Gene-editing trick discovered

natural enzyme called Cas-9 that can target a section of a genetic code and snip out mutated or damaged segments.

The problem is this is not a perfect science. CRISPR’s genome-snipping isn’t always in the right place. Scientists claim progress in limiting off-target effects, but CRISPR hasn’t matured enough as a technique to be used in human therapies.

“It’s so good at cutting the genome, it tends to make cuts at the wrong place, too. I think our technology is much harder to make, but we believe it’s much more specific, with less off-target effects,” said Peter Glazer, a Yale University geneticist and co-author of the study.

Mr. Glazer teamed with other scientists, including Danith Ly, a professor of chemistry at Carnegie Mellon, to develop a technique that takes synthetic genetic material — PNAs, for peptide nucleic acids — and injects it directly into the bloodstream of a mouse with a blood disease. The new study reports on the results of experiments on mice with the blood disease thalassemia, which is relatively common in humans and inhibits levels of hemoglobin. Thalassemia is caused by a single mutation in the genetic code. The synthetic PNAs are designed to bind with the mice DNA to form a kind of bump on the genome.

Indeed, instead of the usual double helix, that stretch of modified DNA will now have a triple helix. A cell has molecules that continually scan the DNA for something that doesn’t look right, and “triplex DNA” is just such an anomaly. So the cell deploys its repair machinery to snip away the offending DNA segment. The laboratory technique developed by Mr. Glazer, Mr. Ly and their colleagues requires a second maneuver, the deployment of a DNA patch that contains a normal, non-mutant version of the hemoglobin gene. Ideally the DNA strand will repair itself by plugging in this nonmutant genetic patch.

This is tricky stuff and only works on a small percentage of cells. But even that level of efficiency is enough to suppress the diseases studied, scientists say.

This has only been done in mice. Will it work in humans?

“That’s the goal, to try to replicate this mouse model in humans,” Mr. Ly said.