A study published last week presents a compact base editor that enables efficient in vivo delivery using a single adeno-associated virus (AAV) vector. The innovation, once approved in clinical trials, could lower the doses of AAV vector needed to achieve therapeutic efficacy and bolster precision medicine approaches using somatic genome editing by making them faster and safer.
The study was published in GEN magazine’s sister journal, GEN Biotechnology, in an article titled, “Adenine Base Editing In Vivo with a Single Adeno-Associated Virus Vector.”
As detailed in GEN in 2019, base editors (BE) are an exciting CRISPR-based platform for precision genome editing, developed in the laboratory of David Liu, PhD, from the Broad Institute, that can engineer a subset of DNA substitutions without cleaving the double helix. However, delivery of BEs in a live organism encounters several challenges, not least of which are safety concerns such as immunogenicity and hepatotoxicity associated with high doses of the viral vector. AAV is the most popular delivery vehicle, but its packaging capacity is limited, making it nearly impossible to fit the coding sequence and regulatory elements of BE constructs in a single vector. To overcome this hurdle, current strategies use two separate AAVs that split the machinery that must be assembled post-translationally to form the full-length editor in target cells. Of course, this creates further challenges.
“Currently, most base editor-guide combinations and their promoters exceed the genomic capacity of AAV vectors that are commonly used for in vivo delivery of genome editors. Dual-AAV delivery strategies often require high viral doses that impose safety concerns,” said the senior author of the GEN Biotechnology study, Erik Sontheimer, PhD, who is professor and vice-chair of the RNA Therapeutics Institute at the University of Massachusetts Chan Medical School.
“In vivo delivery is a major hurdle for gene editing technologies to reach their full potential in the clinic. Much research is therefore directed towards the development of strategies to maximize levels of precise and safe delivery of editing agents alongside the requirement for efficient on-target gene editing. This will ultimately enable the progression and wider deployment of advanced cell and gene therapies for the treatment of plethora diseases,” said Jennifer Harbottle, PhD, a senior scientist at Horizon Discovery, whose work encompasses gene editing and advanced cell and gene therapy. (Harbottle was not involved in the current study.) “It is exciting to see new developments in this area.”
Adenine bases editors (ABEs) include a single-guide RNA (sgRNA) on an inactivated Cas9 protein that is fused to a deaminase enzyme. The complex changes an adenine(A)-thymine(T) base-pair to guanine(G)-cytosine(C).
In earlier studies, Sontheimer’s team developed a compact Cas9 from the bacterium Neisseria meningitidis (Nme2Cas9) that permits strong editing in mammalian cells. This led Sontheimer to envision a compact BE based on Nme2Cas9 that could override the need for two AAV vectors to package the entire editing machinery, thereby improving the safety and efficacy of a range of therapeutic applications.
The current study engineered this compact new ABE (Nme2-ABE) using Nme2Cas9. Compared to ABEs that contain the more popular Streptococcus pyogenes Cas9, the new Nme2-ABE targets a distinct range and editing window, generates fewer off-target edits, and can efficiently correct mutations in both mouse and human genomes.
“The work advances the field of base editing both by creating a more economically sized ABE for AAV delivery and increasing the targeting space of BEs through the development of Nme2Cas9 ABEs for in vivo application,” said Nicole Gaudelli, PhD, director and head of gene editing platform technologies at Beam Therapeutics, who was not involved in the current study. (Gaudelli led the initial development of ABE while a postdoc in Liu’s lab.)
“The work overcomes the need to encode an ABE in two AAVs, by using a CRISPR-Cas enzyme that has a smaller coding sequence. The ABE-Nme2 construct can be packaged into a single AAV along with its gRNA and all regulatory elements. They observed reasonable levels of A>G editing in vivo, in one case enough to lead to phenotypic correction,” observesd Ben Kleinstiver, PhD, assistant professor of pathology at Massachusetts General Hospital and Harvard Medical School, also not involved in the current study.
A C-G to T-A point mutation in the gene MeCP2 inserts a premature stop codon, truncating the protein and resulting in a rare neurological disorder called Rett syndrome, which is characterized by movement disabilities and growth restrictions. Sontheimer’s team electroporated Nme2-ABE mRNA with a synthetic sgRNA into a fibroblast cell line derived from a Rett syndrome patient and successfully edited the target adenine.
“To circumvent the need for multiple vectors, associated toxicity, and the manufacturing burden of increasing viral dosage for effective editing, Sontheimer’s lab leveraged the compact size of Nme2-Cas9 to develop Nme2-ABE and showcased this in vivo with a successful functional outcome, the rescue of a pathogenic phenotype,” Harbottle added.
“The adenosine deaminase part of ABE-Nme2 seemed to have a bit of an incompatibility issue with the new Cas protein, as editing overall is lower and there’s an unpredictable ‘editing window,’ but they managed to make do with it and correct some disease-relevant mutations in human patient-derived cells (Rett syndrome) and a mouse model of tyrosinemia. The adenosine deaminase enzyme was previously evolved to be compatible with SpCas9, so perhaps this editor could benefit from some directed evolution to make the deaminase more compatible with Nme2-Cas9,” said Alexis Komor, PhD, assistant professor of chemistry and biochemistry at the University of California, San Diego. (Komor, along with Gaudelli, developed the first base editor as a postdoc in Liu’s lab).
Encouraged by their results in cell lines, Sontheimer’s team injected the editing machinery including the new ABE (Nme2-ABE8e) packaged in AAV9 into the tail veins of adult mice harboring a mutation in the fumarylacetoacetate hydrolase (FAH) gene that results in the build-up of toxins that damage the liver, kidney, and central nervous system. A month following the injection, the researchers noted positive FAH expression in up to 6% of liver cells. Although quite low, earlier estimates have shown repairing one in 100,000 liver cells rescues the phenotype, suggesting efficiencies above the therapeutic threshold.
Sontheimer said, “Importantly, we show that in vivo delivery of Nme2-ABE and its guide RNA by a single-AAV vector can efficiently edit mouse genomic loci.”
“The team has demonstrated that full-length BEs can be encoded in a single AAV, and that Nme2-ABE can perform appreciable levels of base editing in vivo via hydrodynamic tail vein injection. This has the advantage of potentially enabling a lower dose of total virus relative to SpCas9 BEs and reducing the cost of manufacturing,” commented Gaudelli.
“This work is a valuable advancement in the field of base editing and I look forward to seeing how this advancement can potentially provide further opportunities for patients suffering from genetic diseases,” she added.
The Sontheimer lab’s study rescues a pathogenic phenotype despite the relatively low genomic editing efficiencies. “This highlights the importance of evaluating the level of editing that is required to exert therapeutic benefit, as this will vary between conditions, and modest editing could benefit patients even when considered sub-optimal in vitro,” said Harbottle.
“However, it also highlights an avenue for future development of the Nme2-ABE8e compact system. It will be interesting to see technological developments that narrow the editing window of Nme2-ABE8e to minimize bystander editing, as well as work towards increasing on-target editing efficiency to broaden the system’s therapeutic potential, further minimize viral dosage, and ultimately encourage translation to the clinic.”
Komor added, “The new editor won’t be universally applicable to all instances in which you would want to install an A to G mutation (because of the more complicated PAM, and unpredictable editing window), but I don’t necessarily believe in a ‘universal’ editor anyways. The inhibition of editing using the anti-CRISPR proteins doesn’t seem super useful to me, since the major off-target issues with base editors are independent of the Cas protein (spurious deamination, which this particular adenosine deaminase has been shown to have). Overall, a nice addition to the genome editing toolbox but it could use some editing efficiency improvements.”
Kleinstiver told GEN: “This is an important proof-of-concept study that should motivate continued development of single-AAV editors with more minimal coding sequences. A few challenges with the ABE8e-Nme2 enzyme are that the efficiency is typically quite reduced compared to standard ABE8e-SpCas9 constructs, and the edit window is quite wide, which can lead to unwanted bystander edits of nearby bases. Continued optimization of these enzymes, and the development of ABEs using other smaller Cas orthologs, will surely lead to highly efficient and precise single AAV editors.”
“Comprehensive genomic safety profiling remains challenging to achieve whilst on-target editing is low; gRNA-dependent off-target site editing, for instance, will likely be below the threshold of detection for amplicon NGS and other off-target identification or analysis methods. Off-target assessment of an optimized, next-generation Nme2-ABE8e system will certainly add confidence for its contribution to the gene-editing toolbox and route towards therapeutic use,” Harbottle said.
Sontheimer and his team continue to tinker with the new ABE to refine its potential applications and accuracy. “We are currently working on further improving the editing efficiency and targeting scope of Nme2Cas9-ABEs,” said Sontheimer. “We are also applying the single-AAV vector delivery system in vivo to correct therapeutically relevant mutations in disease mouse models.