A functional promoter polymorphism of the δ‐globin gene is a specific marker of the Arab‐Indian haplotype

Abstract

Most sickle cell anemia (SCA) patients indigenous to the Eastern Province of Saudi Arabia have their HbS gene on the Arab-Indian (AI) HBB gene cluster haplotype. Their fetal hemoglobin (HbF) levels are near 20% and they have milder disease compared with SCA where the HbS gene is on African origin HBB haplotypes [1-9]. The AI haplotype is characterized by an Xmn1 restriction site at position −158 5′ to HBG2 (rs7482144), a Hinc2 site 5′ to HBE (rs3834466) and other polymorphisms [10]. The causal elements that modify HbF might be in linkage disequilibrium with the βS globin gene in this Saudi population. We first performed homozygosity mapping using genome-wide single nucleotide polymorphisms (SNPs) in AI HbS homozygotes [11, 12] and identified a single large autozygous region including the HBB cluster and surrounding genes. By next generation sequencing, we examined this region in these same individuals and identified several variants that included a SNP in the HBD promoter region at position −68 bp 5′ to HBD (CCAAC > TCAAC). We found this SNP only when the HbS gene was on an AI haplotype and not in SCA with other haplotypes. This SNP was functional in reporter assays in K562 cells and is an AI haplotype-specific marker. Table I summarizes the patient characteristics. Using genome-wide SNP data from a limited number of cases, a region of autozygosity was found only in AI HbS homozygotes on chromosome 11 (coordinates 5,196,450–5,323,071). The region contains HBD, HBG1, HBG2, HBE1, and the Xmn1 5′ HBG2 restriction site (rs7482144). By targeted deep sequencing of 400 kb of chromosome 11 (coordinates 5,143,424–5,543,424; average coverage 42x) in 4 AI patients 1,195 variants were found. A homozygous C-T variant −68 bp 5′ HBD with high genotyping and mapping quality that was not in dbSNP build 135 or 1,000 Genomes, was present. Resequencing of 15.9 kb of chr11 (coordinates 5,253,531–5,269,435) by Sanger sequencing detected three new SNPs of which one was the −68 C > T SNP. We focused on this SNP because of its location within the Corfu deletion region and its location in the HBD promoter. The C > T SNP in the HBD promoter was found only in individuals with the AI haplotype. Saudi sickle cell trait carriers with the AI haplotype were heterozygous for this SNP; while siblings without HbS did not carry this mutation. Among 25 AI HbS-β0 thalassemia patients, 16 were heterozygous at this site (C/T) and 9 were homozygous (T/T). All AI HbS-β0 thalassemia patients who were homozygous T/T were also homozygous for the AI haplotype (Table I). Fifteen African American SCA patients with unusually high HbF, 54 Saudi SCA patients from the Southwestern Province (SW)—mainly Benin but including subjects with the Senegal haplotype—19 SW HbS-β0 thalassemia patients, 16 SW sickle cell trait cases, and 25 normal Saudi controls did not carry the −68 HBD SNP. This SNP was not found in 1,094 individuals in 1,000 Genomes May 2011 release. It is important to note that hemoglobin electrophoresis results in Table I were performed using different methods, so direct comparison of HbF and HbA2 between different groups will not be accurate. In addition, the effect of coinheritance of α-thalassemia, or presence of iron deficiency anemia on Hb A2 level was not assessed. Finally, HbA2 levels are artifactually high when HbS is present because of the co-elution of minor HbS species. For these reasons, it is not possible to estimate the effects of the −68 C-T SNP on these subjects HbA2 levels. Reduced expression of HBD relative to HBB in normal individuals is partly a result of a degenerate CCAAT box in the HBD promoter (CCAAC). The CCAAC motif is the site of the −68 C > T SNP (TCAAC) [13-15]. When we compared the activity of the wild-type HBD promoter with the promoter containing the −68 C > T SNP the variant promoter was associated with a significant decrease in the expression of a reporter construct suggesting that it could further impair already enfeebled HBD expression (Fig. 1). Although HBG is expressed at high levels in K562 cells, endogenous HBD is also expressed [16]. The expression studies were designed solely to test the hypothesis that the −68 C > T SNP downregulates the expression of the HBD promoter. Normal (p-Nδ-luc) and −68 (p-Mδ-luc) HBD promoter luciferase constructs along with the control plasmid pRL-SV40 were electroporated into K562 cells and incubated for 48 hr. Cells were lyzed for luciferase activity assay and RNA isolation. Luciferase activity was normalized for Renilla luciferase activity encoded by pRL-SV40. Mean luciferase activity expressed as relative luciferase units (RLU) is presented (n = 3). The luciferase activity of p-Mδ-luc is significantly reduced compared with p-Nδ-luc (*P < 0.027). The bottom of the figure shows the expression of endogenous γ- and δ- mRNA in K562 cells electroporated with the plasmid constructs. The literature provides further evidence for a functional role of the −68 C > T SNP. Its presence was associated with δ+ thalassemia in one individual with reduced HbA2 of 2% and a slightly increased HbF of 1.3% [13]. Moreover, mutations at positions −30, −31, −36, −55, −65, −76, and −77 in the HBD promoter were reported in HbVar database (http://globin.cse.psu.edu/) to cause δ thalassemia [14, 17-19], and HbF levels of 3.3–4.7% have been noted in some hematologically normal individuals with homozygous δ0 thalassemia [19, 20]. A mechanism for increased HbF in the presence of less common HBD promoter mutations is unknown. Any role for the −68 C-T SNP as a modifier of HbF in AI haplotype HbS sickle cell disease is unknown. Perhaps HBD promoter SNPs reduce the interaction of the locus control region and the transcription apparatus with this promoter permitting enhanced interactions with HBG promoters [21]. The paradox of the Corfu deletion first suggested the potential of the HBD-HBG1 intergenic area, the site of the −68 C-T SNP, as a silencer of HBG expression [22]. One potential functional area is the polypyrimidine (PYR) binding site about 960 bp upstream of HBD; however, polymorphisms of this site were not present in AI haplotype carriers [23, 24]. Another candidate is a 3.5 kb region near the 5′ portion of HBD that binds the HBG silencer, BCL11A and also GATA-1 and HDAC1. Deletion of this region delayed HBG to HBB switching [25]. One or more HBG regulatory regions linked to the HBB gene-like cluster, perhaps with transacting elements and expanded erythropoiesis due to hemolysis, are likely to account for high HbF in the AI haplotype. It is not known if the −68 C-T SNP effects HbF level. Five families from the Eastern Province included seven cases with AI SCA, seven individuals with sickle cell trait and the AI haplotype, and two individuals with normal hemoglobin. Four HbS homozygotes from these families were used for autozygosity mapping following genome-wide SNP analysis using the Illumina Human610-Quad SNP array and were analyzed using PLINK [26]. Follow-up studies were done in 46 AI SCA patients, 25 HbS-β0 thalassemia patients where the HbS gene was associated with the AI haplotype, 25 normal Saudi subjects, 54 SW SCA patients where the HbS gene was on the Benin, Bantu, or Senegal haplotype, 19 SW HbS-β0 thalassemia patients, 16 SW sickle cell trait individuals, and 15 African American SCA patients with HbF > 10% with different HBB haplotypes. Age, gender, CBC, and hemoglobin fractionation by high-performance liquid chromatography or capillary electrophoresis were determined in all cases. DNA was extracted from peripheral blood and the HBB gene cluster haplotype was determined using restriction fragment length polymorphisms as previously described [27]. The study was approved by the IRB of participating institutions. A custom DNA enrichment array (Agilent, Santa Clara, CA) was designed for 400 kb of chr11 (5,143,424–5,543,424 build GRCh37). Sequencing in four AI haplotype HbS homozygotes who were the subject of autozygosity mapping was performed on Illumina GA-II instrument with single-end reads (Illumina, San Diego, CA). These regions were also sequenced by Sanger sequencing using primer sets shown in Supporting Information Table S1. We sequenced 15.9 kb between chr11 (5,253,531–5,269,435 build GRCh37; 41,052–56,923 based on HBB record U01317) in seven cases with AI SCA, seven individuals with sickle cell trait and heterozygous for the AI haplotype and two normal Saudi individuals. SNPs within this region were considered unique to the AI haplotype if SCA cases were homozygous, HbS trait subjects were heterozygous, and normal subjects did not have these SNPs (Supporting Information Table S2). DNA from normal subjects and AI haplotype HbS homozygotes was used to amplify HBD promoters using specific primers (Supporting Information Table S1) generating a fragment of 826 bp (−772/+54) that was cloned into p-GEM-T Easy vector (Promega, Madison, WI). The amplified HBD promoter fragments were released with Not1 digestion, end-filled and cloned into basic pGL3 luciferase vector (Promega) at the Sma1 site in the multicloning site [28]. The promoter orientation and −68 C > T polymorphism were confirmed by sequencing. Plasmid DNA was prepared and introduced into K562 cells. Cells were incubated for 48 hr and lyzed for luciferase assay and RNA isolation [29]. Luciferase gene expression was measured using a dual-luciferase reporter assay system and a Promega single tube luminometer. The cDNA was used in PCR reaction using specific primers (Supporting Information Table S1) for HBG to generate a fragment of 239 bp and for HBD to generate a fragment of 190 bp. The PCR products were visualized on a 1.5% agarose gel. Additional Supporting Information may be found in the online version of this article. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article. Abdulrahman Alsultan*, Duyen A. Ngo , John J. Farrell §, Idowu Akinsheye , Nadia Solovieff , Hazem A. Ghabbour*, Amein Al-Ali¶, Ahmed Alsuliman**, Muneer Al-Baghshi , Waleed Albu-Ali , Mohammed Alabdulaali**, Clinton T. Baldwin , Lindsay A. Farrer §, Hong Luo , Efthymia Melista , Surinder Safaya , Maxwell Nwaru Jr. , David H.K. Chui , Martin H. Steinberg , * Sickle Cell Disease Research Center and Department of Pediatrics, College of Medicine, King Saud University, Riyadh, Saudi Arabia, Center of Excellence in Sickle Cell Disease, Department of Medicine, Boston University School of Medicine, Boston, Massachusetts, Division of Genetics, Department of Medicine, Boston University School of Medicine, Boston, MA, § Department of Biochemistry, College of Medicine, University of Dammam, Dammam, Saudi Arabia, ¶ Department of Medicine, King Fahad Hospital, Hofuf, Saudi Arabia, ** HematologyDivision, Maternity and Children hospital, Al-Ahsa, Saudi Arabia, Department of Pediatrics, College of Medicine, King Faisal University, Al-Ahsa, Saudi Arabia, Center for Human Genetics, Boston University School of Medicine, Boston, Massachusetts