BCL11A and HBS1L-MYB polymorphisms, in association with hydroxyurea and sex, modulate fetal hemoglobin levels in individuals with sickle cell anemia in the western region of Bahia, Brazil

Abstract

Background

Sickle cell anemia (SCA) is a condition caused by a mutation in the HBB gene, leading to the production of hemoglobin S in red blood cells. Hydroxyurea (HU), used in treatment, increases fetal hemoglobin (HbF) levels, thereby reducing erythrocyte sickling, as do single-nucleotide polymorphisms (SNPs) in the BCL11A gene and the HBS1L-MYB intergenic region. In this study, we investigated the combined effects of HU therapy, β^S-globin haplotypes, and SNPs in BCL11A and the HBS1L-MYB intergenic region on HbF levels in a cohort from Western Bahia, Brazil.

Methods and results

The DNA of 43 individuals with SCA was analyzed for SNPs rs4671393, rs7557939, rs1427407, and rs11886868 (BCL11A), and rs4895441, rs9402686, and rs11759553 (HBS1L-MYB). Individuals receiving HU therapy presented significantly higher HbF levels than untreated individuals (p = 0.028). Factorial ANOVA demonstrated a significant interaction between HU use, β^S haplotypes, and sex on HbF variability (p < 0.001). Significant associations with HbF levels were observed for rs7557939 in BCL11A under overdominant models in untreated individuals, and for rs11759553, rs4895441, and rs9402686 in the HBS1L-MYB region under dominant and recessive models. Multiple linear regression indicated that HU use, sex, rs1427407, rs7557939, and rs4895441 explained 46.5% of the variability in HbF (R² = 0.465; p = 0.001). Linkage disequilibrium (LD) analyses revealed non-random associations between rs4671393 and rs1427407; and between rs9402686, rs4895441, and rs11759553. The transcription factors GATA-1, SMARCA4, and CBFA2T3, which are essential for regulating hematopoietic cell production and differentiation, were associated with these variants.

Conclusions

These findings reinforce the contribution of genetic variants in BCL11A and HBS1L-MYB to HbF modulation in SCA. The identification of genetic markers associated with increased HbF levels may contribute to future strategies for individualized therapeutic monitoring in SCA.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11033-026-12151-9.

Introduction

Sickle cell anemia (SCA) is a genetic hematological disorder that affects thousands of people worldwide [1]. It is characterized by the homozygous presence of the β^S allele caused by a point mutation in the β-globin gene (HBB: c.20 A > T; rs334), resulting in the production of hemoglobin S (HbS), which, under deoxygenation conditions, undergoes polymerization, leading to a reduction in the deformability of erythrocytes and the characteristic sickling of red blood cells [2].

The condition leads to signs and symptoms that may include chronic anemia, jaundice, pain crises, and aseptic necrosis due to vaso-occlusive events, fatigue, splenic sequestration, acute chest syndrome, immunodeficiency, increased risk of stroke, as well as permanent vision damage [3].

The drug conventionally standardized by the United States Food and Drug Administration (US FDA) for the treatment of the disease is hydroxyurea (HU), which aims to increase fetal hemoglobin (HbF), thereby reducing HbS polymerization and erythrocyte density and inhibiting the sickling of red blood cells [4]. However, the effects of HU remain heterogeneous, and there is no concrete explanation for this event [5]. Despite these gaps, research points to a strong relationship between the presence of polymorphisms in genes related to HbF modulation and the response to HU treatment, including the B-cell lymphoma/leukemia 11 A gene (BCL11A) and the region between the HBS1-like translational GTPase and v-myb avian myeloblastosis viral homolog genes (HBS1L-MYB) [6–9].

BCL11A (2p16) is a gene that encodes the BCL11A transcription factor, involved in hemoglobin switching, which acts as a negative modulator of fetal hemoglobin levels in the erythroid cells of adult individuals [10]. Increased HbF levels have been widely associated with a milder clinical presentation of sickle cell disease due to their inhibitory effect on erythrocyte sickling [11]. Studies have linked single-nucleotide polymorphisms (SNPs) in the BCL11A gene to modulation of HbF expression [12].

Similarly, a region on the long arm of chromosome 6, between the HBS1L and MYB genes, is also involved in the increase in HbF. The HBS1L gene was first described in a study aimed at identifying pancreatic oncogenes, and its role in erythrocyte proliferation remains poorly understood [13]. The MYB gene is described as a proto-oncogene, a regulator of the production and differentiation of stem and progenitor cells in human bone marrow, and is mainly involved in hematopoietic disorders [14]. Data indicates that SNPs residing in the region between these two genes may be important modulators of HbF expression, further contributing to the development of SCA treatment [15].

In this study, we evaluated the association between polymorphisms in BCL11A (rs4671393, rs1427407, rs7557939, and rs11886868) and in the HBS1L-MYB intergenic region (rs9402686, rs4895441, and rs11759553) on HbF variation, considering haplotypes and HU use. The SNPs were selected based on previously reported evidence demonstrating their functional relevance in HbF regulation and their association with clinical variability and response to hydroxyurea treatment [16–20].

Materials and methods

Biological sample

For this study, 10 mL of peripheral blood was collected from 43 individuals with SCA treated at Barreiras’ Sickle Cell Disease Program (Western Bahia), based at the Leonídia Ayres de Almeida Municipal Basic Health Unit, Western Bahia, Brazil, and divided according to HU use (HU+) or non-use (HU-). The samples were transported under refrigeration to the Infectious Agents and Vectors Laboratory (LAIVE) at the Federal University of Western Bahia, where they were used for leukocyte genomic DNA extraction with Chelex^® 5% reagent and for confirmation of homozygous HbS by PCR-RFLP [21]. HbF levels were obtained from the patients’ medical records. The values correspond to laboratory results determined by HPLC and recorded during routine clinical evaluations in the medical records.

Given that HbF levels can be influenced by the β^S haplotype profile, the patients included in this study were also evaluated for their haplotype profile and its relationship with hematological and biochemical parameters, as well as with the use or non-use of HU; these data have been previously described [22]. β^S haplotype profiles were determined by PCR-RFLP [23], and atypical haplotypes were characterized as described in the literature [24].

The hematological data were obtained from laboratory records available in the patients’ medical charts, including routine tests performed during clinical follow-up. The sample excluded individuals who had received a blood transfusion in the 3 months prior to the day of collection, individuals in pain crisis, those who had undergone bone marrow transplantation, pregnant women, those using medications that alter hematological parameters, related individuals, and those who did not have data in their medical records regarding HbF levels. All experimental procedures were approved and registered by the research ethics committee of the Federal University of Western Bahia (UFOB) under number CAAE 29075620.6.0000.5026 in accordance with the National Health Council resolution number 466/2012 (Ministry of Health-Brazil).

Genotyping of the BCL11A gene SNPs

SNPs rs4671393 (A > G), rs7557939 (G > A), and rs1427407 (T > G) were genotyped using the TaqMan^® system (Applied Biosystems), using the TaqPath™ ProAmp™ Master Mix PCR kit and analyzed in a QuantStudio 5 thermocycler (Applied Biosystems). The SNP rs11886868 (T > C) was genotyped by Tetra-Primer ARMS-PCR, designed for this study [25], under the following reaction conditions: 1X buffer, MgCl₂ at 6mM, dNTPs at 0.5 mM, Forward outer, Reverse outer, Forward inner, and Reverse inner primers (Supplementary Table S1) at 0.2 µM each, DNA at 100 ng/µL, and ultrapure water to a final volume of 10 µL. The reactions were performed in a SimpliAmp thermal cycler (Applied Biosystems). The amplification product was analyzed by 2% agarose gel electrophoresis and stained with ethidium bromide.

Genotyping of SNPs in the HBS1L-MYB intergenic region

PCR-RFLP was used to detect rs4895441 (A > G). The reactions were performed to a final volume of 25 µL using 100 ng of genomic DNA, 1X PCR Buffer, 0.5 mM of each dNTP (Invitrogen), 0.1 U/µL of Platinum™ Taq DNA Polymerase (Invitrogen), and oligonucleotide primers at a concentration of 0.2 µM [26]. The amplification products were analyzed on a 2% agarose gel and stained with ethidium bromide. The RsaI enzyme was used to digest rs4895441 according to the manufacturer’s specifications, and the digestion products were visualized on a 3% agarose gel stained with ethidium bromide. The SNPs rs9402686 (G > A) and rs11759553 (A > T) were detected by the TaqMan^® system (Applied Biosystems), using the TaqPath™ ProAmp™ Master Mix PCR reaction kit and analyzed in a QuantStudio 5 thermocycler (Applied Biosystems).

Statistical analysis

Quantitative data were initially evaluated for normality and homogeneity of variances using the Shapiro-Wilk and Levene tests, respectively. Categorical variables, including sex and polymorphism frequencies, were evaluated using Pearson’s chi-square or Fisher’s exact tests. Hardy-Weinberg equilibrium (HWE) was assessed using HW_TEST software (version 1.1). HbF levels were also compared using genetic models: dominant, recessive, allelic, additive and overdominant [27–29], in this case, comparisons between two groups were performed using the t-student test or Mann-Whitney U test, while comparisons between three or more groups were conducted using the Kruskal-Wallis’s test or one-way ANOVA. Linear regression analysis was used to assess the influence of sex, use or non-use of HU, and the other variables on HbF levels. Correlation analyses were performed using Pearson’s correlation coefficients. A two-way ANOVA was performed to assess the interaction effects between variables on HbF levels. A square root transformation (sqrt transformation) was applied to the HbF values prior to the parametric analyses. Data that followed a normal distribution were described as mean ± SD, and those that did not follow a normal distribution were described as median and minimum and maximum values. For all analyses, the significance level used was p < 0.05 and was analyzed using Jamovi software (version 2.3).

Since linkage disequilibrium analysis depends on the physical proximity of variants to estimate non-random allelic associations and define haplotype blocks [30], the physical genomic positions of the SNPs were initially obtained using the online tool e! GRCh37 (https://grch37.ensembl.org/index.html). Subsequently, LD analysis was performed using Haploview version 4.2 software to calculate the LD coefficient (D’) and to construct haplotype block maps of the investigated variants. Finally, data related to the genomic location and functionality of the variants were found in the HaploReg version 4.2 (https://pubs.broadinstitute.org/mammals/haploreg/haploreg_v4.1.php) and RegulomeDB version 2.2 (https://regulomedb.org/regulome-search) databases.

Results

Characterization of the group, β^S haplotype, HWE, genotype, and allele frequency

The general characterization of the groups investigated in the study is in Table 1. The median age is 27 years for individuals who use HU (HU+) and 20 years for those who do not (HU-). HbF levels do not correlate with the age of the study participants (Pearson’s r = 0.160, p = 0.305). The HU+ group had a mean HbF% of 17.3% ± 7.89, and the HU- group had a mean HbF% of 11.7% ± 7.95. These data indicate that SS individuals using HU have significantly higher HbF values than individuals who do not use the medication (t-test = 2.28; p = 0.028).

Table 1.

Descriptive analysis of the individuals evaluated

	HU+ patients (N = 26)	HU- patients (N = 17)	p-value
	Mean ± SD	Mean ± SD
Age (years)	27.9 ± 12.78	21.8 ± 13.87
Min – Max	8–49	4–56	0.145
Gender
Male, n (%)	11 (42.30%)	9 (52.95%)
Female, n (%)	15 (57.70%)	8 (47.05%)	0.545
Hemoglobin
HbF (%)	17.3 ± 7.89	11.7 ± 7.95	0.028 *
Med	17.3	10.8
Min – Max	3.7–32	2–29.5
HbF by gender
Male (%)	17.6 ± 9.07	7.34 ± 4.56	0.005*
Med	19	6.6
Min – Max	3.7–29.3	2–17.3
Female (%)
Med	17.1 ± 7.24	16.6 ± 8.30	0.879
Min – Max	16.3	14.5
	6–32	3.5–29.5
White blood cells (mm³)	7.539 ± 1.71	11.521 ± 1.16	0.011*
Eosinophils (mm³)	315 ± 250	667 ± 719	0.073
Segmented cells (mm³)	4.001 ± 2542	6.078 ± 4202	0.067
Lymphocytes (mm³)	3.210 ± 1.513	3.987 ± 2.139	0.195
Monocytes (mm³)	722 ± 680	1.176 ± 661	0.055
Platelets (mm³)	338.132.000 ± 183.287	382.253.000 ± 164.786	0.401
Haplotypes
Bantu/Bantu	11 (42.3%)	5 (29.4%)
Bantu/Benin	10 (38.46%)	6 (35,3%)
Benin/Benin	4 (15.3%)	3 (17,6%)
Bantu/Atypical 1	0 (0%)	1 (5.9%)
Bantu/Atypical 2	0 (0%)	1 (5.9%)

SD = standard deviation, HbF = fetal hemoglobin, HU + = uses hydroxyurea, HU- = does not use hydroxyurea, Min = minimum, Max = maximum, Med = median, p* = <0.05 (values in bold indicate significant difference)

The individuals were also evaluated regarding the β^S haplotype profile, and the data are also described in Table 1. All participants in the study have Bantu and Benin chromosomes in a homozygous or heterozygous state, and only two carry chromosomes with an atypical haplotype profile (according to the classification described by Okumura et al., 2019 [24]). Therefore, none of the participants exhibits the XmnI/HBG2 (11p15.4) polymorphism, an important modifier that influences increased HbF levels [31–32].

A factorial ANOVA analysis indicated an interaction between HU use, haplotype profile, and sex on the HbF levels of the individuals studied (general model: F = 4.462; df = 11; p < 0.001). The model indicated an interaction between haplotype profile and HU use on HbF levels (F = 8.1482; df = 2; p = 0.002), with Benin/Benin HU- individuals presenting higher HbF levels compared to Bantu/Bantu HU- individuals (F = 15.35, df = 1; p = 0.0078), Bantu/Bantu HU+ individuals having higher HbF levels than Bantu/Bantu HU- (F = 22.9075; df = 1; p = 0.00029), and Bantu/Bantu HU- individuals with lower HbF levels than Bantu/Benin HU- (F = 8.9389, df = 1; p = 0.015). Therefore, overall, the data indicated that Bantu/Bantu individuals who do not use HU have the lowest HbF levels among the groups.

When gender was included in the analysis, the model remained significant (F = 7.638; df = 2; p = 0.0002). In general, females have higher levels of HbF than males. It should be noted that female individuals who use HU and have the Bantu/Bantu haplotype profile have approximately 3.8-fold higher HbF levels than men who do not use HU and who are homozygous for the Bantu haplotype (p = 0.006).

Therapeutic adherence among patients taking hydroxyurea (HU+) was inferred from clinical follow-up records and the observation of leukopenia, which is commonly seen in those taking the medication [33, 34]. HbF levels were initially compared between the HU + and HU- groups, without accounting for interactions among genetic variants, to evaluate drug response independently; higher HbF levels were observed in the HU+ group (Table 1, p = 0.028). Analysis by sex demonstrated a significant difference in HbF levels between males (p = 0.005), with individuals in the HU+ group having a higher mean HbF level than in the HU− group, corresponding to an increase of approximately 240%. The allele and genotype frequencies for each SNP were evaluated in the HU + and HU- groups using Pearson’s chi-square test, which revealed no significant differences between the groups (Table 2). HWE data showed that all SNPs were in equilibrium.

Table 2.

Genotype and allele frequencies, frequency analyses, HWE, and Pearson’s χ2 for the SNPs investigated

Gene		SNPs	Genotype / Allele	Frequency	HWE (p)	HWE χ2	Pearson χ2
BCL11A	Genotype frequency	rs4671393	GG	0.512
			AA	0.046	0.402	0.7	0.398
			AG	0.442
		rs1427407	GG	0.488
			TT	0.024	0.106	2.61	0.681
			TG	0.488
		rs7557939	GG	0.186
			AA	0.232	0.278	1.18	0.32
			GA	0.582
		rs11886868	TT	0.698
			CC	0.069	0.128	2.32	0.737
			TC	0.233
	Allele frequency	rs4671393 Minor allele: A	G	0.732
			A	0.268
		rs1427407 Minor allele: T	G	0.732
			T	0.268
		rs7557939 Minor allele: G	G	0.477
			A	0.523
		rs11886868 Minor allele: C	T	0.814
			C	0.186
HBS1L-MYB	Genotype frequency	rs9402686	GG	0.673	0.204	1.61	0.165
			AA	0.058
			AG	0.269
		rs4895441	GG	0.058	0.073	3.21	0.173
			AA	0.692
			AG	0.25
		rs11759553	AA	0.488	0.589	0.29	0.544
			AT	0.396
			TT	0.116
	Allele frequency	rs9402686 Minor allele: A	G	0.8
			A	0.2
		rs4895441 Minor allele: G	G	0.18
			A	0.82
		rs11759553 Minor allele: T	A	0.68
			T	0.32

HWE = Hardy-Weinberg equilibrium, SNP = single nucleotide polymorphism

Comparative analysis of HbF values and BCL11A SNPs

The analysis of the percentage variation of HbF according to the genetic models for all SNPs evaluated was performed considering the following criteria: (1) All individuals analyzed, regardless of sex and use of HU; (2) Use or non-use of HU; (3) Based on sex, regardless of the use of HU. The data for the analysis of BCL11A gene SNPs for each criterion are presented in Supplementary Table S2.

Criterion 1

As shown in Table S2, no statistically significant differences in HbF levels were observed for any SNP under the dominant model in Criterion 1.

Criterion 2

The results of a significant association were observed for rs7557939 under the overdominant model (p = 0.007), with a mean of 16.85 ± 7.96 for GG + AA compared with GA (7.12 ± 4.57), in the HU- group (Fig. 1a).

Fig. 1.

Analysis of the different dominance models applied to SNPs of the BCL11A gene and the HBS1L-MYB intergenic region a) Comparison between HbF levels and the overdominant model for SNP rs7557939 in the HU- group, b) HbF levels and the recessive model for SNP rs11759553, c) HbF levels and the recessive model for SNP rs4895441 in the HU+ group, d) HbF levels and the recessive model for SNP rs9402686 in the HU+ group, e) HbF levels and the dominant model for SNP rs11759553 in the HU+ group, f) HbF levels and the recessive model for SNP rs4895441 in males, g) HbF levels and the recessive model for SNP rs9402686 in males. Image edited using https://www.biorender.com

Criterion 3

Analysis of genetic models for SNPs indicates no statistically significant differences between sexes for the BCL11A polymorphisms evaluated.

Comparative analysis of HbF values and SNPs in the HBS1L-MYB intergenic region

The evaluation was performed according to the criteria mentioned above. The data for the analysis of HBS1L-MYB intergenic region SNPs for each criterion is presented in Supplementary Table S3.

Criterion 1

A significant association was observed for rs11759553 under the dominant model (p = 0.025, Fig. 1.b), with higher HbF levels in the AT + TT genotypes (17.83 ± 7.77) compared with the AA genotype (12.25 ± 8.04).

Criterion 2

A significant association was identified in the HU+ group for rs4895441 under the recessive model (p = 0.032, Fig. 1c), with higher HbF levels in individuals with the minor allele (GG) (28.60 ± 0.92) compared to AA + AG carriers (16.45 ± 7.46). Similarly, rs9402686 showed a significant association under the recessive model (p = 0.040, Fig. 1d), with higher median HbF levels in individuals with the minor allele (AA) [28.65 (28.0–29.3)] compared to GG + GA carriers [16.45 (3.7–32.0)]. Additionally, rs11759553 was significantly associated with HbF levels under the dominant model in the same group (p = 0.015, Fig. 1e), with higher HbF levels observed in the presence of the minor allele (20.98 ± 7.06) compared to homozygous dominant individuals (AA; 13.68 ± 7.14).

Criterion 3

Significant associations were identified in male individuals for rs4895441 in the recessive model (p = 0.032, Fig. 1f), with higher median HbF levels in individuals with the minor allele (GG) [28.65 (28.0–29.3)] compared with those with the dominant allele [7.95 (2.0–29.0)]. Likewise, rs9402686 in males was also associated with HbF levels in the same model (p = 0.038, Fig. 1.g), with higher HbF values observed in individuals with the minor allele, AA [28.65 (28.0–29.3)], compared to GG + GA carriers [7.95 (2.0–29.0)].

Variation in HbF levels by sex, HU use, and analyzed polymorphisms

Multiple linear regression models (with dummy variables) were fitted to assess the impact of the variables under investigation on HbF levels. The model proved significant when the variables “use of HU,” “sex,” “rs4895441,” “rs1427407” (considering the additive model for both polymorphisms), and “rs7557939” (recessive model) were included, and indicated that 46.5% of the variation in HbF levels is explained by the aforementioned predictor variables (R²: 0.465, F = 4.35, df1 = 7, df2 = 35, p = 0.001) (Table 3).

Table 3.

Linear regression analysis to assess the influence of sex, HU use, and variants on HbF levels (n = 43)

		Overall Model Test
Model	R 2	F		df 1	df2	p-value
	0.465	4.35		7	35	0.001
Preditor				95%CI
	Bb		ββ	Range			p-value	VIF	Tolerance
Female-Malea	0.735		0.645	0.0992-1.191			0.022	1.1	0.911
HU+ - HU− a	0.966		0.848	0.2448–1.451			0.007	1.19	0.841
HBS1L-MYB rs4895441 (Minor allele: G ) - AA-GGa	-1.964		-1.724	-2.4867			0.008	1.1	0.913
	-1.367		-1.199	-2.7991			0.091
HBS1L-MYB rs4895441 (Minor allele: G) - AG-GGa
BCL11A rs1427407 (Minor allele: T) - TG-TTa	2.704		2.373	0.5265–4.219			0.013	1.07	0.934
BCL11A rs1427407 (Minor allele: T) - GG-TTa	2.692		2.362	0.5716–4.153			0.011
BCL11A rs7557939 (Minor allele: G) - GG-AA + AGa	1.024		0.899	0.1518–1.646			0.02	1.17	0.853

^a: reference level; ^b: the magnitude of the effect (unstandardized regression coefficient – B and standardized regression coefficient – β) indicates the estimated change in HbF levels relative to the reference level; Multicollinearity (VIF < 10; tolerance > 0.2)

The analysis showed that HbF levels are significantly higher in: (i) females, compared to males; (ii) individuals who use HU; (iii) carriers of the TG and GG genotypes of the rs1427407 variant compared to homozygous T allele carriers; (iv) carriers of the GG genotype of the rs7557939 variant compared to individuals with AA and AG genotypes. On the other hand, homozygous carriers of the A allele of the rs4895441 variant have lower levels of HbF compared to carriers of the GG genotype (Table 3).

Analysis of linkage disequilibrium and functionality of SNPs in the BCL11A gene and the HBS1L-MYB intergenic region

The LD analysis showed that four pairs of variants had significant LD (Fig. 2). The SNP pairs rs4671393 and rs1427407 (BCL11A), as well as rs9402686, rs4895441, and rs11759553 (HBS1L-MYB), show strong LD.

Fig. 2.

Block map for BCL11A SNPs and the HBS1L-MYB intergenic region. A) Proximity relationship between SNPs rs4671393 (1), rs1427407 (2), rs7557939 (3), and rs11886868 (4); B) relationship between SNPs rs9402686 (1), rs4895441 (2), and rs11759553 (3). The numbers within the blocks indicate the R² for the analysis of variant pairs. The colors represent the degree of linkage disequilibrium: the more intense the red, the stronger the linkage disequilibrium. Image edited using https://www.biorender.com

Searches for location and functionality (Table 4) revealed that SNP rs4671393 is in a regulatory region of the genome that has been identified as a possible binding site for SMARCA4, a transcription factor, component of the SWI/SNF complex, involved in chromatin remodeling and transcriptional regulation, as evidenced by ChIP-seq data (RegulomeDB). The same SNP was identified in an active regulatory region in primary hematopoietic cells and B cells. Rs1427407 was also identified in primary hematopoietic cells and B cells. ChIP-seq analysis of this variant indicated an interaction with GATA-1, a transcription factor essential for the differentiation of erythroid and megakaryocyte cells. It is important to note that studies have shown that this region corresponds to an erythroid enhancer located in intron 2 of BCL11A, where its activity is modulated by genetic variations, primarily by rs1427407, which has been identified as a functional variant capable of altering the binding of erythroid transcription factors, impacting the expression of BCL11A and, consequently, HbF levels [35].

Table 4.

Location and functionality data for SNPs in the BCL11A and HBS1L-MYB genes

SNP ID	POSITION	HaploReg	RegulomeDB
rs4671393	chr2:60493815–60,493,816	Primary hematopoietic cells / B cells	SMARCA4
rs1427407	chr2:60490907–60,490,908	Primary hematopoietic cells / B cells	GATA-1/Hematopoietic progenitor cell
rs9402686	chr6:135106678–135,106,679	Information unavailable	Venous blood
rs4895441	cr6:135105434–135,105,435	CD8 + T cell / Fetal thymus	CBFA2T3 and HMBOX1 / Venous blood
rs11759553	cr6:135101157–135,101,158	CD8 + T cell / Fetal thymus	CBFA2T3 / Venous blood

SNP: Single nucleotide polymorphism; eQTL: Expression of Quantitative Locus Trait

The rs9402686 showed a binding site for POLR2H in venous blood-derived cells, an essential subunit of RNA polymerase II. The rs4895441 was associated with CD8 + T cells and fetal thymus and showed ChIP-seq binding with CBFA2T3 and HMBOX1, both transcription factors involved in hematopoiesis and cell differentiation. The rs11759553 was identified in CD8 + T cells and fetal thymus, like rs4895441. The SNP also showed ChIP-seq binding to the same transcription factor identified in the analysis of rs4895441, CBFA2T3.

Discussion

This study demonstrated associations between HU use, β^S haplotypes, and genetic variants in BCL11A and in the HBS1L-MYB region, which were associated with increased HbF levels, reinforcing that both genetic background and HU therapy contribute to interindividual variability in treatment response. Significant interactions were observed between HU therapy and β^S haplotypes, indicating that both pharmacological and genetic factors contribute to HbF variability in this population. These findings reinforce previous evidence regarding the role of HbF-modulating polymorphisms and provide additional data on their combined effects in a population from Bahia, Brazil.

In this study, individuals receiving HU had significantly higher HbF levels than untreated individuals, reinforcing HU as the primary pharmacological therapy for inducing HbF in patients with SCA. These results are consistent with previous research demonstrating that HU therapy significantly increases HbF levels in individuals with SCA. These findings are consistent with previous studies demonstrating significant increases in HbF following HU therapy in both adult and pediatric patients with SCA [36–38]. Although the exact mechanisms underlying HU-induced HbF production are not yet fully understood, these results reinforce the importance of treatment adherence and highlight the role of HU as a key disease-modifying agent [39], remaining the primary therapeutic strategy for increasing HbF levels and reducing disease severity, including vaso-occlusive crises and hospitalization rates [40].

It is also known that the β^S haplotype profile modulates HbF levels in SCA. In this study, the haplotype profile, sex, and HU use showed a statistically significant interaction. Male individuals carrying Bantu chromosomes who do not use HU have the lowest levels of HbF. The Bantu haplotype has been associated with more severe clinical conditions due to its link with low HbF levels and increased oxidative stress, inflammation, and DNA damage [41–43]. Supporting this observation, Santana et al., 2025 described a significant increase in levels of biochemical parameters, such as platelets, AST (aspartate aminotransferase), DB (direct bilirubin), IB (indirect bilirubin), and LDH (lactate dehydrogenase) in Bantu/Bantu individuals not using HU, compared to drug users with the same haplotype, among the same patients investigated in this study [22]. Elevated levels of these markers have been associated with vaso-occlusive processes, as well as with various clinical complications, such as thrombosis, pulmonary hypertension, leg ulcers, renal dysfunction, priapism, and an increased risk of death [44–48]. Taken together, these results underscore the critical interaction between haplotype profile and HU therapy in modulating HbF levels and disease severity.

Considering that HbF variability is multifactorial, we further evaluated whether biological sex could influence the interaction between β^S haplotypes and HU response. When sex was incorporated into the analysis, females presented significantly higher HbF levels than males, particularly among individuals receiving HU therapy. In this study, women with the Bantu/Bantu haplotype who received HU exhibited higher HbF levels than untreated males with the same haplotype profile. Di Mauro et al. demonstrated that males with SCD tend to exhibit a reduced response to HU therapy and an increased inflammatory burden compared to females, reinforcing the influence of biological sex on disease modulation and treatment response [49]. A large cohort study observed significantly higher HbF levels in females than in males, even among patients receiving HU therapy, supporting the hypothesis that sex-related mechanisms may influence HbF regulation [50]. These findings suggest that HbF modulation may result from a complex, multifactorial interaction rather than from the isolated effect of a single variable.

The available evidence indicates that sex effects on HbF levels operate independently of haplotype, with the X-linked F-cell production (FCP) locus as the dominant factor modulating haplotype effects on HbF expression [51]. The X-linked FCP locus accounts for approximately 40% of HbF variation in SCA, serving as a primary determinant of how other genetic factors, including β-globin haplotypes, influence final HbF levels [52]. However, direct haplotype-specific analyses of sex differences are limited in the literature. These findings should be interpreted with caution, as they may reflect the influence of the underlying genetic structure, including haplotype distribution, rather than a direct biological effect of sex. Additionally, recent data from Arcanjo et al. (2026), in a cohort of 409 patients with SCA in Brazil, demonstrated a higher frequency of severe clinical complications such as stroke, avascular necrosis, leg ulcers, priapism, and acute chest syndrome in male individuals [53]. These findings, together with the results of the present study, suggest that genetic and possibly therapeutic factors, such as HU use, may contribute to differences in the clinical presentation of the disease between the sexes.

The influence of polymorphisms on the genetic modulation of HbF levels has been extensively studied, particularly the impact of variants in the BCL11A gene and the HBS1L-MYB intergenic region [54–56]. In this study, we identified that carriers of alleles with variants in different polymorphisms in BCL11A and HBS1L-MYB may exhibit higher levels of HbF, which could affect the clinical condition of individuals with homozygous HbS.

Among the BCL11A SNPs, only significant associations involving rs7557939 were identified in untreated individuals (HU-), particularly under overdominant genetic models, with respect to criterion 2 of this study. A significantly greater effect on HbF levels was observed in the presence of the minor allele (GG) of the rs7557939 variant associated with the major allele (A) compared to AG genotypes (p = 0.007). Although this variant has not yet been widely highlighted for its association with increased HbF levels compared to other SNPs, the results from this population are consistent with studies by other authors that also observed increased HbF levels associated with the G allele of this variant in individuals with SCA [57].

Other studies have also linked the GG genotype to higher HbF levels, reinforcing the role of this genetic variant in increasing HbF [58]. These findings are biologically plausible given the central role of BCL11A in the silencing of γ-globin and the transition from fetal to adult hemoglobin [59]. Variants located in intronic erythroid enhancer regions of BCL11A, such as rs7557939, may alter transcription factor binding and erythroid-specific regulatory activity, directly influencing HbF expression [60].

In this context, the association observed for rs7557939 in individuals not using HU may indicate that this variant contributes to endogenous HbF production rather than exclusively to the pharmacological response. These findings reinforce the multifactorial nature of HbF regulation and underscore the importance of BCL11A variants as modulators of hematological variability in SCA.

GWAS studies have identified polymorphisms in the HBS1L-MYB intergenic region as the most relevant modifiers of HbF levels, with particular emphasis on the HMIP-2 block [61, 62] (references). In this study, individuals carrying allelic variants of HBS1L-MYB, analyzed under different criteria for the polymorphisms rs11759553, rs4895441, and rs9402686, showed significant variation in HbF levels, indicating that carriers of minor alleles for these variants have higher HbF levels (Supplementary Table S2).

In the HBS1L-MYB intergenic region, the rs11759553 variant showed statistically significant associations with increased HbF levels, especially in the presence of the minor allele (T). This finding is consistent with studies conducted in geographically close populations, suggesting shared genetic characteristics that may influence the progression of SCA in Brazil [63, 64]. Furthermore, this variant has been associated with improvements in hematological parameters, such as hematocrit and hemoglobin concentration, particularly in individuals receiving HU, thereby alleviating disease symptoms [65]. Thus, the results of this study reinforce the importance of this genetic factor in individuals with SCA, as it can both modulate the response to medication and mitigate disease progression.

The rs9402686 showed a significant association with HbF levels in this study. Some studies have shown that the same polymorphism, in a specific sample of Nigerian patients, was strongly associated with higher levels of HbF [66]. In the landmark Cooperative Study of Sickle Cell Disease (CSSCD) and Brazilian cohorts, HBS1L-MYB SNPs, including rs9402686, were strongly associated with HbF variation [67]. Carriers of the minor allele at rs9402686 (AA) exhibit improved clinical parameters, including a reduced reticulocyte count, higher peripheral oxygen saturation, and a lower risk of acute chest syndrome [68]. These clinical benefits appear to be mediated primarily by higher concentrations of HbF. The mechanism underlying these findings appears to be that the increased HbF at rs9402686 reduces transcription factor binding efficiency, thereby decreasing MYB expression and, consequently, increasing HbF production [69, 70].

In this study, we also found that users of hydroxyurea and carriers of the minor allele at rs4895441 have higher HbF levels. It is known that rs4895441, located in the HMIP-2 block, negatively regulates MYB [71]. It is well known that c-MYB is involved in regulating HbF production [72]. The MYB gene, which encodes the c-MYB transcription factor, plays an important role in regulating globin gene expression, cell cycle progression, and erythropoiesis [73, 74]. Additionally, c-MYB can activate repressor genes of γ-globin and inhibit HbF expression [75]. Interestingly, Cardoso and colleagues (2014), evaluating polymorphisms in HMIP-2 in individuals with SCA in northern Brazil, identified that among the variants rs2838451 (HMIP-1), rs4895441, and rs9399137 (HMIP-2), only rs4895441 was significantly associated with HbF levels [76]. Possibly, the relationship of this variant with c-MYB expression is the reason why individuals carrying the minor allele of rs4895441 exhibit higher HbF levels in response to HU use.

Although the individual identification of the relationship between BCL11A and HBS1L-MYB variants is significantly associated with HbF variation, in our analysis, sex and the use of hydroxyurea (HU) in conjunction with the rs1427407 and rs7557939 variants in BCL11A and rs4895441 in HBS1L-MYB are responsible for 46.5% of the variation in HbF levels, confirming the importance of these variables evaluated independently. The model indicates that females have higher HbF levels than males, and that HU use combined with inheritance of lower-allele variants significantly increases HbF levels in sickle cell anemia (SCA). Part of the HbF variability cannot be explained by the model, and this likely reflects interactions with other genes, epigenetic modifications, or variants of smaller effect [77]. Despite this, these findings reinforce the scope of considerations for the management, optimization, and clinical monitoring of SCA patients.

The presence of LD between the rs4671393 and rs1427407 variants of the BCL11A gene and the rs11759553, rs9402686, and rs4895441 variants of the HBS1L-MYB intergenic region suggests a possible functional interaction that may help regulate HbF production, thereby favoring its expression [78].

The variants in BCL11A in LD are in potentially regulatory regions associated with SMARCA4 and GATA-1 binding sites, which are essential for the regulation of hematopoietic cell production and differentiation, suggesting a possible functional role that requires experimental validation. SMARCA4, also known as BRG1, is an important transcription factor that acts in chromatin remodeling, exposing hematopoietic genes, and acting as a differential regulator of the self-renewal and renewal of hematopoietic stem cells, which was evidenced after its silencing by RNAi, which led to impaired self-renewal capacity of these cells [79].

In addition, it acts as a differential regulator of GATA-1 transcription, a key transcription factor in erythropoiesis that recruits co-regulatory complexes to drive gene activation and repression during the production of primary erythroid cells [80]. GATA-1 forms a complex with the DNA-binding transcription factors SCL/TAL1 and the structural proteins LMO2 and LDB1 to mediate the enhancer’s interaction with target promoters for gene activation [81, 82]. GATA-1 and LDB1 recruit the chromatin modifiers CBP/P300 and the BRG1 catalytic subunit of SWI/SNF to the β-globin locus control region (LCR) to promote erythroid differentiation and the formation of higher-order chromatin loops [88, 83, 84]. SMARCA4 (BRG1) and GATA-1, together with various factors, contribute to the regulation of LCR-γ-globin chromatin loop formation [85].

There was also evidence of linkage disequilibrium (LD) among the variants rs4895441, rs4902686, and rs11759553, which may also suggest that these SNPs are associated with the expression of genes essential for HbF production in the individuals who participated in this study [86]. Therefore, the LD patterns observed among these SNPs may reflect haplotype blocks that contribute to the regulation of HbF expression, offering potential targets for therapeutic interventions to increase HbF levels in patients with SCA.

All the HBSL1-MYB contributed to the increase in HbF under the three criteria analyses. The rs4895441 (involved with the regression model that explains 46.5% of the variation in HbF found) and rs11759553 are involved in key binding sites with transcription factors for hematopoietic cell production, such as CBFA2T3 (translocated alpha-2 subunit of the nuclear binding factor), also known as ETO2, a key transcriptional co-regulator in hematopoiesis and the regulation of differentiation of cell lines that contribute to the formation of blood components, such as erythrocytes and platelets [87]. The loss of ETO2 in murine and human erythroid progenitors results in delayed hemoglobin switching and the persistence of embryonic/fetal hemoglobin in adults, potentially representing a target for drugs used to treat β-hemoglobinopathies [88, 89]. To our knowledge, few studies have evaluated these haplotypic associations in Brazilian individuals with SCA.

From a clinical perspective, these findings suggest that the genetic profile of polymorphisms associated with HbF and β^S haplotypes may help predict individual response to HU therapy. These findings may, in the future, contribute to personalized medicine strategies, although validation in larger cohorts is necessary. Furthermore, identifying individuals with high-risk haplotypes may enable earlier interventions and improved clinical management.

Conclusions

The study highlights the relevance of genetic factors in modulating HbF levels in patients with SCA. Significant associations were identified between specific SNPs, particularly rs4895441, rs1427407, and rs7557939, when considered in relation to sex and HU use, indicating that these variables together modulate HbF levels and account for approximately 46% of the variation in HbF. Analysis of genetic variants revealed LD patterns, suggesting functional interactions that favor the regulation of HbF production. From a clinical standpoint, identifying genetic variants associated with hydroxyurea response can help healthcare professionals better predict therapeutic outcomes and optimize treatment strategies for patients with sickle cell anemia.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Manoel Ferreira de Magalhães Filho: Visualization, Investigation, Formal Analysis, Writing-Original Draft. Ilana Luize Rocha Santan: Investigation. Pâmela Lourdes Pereira da Silva: Investigation. Larissa Paola Rodrigues Venancio: Conceptualization, Resources, Methodology, Project administration, Funding acquisition, Supervision, Writing- Original Draft.

Funding

The Article Processing Charge (APC) for the publication of this research was funded by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) (ROR identifier: 00x0ma614). This study was supported by the Bahia State Research Support Foundation (Fundação de Amparo à Pesquisa do Estado da Bahia - FAPESB), PET0006/2022 - LPRV, PDPG III 38/2022 CAPES/FAPESB, and FINEP. This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - finance Code 001.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.

Consent to participate

All participants provided written consent to participate in the research, including those under 18 years of age, for whom written consent was obtained from their parents.

Consent to publish

Not applicable.

Ethical statement

This study has been approved by the Research Ethics Committee of the Federal University of Western Bahia - UFOB (CAAE 29075620.6.0000.5026) in accordance with the National Health Council Resolution number 466/2012 (Ministry of Health – Brazil).