In a short correspondence to Nature Methods Schaefer et al. (2017) have reported some results that are echoing around the gene editing community and insect gene editors should be aware of their findings.
Off-target effects and second-site mutations have been the bane of geneticist since they started systematic mutagenesis screens in the middle of the twentieth century. Recently, as dsRNA and siRNA-based technologies became popular and genetics became ‘democratized’ the awareness and concern for off-target effects grew (but perhaps not fully appreciated). As CRISPR/Cas9 has grown in popularity, its specificity and fidelity have been questioned and evaluated.
Schaefer et. al. (2017) approach this question of CRISPR/Cas9 specificity and off-target effects using whole genome sequencing and focusing specifically on single nucleotide variations – point mutations.
To date efforts to address and quantify CRISPR/Cas9 off-target effects, and there have been many, have focused on indels (insertions and or deletions). Algorithms are available to predict off target sites and these, when used in conjunction with sgRNA design parameters can be helpful in minimizing off-target effects – at least so it has been thought.
But Schaefer et al. look specifically point mutations. They did this using four sgRNAs to mutagenize mice. They took two of the recovered and planned mutants and sequenced their genomes to a depth of 50x. A non-mutagenized control was also sequenced to a depth of 30x.
As expected they found new indels. Each mutant mouse had some 150 indels (off-target) and 117 of them were unique to and shared among the two mice indicated nonrandom targeting. The rate of indel creation was high above background, inferring CRISPR/Cas9 as being the direct or indirect cause.
Surprisingly of the top 50 in silico-predicted off-target sites, none were mutated indicating that our various in silico prediction algorithms need improvement.
But it was their findings of point mutations that are new and interesting. About 1700 point mutations were discovered in each mouse and 1400 were shared among the mutated mice but not the control or any of the genomes of the 36 strains of mice now available. Again, an indication of significant non-random mutagenesis. All of these data implicate CRISPR/Cas9 as being the cause of these mutations. They examine the locations of these mutations and discuss their potential for resulting in phenotypes. This was sobering.
So, Schaefer et al. have seemingly discovered the unpredicted generation of point mutations by CRISPR/Cas9 and it appears to be significant.
This is a cautionary tale that insect biologists should heed. The gold standard of geneticists for attributing a particular mutation to a particular phenotype and definitively ruling out the role of secondary mutations that are certainly there following any mutagenesis with EMS or other mutagens including now CRISPR/Cas9 is genetic reversion to the starting phenotype. This standard is rarely, if ever, applied to contemporary Cas9 mutated insects. It is a high and stringent standard. If gene editing is as easy and quick and cheap as one would come to believe it is from reading even some of the scientific literature, perhaps ‘reverse edited’ insects (revertants) should be expected as support for claims of mutation cause and effect.