It might seem that scientists have never met a chunk of DNA they couldn’t edit in mice or isolated cells using CRISPR — from mutations causing deafness to those for Duchenne muscular dystrophy. In fact, they are learning what every pencil- or Word-wielding editor knows: It’s much easier to improve something that’s in terrible shape than writing that’s near perfect.
In genome-editing, the challenge for CRISPR-wielding scientists is to edit only one of the two copies, or alleles, of every gene that people have, repairing the ever-so-slightly broken one and leaving the healthy one alone.
Now, in one of the first research papers scheduled for publication in the first journal dedicated to research on CRISPR, scientists in Boston report “allele specific” editing of a gene that, when mutated, destroys the eye’s photoreceptors and causes the form of blindness called retinitis pigmentosa.
The achievement might one day help people with retinitis pigmentosa, which affects about 100,000 people in the U.S. But its greater significance is as a proof-of-concept. The hope is that the same trick might work in the hundreds of diseases, including Huntington’s disease and Marfan syndrome, where inheriting a single mutated gene (from mom or dad) is enough to cause problems despite the presence of a healthy copy, too.
“You want to target only the mutant allele without messing up the healthy one,” said Linzhao Cheng, of Johns Hopkins University School of Medicine, who is developing allele-specific techniques for blood disorders. “But the alleles might differ in only one nucleotide,” one of the molecular “letters” that spell out the genetic code. “That makes allele-specific editing probably the most challenging situation for CRISPR.”
The Boston scientists, led by Dr. Qin Liu of the Ocular Genomics Institute at Massachusetts Eye & Ear Infirmary, aimed to remove the misspelled copy of the gene for rhodopsin, which makes up the rods (of rods and cones fame) in the eye. The misspelling consisted of a single wrong nucleotide. That seemingly minor glitch, called P23H, is enough to produce a rogue rhodopsin that is toxic to the healthy rhodopsins produced by the healthy copy.
“It just kills the photoreceptors,” said Dr. Stephen Rose, chief research officer at the Foundation Fighting Blindness, which helped fund Liu’s research. “But what if you could repair that one mutation and turn it back to the normal form? That’s the holy grail, to wave a magic wand and change a single wrong nucleotide to the right one.”
That’s what Liu and her colleagues report doing in the paper to be published in The CRISPR Journal, whose first issue is due this month. They built standard CRISPR molecules: a target-finding molecule called a guide RNA and a snip-the-nucleotide enzyme, in this case a version of Cas9. They injected their CRISPR molecules under the retinas of days-old mice bred to have one good rhodopsin gene and one mutated copy.
The editing flopped. The target-finding molecule couldn’t tell the healthy gene from the one-letter-off copy.
Back at the drawing board, the scientists, who included Editas Medicine (EDIT) co-founder J. Keith Joung of Massachusetts General Hospital, created target-finding molecules that looked for shorter regions of DNA, hoping to avoid editing the healthy gene. That produced better results: Cas9 edited only the mutant allele. But it did so in very few of the cells. As long as there is a lot of mutant rhodopsin compared to healthy rhodopsin, the mutant proteins will kill the eyes’ rods.
The third time was the charm.
In addition to using the short target-finding molecules, the scientists also tweaked Cas9 so it made a beeline for tiny DNA mile markers (called PAMs). The mile markers nearest the disease-causing allele are, luckily, different from those near the healthy one. Including the go-to-PAM instruction in their CRISPR produced accurate editing and a lot of it: There were nearly three times as many healthy rhodopsin molecules as mutant ones, compared to similar numbers of healthy and mutant rhodopsin in cells that had not been CRISPR’d. That translated into healthier eyes, with treated mice having five or six rows of photoreceptors compared with three or four in untreated mice. (The mice were not tested for eyesight, however.)
As always with CRISPR, there is a danger of editing unintended regions of DNA. The scientists checked potential “off target” sites; nine were fine, and one was inadvertently edited in 3 percent of treated cells, though with no apparent ill effects.
“It’s nice work,” said biologist Tara Moore of Ulster University, who is developing allele-specific CRISPR editing for eye diseases. Exploiting DNA’s tiny mile markers, the PAMs, offers the best shot at allele-specific editing, she said: “Otherwise it’s “a challenge,” and the chance of hitting the disease-causing DNA but sparing the healthy copy “is low.”
The mouse study raises hopes that allele-specific editing might work not only for the mutation in retinitis pigmentosa but also “for most, if not all, human dominant alleles,” the scientists wrote.
Liu said she was not permitted to speak to reporters about the paper until The CRISPR Journal published it. The study was also funded by the National Institutes of Health and Mass. General.
Optimism over genome-editing to prevent or cure diseases may turn out to be overblown, but for now scientists feel like the sky’s the limit. “Over time, one way or the other, we’ll get a complete [genome-editing] toolbox that let’s us change anything we want and only what we want,” said Hopkins’s Cheng.