Fetal‐haemoglobin enhancing genotype at BCL11A reduces HbA2 levels in patients with sickle cell anaemia

Abstract Understanding the interplay of genetic factors with haemoglobin expression and pathological processes in sickle cell disease is important for pharmacological and gene‐therapeutic interventions. In our nascent study cohort of Nigerian patients, we found that three major disease‐modifying factors, HbF levels, α‐thalassaemia deletion and BCL11A genotype, had expected beneficial haematological effects. A key BCL11A variant, while improving HbF levels (5.7%–9.0%), also led to a small, but significant decrease in HbA2. We conclude that in general, interventions boosting HbF are likely to reduce HbA2 in patients’ erythroid cells and that such therapeutic strategies might benefit from a parallel stimulation of HbA2 through independent mechanisms.


BACKGROUND
Sickle cell anaemia (SCA, homozygosity of the β S globin mutation), one of the commonest severe genetic disorders, is most prevalent in Africa.
Every year, 150,000 affected children are estimated to be born in Nigeria alone, an 'epicentre' of this debilitating disease. Insight into the causes of variable severity, which is strongly influenced by patients' genetic background, has stimulated the development of new therapeutic approaches. Attention is presently focused on fetal haemoglobin (HbF, α 2 γ 2 ) persistence as an alleviating factor, but also a raised presence of the minor adult haemoglobin, HbA 2 (α 2 δ 2 ), can inhibit HbS polymerisation [1,2], the key pathological mechanism.
In general, HbA 2 levels in a person are altered when the production of the major adult globin chains is impaired. For instance, in carriers of α thalassaemia mutations, they are decreased due to enhanced competition of the δ chain with the β chain, which has a higher affinity for pairing with the rate-limiting α chains. In the presence of β thalassaemia mutations [3] and β globin gene (HBB) regulatory variants [4], on the other hand, HbA 2 % is increased because of reduced β chain competition. Genetic variation at the HbF modifier locus HBS1L-MYB was found to also boost HbA 2 levels in non-anaemic subjects, while HbF modifier variants at the BCL11A and Xmn1-HBG2 loci had no such effect [4]. In sickle cell disease however, due to the effects of the mutation on the properties of the β chain, the relationship between HbF and HbA 2 levels, and with it the effect of genetic HbF modifiers on HbA 2 levels, is fundamentally altered [5]. The defective β globin chain, β S (forming HbS), has reduced affinity to the α chain, compared not only to the native β chain, but also to the γ and δ chains. In this situation, any increase in γ globin presence will lead to the replacement of a weaker competitor (β S ) with a stronger one (γ), leading to a reduction of δ chain inclusion into haemoglobin and to a reduction in HbA 2 levels. Therefore, in contrast to the direct correlation we observed in non-anaemic adults [4], levels for HbF% and HbA 2 % appear to have an inverse, or antagonistic, relationship in Hb SS erythrocytes [6]. A study of American SCD patients [5]

PATIENTS AND METHODS
Two hundred and forty-four patients with sickle cell anaemia (HbS homozygous), 5-46 years of age, neither on hydroxyurea nor transfusion, were recruited at the Lagos University Teaching Hospital (approved as ADM/DCST/HREC/1686) as part of a cross-sectional genetic study that has been described previously [9]. Our patient population is characterised by the absence of genetic HbF modifier XmnI-HBG2 (Senegal or Arab-Indian β-globin haplotypes) and by the presence of thalassaemia deletions (-α 3.7 ) in 33% of patients.
Genotyping for rs1427407 (TaqMan assay) was carried out as previously described [9], and the -α 3.7 thalassaemia deletion was assayed by multiplex PCR (polymerase chain reaction) [

RESULTS AND DISCUSSION
The effects on haematological profiles of our SCA patients were recorded for the three major disease-modifying factors HbF, α thalassaemia and BCL11A genotype (Table 1).
Higher HbF levels were associated with improved blood parameters in the patients, that is, raised total haemoglobin and reduced leukocyte counts (Table 1). A previously reported [5,6] depression of HbA 2 in the presence of more HbF was also detected.
Co-inheritance of the α-thalassaemia deletion (-α 3.7 ), observed in 32.9% of patients (α-/ αα: 29.3%, α-/ α-: 3.6%), led to a typical [13,14] decrease of red blood cell size and haemoglobin content (Table 1). We did not detect the frequently reported [13] improved total haemoglobin or a reduction in haemolysis indicators (reticulocyte count, LDH, data not shown), but did measure a marked increase in red blood cell numbers. A tendency towards improved haematocrit (p = 0.037) is not significant when multiple testing is considered but is nevertheless in keeping with the disease-alleviating character of alpha thalassaemia deletions.
Significantly elevated HbA 2 levels were seen in α deletion-carriers (Table 1), a paradoxical effect known for sickle cell disease, where δ chains have a competitive advantage over the low-affinity β S chain when α chains are in limited supply [15,16].
BCL11A genotype affected HbF levels significantly, as expected.
Presence of the minor allele ('T') of the key BCL11A enhancer variant rs1427407 led to an improvement of anaemia (increased Hb levels, p = 0.009), as has been reported for other SCD patient populations [8,17,18] and also to a tendency towards decreased platelet (p = 0.036) and reticulocyte (p = 0.024) numbers. While patients carrying rs1427407-T had higher average HbF levels -9.0% of total haemoglobin, compared to 5.7% in patients lacking this allele (p < 0.0001) -they showed a decrease of the presence of HbA 2 from 4.0% to 3.6% (p < 0.0001). The loss of HbA 2 is relatively small, and an overall HbS-diluting effect of this BCL11A enhancer variant is still evident. Nevertheless, our data appear to show that genetic (and possibly also pharmacological) alteration of HbF production in sickle cell disease is likely to be accompanied by some reduction in HbA 2 . HbF (heterocellular, i.e., restricted to F cells [19]) and HbA 2 (pancellular [20]) have a different distribution pattern across the population of red blood cells of a patient. Present (hydroxyurea) and future treatment strategies that increase HbF and F cell release could be complemented with a therapeutic approach to boost HbA 2 through an independent pathway, such as gene-regulatory elements detected through HbA 2 -associated genetic variants within the beta globin gene locus [4].

ACKNOWLEDGEMENTS
This study was funded by a grant to T. Adeyemo from the Central

DATA AVAILABILITY STATEMENT
Summary data supporting the findings of this study are available upon reasonable request from the corresponding author. Individualised patient data cannot be shared due to privacy and ethical restrictions.