Epigenetic dissection of human blood group genes reveals regulatory elements and detailed characteristics of KEL and four other loci.

BACKGROUND: Blood typing is essential for safe transfusions and is performed serologically or genetically. Genotyping predominantly focuses on coding regions, but non-coding variants may affect gene regulation, as demonstrated in the ABO, FY and XG systems. To uncover regulatory loci, we expanded a recently developed bioinformatics pipeline for discovery of non-coding variants by including additional epigenetic datasets. METHODS: Multiple datasets including ChIP-seq with erythroid transcription factors (TFs), histone modifications (H3K27ac, H3K4me1), and chromatin accessibility (ATAC-seq) were analyzed. Candidate regulatory regions were investigated for activity (luciferase assays) and TF binding (electrophoretic mobility shift assay, EMSA, and mass spectrometry, MS). RESULTS: In total, 814 potential regulatory sites in 47 blood-group-related genes were identified where one or more erythroid TFs bound. Enhancer candidates in CR1, EMP3, ABCB6, and ABCC4 indicated by ATAC-seq, histone markers, and co-occupancy of 4 TFs (GATA1/KLF1/RUNX1/NFE2) were investigated but only CR1 and ABCC4 showed increased transcription. Co-occupancy of GATA1 and KLF1 was observed in the KEL promoter, previously reported to contain GATA1 and Sp1 sites. TF binding energy scores decreased when three naturally occurring variants were introduced into GATA1 and KLF1 motifs. Two of three GATA1 sites and the KLF1 site were confirmed functionally. EMSA and MS demonstrated increased GATA1 and KLF1 binding to the wild-type compared to variant motifs. DISCUSSION: This combined bioinformatics and experimental approach revealed multiple candidate regulatory regions and predicted TF co-occupancy sites. The KEL promoter was characterized in detail, indicating that two adjacent GATA1 and KLF1 motifs are most crucial for transcription.

Feasibility for non-invasive prenatal fetal blood group and platelet genotyping by massively parallel sequencing: A single test system for multiple atypical red cell, platelet and quality control markers.

Non-invasive prenatal tests (NIPT) to predict fetal red cell or platelet antigen status for alloimmunised women are provided for select antigens. This study reports on massively parallel sequencing (MPS) using a red cell and platelet probe panel targeting multiple nucleotide variants, plus individual identification single nucleotide polymorphisms (IISNPs). Maternal blood samples were provided from 33 alloimmunised cases, including seven with two red cell antibodies. Cell-free and genomic DNA was sequenced using targeted MPS and bioinformatically analysed using low-frequency variant detection. The resulting maternal genomic DNA allele frequency was subtracted from the cell-free DNA counterpart. Outcomes were matched against validated phenotyping/genotyping methods, where available. A 2.5% subtractive allele frequency threshold was set after comparing MPS predictions for K, RhC/c, RhE/e and Fya /Fyb against expected outcomes. This threshold was used for subsequent predictions, including HPA-15a, Jka /Jkb , Kpa /Kpb and Lua . MPS outcomes were 97.2% concordant with validated methods; one RhC case was discordantly negative and lacked IISNPs. IISNPs were informative for 30/33 cases as controls. NIPT MPS is feasible for fetal blood group genotyping and covers multiple blood groups and control targets in a single test. Noting caution for the Rh system, this has the potential to provide a personalised service for alloimmunised women.

Aggregates in blood filter chambers used from the plasma donations of anti-D donors: evaluation for monoclonal antibody discovery using phage display.

BACKGROUND: RhD-immunoglobulin (RhIg) prevents anti-D alloimmunisation in D-negative pregnant women when the fetus is D-positive, reducing the incidence of haemolytic disease of the fetus and newborn. Manufacturing RhIg is reliant on the limited supply of plasma donations with anti-D antibodies. Monoclonal antibody (mAb) development platforms such as phage display, require blood samples to be collected from anti-D donors, which may be a complicated process. The blood filter chamber (BFC) discarded after an anti-D donor's donation might provide a source of Ig-encoding RNA. This study aims to evaluate whether used BFCs are a suitable source of Ig-encoding RNA for phage display. MATERIAL AND METHODS: Haemonetics PCS2 BFCs were obtained from 10 anti-D donors for total RNA extraction, cDNA synthesis and amplification of VH and VL IgG sequences for assembly of single-chain variable fragments (scFvs). A scFv-phage display library was constructed and 3 rounds of biopanning were performed using D-positive and D-negative red blood cells (RBCs). Positive phage clones were isolated, Sanger sequenced and, where possible, reformatted into full-length human IgGs to define specificity. The BFC aggregates from 2 anti-D donors underwent a Wright-Giemsa stain and hematological cell count. RESULTS: Of 10 BFCs, a sufficient yield of total RNA for library construction was obtained from BFCs containing cellular aggregates (n=5). Aggregate analysis showed lymphocytes were the cellular source of Ig-encoding RNA. From the 5 samples with aggregates, scFvs were assembled from amplified IgG variable regions. The library constructed from 1 of these samples resulted in the isolation of clones binding to D-positive RBCs with IGHV3 gene usage. Of the 4 reformatted IgG, 3 were anti-D and 1 had undefined specificity. DISCUSSION: BFC aggregates are a new and convenient source of Ig-encoding RNA which can be used to construct Ig gene libraries for mAb isolation and discovery via antibody phage display.

Severe hemolytic disease of the fetus and newborn due to allo-anti-D in a patient with a partial DEL phenotype arising from the variant allele described as RHD*148+1T (RHD*01EL.31).

BACKGROUND: RhD DEL variants may show complete or partial expression of RhD epitopes. There have been only rare reports of anti-D causing hemolytic disease of the fetus and newborn (HDFN) in this context. We report a case of severe HDFN associated with a recently described DEL variant. CASE REPORT: A multiparous woman presented with an allo-anti-D and showed incongruent phenotyping and genotyping results on initial study. Further investigations identified the RHD mutation, defined as RHD*148+1T and named RHD*01EL.31, which had been previously associated with a DEL phenotype. Extended RhD phenotyping by adsorption-elution showed that there was reactivity with four of nine monoclonal anti-D antibodies, suggesting a partial DEL phenotype. The first child showed no clinical evidence of HDFN, although the cord direct antiglobulin test was positive. The second child developed fetal anemia treated with intrauterine transfusion, and neonatal hyperbilirubinemia requiring exchange transfusion. CONCLUSION: The RHD allele, RHD*148+1T, results in a partial Del phenotype, and the anti-D formed in pregnant women with this phenotype is capable of causing severe HDFN.

A DEL phenotype attributed to RHD Exon 9 sequence deletion: slipped-strand mispairing and blood group polymorphisms.

BACKGROUND: The RhD blood group antigen is extremely polymorphic and the DEL phenotype represents one such class of polymorphisms. The DEL phenotype prevalent in East Asian populations arises from a synonymous substitution defined as RHD*1227A. However, initially, based on genomic and cDNA studies, the genetic basis for a DEL phenotype in Taiwan was attributed to a deletion of RHD Exon 9 that was never verified at the genomic level by any other independent group. Here we investigate the genetic basis for a Caucasian donor with a DEL partial D phenotype and compare the genomic findings to those initial molecular studies. STUDY DESIGN AND METHODS: The 3'-region of the RHD gene was amplified by long-range polymerase chain reaction (PCR) for massively parallel sequencing. Primers were designed to encompass a deletion, flanking Exon 9, by standard PCR for Sanger sequencing. Targeted sequencing of exons and flanking introns was also performed. RESULTS: Genomic DNA exhibited a 1012-bp deletion spanning from Intron 8, across Exon 9 into Intron 9. The deletion breakpoints occurred between two 25-bp repeat motifs flanking Exon 9 such that one repeat sequence remained. CONCLUSION: Deletion mutations bordered by repeat sequences are a hallmark of slipped-strand mispairing (SSM) event. We propose this genetic mechanism generated the germline deletion in the Caucasian donor. Extensive studies show that the RHD*1227A is the most prevalent DEL allele in East Asian populations and may have confounded the initial molecular studies. Review of the literature revealed that the SSM model explains some of the extreme polymorphisms observed in the clinically significant RhD blood group antigen.

Targeted exome sequencing defines novel and rare variants in complex blood group serology cases for a red blood cell reference laboratory setting.

BACKGROUND: We previously demonstrated that targeted exome sequencing accurately defined blood group genotypes for reference panel samples characterized by serology and single-nucleotide polymorphism (SNP) genotyping. Here we investigate the application of this approach to resolve problematic serology and SNP-typing cases. STUDY DESIGN AND METHODS: The TruSight One sequencing panel and MiSeq platform was used for sequencing. CLC Genomics Workbench software was used for data analysis of the blood group genes implicated in the serology and SNP-typing problem. Sequence variants were compared to public databases listing blood group alleles. The effect of predicted amino acid changes on protein function for novel alleles was assessed using SIFT and PolyPhen-2. RESULTS: Among 29 unresolved samples, sequencing defined SNPs in blood group genes consistent with serologic observation: 22 samples exhibited SNPs associated with varied but known blood group alleles and one sample exhibited a chimeric RH genotype. Three samples showed novel variants in the CROM, LAN, and RH systems, respectively, predicting respective amino acid changes with possible deleterious impact. Two samples harbored rare variants in the RH and FY systems, respectively, not previously associated with a blood group allele or phenotype. A final sample comprised a rare variant within the KLF1 transcription factor gene that may modulate DNA-binding activity. CONCLUSION: Targeted exome sequencing resolved complex serology problems and defined both novel blood group alleles (CD55:c.203G>A, ABCB6:c.1118_1124delCGGATCG, ABCB6:c.1656-1G>A, and RHD:c.452G>A) and rare variants on blood group alleles associated with altered phenotypes. This study illustrates the utility of exome sequencing, in conjunction with serology, as an alternative approach to resolve complex cases.

Non-invasive fetal RHD genotyping for RhD negative women stratified into RHD gene deletion or variant groups: comparative accuracy using two blood collection tube types.

Non-invasive fetal RHD genotyping in Australia to reduce anti-D usage will need to accommodate both prolonged sample transport times and a diverse population demographic harbouring a range of RHD blood group gene variants. We compared RHD genotyping accuracy using two blood sample collection tube types for RhD negative women stratified into deleted RHD gene haplotype and RHD gene variant cohorts. Maternal blood samples were collected into EDTA and cell-free (cf)DNA stabilising (BCT) tubes from two sites, one interstate. Automated DNA extraction and polymerase chain reaction (PCR) were used to amplify RHD exons 5 and 10 and CCR5. Automated analysis flagged maternal RHD variants, which were classified by genotyping. Time between sample collection and processing ranged from 2.9 to 187.5 hours. cfDNA levels increased with time for EDTA (range 0.03-138 ng/µL) but not BCT samples (0.01-3.24 ng/µL). For the 'deleted' cohort (n=647) all fetal RHD genotyping outcomes were concordant, excepting for one unexplained false negative EDTA sample. Matched against cord RhD serology, negative predictive values using BCT and EDTA tubes were 100% and 99.6%, respectively. Positive predictive values were 99.7% for both types. Overall 37.2% of subjects carried an RhD negative baby. The 'variant' cohort (n=15) included one novel RHD and eight hybrid or African pseudogene variants. Review for fetal RHD specific signals, based on one exon, showed three EDTA samples discordant to BCT, attributed to high maternal cfDNA levels arising from prolonged transport times. For the deleted haplotype cohort, fetal RHD genotyping accuracy was comparable for samples collected in EDTA and BCT tubes despite higher cfDNA levels in the EDTA tubes. Capacity to predict fetal RHD genotype for maternal carriers of hybrid or pseudogene RHD variants requires stringent control of cfDNA levels. We conclude that fetal RHD genotyping is feasible in the Australian environment to avoid unnecessary anti-D immunoglobulin prophylaxis.

Evaluation of targeted exome sequencing for 28 protein-based blood group systems, including the homologous gene systems, for blood group genotyping.

BACKGROUND: Blood group single nucleotide polymorphism genotyping probes for a limited range of polymorphisms. This study investigated whether massively parallel sequencing (also known as next-generation sequencing), with a targeted exome strategy, provides an extended blood group genotype and the extent to which massively parallel sequencing correctly genotypes in homologous gene systems, such as RH and MNS. STUDY DESIGN AND METHODS: Donor samples (n = 28) that were extensively phenotyped and genotyped using single nucleotide polymorphism typing, were analyzed using the TruSight One Sequencing Panel and MiSeq platform. Genes for 28 protein-based blood group systems, GATA1, and KLF1 were analyzed. Copy number variation analysis was used to characterize complex structural variants in the GYPC and RH systems. RESULTS: The average sequencing depth per target region was 66.2 ± 39.8. Each sample harbored on average 43 ± 9 variants, of which 10 ± 3 were used for genotyping. For the 28 samples, massively parallel sequencing variant sequences correctly matched expected sequences based on single nucleotide polymorphism genotyping data. Copy number variation analysis defined the Rh C/c alleles and complex RHD hybrids. Hybrid RHD*D-CE-D variants were correctly identified, but copy number variation analysis did not confidently distinguish between D and CE exon deletion versus rearrangement. CONCLUSION: The targeted exome sequencing strategy employed extended the range of blood group genotypes detected compared with single nucleotide polymorphism typing. This single-test format included detection of complex MNS hybrid cases and, with copy number variation analysis, defined RH hybrid genes along with the RHCE*C allele hitherto difficult to resolve by variant detection. The approach is economical compared with whole-genome sequencing and is suitable for a red blood cell reference laboratory setting.

A D+ blood donor with a novel RHD*D-CE(5-6)-D gene variant exhibits the low-frequency antigen RH23 (D(W) ) characteristic of the partial DVa phenotype.

BACKGROUND: Blood donors whose red blood cells (RBCs) exhibit a partial RhD phenotype, lacking some D epitopes, present as D+ in routine screening. Such phenotypes can exhibit low-frequency antigens (LFAs) of clinical significance. The aim of this study was to describe the serologic and genetic profile for a blood donor with an apparent D+ phenotype carrying a variant RHD gene where D Exons 5 and 6 are replaced by RHCE Exon (5-6). STUDY DESIGN AND METHODS: Anti-D monoclonal antibodies were used to characterize the presentation of RhD epitopes on the RBCs. RHD exon scanning and DNA sequencing of short- and long-range polymerase chain reaction amplicons were used to determine the RHD structure and sequence. Extended phenotyping for LFAs RH23 (D(W) ) and Rh32 was performed. RESULTS: The donor serology profile was consistent with partial RhD epitope presentation. The donor was hemizygous for an RHD variant allele described as RHD*D-CE(5-6)-D hybrid. The RHCE gene insert is at least 3.868 kb with 5' and 3' breakpoints between IVS4 + 132-c.667 and IVS6 + 1960-IVS6 + 2099, respectively. The sequence for this hybrid was assigned GenBank Accession Number KT099190.2. The RBCs were RH23 (D(W) )+ and Rh32-. CONCLUSION: A novel RHD*D-CE(5-6)-D hybrid allele encodes a partial RhD epitope and carries the LFA RH23 (D(W) ). This and the epitope profile resemble the partial DVa phenotype. Given that RBCs from this individual lack some RhD epitopes, there is an alloimmunization risk if the donor is exposed to D+ RBCs. Conversely, transfusions of RH23 (D(W) )+ cells to RH23 (D(W) )- recipients also pose an alloimmunization risk.