As the technology to create genetically modified babies moves closer to practice, what questions should we ask before such procedures are contemplated? Amber Dance investigates.
In February of 2017, Lisa Salberg began a new life, with a new heart to replace the one that had beat inside her chest since birth. In June, she had another life-changing experience, as she witnessed the dissection of her old heart—sick and withered as a result of inherited hypertrophic cardiomyopathy (HCM). Salberg wondered, “How did I not die?”
HCM has run in her family since the early 1900s. Her great-uncle, her great-grandmother, her grandfather, her father, aunt and uncle—all had it. So, too, do Salberg's cousin, niece and nephew, and her daughter, who got a defibrillator implanted at the age of 10. In August, researchers at Oregon Health & Science University in Portland reported the first experiments aiming to edit out a particular HCM allele in a handful of human embryos using the gene editing tool CRISPR–Cas9 (ref. 1). “I love the idea that I might have a grandchild free of hypertrophic cardiomyopathy,” says Salberg, founder and CEO of the Hypertrophic Cardiomyopathy Association in Denville, New Jersey. But she recognizes that research in this area is just beginning, and clinical success remains far down the road.
Even if current challenges in optimizing gene-editing technology can be addressed and our knowledge of the effect of genetic variants in the human population advances, the prospect of human germline modification engenders a deeper debate. On one side shines the hope of eliminating devastating diseases like HCM, Huntington, or Tay–Sachs. But on the other looms the dark specter of eugenics, the fear of a world in which a wealthy, genetically perfected class dominates poorer have-nots, a world in which diversity is stifled and those with unrepaired disabilities lose support services. What will be the long-term psychological and physiological impacts for offspring and their families of germline modifications? What technical challenges still require solving? When might it be appropriate or necessary to modify the genes of embryos? And where do we draw the line between just healing our bodies and enhancing them?
A few examples exist in which germline modification may have advantages over prenatal genetic diagnosis (PGD) to avoid disease in a child (Box 1). But before such procedures become mainstream, society needs to seriously weigh the consequences of allowing gene modification and enhancement in human reproductive cells and embryos.
Germline engineering of human embryos has been something of a taboo for decades. As far back as 1990, the French National Consultative Bioethics Committee opined that gene therapy should be limited to somatic cells. Similarly, in that same year, Germany's Embryo Protection Law prohibited germline genetic modification. The UK followed when, in 1992, its Clothier Committee on Ethics of Gene Therapy recommended a ban on germline gene therapy due to scanty knowledge on potential risks. Other nations, including Canada and South Korea, have instituted bans on genome modification.
Yet until the rapid uptake of CRISPR–Cas9 gene editing in recent years, germline gene modification looked like a difficult feat—though it was shown to be technically feasible in rats in 2009, using zinc-finger nucleases2. And there were hints earlier on that gene therapy could alter sex cells. In 2001, the US Food and Drug Administration (FDA) paused a trial of traditional somatic-cell gene therapy for hemophilia after a patient's semen, though not his sperm, tested positive for sequences from the gene-delivery virus. The FDA then allowed the trial to continue, with regular testing of semen samples3.
In their paper from August of this year, Shoukrat Mitalipov of the Oregon Health & Science University and colleagues reported on germline editing studies that follow on those carried out by Chinese scientists4,5,6. Collectively, their experiments provide a foundation on which further scientific investigation of human embryo engineering research can build. Additionally, in September, researchers reported in Nature their successful use of CRISPR–Cas9 gene editing to study the role of the OCT4 gene in human embryo development, illustrating the value of the technique for basic research7.
Mitalipov and colleagues used sperm from a man with HCM, who was heterozygous for a four-base-pair deletion in MYBC3, to fertilize eggs from healthy donors. To avoid mosaicism—in which some but not all the cells in the embryo are modified—they injected the CRISPR–Cas9 editing materials at the same time as the sperm, so that endonuclease would edit the embryo at the single-cell stage.
As expected from a heterozygous sperm donor, in an un-edited, control condition, half of the embryos (9 out of 19) contained the HCM mutation, and half did not. In contrast, of the CRISPR–Cas9-treated set of embryos, 42 of 58 (72.4%) contained two wild-type alleles. It's impossible to know exactly how many were wild type before the editing, but that suggests about half of the embryos carrying the mutation were altered or corrected. Only one altered/corrected embryo showed any evidence of mosaicism, and the researchers were unable to detect any off-target editing.
The experiments of Mitalipov and co-workers confirm that much remains to be understood about the use of DNA endonucleases in human embryos. One bright point of the study is that it suggests that genetic chimerism may not be as much of an issue as was first suggested; that said, the low efficiency of homology-directed repair (HDR)—in which the endonuclease stitches a piece of donor DNA into the recipient's genome at the target site—remains a key problem to be solved before any such therapy reaches clinical practice. Another concern is that CRISPR–Cas9 activity often leads to deletions associated with another DNA-repair mechanism: nonhomologous end joining (NHEJ).
Of course, Mitalipov and co-authors note that their techniques would require improvements to efficacy and reliability before going anywhere near a fertility clinic. One surprise was that although the researchers provided a template of the corrected gene to the embryos, they found little evidence of that version in their embryos. Instead, they hypothesize that embryos replaced the mutation with sequences from the healthy maternal genome.
However, some question that conclusion. The authors of a preprint in the bioRxiv repository argue that if the Cas9 editing led to a large deletion in the mutant chromosome, Mitalipov and colleagues' PCR primers would have missed the MYBC3 locus entirely. What looked like the presence of only maternal code could be due to a lack of any PCR products from the edited allele8.
Even if the endonuclease makes the desired change at the right site, there's another challenge in planning human gene editing, whether of somatic or germ cells: CRISPR guide RNAs may also trigger genetic changes at off-target sites elsewhere in the recipient genome. And the location where those off-target sites occur will likely vary in different recipients, due to differing genetic backgrounds, an issue raised by David Scott and Feng Zhang of the Broad Institute of MIT and Harvard in Cambridge, Massachusetts, in a recent Nature Medicine paper9. Polymorphisms in a target site could render gene editing therapy ineffective, while variations that create a second, pseudo-target site elsewhere in the genome could lead to unwanted editing.
“When you treat disease at the level of DNA, you have to be aware that everybody has small differences in the sequence of their genome,” explains Zhang. He and Scott used the Exome Aggregation Consortium10 and 1000 Genomes11 datasets to figure out how CRISPR–Cas9 editing targeted at 12 different disease genes would affect people of diverse genomes. Most off-target interactions would be uncommon, which they suggest makes it all the more important to screen patient's entire genomes for those rare variants before trying to edit them.
Despite these caveats, according to Salberg, her community is excited at the prospect of someday editing out the HCM gene in in vitro fertilization (IVF) clinics. If germline gene editing were to become available, Salberg estimates that 40–50% of members of the HCM Association would look into it. But, she adds, access could be an issue. She expects that such a treatment is likely to be available, at first, at only specialized centers. Those in rural areas might not be able to take advantage of it right away. And, of course, price would be an issue for many. Salberg hopes insurers, or even the federal government, would choose to foot the bill, suggesting that preimplantation editing would likely be less expensive than caring for a person with HCM.
Social and ethical challenges
Even if the above technical issues can be addressed, germline engineering raises myriad ethical and societal issues. And it poses a much bigger risk to societies than somatic cell engineering because the changes introduced into individuals would be passed on to later generations. Another concern is that if such changes in the germline were carried out on a sufficiently large scale, alleles might be lost from the human gene pool that might later be discovered to be beneficial in some context.
One illustration of this concept is the sickle-cell anemia allele. Individuals with this allele can suffer blood vessel blockages and organ damage, even death, particularly among young children. But the allele has persisted in the human population through millennia in part because it confers resistance to the malaria parasite. This is beneficial if one lives in sub-Saharan Africa, where malaria is prevalent; not so much in Slough, UK. Of course, for sickle cell, we know the benefits of the disease allele. But for many other variants we may deem undesirable, protective features may be unknown, particularly if we do not look in the right genetic backgrounds and environments. Indeed, biochemist Jennifer Doudna of the University of Berkeley, one of the originators of CRISPR–Cas9 technology, points to the difficulty of identifying these features preclinically. What's more, Doudna adds, the fallout from editing a gene might not even be apparent until decades later.
Experts also caution that by changing the gene pool, germline gene modification could also lead to unexpected social consequences, just as China's one-child policy has in that country. Introduced in 1979 (and lifted in 2015, so that most couples are now entitled to two children), the policy had as its goal population control. Birth rates fell, and one-quarter of the population will be too old to work by 2050. And because of selection by couples for boys, males outnumber females by millions, a fact that has been blamed for a rise in violent crime in the nation.
Opponents to germline genetic enhancement also cite ethical concerns it raises about access. Should germline editing become practical in consumer-driven societies, the wealthy might then have the ability to endow their offspring with genetic advantages not available to other strata in society—a phenomenon termed 'consumer eugenics'. This might be athletic prowess, physical attributes or, one day, perhaps intellectual prowess.
On the flip side, some ethicists point out that human society has already crossed some of these boundaries. Education, nutrition and access to medicine are generally better for wealthier children. And outside of genetics, people are already using chemicals, protein medicines or devices to ameliorate certain diseases or to attempt to extend healthspan and lifespan. And cosmetics are in widespread use to temporarily enhance appearance.
What's different about germline editing is that it happens before the recipient could ever consent, and could potentially affect the entire gene pool. “You are not only modifying someone, but depriving that person's progeny of a kind of open-endedness about their futures that all of humanity, potentially, is entitled to expect,” says Sheila Jasanoff, a professor of science and technology studies at the Harvard Kennedy School of Government in Cambridge, Massachusetts. “All those other interventions only change the person who is being modified, and are usually done in adulthood at the person's own behest.”
Plus, germline gene changes are permanent. “Medications can be stopped, implants can often be removed, surgeries can sometimes be reversed or repaired,” points out Marcy Darnovsky, executive director of the Center for Genetics and Society in Berkeley, California. “None of that is ever likely to be true of germline genetic alterations.” A second long-held concern is that manipulation of the human germline would meddle with or contravene nature. Natural-born humans, as a species, can grow only so tall, run only so fast, live only so long. By moving the bounds on those parameters, genetic engineering could alter the very nature of humankind. This is often linked to another concern—the 'slippery slope' argument. Some say that once any type of genetic change in the human germline becomes acceptable, there lies the road to changing the nature of what it is to be human.
The community speaks
A December 2015 conference in Washington, DC, convened by the US National Academies, the Chinese Academy of Sciences and the UK Royal Society, brought together more than 500 researchers, ethicists, lawyers and advocacy groups from around the world, with the aim of producing guidelines for human gene editing. Although the summit agreed that research on somatic gene editing should continue under current regulatory control, its conclusions about germline editing were more cautious: “It would be irresponsible to proceed with any clinical use of germline editing unless and until (i) the relevant safety and efficacy issues have been resolved, based on appropriate understanding and balancing of risks, potential benefits, and alternatives, and (ii) there is broad societal consensus about the appropriateness of the proposed application”12.
A separate committee, convened by the US National Academies of Sciences, Engineering, and Medicine, issued a report earlier this year13. As in the previous meeting, they suggested that existing regulatory mechanisms could appropriately regulate lab research and somatic cell editing. They noted that human germline editing remains “highly contentious,” but stated that it might be permissible if it meets strict criteria, including an absence of alternatives and careful oversight.
Not everyone is confident of the adequacy of current regulatory bodies to manage gene editing. “We have a patchwork of voluntary and regulatory structures,” says Jasanoff. Those include Institutional Review Boards, which focus on risk to human subjects, she says, and the FDA, which is primarily concerned with the safety and efficacy of drugs. Gene editing “falls in between,” says Jasanoff.
Many governments draw a line between somatic and germline editing. About 40 nations prohibit germline modification for reproductive use, but not for research, says Darnovsky. In Europe, 29 states have ratified the Oviedo Convention, a treaty that permits genome modification only for research, prevention, diagnosis or therapy and prohibits any modification that would pass to a person's descendants. A UNESCO panel has called for at least a temporary moratorium on germline editing. Although the US does not prohibit research on editing in human embryos, it won't pay for it with federal research funds.
The American Society of Human Genetics also came down against germline editing in an August statement, which was co-developed or co-signed by ten other international groups14. Their position is that although germline gene editing is acceptable for research purposes, it's not appropriate, for now, in any procedure that would result in a pregnancy. The societies do suggest that clinical applications might be permissible in the future, if there were solid medical, ethical and societal justifications. The group also called for public funding for such research, suggesting that otherwise the studies will still take place, but move to locations with less regulation and transparency—with all the attendant risks.
Another set of researchers argued against germline editing in the pages of Nature in 2015 (ref. 15). Co-authors Edward Lanphier and Fyodor Urnov, who worked at Sangamo BioSciences in Richmond, California, fear that germline editing could have unpredictable effects. What's more, they suggest, public disapproval of germline modification could lead to restrictions on somatic cell editing, which they see as promising, and which forms the core of their business.
From therapy to enhancement
The US National Academies committee concluded: “Genome editing for purposes other than treatment or prevention of disease and disability should not proceed at this time”13. Such statements haven't stopped at least one biotech company from attempting to defy natural aging (Box 2).
But the distinction between germline therapy and germline enhancement can be difficult to make. “I don't know that there's a way to draw the line there,” says Darnovsky. “It's subjective, it's fuzzy—What counts as a serious disease?”
It's a moving definition, points out Charis Thompson, a professor at the Center for Science, Technology, and Medicine in Society, at the University of California in Berkeley, and a professor of sociology at the London School of Economics. Homosexuality was listed in the Diagnostic and Statistical Manual of Mental Disorders (DSM) until 1987. Getting depressed in the wintertime wasn't a disease until the definition of seasonal affective disorder. Infertility itself wasn't a medical condition until doctors could do something about it.
The decision as to what's treatment-worthy may well fall to bodies like the UK National Health Service or to private insurers, who choose what to pay for. But even if regulations in many nations prohibit germline editing or gene modification for the purposes of some definition of enhancement, that doesn't mean the procedures won't occur (Fig. 1). Darnovsky notes that once a treatment is approved, there's generally little oversight to prevent it from being used off-label.
And harsh rules might mean that procedures will go underground and become “only available to the super-rich,” suggests Anas Rana, a computational biologist at the University of Oxford in the UK. For example, in 2016 an American fertility doctor went to Mexico to perform a controversial procedure on a Jordanian couple that uses two eggs—one to provide the nucleus, one for the cytoplasm and mitochondria—to avoid passing on a mitochondria-based disease. He told New Scientist he went to Mexico because in that country “there are no rules,” (News Feature, p. 1012).
For her part, Salberg at the HCM Association worries that if clinics start offering designer-baby-type editing, it could lead to an outright ban on all germline therapy, even for folks like her community with a life-threatening disease.
What is disease, what is disability?
Another question: What constitutes a condition 'deserving' of genetic modification? Are there certain kinds of disability that should be eliminated at a parent's wish?
Disability is difficult to define. It can even be context specific, notes Alta Charo, a bioethicist at the University of Wisconsin Law School who co-chaired the National Academies committee for the recent report. For someone in a hunter–gatherer community, poor eyesight is a disability. For someone who can pick up eyeglasses at their local Walmart, it's more of an inconvenience.
And one person's disability can be another person's culture or community—for example, in the case of close-knit communities like the deaf. Some deaf parents even request PGD so that they can ensure their children will be born deaf, and thus take part in their culture and lifestyle.
Down syndrome offers a current case study of what happens when parents can control their offspring's genetics. Caused by trisomy 21, this condition can be detected in utero, and women often have the option to end their pregnancy. The rate at which women make that choice varies widely, by demographics and geography. One 2011 study found that within the US, Midwesterners were most likely to carry Down syndrome fetuses to term, whereas those in the West were less likely to do so. Age, education, ethnicity and marital status also influenced the rate of Down syndrome births16. Another study, based on US birth data from 2010, estimated nearly 900 Down syndrome fetuses were likely terminated in nine states. The estimated termination rate ranged from as low as 26%, in Indiana and Michigan, to as high as 52%, in New Jersey17.
Iceland, in contrast, has nearly eliminated Down syndrome altogether; the nation of 330,000 averages just two Down syndrome births per year. The majority of women undergo a prenatal screening test, and nearly all choose to terminate. There, it's seen as a chance to avoid suffering due to the condition.
But disabled people can have both high quality of life and value, says Rosemarie Garland-Thomson, co-director of the Disability Studies Initiative at Emory University in Atlanta. She herself was born with one arm shorter than the other, and six fingers altogether. “I have a very high quality of life,” says Garland-Thomson. “It's because I was able to go to school, and because I was able to access the kinds of technology that I need to do my work.” She's learned to be resourceful, for example, by using dictation instead of typing. Supporting her statements, a 1999 study reported that more than half of people with moderate to severe disabilities rated their own quality of life as good or excellent18.
Garland-Thomson says there's value in diversity, including all brain and body types, for society. So does Ethan Weiss, a cardiologist at the University of California, San Francisco. His daughter, Ruthie, was born with albinism that caused her to be legally blind, but that hasn't stopped her athletic or intellectual pursuits. He sees Ruthie's classmates learning compassion. For example, they know she needs to sit right at the front of the room to be able to see some of what a teacher or presenter is saying. When other kids are around, and claim Ruthie's in the way, students who know her quickly leap to her defense, says Weiss.
Eliminating a large number of people with disabilities would likely lead to a reduction in services that assist them, says Garland-Thomson, making it more difficult for people who still have disabilities to live their lives to the fullest. “We want to make a world where everyone can flourish as they are,” she says. “We don't do that when we try to imagine a kind of 'standard' person...What people need is an environment in which they can flourish, rather than a body in which they can flourish.”
Is society ready?
Though Garland-Thomson, bioethicists, scientists, genetic counselors and others are debating the pros and cons of genetic modification in conferences and the pages of scientific journals, what matters most is the opinion of a broader swathe of humanity, about whether germline gene-editing and related technologies are worthwhile, desirable, or unconscionable. The challenge is to obtain that “broad public consensus” called for by the 2015 summit and 2017 report.
“Nobody knows what the heck that means,” notes Darnovsky. Nor is it clear how to achieve it (Feature, p. 1029). Presumably, Darnovsky says, that consensus would require more than a handful of expert committees.
In a recent Science study, American researchers used survey data collected by YouGov to analyze how 1,600 US adults see genome editing19. Two-thirds approved of both somatic and germline procedures for therapeutic purposes. In terms of genetic enhancement, a split was apparent: only one-quarter supported germline alterations, whereas nearly 40% were in favor of somatic enhancement (Fig. 2).
That's just one source of necessary data; there are many opinions that might feed into a “broad consensus.” For example, opinions on human gene editing vary along geographic and religious lines.
Nations often fall into one of two groups, says Charo. Some, like the US, are focused on the politics of abortion and embryo rights. Others she terms “pro-natal.” These countries, such as Japan, are eager to expand their populations and frequently embrace technology that promotes the birth of healthy babies.
Israel is another pro-natal nation, notes Ephrat Levy-Lahad, a physician and director of the Medical Genetics Institute at Shaare Zedek Medical Center in Jerusalem. Government health insurance covers in vitro fertilization for up to two children, and about one in 25 babies owe their existence to the procedure, she says. The fertility rate hovers around three children per woman. Israelis tend to trust technology and medicine, she says, which might make safe and effective gene editing, for therapeutic purposes, appealing to many in the future.
The religious situation in Israel also contributes to the pro-technology attitude, Levy-Lahad adds. “In Judaism and Islam, life does not begin at conception, so issues like IVF or early-embryo manipulation, things like that are much more acceptable,” she says. In contrast, religions such as Catholicism disallow embryo manipulation. In nations like the US, the Christian-leaning beliefs of many lead to a concern about “playing God.” In the US YouGov data, those who reported that religion guided their daily lives were less supportive of either treatment or enhancement by gene editing.
Other cultural factors also play into whether people tend to accept genetic editing. In China, where genetic diseases are stigmatized and disability support is low, PGD is gaining in popularity. Clinics are limited to applying it in the case of infertility or to avoid serious conditions20.
Preparing for our genetic future
For now, it's not yet possible to safely and accurately alter the human germline, both technically and, in many places, legally. Therapeutic somatic modification, for the most part, is now in human testing. “I don't think we will see any enhancement uses—even somatic—any time soon, and certainly not in the US,” says Charo. In part, that's because many traits people might want to change are polygenic. “But is it possible flim-flam clinics will advertise such enhancements through editing? Sadly, I think it is.” For now, humanity has a crucial “window of time” in which to consider the consequences, says Charo.
And it's a window that won't slam shut anytime soon, adds Doudna. “We're going to have to continue to evaluate this as the technology evolves,” she says. “It's incredibly important to keep this discussion going.”