A splice site variant defining the novel <i>RHD*01(487‐3G)</i> allele in <i>trans</i> to <i>RHD*DAR1.2</i>

A 53-year-old Mexican American female with rheumatoid arthritis and cirrhosis secondary to autoimmune hepatitis presented with melena and hematemesis, abdominal pain, and weakness. She was found to have multiple varices in the proximal and middle esophagus. The patient was anemic with a hemoglobin of 8.3 g/dL, which was 2 g/dL lower than her baseline. The patient typed as Group A and D typing was weak and variable by Galileo NEO automated testing (see Table 1). The patient was treated as group A, D-negative and transfused with two units of group A, D-negative RBCs. Her sample was investigated to determine the RHD underlying the weak phenotype and to further characterize the sample serologically. 0: LHM174/102, LHM70/45 +w: LHM169/81, LHM59/19, LDM1 3–4+: LHM76/58, LHM76/59, LHM50/2B, ESD1, LHM76/55, LHM77/64 Initial D typing was performed on the Galileo Neo automated instrument using D4 and D5 Anti-D (Immucor, Norcross, GA) (Table 1). Further serological testing was done in tube with multiple Anti-D reagents including BioClone (Ortho Clinical Diagnostics, Raritan, NJ), Seraclone (Bio-Rad, Hercules, CA), Gamma-clone (Immucor, Norcross, GA), and the ALBAclone RhD Variant Investigational kit (Quotient Suisse SA, Eysins, Switzerland). RhCE testing was not routinely performed for patients with weak reactivity to D antigen and a negative antibody screen. Genomic DNA was isolated from white blood cells (buffy coat) (QIAGEN, Valencia, CA). RHD genotyping was performed with RHD Beadchip (Werfen, Warren, NJ), hybrid box assay for RHD zygosity,1 and amplification and Sanger sequencing of RHD exons 1 to 10 and flanking intron regions. To determine cis/trans relationship of detected variants in intron 3 and exon 4, targeted RH next-generation sequencing (NGS) short-read technology was performed on MiSeq Illumina platform (San Diego, CA)2 and analyzed with custom-developed software. SpliceAI, deep learning-based tool,3 was used to evaluate the effect on splicing of an unreported variant. When testing the patient sample by both automated and manual methods, RBCs had variable reactivity (Table 1). The results from the Galileo NEO were weak with D4/D5 Anti-D reagents. Results were variable when tested in tube with different reagents at immediate spin and antihuman globulin phases. By RHD BeadChip, the sample was heterozygous for RHD and RHD*DAR.4 The hybrid box was absent in support of two RHD alleles. As a sample with conventional RHD would not be expected to have a weak D phenotype, it prompted further testing. The pattern of serologic reactivity for the ALBAclone variant RhD kit was not consistent with that listed in the interpretative guide for DAR. Two monoclonal reagents (LHM169/81, LHM59/19) which should have been nonreactive were reactive (+w) (Table 1). By Sanger sequencing, the sample was heterozygous for RHD c.602C>G (p.Thr201Arg), c.667T>G (p.Phe223Val), c.744C>T (p.Ser248=), c.957G>A (p.Val319=), and c.1025T>C (p.Ile342Thr), associated with RHD*DAR1.2 (RHD*09.01.02), and for a variant, c.487-3C>G, located in the intron 3 acceptor splice site region. This variant has not been previously reported in clinical settings but is listed on the genome aggregation database (gnomAD v4.1.0, rs746498767, frequency 0.0001066 in Admixure America and 0.000004354 overall).5 Analysis of the targeted RH NGS reads showed the c.487-3G is in trans to c.602G (associated with RHD*DAR1.2) (Figure 1). The SpliceAI reported an acceptor loss score of 0.89 (with 0 indicating no predicted impact on splicing and 1, a high predicted impact) for c.487-3C>G. A fresh blood sample was not available to perform RhD-cDNA analysis. We identified in this patient two altered RHD alleles, known RHD*DAR1.2 (RHD*09.01.02) encoding a partial D and a novel allele RHD*01(487-3G). The SpliceAI acceptor loss score of 0.89 suggests a strong probability that variant c.487-3C>G results in the of loss of acceptor splice site, potentially resulting in reduced levels of RhD transcripts. RHD*DAR is more common in individuals of African descent6 and here, it was detected in Mexican American woman, highlighting population admixture. Individuals with RHD*DAR1.2 may form alloanti-D and depending on the patient's gender, age, and diagnosis should be regarded as D-negative for transfusion management.6 In this case with a conventional RHD sequence with variant in splice site region, it is difficult to predict whether any conventional RhD protein may be present on the RBCs. Therefore, it was decided to treat her as D-negative. The authors have disclosed no conflicts of interest.