Above the cut: HbF induction by multiplex prime editing Free

In this issue of Blood, Chalumeau et al¹ fine-tune the application of precise prime-editing technology for high-level induction of γ-globin in edited cells with the goal of improving curative therapies for β-hemoglobinopathies, such as β-thalassemia and sickle cell disease (SCD).

The fetal globin genes HBG1 and HBG2 (HBG1/2) encode γ-globin, which takes the place of β-globin in fetal hemoglobin (HbF) and prevents sickling, making its induction the mainstay of recent advanced therapy development for β-hemoglobinopathies. Although HbF is naturally silenced after birth, its permanent reactivation by genome editing in hematopoietic stem and progenitor cells (HSPCs) is a promising strategy for curative treatment of patients with β-hemoglobinopathy and is the mechanism of action for Casgevy (exagamglogene autotemcel). Casgevy, like other first-generation DNA editing approaches, transiently introduces targeted double-strand breaks (DSBs) in the genome as the basis for permanent genome modification and HbF activation. Although treatment has proven to be safe to date,² DSB-dependent editors inherently introduce variable insertions and deletions at the intended, on-target site, facilitate recombination events with spontaneous DSB events elsewhere in the genome, and may induce additional reactive DSB events at sequence-similar off-target sites.

Newer generations of genome editors, such as base editors (BEs) and the more versatile prime editors (PEs),³ are usually DSB-independent and thus inherently safer, particularly for genome modification with several concurrent editing sites.⁴ Such multiplex combinatorial editing has emerged lately as an effective means of inducing γ-globin and, for base editing, uses separate editors for each base change. In contrast, PEs employ a prime-editing guide RNA (pegRNA), which, in addition to the target-recognition site and other features, encodes a repair template that may contain multiple proximal base changes or insertions and deletions for a single editor. Aiming to enhance HbF induction based on multiple activating modifications in the HBG1/2 promoters and to increase the predictability and safety of editing outcomes when compared with first-generation editors and BEs, Chalumeau et al thus sought to optimize multiplex prime editing as therapy for β-hemoglobinopathies.

Employing K562 human erythroid cells for optimization and HSPCs for the assessment of HbF induction, the authors demonstrated that several factors enhance the multiplex editing capability of PEs in the HBG1/2 promoters. One such factor is the proximity of the intended mutations to the nick site and another factor is the extension of the pegRNA repair template to lower the impact of pegRNA 3ʹ flap resection. Likewise, template sections that encode small deletions were stabilized for correct target annealing by elongation of the pegRNA. Finally, inclusion of the tevopreQ1 motif at the pegRNA 3ʹ end was used to create an altogether stabilized enhanced pegRNA.⁵ 

Evaluation of multiplex edits included both the removal of binding sites for repressors, such as BCL11A and ZBTB7A/LRF, and the creation of binding sites for activators, such as KLF1, TAL1, and GATA1. In K562 cells, precision editing reached 50%, whereas for SCD HSPCs, the highest efficiency achieved was limited to 7%. The latter was moreover obtained with a nuclease-based PE, PEnmax, which unusually induces DSBs instead of DNA nicks as the basis for prime editing. Employed as a means of boosting performance for ineffective pegRNAs, PEnmax as a DSB-based editor therefore has an increased risk for indel and recombination events. One such recombination event is the excision of the 4.9-kb sequence that separates the HBG1 and HBG2 genes and was readily detectable in both K562 cells and HSPCs using nickase-based PEmax and even more so using DSB-based PEnmax. Such excision events varied reproducibly for different primary samples and did not correlate with the same-sample efficiency of desired edits, raising the hope that, under suitable conditions, the desired edits could be elevated without simultaneous increases in recombination events. Off-target events were assessed in K562 cells using genome-wide, unbiased identification of DSBs enabled by sequencing,⁶ which revealed minimal off-target activity at sites with 3 or more mismatches.

At the functional level, edited HSPCs retained their growth and multilineage potential. Importantly, combinatorial application of cis regulatory modifications led to elevations in HbF induction, so that, for example, the creation of GATA1 and KLF1 binding sites, along with the disruption of the BCL11A site, led to significantly higher γ-globin expression than the individual edits. When compared with PEs in HSPCs, BEs had much higher editing efficiency but also had difficulties achieving consistent multiplex editing. Clonal analyses of edited HSPCs therefore showed HbF expression for PEs equal or superior to that observed for BEs. Although spontaneous in vivo enrichment in corrected erythroid cells in β-hemoglobinopathies⁷ may thus particularly favor multiplex PE-edited cells and reduce the number of corrected cells needed for successful therapy, the bulk PE editing efficiency in HSPCs remains a key limiting factor.

Improvements and further validation will therefore be needed to move multiplex prime editing from concept to clinic. Besides raising nick-based multiplex prime-editing efficiency in HSPCs to levels demonstrated elsewhere for simplex edits,⁸ greater robustness is needed for PE application, indicated by the high level of interpatient variability. This might be achieved by further optimization of pegRNA design, application of prime-editing enhancers, such as the small RNA-binding exonuclease protection factor La,⁹ and reducing unwanted deletion and recombination events with the application of inhibitors for components that facilitate undesired repair pathways, such as Polθ and DNA protein kinase.¹⁰ Finally, ex vivo–edited SCD HSPCs will need to be assessed in chimeric mouse models to ensure correction of clinically relevant long-term repopulating cells. This might be supported by alternative delivery modalities, such as by targeted nanoparticles, which would moreover ready multiplex prime editing for a shift from ex vivo to in vivo application.

Taken together, Chalumeau et al provide a convincing proof of concept of multiplex prime editing of the HBG1/2 promoters as a therapeutic strategy in β-hemoglobinopathies. Their study demonstrates that combinatorial prime editing of cis regulatory elements can achieve high levels of HbF expression in edited cells. Although its current implementation is still held back by efficiency and safety concerns, conceptually multiplex prime editing has the potential to emerge as the most powerful universal strategy for one-and-done therapy of β-hemoglobinopathies.

Conflict-of-interest disclosure: The author declares no competing financial interests.