TO JENNIFER DOUDNA, mitochondrial replacement therapy looked like a faint foreshadowing of what CRISPR might deliver. Doctors like Zhang were only replacing faulty genes in embyros with healthy ones. CRISPR might allow them to rewrite any of the twenty thousand or so protein-coding genes sitting on an embryo’s chromosomes. And that change could be inherited by their descendants.
The last thing Doudna wanted was for CRISPR to replay the botched history of mitochondrial replacement therapy. That treatment had crept into practice without any public discussion of its ethics, and when the conversation finally got off to a late start, it was distorted by lurid visions of Frankenstein and charged language about three-parent babies. In 2014, Doudna decided that she would try to avoid such a debacle by starting a public conversation.
It was not a role Doudna relished. She felt comfortable talking about the inner workings of bacteria, not about the potential dangers of retooling human embryos. Her new role “felt foreign,” she later said. “Almost transgressive.”
Doudna started small. In January 2015, she hosted a meeting at a cozy inn in the Napa Valley, about an hour from Berkeley. Among the eighteen people who assembled in wine country to talk about CRISPR were David Baltimore and Paul Berg, two eminent biologists who had led similar meetings in the 1970s to discuss recombinant DNA. Now as then, Weismann’s barrier split the conversation in two. The Napa group divided their time talking about tinkering with somatic cells and germ cells.
CRISPR might prove superior to viruses as gene therapy, the researchers speculated, because doctors could use it to fix somatic cells with more precision. As for the germ line, some people at the Napa meeting weren’t bothered by the prospect of using CRISPR to alter it. Others considered Weismann’s barrier a line never to be crossed.
Despite their differences, everyone at the meeting knew that they couldn’t just let their differences lay idle. There wasn’t time. There were rumors that scientists in China had already used CRISPR on a human embryo. A paper from the scientists was supposedly circulating among journals for publication. Any day now, the news might break. Most of the participants at the Napa meeting agreed to coauthor a commentary, which they’d submit to a journal. On March 19, Doudna and seventeen fellow scientists published a piece in Science called “A Prudent Path Forward for Genomic Engineering and Germline Gene Modification.”
The scientists didn’t call for an outright ban on human germ line engineering, but they did strongly discourage it for now. They also proposed that a public meeting take place, bringing experts together from around the world to drill deeper into the risks and benefits of the new technology. Even such a big gathering as that would not be enough to settle matters, Doudna and her coauthors warned. “At present, the potential safety and efficacy issues arising from the use of this technology must be thoroughly investigated and understood before any attempts at human engineering are sanctioned, if ever, for clinical testing,” they declared.
The Science piece drew so much attention that the National Academy of Sciences agreed to host an international meeting just a few months later, and the Royal Society and the Chinese Academy of Sciences signed on to participate. Things were unfolding as Doudna had hoped, at least for a few weeks.
In April, she discovered that the rumors she’d been hearing were true. Junjiu Huang, a biologist at Sun Yat-sen University, and his colleagues reported that they had crossed the line. They had used CRISPR to alter human embryos.
Depending on how you looked at it, Huang’s experiment was a historic achievement or a botched nonstory. As the Chinese scientists explained in the journal Protein & Cell, they set out to alter the HBB gene, the gene in which mutations can cause beta-thalassemia. They designed the experiment to sidestep ethical concerns about tinkering with viable embryos. When fertility clinics fertilize eggs, they sometimes make mistakes, allowing two sperm to fuse to a single egg. These embryos end up with three sets of chromosomes—hence their name, triploid—and they fail to develop for more than a few cell divisions. Huang and his colleagues got hold of dozens of triploid embryos to study, confident in the knowledge that these embryos could never be used to start a pregnancy even if someone wanted to.
Huang and his colleagues built CRISPR molecules that could cut part of HBB genes, allowing them to be replaced with a new stretch of DNA. They injected the mix into the triploid embryos and waited for them to divide into eight cells. Huang and his colleagues analyzed fifty-four of the embryos to see how well the CRISPR molecules had worked. They had only managed to cut the HBB genes in twenty-eight. And of those embryos, a smaller fraction had replaced the old DNA with the new material. In other embryos, the cells had accidentally copied similar genes elsewhere in their DNA.
A number of the embryos, Huang and his colleagues found, ended up as mosaics. Some cells in the mosaic embryos had an altered version of the HBB gene, and some didn’t. It turned out that the CRISPR molecules needed a lot of time to find their targets in the human DNA. By the time they found the HBB gene, the single-cell zygote had divided into several new cells, some of which ended up without any of the CRISPR molecules inside.
When news of Huang’s paper broke, I asked Doudna what she thought of it. She told me that it “simply underscores the point that the technology is not ready for clinical application in the human germline.”
Doudna was choosing her words carefully. From the Napa meeting onward, she wanted to avoid turning public opinion against CRISPR in general. Her fellow scientists would have to restrain their curiosity and not carry out experiments that might seem grotesque or reckless. On the other hand, Doudna didn’t want to rule out germ line engineering altogether. Perhaps someday in the future, it would be worth considering.
Her careful parsing couldn’t stop the story from spiraling out of control. It caused such a worldwide commotion that Francis Collins, the director of the National Institutes of Health, released a blunt statement a few days later invoking the policies put in place in the 1980s. “The concept of altering the human germline in embryos for clinical purposes has been debated over many years from many different perspectives, and has been viewed almost universally as a line that should not be crossed,” he declared. The NIH was pouring over $250 million a year into gene therapy research in the hopes of curing diseases by changing genes in somatic cells. But they would not pay anyone to leap to the germ line, full stop.
When scientists like Collins talked about a line that should not be crossed, they made it sound blindingly clear. Yet scientific research was only making it harder to find. Collins spoke of altering embryos “for clinical purposes,” for example. One could argue that an experiment like Huang’s had no clinical purpose whatsoever, because the triploid embryos he used could never develop into children. Was his experiment beyond the line anyway, because someone might use the knowledge he gained to alter an embryo’s HBB gene and eliminate beta-thalassemia from a line of descendants?
In September 2015, the line got even harder to discern. A scientist named Kathy Niakan at the Francis Crick Institute in London applied to the British government for permission to use CRISPR on human embryos. Niakan planned to use CRISPR to shut down genes believed to be crucial for the early development of embryos. By seeing how the embryos then developed could give Niakan clues about the jobs the genes carry out. She would study the embryos only up till they were about a week old and then destroy them. But, unlike Huang’s experiment, Niakan would be carrying out CRISPR on viable embryos. Did their viability make her research an affront to all that is decent, even if she destroyed them when they were still a microscopic clump of altered cells?
The news about Niakan and Huang brought an intense urgency to the International Summit on Human Gene Editing when it opened on December 1, 2015. At the National Academy of Sciences in Washington, David Baltimore welcomed the five hundred attendees with a provocative kickoff, echoing Hotchkiss’s words fifty-one years beforehand.
“Today, we sense that we are close to being able to alter human heredity. Now we must face the questions that arise,” Baltimore said. “The overriding question is when, if ever, will we want to use gene editing to change human inheritance?”
The conference never managed to live up to Baltimore’s provocation. “It was an extremely low-key meeting,” the reporter Sharon Begley observed, “with more and more empty seats as it went on.”
For some scientists, the low-key tone seemed like an intentional strategy. They hewed to jargon-rich prepared remarks, not wanting to stumble over any ethical trip wires. At the end of the meeting, Baltimore, Doudna, and the rest of the organizing committee came on stage to deliver a consensus statement. It didn’t go much beyond what the Napa group had agreed to eleven months earlier. The committee endorsed CRISPR for somatic cells—gene therapy, in other words—as well as CRISPR-based research on early human embryos. They came out against germ line modification in embryos used to establish a pregnancy. But they didn’t close the door. “The clinical use of germline editing should be revisited on a regular basis,” they declared.
For some of the scientists at the meeting, the real question about germ line editing was whether it was even worth trying. “If we really care about helping parents avoid cases of genetic disease,” said Eric Lander, the director of the Broad Institute in Cambridge, Massachusetts, “germline editing is not the first, second, third, or fourth thing that we should be thinking about.”
Lander argued that parents would be better helped in almost every case by preimplantation genetic diagnosis. If they risked giving their child a hereditary disease, they could get the help of doctors to sort through embryos—or perhaps even eggs and sperm—to make sure their children were not born with the disease. Only after parents tried these measures without success would it make sense to edit the germ line.
A few of the speakers came out strongly against even this restrained approach to CRISPR. Marcy Darnovsky, the executive director of the Center for Genetics and Society, painted a dark future very different from Lander’s. She envisioned an unregulated marketplace where germ line editing would run rampant. “It’s a radical rupture with past human practices,” she warned.
The risks—of CRISPR missing its target and rewriting a different gene, for example—were simply too great. And not only was it dangerous to change a child’s DNA, Darnovsky argued, but it was simply wrong. An embryo could not give consent for the procedure. And altering the child’s genes was, when you came down to it, an affront to the child’s individuality. Even if CRISPR worked splendidly, its very success might create unprecedented social woes. Darnovsky could picture a world in which the rich engineered their children’s genes to escape the burden of disease, while poorer children could not. And once parents started trying to get rid of certain traits in their children, where would they stop?
Deafness is not lethal, for example, but when Harry Laughlin of the Eugenics Record Office published his Model Eugenical Sterilization Law in 1922 he put it on the list. Would fertility doctors offer to knock out deafness mutations? Would other conditions, such as dwarfism, be judged intolerable burdens, too? Disabled communities might feel besieged before long. And once parents got used to altering more and more aspects of their children’s biology, Darnovsky predicted, they might well start changing genes to enhance their children.
“The temptation to enhance future generations is profoundly dangerous,” Darnovsky warned. “I ask you to think about how business competition might kick in with fertility clinics offering the latest upgrades for your offspring. Think about how the market works.”
Most of the scientists at the meeting shied away from even mentioning enhancement. One exception was a towering, long-bearded geneticist from Harvard named George Church. Enhancement was coming, Church said, and it would begin not with embryos but with old people.
Here’s one way that might happen. Nine percent of older people suffer from brittle bones due to osteoporosis. Cells in their skeleton start to break down the surrounding bone, releasing the minerals into the bloodstream. Some drugs can slow down the decline by sticking to cells, making it harder for them to make contact with bone. But there may be a better way to treat the disease.
One reason that cells break down bone is that in old age they make less of a protein called TERT. In 2012, researchers at the Spanish National Cancer Research Centre coaxed old mice to make extra TERT by giving them an extra copy of the gene. Their osteoporosis reversed, and their bones strengthened. It’s conceivable that doctors could use CRISPR gene therapy to treat people as well. CRISPR molecules would home in on the TERT gene in bone cells and edit it. The gene would behave as it did when the patients were younger, strengthening their bones.
But gene therapy for TERT could do a lot of other kinds of good, too. The Spanish researchers who tried it out on mice found that it also reversed aging in their muscles, their brains, and their blood. It extended the life span of old mice by 13 percent. When the scientists treated younger mice with TERT gene therapy, the animals lived 24 percent longer.
If CRISPR-based gene therapy got approved for osteoporosis in humans, people might soon clamor to get the treatment for a longer, healthier life. And if it proved to be more effective the earlier it was administered, then some parents might want their children to have that benefit from the very start. Why end life with good TERT genes when you can start with them?
TERT was just one candidate among many that Church saw for enhancement. Editing genes to treat wasting muscles might lead to people enhancing their strength. Researchers are discovering ways to fight the decline in memory and learning that comes with Alzheimer’s. Gene therapy could conceivably deliver the same results. And editing those genes in healthy people might enhance their cognition. Parents might choose to edit those same genes in their children to give them an edge in school and at work.
“I think enhancement will creep in the door,” Church told the audience.
In one sense, Doudna’s campaign was a success. She had succeeded in getting a conversation started. By 2016, CRISPR had cracked open its cocoon and was a full-blown media butterfly, the subject of regular coverage on television and in the newspapers. Doudna hopped from city to city to explain the awesome tools she had helped invent and to provoke discussions about its use.
That conversation’s center of gravity soon shifted in a profound way. The National Academy of Sciences brought together a committee of twenty-two experts—including biologists, bioethicists, and social scientists—to grapple with the science, ethics, and governance of human genome editing. In February 2017, they released a 260-page report on their deliberations. Instead of kicking the germ line editing can even farther down the road, they came to a startling agreement. They endorsed clinical trials for treating serious diseases for which there were no other alternative treatments.
In their report, the committee asked readers to consider the case of Huntington’s disease. Karen Mulchinock was able to prevent her children from inheriting the disease by using preimplantation genetic diagnosis. Because she carried only one defective copy of the HTT gene, she could pick out embryos that inherited her good copy. If two people with Huntington’s have children, however, their children run a 25 percent risk of inheriting the disease-causing copy of the gene from both parents. In these rare cases, preimplantation genetic diagnosis offers no help. Genome editing might. “The number of people in situations like those outlined above might be small, but the concerns of people facing these difficult choices are real,” the committee observed.
Shortly after the genome editing committee released their report, Shoukhrat Mitalipov published an experiment that suggested that these clinical trials could very well succeed. Mitalipov followed Junjiu Huang’s example and tried to erase a genetic disorder by editing human embryos. Unlike Huang, however, Mitalipov and his colleagues manipulated viable embryos rather than doomed ones. They did everything they could to keep the embryos viable. And while Huang’s team used relatively crude CRISPR tools, Mitalipov took advantage of newer, more precise ones.
For their experiment, Mitalipov chose a genetic disease of the heart known as hypertrophic cardiomyopathy. Mutations to a gene called MYBPC3 cause the heart’s walls to thicken and falter. Without warning, people with the disease may die of a sudden heart attack. The disorder is dominant, meaning that a child need inherit only one copy of the faulty gene to develop a faulty heart.
Mitalipov’s team got sperm from a man suffering from hypertrophic cardiomyopathy and used it to fertilize healthy eggs. They also delivered CRISPR molecules tailored to seek out the mutation on the MYBPC3 gene. Out of fifty-four embryos treated this way, thirty-six ended up with two healthy copies of the gene. Another thirteen ended up as mosaics.
By the time Mitalipov’s CRISPR molecules made their way inside the fertilized eggs, they had already made a new copy of their DNA. The CRISPR molecules apparently edited only one copy, so that when the embryo divided into two cells, one had the defective MYBPC3 gene and the other didn’t. As the embryo continued to grow, new cells inherited one version of the gene or the other.
To avoid making mosaics, Mitalipov’s team tried a new method. They edited the mutation-carrying sperm, and then used them to fertilize eggs. In these trials, 72 percent of the embryos lost the mutation. Mitalipov and his colleagues found that their edited embryos developed normally for eight days. If they had implanted those embryos instead in a woman’s uterus, they might well have developed into babies with healthy hearts.
I was on vacation with my family in a little English village in August 2017 when the news broke of Mitalipov’s experiment. I had hoped to take a break from CRISPR, filling my days with footpath walks and visits to castles. But one day I walked into the local grocery store and spotted a newspaper with four-cell human embryos scattered across its front page.
“The Cells That Could End Genetic Disease,” it blared. That evening, I turned on the television at our cottage, only to encounter Mitalipov talking about his experiment. CRISPR was becoming inescapable.
For all the attention the world gave Mitalipov’s research, he didn’t promise much from it. Parents who carry variants for hypertrophic cardiomyopathy can already use preimplantation genetic diagnosis to identify embryos that won’t develop the disease. CRISPR might help them deal with simple Mendelian inheritance, which leaves them with only 50 percent of their embryos to implant, lowering the odds of a successful birth.
“Gene correction would rescue mutant embryos, increase the number of embryos available for transfer and ultimately improve pregnancy rates,” Mitalipov and his colleagues wrote in their paper.
By narrowing his focus, Mitalipov made the ethical challenges of CRISPR seem manageable. He just wanted to improve the odds that parents could have healthy families. Likewise, other scientists who contemplated germ line engineering with CRISPR simply wanted to cure hereditary diseases in the womb, rather than wait to treat somatic cells later in life.
Within this narrow focus, a lot of the ethical concerns that have been raised about CRISPR seem less like dystopian nightmares than the everyday challenges that conventional medicine already poses. If CRISPR turns out to work reliably, we might well face a world where hereditary diseases are a bigger burden on those who can’t afford it. But the cost of medicine has been a grave problem for generations, and many recent advances have made this inequality more dire. As gene therapy has inched its way to the clinic, companies have begun floating astonishing price tags for the treatment. A single shot of gene-carrying viruses might cost a million dollars or more. Yet no one has responded to this figure by demanding that gene therapy be banned. There’s nothing wrong with gene therapy in itself, only with the ability of some people to get it while many others can’t. That’s a problem of politics, of economics, of regulations. If we are worried some people can’t get CRISPR, then the solution is obvious: CRISPR for everyone.
The question of consent isn’t new to germ line engineering either. We don’t require that children give their consent in order to get vaccines or antibiotics. That’s what parents are for. If conventional medicine fails to help sick children, their parents may give their consent to experimental treatments, knowing full well their children may not be cured and may even suffer as a result. Early in the history of gene therapy research, parents started enrolling their children in studies. It was a profound decision for the parents to make, weighing the grave diseases their children suffered against the possibility that some side effect would emerge. It was especially serious for gene therapy studies, since the children would be carrying genes in their cells possibly for the rest of their lives. Yet no one has responded to these difficult ethical choices by calling for all gene therapy to be banned.
As for the ethics of enhancements, we already live in a world in which parents try to enhance their own children’s prospects. And many of those enhancements are already spread out unfairly. In 2010, American parents whose income was in the top 10 percent spent more than $7,000 on young children, including books, computers, and musical instruments. Parents in the bottom 10 percent spent less than $1,000.
In some cases, societies have managed to spread enhancements beyond the wealthy few. Vaccinations enhance our immune systems by priming them to fight measles and other diseases. The world has committed itself to getting as many children vaccinated as possible. That’s a noble achievement. But, as a society, we still fall shamefully short in other respects.
While some enhancements should be spread more fairly, other enhancements are simply misguided. Some parents insist on having their short children treated with human growth hormone—not because they’re sick, but because their parents want to give them social advantages that come with being tall. This push for enhancement can lead instead to insecurity, as children feel inadequate in their parents’ eyes, despite being perfectly healthy.
If we treated embryo editing in this way—as very early gene therapy—we would probably muddle our way to a new normal. We’d allow it much as we allow in vitro fertilization. We’d debate whether this or that use of CRISPR is acceptable. We’d ban some, approve others. Altering some genes might turn out to have dangerous side effects, and regulations would need to be put in place to keep children safe. Somebody would certainly sneak off and try a reckless treatment, and we’d try to make sure no one ever did so again. And, in time, CRISPR would become a responsible form of medicine.
But editing embryos is not merely another form of medicine. As David Baltimore made plain at the outset of the international meeting in 2015, what made it so unsettling was the possibility that we could use CRISPR to alter the future of human heredity. We were climbing into the chariot of the sun without knowing if we were wise enough to control its course.
It’s heredity that matters most, and yet we have given surprisingly little reflection about why it matters so much, or how our actions would actually alter it.
One of the few people to think through the ethics of altering heredity was a theologian named Emmanuel Agius. In 1990, years before CRISPR even had a name, Agius argued that germ line editing would rob future generations of their inheritance.
“The collective human gene pool knows no national or temporal boundary, but is the biological heritage of the entire human species,” Agius said. “No generation has therefore an exclusive right of using germ line therapy to alter the genetic constitution of the human species.”
But what would it really mean to alter the genetic constitution of a new species? People have sketched out many scenarios: a utopia without disease; a dystopia where the rich enjoy genetically enhanced intelligence and health while the poor endure nature’s miseries. Some have even claimed that we will turn Homo sapiens into a new species altogether.
These are dreams. Sometimes dreams prove prophetic. Sometimes they prove to be fantasies. Hermann Muller’s dream of Germinal Choice got some parts of the future right, and some parts wrong. He assumed a socialist government would protect the future of the human gene pool. He may well have been shocked if he had lived long enough to see sperm banks, in vitro fertilization, and preimplantation genetic diagnosis take hold thanks to capitalism instead. Parents did not volunteer for duty as he envisioned. They became consumers.
Whichever future we end up in, the path will have to start from where we are today. Preimplantation genetic diagnosis might be the beginning of a major shift in how children are born. In 2014—thirty-six years after the first test-tube baby was born—only 65,175 babies in the United States were born from in vitro fertilization. That’s about 1.6 percent of the American babies born that year. Of those, only a small fraction went through preimplantation genetic diagnosis (it’s hard to get precise numbers). Worldwide, there might be tens of thousands of children who went through it. Together, they make up only a hair-thin fraction of the 130 million babies born each year worldwide. But with each year, more parents are choosing the procedure—in some cases, encouraged by their governments. In a 2010 study, researchers investigated how much money would be saved if a couple used preimplantation genetic diagnosis to avoid having a child with cystic fibrosis. A $57,500 procedure could avoid $2.3 million in medical costs over a lifetime.
Today, parents typically use preimplantation genetic diagnosis if they already know their children are at risk of particular diseases. As DNA sequencing speeds up, it will become possible for doctors to scan every gene in an embryo, detecting hereditary diseases that parents may not know they carry. It would be a seductive offer. I can imagine how it would work by looking at my own genome. It turns out, for example, that one of my copies of a gene called PIGU has a mutation that puts people at greater risk of skin cancer. If I had my druthers, I’d prefer my children to inherit my good copy and not my bad one. Either way, they’d still be inheriting my DNA.
And it would be hard to stop there. I have also discovered I have a variant of a gene called IL23R that dramatically lowers my risk of certain disorders. Those disorders include Crohn’s disease, a chronic inflammation of the gut, and ankylosing spondylitis, which fuses the vertebrae in the spine, causing chronic pain and forcing people to hunch forward. What they have in common is the immune system going haywire and attacking the body’s own tissue. No one knows exactly what triggers this attack, but it appears my variant of IL23R—found in only 8 percent of people with European ancestry—tamps down the immune system’s communication network. My variant is so potent, it turns out, that drugmakers used its biology as the basis for drugs for autoimmune disorders. As a parent, I would do anything I could to lower the risk that my children get diseases that put their back into howling pain or that give them a lifetime of intestinal distress. To give my children my protective copy of IL23R, rather than the ordinary one, would be the least I could do.
If governments allowed it, some parents might ask if they could pick out their variants that could influence other traits in their children. Preimplantation genetic diagnosis on its own would usually fail to produce big results. But there might be exceptions. Scientists have found a mutation on a gene called STC2 that alters a hormone our bodies make, called stanniocalcin. One copy of the mutant version can make a person three-quarters of an inch taller. But only one in a thousand people carry it.
It’s hard to predict how far parents would go as they picked out natural variations. A physicist named Stephen Hsu at Michigan State University has claimed that parents could raise their children’s intelligence by selecting from embryos. Their doctors could check for which versions of a hundred genes influencing intelligence each embryo had. The embryo with the highest score could be implanted. Hsu estimated that this selection might, on average, raise a child’s IQ score by five to ten points.
Geneticists generally scoff at Hsu’s claims. We still know precious little about the genes that influence intelligence. While scientists have zeroed in on some of the genes that likely play a role, it’s entirely possible that the true players are nearby genes or gene switches. And since we know so little about how genes for intelligence interact with the environment, picking out certain alleles to give to embryos could wind up having no effect at all.
That skepticism didn’t stop Hsu. In 2011, he joined researchers at a Chinese DNA-sequencing center called BGI to found their Cognitive Genomics Lab. They set out to get DNA from two thousand of the world’s smartest people and find variants they shared. In 2013, reporters got wind of the project and described it in breathless tones. “Why Are Some People So Smart? The Answer Could Spawn a Generation of Superbabies” was the headline of a Wired article. “China Is Engineering Genius Babies,” Vice announced.
Vice claimed that the BGI team was close to finding intelligence alleles and that China had “developed a state-endorsed genetic-engineering project.” Wired’s John Bohannon suggested that a generation of superbabies might be spawned if a government like Singapore’s encouraged parents to use preimplantation genetic diagnosis to pick embryos with high genetic scores for intelligence. Hsu himself found the coverage outrageous. In an interview with the journalist Ed Yong, he simply said, “That’s nuts.”
But Hsu had Muller-size dreams of his own. Imagine that preimplantation genetic diagnosis for intelligence genes became widespread in a country. Now imagine that the children produced from that selection used the procedure on their own children. In a 2014 essay for Nautilus, Hsu argued that this would be no different from what happens when cattle breedings select animals for size or milk yield. With so many variants influencing intelligence, it would be possible to raise intelligence test scores for generations until today’s tests would no longer be able to measure it. “Ability of this kind would far exceed the maximum ability among the approximately 100 billion total individuals who have ever lived,” Hsu promised.
In 2017, I e-mailed Hsu to see how his dream was faring. Six years had passed since the Chinese intelligence project had launched. And in that time I had heard of no concrete results. When I contacted Hsu, he told me that BGI had sequenced about half of the two thousand people in the study. Then they got in a business dispute with the company supplying them with their DNA-sequencing equipment.
“As a peripheral consequence,” Hsu said, BGI “cut our project off a few years ago. So, we still to this day have not sequenced all of our samples.”
The first generation of superbabies would have to wait.
Preimplantation genetic diagnosis already allows parents to choose which of their own variants their children can inherit. It may open the door to CRISPR, just as Mitalipov has proposed. If that does happen, it will not alter heredity any more than we’re already altering it by blocking some disease-causing mutations from getting into future generations. At least not at first.
If CRISPR became a standard tool in fertility clinics, people might lose their suspicions of it—just as people lost their suspicions of in vitro fertilization in the 1980s. Before long, people might be willing to entertain a new use for CRISPR. Doctors might edit beneficial changes into an embryo’s genes. They might protect children from Crohn’s disease by rewriting the IL23R gene. There are other rare variants that show signs of protecting people from Alzheimer’s disease, various kinds of cancer, and infectious diseases like tuberculosis.
None of these variants are artificial, since they were discovered in people’s DNA. Parents could give their children all the advantages that scientists have found in our species’ genetic variations. But more variants keep coming to light with more research. If this practice became popular enough, the Australian philosopher Robert Sparrow has speculated, parents might hold off having children, in the same way people wait to buy a phone until a new model is released. Sparrow wonders if future generations might find themselves stuck in an “enhanced rat race.”
The choices that parents make about editing embryos would not just affect their children. The alterations could be inherited by their grandchildren. For parents with Huntington’s disease, it would probably be a great relief to know their descendants wouldn’t be tormented by a faulty HTT gene—unless, of course, they inherited it from another ancestor.
But when we try to look far forward, over the course of many generations, heredity doesn’t necessarily work the way we imagine it to. Introducing a single edited gene into the human gene pool does not guarantee that it will take over the human species as Agius promised. In fact, the science of population genetics has found that it’s far more likely for a new variant to eventually disappear. Bequeathing an Alzheimer’s-fighting variant to your children may seem like a wonderful gift, but it’s not a gift that can be reliably passed down through the generations. Imagine your daughter, equipped with two copies of the variant, marries a man who lacks them. Their children will inherit only one copy apiece. When they have children with people who don’t have a copy, many of your great-grandchildren will probably not have any protection left at all.
Natural selection won’t raise the allele’s odds of surviving, either. While we may all want to avoid Alzheimer’s disease, evolution doesn’t care about our desires. Alleles get spread over the generations if they help people reproduce more. An allele that lowers the odds of dementia at age seventy doesn’t help at all. Within a few generations, the variant you paid so dearly to give your children might disappear entirely.
When people wring their hands about what genetic engineering might do to the human gene pool, they often forget that it’s actually more like a human gene ocean. If I strain my science-fiction faculties to their limit, I can imagine a worldwide dictatorship that forces every parent on Earth to submit to CRISPR and introduce the same variants into every child. But just because I can imagine the movie doesn’t mean I think it’s likely.
Gene-pool arguments are flawed for another reason. They treat the collective DNA of our species as if it were inscribed in stone tablets long ago and passed down unchanged ever since. In fact, the human gene pool has always been changing, and will continue to change, regardless of what we do to it. Each one of the 130 million babies born each year gains dozens of new mutations. Some will gain mutations so toxic that they will never get the chance to have children of their own, while others will choose not to. The rest will pass down some of those new mutations to future generations. Some mutations will lead to slightly larger families, on average, and those will become more common in the human gene pool. Over time, other mutations will fade back. The variants that succeed in one part of the world will sometimes be different from those in other places. Some variants are beneficial at high altitudes but not low; some tend to spread in places with malaria and not in places free of the parasite.
Amidst all this churning change, another transformation has also been steadily occurring in our species. As Muller feared, the human gene pool is indeed gaining a burden of harmful mutations. Muller first proposed the concept of a mutation load at a time when scientists knew next to nothing about the biological details of mutations. For the most part, Muller just relied on math. It wasn’t until long after his death in 1967 that biologists began making precise measurements of the mutation load by surveying people’s DNA. It turns out that our species does carry a substantial burden of harmful genetic variants. While extreme mutations are rare, mildly harmful ones are abundant. They are accumulating in our DNA as we find more ways to shield ourselves from suffering and death.
In 2017, Alexey Kondrashov, a geneticist at the University of Michigan, got so worried about the emerging research on our mutation load that he published a book-length warning, called Crumbling Genome. It’s possible, Kondrashov said, that each generation will inherit a more burdened gene pool than the previous one. Depending on how quickly the mutation load grows, it might someday drag down our collective health.
Muller’s Germinal Choice plan might sound absurd, but Kondrashov believes that the mutation load is a threat we cannot ignore. He suggests there are some ethically uncomplicated things we might do today to defend against it. As men get older, their sperm accumulate more mutations. If they freeze sperm as young men, they can pass on less of a burden to future generations. If the mutation load gets worse despite such measures, our species might have to use CRISPR or some other gene editing tool to plug the rising tide.
“I hope that ‘War on Mutation’ is declared soon,” Kondrashov wrote.
The future probably won’t match the most extreme visions we can dream up. But it will disorient us. It will take what we thought were iron laws of heredity and stretch them in strange figures. In fact, the disorientation has already begun.
In the early 2000s, for example, fertility doctors began producing so-called savior siblings. When children developed leukemia or some other disease requiring a bone marrow transplant, some families would go through rounds of in vitro fertilization until they produced a baby with just the right combinations of HLA alleles to be a donor.
In 2011, a seventeen-year-old Israeli girl named Chen Aida Ayash was killed in a car accident. After her death, her parents asked for doctors to collect some eggs from her cadaver. They had to go to court to get permission, explaining to a judge that they wanted to fertilize Chen’s eggs, after which Chen’s aunt would bear them to term. After her own death, Chen would give her parents grandchildren.
These cases are carrying us into realms where old customs and rules start to sputter and fail. The words that we used to use to talk about heredity lose their old meanings, or take on new ones. And when people fight about those words, judges struggle to figure out who is right. In 2012, the US Supreme Court found itself in such a bind when they heard a case brought by a Florida woman named Karen Capato. Her husband, Robert, was diagnosed in 1999 with esophageal cancer. He immediately began depositing his sperm in a sperm bank, so that if his chemotherapy left him infertile, Karen could still become pregnant through in vitro fertilization. The treatment failed, and Robert died in 2002.
Karen did not have his frozen sperm destroyed after his death. Nine months later, she used some of it to fertilize her eggs and gave birth to twins. Karen filled out paperwork so that the twins could get Social Security benefits for their father’s death. But the Florida state government rejected her application. They pointed to their state laws, and held that children conceived after the death of their father couldn’t inherit his personal property.
After hearing Karen Capato’s appeal, the Supreme Court ended up ruling against her. But their 9–0 decision came only after hours of maddening oral arguments. The judges and the lawyers got bogged down in debates over the definition of a child. It was clear that the congressmen who laid down the rules about inheritance in the Social Security Act in 1939 couldn’t have imagined children being conceived months after their father’s death. “They never had any inkling about the situation that has arisen in this case,” grumbled Associate Justice Samuel Alito.
If genetic engineering ever becomes commonplace, the Supreme Court will probably find itself in even harder quandaries, where old laws provide even worse guidance to the new ways of tinkering with heredity.
A few cases have been brought by children against their parents for allowing them to be born with congenital diseases. According to these “wrongful birth” lawsuits, the parents were negligent for ignoring tests on the fetus before birth and going ahead with it anyway. Some ethicists now wonder if children in the future may sue their parents for not using mitochondrial replacement therapy to cure Leigh syndrome or some other devastating mitochondrial disease. If parents have the genome of an embryo sequenced and choose not to edit out a variant that puts people at a high risk of dementia, their children might hold them accountable.
It’s hard to say if such children would win. In some forms of mitochondrial replacement therapy, the nucleus from an unfertilized egg is moved to a new egg. Only then do doctors fuse a sperm to it. For a child in this case to claim they were harmed by coming into existence, they have to show they’re worse off as a result of the procedure. But if not for the therapy, somebody else would have been born—in other words, an embryo that inherited a different combination of genetic variants from its parents.
As a society, we are probably not prepared to handle these ethical dilemmas. But there are even more profound challenges to our concepts of heredity coming fast over the horizon. Fertilizing eggs months after a father’s death seems strange because it stretches the timing by which one generation produces the next. But the process of heredity that takes place is utterly conventional. Karen and Robert Capato, for instance, produced lineages of cells in their bodies that gave rise to germ cells. They shuffled their chromosomes through meiosis as cells have done for billions of years. The germ cells combined, joining their genes together to produce embryos. And then the embryos developed into children with germ cells of their own.
Not even gene editing with CRISPR changes this series of events. Once scientists began using CRISPR on ground-cherries and mice, their descendants still inherited DNA. The only difference was that some of the variants they inherited were put in place by people, rather than through spontaneous mutations. It’s as if scientists were rerouting a river: Even in its new configuration, the river still flows.
But some recent advances in research may alter heredity itself in a far more profound, puzzling way. In one of them, scientists accidentally broke through Weismann’s barrier.
In 1999, a Japanese biologist named Shinya Yamanaka opened a new lab at the Nara Institute of Science and Technology, hoping to find a way to make a mark for himself in a crowded field. Before coming to Nara, Yamanaka had discovered some genes that were active in the early embryos of mice. Many other scientists were also studying mouse embryos, figuring out how embryonic cells take on different identities. They pinpointed proteins that could push a lineage of stem cells to become muscles or neurons or other types of tissue. In the 1990s, research on embryonic cells raised hopes for a new way to treat diseases. Scientists could pluck a single cell from an embryo made in a fertility clinic and use it to make a colony of embryonic cells in a laboratory dish. With the right chemical signals, the embryonic cells could keep dividing into new embryonic cells for six months. A number of scientists began predicting that this method would make it possible to grow tissue on demand. People with Parkinson’s disease could get transplants of healthy neurons. After a heart attack, doctors could repair a patient’s damaged cardiac muscle with new cells.
Yamanaka thought that if he joined the chase, he’d get trampled into obscurity. So he decided to turn around and head in the opposite direction. Instead of figuring out how to turn embryonic cells into adult cells, Yamanaka would try to turn adult cells back into embryonic cells.
No one else was trying to pull off this trick, and with good reason. Turning back developmental time seemed impossible. If you trace a branch of the human body’s pedigree from the fertilized egg to any cell in the adult body, you travel a long, twisted route. There may be hundreds or thousands of branching points along the way where one cell divided in two. And within each generation of cells, there was a flurry of biochemistry that made it possible for a different flurry to take over in daughter cells. To turn an adult skin cell back into an embryonic cell would seem to require traveling through all that history to the beginning, running all that biochemistry backward.
But Yamanaka suspected that our inner heredity might not be so hard to override after all. A few experiments carried out over the years gave him some hope. In 1960, for example, a British biologist named James Gurdon destroyed the nucleus in a frog’s egg and replaced it with the nucleus from the animal’s intestines. The egg began dividing, and eventually it grew into another frog. With this experiment, Gurdon had cloned the first animal. And in the process, he also showed that the genes in an adult cell could be reprogrammed to build an embryo all over again. In 1996, the Scottish biologist Ian Wilmut and his colleagues achieved much the same thing, this time in a sheep, creating a clone they dubbed Dolly.
Yamanaka wondered if there might be a simpler way to reprogram an adult cell to make it embryonic. To understand what makes embryonic cells embryonic, he looked for genes that were active only early in life and became silent in adulthood. Yamanka discovered some genes for proteins that acted like master switches, grabbing onto many genes in the cells and either shutting them down or turning them on. Yamanaka contemplated the idea of flooding adult somatic cells with proteins like these. They might seize control, forcing the cells back to an embryonic state once more.
It was, Yamanaka knew, a long shot. While he was aware of a few proteins that were active in embryonic cells, he had no idea how many others he would have to manipulate. There might be dozens, even hundreds. “We thought at that time that the project would take 10, 20, 30 years or even longer to complete,” Yamanaka said.
Yamanaka organized his lab to start hunting for the proteins in mouse embryos. Five years of searching brought them two dozen. The scientists then tested each of the genes to see if it could reprogram an adult cell. They would add extra copies of a given gene to a skin cell from an adult mouse. The extra genes would flood the cell with extra copies of their protein. But the adult cell always stubbornly remained adult.
As the disappointments piled up, a graduate student named Kazutoshi Takahashi suggested that they stop testing the proteins one at a time. Instead, they should flood cells with all twenty-four of their proteins at once. Perhaps the combination of all the proteins might be able to deliver a little nudge to the cells. Even such a tiny sign of hope would tell them their work wasn’t in vain.
Yamanaka gave his blessing to the experiment, although he was sure Takahashi would fail. Takahashi inserted all twenty-four genes into the skin cells and waited to see what happened. Four weeks later, Takahashi came to Yamanaka with news. The adult skin cells had turned themselves into what looked like full-blown embryonic cells.
“I thought this might be some kind of mistake,” Yamanaka said. He had Takahashi rerun the experiment many times over. Time and again, the cells turned embryonic.
It was impressive enough that the cells looked like embryonic cells and made the key embryonic cell proteins. But Yamanaka wondered if they could behave like embryonic cells, too. His team injected a few of the reprogrammed cells into early mouse embryos to find out. The embryos developed into healthy adults, and the scientists found that the reprogrammed cells had given rise to normal adult cells scattered throughout the body.
This success led Yamanaka to wonder if flooding cells with all twenty-four proteins was overkill. He launched a new experiment, creating cocktails containing only some of the proteins and leaving others out. His lab found they needed only four proteins. Working with James Thomson at the University of Wisconsin–Madison, Yamanaka demonstrated that human cells became embryonic with the same simple recipe.
In his reports on the experiments, Yamanaka referred to his reprogrammed cells as induced pluripotent stem cells. Other scientists began testing out these cells, hoping they would prove even better than embryonic cells for medical treatments. It was easy to imagine doctors taking skin cells from patients, reprogramming them, and then coaxing the induced pluripotent stem cells into any type of adult cell they needed. Because the cells belonged to patients themselves, there wouldn’t be any worry about rejection of foreign tissue.
In 2012, Yamanaka won the Nobel Prize. The prize not only recognized the practical promise of induced pluripotent stem cells; it also honored his discovery of a new way to think about time. August Weismann had pictured the body as a branching tree of cells, the branches splitting through time. We could split our development into milestones: day 1, fertilization; day 2, two totipotent cells; and so on through the calendar of life. Each milestone had to come after the previous ones, because it depended on them. The heart could not appear before the three germ layers, because the heart had to form from one of those layers. Time gradually stiffened our inner heredity, committing each lineage to a single fate till death.
Yamanaka showed that time is not actually essential to the difference between an embryonic cell and a cell in the gall bladder or a hair cell in the ear. Our ancestors evolved a way to develop over time, to build one biochemical reaction on another in lineages of cells. But we can just push cells from one state to another.
Yamanaka didn’t just undermine the power of time with his research; he also undermined some long-held beliefs about the germ line. The germ line has come to be seen as an all-important thread of heredity that is the sole link from one generation to the next. But this is a convenient fiction. When sperm and egg combine, they produce an embryo that has no distinct germ cells at all. Any cell in the embryo can, at that stage, give rise to new germ cells (or any other kind of cell). The germ line breaks, in other words, and only later in an embryo’s life is it rebuilt. By turning somatic cells into germ cells, Yamanaka could sneak around Weismann’s barrier.
Induced pluripotent stem cells behave much like the earliest cells in the embryo before the germ line reappears. With the right signals, they can develop into germ cells, just as they can become other types of tissue. In 2007, Yamanaka and his colleagues injected induced pluripotent stem cells into male mouse embryos, and found that some of the injected cells developed into sperm. The chimeric mice could even father mouse pups of their own with these sperm, which had come from a different mouse.
In order for the induced pluripotent stem cells to become sperm, a mouse’s body sent it a particular series of chemical signals, guiding their development. Yamanaka’s experiment led other researchers to wonder if they could deliver the same signals to cells sitting in a dish rather than in a mouse. In 2012, the Japanese biologist Katsuhiko Hayashi managed to coax induced pluripotent stem cells to develop into the progenitors of eggs. If he implanted them in the ovaries of female mice, they could finish maturing. Over the next few years, Hayashi perfected the procedure, transforming mouse skin cells into eggs entirely in a dish. When he fertilized the eggs, some of them developed into healthy mouse pups. Other researchers have figured out how to make sperm from skin cells taken from adult mice.
Translating those results into experiments on human cells has proven hard. Some researchers have managed to turn a man’s skin cells into a precursor of sperm called spermatids. But these transformed cells don’t easily undergo meiosis, shuffling their DNA and pulling it into two sets.
Nevertheless, the success that Yamanaka and other researchers have had with animals is grounds for optimism—or worry, depending on what you think about how we might make use of this technology. It’s entirely possible that, before long, scientists will learn how to swab the inside of people’s cheeks and transform their cells into sperm or eggs, ready for in vitro fertilization.
If scientists can perfect this process—called in vitro gametogenesis—it will probably be snapped up by fertility doctors. Harvesting mature eggs from women remains a difficult, painful undertaking. It would be far easier for women to reprogram one of their skin cells into an egg. It would also mean that both women and men who can’t make any sex cells at all wouldn’t need a donor to have a child. A man left infertile by chemotherapy, for example, could use a skin cell to make sperm instead.
Some researchers think that in vitro gametogenesis will trigger an explosion in the test-tube-baby business. Henry Greely, a bioethicist at Stanford University Law School, explored this possibility in his 2016 book The End of Sex and the Future of Human Reproduction. Greely speculated about a future world “where most pregnancies, among people with good health coverage, will be started not in bed but in vitro and where most children have been selected by their parents from several embryonic possibilities.”
Today, parents who use in vitro fertilization can choose from about half a dozen embryos. In vitro gametogenesis might offer them a hundred or more. Shuffling combinations of genes together so many times could produce a much bigger range of possibilities.
Even after ruling out the embryos with disease-causing mutations, parents would still have many embryos left to choose from. They might pick embryos with variants that could affect the color of their children’s eyes. Or they might follow Stephen Hsu’s call, and pick out embryos that have a combination of variants that have been linked to higher intelligence scores.
But the implications of in vitro gametogenesis go far beyond these familiar scenarios—to ones that Hermann Muller never would have thought of. Induced pluripotent stem cells have depths of possibilities that scientists have just started to investigate. Men, for instance, might be able to produce eggs. A homosexual couple might someday be able to combine gametes, producing children who inherited DNA from both of them. One man might produce both eggs and sperm, combining them to produce a family—not a family of clones, but one in which each child draws a different combination of alleles. It would give the term single-parent family a whole new meaning.
The possibilities go on. Instead of three-parent babies, one can envision a four-parent child. It might be possible someday for four people to swab their cheeks and have induced pluripotent stem cells produced. Scientists could then turn the cells into sperm or eggs, which could then be used to make embryos. Two people would produce one set of embryos, and the other two would produce a second.
At the earliest stages of development, when the embryos were just balls of cells, scientists could remove cells from each one and coax them to develop in a dish into more eggs and sperm. And those could be used to produce a new embryo. If that embryo was then implanted in a surrogate mother and allowed to develop, the child would inherit a quarter of its DNA from each of the donors.
We haven’t reached the age of multiplex parenting just yet. But it’s close enough that philosophers have been thinking seriously about what it would signify. It would make mitochondrial replacement therapy look like ethical child’s play. It would also leave children struggling to make sense of their own heredity. With some help from humans, any somatic cell can now gain the germ line’s immortality and give rise to a new organism.
Hayashi’s experiments push our language of kinship to the breaking point. The mouse pups he produced have a mother of sorts, although they descended from her skin rather than from one of her original eggs. The same might be said of a human child born through in vitro gametogenesis. But once scientists start carrying out rounds of fertilization in their labs, it will be hard to say exactly what their pedigree is. Can your parents be eight-cell embryos? Since those original embryos won’t be implanted, they will never become human beings. Robert Sparrow has argued that the embryos produced this way would be orphaned at conception.
Strange possibilities have a way these days of becoming real. To make sense of them and to make ethical judgments, we need a deep sense of heredity, of the full scope of what the word means. We have to recognize Mendel’s Law as one of many ways that genes move naturally from ancestors to their offspring, something that we are learning to manipulate. We have to recognize that the cells in our bodies have ancestors and descendants of their own—ones that can become mosaics or mingle together in chimeras. We have to loosen the boundaries of what we call heredity, to consider other ways in which today correlates with yesterday—be they molecules other than DNA that slip into future generations, or microbes that hitchhike along as well, or the tools and traditions of human culture, or even the environment into which our children are born. Only then will we have the language to talk about the ways in which we can control heredity for our benefit, and the dangers that we leave to the future.