Who should donate blood? Policy decisions on donor deferral criteria should protect recipients and be fair to donors.

An important element in the development of voluntary blood donation schemes throughout the world has been the attention given to minimising the risk to recipients of donated blood, primarily the risk of transfusion transmitted infections. In response to the appearance of human immunodeficiency virus (HIV) in the 1980s a range of national policies emerged that excluded populations at high risk of contracting HIV from donating blood, with a particular focus on men who have sex with men (MSM), the primary reason being the protection of recipients of donated blood. Recently some countries, including the UK, have revised their policies, informed by advances in screening tests, epidemiological evidence of transmission rates and an increasing concern about unfair discrimination of specific groups in society. Policy makers face a difficult task of balancing safety of recipients; an adequate blood supply for those who require transfusion; and societal/legal obligations to treat everyone fairly. Given that no transfusion is risk free, the question is what degree of risk is acceptable in order to meet the needs of recipients and society. Decisions about acceptance of risk are complex and policy makers who set acceptable risk levels must provide ethically justifiable reasons for their decisions. We suggest it is possible to provide a set of reasons that stakeholders could agree are relevant based on careful evaluation of the evidence of all relevant risks and explicit acknowledgement of other morally relevant values. We describe using such a process in the Safety of Blood Tissue and Organs (SaBTO) review of donor deferral criteria related to sexual behaviour.

Inhibition of MT1-MMP activity using functional antibody fragments selected against its hemopexin domain.

The membrane associated MMP, MT1-MMP, is a critical pericellular protease involved in tumour cell invasion and angiogenesis and is highly up-regulated in numerous human cancers. It therefore represents an exciting new therapeutic cancer-specific target. We have generated recombinant human scFv antibodies against the non-catalytic, hemopexin domain of MT1-MMP that modulate its interactions with collagen. One of these is an effective inhibitor of the invasive capacity of cancer cells and of angiogenesis in model systems. This demonstrates that targeting sites outside the catalytic domain presents a potential novel approach to proteinase inhibition that could have applications in cancer therapeutics.

Immunologic and structural analysis of eight novel domain-deletion beta3 integrin peptides designed for detection of HPA-1 antibodies.

BACKGROUND: The single-nucleotide polymorphism (SNP) rs5918 in the ITGB3 gene defines the human platelet antigen-1 (HPA-1) system encoding a Leu (HPA-1a) or Pro (HPA-1b) at position 33. HPA-1 antibodies are clinically the most relevant in the Caucasoid population, but detection currently requires alpha(IIb)beta3 integrin from the platelets of HPA-genotyped donors. OBJECTIVES: We set out to define the beta3 integrin domains required for HPA-1a antibody binding and produce recombinant soluble beta3 peptides for HPA-1 antibody detection. METHODS: We designed two sets (1a and 1b) of four soluble beta3 domain-deletion peptides (deltaSDL, deltabetaA, PSIHybrid, PSI), informed by crystallography studies and computer modeling. The footprints of three human HPA-1a-specific phage antibodies were defined by analyzing binding patterns to the beta3 peptides and canine platelets, and models of antibody-antigen interfaces were derived. Specificity and sensitivity for HPA-1a detection were assessed using sera from 140 cases of fetomaternal alloimmune thrombocytopenia (FMAIT). RESULTS: Fusion of recombinant proteins to calmodulin resulted in high-level expression in Drosophila S2 cells of all eight beta3 peptides. Testing of FMAIT samples indicated that deltabetaA-Leu33 is the superior peptide for HPA-1a antibody detection, with 96% sensitivity and 95% specificity. The existence of type I and II categories of HPA-1a antibodies was confirmed by the study of HPA-1a phage antibody footprints and the reactivity pattern of clinical samples with the four beta3-Leu33 peptides, but there was no correlation between antibody category and clinical severity of FMAIT. CONCLUSIONS: Soluble recombinant beta3 peptides can be used for detection of clinical HPA-1a antibodies.

Three novel beta3 domain-deletion peptides for the sensitive and specific detection of HPA-4 and six low frequency beta3-HPA antibodies.

BACKGROUND: Antibodies against human platelet antigens (HPA) are clinically important in fetal-maternal alloimmune thrombocytopenia, refractoriness to platelet transfusions and post-transfusion purpura. Of the 16 HPAs, nine are located on the beta3 subunit of the alphaIIb beta3 integrin. Antibody detection is generally based on platelet-derived alphaIIb beta3 from HPA-genotyped donors. Recombinant allelic beta3 peptides, expressed at high levels would improve consistency in antibody detection, but the expression of soluble and monomeric integrins expressing complex dependent epitopes has previously proved challenging. OBJECTIVES: We aimed to generate three recombinant beta3 peptides for the detection of antibodies against HPA-4, HPA-8bw and five of the six remaining low frequency beta3 alloantigens. METHODS: The removal of the specificity-determining loop from the betaA domain and fusion of truncated beta3 to calmodulin was exploited to obtain expression of monomeric protein. Using site-directed mutagenesis, the mutations for HPA-4b and HPA-8bw were introduced in the ITGB3*001 haplotype. A third peptide for the detection of antibodies against HPA coded by non-synonymous single nucleotide polymorphisms of low frequency was generated by the introduction of five mutations forming the basis of HPA-6bw, -7bw, -10bw, -11bw, and -16bw antigens. RESULTS: Reactivity of the three peptides with beta3-specific murine monoclonal antibodies and human HPA-1a phage antibodies confirmed the structural integrity of the recombinant fragments, and reactivity with a unique panel of polyclonal anti-HPA sera confirmed expression of the relevant HPA epitopes. CONCLUSIONS: These data demonstrate that beta3 integrin domain-deletion fragments are suitable molecular targets for HPA antibody detection.

Severe fetomaternal alloimmune thrombocytopenia due to anti-human platelet antigen (HPA)-1a in a mother with a rare and silenced ITGB3*0101 (GPIIIa) allele.

BACKGROUND AND OBJECTIVES: Fetomaternal alloimmune thrombocytopenia (FMAIT) is caused by maternal antibodies against a human platelet antigen (HPA) present on fetal, but absent from maternal platelets. We identified and characterized a case of FMAIT due to anti-HPA-1a in a mother with an HPA-1a1b genotype. MATERIALS AND METHODS: The first child of a 29-year-old mother presented with a petechial rash and a platelet count of 8 x 10(9) per l. Upon routine serological investigation, a discrepancy between the HPA-1a genotype and phenotype prompted the sequencing of the 15 exons of the ITGB3 (integrin beta3, GPIIIa and CD61) gene in the mother. RESULTS: The mother was genotypically HPA-1a1b heterozygous but phenotyped as HPA-1a negative. Sequencing of the ITGB3 exons confirmed HPA-1a1b heterozygosity, but also identified a novel single nucleotide insertion in exon 10 leading to a frameshift and premature termination at amino acid 471 of ITGB3. Maternal anti-HPA-1a was detected but with a pattern typical for a low-affinity antibody. Three transfusions of HPA-1a and -5b negative neonatal platelet concentrates were required to return to a safe platelet count. CONCLUSION: A rare ITGB3 allele was uncovered by the investigation of a severe case of alloimmune thrombocytopenia in a mother with HPA-1a antibodies who genotyped as HPA-1a1b.

Definition of novel GP6 polymorphisms and major difference in haplotype frequencies between populations by a combination of in-depth exon resequencing and genotyping with tag single nucleotide polymorphisms.

BACKGROUND: Common genetic variants of cell surface receptors contribute to differences in functional responses and disease susceptibility. We have previously shown that single nucleotide polymorphisms (SNPs) in platelet glycoprotein VI (GP6) determine the extent of response to agonist. In addition, SNPs in the GP6 gene have been proposed as risk factors for coronary artery disease. METHODS: To completely characterize genetic variation in the GP6 gene we generated a high-resolution SNP map by sequencing the promoter, exons and consensus splice sequences in 94 non-related Caucasoids. In addition, we sequenced DNA encoding the ligand-binding domains of GP6 from non-human primates to determine the level of evolutionary conservation. RESULTS: Eighteen SNPs were identified, six of which encoded amino acid substitutions in the mature form of the protein. The single non-synonymous SNP identified in the exons encoding the ligand-binding domains, encoding for a 103Leu > Val substitution, resulted in reduced ligand binding. Two common protein isoforms were confirmed in Caucasoid with frequencies of 0.82 and 0.15. Variation at the GP6 locus was characterized further by determining SNP frequency in over 2000 individuals from different ethnic backgrounds. CONCLUSIONS: The SNPs were polymorphic in all populations studied although significant differences in allele frequencies were observed. Twelve additional GP6 protein isoforms were identified from the genotyping results and, despite extensive variation in GP6, the sequence of the ligand-binding domains is conserved. Sequences from non-human primates confirmed this observation. These data provide valuable information for the optimal selection of genetic variants for use in future association studies.

Analysing the platelet transcriptome.

Historical studies, using Northern blot hybridization, RT-PCR and cDNA library construction have demonstrated the presence of a variety of mRNA molecules in platelets. The development of microarray technology has allowed further characterization of the transcripts represented in the platelet transcriptome. In this review, these studies will be summarized and their findings in relation to the study of platelet function and the identification of disease risk genes discussed.

Studies of single-chain antibody expression in quiescent Escherichia coli.

Quiescent Escherichia coli cells are generated by overexpressing the Rcd transcript in an hns-205 mutant host. The resulting nongrowing, metabolically active cells were used here to express a single-chain antibody fragment (scFv) in shake flask and fermentor cultures. The expression system is based on two plasmids; one carries the product gene expressed from lambdaP(L) under the control of the cI857 temperature-sensitive repressor, while the second expresses Rcd from lambdaP(R). Shifting the culture from 30 to 42 degrees C induces Rcd expression and product expression simultaneously. Our scFv carried a PelB leader, and 90% of the protein was secreted into the culture supernatant. In a batch culture, the supernatant concentration of scFv in the quiescent-cell culture (optical density at 600 nm [OD(600)] of 3.5) was 37 mg x liter(-1), compared to a maximum of 13 mg x liter(-1) in the control culture (final OD(600) of 20). In a fed-batch fermentor culture, quiescent cells were held at an OD(600) of 20 for 24 h and the extracellular scFv concentration reached a maximum of 150 mg x liter(-1). A control culture with a similar feed reached an OD(600) of 80, but despite the higher density, the extracellular scFv concentration did not exceed 35 mg x liter(-1). Quiescent cells at an OD(600) of 50 exhibited a small decline in the specific product formation rate, but nevertheless, an extracellular scFv concentration of 160 mg x liter(-1) was achieved in 8 h. The rate of extracellular accumulation was 10-fold greater in the quiescent culture than in the control culture. This study demonstrates that it is possible to establish high-density quiescent E. coli cultures that are capable of efficient synthesis, folding, and export of proteins.

Molecular studies of anti-HLA-A2 using light-chain shuffling: a structural model for HLA antibody binding.

Human leukocyte antigen (HLA) A2 is one of the most immunodominant HLA antigens. Through a process of light-chain variable domain (VL) shuffling, we analyzed the VL domains' role in anti-HLA-A2/A28-binding site diversity. This was achieved by combining a VH3-30-encoded HLA-A2/A28-specific heavy-chain variable domain with 10(4) non-immune VL domains. Twelve HLA-A2/A28-specific antibodies were subsequently identified. VL gene analysis demonstrated an absence of Vlambda domains and that all have VkappaI-encoded light chains. The affinities correlated with the VkappaI gene present, with the seven highest affinity antibodies using Vkappa domains encoded by the O18 gene segment. A 300-fold difference in affinity was observed between the 12 antibodies, and homology modeling demonstrated a correlation between electrostatic surface potential of the antigen-binding site and affinity for HLA. Overlap between the T-cell receptor-binding site and that of the antibodies was indicated by inhibition of cytotoxic T-lymphocyte killing of peptide-pulsed target cells. A model of antibody binding to HLA-A2 suggested contact with both alpha helices of the HLA molecule, such that the antigen-binding site spans the peptide-binding groove. These data increase the understanding of antibody recognition of HLA and may facilitate the production of clonotypic antibodies with peptide-specific binding.

The isolation and characterisation of human monoclonal HLA-A2 antibodies from an immune V gene phage display library.

Molecular cloning techniques and V gene phage display have revolutionised the production of human monoclonal antibodies. Antibodies of a defined specificity can be obtained by selecting phage display libraries on antigen in a process known as panning. We have applied these techniques to the isolation of three HLA-A2-specific single chain variable domain fragments (scFv) from a patient alloimmunised by blood transfusion. Analysis of specificity with cells of HLA genotyped donors revealed the following: i) in addition to the major reactivity with HLA-A2, cross-reactivity with the HLA-A28 epitope; and ii) inhibition of scFv binding to the antigen by the patients' antibodies. The heavy chain variable genes of all three were derived from the germline gene Cos-3, carry the hallmarks of somatic hypermutation, and are most likely derived from clonally related B cells. The light chain variable domains were encoded by DPK1 and DPK8 from the VkappaI family. These data show that phage display can be used to clone HLA-specific alloantibodies that recognise the native antigen from alloimmunised patients.

Introduction to antibody engineering and phage display.

The vertebrate immune system, capable of rapidly producing highly specific antibodies upon immunisation, has been used to produce murine monoclonal antibodies (mAbs) via immortalisation and isolation of antibody-secreting cells. These mAbs have had a profound impact in the fields of diagnostics and therapeutics. However, the therapeutic use of murine mAbs is complicated by their immunogenicity. To circumvent this immunogenicity, antibody engineering techniques which render murine mAbs more compatible with the human immune system have been devised. Over the last decade, the technique of antibody phage display, which facilitates the production of human mAbs, has been developed. This review serves as an introduction to antibody engineering, phage display and the development of antibody fragments into viable diagnostic and therapeutic reagents.

Rapid phenotyping of HPA-1a using either diabody-based hemagglutination or recombinant IgG1-based assays.

BACKGROUND: The HPA-1 system is carried on the beta3 integrin. HPA-1a (Zw(a), Pl(A1)) is immunogenic in an HPA-1b homozygote (HPA-1b1b). In pregnancy, 1 of 365 women forms anti-HPA-1a, which causes severe thrombocytopenia in 1 in 1100 neonates. Identification of women at risk of forming anti-HPA-1a and the screening of donors to obtain HPA-1a-negative platelets for therapy need reliable, low-cost, automated assays. STUDY DESIGN AND METHODS: A diabody with dual specificity for HPA-1a x D and an IgG1 anti-HPA-1a have been constructed by the use of the genes encoding the first anti-HPA-1a fragment. With these reagents, two complementary HPA-1a phenotyping assays have been developed. RESULTS: This diabody was used in a simple hemagglutination technique to perform HPA-1a phenotyping on soluble glycoprotein IIb/IIIa from EDTA plasma samples. Over 1000 unselected donors have been correctly HPA-1a-phenotyped by use of the diabody. The human recombinant IgG1 anti-HPA-1a was produced in a rat myeloma cell line and was fluorescein labeled for use in a whole-blood flow cytometric HPA-1a phenotyping assay. This IgG1 anti-HPA-1a shows a clear differential between HPA-1a-positive and HPA-1a-negative platelets at nM antibody concentrations. CONCLUSIONS: The two recombinant reagents described are highly suitable for screening and confirmatory HPA-1a phenotyping. They permit rapid determination of the HPA-1a phenotype and are amenable to automation.