Gene Editing Primed for Success

Mutations in key genes involved in the development or function of the retina result in inherited retinal diseases (IRDs). To date, over 325 genes and genetic loci have been identified as contributors to IRDs, with most of these genes having more than one identified mutation that is causative of disease. Adding to the complexity, different mutations in the same gene can lead to different clinical manifestations, such as mutations in RPE65, which can cause either autosomal recessive Leber congenital amaurosis (LCA), autosomal recessive retinitis pigmentosa (RP), or autosomal dominant RP.

Our ability to identify candidate genes outpaces our ability to develop molecularly targeted treatments for IRDs, and, given the aforementioned numbers of both genes and mutations, it will be a great challenge to develop small molecule therapies for each mutated protein. Thus, researchers must develop new therapeutic approaches for IRDs. This new approach is exemplified by the development of Luxturna (voretigene neparvovec) – the first FDA-approved gene augmentation therapy. Approved for the treatment of LCA, Luxturna was a triumph of translational research, providing proof of concept for gene augmentation therapy and sequencing-informed treatment. However, gene augmentation therapy is less likely to be able to treat diseases that act in a dominant manner, as expression of the wildtype gene may not be able to ameliorate the toxic effects of a dominant negative mutant protein. Another limitation of gene augmentation therapy is its inability to package large transgenes, such as USH2A, MYO7A, and ABCA4, as cargoes in gold standard viral vectors. In light of this, other therapeutic strategies – such as direct genome editing – need to be considered to address these unmet needs for IRD treatment.

Prime editing
 

First developed in the lab of David Liu at the Broad Institute of MIT and Harvard, prime editing (PE) is a novel third-generation CRISPR/Cas9 gene editor. PE offers many advantages over traditional, nuclease based CRISPR/Cas9 strategies. These include: the avoidance of double-stranded breaks, no dependence on mitosis and active cell cycling for efficient repair, and purer, more defined editing outcomes. In (relatively) simple terms, the process uses at least one prime editing guide RNA (pegRNA) to target a partially inactivated Cas9 protein to a defined location in the genome where a single-stranded DNA break is made. Next, a reverse transcriptase, which is fused to Cas9, installs an intended edit templated by the pegRNA by reverse transcribing the template into DNA, which is then incorporated into the genome. The mechanism of PE enables the correction of all 12 types of point mutations, as well as the reversion of small insertions and deletions.

Prime editor engineered virus-like particles
 

A major challenge in the application of PE and other genome editors is the need to safely and efficiently deliver genome editing components. The use of viral vectors, such as adeno-associated viruses and lentiviruses, come with significant drawbacks. When CRISPR/Cas9 components are expressed over a sustained period – such as from a cell that has been transduced by a virus – there is an increased risk of unintended off-target and bystander editing outcomes. There is also a low risk of viral genome insertion into the genome, either in a random (and potentially oncogenic location) or into the nicked DNA strand. An ideal genome editing delivery platform should deliver CRISPR/Cas9, PE, and guide RNA components transiently, thereby limiting risks.

Our previous research demonstrated that engineering the Moloney murine leukemia retrovirus structure enables the packaging of base editor protein and guide RNA without an associated viral genome. This results in an engineered virus-like particle (eVLP), which has a high cargo capacity, a modifiable tissue/cell-type tropism, and transient delivery of genome editors without risk of genome integration. We showed that subretinal delivery of base editor eVLPs in vivo corrected a mutation in rd12 mice; these mice carry a natural nonsense mutation in Rpe65 – an enzyme expressed in the retinal pigment epithelium (RPE) that is critical for the retinoid visual cycle, and the same enzyme that is replaced by Luxturna. The editing also resulted in protein rescue and substantial physiology rescue of the electroretinography (ERG) response, indicating partial restoration of visual function.

However, the base editor eVLP architecture did not support efficient PE in vitro. To improve editing efficiency, we engineered and optimized several components of the eVLPs, including optimization of nuclear export and import signals, engineering the protease release site, and improving recruitment and packaging of both PE protein and pegRNAs. This led to two high-efficiency and complementary architectures: the PE v3- and PE v3b-eVLPs.

We chose two mouse models of IRDs to test the in vivo editing capability of these PE-eVLPs. Once again, we corrected the rd12 mouse model of Rpe65 and observed efficient RPE editing and physiological rescue. We also investigated the use of PE-eVLPs in the rd6 model of Mfrp, where mutations are associated with RP. This naturally occurring mouse line has a four base pair deletion in an mRNA splicing site, leading to aberrant mRNA splicing and no expression of MFRP protein. This deletion mutation is uncorrectable by base editing and provides a powerful proof-of-concept for IRD-causing insertion and deletion mutations. Again, we observed efficient editing in the RPE and restoration of wildtype MFRP protein. Importantly, for both genes and mouse models, we did not detect any off-target editing in the top ten sites we determined by CIRCLE-seq to be most susceptible to off-target editing in the RPE of animals treated with PE-eVLPs. This further supports the safety of transient PE-eVLP delivery for the treatment of IRDs.

The future of IRD treatment
 

Genetic medicine is poised to become an exciting treatment modality for IRDs. With further advances in PE and PE-eVLPs, we can envision the successful correction of almost all disease-causing mutations that lead to LCA, RP, and other IRDs. Though more careful study of the immunogenicity and off-target editing effects of PE-eVLPs need to be explored, prime editors will soon enter clinical trials, as Prime Medicine recently announced in a press release last year.

In our work, we demonstrated efficient in vivo editing of the RPE. A major unmet goal for the development of PE-eVLPs will involve the targeting of PE-eVLPs to rod and cone photoreceptors, which we did not investigate in this study, but constitute a major therapeutic target for IRD research. Nevertheless, as more disease-causing genes and mutations are discovered by sequencing of affected patients, researchers can design precision medicine PE therapies to correct these mutations and restore sight.