Antigens of the Lutheran blood group system on human erythrocytes are carried on two isoforms of the Lutheran glycoprotein [Lutheran (LU) and B-cell adhesion molecule (B-CAM)]. In rare cases, Lutheran blood group antigens are not expressed on erythrocytes, a condition known as the Lu(a-b-) phenotype. This can be caused by homozygosity for recessive alleles at the BCAM (LU) locus (Karamatic Crew et al, 2007), the dominant In(Lu) phenotype resulting from heterozygosity for inactivating mutations in the erythroid transcription factor gene KLF1 (Singleton et al, 2008) or inheritance of an unidentified X-linked suppressor gene (termed XS2, Norman et al, 1986). Inheritance of XS2 has been reported in only one family. Affected members in this family had normal red cell indices and occasional macrothrombocytes in the peripheral circulation (Norman et al, 1986). We hypothesized that the X-linked type of Lu(a-b-) might be due to a mutation in a transcription factor. GATA1 (GATA binding protein 1) was a likely candidate as it is essential for both erythroid and megakaryocyte differentiation (Pevny et al, 1991; Shivdasani et al, 1997) and the gene is located on the X-chromosome (Xp11·23). Sequencing of the GATA1 gene in the male proband of the XS2 family revealed a missense mutation in the termination codon, which is the likely cause of the erythrocyte Lu(a-b-) phenotype and mild platelet disorder evidenced by thrombocytopenia and occasional macrothrombocytes.
The male Lu(a-b-) proband reported by Norman et al (1986) is now 67 years old, is not anaemic but is thrombocytopenic and has occasional macrothrombocytes (diameter 4–5 μm) and a history of bruising easily (Table SI). Serological examination of his red cells revealed suppression of Lutheran blood group antigens (Lu2, Lu3, Lu6, Lu8, Lu18 and Lu21) and weak P1 expression. Expression of AnWj, Inb and MER2 antigens was normal and i activity was enhanced (data not shown). These results are consistent with the original findings (Norman et al, 1986). Flow cytometric analyses using monoclonal antibodies confirmed the extremely low levels of Lutheran blood group glycoprotein and the presence of normal levels of CD44 on the proband's erythrocytes (Fig 1A, Methods S1), indicative of Lu(a-b-) red cells of the X-linked and not the In(Lu) type. Other markers (GPA, GPC, CD59, Fy3 and Coa) were also normal. Red cells of the proband's sister showed normal expression of Lutheran and other antigens (data not shown).
Genomic DNA was extracted from a blood sample from the male XS2 proband. The five coding exons of GATA1 were amplified by polymerase chain reaction (PCR) and sequenced (Table SII, Methods S1). A hemizygous 1240T>C mutation was found in the termination codon, converting TGA to a codon for arginine (CGA; Fig 1B). The mutation predicts a translated GATA1 protein containing an additional 41 amino acids at the carboxy terminus (X414R; Fig 1C, D). As the mutation destroys a BspHI restriction enzyme site, restriction fragment length polymorphism-PCR was used to screen other samples. No mutation was found in DNA extracted from blood obtained from the proband's unaffected sister, or from 78 random blood donors and 24 In(Lu) samples (data not shown), consistent with the mutation causing the X-linked Lu(a-b-) phenotype.
Several human phenotypes have been associated with mutations in the GATA1 gene, including anaemia, dyserythropoiesis, abnormal platelets, thalassaemia and congenital eythropoietic porphyria (Ciovacco et al, 2008). These result from missense mutations in the N-terminal zinc finger domain and most interfere with the binding of GATA1 to its partner protein FOG1 (Friend of GATA1). Mutations resulting in the use of an internal Methionine and producing an N-terminally truncated form of GATA1 (GATA1s) are associated with Trisomy 21 (Fig 1D, Ciovacco et al, 2008). We have identified a novel mutation in the GATA1 gene of an individual with the rare X-linked form of Lu(a-b-) blood group phenotype. This is the first inherited missense mutation that occurs outside of the zinc finger domains and suggests a functional role for the C-terminus. Consistent with this, a recent study in mice has identified a second activation domain at the C-terminus of GATA1 protein (Kaneko et al, 2012).
The ultrastructure of platelets from individuals with different GATA1 mutations has been extensively studied and many structural defects identified (White & Thomas, 2009). Macrothrombocytes in the XS2 proband's peripheral blood were therefore examined by electron microscopy, revealing several unusual features. Most macrothrombocytes contained areas of flattened cisternae, similar to smooth endoplasmic reticulum (Fig 2A), reminiscent of the tubular membrane sheets previously described in G208S GATA1-mutant platelets (White et al, 2007). Numerous Golgi-like saccules and vesicles (Fig 2B) and multi-vesicular bodies were also present (Fig 2C) and many thrombocytes contained an area of densely packed fine vesicles (fine interwoven cisternae; Fig 2D).
Our results show that a rare blood group phenotype is associated with a novel mutation in GATA1 and that mutations outside of the zinc finger domains can alter GATA1 functions in erythroid and megakaryocyte differentiation. The X414R mutation has mild effects on erythropoiesis but results in marked reduction in expression of the Lutheran glycoprotein. Lutheran is expressed at the erythroid cell surface very late in erythropoiesis (Southcott et al, 1999) and its absence or gross reduction on erythrocytes appears to be a sensitive indicator of dyserythropoiesis, as evidenced by the present study and our previous work demonstrating haploinsufficiency for KLF1 causing the In(Lu) type of Lu(a-b-) phenotype (Singleton et al, 2008).
The case described here is unique in that the GATA1 mutation was revealed through identification of the rare Lu(a-b-) phenotype rather than disease. Previous investigators describing mutations in GATA1 in association with thrombocytopenia and dyserythropoiesis have not examined their patients' erythrocytes for expression of Lutheran glycoprotein. We would expect the erythrocytes of some or all of the patients from earlier studies to have Lutheran glycoprotein deficiency. Consequently, we suggest a simple serological or flow cytometric screening test to determine the level of expression of the Lutheran glycoprotein on erythrocytes. This would provide a useful additional laboratory tool suggesting the presence of novel mutations in erythroid transcription factors in individuals with uncharacterized dyserythropoiesis and thrombocytopenia.