Why the paper on the CRISPR babies stayed secret for so long

More than a year after the birth in China of twin girls known as Lulu and Nana, the world’s first gene-edited babies, the affair is still shrouded in secrecy. US researchers and universities have given incomplete or equivocal accounts of their involvement with He Jiankui, the Chinese biophysicist who used CRISPR to make changes to the girls’ DNA while they were still embryos. In China, if you distribute a news story to WeChat asking what happened to the twins, state censors will issue a takedown notice.

No reason is given. No appeal is possible.

The silence hasn’t served only to conceal what really happened to the girls. It is hiding the scientific facts themselves. Starting late last year, manuscripts written by He describing the creation of the twins were considered for publication by at least two supremely influential journals: Nature and JAMA, the Journal of the American Medical Association. Neither has published his work.

The reason isn't only that He's project trampled ethics rules. Another major obstacle to a full account is that He has not been seen or heard from for months. He didn’t make it to his home village for Chinese New Year in February, his father told us. His lab and data, according to one insider, were seized by Chinese authorities last December, and his original team of 10 has scattered to the four winds. An American collaborator, Michael Deem of Rice University, is the subject of an investigation by that institution; it has come to no public conclusion or disclosed any findings. So there may be nobody who can answer questions, expand upon the data, or carry out follow-up experiments, as scientific review by a journal often demands.

Although the reaction to the CRISPR babies was overwhelmingly negative, the future that the unpublished manuscripts unveil—a future of genetically engineered humans—is coming faster than many people realize. Genome-writing techniques are improving at a blazing pace. Select researchers remain keen to employ them in human embryos, tempted by the chance to prevent disease or improve heredity. The fear is they will do it again in secrecy, in some other country with lax oversight, and repeat He’s mistakes.

Today, MIT Technology Review is reporting excerpts from He’s unpublished manuscript about the creation of the twins for the first time. You can read those excerpts here, together with comments from medical and legal experts on why they think He’s research was seriously flawed. In another piece, Kiran Musunuru, a gene-editing specialist at the University of Pennsylvania, argues that these flaws are precisely why the contents of He’s paper need to be made public.

And in this story, we trace the journey of the manuscript, and the ethical and legal quandaries that have kept it out of the public eye.

Near dawn on the morning of November 26, 2018, in China, He Jiankui was about to awake to a nightmare. News was going to leak about a secretive plan he’d pursued for two years to create the first gene-edited babies. They’d been born the month before, and he’d prepared a scientific manuscript, titled “Birth of Twins After Genome Editing for HIV Resistance." As he slept, he might have dreamed of being toasted for a medical first and seeing his name on a paper that, decades from now, would be still be collecting citations.

The twins, given the pseudonyms Lulu and Nana, had developed from embryos altered with the gene-editing tool called CRISPR. The edits were intended to disable a gene called CCR5, a change that could theoretically make the girls resistant to HIV. After their birth, He had reason to make what he thought was a big success publicly known, but some of his associates worried he was moving too quickly and tried to get him to slow down. According to an insider with knowledge of events, a US advisor, Stephen Quake of Stanford University, told He to try for publication in Nature, by some measures the most important journal in the world, where Quake offered contacts. Quake did not reply to a request for comment.

By about November 19, a week before a major genome-editing summit in Hong Kong where He planned to speak, the Chinese team had submitted its research to Nature for review. Publication in the journal would attest to the quality and importance of the work, even if the peer-review process meant it wouldn’t appear for several months. During that time, remaining holes in the science could be plugged. Some of those gaps were significant: He still couldn’t fully rule out the possibility that he’d introduced errors, or “off target” changes, into the twins’ DNA. Proving that he hadn’t would take time.

Meanwhile, He provided copies of his manuscripts to the Associated Press and allowed its photographers into his lab. The news agency distributed copies, under strict confidentiality, to at least three senior American researchers to gather their reactions for an eventual story.

One of them was Musunuru, the gene-editing specialist whose op-ed we have also published. He was aghast at what he saw. Musunuru felt there was evidence in the data that the editing had gone awry, and in ways that the paper about the twins didn’t acknowledge. That Thursday, Thanksgiving, his friends and relatives wondered what was wrong; he made excuses, since the news was not yet public.

Nature, according to three people, initially concluded that the research was worth sending out for peer review. (A spokesperson for Nature declined to comment, citing the confidential process of submissions.) But it was He’s efforts to dot his i’s for publication that tripped up his plans. Nature’s policies require that before authors submit a manuscript, they register any human clinical trials in a public repository. On November 8, He retroactively posted a description of the trial, its ethics approval, and key genetic data on the twins to the China Clinical Trial Registry, an online catalogue of trials under way. MIT Technology Review found the documents and published a story, titled “Exclusive: Chinese scientists are creating CRISPR babies,” on November 25. It described an attempt to create people immune to HIV, with pregnancies at least six months along.

After we published the story, He was compelled to make his dramatic claims public almost immediately via a series of YouTube videos; the AP also published the detailed report it had been working on. He announced that the twin girls had been born and explained his motives: HIV is highly stigmatized in China, and he believed making people genetically immune from birth would help stop the epidemic.

An early news item in Chinese state media—since deleted—hailed the project as a big national success. But the opinions that rolled in over the next 24 hours were devastating. One of He’s heroes, Feng Zhang of MIT (who He believed was the true inventor of CRISPR), quickly called for a moratorium on CRISPR babies, citing his deep concern about the “lack of transparency” in the Chinese trial. He’s home institution, the Southern University of Science and Technology in Shenzhen, also distanced itself, saying it wasn’t even aware of the research.

On November 26, a Monday, the editors at Nature concluded that the submission they had received had become untouchable and canceled their review. The reason they gave, said one person copied on some of the correspondence, was that He’s home institution had disavowed knowledge of the work. Nature would later, in an unsigned editorial, accuse He of going “rogue” and flouting “conventions of safety and research ethics,” and wonder who could have prevented it by “raising the alarm.”

Though rebuffed by Nature, He would soon have a very public opportunity to show his work to fellow scientists. On November 28, he spoke at the Second International Summit on Human Genome Editing in Hong Kong, offering key conclusions and showing extensive data directly from his manuscript on screen. He walked cautiously onto the stage and then spoke: “First, I must apologize. These results leaked unexpectedly," he said. Although—or perhaps because—he was by then being called “China’s Frankenstein,” following scientific protocol was at the front of his mind.

Unlike other speakers, He did not give the organizers a copy of his slides, according to the National Academies, which co-hosted the summit. But there were nearly two million people watching online and film crews present, too. A set of screen captures would quickly become the main source of information for other scientists trying to understand his results.

Afterwards, fleeing the venue with concerns for his security, He would return to Shenzhen. Even as he did so, his team was already exploring a new option: placing a draft of his research on the preprint server bioRxiv, a fast-growing website where researchers post their work before peer review. Tens of thousands of scientists use bioRxiv to gather feedback, scoop competing labs, or communicate research quickly, without waiting for it to appear in a journal.

As the day progressed, Ryan Ferrell, an American working with He on his public relations plans, emailed with editors at bioRxiv. Ferrell wanted to know if He’s research could quickly find a place there. It would be a decisive and rapid way to make the data public for all to see, letting the world’s experts weigh in.

But editors at bioRxiv said they might also reject the work if it described anything illegal or unethical. They told Ferrell they would need 48 hours to reach a decision, but they never heard back from him.

“You have to draw a line somewhere,” Richard Sever, one of bioRxiv’s editors, would later say on Twitter. “You really want this? ‘Dear Dr. Mengele, Thank you for your submission. Your paper will be online shortly.’” It was a reference to the Auschwitz death camp doctor Josef Mengele, who sought out and murdered twins as part of what he viewed to be scientific research. “Some people raise the question of censorship. But we don’t want [this] to be a place where people can put things that no one else will touch,” says Sever.

He quickly developed a new publishing option, one that hasn’t been previously reported. By December 3, a manuscript about the twins was out for review by JAMA, a publication owned by the largest US doctors’ association. According to one source, JAMA actively solicited the bombshell manuscript.

In the editor-in-chief of JAMA, Howard Bauchner, a pediatrician, He found someone willing to keep an open mind. While some people think editing embryos is just wrong, Bauchner is among those who believe that it may eventually be good medicine, a totally new way to prevent disease. “I think it’s inevitable that it will move ahead, and I think it should move ahead,” he says. Organ transplantation, which saves tens of thousands of lives a year, and in vitro fertilization, which leads to half a million births a year, were accompanied by profound ethical debates, which waned as they proved successful. “Oftentimes early scientific breakthroughs are seen as unethical, and then over time that changes,” says Bauchner.

Still, JAMA took unusual precautions with He’s manuscript. The journal asked 11 experts to review it, according to one of them, George Church of Harvard Medical School. That’s about three times as many as usual.

The review hinged on both scientific and ethical issues. On paper, He had taken elaborate steps to inform the couples in the trial about what was going on, and let them choose whether to participate. But there were inexplicable irregularities, too. For example, Church says it’s not clear whether the twins were conceived and born at a different hospital from the one that approved the ethics of the experiment. In a short statement published in January—still the only comment Chinese authorities have given on their investigation—the Xinhua News Agency referred to a “fake ethical review certificate.”

Might He’s team have been making a well-intentioned attempt to protect the patients’ privacy? Perhaps. But it meant the scientists were capable of deceit, which would be a serious obstacle to getting the research published. “This could set a precedent for publishing papers with less-than-ideal human-subjects approval,” says Church. That is, it could encourage the next He.

According to Church, JAMA rejected the manuscript last January. But we have also heard that the review might be ongoing in some way, or that JAMA might be seeking to publish it in some atypical format, perhaps simply as a “rejected” manuscript.

Bauchner, the JAMA editor, said he could not comment on confidential submissions. But he outlined a dilemma. Some people believe that the ethical questions surrounding He’s research make it unpublishable, perhaps forever, even if it contains important information about a technology likely to be used again.

“If someone thinks it’s scientifically valid but not ethical, does that mean the study will stay in the shadows?” wonders Bauchner. “It’s such an interesting question.”

Journals have increasingly been forced to act as a sort of ethical police. Two years before the CRISPR babies affair, Nature had already vowed to subject research on editing laboratory embryos to extra rounds of ethical scrutiny. Now, in the wake of the scandal, journals are set to take on a more formalized policing role.

This summer a panel of experts proposed that the World Health Organization set up an international registry for all gene-editing experiments, and that journals refuse any papers from projects not on the registry. The WHO is even discussing how whistle-blowers might help report baby-making attempts. Such rules could be a relief to editors—they’d no longer have to make right-or-wrong decisions entirely on their own, under the secrecy of confidential review. Magdalena Skipper, the editor of Nature, tweeted that the plan for a registry was “very satisfying.”

Journal editors and scientists to whom He sent his research paper have kept it confidential. But earlier this year a source sent us Microsoft Word versions of two draft manuscripts—one about the twins, the other discussing related laboratory research on embryos—with no conditions attached. Other versions, including a combined version, may also exist.

Much what is in the twins manuscript is already public, thanks to the media coverage and He’s own brief presentation of the data at the Hong Kong summit last year. But should the rest be published or simply made public? We polled a dozen experts. They had mixed opinions, but what they had in common was that all opposed its publication in a scholarly journal, and all wanted to see it themselves.

“I am curious to see [the manuscript] but I hope it doesn’t get published in a legitimate journal,” says Paula Amato, an IVF doctor at the Oregon Health Sciences University. She works with a team editing embryos, but says it remains premature to create children from them. “I agree with journals being a gatekeeper and not publishing unethical research, because publishing in a high-impact journal comes with notoriety and publicity, and they are widely read,” she says. “I think the problem is that he did it at all.”

Rudolf Jaenisch, a professor of biology at MIT, was among the scientists He told of his intentions early on, before the pregnancies began. Jaenisch says he thought the idea “was nuts” and said so. He, who kept detailed records about who supported his plans, never told Jaenisch any more. I showed him our copy of the manuscript about the twins, which he hadn’t seen. “It’s not very good,” he said. “No self-respecting journal would publish it.”

One argument in favor of placing the complete text of the manuscripts in the public record is that it would allow experts, governments, and anyone else with an internet connection to read it and improve their understanding of what occurred in China. As Musunuru writes in the accompanying op-ed, “He’s work reveals serious, unsolved safety concerns. It’s not clear that any effort to directly edit human embryos, even if done ethically and with full social approval, can reliably avoid these problems … It’s time for the scientific community to fully understand what happened with Lulu and Nana and to avoid stumbling down a path toward further ill-starred experiments with clinical germline gene editing.”

For all the information it contains, the twins manuscript excerpted in the accompanying item is not going to assuage skeptics. Many say what's needed now is an independent scientific effort to determine what really happened in China, starting with some hard proof that the edited girls exist.

“The first piece of data we need to have is verification from authorities in China that these girls JK [He] says he modified exist. As far as I know, nobody has said we took a drop of blood and verified the CCR5 edit,” says Dana Carroll, a gene-editing scientist at the University of Utah. “It would be simple to do—very straightforward. What I want to know is whether he did it.”

Whether that should happen—or whether the twins should be left alone, and possibly never even learn they were the subject of an infamous experiment—is a matter of debate. Without robust public scrutiny of He’s existing scientific claims and data, it’s harder to say what responsibilities the Chinese state has to monitor the children and make its findings known. Allowing all scientists to see He’s manuscript “might influence whether or in what way one might best follow up on the health of the twins,” says Church, including whether such follow-up “should [be] encouraged or not.”

He had extensive plans for this type of follow-up, according to the “discussion” section of his manuscript—a final part of a paper in which scientists can speculate and outline next steps. In that section, the authors outlined their plan to “further confirm” whether the girls are resistant to HIV.

In truth, they had no confirmation to begin with, but they did intend to gather evidence. They planned to do that by taking blood from the girls and exposing their T cells, a part of the immune system, to the virus and observe whether these cells resisted it. And because they saw a need “to monitor for any potential long-term consequences of genome editing,” they intended “to monitor the health of the twins for 18 years,” or until they became adults. In fact, they’d planned to pay for private health insurance until that time. Wherever He is now, and whatever he ends up doing, it seems unlikely that he will be able to complete the experiment.

With reporting by Michael Standaert

Correction: The World Health Organization is studying the role of scientific whistle-blowers in genome editing but has not proposed any specific mechanism. An earlier version of this article incorrectly said the WHO was developing plans for a hotline.