Keeping fetal hemoglobin in the loop

Sickle cell disease (SCD) is caused by a point mutation in the adult type β-globin gene. The faulty β-globin chain triggers hemoglobin polymerization, promoting red blood cell sickling. Altered red blood cell shape causes occlusion of small blood vessels, leading to multi-organ damage and limiting life expectancy to 40–50 y of age. Although all SCD patients bear the same mutation, the severity of the clinical presentation varies among individuals. Milder courses of the disease are experienced by patients who continue to express in adult life elevated levels of the fetal form of hemoglobin, called γ-globin.¹ Since γ-globin has anti-sickling properties, reactivating its expression in SCD patients has been a major goal over the past decades.²

The human β-type globin genes, comprising one embryonic (ϵ), 2 fetal (^Gγ, ^Aγ) and 2 adult forms (δ, β) are arranged along the chromosome in the order in which they are expressed throughout development. The expression of these genes requires the locus control region (LCR), a powerful enhancer that physically contacts the globin gene promoters via chromatin looping in a developmental stage appropriate manner. LCR-promoter contacts require the transcription co-factor Ldb1, which associates with both regions and facilitates their interaction, presumably via self-association. Previous studies addressed the cause-effect relationship of LCR-promoter looping and transcription activation: artificial zinc finger DNA binding domains (ZF) designed to bind the β-globin promoter were fused to Ldb1 (ZF-Ldb1) and introduced into immature erythroblasts, which lack the LCR-promoter loop. ZF-Ldb1 was sufficient to trigger a looped LCR-promoter interaction and activate β-globin transcription,³ establishing forced chromatin looping as a potential strategy to induce specific genes within the β-globin locus.

In the most recent study, ZF-Ldb1 fusion proteins were targeted to the embryonic or fetal globin promoters in adult erythroid cells.⁴ This stimulated their interaction with the LCR and activated transcription with a concomitant reduction in adult β-globin expression, consistent with a mechanism by which the globin genes interact with the LCR in a mutually exclusive manner and compete for its activity. Of note, in cultured human adult erythroblasts the ZF-Ldb1 fusion protein targeting the γ-globin gene activated its expression to 85% of total β-type globin (γ-globin + β-globin) transcription, which is significantly higher than would be needed to alleviate symptoms in SCD patients.

A key question now is whether forced chromatin looping can be applied in a clinical setting. The first step in this direction involves the use of SCD mice that express only human globin genes, including the form that contains the sickle cell mutation.⁵ The approach will be to transduce haematopoietic stem cells (HSC) with ZF-Ldb1 fusions driven by an erythroid-specific promoter and transplant them into lethally irradiated host animals, followed by assessment of globin expression profiles and organ pathologies. If SCD can be ameliorated in mouse models, this might pave the way for studies in human patients.

How does the forced chromatin looping approach compare to alternative forms of SCD treatment? Correcting the SCD mutation in HSC using genome-editing methods would be the ideal solution since it would be permanent and not fraught with complications associated with transgene insertion. However, in order to be clinically useful, genome editing in HSC would have to be achieved with very high efficiency and little genotoxicity (potentially including mutagenesis and chromosomal translocations caused by genome editing tools). Additional genome editing approaches include interference with the fetal globin repressor Bcl11a.⁶ Increasing fetal hemoglobin production with a virally transduced γ-globin transgene is also an attractive strategy.⁷ This approach necessitates very high expression to compete with the mutant β-globin chain, requiring powerful enhancers and multiple viral copies per genome, both of which might have detrimental effects on the host genome. The forced looping approach described above circumvents some but not all of these limitations. ZF-Ldb1 mRNA expression levels required for high γ-globin induction are less than 1% of β-globin mRNA levels,⁴ thus allowing use of a weaker erythroid promoter and fewer viral copy numbers, lowering the potential for unwanted effects on the host cells. Second, activation of γ-globin is accompanied by reduced β-globin levels, further improving anti-sickling efficacy.

Several important questions about the basic science and clinical applicability of forced chromatin looping remain. For example, although it is well known that nuclear factors such as Ldb1 regulate higher order chromatin structure, the converse question, i.e., the impact of chromatin looping on nuclear factors, is largely unknown. How do newly formed chromatin loops at least partially overcome the effects of the γ-globin silencer complexes? From a clinical perspective it will be essential to examine possible off-target effects of ZF proteins. Although erythroid maturation and the expression of several erythroid genes examined in the Deng et al. 2014 study⁴ were not significantly impaired by the ZF-fusion proteins, genome-wide transcriptome analyses are required to address this question. In addition, the system needs to be optimized for ZF expression levels and target sequence recognition to maximize γ-globin induction and minimize off-target effects. Finally, it will be worth investigating whether forced looping can be used in combination with pharmacologic fetal globin inducers to improve therapeutic benefit while reducing potential toxicities.