Abstract 1627

In the Eastern Province of Saudi Arabia, sickle cell anemia (HbSS) is associated with the Saudi Indian (SI) HBB-gene cluster haplotype, high levels of fetal hemoglobin (HbF) and milder disease, when compared with Southwestern Province HbSS patients who have lower HbF levels and different HBB haplotypes. An association between HbF and the Xmn1 restriction site in the HBG2 promoter present in both the SI and African-derived Senegal haplotypes is well known, but the causal elements of this association are unknown. Moreover, among individuals with the SI haplotype, only HbSS patients have high HbF while individuals with sickle cell trait (HbAS) or normal hemoglobin (HbAA) do not. Furthermore, HbF levels are far higher in SI haplotype patients, as shown below, compared with Senegal haplotype homozygotes. For example African patients homozygous for the Senegal haplotype had 12.3±5.3% HbF. To better understand the genetic basis for high HbF in SI haplotype HbSS cases, we compared sequences in the HBB gene cluster in patients with SI and Senegal haplotypes. We hypothesized that the causal elements that modify HbF in Saudi patients are in linkage disequilibrium (LD) with the βS globin gene in this population. Accordingly, we studied 5 Saudi families from the Eastern Province. Seven SI haplotype patients with HbSS (median age 5 yrs, range 2.5–49 yrs) were homozygous for the Xmn I site and had Hb 9.7 ± 1.6 g/dL, MCV 76.5 ± 8.3 fl and median Hb F 30.3 (range 18–41). Seventeen SI haplotype individuals had HbAS (median HbF 1.2, range 0–4.2); and 2 were normal. We first determined the genotypes of 3 known HbF QTLs, BCL11A (rs766432); HBS1L-MYB (rs7775698 and rs9399137); and OR51B5/6 (rs5006884). There were no consistent genotypes among these 7 patients to explain their universal high HbF. Next, we performed homozygosity mapping using Illumina Human610-Quad SNP array and identified runs of homozygosity (RoH) of variable length (from 160 kb to nearly 2 mb) within and surrounding the HBB cluster only in HbSS patients. RoH were absent elsewhere in the genome in HbSS. The RoH that was shared by all HbSS patients was 126.6 kb in chr11:5153026-5279647 (NCBI36/hg18) and contains SNPs from rs11036090 to rs7118113 of the Illumina Human610-Quad SNP array. This region contains: OR51B4, the complete HBB cluster, and OR51V1. Homozygosity mapping in 6 Senegal haplotype homozygotes showed a slightly larger RoH from chr11:4909490-5314457 and SNPs rs840713-rs10837822. Both the Saudi patients and Senegal homozygotes had the same homozygous genotypes for the overlapping region of chr11:5205580-5235931 ranging from rs11036364 to rs5010981.To identify potential genetic modifiers of HbF level in the region detected in the Saudi cases, we sequenced areas within or near the Corfu deletion that is known to cause HPFH, the HBD-HBG1 intergenic region, and core regions of HS- 2, 3, and 4 in the LCR. Core regions of HS-3 and HS-4 were identical to the reference sequences. In the core of HS-2, the 10TA.2CA.2TA.CG.12TA motif was present. This motif is known to be associated with the SI haplotype but not with any other haplotypes. Within the region of the Corfu deletion, many polymorphisms were identified highlighting the complexity of SI haplotype and HBB haplotypes in general. Many of these polymorphisms lead to creation or abolition of transcription factor binding sites when this was examined in silico using the TFBS search program ConSite (consite.genereg.net). Some of these putative sites bind transcription factors presumed to have regulatory roles in globin gene expression. Complete sequencing of the 126.6 kb interval with comparison to other HBB haplotypes associated with high and low HbF might focus attention on areas of interest that can be examined in functional studies.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.