The concept behind gene editing is that you can make very precise and controlled changes to genomes with the potential of curing a large number of diseases. CRISPR/Cas9 has allowed us to edit genes in a way that is cheaper, faster and arguably more efficient than the technologies that preceded it. It was also thought that CRISPR/Cas9 was very safe in that it would rarely go “off-script” and cause changes to DNA in other locations of the genome. However, a recent study by Schaefer et al. 2017 suggests CRISPR/Cas9 may introduce hundreds if not thousands of unintended changes to the genome.
Every genome editing technology will have some level of off-target effects where unwanted mutations are introduced. If the number of these mutations is low this is unlikely to be problematic. However, the higher the number the greater chance that these mutations could lead to cancer.
As CRISPR/Cas9 is guided by a short RNA sequence most off-target mutations are likely to occur in regions of DNA whose sequence closely resembles the target sequence. As such, most people using CRISPR use algorithms to predict where mutations are likely to occur and then they look at these specific regions for any unwanted alterations following genome editing procedures. This method is far easier and cheaper than looking at the whole genome.
Iyer et al. 2015 however used whole-genome sequencing in 2015, and reported that off-target mutations were rare in mice treated with CRISPR to delete parts of a specific gene. In this study, whole-genome sequencing was used to detect insertions or deletions in the genome and not single changes of the DNA from one base pair to another (e.g. A to G).
In the latest study by Schaefer et al. the team used whole-genome sequencing to detect both insertions and deletions together with single base changes. The group previously used CRISPR/Cas9 to repair a gene called Pde6b in mice which, when mutated causes blindness (Wu et al. 2016). In the present study, they used whole-genome sequencing to assess whether this treatment introduced any unwanted mutations.
They report that mice treated with CRISPR had over 100 insertions or deletions (indel) to their DNA and over 1,500 instances were single base pair mutations were found which were not seen in untreated mice. Of concern, almost half of these mutations were associated with known genes and a small fraction were expected to impact on their function. Crucially, none of these mutations were predicted by the type of algorithms typically used to tell scientists where to look for mutations.
The results of this study by Schaefer et al. 2017 were so shocking that stock prices of Intellia Therapeutics and Editas took a hit upon the publication of the study. Both companies were set-up with the aim of turning CRISPR technology into a medicine.
Before dismissing CRISPR/Cas9 as a non-starter it is important to note that this is a single study in which only 2 mice were assessed. Also, CRISPR/Cas9 here was used to target a single gene using a single guide RNA meaning the high level of mutation rate could be a peculiar phenomenon specific to the guide RNA used and not CRISPR/Cas9 in general. These factors make it difficult to draw firm and general conclusions with regards to safety of CRISPR/Cas9.
Also, with the current study design it is difficult to pinpoint the CRISPR system as the cause of the mutations seen in the treated mice. We all develop mutations as our cells divide and mice are no different. Because of this, single base differences in our genomes are very common- these are known as single nucleotide polymorphisms and for the most part are not thought to be involved in disease. Given this natural variation we cannot be certain that the mutations reported in the treated mice were due to CRISPR/Cas9 and did not occur naturally. The ideal scenario would have been to sequence the genome of the same mice before and after CRISPR treatment.
Nevertheless, the number of mutations found in the CRISPR treated mice was substantially higher than would have been expected from natural variation and the single nucleotide changes seen in CRISPR treated mice were not seen in a large number of untreated mice whose genomes have been deposited in databases.
In conclusion, this new study has generated some unexpected results but is by no means going to spell the end of the CRISPR revolution underway in personalised medicine. Indeed stock prices of Intellia Therapeutics and Editas rapidly recovered. The key take home message from this study is that we need to proceed with caution and ensure whole-genome sequencing is the norm for assessing off-target mutation rates with each CRISPR/Cas9-guide RNA combination we use. The first clinical trial to employ CRISPR is already underway in China and one is planned to start in the U.S. next year with others no doubt to follow shortly after.
Iyer, V., et al. (2015). “Off-target mutations are rare in Cas9-modified mice.” Nat Methods 12(6): 479.
Schaefer, K. A., et al. (2017). “Unexpected mutations after CRISPR-Cas9 editing in vivo.” Nat Methods 14(6): 547-548.
Wu, W. H., et al. (2016). “CRISPR Repair Reveals Causative Mutation in a Preclinical Model of Retinitis Pigmentosa.” Mol Ther 24(8): 1388-1394.