Manipulation of Developmental Gamma-Globin Gene Expression: an Approach for Healing Hemoglobinopathies

β-Hemoglobinopathies are the most common monogenic disorders, and a century of research has provided us with a better understanding of the attributes of these diseases. Allogenic stem cell transplantation was the only potentially curative option available for these diseases until the discovery of gene therapy. The findings on the protective nature of fetal hemoglobin in sickle cell disease (SCD) and thalassemia patients carrying hereditary persistence of fetal hemoglobin (HPFH) mutations has given us the best evidence that the cure for β-hemoglobinopathies remains hidden in the hemoglobin locus.

ABSTRACT

β-Hemoglobinopathies are the most common monogenic disorders, and a century of research has provided us with a better understanding of the attributes of these diseases. Allogenic stem cell transplantation was the only potentially curative option available for these diseases until the discovery of gene therapy. The findings on the protective nature of fetal hemoglobin in sickle cell disease (SCD) and thalassemia patients carrying hereditary persistence of fetal hemoglobin (HPFH) mutations has given us the best evidence that the cure for β-hemoglobinopathies remains hidden in the hemoglobin locus. The detailed understanding of the developmental gene regulation of gamma-globin (γ-globin) and the emergence of gene manipulation strategies offer us the opportunity for developing a γ-globin gene-modified autologous stem cell transplantation therapy. In this review, we summarize different therapeutic strategies that reactivate fetal hemoglobin for the gene therapy of β-hemoglobinopathies.

INTRODUCTION

Red blood cells (RBCs) are biconcave-shaped cells with capacious cytoplasm, lacking nuclei and cell organelles, to provide ample room for the accumulation of hemoglobin required to carry out gaseous exchange for 120 days. Adult hemoglobin is a heterodimeric protein composed of dimers of α-globin and β-globin chains with heme as a prosthetic group (1). The life-threatening conditions called hemoglobinopathies arise when the integrity of the hemoglobin structure or the synthesis of globin chains is compromised due to mutations in the globin genes. β-Hemoglobinopathies, particularly, sickle cell disease (SCD) and β-thalassemia, are a group of inherited monogenic recessive disorders marked by defective or decreased production of β-globin chains, respectively (2). Globally, β-hemoglobinopathies are highly prevalent in the Mediterranean basin, Africa, Middle East, South and Southeast Asia, and the Pacific Islands. Advancements in transportation and globalization eased the migration of individuals, shifting the demographics of β-thalassemia and SCD to regions where globin chain disorders were not endemic (3). Every year, between 300,000 and 400,000 newborns are diagnosed with a spectrum of hereditary hemoglobinopathies (4). A survey from the World Health Organization on the global epidemiology of hemoglobin disorders showed that β-hemoglobinopathies account for 3.4% of mortality among children under the age of 5 years. β-Thalassemia and sickle cell disorders pose a significant health burden in India. Based on another survey by March of Dimes Global Report on Birth Defects, the incidence rate of β-hemoglobinopathies in India is 1.2 per 1,000 births, which translates to 32,400 children with β-hemoglobinopathies per 27 million births per year in India. The prevalence rate of β-thalassemia carriers in India ranges from 3% to 4%, which translates to 35 to 45 million carriers in a vast population of 1.21 billion people (5). Currently, allogeneic stem cell transplantation remains the only curative treatment available for a very small fraction of patients due to HLA restrictions. However, gene therapy with autologous stem cell transplantation has the potential to become the first line of treatment in the near future. Persistent production of gamma-globin (γ-globin) by gene-modified autologous hematopoietic stem cells could be an effective gene therapy strategy to alleviate the severity of β-hemoglobinopathies. This review deals with the clinically relevant and possible gene manipulation strategies for increasing fetal hemoglobin (HbF).

β-HEMOGLOBINOPATHIES AND THEIR PATHOPHYSIOLOGY

β-Thalassemia is associated with decrease in (β⁺) or absence of (β⁰) the synthesis of the β-globin chain. More than 200 different types of mutations, including frameshift, splicing, promoter mutations, and deletions in the β-globin gene, result in this clinically heterogeneous condition. Based on the clinical outcomes of different mutations and their associated severities, β-thalassemia is categorized as major, minor, or intermedia. β-Thalassemia major is the most severe form requiring chronic red blood cell transfusion, the minor form is asymptomatic due to heterozygosity in the thalassemia-causing mutations, and the intermedia form is heterogeneous, with severity ranging from asymptomatic to transfusion dependent (6, 7). In β-thalassemia, the levels of adult hemoglobin (α2β2) are greatly affected by an imbalance in the production of protein monomers. The equilibrium loss due to decreased β-globin production causes the accumulation of normally produced α-globin chains. The excess free insoluble α-globin chains precipitate in mature erythroblasts and induces apoptosis, resulting in ineffective erythropoiesis (8, 9). Severe anemia as a result of apoptosis and ineffective erythropoiesis triggers extramedullary hematopoiesis (EMH) in the liver, spleen, lymph nodes, and other sites of the body. EMH causes severe abdominal discomfort, bone pain, and compression of the vital organs. Treatment involves lifelong red blood cell transfusions along with iron chelation therapy, which are aimed at compensating the ineffective erythropoiesis and reducing the anemia. Splenectomy is commonly practiced to treat transfusion-dependent β-thalassemia patients, as it decreases the frequency of transfusions by increasing the RBC survival and the amount of circulating red cells in the blood (10).

SCD comprises a group of inherited recessive blood disorders caused by different point mutations in the β-globin locus that result in the production of defective β-globin chains. These include sickle hemoglobin (HbS) in sickle cell anemia (β6Glu→Val), hemoglobin C (HbC; β6Glu→Lys), sickle/β⁺ thalassemia (HbSB⁺), and sickle/β⁰ thalassemia (HbSB⁰) (11, 12). Sickle cell anemia (SCA) is the most predominant form that occurs due to the substitution of T→A in the 6th codon of the β-globin gene, resulting in the replacement of hydrophilic glutamic acid (Glu) by hydrophobic valine (Val) residue. Intraerythrocytic deoxygenation of sickle hemoglobin (HbS) tetramers in a hypoxic environment, particularly near the tissues with high demand for oxygen, exposes the hydrophobic groups present on the HbS tetramers. This results in hydrophobic interactions between the motifs of the HbS tetramers leading to polymerization. Polymers of HbS tetramers greatly alter the morphology of the RBCs, causing loss of flexibility, membrane deformities, sickling, premature hemolysis, and impaired rheology. Sickled RBCs aggregate with the endothelial cells, platelets, and neutrophils to produce a mass of cells that causes vaso-occlusion-associated organ damage (13). Hydroxyurea (HU) was shown to be clinically effective in SCD by inducing the expression of γ-globin, and the response is highly correlated with the presence of XmnI polymorphism (14). Nonresponders of HU require red cell transfusions to increase blood oxygenation and lower blood viscosity. The exchange of RBCs during transfusions also reduces the concentration of sickled RBCs. Individuals heterozygous for the sickle mutation show signs and symptoms of sickle cell anemia on exposure to unfavorable conditions such as hypoxia, dehydration, hypothermia, hyperthermia, and increased sympathetic tone that induces sickling (15).

HEMOGLOBIN SWITCHING

Hemoglobin switching is the phenomenon of the transitioning of the physical and functional properties of hemoglobin during the development of an individual. The switching is associated with the change in the expression pattern of β-like globin genes present on chromosome 11, and it is developmentally regulated by various transcription and epigenetic factors (16). In the initial phase of the first trimester of pregnancy, embryonic globin (ε) chains in the β-globin cluster are highly expressed in the erythroid cells derived from the primitive yolk sac. The first hemoglobin switch involves a transition in the expression of ε-globin to γ-globin derived from fetal liver erythroid cells. The dimers of γ-globin produced from chromosome 11 combine with the dimers of α-globin produced from chromosome 16 to form a heterodimer called fetal hemoglobin (HbF). The second switch, also called the fetal-to-adult hemoglobin switch, is marked by a downregulation of the γ-globin chains with an upregulation in the expression of β-globin chains, which takes place around the time of birth. This results in the production of adult hemoglobin (HbA), composed of dimers of α-globin and β-globin subunits (α2β2), and HbA2, composed of dimers of α-globin and δ-globin subunits (α2δ2) (17) (Fig. 1). Hemoglobin switching is regulated by autonomous stage-specific repression or activation of β-like globin genes by various transcription and epigenetic factors. Also, the locus control region (LCR) present at the 5′ end of the β-globin cluster increases the expression of stage-specific globin genes severalfold (16). Earlier, it was speculated that the sequential expression of globin genes is mediated by long-range LCR-promoter interactions, but murine LCR knockout studies revealed the dispensable role of LCR in hemoglobin switching (18).

FIG 1.

Human β-like globin gene regulation during development. The two monomers of α-like globin expressed from chromosome 16 pair with two monomers of β-like globin to form a functional hemoglobin tetramer. The components of the hemoglobin change during development. The embryonic (ε) part of β-like globin is expressed during the first 6 weeks of gestation and forms embryonic hemoglobin by pairing with α-globin. After 6 weeks, ε-globin expression declines and is replaced by γ-globin to form fetal hemoglobin. The final hemoglobin switching takes place a few weeks before birth when the γ-globin expression is replaced by β-globin to form adult hemoglobin.

BENEFICIAL MUTATIONS ALTERING DEVELOPMENTALLY REGULATED EXPRESSION

Hereditary persistence of fetal hemoglobin (HPFH) is a benign condition resulting in the persistence of fetal γ-globin into adulthood. The steadfast expression of fetal γ-globin is due to large deletions that are downstream of the A_γ-globin gene in the β-globin cluster (deletional HPFH mutations) and to mutations in the promoters of γ-globin genes (nondeletional HPFH mutations). The coinheritance of SCD or thalassemia mutations with HPFH mutations ameliorates disease severity by the continuous expression of γ-globin compensating for the decreased or defective β-globin chains (19).

Deletional HPFH mutations.

With the insights acquired from hemoglobin gene ontogeny and genome-wide association studies, deletional HPFH is characterized by the deletion of genomic sequences downstream of the A_γ-globin gene comprising ψβ-, δ-, and β-globin genes. The deletions vary in size, with the smallest deletion of 13 kb in Sicilian HPFH to as much as 85 kb in Black HPFH. To date, at least nine HPFH deletions and the associated HbF levels have been reported (Fig. 2) (Table 1).

FIG 2.

Deletional HPFH. Schematic diagram of large deletions in the β-globin cluster responsible for persistent fetal hemoglobin production in adulthood. Genomic coordinates indicate the location of β-LCR and β-like globin genes present in chromosome 11. The gray bars mark the sizes of different HPFH deletions in the β-globin cluster, and the vertical bars correspond to specific β-like globin genes. HPFH deletions ranging from 12.9 kb to 84.9 kb delete the A_γ gene, δ-globin gene, β-globin gene, and 3′ β enhancer.

TABLE 1.

Fetal hemoglobin levels in individuals homozygous and heterozygous for HPFH deletions with or without coexisting thalassemia mutations

HPFH type	Deletion size (kb)	Deleted globins	% of HbF in HPFH+/−	Aγ (%)	Gγ (%)	Hb (g/dl)	Reference(s)
HPFH-1 (Black)	84.9	δ, β	21.7–27.9				117–124
HPFH-2 (Ghanaian)	83.6	Pseudo β, δ, β	21.6–27.2		27.5–37.1		121, 125, 126
			100 (HPFH+/+)		45–60	15.5–18
HPFH-3 (Indian)	47.7	Pseudo β, δ, β	21.6–23.6		65.8–72.8		121, 127, 128
HPFH-4 (Italian)	40	δ, β	21–30		31–39	11.8–16.6	129
HPFH-5 (Sicilian)	12.9	δ, β	16–20	83–85		13.6–14	130
HPFH-6 (Southeast Asian)	27	δ, β	15.1–30.4			9.2–14.4	131
			65.6 (HPFH/Thal+/−)			8.9
HPFH-7 (Kenyan)	22.7	Aγ, pseudo β, δ, β	11.1–12.6		96–98	12.4–12.7	132, 133
French HPFH	20	δ, β	35.7			14.7	134
Algerian HPFH	24	δ, β	41			10	134

It is interesting how some large deletions in the β-globin cluster can result in an HPFH phenotype and a few result in δβ-thalassemia. Mapping the breakpoints in both phenotypes has resulted in the identification of a 3.5-kb regulatory region that is crucial for silencing HbF expression. This regulatory region is observed to be the segregating factor between HPFH and δβ-thalassemia. The regulatory region is intact in δβ-thalassemia deletions causing mild increase in HbF levels and deleted in HPFH phenotypes resulting in robust levels of HbF. Initial experiments showed that the regulatory region harbors the binding site for BCL11A (20). However, high-resolution chromatin immunoprecipitation (ChIP) experiments showed that BCL11A does not bind to the region (21). Gene editing-mediated deletion of the 3.5-kb or the 1.7-kb putative repressor region (PRR) present in the γδ intergenic region did not show an expected increase in γ-globin levels. This indicates that the induction of HbF by HPFH deletions is due to the disruption of both the PRR and downstream regions of the β-globin gene (22). Genetic deletion of a 13.6-kb region in HUDEP-2 cell lines, which recapitulated a robust HbF phenotype, showed that the 3′ β-globin enhancer elements present downstream of the β-globin gene are juxtaposed to the γ-globin gene and also showed that γ-globin promoters are hyperacetylated to allow the recruitment of transcription activators (23). The open chromatin framework promotes the interaction of the LCR with γ-globin promoter and increases γ-globin expression severalfold. However, the 3′ β-globin enhancer elements downstream of the β-globin gene are also excised in many of the HPFH deletions, suggesting the mechanism of HbF reactivation may be different for different HPFH deletions.

Nondeletional HPFH mutations.

The persistence of fetal hemoglobin expression in adulthood as a result of beneficial point mutations or short deletions in the γ-globin promoter is called nondeletional HPFH. Nondeletional HPFH activates HbF either by mutations that disrupt the binding sites of γ-globin repressors or those that favor the binding of transcriptional activators of γ-globin. At least 10 different mutations are shown to reactivate HbF to clinically beneficial levels. The mutations are grouped around the −115 cluster, −175 region, and −200 cluster in the γ-globin promoter (Table 2).

TABLE 2.

Fetal hemoglobin levels in homozygous and heterozygous individuals for nondeletional HPFH mutations with or without coexisting SCD or thalassemia mutations

Site	Mutation	% of HbF in HPFH+/−	Aγ (%)	Gγ (%)	Hb (g/dl)	Reference(s)	Mechanism
Mutations disrupting repressor binding site
−114	C→T (Aγ)	2.9–4.7	90.3–91.4	8.6–9.7		27	Disruption of BCL11A binding site
	C →T (Gγ)	11–14		86–92		24
	C→G (Gγ)	8.6		90	14.2	26
	C→A (Gγ)	0.6–3.5				25
		3.7–11.2 (HPFH+ Thal+/−)
13-bp deletion	−102→−114 (Aγ)	30.1–31.8 (HPFH+ HbS+/−)	81.2–90.1	18.7–19.8	9.3–13.4	112
−117	G→A (Aγ)	10–20			11.3–15.2	28, 29
		26–43 (HPFH+ Thal+/−)			8.2–11.3
−195	C→G (Aγ)	4.5–7	86.1–90.7	13.9	15	30	Disruption of LRF binding site
−196	C→T (Aγ)	12–16				113
		38–40 (HPFH+ Thal+/−)
	C→T (Gγ)	8.6			14.3	31
−197	C→T (Aγ)	5.9–6			12.3–15.5	32
−201	C→T (Aγ)	10.2			14.5	31
−202	C→T (Aγ)	28.6 (HPFH+ HbS+/−)	68.4	31.6	11.1	33
	C→G (Gγ)	19.9–23.5 (HPFH+ HbS+/−)				35
Mutations creating activator binding site
−175	T→C (Aγ)	36.7–38.5		18.1–37.9		114, 115	Recruitment of TAL1, GATA1, LMO2, LDB1
		37.3 (HPFH+ HbC+/−)		19	15
		40.4 (HPFH+ HbS+/−)	29.9	34.6	14
	T→C (Gγ)	17–22.1			11.7–13	39–41
		29.7 (HPFH+ HbS+/−)		100	12
−198	T→C (Aγ)	3.5–14.2			12.1–16.2	36, 37, 116	Recruitment of KLF1
		18.3–21.4 (HPFH+/+)	90–95

BCL11A, one of the repressors of γ-globin, binds to the −115 cluster. Several beneficial mutations such as −114 C→T/G/A, −117 G→A (24–27), and a 13-bp deletion from −102 to −114 disrupt the binding site of BCL11A, resulting in derepression of γ-globin (28, 29). The −200 cluster recruits the fetal globin repressor LRF, and the naturally occurring beneficial point mutations such as −195 C→G/T, −196 C→T, −197 C→T, 201 C→T, and 202 C→T/G disrupt the LRF binding sites and derepresses γ-globin (30–35). Few point mutations create novel binding sites for the transcription activators in the γ-globin promoter, such as −175 T→C, which recruits T-cell acute lymphocytic leukemia protein 1 (TAL1), GATA1, and Lim domain-binding protein 1 (LDB1), and −198 T→C (36, 37), which recruits Krueppel-like factor 1 (KLF1) and, in turn, promotes LCR interactions for fetal globin activation (38–41) (Fig. 3).

FIG 3.

Nondeletional HPFH. Schematic diagram of the γ-globin promoter harboring mutations/small deletions in individuals with persistent hemoglobin production. Three different areas of beneficial mutations are shown: clusters located at −200 and −115 and a single point mutation at the −175 nucleotide from the transcription start site. The small deletions and indels in the −115 and −200 cluster disrupt the binding domains of BCL11A and LRF, respectively. The point mutation in the −175 nucleotide recruit transcription activator TAL1, GATA1, and LDB1, whereas −198 T→C recruits transcription activator KLF1.

Genetic polymorphism involved in HbF activation.

Few individuals present with modest increases in HbF levels, even in the absence of γ-globin-inducing deletional or nondeletional HPFH mutations. Genome-wide association studies (GWAS) revealed three major quantitative trait loci (QTL) which account for 20% to 50% of phenotypic differences in HbF levels among individuals of different ethnic groups. These include chromosome 11p16 (chr.11p16) XmnI-G_γ, chr.6q23 HBS1L-MYB intergenic region, and chr.2p15 BCL11A (42–44).

The XmnI (−158 C→T) polymorphism present in the γ-globin promoter is common among people from different ethnic groups, with a frequency as high as 32% to 35%, and it increases the expression of G_γ under erythropoietic stress conditions such as SCD and β-thalassemia, resulting in decreased disease severity and frequency of blood transfusion (45, 46). However, the molecular mechanisms involved in increased HbF levels are still unknown (47).

The HBS1L-MYB intergenic polymorphism (HMIP) locus present in chr.6q23 is located 79 kb from the HBS1L gene and 45 kb upstream of the MYB gene, distributed as three blocks named HMIP1, HMIP2, and HMIP3. Two single nucleotide polymorphisms (SNPs) present in HMIP2 are strongly associated with increased HbF levels. The HbF levels in healthy individuals homozygous for SNPs range from 1.1% to 3%, and in heterozygous beta-thalassemic patients, the HbF levels range from 10% to 24% (48). The SNPs present in the erythroid cell-specific enhancer located in the intergenic region hinder the binding of LDB1, GATA1, TAL1, and KLF1. This results in decreased erythroid cell-specific enhancer interactions with the MYB promoter. Downregulation of MYB decreases MYB-induced expression of BCL11A and thus increases HbF synthesis (49).

Few SNPs associated with high HbF levels were also identified in intron 2 of BCL11A (42). Intron 2 consists of three DNase I-hypersensitive sites, termed h+55, h+58, and h+62 based on the distance from the transcription start site. SNPs in h+62 disrupt the binding site of transcription factors such as GATA1 and TAL1, leading to a reduction in the expression of BCL11A only in the erythroid lineage increasing the HbF levels (50).

THERAPEUTIC STRATEGIES FOR HbF REACTIVATION

Pharmacological induction of fetal hemoglobin.

Silencing of γ-globin genes by methylation plays a critical role in transition to β-globin (adult hemoglobin) from γ-globin (fetal hemoglobin), resulting in fetal-to-adult hemoglobin switching. The clinical observations in SCD and thalassemia patients have indicated that preventing the switching of HbF to HbA or reswitching of HbA to HbF can ameliorate the disease severity. Thereby, most of the therapeutic approaches focus mainly on fetal hemoglobin induction.

(i) 5-Azacytidine.

Nucleoside analog 5-azacytidine is a potent demethylating agent shown to inhibit the methylation pattern linked with the differentiation of mouse embryonic cells (51). The induction of γ-globin by 5-azacytidine was also demonstrated in the erythroid cells of anemic adult baboons (52). Treatment with 5-azacytidine in thalassemia patients ameliorated the globin chain imbalance by increasing peripheral blood fetal hemoglobin levels (53). It also decreased HbS polymerization in SCD patients by γ-globin reactivation (54). Irrespective of the success laid by 5-azacytidine for the treatment of β hemoglobinopathies, its cytotoxic effects are a major concern (55).

(ii) Hydroxyurea.

Hydroxyurea (HU), which initially was a drug of choice for solid tumors and leukemia due to its antineoplastic activity, also showed reactivation of fetal hemoglobin in an anemic monkey without altering much of the methylation pattern in the γ-globin promoter (56, 57) The SCD patients treated with HU showed increased fetal erythrocytes without any major marrow toxicity (58). In thalassemia patients, HU treatment produced a 33% increase in HbF levels with a decrease in circulating reticulocytes and improved balance in the α/β-like globin chain ratio (59). However, SCA and thalassemia patients respond differently to different pharmacological drugs that stimulate fetal globin expression. SCA patients respond better to hydroxyurea than thalassemia major patients (60). The proposed model for HbF reactivation by HU lies in the activation of p38 mitogen-activated protein kinase (MAPK) by the cellular stress sensors through a series of proteins involved in the signaling (61). However, the exact mechanism by which HU activates γ-globin is not clearly understood.

(iii) Short-chain fatty acid derivatives: butyrate.

A systemic increase in the levels of α-amino n-butyric acid was noticed in fetuses developing in the hyperinsulinemia environment of their insulin-dependent diabetic mothers (62). The fetuses also showed delayed fetal-to-adult hemoglobin switching. Consistent with this, fetal γ-globin is activated in butyrate-treated erythroid progenitors derived from the cord blood of normal infants and infants of insulin-dependent diabetic mothers (63). In vivo studies also showed a delayed fetal-to-adult globin switching. The adult globin switch is halted when butyrate is administered in the early phase of switching (64). In thalassemic patients, butyrate treatment increased the hemoglobin levels from 4.7 to 10.2 g/dl with minimal side effects (65). However, the treatment not only induces the expression of γ-globin chains but also increases the expression of the α-globin chain in thalassemia (66). Therefore, the beneficiary role of butyrate in restoring the α/β-like globin balance is considerably diminished due to an increase in the α-globin chains.

(iv) Pomalidomide.

Pomalidomide, a thalidomide analog, is an FDA-approved immunomodulatory drug to decrease or eliminate the need for blood transfusions in patients with multiple myeloma and myelodysplastic syndromes by restoring erythropoiesis. In vitro erythroid differentiation of hematopoietic stem and progenitor cells (HSPCs) from normal donors or SCD patients with pomalidomide showed increased proliferation of early stage erythroid progenitors along with an increase in fetal hemoglobin and total hemoglobin levels (67). The increase in HbF levels was 2-fold higher than with 5-azacytidine, HU, and butyrate. In addition, pomalidomide treatment decreased the number of apoptotic erythroid progenitor cells compared to that with known HbF inducers. In vivo efficacy of pomalidomide in sickle cell mice showed an increase in HbF from 6.24% to 9.51% with increased reticulocyte counts by promoting erythropoiesis (68). The transcriptional reprogramming of adult erythroid progenitors to fetal-like erythroid progenitors by altering the expression of γ-globin repressors such as BCL11A and SOX6 is observed to be the mechanism for HbF reactivation (69).

(v) Methyltransferase inhibitors.

Altering the epigenetic pattern and the chromatin accessibility of γ-globin promoters is one of the exciting strategies for γ-globin reactivation. Euchromatic histone-lysine N-methyltransferase 1 and 2 (EHMT1/GLP and EHMT2/G9a) catalyzes the methylation of 9th lysine residue on the histone 3 protein (H3K9Me2), which is a mark of transcriptional repression of its associated genes (70). Treatment of adult human erythroid cells with UNC0638, a selective inhibitor of G9a methyl transferase, induces HbF reactivation by facilitating the occupancy of LDB1, resulting in LCR interaction with γ-globin promoters. Even though UNC0638 induces HbF synthesis with low toxicity in vitro, the poor pharmacokinetic properties hinder clinical translation (71). Genome-wide inhibition of G9a methyl transferase modifies chromatin organization and activates several repressed genes, thereby altering the physiological transcriptome.

(vi) Speckle-type POZ protein.

CRISPR-Cas9 screening based on protein domain identified a BTB domain protein called speckle-type POZ protein (SPOP), as a novel regulator of γ-globin. SPOP repress γ-globin by complexing with CUL3 ubiquitin ligase. Targeted disruption of the SPOP interacting domain and short hairpin RNA (shRNA)-mediated depletion increases γ-globin levels in HUDEP-2 cell lines and primary erythroid cells. Overexpression of a dominant negative form of SPOP (Y87N) increases γ-globin levels, and the activation is independent of BCL11A and LRF. Depletion of SPOP along with pomalidomide treatment results in a 2-fold increase in γ-globin due to activation of HbF through two independent pathways. Small molecules targeting the interaction of SPOP-CUL3 ligase can be an effective agent for pharmacological induction (72).

(vii) Heme-regulated inhibitor.

Domain-focused CRISPR-Cas9 screening revealed heme-regulated inhibitor (HRI) as a repressor of γ-globin. CRISPR-Cas9-mediated disruption of HRI activates γ-globin in HUDEP-2 cell lines. In addition, shRNA-mediated knock down of HRI in CD34⁺-derived primary erythroid cell increases γ-globin by depletion of BCL11A, indicating that BCL11A acts as an effector of HRI (73). Combining HRI with HbF-inducing pharmacological agents might synergistically provide therapeutic benefits.

Galvanizing fetal globin for gene therapy.

The arrival of facile genome editing tools such as clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (Cas9), zinc finger nucleases (ZFNs), and transcription activator-like effector nuclease (TALEN) are transforming gene therapy. After decades of sobering experience with the available transient therapy provided for treating hemoglobinopathies, genome editing tools heightened the hopes to have a safe and long-term cure for patients. Gene therapy for hemoglobinopathy has pros and cons. The positive aspect is the utilization of patient’s own HSPCs as the target for autologous ex vivo gene therapy, thus bypassing the limitation of finding an HLA-matched donor and the risk of graft-versus-host disease (GvHD) associated with allogeneic stem cell transplantation (Fig. 4). However, the concern is that the editing of HSPCs is a difficult process due to the toxicity associated with gene editing reagents and the difficulties in finding a suitable delivery system with high editing efficiency. Recently, many investigators have come up with advanced gene editing protocols with modifications using either lentivirus-based delivery or a ribonuclear protein (RNP)-based system. The prime goal is to target the cis and trans regulatory factors of γ-globin expression to reactivate the silenced fetal globin gene (Fig. 5).

FIG 4.

Autologous transplantation of gene-edited HSPCs. HSPCs were obtained by granulocyte colony-stimulating factor (G-CSF) and or plerixafor based mobilization and then enriched by immunoselection of CD34 markers. The HSPCs were manipulated by gene editing tools such as CRISPR-Cas9/ZFN/TALENs to correct the disease-causing mutations or activating γ-globin for the reversal of the disease phenotype. The gene-edited HSPCs which are transplanted in the patients are expected to engraft and produce erythroid cells with near-normal levels of hemoglobin.

FIG 5.

Therapeutic approaches for γ-globin reactivation. Several gene therapy strategies are shown that activate fetal γ-globin by creating HPFH mutations, BCL11A erythroid cell-specific enhancer disruption, shRNA-mediated knockdown of γ-globin repressors, forced LCR looping, and γ-globin supplementation. In SCD, the γ-globin chains replace sickle β-globin chains and decrease sickling and its associated complications. In thalassemia, γ-globin chains compensate for the loss of β-globin chains by forming functional fetal hemoglobin tetramers with free α chains and prevent free α-globin-mediated damage to the RBCs.

(i) Introducing deletional HPFH mutations for gene therapy.

Creating HPFH deletions in the HSPCs of homozygous sickle and thalassemia patients provides a potential therapeutic approach for the treatment of β-hemoglobinopathies. Advancement in gene editing technology paves the way for recreating such large deletions for HbF induction. Staphylococcus aureus Cas9 (SaCas9)-mediated excision of a 13-kb segment from the β-globin cluster encompassing δ- and β-globin genes displayed robust reactivation of γ-globin chains (74). Similarly, CRISPR-Cas9-mediated 13.6-kb deletion comprising the putative γ-δ intergenic region and δ- and β-globin genes in the β-globin cluster is associated with HbF activation in HUDEP-2 erythroid progenitor cell lines and in HSPCs. The targeted 13.6-kb deletion also reverses the sickling phenotype in erythroid cells derived from HSPCs of SCD patients. Chromosome conformation capture (3C) analysis showed interactions of the HS2 of LCR to the A_γ-globin promoters on 13.6-kb deletion. The reactivation of γ-globin is also due to the juxtaposition of the β-globin intronic enhancer and 3′ HS-1 to the γ-globin gene (23). While HPFH deletions are shown to reactivate γ-globin expression, the long-term persistence of HSPCs harboring such large deletions has yet to be demonstrated in a suitable animal model. The need of such experiments is emphasized by a finding which showed that the frequency of HSPCs with a shorter deletion of 13 bp declined posttransplantation. The reduction might be due to the higher editing rates in committed progenitor cells or to the decreased engraftment of long-term repopulating hematopoietic stem cells (LT-HSCs) that underwent double-strand break (DSB)-mediated DNA repair (75). Gene editing by sorting the LT-HSCs or transient suppression of p53 by coelectroporation of a dominant negative form of P53 (GSE56) mRNA during the editing process may increase the engraftment of HSPCs mimicking large HPFH deletions (76, 77).

(ii) Introducing nondeletional HPFH mutations.

The nondeletional HPFH mutations can be created in the γ-globin promoter by exploiting a homology-directed repair (HDR) donor template harboring the mutation to be introduced. Introduction of −175 T→C and −198 T→C (British HPFH) point mutations using CRISPR-Cas9 along with single-stranded oligodeoxynucleotides (ssODNs) are shown to reactivate HbF in HUDEP-2 cell lines. These studies also helped to understand the mechanism behind HbF reactivation. The −175 T→C mutation creates a binding motif for TAL1, and the γ-globin promoter-bound TAL1/LDB1 complex promotes the looping of LCR to the γ-globin promoter, resulting in the expression of fetal globin chains (78). The −198 T→C mutation in the γ-globin gene creates a de novo binding motif for transcription activator KLF1, which promotes LCR looping to γ-globin promoters (79). Similarly, the −113 A→G mutation creates a de novo binding motif for GATA1 (80). These HDR-based introductions of HPFH point mutations have yet to be demonstrated in the HSPCS. Such studies are challenging in the HSPCs due to poor HDR rates and donor-mediated toxicity. The HDR-based sickle mutation correction in the long-term repopulating HSPCs was observed to be low, highlighting the difficulties in using HDR-based approaches in HSPCs. However, a recent breakthrough study showed the correction of the sickle mutation at an efficiency of >25% both in vitro and in vivo, suggesting that similar protocols can be adopted for introducing point mutations at the γ-globin promoter (81, 82).

Exploiting the nonhomologous end joining (NHEJ)-mediated DNA repair pathway to disrupt the repressor binding motifs tends to be a simple and efficient approach for reactivating γ-globin. CRISPR-mediated editing of −102 to −114 in the γ-globin promoter disrupts the TGACCA motif that recruits transcription repressor BCL11A and results in robust reactivation of fetal globin. This approach recapitulates the naturally occurring 13-nucleotide (nt) HPFH deletion and reverses the sickling phenotype in erythroid cells derived from HSPCs of SCD patients (75, 83). The CRISPR-Cas9-mediated disruption of the bp −200 cluster region in the γ-globin gene promoter activates the fetal globin by inhibiting the binding of γ-globin repressor LRF. Moreover, xenotransplanted HSPCs with guide RNA (gRNA) targeting the bp −200 cluster showed high indel efficiency without any impairment in the engraftment and differentiation potential of the HSPCs (84). Editas Medicine has performed a CRISPR-Cas9 screening of the entire β-globin cluster and identified novel targets in the γ-globin region for reactivating HbF (85).

The sequence homology between G_γ and A_γ results in simultaneous generation of double-stranded breaks leading to the excision of the 4.9-kb intergenic region. The consequence of such deletion is not known. A recent study reported that the 4.9-kb deletion in the γ-globin promoter is not deleterious, and the deletion-mediated activation of HbF is driven by the A_γ-globin gene (86).

(iii) Targeting γ-globin repressors for fetal globin reactivation.

BCL11A is one of the major transcriptional repressors of γ-globin. Targeting BCL11A is an emerging therapeutic approach for HbF activation. However, the essential role of BCL11A in B-lymphopoiesis and maintenance of HSPCs emphasizes BCL11A as an indispensable factor for hematopoiesis (87, 88). The erythroid lineage-specific BCL11A depletion strategy has evolved to preserve the HbF regulation-independent functions of BCL11A.

(a) BCL11A erythroid cell-specific enhancer disruption. CRISPR-Cas9-mediated in situ saturating mutagenesis analysis has provided us the critical insights on the specific sequences dictating the expression of BCL11A in erythroid cells. The h+58 DNase I hypersensitive site in intron 2 of the BCL11A locus precisely modulated the BCL11A expression in erythroid cells (89). ZFN-mediated disruption of the BCL11A coding region (exon 2) and the GATA motif in the h+58 DNase I hypersensitive site resulted in HbF reactivation. However, enucleation rate and the engraftment in mice were severely affected by BCL11A exon disruption but not by erythroid enhancer disruption. The enhancer editing in the HSPCs from a thalassemia patient improved globin chain balance and promoted effective erythropoiesis (90). CRISPR-based BCL11A erythroid-specific enhancer editing reversed the sickling phenotype in SCD patient cells (91). A recent study by Editas Medicine revealed that the BCL11A enhancer-edited HSPCs showed erythroid cell-specific defects in an immunocompromised NBSGW mouse model (92). The preliminary data from the Sangamo Therapeutics phase 1/2 clinical trial ST-400 which uses ZFN to disrupt autologous BCL11A erythroid cell-specific enhancer demonstrated the activation of fetal globin in the peripheral blood of three patients from the fourth week posttransplantation (Table 3). However, a rapid reduction in fetal hemoglobin was observed in all three patients in the subsequent weeks; thus, the patients had to undergo red blood cell transfusion after 7 to 8 weeks posttransplantation (93). A long-term follow-up will provide more clarity on the sustained HbF production in these patients. The CRISPR Therapeutics and Vertex Pharmaceuticals trial using Cas9 to disrupt the BCL11A enhancer sequences demonstrated the safety and efficacy of the approach in a patient with transfusion-dependent β-thalassemia (TDT) and another patient with SCD. After 9 months of transplantation, the patient with TDT attained fetal hemoglobin levels of 10.1 g/dl and total hemoglobin of 11.9 g/dl and was transfusion independent. The SCD patient who had a history of frequent vaso-occlusive crises displayed total hemoglobin levels of 11.3 g/dl without any reoccurrence of vaso-occlusive crises postinfusion of the drug product on a 4-month follow-up (94).

TABLE 3.

Clinical trials on γ-globin gene addition and gene editing-mediated γ-globin reactivation strategies

Sponsor	Identifier	Intervention	Disease	Phase	Estimated no. of participants	Study duration
Aruvant Sciences, Cincinnati Children’s Hospital Medical Center	NCT02186418	γ-Globin lentivirus vector (ARU-1801)	Sickle cell disease	1/2	10	July 2014–June 2035
Boston Children’s Hospital	NCT03282656	BCL11A shRNA lentivirus vector	Sickle cell disease	1	15	February 2018–February 2024
Sangamo Therapeutics	NCT03432364	ZFN-mediated BCL11A enhancer disruption (ST-400)	Transfusion-dependent β-thalassemia	1/2	6	March 2018–March 2023
Vertex Pharmaceuticals Inc. and CRISPR Therapeutics	NCT03655678	CRISPR-Cas9-mediated BCL11A enhancer disruption (CTX001)	Transfusion-dependent β-thalassemia	1/2	45	September 2018–May 2022
	NCT03745287		Sickle cell disease	1/2	45	November 2018–May 2022

(b) Downregulation of γ-globin repressors. shRNA-mediated downregulation of BCL11A has evolved as a clinically advanced strategy for HbF reactivation. The knockdown of BCL11A by a microRNA-adapted shRNA (shRNA miR) approach resulted in an increase in the HbF reactivation. However, it also resulted in the impairment of long-term engraftment potential of mouse and human HSPCs (95). The adverse effects associated with the knockdown of BCL11A led to the development of lineage-specific shRNA miR. Erythroid lineage-specific expression of shRNA miR using lineage-restricted polymerase II promoter activated γ-globin without compromising the engraftment potential of the HSPCs. The transplantation of shRNA transduced murine HSPCs into Berkeley sickle mice ameliorated the features of SCD and its pathophysiological markers (96). This study is currently in a phase 1 clinical trial, and recent data from 3 subjects with severe SCD showed that the percentage of fetal globin ranges between 23.7% and 40.7%, with F⁺ cells constituting 67.7% to 76.4% of the RBCs in the peripheral blood and the HbF content of each F⁺ cell ranging between 10.3 and 17.2 pg (97). The knockdown of LRF by shRNA in HSPCs increased the synthesis of γ-globin chains with loss of β-globin chains and with the defects associated with erythroid differentiation. Even though the γ-globin activation met the therapeutic levels required to ameliorate the β-globin disorders, the translational potential of this approach is questioned due to the delay in erythroid differentiation, emphasizing the crucial role played by LRF in erythroid maturation (98).

Supplementing γ-globin by lentiviral vectors.

The potential role of γ-globin chains in ameliorating the disease severity of thalassemia and SCD is due to its functional similarities with the β-globin chains. Gene addition strategies that supplement γ-globin to compensate for the decreased or defective β-globin chains have revolutionized gene therapy for β-hemoglobinopathies. Earlier studies have demonstrated a robust production of γ-globin chains in murine and human erythroid cells by the expression of the γ-globin gene using retroviral and adeno-associated viral vectors driven by the α- and β-globin regulatory elements. The amelioration of disease phenotype in a murine sickle cell disease model was demonstrated by the γ-globin lentiviral vector driven by the 130-bp promoter of the β-globin gene and 3.1-kb regulatory sequences (99, 100). Sustained production of HbF by the β-globin regulatory sequence was observed in the erythroid cells of sickle mice, and the mice were devoid of phenotypic abnormalities and complications associated with SCA. Amelioration of SCA occurred when HbF surpassed 10% and F⁺ cells constituted approximately two-thirds of circulating erythrocytes, with γ-globin chains occupying one-third of the total globin chains (101). Preliminary results from a phase 1/2 clinical trial conducted by Aruvant Sciences in two SCA patients transduced with modified γ-globin lentivirus vector showed 20% of vector-derived HbF and 10.6 g/dl of total Hb within 1 year of transplantation (102). The data from long-term follow-up are awaited to confirm the persistence of gene-modified HSPCs.

Base editor-mediated γ-globin reactivation.

The extensive nuclease activity of Cas9 possess the risks for off-target gene modification. The attempts to create beneficial nondeletional HPFH mutations in the γ-globin promoter by CRISPR-Cas9 are associated with a 4.9-kb deletion removing the A_γ coding region. Better controlled nuclease activity could prevent such unwarranted genomic deletions. Base editors, developed by tethering the cytidine deaminase or adenosine deaminase to Cas9 nickase, enabled irreversible and targeted transition conversions without inducing double-stranded breaks in the genome (103, 104). Beam Therapeutics screened the γ-globin promoters with cytosine and adenine base editors and identified three regions effectively contributing to HbF elevation by base conversions that mimic the naturally occurring HPFH mutations (105). Impressively, the desired base conversions reached 94%, with an increase in the γ-globin chain of up to 63%. Another elegant and extensive γ-globin promoter screening by adenosine and cytosine base editors in HUDEP-2 cell lines identified several novel HPFH-like mutations (106). A recent study demonstrated the activation of γ-globin using hA3A-BE3 base editors by targeting the cytosine bases in the −114 and −115 region of the γ-globin promoter (107). In addition to this, the +58 region in the BCL11A erythroid cell-specific enhancer was extensively screened using the hA3A-BE3 base editors, with increased activation of HbF upon base editing of the core regions. The activation of HbF was similar to the levels of HbF induced by the Cas9-mediated indels. Base editing also activated HbF in erythroid cells derived from SCD and thalassemia patient HSPCs with efficient editing in long-term repopulating HSPCs (108). Base modification of the SCA mutation in HSPCs to a naturally occurring asymptomatic variant has been reported with an editing efficiency of up to 40%. As the base editors were designed only to create transition mutations, the correction of A→T transversion from sickle (CAC) to the wild type (CTC) was not possible. However, the conversion of the adenine residue of valine led to the production of a naturally occurring variant called Hb G-Makassar (CAC→CGC), which results in normal hematological parameters (105). Base editing of noncoding regions of the genome (promoters, introns, and intergenic regions) can be a potential strategy for HbF activation. However, the application of base editors for correcting disease-causing mutations is greatly narrowed due to the possible deleterious effects caused by bystander edits and possible off-target base conversions in the RNA.

Forced LCR looping.

LCR is a long-range cis-regulatory element that enhances the expression of globin chains through promoter interactions. The chromatin looping of LCR with the promoter is mediated by several transcription factors, including GATA1, KLF1, FOG1, and LIM domain-binding protein 1 (LDB1). The artificial zinc finger (ZF) β-globin promoter DNA-binding peptide tethered with the self-association (SA) domain of LDB1 induced the expression of the β-globin gene in murine erythroid progenitors, providing the proof of concept for increasing the expression of a gene of interest by inducing enhancer-promoter interactions through forced chromatin looping (109). Similarly, The ZF-SA complex targeting both the promoters of the human γ-globin gene induced the interaction of LCR with the γ-globin promoters, increasing HbF levels. This strategy decreased sickle hemoglobin levels in the erythroid cells derived from HSPCs of the SCD patient (110, 111). This strategy uses viral transduction of a ZF-SA construct in the HSPCs. The long-term consequences of overexpression of the SA domain in HSPCs have yet to be explored.

CONCLUSION

The mechanistic understanding on the factors involved in temporal regulation of globin genes and the clinical demonstrations on the beneficial effects of γ-globin in diluting the concentration of HbS and rescuing erythropoiesis defects have allowed us to use γ-globin activation as a therapeutic approach for β-hemoglobinopathies. The one-time gene therapy approaches for γ-globin activation are emerging as an alternative to transient pharmacological inductions. The γ-globin activation strategy possesses several advantages over β-globin manipulation. The γ-globin repressor downregulation constructs are smaller than β-globin constructs, allowing high-titer lentiviral production and efficient transduction and thus reducing the cost of production of gene-modified HSPCs. Similarly, gene editing-mediated γ-globin activation can be achieved without the use of homologous sequence donor, which is a prerequisite for β-globin correction, and thus reduces the toxicity and the cost associated with the donors. All these factors make γ-globin activation the preferable approach for gene therapy over the β-globin modification approaches. While the reports on the long-term persistence of globin-modified HSPCs and sustained expression of γ-globin are encouraging, we still lack a detailed understanding on the minimum frequency of gene-modified HSPCs and F cells required for the correction of disease pathophysiology. The risks associated in harvesting HSPCs from inflamed bone marrow, the toxicities associated with conditioning, and the infection risks during the recovery of hematopoiesis make gene therapy a complex procedure. Also, gene therapy cannot be repeated if the first attempt fails in a given patient. As younger patients are a prime target for gene therapy, the above-described factors make it a risky option. In the current form, gene therapy might help a small slice of the population while being risky and expensive. Considering that hemoglobinopathy patients are more prevalent in economically weak countries, gene therapy must evolve to match the feasible drug treatment options to help a broader range of patients.