Longer-term efficiency and safety of increasing the frequency of whole blood donation (INTERVAL): extension study of a randomised trial of 20 757 blood donors.

BACKGROUND: The INTERVAL trial showed that, over a 2-year period, inter-donation intervals for whole blood donation can be safely reduced to meet blood shortages. We extended the INTERVAL trial for a further 2 years to evaluate the longer-term risks and benefits of varying inter-donation intervals, and to compare routine versus more intensive reminders to help donors keep appointments. METHODS: The INTERVAL trial was a parallel group, pragmatic, randomised trial that recruited blood donors aged 18 years or older from 25 static donor centres of NHS Blood and Transplant across England, UK. Here we report on the prespecified analyses after 4 years of follow-up. Participants were whole blood donors who agreed to continue trial participation on their originally allocated inter-donation intervals (men: 12, 10, and 8 weeks; women: 16, 14, and 12 weeks). They were further block-randomised (1:1) to routine versus more intensive reminders using computer-generated random sequences. The prespecified primary outcome was units of blood collected per year analysed in the intention-to-treat population. Secondary outcomes related to safety were quality of life, self-reported symptoms potentially related to donation, haemoglobin and ferritin concentrations, and deferrals because of low haemoglobin and other factors. This trial is registered with ISRCTN, number ISRCTN24760606, and has completed. FINDINGS: Between Oct 19, 2014, and May 3, 2016, 20 757 of the 38 035 invited blood donors (10 843 [58%] men, 9914 [51%] women) participated in the extension study. 10 378 (50%) were randomly assigned to routine reminders and 10 379 (50%) were randomly assigned to more intensive reminders. Median follow-up was 1·1 years (IQR 0·7-1·3). Compared with routine reminders, more intensive reminders increased blood collection by a mean of 0·11 units per year (95% CI 0·04-0·17; p=0·0003) in men and 0·06 units per year (0·01-0·11; p=0·0094) in women. During the extension study, each week shorter inter-donation interval increased blood collection by a mean of 0·23 units per year (0·21-0·25) in men and 0·14 units per year (0·12-0·15) in women (both p<0·0001). More frequent donation resulted in more deferrals for low haemoglobin (odds ratio per week shorter inter-donation interval 1·19 [95% CI 1·15-1·22] in men and 1·10 [1·06-1·14] in women), and lower mean haemoglobin (difference per week shorter inter-donation interval -0·84 g/L [95% CI -0·99 to -0·70] in men and -0·45 g/L [-0·59 to -0·31] in women) and ferritin concentrations (percentage difference per week shorter inter-donation interval -6·5% [95% CI -7·6 to -5·5] in men and -5·3% [-6·5 to -4·2] in women; all p<0·0001). No differences were observed in quality of life, serious adverse events, or self-reported symptoms (p>0.0001 for tests of linear trend by inter-donation intervals) other than a higher reported frequency of doctor-diagnosed low iron concentrations and prescription of iron supplements in men (p<0·0001). INTERPRETATION: During a period of up to 4 years, shorter inter-donation intervals and more intensive reminders resulted in more blood being collected without a detectable effect on donors' mental and physical wellbeing. However, donors had decreased haemoglobin concentrations and more self-reported symptoms compared with the initial 2 years of the trial. Our findings suggest that blood collection services could safely use shorter donation intervals and more intensive reminders to meet shortages, for donors who maintain adequate haemoglobin concentrations and iron stores. FUNDING: NHS Blood and Transplant, UK National Institute for Health Research, UK Medical Research Council, and British Heart Foundation.

Silencing of RhoA nucleotide exchange factor, ARHGEF3, reveals its unexpected role in iron uptake.

Genomewide association meta-analysis studies have identified > 100 independent genetic loci associated with blood cell indices, including volume and count of platelets and erythrocytes. Although several of these loci encode known regulators of hematopoiesis, the mechanism by which most sequence variants exert their effect on blood cell formation remains elusive. An example is the Rho guanine nucleotide exchange factor, ARHGEF3, which was previously implicated by genomewide association meta-analysis studies in bone cell biology. Here, we report on the unexpected role of ARHGEF3 in regulation of iron uptake and erythroid cell maturation. Although early erythroid differentiation progressed normally, silencing of arhgef3 in Danio rerio resulted in microcytic and hypochromic anemia. This was rescued by intracellular supplementation of iron, showing that arhgef3-depleted erythroid cells are fully capable of hemoglobinization. Disruption of the arhgef3 target, RhoA, also produced severe anemia, which was, again, corrected by iron injection. Moreover, silencing of ARHGEF3 in erythromyeloblastoid cells K562 showed that the uptake of transferrin was severely impaired. Taken together, this is the first study to provide evidence for ARHGEF3 being a regulator of transferrin uptake in erythroid cells, through activation of RHOA.

Identification of the primary collagen-binding surface on human glycoprotein VI by site-directed mutagenesis and by a blocking phage antibody.

Glycoprotein (GP) VI is the major receptor responsible for platelet activation by collagen, but the collagen-binding surface of GPVI is unknown. To address this issue we expressed, from insect cells, the immunoglobulin (Ig)-like ectodomains (residues 1-185) of human and murine GPVI, called hD1D2 and mD1D2, respectively. Both proteins bound specifically to collagen-related peptide (CRP), a GPVI-specific ligand, but hD1D2 bound CRP more strongly than did mD1D2. Molecular modeling and sequence comparison identified key differences between hD1D2 and mD1D2. Ten mutant hD1D2s were expressed, of which 4 had human residues replaced by their murine counterpart, and 6 had replacements by alanine. CRP binding studies with these mutants demonstrated that the exchange of lysine at position 59 for the corresponding murine glutamate substantially reduced binding to CRP. The position of lysine59 on the apical surface of GPVI suggests a mode of CRP binding analogous to that used by the related killer cell Ig-like receptors to bind HLA. This surface was confirmed as critical for collagen binding by epitope mapping of an inhibitory phage antibody against GPVI. This anti-GPVI, clone 10B12, gave dose-dependent inhibition of the hD1D2-collagen interaction. Clone 10B12 inhibited activation of platelets by CRP and collagen in aggregometry and thrombus formation by the latter in whole blood perfusion. Antibody 10B12 showed significantly reduced binding to the hD1D2-E59, and, on that basis, the GPVI:10B12 interface was modeled.

A tyrosine703serine polymorphism of CD109 defines the Gov platelet alloantigens.

The biallelic platelet-specific Gov antigen system-implicated in refractoriness to platelet transfusion, neonatal alloimmune thrombocytopenia, and posttransfusion purpura-is carried by the glycosylphosphatidylinositol (GPI)-linked protein CD109. The recent identification of the human CD109 complementary DNA (cDNA) has allowed the molecular nature of the Gov alleles to be elucidated. By using reverse transcriptase-polymerase chain reaction (RT-PCR) to amplify CD109 cDNAs from 6 phenotypically homozygous Gov(aa) and Gov(bb) individuals, we have determined that the Gov alleles differ by an A to C single nucleotide polymorphism (SNP) at position 2108 of the coding region, resulting in a Tyr/Ser substitution at CD109 amino acid 703. Allele-specific PCR sequence-specific primers (SSP), PCR-restriction fragment length polymorphism, and real-time PCR studies of 15 additional donors (5 Gov(aa), 5 Gov(bb), and 5 Gov(ab)) confirmed that this SNP correlates with the Gov phenotype. In addition, Chinese hamster ovary cells transiently expressing nucleotide 2108 A>C CD109 cDNA variants were recognized specifically by allele-specific Gov antisera, indicating that this polymorphism defines the Gov alloantigenic determinants. Real-time PCR was then used to genotype 85 additional Gov phenotyped donors. In all but 3 cases, genomic testing concurred with the Gov phenotype. Repeat testing corrected 2 of these discrepancies in favor of the genotyping result. The third discrepancy could not be resolved, likely reflecting low-level CD109 expression below the sensitivity of the phenotyping assay. We conclude that the Gov alleles are defined by a 2108 A>C SNP that results in a Tyr703Ser substitution of CD109 and that genotyping studies are more accurate for Gov alloantigen determination than are conventional serologic methods.