Scientists using CRISPR discover hundreds of unintended mutations
The CRISPR gene editing technology has taken the world of biology by storm. It's provided a simple, cheap, easy to use tool for scientists that allows them to manipulate DNA in ways they'd only previously imagined. And one of its big selling points has been that it is precise. It targets the DNA right where you want it – a bulls-eye every time.
But – and you knew there was going to be a but – there may be a fly in the ointment. A new study released this week suggests CRISPR might not be as precise as we thought. To the great surprise of Dr. Vinit Mahajan, a vitreoretinal surgeon and associate professor of ophthalmology at Stanford University in California, he and his team found that in some ways, it can be downright sloppy. And that's a major concern when it comes to genetic modification.
This interview has been edited for length and clarity.
Bob McDonald: Tell me about the experiment you were doing with CRISPR.
Dr. Vinit Mahajan: A few years ago, we wanted to use the CRISPR system to correct a gene that causes blindness in mice as a model for humans. CRISPR is incredible technology. We were able to cure the blindness in the mice. We had tissues from the mice left over. And more recently, we wanted to do our due diligence and see if CRISPR was having any off-target effects; meaning did the CRISPR go to places that we didn't want it to go and do something to the genome of the mice. My graduate student was on it and we had the expectation that based on the news and things like that, that it would be pretty clean. She was surprised to find that there were a number of unintended mutations using the method that we used to look for these.
BM: Unintended mutations? Well, tell me about that.
VM: We used a method called whole genome sequencing, in a very unbiased way, and looked at the entire genome to see if there were any DNA changes elsewhere outside of that blindness causing gene.
BM: What did you find?
VM: Well we found that if we used the current computer algorithms that are supposed to help us predict where those off-target effects might occur, we didn't find any at those locations. Instead we found deletions and changes in different letters of the genome, at lots of different places, that were not predicted.
- Nature Methods paper: Unexpected mutations after CRISPRCas9 editing in vivo
- Nature Methods editorial, "CRISPR standards"
- CRISPR Repair Reveals Causative Mutation in a Preclinical Model of Retinitis Pigmentosa
BM: So are you saying that these were further away from the site where you were doing your editing?
VM: They were further away. And frequently what we expect is that the code of letters that we're targeting, the CRISPR-Cas9 would end up at something that looked similar. And the sequences that we found that were getting off-targeted actually didn't look that similar.
BM: Why do you think it went to sequences that were different than the one that you were going after?
VM: I honestly don't have the answer to that. And I think it's something that the community of scientists are going to have to look at and try to figure out and better understand. The technology is moving so fast.
BM: Well did these off-target unintended mutations that happened have any effect on the mice?
VM: Well so we looked at the mice and in general, the mice looked OK.
BM: Well these genes that you weren't intending to edit, do you know what they do?
VM: Some of the genes that got targeted, we have guesses at what they do. We did not investigate how those genes were functionally affected, but 99 per cent of our genome does not actually code for a protein. In the old days we used to call this part of our genome the "junk DNA." But we found out over the years that this DNA that actually doesn't code for protein is actually important. It codes for RNAs. It codes for other things that are important for the health of the cell. And that's where we saw a lot of these mutations.
BM: Well what are the risks of the fact that CRISPR is no longer as precise as you thought?
VM: I think it can be very precise at targeting what you want, but it's not perfect. And as a physician, I always say it comes down to risk-benefit. There are some diseases and conditions that are so severe, without any kind of treatment, where it makes sense to use gene editing. In other cases, the risk might be so high that it doesn't make sense, like designer babies.
BM: Well how then do you think this should change how scientists think about using CRISPR?
VM: I think we need to begin to explore whether whole genome sequencing is a good method for looking at off target effects and comparing it to some of the other great methods that have been proposed.