NEW YORK — Genome editing with CRISPR-Cas9 can generate unintended mutations at both on- and off-target sites, a new study in zebrafish has found.
CRISPR-Cas9 has become a widespread tool in biology, allowing researchers to make genetic modifications in the genomes of microorganisms, plants, and animals. While it could also have applications in healthcare, ethical and other concerns need to be addressed, including uncertainty surrounding the frequency of off-target and other unwanted alterations.
Researchers from Uppsala University have examined the effect of genome editing with CRISPR-Cas9 in a living model organism, the zebrafish Danio rerio, over two generations. Previous studies, they noted, have had conflicting findings about the prevalence of unwanted CRISPR-Cas9 editing outcomes.
As they reported in Nature Communications on Wednesday, the Uppsala researchers found CRISPR-Cas9 editing can lead to structural variants affecting both on- and off-target sites within not only the edited generation of fish, but also their offspring.
"Knowing these unexpected mutations are heritable is important, since they can have long-term consequences for future generations. But that can happen only if you change the genome of embryos or germ cells," first author Ida Höijer, a researcher at Uppsala University and SciLifeLab, said in a statement.
Höijer and her colleagues edited fertilized zebrafish eggs at the single-cell stage using four different guide RNAs. The guide RNAs targeted ldlra, nbeal2, sh2b3, and ywhaqa, all orthologs of human genes implicated in cardiometabolic risk and disease. The researchers collected samples from various zebrafish developmental stages, such as when the founder generation of fish reached the larval stage and adulthood and when the next generation reached the larval and juvenile stages, for analysis.
By applying a long-read sequencing approach to more than 1,100 fish, the researchers found editing led to insertions and deletions of various sizes at both on-target and off-target sites. In particular, editing had on-target efficiencies of 92.6 percent for ldlra, 96.7 percent for nbeal2, 92.6 percent for sh2b3, and 93.6 percent for ywhaqa. Off-target editing was detected at three sites and had lower efficiencies.
Nearly 70 percent of first-generation fish showed on-target editing and about a quarter of those founder fish had off-target editing in at least 10 percent of DNA molecules, though the amount varied by individual. But while most on-target events were insertions or deletions, about 7 percent were structural variations of 50 or more bases.
These changes — both on- and off-target — could further be passed on to the next generation of fish, the researchers found. Next-generation fish harbored a higher proportion of edited alleles as well as more structural variants. More than a quarter of these fish carried an off-target mutation and 9 percent had a structural variant, though those SV were all at on-target sites.
The findings suggested to the researchers that validation and verification of CRISPR-Cas9-based genome editing is a needed step, especially if clinical applications are being considered. They further proposed a three-step approach relying on an in vitro method to detect Cas9 cleavage sites, long-read sequencing of those on- and off-target sites, and an orthogonal method to identify all CRISPR-Cas9-induced edits.
"CRISPR-Cas9 can be an amazingly valuable tool in healthcare. But we need to minimize the risk of unwanted effects, and we can do this by carefully validating the modified cells with the latest DNA sequencing technologies," senior author Adam Ameur, an associate professor at Uppsala and SciLifeLab added in a statement.