When He Jiankui announced the birth of twin girls whose DNA he had modified when they were embryos using the CRISPR gene-editing tool, he justified his actions on the ground that he had given the two girls lifetime immunity from HIV infection. The Chinese scientist claimed that he had altered a gene called CCR5, which allows the AIDS-causing virus to infect an important class of cells in the human immune system.
Not only was He ethically wrong in doing this work, but its scientific basis was even weaker than generally recognized.
He’s experiment has been widely, roundly, scorchingly, and appropriately condemned. The risks to the babies born nine months later were vastly higher than the benefits to them or to science; the informed consent process for their parents was deeply flawed; China’s legally required regulatory process had apparently been corrupted; the experiments had been wrongly kept secret; and a worldwide scientific consensus against human germline genome editing, built in publications and statements over more than three and a half years, had been ignored.
He’s stated goal was to give the children born from his experiment genes for a protein called CCR5 with 32 base pairs deleted. This mutation, called CCR5∆32, seems to protect some northern Europeans from infection with HIV. Yet soon after the announcement of his “success” in November 2018, some experts noted that the likely benefits were smaller than advertised, since He had given neither of the twins that specific gene edit. According to He’s own analysis, his CRISPR treatment yielded mutations in CCR5 that had never before been seen in humans — meaning the effectiveness of the edit against HIV infection and its safety were unknown.
Adding to his sins, He knew after the gene editing — but before the embryos were transferred to their mother’s uterus — that one of the twins, whom he called Lulu, had a normal CCR5 allele on one copy of her chromosome 3 and an edited version on the other copy, greatly undercutting any protection she might have received. The other twin, called Nana, had two edited copies, but neither of them was CCR5∆32. We can only hope that the girls, whose real names and identities have been shielded, remain healthy.
But it gets worse. Many scientific commentators, as well as He himself, quickly jumped at the idea that people without functioning CCR5 proteins cannot become HIV infected. That’s not correct. Here’s why:
CD4 cells are a type of T cell, white blood cells that fight infection and play other important roles in the immune system. They get their name from the CD4 protein they carry on the outside and the inside of their cell membranes. CD4 proteins serve as receptors for external molecules, like the proteins on a virus. CD4 cells appear to be the most crucial type of T cell for HIV infection and its progression to AIDS.
Most HIV infections begin when the virus latches onto the CD4 protein on the surface of CD4-positive T cells. But for infection to occur, the virus also needs to attach itself to a second cell surface protein. This second protein is often CCR5, but it can also be another protein called CXCR4. When HIV is connected to both CD4 and one of the other receptors, it fuses into the cell.
The most common strain of HIV in humans, dubbed R5 HIV-1 (for the CCR5 receptor), appears to require CCR5 in order to infect CD4 cells. But another strain, called X-4 HIV-1, uses CXCR4 instead of CCR5. Some strains of X-4 HIV-1 can use either CCR5 or CXCR4. And the less common and less deadly strain known as HIV-2 uses many receptors in addition to CD4.
The bottom line is that the proteins made by the CCR5∆32 gene variant that He tried to create in Lulu and Nana do not necessarily prevent infection with HIV. This variant reduces infection of CD4 cells with R5 strains of HIV, and it may even prevent infection. But CCR5∆32 doesn’t interfere with the ability of X-4 and some HIV-2 strains to infect CD4 cells. Even the R5 strain can infect important immune cells known as dendritic cells that can serve as a reservoir for HIV and, while there, possibly mutate from the R5 version of HIV into a CXCR4 version, which could then go back and infect CD4-positive T cells.
There is good evidence that the absence of functional CCR5 proteins is not completely protective against HIV. In 2012, the world was excited by “the Berlin patient,” who was later identified as Timothy Ray Brown. After living with HIV for 10 years, Brown developed acute myeloid leukemia, for which he received two bone marrow transplants from the same donor — who happened to carry the CCR5∆32 mutation that He tried to produce in Nana and Lulu. Brown was eventually able to stop taking anti-retroviral drugs and is viewed as the only person to have been “cured” of HIV.
In early 2019, a second man with HIV received a bone marrow transplant for a different type of blood cancer from a donor with no functioning copies of CCR5. He, too, is off anti-retroviral drugs, and is apparently cured of AIDS. Like the Berlin patient, this “London patient” could still be infected with X-4 HIV-1. In at least one case, the “Essen patient,” a man with HIV who received a bone marrow transplant from a CCR5∆32 donor later died from infection with the X4 strain of HIV.
As the New York Times wrote in discussing the London patient:
“One important caveat to any such approach is that the patient would still be vulnerable to a form of H.I.V. called X4, which employs a different protein, CXCR4, to enter cells. ‘This is only going to work if someone has a virus that really only uses CCR5 for entry — and that’s actually probably about 50 percent of the people who are living with H.I.V., if not less,’ said Dr. Timothy J. Henrich, an AIDS specialist at the University of California, San Francisco.”
My message here is that CCR5∆32 does not guarantee immunity to HIV infection. It works only for strains of HIV-1 that use both the CD4 and CCR5 cell surface receptors to infect T cells. Even if the version of CCR5 that He gave Nana does indeed protect her, it would only be from infection with the R-5 strains of HIV — not from X-4 strains and also not from other variants that the virus will inevitably evolve.
The benefit/risk ratio for these girls, already terrible, now looks even worse. A few reporters seem to have noticed, or at least commented on, the importance of CXCR4, including early news stories from The Atlantic, Nature, and Bloomberg. But those stories largely disappeared.
Although I had been carefully following the story of Lulu and Nana since He announced their birth in late November 2018, it was only after noticing some of the caveats in the stories about the London patient in early March that I become aware of this issue. And that has made me wonder why we, including me, bought into the CCR5 story?
For one thing, it’s a good story. For another, it was He’s story, as well as one that had previously been described by other scientists. And we all knew (or we thought knew) about the Berlin patient.
But I think another factor contributed. It was a story of starkly conflicting interests — immunity from dreaded HIV set against the risks of a new technology. It made for a good exam question, thought experiment, or commentary. But those two little girls are not thought experiments. They are real people, the results of a real experiment marred by many deep flaws, one of which was to ignore entirely the existence of HIV strains that don’t require CCR5.
Would a serious advance review of He’s experiment have caught this issue? I can only hope so. It certainly should have if the reviewing group included expertise in HIV infection and not just in CRISPR editing. This is a crucial lesson in the importance of a probing prior review before any experiment involving human subjects and especially before any first-in-human experiment. It is also a sobering lesson, for me and others, about retrospective review of such experiments.
Nature is complicated, viruses are tricky, assumptions are dangerous — and He Jiankui’s experiment was, amazingly, even worse than I first thought.
Henry T. Greely, J.D., is professor of law and professor by courtesy of genetics at Stanford University, where he directs the Stanford Center for Law and the Biosciences and chairs the steering committee for the Stanford Center for Biomedical Ethics. He thanks his research assistant, James Rathmell, for his help with the article.