Two search strategies were conducted in the PubMed, SciELO, and LILACS databases to develop this review. Manual searches were also conducted by checking references of potentially eligible studies. The searches were made in July and August 2023, in English and Portuguese, using the descriptors ‘Fetal Hemoglobin’, ‘Transcription Factors’, and ‘Hemoglobin Regulation’ and the Boolean operators AND and OR. The first search strategy aimed at identifying TFs related to γ-globin regulation within a timeframe of ten years (2013–2023). The second search strategy had two aims: (1) to select the most investigated TFs involved in Hb F regulation and (2) to obtain more functional data about these TFs. To achieve this, all the TFs identified in the first search were searched independently using the terms ‘Fetal Hemoglobin’ and the AND operator (e.g. ‘BCL11A’ AND ‘Fetal Hemoglobin’). The timeframe was reduced to five years (2018–2023) to select the most recent data. Figure 1 demonstrates the search strategies employed.

Figure 1.

Flowchart demonstrating the search processes applied to identify potentially eligible studies.

(0.18MB).

TF: transcription factors; SCA: Sickle cell anemia

Experimental gene editing studies and cohort studies analyzing HPFH polymorphisms were eligible to identify the TFs. Studies about regulatory components other than TFs, such as microRNAs (miRNAs), long non-coding RNAs (LncRNAs), and RNA Binding Proteins (RBPs), were not eligible. Additionally, studies that investigate pharmacological agents, physiological components other than TFs, and TFs investigated in less than five eligible studies were also excluded.

The results from the first search were exported to the Rayyan tool and subsequent article selection was based on the title and abstract. After duplicate studies were removed and eligibility criteria applied, each study was labeled with the investigated TF and the respective reasons for inclusion or exclusion. Following this, studies from the second search strategy were exported to Rayyan, and a similar selection and labeling process was performed. Finally, included studies were analyzed by reading the article in full to assess eligibility. The study selection process is presented in the flowchart in Figure 2.

Figure 2.

Flowchart demonstrating the identifying, screening, and selecting process of the studies.

(0.59MB).

Hemoglobin (Hb) synthesis is a crucial process in erythropoiesis that enables RBCs to bind to oxygen molecules. Hb structure contains four globin subunits transcribed by genes located on chromosomes 11 and 16. On chromosome 11, the HBB locus encodes epsilon (ε), gamma (γ), delta (δ), and beta (β) globins. In this context, the locus control region (LCR) interacts with globin genes through chromosomal looping to recruit transcription interactions at different stages of physiological development. 4,6,12,15

During the fetal period, Hb needs to bind with a higher affinity to oxygen molecules. Consequently, the LCR interacts with the HBG1 and HBG2 genes, leading to γ-globin synthesis and Hb F (α²γ²) expression. The second switching process occurs when the LCR moves from HBG1/2 to the HBB gene, leading to adult hemoglobin (Hb A - α²β²) synthesis through β-globin coding. Although fetal-to-adult hemoglobin switching is a natural and fundamental process in human development, β-globin expression can be problematic. Individuals harboring HBB mutations often synthesize abnormal hemoglobins, causing hematological diseases known as hemoglobinopathies. 1,2,4,6,12,14

During Hb F repression, KLF1 and MYB factors positively regulate BCL11A and LRF expression. BCL11A and LRF repressors bind to HBG1/2 promoters and recruit the NuRD complex to induce chromatin compaction. Doerfler et al.12 demonstrated that compacted chromatin eliminates the NF-Y and GATA1 activators of HBG1/2 promoters, further contributing to Hb F silencing (Figure 3). LCR interaction with HBG1/2 genes induces γ-globin expression by inhibiting BCL11A and LRF repressors and enabling NF-Y and GATA1 activators to directly bind to HBG1/2 promoters. Some studies suggest that GATA1 also indirectly regulates γ-globin through BCL11A and KLF1 interactions, but this transcriptional relationship is not yet fully understood. 4,6,11,14,15,52

Figure 3.

Transcriptional repression of Hb F: illustration demonstrating locus control region (LCR) inducing fetal-to-adult hemoglobin switching through chromosomal looping.

(0.24MB).

HRI is a kinase that phosphorylates eIF2α to regulate protein synthesis. In erythropoiesis, when the integrity of the HBB gene is impaired, HRI inhibits eIF2α phosphorylation, resulting in ATF4 repression. Boontanrart et al.14 and Huang et al.57 demonstrated that ATF4 downregulation inhibits BCL11A expression. Thus, ATF4 repression leads to MYB inhibition and compensatory γ-globin synthesis due to β-globin deficiency (Figure 4). These findings may explain why some HPFH variants occur through mutations in the HBB gene and the HB S1L-MYB intergenic region. However, while Huang et al.57 demonstrated that ATF4 directly interacts with BCL11A to repress Hb F, Boontanrart et al.14 demonstrated a specific interaction of ATF4 with MYB to inhibit BCL11A and LRF repressors. These divergent findings reinforce the need for further investigation regarding ATF4 molecular interactions. 14,17,56-60

Figure 4.

Transcription Activation of Hb F: stress erythropoiesis and a locus control region (LCR) performing chromosomal looping to attract GATA1 and NF-Y activators.

(0.28MB).

Collectively, these data demonstrate that Hb F is cooperatively regulated by transcription interactions, and TFs that regulate physiological homeostasis can induce Hb F synthesis when Hb A integrity is impaired. 4,6,12,14,17,52,56-60

Although randomized clinical trials are recommended to assess the feasibility of therapeutic interventions, only preclinical studies were identified in the literature. Among the identified studies, only Frangoul et al.3 and Esrick et al.10 explored the feasibility of interventions in human models. BCL11A was the only TF investigated in human experiments, demonstrating that the feasibility of various molecular targets is still uncertain. Despite this, preclinical studies have provided fundamental results that should be further evaluated to guide new clinical interventions.

KLF1 is a transcriptional activator of Hb F in −198 T > C, −123 T > C, and −124 T > C mutations. 7,22,43 However, most studies demonstrated that KLF1 repression increases Hb F levels, suggesting that KLF1 may play roles in both repression and activation of γ-globin. 8,12,18,48KLF1 polymorphisms are associated with the rare blood type In(Lu) and other RBC abnormalities.11,50 Lamsfus-Calle et al.11 demonstrated that KLF1 cleavage can result in dysregulated genes. Moreover, studies demonstrated that KLF1 haploinsufficient individuals exhibit varied Hb F levels, 11,49 possibly due to other regulatory components associated with KLF1 functionality and/or a specific mutation inherited by these individuals. Collectively, studies demonstrated that KLF1 is not yet a safe and effective therapeutic target for Hb F induction. 11,50

Research has demonstrated that specific HPFH mutations can create new binding motifs for γ-globin activators. 12,52,22 As a result, researchers have employed gene editing tools to replicate polymorphisms that upregulate NF-Y and GATA1 activators. Because NF-Y and BCL11A regulate γ-globin through divergent pathways, BCL11A editing can promote NF-Y binding at the CCAAT motif and result in γ-globin synthesis. Furthermore, the −110 A > C mutation stabilizes NF-Y at the CCAAT motif, and the −113 A > G mutation inhibits the CCAAT motif and creates a new GATA motif to activate Hb F transcription.12,52 Additionally, the disruption of the GATA motif impairs the pancellular distribution of Hb F in RBCs suggesting that GATA1 may be indispensable for γ-globin synthesis. 6,12,15,52

There are still gaps regarding the efficacy and safety of artificial mutations. Initially, for effective inhibition of Hb S polymerization, the artificial polymorphisms must trigger the pancellular HPFH phenotype. 1-3,52,53 Moreover, the use of nucleases to induce double-strand breaks may activate p53-dependent damage responses and lead to the occurrence of large deletions. 17,21,42 Cheng et al.17, Ravi et al.21 and Antoniou et al.42 demonstrated that base editors may be safer tools for inserting polymorphisms into the genome. The safety of the intervention can also be influenced by transcriptional targets. While GATA1 dysregulation can cause hematopoietic maturation defects, NF-Y disruptions may impair transcriptional events. 55,60 Additionally, Woodard et al.43 demonstrated that transgenic mice may not faithfully replicate the human HPFH due to species differences in the genome. Overall, the studies demonstrated that the safety and efficacy of artificial mutations remains uncertain in human models. 17,21,42,43,55,60

Several cohort studies have identified polymorphisms in the HB S1L-MYB intergenic region. 14,16,28,37 As previously mentioned, the safety and efficiency of artificial mutations are still uncertain in human models. 11,17,21,42 Although MYB cleavage can result in Hb F synthesis, studies evidenced that this TF is crucial for hematopoietic integrity. Clarke et al.17 demonstrated that MYB haploinsufficiency in mice results in various hematological disorders, suggesting that MYB is not a safe molecular target for clinical applications. However, Boontanrart et al.14 demonstrated that editing the HB S1L-MYB intergenic region results in Hb F expression because it prevents ATF4 binding, suggesting that ATF4 may be an efficient molecular target to inhibit MYB and activate γ-globin transcription.

ATF4 repression results in Hb F synthesis by inhibiting MYB and BCL11A. Because ATF4 is regulated by the erythroid-specific kinase HRI, this approach may offer a more precise way to repress MYB. 14,56-59 However, there are divergences between the studies by Boontanrart et al.14 and Huang et al.57 regarding the mechanism of action and the effectiveness of ATF4 as a molecular target. While Huang et al.57 demonstrated that ATF4 interacts directly with BCL11A and is not an effective molecular target in mice, Boontanrart et al.14 demonstrated that ATF4 binds to the HB S1L-MYB intergenic region and is an effective target in human cells. These findings highlight the need for further studies to elucidate the functionality and effectiveness of ATF4.

Because β-globin deficiency results in γ-globin expression due to stress erythropoiesis, artificial HBB polymorphisms may also be effective to trigger the HRI-eIF2α-ATF4 pathway. However, HBB deficiency is related to hematological disorders such as β-Thalassemia. Grevet et al.56 demonstrated that HRI depletion results in Hb F expression, possibly due to subsequent inhibition of ATF4. However, Peslak et al.56 demonstrated that >80% of HRI cleavage is necessary to achieve elevated Hb F levels. Although the study indicated that HRI depletion is well-tolerated under homeostatic conditions, it is crucial to assess HRI deficiency in pathological conditions, given that this kinase functions primarily during stress events. Zhang et al.59 demonstrated that, although healthy mice exhibit normal erythropoiesis, HRI depletion under iron deficiency conditions disturbs erythropoiesis. The study also revealed that ATF4 plays a crucial role in reducing oxidative stress, and ATF4 depletion is associated with neonatal mouse mortality. These data support the hypothesis that inhibiting homeostatic regulators may be a risky strategy, especially in individuals with hemoglobinopathies.

Polymorphisms that disrupt LRF binding motifs are associated with Hb F expression. 7,17,18,25,42 Combining LRF repression with the recruitment of a transcriptional activator can trigger higher levels of Hb F. These findings have been validated by studies that replicated the −198 T > C mutation, which results in LRF repression and creates a new binding motif for KFL1. 7,21,42

LRF and BCL11A cooperatively regulate Hb F through similar mechanisms. However, the diverse physiological functionality of LRF implies that BCL11A may be a safer target. Ravi et al.21 showed that LRF cleavage can result in erythroid differentiation defects. Furthermore, Bagchi et al.37 demonstrated that the integrity of erythroid cells remained unaffected after BCL11A editing, whereas LRF cleavage resulted in erythroid differentiation defects. These data demonstrate that LRF is not currently a safe molecular target and that future studies should focus on developing more specific strategies to inhibit LRF functionality.

The BCL11A gene encodes a TF especially crucial in the brain, B-cell development and the fetal-to-adult hemoglobin switch. Individuals harboring BCL11A mutations can express Dyas-Logan syndrome, a delayed intellectual disorder associated to HPFH. 27,45 Despite the limitations, BCL11A is a promising molecular target to induce γ-globin expression due to the possibility of specifically inhibiting this gene function in the fetal-to-adult Hb switching process. 3,4,11,12,21,23,33 Insertions or deletions at the TGACCA site of HBG1/2 promoters or in BCL11A erythroid-specific enhancers proved to be highly specific editing strategies for Hb F synthesis. Editing BCL11A erythroid-specific enhancers has been applied in rhesus monkeys and human models 3,10,33, while editing the TGACCA motif has only been employed in vitro and in mice. 4,6,11,12,21 Although the sample of clinical experiments is still small, strategically inhibiting BCL11A reduces SCA symptoms without affecting other physiologic functionalities of this TF. However, clinical experiments also demonstrated that myeloablative chemotherapy that precedes ex vivo interventions is highly toxic for patients. 3,11

BCL11A is the exclusive molecular target employed in human experiments due to its effectiveness and safety results in experimental studies. 3,4,6,10-12,20,22,31,37 Despite the diverse functionalities exhibited by all the TFs, strategically editing BCL11A was the sole intervention that did not affect parallel physiological events. 3,9,11,32,37 The aspects highlighted by studies reinforce that BCL11A stands out as the safest and most effective molecular target for Hb F synthesis. Nevertheless, it is crucial to emphasize that implications related to the cerebral and lymphocytic functionality of BCL11A underscore the need for meticulous editing analyses to enable clinical applications. 3,4,11-13,22,24

Gene editing is revolutionary for new therapeutic approaches. However, it is still necessary to develop safe and effective methods to deliver editing tools into host cells. Preclinical studies have employed ex vivo interventions, where HSCs are extracted, gene-edited, and subsequently reintroduced into the patient after myeloablative chemotherapy. Although ex vivo interventions are effective and provide greater control over editing procedures, cell extraction and myeloablation are invasive and potentially toxic. In contrast, in vivo interventions offer a safer, more accessible, and simpler alternative, as they eliminate the requirement for myeloablation and HSC harvesting. (Figure 5) 3,10,37,41,61-63

Figure 5.

Ex vivo and In vivo interventions employed for gene editing Created with Biorender.com.

(0.25MB).

The feasibility of clinical approaches depends on the strategy applied to introduce editing tools into host cells. Microinjection of the components or the insertion of electrical pulses into cells (electroporation) are highly effective physical methods often employed in genetic interventions. Although Frangoul et al.3 demonstrated that electroporation is efficient in human models, it is known that this method can be toxic to HSCs. Furthermore, the viability of physical methods may be restricted to ex vivo interventions, making biological methods a promising alternative for in vivo implementation. Lentivirus, adenovirus, and adeno-associated viruses (AAV) are commonly employed in biological delivery methods, particularly in vivo genetic interventions. The robust infection capacity of these viruses enhances their effectiveness as carriers of genetic tools into target cells, and several studies have underscored the efficiency of viral vectors in genetic interventions. 9,37,41,61-63 However, each viral method presents specific challenges and advantages. AAV, for example, has a reduced packaging capacity, which can be a challenge to deliver CRISPR/Cas9 ribonucleic complexes into HSCs. Nevertheless, a major concern related to viral vector delivery methods is associated to the potential risks of carcinogenesis and immunogenicity. 9,61-63

Multiple studies have explored non-viral biological delivery methods and assessed their safety and efficacy in host cells. 61-63 Lattanzi et al.62 demonstrated that non-viral ribonucleoprotein delivery may be a non-toxic and highly effective delivery method. Similarly, Cruz 61 demonstrated that Poly(lactic-co-glycolic acid) (PLGA) nanoparticles exhibit comparable efficiency to viral vectors. However, the study also underscored certain limitations, including the challenge of compounds reaching the bone marrow in in vivo interventions. These findings indicate that non-viral biological delivery has the potential to be safer than viral methods, but further efforts are required to optimize the efficiency of non-viral approaches. 61-63

Finally, it is crucial to empathize that SCA has a high incidence in low-income countries. These aspects underscore the importance of developing cost-effective and simpler approaches. Although most preclinical studies have utilized ex vivo interventions with physical or viral delivery methods, it is essential to optimize in vivo non-viral delivery to provide more accessible and economical interventions to patients. 1-4,10,20,40,41