CRISPR gene editing is not quite as precise and as safe as thought

Authored by newscientist.com and submitted by recentfish
image for CRISPR gene editing is not quite as precise and as safe as thought

A study of CRIPSR suggests we shouldn’t rush into trying out CRISPR genome editing inside people’s bodies just yet. The technique can cause big deletions or rearrangements of DNA, says Allan Bradley of the Wellcome Sanger Institute in the UK, meaning some therapies based on CRISPR may not be quite as safe as we thought.

The CRISPR genome editing technique is revolutionising biology, enabling us to create new varieties of plants and animals and develop treatments for a wide range of diseases.

The CRISPR Cas9 protein works by cutting the DNA of a cell in a specific place. When the cell repairs the damage, a few DNA letters get changed at this spot – an effect that can be exploited to disable genes.

At least, that’s how it is supposed to work. But in studies of mice and human cells, Bradley’s team has found that in around a fifth of cells, CRISPR causes deletions or rearrangements more than 100 DNA letters long. These surprising changes are sometimes thousands of letters long.

This finding is not a problem for the purposes for which CRISPR is currently being used. But some groups are developing treatments that would involve using CRISPR to edit billions of cells inside the human body. If Bradley is right, there’s a chance that a few of these cells might turn cancerous.

“There’s a risk of causing cancer sometime in a patient’s lifetime,” says Bradley. “We need to understand more before rushing into human clinical trials.”

“I do believe the findings are robust,” says Gaetan Burgio of the Australian National University, an expert on CRISPR who has debunked previous studies questioning the method’s safety. “This is a well-performed study and fairly significant.”

The finding comes just a month after another team reported that there was a cancer risk from gene-editing embryonic stem cells, for a quite different reason. It was also claimed last year that CRISPR causes lots of mutations far from the target site, but Burgio and others pointed out that the study was seriously flawed and it has since been retracted.

“Overall, this adds more safety challenges to CRISPR-based gene therapies but it doesn’t compromise its use,” says Burgio of the new findings.

Many variations of the CRISPR technique are currently being developed, including base editors that alter specific DNA letters and nickases that do not completely sever the DNA. These methods may prove to be far safer. Some groups are also using modified forms of CRISPR to control how active a gene is instead of directly altering a gene’s DNA.

So why have the thousands of teams using CRISPR failed to discover this before? Because they have been looking for small mutations in a narrow region around the target site. If that whole region is deleted, this approach makes it appear as if there have been no mutations at all. “You find what you look for,” says Bradley.

Journal reference: Nature Biotechnology, DOI: 10.1038/nbt.4192

maybeCarmenSanDiego on July 16th, 2018 at 20:19 UTC »

so it's kinda like deleting that strangely named process on your computer and your computer suddenly not working one week later.

SirT6 on July 16th, 2018 at 17:13 UTC »

This is a pretty interesting result. I spent a lot of time in grad school thinking about DNA repair, so I can't say I am terribly surprised. I suspect there will be a rush by the science media to spin this as doomsday news for the field. I think a more nuanced take is warranted. I'm copying my comment from another thread (apparently it gets removed if I say where...):

So some quick basics:

CRISPR/Cas9 is one of the hottest technologies in biotech. It is no understatement to say that it has massively changed the way basic research is done - from screening, to making mouse models, to creating models of human disease. It's easy to use, and can be applied in thousands of different ways.

The success at the lab bench has gotten people excited about the prospect of using CRISPR/Cas9 in the clinic to treat human genetic diseases, and other conditions. There are three major CRISPR companies: Editas, Intellia and CRISPR Therapeutics. They have all licensed the technology from different academic labs (lots of drama!), and are working in partnership with other big pharma companies like Novartis, Regeneron and Vertex (just to name a few).

It's been a bit of a wild ride over the past 12 months for CRISPR/Cas9 as a medicine. Reports of concerns over immunogenicity, off-target mutations and clinical holds have rocked the boat a bit. Yet, optimism remains high (just look at the stock prices of these companies).

What did this current paper show?

CRISPR/Cas9, in most iterations, works by creating a DNA double strand break. There have long been concerns that this would create room for mutations.

What this paper shows is that you can detect certain complex forms of mutations at the target site, that you would otherwise miss by conventional testing methods.

Cumulatively, across the studies they found that between 20-35% of all clones exhibited evidence of large/complex genomic rearrangement that might be missed by conventional sequencing.

What does this mean for basic biology?

This is something to be aware of. And any good paper should properly validate their finding using several orthogonal approaches to rule out the possibility that mutations are driving the phenotype.

I suspect a lot of the potential for rearrangement will come down to methodology: what type of cells were used? how much Cas9 complex was present? how optimized were the guides? etc.

What does this mean for CRISPR in medicine?

Here the concern is more real. Inducing mutations in patient genomes is risky, opening up the possibility for tumorigenesis and similarly nasty outcomes.

Intellia, one of the major CRISPR companies, has already commented on the paper:

Similarly, at Intellia, another firm in Cambridge that is developing CRISPR–based therapies, scientists have looked for large deletions in gene-editing studies in mouse livers. So far they have found no evidence of such deletions, says senior vice president Thomas Barnes. This, he says, may be because the cells his team works with do not divide often. Bradley and colleagues' study, by contrast, used actively dividing cells.

I'm inclined to trust these companies at this point (they can face stiff penalties for lying - especially stay tuned to how they answer questions during end of quarter conference calls next month). Further, there are a number of in vivo mechanisms that might work against some of the in vitro results reported here.

Either way, it is an interesting result, though. And one that definitely bears paying attention to. This data also probably warrants a close look by third parties. There was a bit of an embarrassing SNAFU last year when a group reported that CRISPR/Cas9 caused thousands of off-target mutations, but upon closer inspection it was revealed their methodology was woefully inadequate and the paper was retracted.

I'll add in more details as I get a chance to read the paper more closely!

wojosmith on July 16th, 2018 at 16:14 UTC »

I've been watching and you had to know it was not going to be that easy. A lot of stuff that goes on is not currently understood. I do think it's encouraging though.