Sickle cell anemia (homozygosity for HbS) and HbSC disease (compound heterozygosity for HbS and HbC) are the most common genotypes of sickle cell disease (SCD). Fetal hemoglobin (HbF), the major genetic modulator of SCD, interferes with deoxyHbS polymerization [1]. Patients with HbSC disease have about half the number of complications as do HbS homozygotes [1], however, their mean HbF levels were 2.4% compared with about 6% in patients with sickle cell anemia [2]. Single nucleotide polymorphisms (SNPs) in quantitative trait loci cis and trans to the β-globin gene cluster are associated with HbF. We analyzed the association of SNPs in BCL11A, (chromosome 2p), in the HBS1L-MYB intergenic region (HMIP) (chromosome 6q), and in the 5′ olfactory receptor (OR) gene cluster of chromosome 11p with HbF in Brazilian patients with SCD.
622 patients were studied. They were not treated with hydroxyurea and were aged ≥6 years. Hemoglobin was analyzed by HPLC. SNPs rs766432 and rs6732518 in BCL11A; rs35959442 and rs11759553 in HMIP; and rs4910755 and 7483122 located in OR51B5-OR51B6 were genotyped by ABI TaqMan Assays. The βS and βC globin haplotypes were investigated with PCR and restriction fragment length polymorphism techniques. Mean HbF was compared according to genotype using the additive genetic model adjusting for age, gender and HbS homozygotes versus HbSC. A dominant genetic model adjusting for age, gender and SCD genotype was used to test the association SNPs with HbF. We tested to see if there was an association between HbF and SNPs in the OR gene cluster region after adjusting for age, gender, SCD genotype and the HBB gene cluster haplotype (Central African Republic [CAR] or Bantu versus absence of a CAR haplotype). To determine if there was a synergistic effect between the SNPs or if the effect exerted on HbF was independent, we developed multivariable linear regression models adjusting for age, gender and SCD disease status and ran stepwise linear regression analyses where we added the most significant SNPs from BCL11A, HMIP one-at-a-time. Data analysis was performed using R version 2.14.1 and the cubic root transformation of HbF [3].
392 patients were HbS homozygotes (HbF 7.4±4.4%) and 230 had HbSC disease (HbF 2.5±2.3%). The association between HbF and SNPs in the BCL11A, HMIP and the OR gene cluster for all patients combined and for HbS homozygotes and HbSC patients separately are shown in Table I a-c (on-line supplementary material). In the combined group, there is a statistically significant association between HbF levels and both BCL11A SNPs and with HMIP SNP rs11759553 after adjusting for age, gender and the SCD genotype. As the number of minor alleles increase, so does HbF level; in HbS homozygotes a similar result was found; in HbSC disease both SNPs in BCL11A and HMIP were associated with HbF. The association between HbF levels and SNPs in the OR genes, after adjusting fo r HBB haplotype was not significant (Table II, on-line supplementary material). In a stepwise linear regression analysis adding the SNPs to a multivariable linear regression model as described, after adjusting for age, gender and SCD genotype rs35959442 is still significantly associated with HbF indicating that its effect on HbF is independent of rs766432 (Table III, on-line supplementary material).
We dichotomized the patients according to the HBB gene cluster haplotype. Brazilians with SCD have a higher prevalence of the CAR haplotype compared with African Americans. This haplotype is associated with lower levels of HbF than other haplotypes [4][5]. Sickle cell anemia patients were heterozygous or homozygous for the CAR haplotype chromosome or lacked a CAR haplotype chromosome; HbSC disease patients had or did not have a CAR haplotype chromosome. Most patients with a CAR haplotype had the minor allele for rs4910755 and rs7483122, both of which are located in the 5′ OR gene cluster of chromosome 11p and could be in linkage disequilibrium with SNPs characterizing this haplotype.
HbSC disease patients usually have low HbF levels. Nevertheless, it is a milder condition than sickle cell anemia even though patients are more susceptible to clinical events related to high bloodviscosity. Although HbF in HbSC disease is lower than in HbS homozygotes, BCL11A and HMIP modulate HbF in both genotypes.
Supplementary Material
Acknowledgments
Financial support: NIH / Fogarty Internacional Center RA237816BAJ; NIH/NHLBI T32HL007501; Conselho Nacional de Pesquisa (CNPq) [307496/2010-4]
REFERENCES
- 1.Steinberg MH, Nagel RL. HbSC disease and HbC disorders. In: Steinberg MH, Forget BG, Higgs DR, Weatherall DJ, editors. Disorders of hemoglobin: genetics, pathophysiology, clinical manegement. Cambridge University Press; Cambridge: 2009. pp. 525–548. [Google Scholar]
- 2.Solovieff N, Milton JN, Hartley SW, et al. Fetal hemoglobin in sickle cell anemia: genome-wide association studies suggest a regulatory region in the 5′olfactory receptor gene cluster. Blood. 2010;115:1815–1822. doi: 10.1182/blood-2009-08-239517. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.R Development Core Team . A language and environment for statistical computing. R Foundation for Statistical Computing; Vienna, Austria: 2011. ISBN 3-900051-07-0, URL http://www.R-project.org/ [Google Scholar]
- 4.Nagel RL, Rao SK, Dunda-Belkhodja O, et al. The hematologic characteristics of sickle cell anemia bearing the Bantu haplotype: the relationship between G gamma and HbF level. Blood. 1987;69:1026–1030. [PubMed] [Google Scholar]
- 5.Goncalves MS, Nechtman JF, Figueiredo MS, et al. Sickle cell disease in a Brazilian population from Sao Paulo: a study of the bea S haplotypes. Hum Hered. 1994;44:322–327. doi: 10.1159/000154238. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
