Precise, Integrated Gene Correction and Integration Offers Therapeutic Benefits

June 5, 2024

By Dava Stewart  

June 5, 2024 | It’s nothing short of amazing how quickly the field of gene editing has progressed. During his keynote address at last month’s PEGS Boston, David Liu, director of the Merkin Institute of Transformative Technologies in Healthcare, vice-chair of the faculty at the Broad Institute of MIT and Harvard, and a Howard Hughes Medical Institute investigator, described how the laboratory evolution of gene editing systems has enabled efficient and programmable gene correction or even whole gene integration. 

Liu describes the latest engineered proteins as “laboratory-engineered molecular machines,” and says they already offer therapeutic benefits. He’s optimistic about improving efficiency and the continuing effort to treat numerous genetic diseases.  

First, a Bit of History: Base Editors

Base editors, which were developed by Liu, correct transition point mutations—purine for purine (A and G) or pyrimidine for pyrimidine (T and C) mutations—without creating double strand breaks (DSBs) in the original DNA. DSBs initiate DNA repair pathways that can cause large deletions and complex rearrangements, so methods that correct genes without creating DSBs are particularly attractive.  

Liu shared two examples of using base editing to correct disease in humans.  

Alyssa, a 13-year-old girl in the UK with lymphoblastic leukemia, received the first treatment created with base editing in 2022. She had already been through chemotherapy and a bone marrow transplant, both of which failed. For the new treatment, a form of CAR-T therapy, Alyssa was given T-cells expressing a chimeric antigen receptor recognizing CD7. The novelty was that base editing was used on these unrelated donor T-cells to triply delete CD7, CD52, and the β chain of the αβ T-cell receptor, allowing them to act as “universal” donor cells and to avoid self-recognition. The base editing involved single cytosine->thymine alterations that create premature stop codons for these genes. Within one month of receiving the treatment, her disease was in remission, and this remains true more than 2 years later. She has returned to school and plans to pursue a career in biomedicine. 

In another study, patients who had very high LDL cholesterol levels and an accompanying high risk of heart attack or stroke were given a single injection of an adenine base editor that was precisely programmed to disable a splice site in PCSK9, ultimately having the same effect of effectively deleting this gene. The treatment lowered patients’ risk of heart attack or stroke by about 55% on average. In this case, base editing treatment addressed both the existing problem of high cholesterol and the potential future problems that high LDL cholesterol could cause, noted Liu.  
The Next Step: Prime Editing 

Along with base editors, prime editing is having an impact. Liu explained that prime editing uses nicked target DNA to prime reverse transcription of an edited sequence encoded in an engineered prime editing guide RNA (pegRNA). The pegRNA guides and demonstrates the desired edit. Using pegRNA, the researcher can specify the reverse template up to 200 base pairs in length. Since somewhere between 95%-97% of mutations or deletions are fewer than 50 base pairs in length, prime editing with pegRNA should be able to correct most of them.  


Liu’s lab has also developed phage-assisted continuous evolution (PACE), which accelerates protein evolution more than 100-fold compared to conventional methods, and used this method to generate improved prime editors. Reverse transcriptases typically have a hard time with longer prime edits, because the pegRNA can fold and create hairpins. Prime editors engineered with PACE are smaller, more efficient, and enable the insertion of longer sequences. 

Combining PACE with base editing and prime editing has collectively had a profound impact on challenging edits and yields results where they are most needed—in vivo, Liu said.  For example, these techniques combined have allowed for the precise re-introduction of the three letters missing from CFTR in the most common cause of cystic fibrosis. The treatment works so well that the results are comparable to Trikafta, the combination of three small molecule drugs that has revolutionized cystic fibrosis treatment, but unfortunately requires a lifetime of doses at a cost of around $300,000 per year. Gene editing, in contrast, requires a single treatment.  

In April, the FDA cleared the first clinical trial for prime editing, only four and a half years after the first paper on the topic was published. Scientists at Prime Medicine, a company for which Liu is a scientific co-founder, developed an ex vivo treatment for chronic granulomatous disease (CGD), a primary immunodeficiency. Hundreds of patients in the US have CGD. The median age of death is between 30 and 40 years, with progressive debilitation from infection, inflammation, and autoimmunity. The only current cure is allogeneic bone-marrow transplant, which carries a significant graft-versus-host risk. The disease is caused by mutations in subunits of NADPH oxidase complex, mostly commonly a two base-pair deletion. Prime editing can correct it, with no off-target edits.  

Outliers and Regulatory Questions

Unlike these examples, some genetic diseases are caused by many different mutations. In the ideal scenario, many patients will carry the same mutation, forming a large cohort that can be treated with a single well-tested therapy. But for diseases with many distinct causative variants, it is difficult to imagine how to practically develop the perhaps hundreds of different gene editing agents.  

For drug development and manufacturing to meet this challenge, the regulatory pathway needs to be modified such that changing the guide RNA doesn’t mean starting all over. Liu says that regulatory agencies are aware of the need to streamline regulatory pathways, as they are standing at “the base of a tidal wave” of new treatments.  


A potential solution exists by combining prime editing with site-specific recombinases to create gene-sized insertions at programmable loci in a method known as prime assisted site-specific integrase gene editing (PASSIGE). Instead of correcting each of several potential individual variants, one therapy could be used to replace all of these with one healthy gene. However, Liu says his team learned that trying to simultaneously introduce a prime editor, pegRNA, recombinase, and the cargo into a cell at the same time doesn’t work well.  
“Efficiency is pretty modest,” he said. Multiple studies found a maximum efficiency of around 10%-20%. The bottleneck was the recombinase, so Liu’s group used PACE to evolve an improved recombinase and improve efficiency. They call this modified protocol using evolved recombinase eePASSIGE. 

Liu’s lab has experimented with six therapeutic target gene integrations using eePASSIGE in human and mouse cells and found large, consistent benefits. Additionally, off-target integration remained low across all tested methods. Evolved recombinases worked better in more therapeutic cell types with around 30% gene integration.  

Another alternative, CASTs, or CRISPR-associated transposases Type 1, allow integration in a single process with no byproducts, but they show very low activity in mammalian cells. Even the most active transposases support less than 0.1% integration, says Liu.  

Using PACE to evolve CASTs allows researchers to tailor rapid application. “The real enabling feature,” says Liu, is that it is “is faster, and more efficiency allows you to fail 100x more, which is useful.” PACE shortens the linear path required for survival through more than 1000 generations of mutation, selection, and replication.  
These evolved transposases have been used on several genomic test sites and could eventually be used to treat a host of diseases. Liu is hopeful that efficiencies can be improved beyond the observed 10-30%, but says, “it can already offer potential therapeutic benefits.”  
Liu says he hopes these laboratory-engineered molecular machines will improve the ability to insert whole genes or large fragments where they are needed without any double-strand breaks or by-products. “I’m getting optimistic about delivery, partly because young scientists are interested in it, and partly because some of these machines have entered clinical trials.”