Accelerating base editing protein therapy for genetic diseases

delivery of proteins for genetic disease

Date: 19th January 2022

The recent development of novel gene editing agents is enabling precise manipulations of genomic DNA in living organisms and is making the possibility of treating the root cause of many genetic diseases a reality.   In particular, CRISPR technology has become a phenomenon in both biomedical and therapeutics research and is already being used in a clinical setting.  However, delivering these editing agents safely in a manner that limits off-targets effects has proven challenging. Whilst, methods to deliver gene editing machinery in vivo as ribonucleoproteins (RNPs) could offer safety advantages over nucleic acid delivery approaches, this strategy has been limited by modest editing efficiencies due to inefficient protein delivery and has limited validation of therapeutic efficacies. Now, researchers have engineered DNA-free virus-like particles (eVLPs) that efficiently package and deliver base editor or Cas9 ribonucleoproteins in vivo with therapeutically relevant efficiencies.

Gene editing has the potential to treat the underlying cause of many genetic diseases and CRISPR technology has revolutionised the field.  Several CRISPR-based therapies are already in clinical use, but the technology is rapidly evolving such that is offers much greater control for making a wide range of precisely targeted edits to the DNA code.  One such breakthrough has been base editors (BEs), these mediate targeted single-nucleotide conversions without requiring double-stranded DNA breaks (DSBs) which minimises undesired consequences of editing such as indels, large deletions or other chromosomal abnormalities.

To date the most robust approaches for delivering BEs in vivo involves the use of viruses, such as adeno-associated viruses (AAVs).  However, viral delivery of DNA-encoding editing agents leads to prolonged expression in transduced cells, which increases the frequency of off-target editing, and leaves open the opportunity for viral genetic material inserting into the host genome.

Virus-like particles have long been considered one of the most attractive delivery technologies to carry molecular cargo as they do not contain viral genetic material.  In addition, delivery of proteins or ribonucleoproteins (RNPs) instead of DNA would be beneficial as the short lifespan of RNPs in cells limits opportunities for off-target editing.  However, current VLP-based approaches for delivering RNPs report only modest editing efficacy and validation of therapeutic levels of in vivo gene editing with VLP-RNPS has not been reported.

Now, researchers at the Broad Institute of MIT and Harvard, US, led by gene-editing pioneer David Liu, systematically engineered different parts of the VLP architecture to overcome delivery efficiency.  The resulting eVLP is the first virus-like particles to deliver therapeutic levels of gene editing proteins to a variety of tissues in adult animals. It could package 16 times more cargo proteins than previous designs, and enabled an 8- to 26-fold increase in editing efficiency.  Base editor eVLPs reduced serum Pcsk9 levels 78% following 63% liver editing in mice, and partially restored visual function in a mouse model of genetic blindness.

 The team started by creating a BE-eVLP using a retroviral capsid as a scaffold.  The BE protein was fused to the C-terminus of murine leukemia virus (MLV) gag polyprotein via a linker that was cleaved by the MLV protease upon particle maturation.  Initial tests showed the BE-eVLP could install mutations in the genome for a potential treatment of β-hemoglobinopathies however, the team wanted to increase the efficiency of the system.

To address this they identified three bottlenecks – cargo packaging, release, and localisation and then engineered the eVLP to overcome these limitations.

First they reasoned that linker-cleavage efficiency might bottleneck BE-VLP editing.  So they used a variety of protease-cleavable linker sequences between the MLV gag and the BE protein.  One in particular exhibited up to 1.5-fold higher editing efficiencies at all doses tested relative to the original version, suggesting that the gag–cargo protein linker sequence is an important parameter of VLP architectures and is an important optimisation parameter.

Next, they turned their attention to improving cargo localisation and loading into eVLPs.  Typically BEs contain one or two nuclear localisation signals (NLSs), which promotes nuclear import of BEs and enhances their access to genomic DNA.  However, in order to be incorporated into VLPs as they form BEs need to be localised in the cytoplasm. So the team added nuclear export signals (NESs) to promote BE loading.  One version containing 3 NESs showed up to a 2.1-fold improvement in editing efficiencies, and showed an increase of the BE protein in the cytoplasm of the producer cells.

The final round of VLP engineering was aimed at improving component stoichiometry, the team hypothesised that the balance between the amount of gag–cargo available to be packaged into VLPs with the amount of MMLV protease required for VLP maturation may be important.  Indeed it was, and an optimal ratio of the two was found, which modestly improved editing efficiencies.

Now armed with an optimised version of BE-eVLP, which could package 16x more cargo proteins than their initial VLP, and enabled up to a a 26-fold increase in editing efficiency, the team put the system to the test by correcting mutations in a range of mouse and human cells.  Here, they observed 95% editing efficiency in some cases, but in all cases saw an increase in editing efficiencies with the BE-eVLPs compared with their original version.  Together, the data showed that BE-eVLPs could mediate high-efficiency base editing with minimal impact to the viability of treated cells and would work in many different cell types.

But how would the system translate in vivo?

First, the team investigated the ability of eVLPs to enable base editing within the mouse central nervous system (CNS). They designed a BE-eVLPs to install a silent mutation in mouse Dnmt1 gene and injected them directly into neonates into the cerebrospinal fluid that bypasses the blood-brain barrier.  They saw 53% and 55% editing in transduced cortex and mid-brain cells, although this corresponded to 6.1% and 4.4% editing of bulk cortex and mid-brain cells.  This suggested that the system could deliver robust levels of active BE RNP per transduction event, but transduction events were low and require improvement.

Next, they investigated their ability to mediate therapeutic base editing in adult animals.  They targeted proprotein convertase subtilisin/kexin type 9 (Pcsk9), a therapeutically relevant gene involved in cholesterol homeostasis. Previous work suggests that disrupting the PCSK9 gene could be a promising strategy for the treatment of familial hypercholesterolemia.  BE-eVLPs were designed to target and disrupt a splicing donor site, and were injected systemically into mice.  They saw 63% editing in bulk liver following treatment with the highest dose, which was 26-fold higher than with their original BE-VLP.   Phenotypic analyses revealed a 78% reduction in serum Pcsk9 protein levels after one week, demonstrating efficient and therapeutically relevant base editing in the mouse liver had occurred.  Furthermore, there were no apparent adverse consequences and no off-target editing was detected.

The final experiment was to use BE-eVLPs to correct a disease-causing point mutation in an adult mouse model of a genetic retinal disorder, Leber congenital amaurosis (LCA).  The team used the system to correct a nonsense mutation, sequencing analysis showed up to 21% correction 5 weeks post injection, with a 1.8-fold higher editing at the target base compared to the orginial BE-VLP, and improved visual function.  Whilst on-target editing was high, the team were also able to minimise off-target editing by using alternative BEs with narrower editing windows.

Conclusions and future applications

The team here have developed an efficient, engineered VLP platform that can safely deliver RNPs for therapeutically relevant ex vivo and in vivo applications. Through identifying and engineering solutions to three distinct bottlenecks to VLP-delivery efficiency, they have improved protein loading within eVLPs by an average of 16-fold and base editing efficiencies by an average of 8-fold compared with initial designs based on previously reported VLP scaffolds.

Liu and colleagues are now expanding the range of organs and cell types that eVLPs can target in animals. They will also continue to characterise eVLPs to better predict and mitigate any unwanted immune responses the particles may produce. They envisage the eVLP architecture will serve as a modular platform for delivering other proteins or RNPs of interest, in addition to BEs and nucleases.

BE technology is evolving rapidly, and the addition of precise delivery systems with minimal off-target effects will be a welcome addition.  The accuracy of BEs has been subject to enhancement, with the aim of minimising bystander edits.  The BE-eVLP platform being modular can accommodate enhanced version of BEs as they arise, ensuring optimal and precise editing.

BEs in general are becoming an evermore popular tools for correcting genetic disease.  They have been used in vivo to repair a single nucleotide mutation in TMC1 deaf mice which was partially able to restore hearing.  In vivo base editing to correct the single progeria mutation, rescued the vascular pathology in progeria-modelled mice and extended the lifespan by nearly two and half times.  Once limited to a single mutation correction, it has also been established that multiplexed precise base editing works efficiently in primates, enabling the editing of up to three target sites simultaneously.

Whilst safety and efficiencies are top of the list of concerns for any new technologies entering the clinic, advances such as here with the eVLP system and machine learning models, such as BE-Hive – also developed by Liu – which determines the best-in-class BE for a particular gene sequence, will allow balance editing efficiencies with precise on-target events, to ensure the safest therapies enter the clinic.  Now, researchers have a new tool to improve therapeutic macromolecule delivery for patients, and the eVLP system will open new avenues for protein therapy and rescuing the most debilitating of genetic diseases.



For more information please see the press release at the Broad Institute


Banskota, S., Raguram, A., Suh, S., Du, S.W., Davis, J.R., Choi, E.H., Wang, X., Nielsen, S.C., Newby, G.A., Randolph, P.B., et al. Engineered virus-like particles for efficient delivery of therapeutic proteins. Cell.