On 25 November 2018, two days before an international summit on human genome editing was due to open in Hong Kong, the MIT Technology Review broke the story that a Chinese biophysicist, He Jiankui, had used CRISPR-Cas9 to edit the genomes of human embryos and that two of those embryos had been brought to term as twin girls.
He confirmed the news within hours, in a series of videos posted to YouTube. He gave the children the names Lulu and Nana, pseudonyms used in his lab’s materials. A third child, born to a different couple in 2019, was later confirmed to exist. The international scientific community condemned the experiment almost immediately. So did the Chinese government, which placed He under house arrest, then formally charged him, and in December 2019 sentenced him to three years in prison for “illegal medical practice.”
The standard short version of the story is that He achieved a scientific first, even if the ethics were wrong. The standard short version, on the scientific record, is misleading in important ways. The experiment did happen, the children do exist, and the genetic edits were real. But the edits did not do what He claimed they did, the medical rationale was thin, and the procedure exposed the children to risks the consensus view considers indefensible.
What He Jiankui was trying to do
The stated goal was to give the embryos resistance to HIV by disabling a gene called CCR5, which produces a receptor that the most common strains of HIV use to enter immune cells. People who naturally carry a particular 32-base-pair deletion in CCR5, known as CCR5-Δ32, are largely resistant to HIV infection. The mutation is common in some northern European populations and confers genuine protective effects, though it has been associated with possible increased susceptibility to other infections such as West Nile virus and influenza.
The medical case for editing the twins was weak from the start. According to contemporaneous coverage in BioTechniques and the broader specialist press, both couples involved in He’s experiment consisted of an HIV-positive father and an HIV-negative mother. Standard sperm-washing techniques and existing antiretroviral protocols already eliminate essentially all risk of HIV transmission in IVF cycles of this kind. The children were not at meaningful risk of inheriting HIV in the first place. The case for editing their embryos to confer lifelong HIV resistance, with all the associated genetic uncertainty, did not exist.
What the edits actually produced
The scientific consensus on what He actually achieved was reached quickly, once researchers outside his lab had access to his data. The clearest published assessment came from Kiran Musunuru, a gene-editing specialist at the University of Pennsylvania, who reviewed an unpublished draft of He’s manuscript and wrote a detailed analysis for MIT Technology Review. The picture his analysis described, and that other specialists later confirmed, contained three serious problems.
The first problem was that He did not actually produce the CCR5-Δ32 mutation he was supposedly mimicking. He produced novel deletions in the CCR5 gene that had never previously been documented in humans. Whether those novel variants confer HIV resistance, no resistance at all, or some other unknown phenotype, was, and remains, unknown.
The second was mosaicism. CRISPR editing of an embryo can fail to reach every cell, and the data Musunuru saw showed that the edits in the twins were not uniform across their cells. Different cells in their bodies carried different versions of the CCR5 gene, including some that were edited, some that were edited differently, and some that were not edited at all. A mosaic body cannot be said to be “resistant to HIV” in any meaningful sense. It is a patchwork.
The third was off-target editing. CRISPR-Cas9 sometimes cuts DNA at locations other than the intended target. The published data showed evidence of off-target edits elsewhere in the genomes of the embryos that were implanted. Where those edits landed, what genes they affected, and what consequences they might have over a human lifetime, are also unknown.
Musunuru’s response, in interviews he gave to the Science History Institute and elsewhere, was that the data showed an experiment that had “gone terribly wrong.” The phrase he used in those interviews was not academic. He described looking at the chromatograms in the draft paper and crying. The twins, in his reading, had been exposed to gene editing that had failed to produce the intended outcome, had introduced unknown other changes, and could not be said to have benefited in any way that scientists could measure or predict.
The ethical case against
The ethical objections were not subtle. Editing the germline, meaning the heritable DNA that a person will pass to any future children, had been a recognised bright line in the gene-editing community for years before He’s experiment. Multiple international statements, summit consensus documents, and national regulations specified that germline editing for reproduction should not proceed without broad scientific consensus, transparent oversight, robust safety data, and a strong medical case. He’s experiment had none of these.
The 2019 court verdict in Shenzhen, reported in detail by Science, found that He and two colleagues had forged ethical review documents and had misled the doctors who implanted the edited embryos. He received the three-year sentence and a 3 million yuan fine. His two collaborators received shorter sentences. The Chinese health ministry banned all three from working with human reproductive technology again.
He was released from prison in April 2022. According to STAT News, within months of his release he began promoting a new lab in Beijing focused on gene therapy for rare diseases, and has periodically attempted to re-enter the international research community through visits, conferences, and visa applications. Hong Kong revoked a work visa he received in 2023 within hours of its announcement. His current activities remain controversial.
What the field decided afterwards
The CRISPR babies experiment forced the genome-editing community into a much more concrete conversation about heritable human genome editing than it had previously had. A 2020 report by an international commission convened by the US National Academy of Medicine, the US National Academy of Sciences, and the UK Royal Society set out specific scientific, regulatory, and consent criteria that would have to be met before any future heritable editing could be considered, and concluded that those criteria are not currently met for any condition. The World Health Organization issued its own framework in 2021. Both documents are explicit that He’s experiment did not meet, and could not have met, the relevant standards.
The deeper conclusion in the field, expressed across multiple published commentaries since 2018, is that the He experiment was not a scientific advance that arrived too early. It was a procedure that did not work as advertised, performed on children who could not consent and who could have been protected from HIV by existing, safer means. The twins, and the third child born in 2019, are the people who will live with the consequences of that decision. They are children. Out of respect for their privacy and the publication’s standard practice on minors, this article will not speculate about their current lives.
What the CRISPR babies case is, in the field’s settled view, is a cautionary record of what happens when a researcher acts outside the consensus framework that the rest of the community had spent a decade trying to build. The science of human germline editing has not stopped. The international guard-rails around it, partly because of what happened in 2018, are considerably more specific than they used to be.
