[Updated 7/16/18 4:00pm ET, see note below] As CRISPR-based therapeutics inch closer to human tests in the U.S., academic researchers have kept up a steady drum beat about potential risks of the gene editing technique. They recently published two separate papers warning about possible dangers of CRISPR therapies, namely an increased risk of cancer. Both sent shockwaves through the sector.
Today comes the latest warning sign. In a paper published in Nature Biotechnology, researchers in the U.K. show that in mouse and human cells, the CRISPR-Cas9 gene editing system can delete large chunks of DNA at and around the site it is supposed to edit, in up to 20 percent of the edited cells, inactivating genes that it shouldn’t touch. The researchers, led by Allan Bradley of the Wellcome Sanger Institute, suggest that this is happening more frequently, and the deleted pieces are bigger, than what’s been previously published in the literature. The team speculates that some of these kinds of large mutations could cause disease such as cancer.
The authors conclude that developers of CRISPR-based therapeutics need to thoroughly check edited cells—by sequencing their entire genomes, for example—before giving them to humans. “I would say our best bet now is to avoid [editing locations along the genome] where large deletions can easily cause harm,” such as near cancer-causing genes, said Michael Kosicki, the paper’s lead author, in an e-mail.
Publicly traded companies Editas Medicine (NASDAQ: [[ticker:EDIT]]), CRISPR Therapeutics (NASDAQ: [[ticker:CRSP]]), and Intellia Therapeutics (NASDAQ: [[ticker:NTLA]]) are all close to testing CRISPR drugs in humans. Shares for all three companies dipped 6 to 7 percent today by the close of the market. [updated with share price information].
There are several caveats to these findings. First, the study was primarily focused on mouse cells. The frequency of large deletions in the human cells they looked at (a type of retinal cell) was lower, in the 3 to 9 percent range, compared to the mouse cells, which ranged up to 20 percent. And the team didn’t actually show that edited cells grew into tumors.
Also, the researchers looked at only four genes along the entire genome, a small number, according to Gaetan Burgio, a geneticist and CRISPR expert at Australian National University. Bruce Conklin, who is using genome engineering to develop therapies at the University of California, San Francisco, says that some parts of the genome are more prone than others to suffering large-scale damage from DNA editing. “It is not clear how findings from Dr. Bradley will apply to other parts of the genome,” wrote Conklin in an e-mail.
Conklin adds that the researchers used high doses of the DNA-cutting enzyme Cas9 (higher than what Intellia says it’s using) for prolonged periods of time, and wonders if they would have gotten the same results with lower doses.
Tom Barnes, senior vice president of innovative sciences at Intellia, notes that the authors studied only a few types of cells, including ones that are actively dividing. Gene editing is known to delete larger chunks of DNA in those kinds of cells, Barnes says. Intellia is targeting quiescent, or non-dividing cells, in the body, which show smaller deletions, he says. And the dividing cells Intellia is editing in the lab for eventual use in patients, such as blood stem cells, so far haven’t turned into tumors after being implanted in mice, he adds. “This [paper] doesn’t cause us to change anything we’re doing,” says Barnes.
Overall, Barnes says this latest finding isn’t surprising. Intellia has already been looking for these kinds of large-scale deletions and so far hasn’t seen evidence of them in the cells they are targeting, including liver cells.
A spokesperson for Editas had a similar comment, saying that Editas scientists have looked at this and “do not believe [it] is specifically problematic in our work to make CRISPR-based medicines.”
Still, Conklin says the paper does raise some important concerns about undesirable mutations. But Burgio says it’s hard to say at this point, without more data, how much of a risk there is for CRISPR-Cas9 gene editing to trigger disease-causing mutations.
This and other studies suggesting possible unintended consequences with first-generation CRISPR-Cas9 editing all go to show why there’s growing interest in next-generation approaches. These include editing RNA, which doesn’t alter the genome, and base editing, which is more precise than CRISPR-Cas9 and involves replacing single letters of DNA rather than larger pieces.
“This explains why many in the gene therapy field are moving towards base editing as more safe than classical CRISPR editing,” wrote Burgio in an e-mail. Beam Therapeutics, for instance, is trying to turn base editing and RNA editing into human medicines. It launched earlier this year with a $87 million Series A round of venture funding.
