Last week, when a British reporter broke the news that American scientists had used Crispr to edit the first human embryos on US soil, he wasted no time in cutting to the big, juicy, highly controversial chase. “One Giant Step For Designer Babies” ran the headline on Steve Connor’s world exclusive in the i, a London-based online newspaper. A similar report appeared at the same time in MIT Technology Review, though with a far more subdued title. But in both stories, details about the exact experiments were scarce, because the academic paper summarizing the work was still under peer review. Now the study is out, published online Wednesday morning in the journal Nature. And there’s a lot more to talk about.
In the past two years years, boundary-pushing reproductive biologist Shoukhrat Mitalipov led researchers at Oregon Health and Science University, the Salk Institute, and Korea’s Institute for Basic Science through a series of experiments designed to correct a genetic defect in viable embryos. A mutation in MYBPC3 causes a heart condition known as hypertrophic cardiomyopathy that affects one in 500 people—the most common cause of sudden death among young athletes. Using Crispr-Cas9, they successfully replaced the flawed gene with a normal one for 42 out of 58 embryos, the most successful demonstration of the technique’s gene editing prowess in the human germline. And while the mutation-correcting mechanism was highly efficient, it wasn’t the one that Mitalipov, or anyone, was expecting.
Before Mitalipov’s team could edit the first embryos in the US, they had to make them. So they took sperm from a guy with a mutation in his MYBPC3 gene and used it to fertilize eggs from 12 healthy females. In addition to the sperm, they also injected each egg with Crispr-Cas9 protein, a guide RNA directing it toward the mutant copy of MYBPC3, and a piece of template DNA, modeled after the normal gene but with a few tags so scientists could find it again later. The idea was for Crispr to cut out the mutant copy and the embryo’s repair machinery to use the supplied template to build a normal gene in its place.
And it worked—surprisingly well. Past Crispr experiments in China have run into issues; sometimes, not every cell in the embryo gets repaired, or Crispr cuts things it shouldn’t. Even earlier attempts by Mitalipov’s team to edit MYBPC3 in stem cells with Crispr ran into similar issues. But when it came to the embryos they injected at the exact time of fertilization, they saw very low rates of either of these failures.
But one thing didn’t work at all the way the scientists expected. Of the 42 successfully corrected embryos, only one of them used the supplied template to make a normal strand of DNA. When Crispr cut out the paternal copy—the mutant one—it left behind a gap, ready to be rebuilt by the cell’s repair machinery. But instead of grabbing the normal template DNA that had been injected with the sperm and Crispr protein, 41 embryos borrowed the normal maternal copy of MYBPC3 to rebuild its gene.
Which is why Mitalipov insisted on the title given to their paper: “Correction of a Pathogenic Gene Mutation in Human Embryos.” “Everyone always talks about gene editing. I don’t like the word editing. We didn’t edit or modify anything,” Mitalipov says. “All we did was unmodify a mutant gene using the existing wild type maternal gene.”
The next step will be to see if they can replicate this “unmodifying” effect in different mutations. The MYBPC3 gene had four messed up base pairs, so it was pretty easy for Crispr-Cas9 to find and replace. But other mutations might be off by just a single letter, which would be harder to fix. There’s always a chance that MYBPC3 will have been a case of beginner’s luck, so they want to make sure the effects are generalizable to other common mutations, like the BRCA genes associated with increased risk for breast and ovarian cancers.
Crispr experts around the globe were quick to celebrate the work while pointing out its many limitations. “This is a remarkable paper that shows how much the field has progressed in just the last year or two,” says Gaetan Burgio, a geneticist at the Australian National University. “But I think for now everyone needs to chill down a bit. The scope is very limited, and it’s unlikely to me that Crispr would be a substitute for preimplantation genetic diagnosis, whatever the authors say.” Burgio is referring to the genetic profiling of embryos prior to implantation via IVF—it’s a way to screen for mutated genes like MYBPC3 and only select the 50 percent of embryos that are normal.
Mitalipov and his coauthors argue that their Crispr technique can get that number up to around 75 percent, maybe even 100. Which would prevent prospective moms, especially older ones, from having to go through multiple rounds of costly, unpleasant egg harvesting.
But validating that kind of treatment would require lengthy clinical trials—something a rider in the current Congressional Appropriations Act has explicitly forbidden the Food and Drug Administration from even considering. Mitalipov said he’d have no problem going elsewhere to run the tests, as he did previously with his three-person IVF work. Before that, he’d need to re-run these experiments in animals, and implant the embryos to assess them at different developmental stages for any abnormalities. Collaborators like Jun Wu at the Salk Institute will likely follow up in another way, with more stem cell studies, to see if the Crispr corrections follow cells through all their different lineages—into neurons and liver cells and heart cells.
If there’s anything Wu and Mitalipov and the rest of their team have learned through all this, though, it’s that stem cells and embryos are not re-created equal. Because the early days of embryonic development are so tumultuous, with lots of dividing and recombining, those cells might have special ways to avoid genetic mishaps—like, say, copying a random piece of DNA that a scientist stuck into a cell. Evolution may have made it harder than anyone thought to subvert its will with superbaby genes.