Late last year, a Chinese researcher shocked the scientific community when he announced that the first gene-edited humans had already been born. He Jiankui barreled past an emerging consensus that the technology wasn’t ready for use and, once it was, should be reserved for otherwise untreatable diseases. Instead of respecting those boundaries, He did much of his work without any clear institutional oversight.
Rather than target an incurable genetic disorder, He Jiankui focused on something for which we have both preventative measures and treatments: HIV infection. He did so by using CRISPR gene editing to damage a gene that encodes a protein that HIV uses to enter human cells; previous studies have shown that mutations in this gene protect against HIV infection. But the same mutation was already known to make humans more susceptible to other diseases, raising the question of whether the gene editing put its recipients at risk.
That question has now been answered with an emphatic “yes.” Researchers have found that adults carrying mutations in the gene see their general mortality rise by 20 percent compared to their peers.
The gene in question is called CCR5, and it encodes a receptor for immune signaling proteins. To perform that function, it has to sit on the surface of immune cells, where HIV can latch on to it and use it to take a ride into the cell’s interior, where it can continue an infection. Research found that some people who had HIV infections that progressed slowly often had mutations in CCR5, often a specific mutation called ∆32, which causes severe damage to the protein.
CCR5-∆32 mutations occur at a low level in the general population, and it is protective both when present in one or both of the copies of the gene. Using the knowledge of its role in HIV infections, researchers have used CCR5-∆32 donors to provide new blood stem cells to two HIV-positive leukemia patients-a transplant that successfully eliminated the virus from these individuals.
He Jiankui’s gene editing was intended to provide HIV protection to the children of an HIV-positive parent. This, to a certain extent, was superfluous; we already know how to prevent parent-to-child HIV transmission and, even in the case that our knowledge failed, HIV infection is treatable. However, earlier research had suggested that people who carry two copies of the CCR5-∆32 mutation have elevated risks of being infected by other viruses. This raised the prospect that the gene editing was not only unnecessary, but harmful.
Two researchers at the University of California, Berkeley (Xinzhu Wei and Rasmus Nielsen) decided to see whether that was the case. To do so, they relied on the UK Biobank, an ambitious effort to gather genetic information and track the health of a half-million UK residents. Not all the data is in place yet, but Wei and Nielsen were able to look up the CCR5 genotype of more than 400,000 individuals. While the CCR5-∆32 mutation is rare in the UK population, this ensured that the study had thousands of people to gather data from.
An early death
It’s not clear for which diseases CCR5-∆32 increases are susceptible, so the researchers simply looked at all-cause mortality. This would also include non-infectious issues, such as if the mutation altered the frequency of autoimmune issues.
The results were dramatic: “∆32/∆32 has an approximately 21 percent higher aggregated death rate before age 76 than the other genotypes.” Put differently, people with two copies of the mutation were 20 percent less likely to reach the cutoff age of 76 years old. The risks appeared to increase with age, too, as the highest difference for those with two copies of the mutation occurred at age 74, when the mortality rate was over 25 percent higher among those with the CCR5-∆32 mutation.
People with one normal and one CCR5-∆32 mutation version of the gene also have significant protection from HIV infections. And, in this study, their mortality rate was indistinguishable from two undamaged copies of the gene.
The authors also looked at the dynamics of the CCR5-∆32 mutation within the population, looking for signs of evolutionary changes in its frequency. This analysis produces a number roughly in line with the mortality data. Why isn’t HIV creating some pressure to increase the frequency of this mutation? Two reasons: drug treatments now allow people with HIV to survive, and the frequency of HIV infection in the UK is only about 0.16 percent, too low to have a strong influence on genetics.
To get a sense of what might be causing people with two copies of the CCR5-∆32 mutation to die early, the researchers looked for other health outcomes associated with the mutation. And they found a lot. Even after correcting for the fact that this involved multiple tests (which increase the risk of spurious associations), they found plenty of health issues associated with CCR5-∆32. This shouldn’t be a shock for something central to immune function.
The clear message of this analysis is that adopting CCR5-∆32 as a form of HIV protection isn’t something we should do lightly. While it might be a reasonable approach for someone who’s already dealing with both leukemia and HIV infection, it’s obvious that the premature gene editing work has left its recipients burdened with a lifetime health risk.
Nature Medicine, 2019. DOI: 10.1038/s41591-019-0459-6 ( About DOIs).