Researchers from the National Institutes of Health, the Chinese Academy of Sciences, and Wenzhou Medical University have recently co-developed a next-generation adenine base editor (ABE) with maximized editing accuracy and adequate editing efficiency in plant and human cells. Compared with the currently available adenine base editor NG-ABE8E, the new base editor can significantly improve editing safety, reducing bystander effects by as much as 7-fold in some human cells.
In recent years, the rapid development of gene therapy technology, especially the progress of gene editing technology represented by CRISPR, has made it possible to cure genetic diseases. However, gene editing using the CRISPR system must first create double-strand breaks (DSBs), which often lead to genome lesions and recombination.
Base editors overcome many of the limitations of the CRISPR system by achieving accurate gene correction through single nucleotide conversion in genomic DNA without DSB generation. A base editor cuts a necessary DNA site in preparation for nucleotide substitutions. When it works, a base editor edits only a given nucleotide. Nevertheless, sometimes it also edits nucleotides adjacent to the target nucleotide within the editing window, leading to an undesired result known as the bystander effect. The off-target bystander editing has hindered the safe application of base editors in human gene therapy.
Previously, researchers have developed an EGFP-based screening system that can measure ABE activity simultaneously at multiple sites. ABE-mediated editing can restore full-length expression of EGFP, producing a large amount of fluorescence enhancement. By analyzing cell fluorescence intensity through flow cytometry (FCM), ABE's editing efficiency and bystander effect can be quantitatively detected.
In this study, the researchers used this method to screen different ABEs. They found that one of the ABE mutants not only showed strong editing efficiency at the target site A6 but also significantly reduced the editing at the bystander sites A9, A11 and A12, suggesting a systematic improvement of editing precision. This new mutant was later named NG-ABE9e, as it contains nine mutations compared to the original NG-ABEmax.
To investigate whether the excellent editing properties of NG-ABE9e are suitable for other cell types, the researchers used NG-ABE9e (PABE9e) to edit 9 genomic loci in rice. They compared NG-ABE9e against a variety of highly efficient ABE variants.
They found that PABE9e shows considerable editing efficiency compared to PABEmax and PABE7.10. In addition, compared to PABE8e, the overall editing efficiency of PABE9e is slightly reduced, but the accuracy is greatly improved, resulting in a five-fold reduction in bystander edits.
The researchers also tested the new base editor in 16 representative endogenous genomic loci and various human cell lines. The study found that NG-ABE9e significantly reduced the frequency of off-target bystander edits at all test sites and in all tested human cell lines. The bystander editing even decreased by more than 7 fold at some sites. This suggests that reduced bystander editing is a unique attribute of NG-ABE9e, and that this ABE variant can edit with precision in different genomic and cellular environments.
To further explore the potential application of NG-ABE9e in gene therapy to correct human disease-causing mutations, the researchers established a disease model of autosomal dominant retinitis pigmentosa (adRP) in the HEK-293 cell line. adRP is caused by a gain-of-function mutation T17M in Rhodopsin (RHO) that leads to vision loss.
In this study, the researchers attempted to leverage NG-ABE8e or NG-ABE9e and the designed sgRNA to introduce an M17T reverting mutation into cells that turns CAT into CGT. They found that both NG-ABE8e and NG-ABE9e mediated efficient base editing (61% and 58%, respectively). However, the perfect editing rate (the probability of correcting the disease-causing mutation without producing a bystander effect) of NG-ABE9e reached 49%, while the NG-ABE8e in the same case only reached 27%. In other words, NG-ABE9e significantly reduced the bystander effect and almost doubled the perfect editing rate. These results indicate that NG-ABE9e is safer, more accurate, and more reliable for correcting human pathogenic mutations, than the currently available adenine base editor NG-ABE8e.