‘Three-parent IVF’ was supposed to prevent mitochondrial disease. Studies are raising doubts

The news out of the U.K. last year that eight healthy children had been born to mothers carrying disease-causing mutations in their mitochondria showed the promise of a pioneering reproductive technique known as “three-parent IVF.” The long-anticipated results seemingly showed the procedure, formally called mitochondrial replacement techniques, or MRT, could reduce the risk of babies inheriting sometimes fatal diseases, and raised hopes that other countries, such as the U.S. where MRT is banned, might soon permit it. But even as regulators and researchers in Australia, the second country to explicitly legalize the specialized technique, prepare to launch their own national experiment with MRT, some scientists and ethicists are rethinking the potential risks and benefits of the technology, which combines genes from a mother and father with mutation-free DNA in the mitochondria of a donor egg. Emerging data from the U.K. as well as several small pilot studies from fertility clinics in other parts of the world are raising questions about whether trying to prevent mitochondrial disease is really the best first use of MRT. “The general public may have the idea that it is very safe and low risk and that’s just not the case,” said Bert Smeets, a professor in clinical genomics who studies mitochondrial diseases at Maastricht University in the Netherlands. Until now, pushback against MRT has mainly focused on its expanding use as a treatment for infertility. A STAT investigation published on Monday showed that IVF clinics globally are pitching it to older women desperate to have biologically related children despite scant evidence that it works for them. But some scientists are now saying that carefully run studies addressing severe forms of infertility may present the most ethical way to develop MRT safely. Smeets serves on the oversight committee that reviews and approves applications for patients to receive MRT at Newcastle University, the only clinic that currently has a license in the U.K. to perform it. Newcastle is where one of the main methods of MRT, known as pronuclear transfer, was first developed in the early 2000s. It involves creating embryos from both the biological mother and the egg donor and then swapping out the nuclear DNA in one for the other. Data published in 2010 by Newcastle scientists showed that in lab experiments, only a tiny amount of the mother’s mutation-bearing mitochondria — less than 2% — showed up in the reconstructed embryos. The technique looked safe and showed promise in preventing mitochondrial disease, they reported, though not very efficiently (meaning a lot of embryos didn’t survive the process). Shortly after U.K. regulators began accepting applications from clinics to offer MRT, this picture grew more complicated. Shoukhrat Mitalipov, a developmental biologist at Oregon Health & Science University, had been developing his own method for MRT, called maternal spindle transfer. It involved putting a mother’s genetic material into a donor egg with healthy mitochondria prior to fertilization, when the egg’s chromosomes are lined up in a structure called the spindle. In 2009, Mitalipov’s lab had reported the births of the world’s first monkeys following maternal spindle transfer. The animals had appeared healthy with no traces of mitochondrial DNA left over from the original eggs. But when his group tried the technique in the lab with human eggs, they noticed something odd. Despite refining the spindle transfer to the point where his lab could swap up to 99% of defective mitochondria with healthy replacements, analyses run by one of his postdocs showed that the few problematic ones that had managed to cling to the spindle began to take over as the fertilized egg continued to divide. With each subsequent generation of cells, they saw fewer and fewer of the healthy mitochondria — the power plants of cells — and more and more of the mutated ones. Before Mitalipov’s team fully understood why or how that happened, they rushed out a paper describing this problem, which they dubbed “reversion.” Their data, published in November 2016, suggested that in roughly 15% of cases, reversion could cause MRT to fail, allowing fatal defects to resurface or new health issues to arise. The revelation caused a stir but ultimately did not prevent the U.K. from moving forward with the Newcastle group’s plan to use pronuclear transfer. The group won a license for the work in 2017. Mitalipov, stymied from pursuing his own clinical trial of spindle transfer in the U.S. by a 2015 congressional ban, began evaluating the technology as a potential fertility treatment. He and his collaborator Nuno Costa-Borges of the Spanish company Embryotools wanted to test the idea that young mitochondria might be able to help overcome poor egg quality — one of the reasons IVF fails. They recruited 25 women under 40 years old who’d had multiple IVF failures to a small study conducted in Greece. The results, published in 2023, showed that spindle transfer could produce pregnancies. But it also showed that reversion wasn’t just theoretical; it happened in the IVF clinic too. Two of the seven children born carried a mix of mitochondria from the biological mother — as high as 51% in one child and 35% in another — and the donor. Two groups have now found that reversion can happen with pronuclear transfer, too. When the Newcastle University team published papers last year about the births of the eight babies, the researchers noted that although the children showed no signs of disease, they found low levels of reversion, between 5% and 16%, in three of the infants. A Ukrainian group from the Nadiya Clinic, led by its director Valery Zukin and embryologist Pavlo Mazur, has also found evidence of reversion. Out of six children whose parents allowed testing, two were found to have reversion — one with maternal mitochondrial levels up to 12% and the other with levels as high as 72% — according to unpublished data shared with STAT. In the case of the Nadiya Clinic, pronuclear transfer was being tested as a fertility treatment, and the mothers’ mitochondria carried no apparent disease-causing mutations, so the finding of reversion is less of a concern, at least in terms of known risks, said Jacques Cohen, a New York-based embryologist who has reviewed the Ukrainian group’s data. But it adds to the growing evidence that even a tiny amount of maternal mitochondria carryover could compromise the efficacy of MRT, which would be a much bigger deal for people trying to avoid passing down mitochondrial diseases to any future children. “To do it in a family affected by mitochondrial disease is to take a very big risk,” Cohen said. “The field is showing that it’s a bad idea.” Cohen, who would like to see a moratorium on the use of MRT for preventing mitochondrial disease, is at one extreme of an ethical debate about what to do with this emerging information. He and others think that it’s most prudent to pause studies involving current iterations of the technology until scientists can figure out how to minimize the risks of reversion. That would allow scientists in the U.K. to continue to follow up on the children already born to see if the amount of faulty mitochondria continues to increase over time and to monitor for any resulting health complications. Guido Pennings, a bioethicist at Ghent University in Belgium, has argued that because of the residual risk of serious disease to the offspring of carriers of mitochondrial mutations, MRT should be developed in carefully controlled trials only for infertile couples. That would enable researchers to sort out the mystery of reversion — insights that could lead to more refined methods of MRT — without exposing unborn children to known risks. It would also allow scientists to study the unknown risks posed by the mismatch between the donor’s mitochondrial DNA and the mother’s nuclear DNA. (Mitochondrial DNA is typically inherited from the mother.) Some research in mice suggests this phenomenon could lead to immune issues or other unintended health complications. There have also been reports that mice carrying a mix of mitochondrial DNA from two animals had behavioral, cognitive, and other abnormalities. Most ethicists and scientists who spoke to STAT wouldn’t go as far as Pennings. But there is an increasing recognition that prohibitions of trials of MRT in the context of infertility should be revisited. In 2024, Esther Braun, of the Institute for Medical Ethics and History of Medicine at Ruhr University in Germany, laid out a case for changing the laws in the U.K. and Australia to allow for clinical trials for treating infertility, highlighting that previous ethical assessments may have underestimated both the risks of MRT for preventing mitochondrial disease and the suffering caused by involuntary childlessness. That shouldn’t be taken as an ethical green light for a burgeoning global MRT tourism industry, she emphasized. “It seems evident that the routine clinical use of a novel medical procedure is impermissible until clinical trials have established that the procedure is effective,” Braun wrote. “However, this does not imply that conducting such clinical trials is itself impermissible.” A spokesperson for the U.K.’s fertility regulator — the Human Fertilization and Embryology Authority — told STAT in an email that its scientific committee has considered research exploring the use of MRT for infertility, but that for now, only people who are at very high risk of passing a serious mitochondrial disease onto their children are eligible for this treatment. Over the past year, a few research teams from around the world have made policy recommendations for how such clinical trials might responsibly proceed. In March, a group led by Smeets in the Netherlands proposed a staged approach to first identify and define types of infertility that MRT might plausibly address, then study its impact in the lab before moving on to small pilots to test its effectiveness only in patients exhibiting those specific infertility conditions. That would allow for the field to more critically evaluate which conditions MRT could safely tackle. The urgency has increased, Smeets said, as MRT is increasingly being adopted by commercial clinics as an all-purpose add-on, despite a lack of clear evidence. “As a biologist, I always want to know that the intervention I’m doing targets the cause of something you want to solve,” Smeets said. “So this might be relevant to do this for certain subgroups of infertility where changing the mitochondria and the maternal factors in the oocyte is really allowing the embryo to develop. But that’s not the age-related cases.” An example of what Smeets’ group has proposed is already underway in neighboring Belgium, at Ghent University. For the past few years, Dominic Stoop, a fertility specialist, and Björn Heindryckx, a reproductive geneticist, have been tracking down the molecular culprits behind the most difficult-to-treat infertility cases. They’ve helped define clinical criteria for diagnosing those kinds of cases, commonly grouped together under the term OZEMA — for oocyte/zygote/embryo maturation arrest. In a not-yet-published study spanning tens of thousands of IVF cycles at the Ghent University Hospital’s fertility clinic and a collaborating clinic in the U.S., Boston IVF, they found that OZEMA seems to occur across all ages more or less equally. Stoop’s team has asked Belgian regulators for approval to study the effect of nuclear transfer technologies in three different patient populations — individuals with a history of mitochondrial disease, OZEMA patients, and women of advanced maternal age. The plan, he told STAT, is to treat seven patients in each group to start. So far, the government has given the green light for the first two, which Stoop hopes to be recruiting for this fall, with the first results out next year. His team is still generating preclinical data to justify doing it for the advanced maternal age group. “The patient population where the ethical committee finds the least problems is, in my opinion, the most risky group,” Stoop said, referring to the families with mitochondrial disease. “For the advanced maternal age and the OZEMA patients, it’s different. There may be a risk we don’t know. But if there’s a risk there, it will also be in the mitochondrial disease patients on top of the risk of these abnormal mitochondria. So these patients need very good counseling before doing it.” Mary Herbert — a former leader of the Newcastle group who relocated to Monash University in Melbourne several years ago — said communicating both known and unknown risks to parents, and monitoring the long-term health of any resulting children, are essential for ethical use of MRT. “We have known for a decade that MRT may not work in all cases,” she told STAT. Her own preclinical research, in addition to data from the Mitalipov group, suggested that MRT could reduce the risk of mitochondrial disease but could not guarantee prevention. That’s how it should be described to potential patients, not “as a means of preventing” mitochondrial disease, she said. So far, clinical data from the Newcastle program have shown that MRT resulted in a 75% to 100% reduction in transmission of disease-causing mitochondrial mutations, she said. Even in the cases where reversion was observed, mutated mitochondrial levels were below the 50% threshold where disease symptoms are expected to arise. Herbert’s group at Monash has been training embryologists on MRT and is preparing to ask the Australian government for permission to begin a clinical trial in women carrying mitochondrial mutations. It would include follow-up studies of any carried-over mutated mitochondrial DNA “to ensure that levels remain stable over time and across different organ systems,” Herbert said. STAT’s coverage of bioethics is supported by a grant from the Greenwall Foundation and the Boston Foundation. Our financial supporters are not involved in any decisions about our journalism.