a T145M mutation in the GPIb-alpha. The speaker emphasized how large a "footprint" the alloantibody places in relation to the size of the target protein (20 x 30 Angstroms-squared). The Immuno Polymorphism Database (IPD)-HPA was shown (from the European Bioinformatics Institute [EBI]), which summarizes information on the human platelet alloantigens.
    
Current research activities aimed at elucidating a biological role for SNPs in platelet physiology were summarized. For example, there is "hard-wired" variation to stimulation by ADP among normal humans, with possible relevance to cardiovascular disease. The aim is to identify those SNPs that would be associated with greater relative risk (>1.5 RR) of cardiovascular events. The aim is to identify the genetic platelet map associated with platelet function and, ultimately, cardiovascular risk. This is performed by identifying platelet-specific transcripts, identifying the SNPs in these genes, and then determining their effects on platelet functions. Microarray techniques are used to facilitate this process, with gene products of erythroblasts and megakaryocytes showing distinct profiles. The Wellcome Trust Sanger Institute high throughput sequencing machines will be used to determine the common alleles of over 500 target genes. GPVI was used to validate the pipeline. Eighteen SNPs were identified by resequencing the GPVI exon. Overall, 65%, 25%, and 2% of GPVI SNPs could be designated as aa, ab, and bb, but 7% of the caucasoid population had rare "hybrid" GPVI alleles. The frequency of the various GPVI isoforms varies among different populations. Comparison of the aa vs bb GPVI genotype with respect to platelet activation generally reveals greater activation (or lower amounts of agonist required) with the aa individuals. Results from clinical studies evaluating an association with cardiovascular disease or bleeding risk are preliminary and non-conclusive.

a) Thrombocytopenia caused by sensitivity to GPIIb/IIIa inhibitors: an update. - Aster RH.
    
Dr. Aster discussed thrombocytopenia due to GPIIb/IIIa inhibitors (new class of antithrombotic agents used in the patients undergoing percutaneous coronary interventions [PCIs]). In general, the thrombocytopenia occurs after the first exposure to the drug, due to the presence of naturally-occurring antibodies. Three agents are approved: abciximab, eptifibatide, and tirofiban. Abciximab is a humanized chimeric Fab fragment that binds to GPIIb/IIIa. Abciximab-induced thrombocytopenia occurs in about 1% of patients, usually within a few hours after the first administration, and is sometimes associated with bleeding (occasionally, anaphylactoid reactions). This syndrome must be distinguished from pseudothrombocytopenia ( the platelet count fall is not usually profound with pseudothrombocytopenia and so, a platelet count of 10 x 109/L or less is the true form). Patients with abciximab-induced thrombocytopenia have IgG antibodies that react against abciximab-coated platelets; however, there is considerable overlap with normal subjects. To distinguish "normal" antibodies from antibodies associated with abciximab-induced thrombocytopenia, inhibition of reactivity by Fab fragments prepared from pooled IVIG preparations is useful ("normal" antibodies are inhibited). The other is to test patient antibodies against both 7E3-coated platelets as well as platelets coated with other GPIIb/III-reactive monoclonal antibodies, e.g., AP3: a higher ratio of binding to 7E3-coated platelets to AP3-coated platelets is seen with antibodies from patients with abciximab-induced thrombocytopenia, compared with "normal" antibodies. The distinction seems to be that abciximab-induced antibodies react with the mouse portion of abciximab, whereas the "normal" antibodies react against the papain-cleavage site of the N-terminus of the abciximab Fab region.
    
There is also a syndrome of "delayed thrombocytopenia" after abciximab (not widely recognized). Antibodies newly formed in response to abciximab exposure develops thrombocytopenia about a week later (which can be severe). Another topic is thrombocytopenia caused by the ligand-mimetic inhibitors of GPIIb/IIIa (tirofiban [mimics R-G-D], eptifibatide [which contains R-G-D]). The acute-onset thrombocytopenic syndrome is similar to abciximab, except that pseudothrombocytopenia has not been reported. The antibodies can generally be readily detected in the presence of the drug (flow cytometry). Pre-treatment samples show similar reactivity as the post-treatment sample; thus, these are also naturally-occurring antibodies. The precise structure of the epitope recognized by these antibodies remains unclear. It is unlikely that the antibodies are specific against the binding site of the drug itself, because the antibodies from different patients differ in their calcium requirements, only some are blocked by abciximab, some tirofiban-dependent antibodies cross-react with eptifibatide and vice versa, etc. Thus, these fiban-dependent antibodies could be specific for conformational changes induced in GPIIb/IIIa elsewhere by the binding of the (fiban) ligands. It is worth noting, however, that fiban-dependent antibodies do not generally react with platelets activated by agonists such as thrombin, collagen, RGD peptide or LIBS-specific monoclonal antibodies. This suggests that each ligand induces its specific family of epitopes (LIBS determinants). There remain unanswered questions: what epitopes on GPIIb/IIIa are recognized by these antibodies? Can a simple screening test to detect these antibodies be derived? If so, would screening before treatment reduce the incidence of severe thrombocytopenia? What is the significance of naturally-occurring antibodies that recognize GPIIb/IIIa?

b) Diagnosis of heparin-induced thrombocytopenia (HIT): problems and proposed refinements of diagnostic criteria. - Warkentin TE.
    
Three uses of scoring systems were described: validation of new assays for HIT, estimation of pretest probability of HIT, and to establish diagnostic criteria for HIT. The rationale for why HIT cannot necessarily be diagnosed solely by clinical or laboratory criteria was explained, with the conclusion being that HIT should be diagnosed by both clinical and laboratory criteria, i.e., a clinico-pathologic syndrome. Thus, one (common) approach to diagnose HIT is for the clinician to estimate the pretest probability of HIT based upon criteria availability at the time of patient assessment, and then to alter this pretest probability based upon the results of laboratory testing for HIT antibodies (Bayesian approach), i.e. pretest probability X likelihood ratio = posttest probability. A scoring system presented in Boston (2002) at the Platelet Immunology SSC based upon 3 clinical parameters (timing of onset of thrombocytopenia; thrombosis or other sequelae of HIT, occurrence of other explanations for the thrombocytopenia) was shown.  However, it had been suggested that thrombocytopenia should be considered to be a fourth criterion. This was added: thus, a 4-item scoring system ("the 4 T’s") was presented: (1) Thrombocytopenia; (2) Timing of onset of thrombocytopenia; (3) Thrombosis (or other sequelae of HIT); and OTher cause for thrombocytopenia apparent. Each item could score 0, 1, or 2; thus, 8 was the maximum score, 6-8 high pretest probability for HIT, 4-5 moderate pretest probability for HIT, and 1-3 low pretest probability for HIT. Regarding the newly added criterion of thrombocytopenia, very low platelet counts counted lower scores because HIT uncommonly gives platelet count falls to <20, and only rarely leads to platelet counts <10. Additionally, a proportional fall of >50% (though platelet nadir >20) was given the maximum score of 2, since such a large-magnitude platelet count fall is suggestive of HIT even if the platelet count does not fall to <150. The operating characteristics of a washed platelet activation assay (platelet serotonin release assay) was shown: a result of 50% release is associated with a likelihood ratio for HIT of about 40 (post-orthopedic surgery patient) and a likelihood ratio of about 7 (post-cardiac surgery patient). Preliminary results of the scoring system based upon 50 patients assessed was presented: the number of patients testing positive (>50% serotonin release) based upon the categories, low, moderate, and high pretest probability were: 3%, 23%, and 100%. It was concluded that a scoring system based upon the 4 clinical criteria proposed could provide useful information for determining pretest probability for HIT.

c) Drug-induced thrombocytopenia: drug-dependent antibody assays and recent advances. - Chong BH.
    
Dr. Chong presented a table of drugs that cause immune thrombocytopenia. Heparin, quinine/quinidine, and GPIIb/IIIa inhibitors are relatively common, compared with others (alprenolol/oxprenolol, ampicillin/methicillin, carbamazepine, cephalothin, co-trimoxazole, chlorothiazine, digoxin, gold, methyldopa, ranitidine/cimetidine, rifampicin, sulfathiazole/sulfasoxazole, valproic acid). Two topics were discussed: drug-dependent antibody assays (methods, etc.), and GPIX epitope mapping. Regarding methods, there are direct (i.e., detect antibodies on platelets) and indirect tests (detect antibodies in patient serum/plasma). The history of antibody testing was reviewed. It covered three periods: phase I (1975: measure platelet changes, such as platelet aggregation/agglutination, complement-mediated lysis, serotonin release, etc.), phase II (1975-1987; indirect PAIgG, e.g., ELISA, flow cytometry, radioisotope assay, etc.); phase III (1987-present; platelet antigen-specific assays, e.g., microtiter well assay, immunobead assay, modified-antigen capture ELISA [MACE], MAIPA). Quality control and standardization issues involve: use of positive controls, especially weak-positive control (which might be difficult to obtain), use of negative control, use of group O platelets (however, this might not be so big an issue), and appropriate dilution of patient serum and secondary antibodies. Important technical issues include drug solubility, whether the drug or metabolite is to be used, concentration of the drug (which may be higher than the therapeutic range), and requirement to include the drug (or metabolite) in the washing buffer. Dr. Chong reported that the MAIPA (using GPIb-IX [n=13] than GPIIb/IIIa [n=3 among the 13 patients]) was more sensitive that the flow cytometry assay, perhaps because lower dilutions can be used in the MAIPA. The specificity of these assays was considered to be very high (approaching 100%). The antigen targets of D-ITP were discussed: GPIb/IX, GPIIb/IIIa, PECAM-1, and GPV.
    
Mapping of the epitope(s) of GPIX for quinine-induced thrombocytopenia was also presented. Chimera between mice and human GPIX constructs were generated so as to map the antigens. SZ1 monoclonal antibody also was used, as it blocks (at the C-terminal site) many quinine-dependent antibodies (as well as rifampicin- and ranitidine-induced thrombocytopenia antibodies). The epitope region was narrowed down to the L80–D151 amino acid region, and further work was concentrated there. One of the constructs (construct 4) had greatly diminished binding. Only amino acids 126 and 131 were changed to make these changes in the construct, with R126 shown to be the most important amino acid responsible for forming the epitope.