In an ethically charged first, Chinese researchers have used gene editing to modify human embryos obtained from an in vitro fertilization clinic.

The 16-person scientific team, based at the Sun Yat-Sen University in Guangzhou, China, set out to see whether it could correct the gene defect that causes beta-thalassemia, a blood disease, by editing the DNA of fertilized eggs.

The team’s report showed the method is not yet very accurate, confirming scientific doubts around whether gene editing could be practical in human embryos and whether genetically engineered people are going to be born anytime soon.

The authors’ report appeared on April 18 in a low-profile scientific journal called Protein & Cell. The authors, led by Junjiu Huang, say there is a “pressing need” to improve the accuracy of gene editing before it can be applied clinically—for instance, to produce children with repaired genes.

The researchers did not try to establish a pregnancy and say that for ethical reasons they did their tests only in embryos that were abnormal.

“These authors did a very good job pointing out the challenges,” says Dieter Egli, a researcher at the New York Stem Cell Foundation in Manhattan. “They say themselves this type of technology is not ready for any kind of application.”

The paper had previously circulated among researchers and had provoked concern by highlighting how close medical science may be to tinkering with the human gene pool (see “Engineering the Perfect Baby”).

In March, an industry group called for a complete moratorium on experiments of the kind being reported from China, citing risks and the chance they would open the door to eugenics, or changing nonmedical traits in embryos, such as stature or intelligence (see “Industry Body Calls for Gene-Editing Moratorium”). Other scientists recommended high-level meetings of experts, regulators, and ethicists to debate whether there are acceptable uses for such engineering (see “Scientists Call for a Summit on Gene-Edited Babies”).

The Chinese team reported editing the genes of more than 80 embryos using a technology called CRISPR-Cas9. While in some cases they were successful, in others the CRISPR technology didn’t work or introduced unexpected mutations. Some of the embryos ended up being mosaics, with a repaired gene in some cells but not in others.

Parents who are carriers of beta-thalassemia could choose to test their IVF embryos, selecting those that have not inherited the disease-causing mutation. However, gene editing opens the possibility of germ-line modification, or permanently repairing the gene in an embryo, egg, or sperm in a way that is passed onto the offspring and to future generations.

That idea is the subject of intense debate, since some think the human gene pool is sacrosanct and should never be the subject of technological alteration, even for medical reasons. Others allow that germ-line engineering might one day be useful but needs much more testing.  “You can’t discount it,” says Egli. “It’s very interesting.”

The Chinese team performed the gene editing in eggs that had been fertilized in an IVF clinic but were abnormal because they had been fertilized by two sperm, not one. “Ethical reasons precluded studies of gene editing in normal embryos,” they said.

Abnormal embryos are widely available for research, both in China and in the United States. At least one U.S. genetics center is also using CRISPR in abnormal embryos rejected by IVF clinics. That group described aspects of its work on the condition that it would not be identified, since the procedure remains controversial.

Making repairs using CRISPR harnesses a cell’s own DNA repair machinery to correct genes. The technology guides a cutting protein to a particular site on the DNA molecule, chopping it open. If a DNA “repair template” is provided—in this case a correct version of the beta-globin gene—the DNA will mend itself using the healthy sequence.

The Chinese group says that among the problems they encountered, the embryo sometimes ignored the template, and instead repaired itself using similar genes from its own genome, “leading to untoward mutations.”

Huang said he stopped the research after the poor results. “If you want to do it in normal embryos, you need to be close to 100 percent,” he told Nature News. “That’s why we stopped. We still think it’s too immature.”