A variant Sp1 (R218Q) transcription factor might enhance HbF expression in β0‐thalassaemia homozygotes

Early in 2016, we reported two dizygotic twin Iraqi-American men (aged 21 years old) who were homozygous for codon 8 (−AA) frame-shift (HBB:c.25_26delAA) β0-thalassaemia mutation (Jiang et al, 2016). They were clinically well, with haemoglobin levels of between 123 and 133 g/l, and 98% of their haemoglobin was fetal haemoglobin (HbF, α2γ2). A careful search for single nucleotide polymorphisms (SNPs) known to be associated with elevated HbF, including those within the HBB locus on chromosome 11p15, BCL11A intron 2 on chromosome 2p16 and HBS1L-MYB intergenic polymorphism (HMIP) on chromosome 6q23, was done in the asymptomatic twins and in 22 other homozygotes for the same β0-thalassaemia mutation but with severe anaemia (Jiang et al, 2016). In addition, the KLF1 gene from one of the twins was subjected to nucleotide sequencing but no mutation was found. Our search revealed that both twins were homozygous for a major HbF quantitative trait locus (QTL) HMIP 3-bp deletion at rs66650371, which was not found in any of the other 22 severe patients. Furthermore, both twins were heterozygous for Hph α-thalassaemia mutation, which was also not present in the other 22 patients. These 2 variant genetic motifs, rs66650371 homozygosity and Hph α-thalassaemia mutation heterozygosity, found uniquely in the twins, were likely to be partially responsible for the their mild disease phenotype. To search for additional variant genetic motifs that might account for the twins’ mild phenotype, exome sequencing was undertaken in 1 twin. In the ‘Red Cell Disease Gene’ pipeline, the twin was found to be heterozygous for an SP1 (Chr12q13) exonic non-synonymous G→A single nucleotide variant, SNP rs764833434, resulting in the substitution of arginine with glutamine in codon 218 (R218Q) of the SP1 transcription factor. Subsequent Sanger sequencing confirmed that both twins were heterozygous for rs764833434. SP1 is an ubiquitous transcription factor. It has 2 serine/threonine-rich and 2 glutamine-rich transcription activation domains at the amino-terminal of the peptide. At the carboxyl-terminal, there are three C2H2 zinc-finger DNA-binding motifs, which recognize GC-rich DNA elements, particularly GGGGCGGGGC. Importantly, there are several SP1 binding sites at the HBG proximal promoters (Gumucio et al, 1991). There are also human HBG proximal promoter mutations that could lead to enhanced SP1 binding, resulting in markedly increased HBG expression (Ronchi et al, 1989; Sykes & Kaufman, 1990; Gumucio et al, 1991; Jane et al, 1993). An amplification refractory mutation system (ARMS) test was used to examine other patients’ DNA samples. The twins’ father and sister, both of who were heterozygous for the β0-thalassemia mutation, were also heterozygous for rs764833434. Their mother, who was heterozygous for the β0-thalassemia mutation, did not have this variant. The apparent lack of effect for rs764833434 to significantly up-regulate HbF in the twins’ father and sister might be due to their being heterozygous for the β0-thalassemia mutation, heterozygous for β-globin haplotype I/IV, and heterozygous for rs7482144 alternate allele A (heterozygous for Xmn I). In contrast, the twins are homozygous for all 3 genetic elements, providing a different and ‘permissive’ erythroid genetic milieu for the variant SP1 to further up-regulate their HbF expression. Rs764833434 was not found in 22 other patients from Iraq, Turkey or Morocco who were homozygous for the same β0-thalassaemia mutation. These patients had severe anaemia and most were transfusion dependent (Jiang et al, 2016). The SNP was not found in 125 other patients who were more than 10 years old with HbF >10%, including 41 patients with sickle cell anaemia, 8 with HbS/β-thalassaemia, 25 β-thalassaemia homozygotes or compound heterozygotes, HbC/β-thalassaemia, HbE/β-thalassaemia and other β-globin genotypes. It was not found in 102 Iraqi thalassaemia intermedia patients (Al-Allawi et al, 2015). Among 121,390 total alleles examined in Africans, East and South Asians, Europeans, and Latinos in the Exome Aggregation Consortium (ExAC) dataset, only 3 alleles with rs764833434 were detected (Lek et al, 2016). The arginine residue in SP1 coded by the reference allele of rs764833434 is conserved in humans, Rhesus monkey, mouse, dog, elephant, chicken and Xenopus, indicating that this arginine is an invariant amino acid residue for the normal function of SP1. The missense mutation, R218Q, occurs and adds an additional glutamine residue near the centre of the first 5′ glutamine-rich transcription activation domain of SP1. Additionally, it is adjacent to a potential phosphorylation site (S220) that has been reported to control SP1 activity (Beishline & Azizkhan-Clifford, 2015). Thus the R218Q mutation is likely to impact SP1 function. Other HbF QTL such as rs766432 in BCL11A or KLF1 mutations can up-regulate HbF expression when present in the heterozygous state (Menzel et al, 2007; Borg et al, 2010). Taken together, this rare variant SP1 found in the uniquely asymptomatic twins is likely to play a role in up-regulating HBG expression in patients who are homozygous for codon 8 (−AA) frame-shift (HBB:c.25_26delAA) β0-thalassaemia mutation. This mutation might interfere with the normal SP1 bioactivity via the phosphorylation pathway. Alternatively, the mutation might result in increased transcription activation potential through its glutamine-rich domain. Further experiments are necessary to confirm these hypotheses. ZJ, HL, JF, VS and DA performed the research, analysed the data, reviewed and revised the paper. ZZ, MS, NA, PG, BF and DC designed the study, analysed data and wrote the paper.