Late last year, scientist He Jiankui announced the birth of the world’s first genetically modified babies. He deployed a technology known as CRISPR to alter the genomes of two embryos that he used to initiate a pregnancy; their mother carried the pregnancy until the two girls, publicly called Lulu and Nana, were delivered prematurely by C-section.
The response of scientists, bioethicists, and the public was immediate and overwhelmingly negative. He’s reckless actions violated Chinese regulations, crossed a bright line that dozens of countries have written into their laws, flouted nearly every rule of medical ethics, and exposed two girls to grave and unknowable risks.
This week, a study emerged highlighting the dangers of what He did, especially based on our current level of understanding. He attempted to disable a gene called CCR5 by producing a variant that has been shown to confer resistance to HIV. But the new study suggests that having non-working copies of CCR5 actually has an overall negative effect on health, specifically a 21 percent increase in mortality before age 76.
These findings drew numerous comments on Twitter and in news stories. A few writers have criticized the study and cautioned against over-interpretation of its results. But whether or not the study’s conclusions ultimately hold up, they reveal how much we still don’t know about genomics and how far we are from being able to call heritable genome editing “safe.”
We don’t actually need more evidence to understand how irresponsible and dangerous He’s actions were. We also already have good reason to conclude that heritable genome modification is unacceptably risky and that medical justifications for it are tenuous. It would not treat any existing patient, and there are already safe and proven techniques, including embryo screening, to prevent passing on inheritable diseases. Editing eggs, sperm, or embryos produces changes that are passed down to all future generations. (This is distinct from somatic gene therapies, which are already being used to produce treatments for existing patients).
The societal risks of heritable genome editing, while often downplayed, are grave. Because altering the traits of future generations could lead to wealthy parents purchasing genetic “upgrades” for their offspring, it would send us in the direction of an even more unequal world built on divisions between genetic “haves” and “have nots.” No matter how sophisticated genome editing techniques become or how many safety criteria they meet, these fundamental social risks cannot be mitigated in a lab.
There are also immediate and considerable risks to children who develop from edited embryos. What would qualify as safe or “safe enough”? So far, discussions of safety have focused on whether CRISPR makes edits in the wrong place or edits incorrectly in the right place. There are also concerns that the changes may be made inconsistently, with edited sequences found in some cells but not others. As the recent study demonstrates, even if desired edits are perfectly made, do we know what all the effects would be? In different populations? How they would interact with other genes? With each individual’s unique genome? Throughout a lifetime and in future generations?
Determining safety would take generations and, even then, we could never really be sure.
He’s experiment introduced additional wrinkles. The link between CCR5 and HIV is fairly well studied. Disabling CCR5 removes the doorway HIV uses to enter and infect cells, but it does so only for some strains of HIV; there are others that don’t need CCR5. Further, the genetic sequence He’s edits produced does not match this well-studied variant of CCR5; in fact, it has never been observed in humans or animals. In other words, no one has any idea whether the variant with which Lulu and Nana are now living will affect HIV immunity or anything else.
That’s a key issue: Genes don’t do just one thing. Most illnesses and traits are influenced by dozens, hundreds, even thousands of DNA variations. Each of our roughly 20,000 genes is linked to many different aspects of our physiology and health. So what else does CCR5 do? A variant that provides protection against HIV also seems to increase susceptibility to a number of more common diseases, like flu and West Nile virus.
CCR5 has also been linked to brain function, which led to some sensational headlines and media speculation that the gene-edited babies might have enhanced brains. There are likely myriad other processes to which CCR5 contributes that we don’t know about yet. To that point: before the recent study, no one had researched whether the CCR5 mutation resulted in better or worse health over a person’s lifetime.
The CCR5 story illustrates a flaw in the logic that underlies gene editing. Efforts to change one gene to affect one illness in a future person ignore the fact that health is the result of infinitely complex interactions within and outside a person’s body. In most cases, the presence or absence of a particular genetic variant is not the sole determinant of a disease or condition.
Some proponents of heritable genome editing argue that “safe” application is just around the corner. But that assertion seems to depend on a much narrower definition of safety than should be required for procedures that would affect generations of people. And even if we could understand all the physiological and health effects that heritable genome modification would produce, we would still need to consider its unacceptably dangerous social consequences.
In other words, does the study give us a reason to oppose heritable genome modification? Sure. Throw it on the pile.
Katie Hasson, Ph.D. is the program director on Genetic Justice at the Center for Genetics and Society, a public interest organization based in Berkeley, Calif.