Polymorphisms Associated with the Arab-Indian Haplotype of Sickle Cell Anemia Are Candidate Fetal Hemoglobin Gene Modulators

Fetal hemoglobin (HbF) inhibits HbS polymerization. Because of this, sufficient HbF in most sickle erythrocytes can lead to a milder disease phenotype. HbF levels differ amongst the β-globin gene (HBB) cluster haplotypes of sickle cell anemia. In the Arab-Indian (AI) haplotype, HbF was about 20% compared with 5-10% in the Bantu, Benin, and Senegal haplotypes. Functional elements linked to the HBB haplotype are likely to regulate the expression of HbF in addition to the effects of trans-acting modulators. To identify cis-acting SNPs in the HBB gene cluster that differentiate the AI haplotype from all others, including the Senegal haplotype—the Senegal haplotype shares some SNPs with the AI haplotype but its carriers have lower HbF—we studied patients with sickle cell anemia who were homozygous for HBB haplotypes by genome-wide SNP association analysis (GWAS; Table). First, we compared the results of GWAS of 42 Saudi AI haplotype homozygotes with GWAS in 71 Saudi Benin haplotype homozygotes. The only variants distinguishing these 2 populations with genome-wide significance (p-values between 9.6E-07 and 2.7E-45) were 223 SNPs in chromosome 11p15 from positions 3.5 to 6.5 mb. This region included the HBB gene cluster, its locus control region (LCR) and the upstream and downstream olfactory receptor gene clusters. The minor allele frequency of SNPs in MYB (chr 6q23), BCL11A (chr 2p16) and KLF1 (chr 19), trans-acting loci that affect expression of the HbF genes, were similar in these 2 cohorts. A novel candidate trans-acting locus was not found, however our power to detect such an association was low. We followed-up these observations by comparing allele frequencies in 303 African American cases homozygous for the haplotypes shown in the Table. Thirteen GWAS-significant SNPs, in addition to rs7482144 and rs10128556, were present in all AI haplotype cases but not in 83 Senegal haplotype chromosomes. The allele frequency of these SNPs was replicated in 62 independent AI haplotype cases. Rs2472530 is in the coding region of OR52A5; rs16912979, rs4910743 and rs4601817 are in the HBB gene cluster LCR; rs16912979 in DNase I hypersensitive site-4 altered motifs for POLR2A, GATA1, and GATA2 binding.The minor allele of rs10837771 causes a missense mutation in OR51B4  an upstream olfactory receptor gene. To see if any of these or other alleles might sometimes be associated with HbF in the Bantu and Benin haplotyes, we selected  homozygotes and compound heterozygous for these haplotypes who had unexplained and uncharacteristically high HbF. Thirty-one African Americans, aged ≥5 yrs. who had a HbF of 21% were compared with 350 similar cases who had a mean HbF of 3%.  Four additional SNPs on chromosome 11, from positions ranging from 5536415 to 5543705 in the UBQLNL/HBG2, region and present in 45-48% of AI haplotype and 3-13% of other haplotypes, were found at higher frequencies in the high HbF group compared with the low HbF group. These SNPs also altered transcription factor binding motifs. Loci marked by SNPs that distinguish the AI from the Senegal and other HBBhaplotypes might contain functionally important polymorphisms and account in part for high HbF in AI haplotype sickle cell anemia, independent of, or in addition to, the effects associated with rs7482144 or rs10128556. They might also be rarely associated with high HbF found in other haplotypes. These observations provide a foundation for mechanistic studies focused on the role of these variants in the expression of their linked HbF genes.

	non-coded allele	genomic location	Saudi AI (n=42)	Saudi ben.ben (n=71)	AA ben.ben (n=264)	AA ban.ban (n=31)	AA sen.sen (n=8)
HbF (%)			17	11	6	6	9
rs10837771	G	exon OR51B4	1	0.02	0.02	0	0
rs4601817	G	LCR	1	0.02	0.04	0	0
rs4910743	C	LCR	1	0.01	0.01	0	0
rs16912979	C	LCR	0	0.96	0.92	0.11	1
rs10488675	G	intron HBE1	1	0.01	0	0	0
rs7482144*	A	HBG2	1	0	0	0	1
rs10128556#	T	Intron HBBP1	1	0	0	0	1
rs7935470	C	OR51V1	1	0.02	0.03	0	0
rs10837582	G	OR51V1	1	0	0.02	0	0
rs11036227	T	OR51V1	1	0	0	0	0
rs10734485	C	OR51A1P	0	0.99	0.97	1	1
rs10837461	A	OR52A1	1	0.01	0	0	0
rs2472522	G	OR52A1	1	0.01	0	0	0
rs2472530	G	exon OR52A5	1	0.01	0	0	0
rs2499948	T	OR52A5	1	0.02	0.01	0.02	0
Allele frequencies in haplotypes of sickle cell anemia. * Xmn1 5’ HBG 2 restriction site. This SNP, not present on the SNP microarray, was genotyped independently; # LD with rs7482144; AA designates African Americans; ben-Benin; ban-Bantu; sen-Senegal.