Source: The New York Times
A new gene-editing tool might create an ethical morass
— or it might make revising nature seem natural.
One day in March 2011, Emmanuelle Charpentier, a geneticist who was studying flesh-eating bacteria, approached Jennifer Doudna, an award-winning scientist, at a microbiology conference in Puerto Rico. Charpentier, a more junior researcher, hoped to persuade Doudna, the head of a formidably large lab at the University of California, Berkeley, to collaborate. While walking the cobblestone streets of Old San Juan, the two women fell to talking. Charpentier had recently grown interested in a particular gene, known as Crispr, that seemed to help flesh-eating bacteria fight off invasive viruses. By understanding that gene, as well as the protein that enabled it, called Cas9, Charpentier hoped to find a way to cure patients infected with the bacteria by stripping it of its protective immune system.
Among scientists, Doudna is known for her painstaking attention to detail, which she often harnesses to solve problems that other researchers have dismissed as intractable. Charpentier, who is French but works in Sweden and Germany, is livelier and more excitable. But as the pair began discussing the details of the experiment, they quickly hit it off. ‘‘I really liked Emmanuelle,’’ Doudna says. ‘‘I liked her intensity. I can get that way, too, when I’m really focused on a problem. It made me feel that she was a like-minded person.’’
At the time, bacteria were thought to have only a rudimentary immune system, which simply attacked anything unfamiliar on sight. But researchers speculated that Crispr, which stored fragments of virus DNA in serial compartments, might actually be part of a human-style immune system: one that keeps records of past diseases in order to repel them when they reappear. ‘‘That was what was so intriguing,’’ Doudna says. ‘‘What if bacteria have a way to keep track of previous infections, like people do? It was this radical idea.’’
The other thing that made Crispr-Cas9 tantalizing was its ability to direct its protein, Cas9, to precisely snip out a piece of DNA at any point within the genome and then neatly stitch the ends back together. Such effortless editing had a deep appeal: In the lab, the process remained cumbersome. At the time, though, Doudna didn’t think much about Crispr’s potential as a gene-editing tool. Researchers had stumbled on such systems in the past, but struggled to harness them. Nonetheless, she says: ‘‘I had this feeling. You know when you pick up a suspense novel, and read the first chapter, and you get a little chill, and you know, ‘Oh, this is going to be good’? It was like that.’’
Doudna arranged for a postdoctoral researcher, Martin Jinek, to collaborate with Charpentier’s team. After months of experimentation, they determined that Crispr relied on two separate kinds of RNA: a guide, which targeted the Cas9 protein to a particular location, and a tracer, which enabled the protein to cut the DNA. But even then, it wasn’t clear whether Crispr was anything more than a curiosity. Unlike most living things — people, animals, plants — the cells of bacteria have no nucleus, and their RNA and DNA interact in a different way. Because of that, Jinek says, it was hard to say ‘‘whether the system would be portable’’ — whether it would work in anything except bacteria. Going over the problem in Doudna’s office, Jinek began sketching the two RNA molecules on the whiteboard. In their natural form, the two are separate, but Doudna and Jinek believed that it would be possible to combine them into a single tool — one that was more likely to work in a wide range of organisms. ‘‘That was the moment the project went from being ‘This is cool, this is wonky’ to ‘Whoa, this could be transformative,’?’’ Doudna says.
The tool Doudna ultimately created with her collaborators paired Crispr’s programmable guide RNA with a shortened tracer RNA. Used in combination, the system allowed researchers to target and excise any gene they wanted — or even edit out a single base pair within a gene. (When researchers want to add a gene, they can use Crispr to stitch it between the two cut ends.) Some researchers have compared Crispr to a word processor, capable of effortlessly editing a gene down to the level of a single letter.
Even more surprising was how easy the system was to use. To edit a gene, a scientist simply had to take a strand of guide RNA and include an ‘‘address’’: a short string of letters corresponding to a particular location on the gene. The process was so straightforward, one scientist told me, that a grad student could master it in an hour, and produce an edited gene within a couple of days. ‘‘In the past, it was a student’s entire Ph.D. thesis to change one gene,’’ says Bruce Conklin, a geneticist at the Gladstone Institutes in San Francisco. ‘‘Crispr just knocked that out of the park.’’
Genetic engineering has wrought spectacular changes in the years since it was first developed in 1973. By breeding mice to have particular mutations, researchers have been able to explore the roots of diseases including cystic fibrosis and diabetes. It has also opened the door to new hybrids: pest-resistant corn with genes taken from bacteria, for instance, and yeast modified to churn out an antimalarial drug. As of 2014, the market for genetically engineered products was worth almost $2 billion — a number that is expected to double over the next five years.
But despite these advances, the process of altering genes has remained laborious and inexact. Engineering a mouse with a single mutation took a dedicated lab almost two years, and even that was something of a crapshoot. Altered genes frequently ended up in random locations, or else in widely varying numbers — no copies in one cell, a dozen copies in another — often with confounding results. One scientist told me that before Crispr, he had to microinject roughly a million cells in order to get one perfect mutation. With Crispr, he could get the same result using just 10 cells.
And mice were the best case. Other animals were far harder to engineer — weirdly, even rats were difficult — and many couldn’t be altered at all, for reasons nobody really understood. ‘‘There’s a reason the mouse became the model animal for human diseases,’’ notes Tom Cech, director of the University of Colorado’s BioFrontiers Institute and a Nobel laureate. ‘‘Before Crispr, trying to genetically modify any other animal was either impossible, or impossible to do with any kind of precision.’’ And because scientists could alter only a single gene at a time, moreover, they could barely scratch the surface of many disorders, like cardiovascular disease, that were thought to involve multiple, or, in some cases, even dozens of genes. ‘‘What most people don’t realize is how limited we were before Crispr came along,’’ Cech says. ‘‘The tools we had were extremely crude.’’
In the era of Crispr (short for Clustered Regularly Interspaced Short Palindromic Repeats, a reference to the gene’s structure), those limitations are already disappearing. In October, Harvard researchers used Crispr to simultaneously alter 62 genes in pig embryos, creating animals that could, at least in theory, grow human organs for transplant. Uncannily, the tool also seemed to work in nearly every organism, from silkworms to monkeys, and also in every cell type: kidney, heart and those, like T-cells, that researchers had previously struggled to modify. In early November, the biotechnology start-up Editas Medicine announced that it planned to test a Crispr-based gene therapy technique in hopes of curing a rare form of blindness, by deleting part of a gene that controls the eye’s photoreceptor cells. But most researchers believe that Crispr’s biggest impact will be in speeding up the drug pipeline. Drug development currently relies, in part, on genome-wide association studies to identify mutations that people with a certain disease have in common. The problem is that those studies typically turn up hundreds of loosely associated mutations, each of which may or may not actually relate to the disease. (Some of them may be caused by the disease.) Before Crispr, it was so difficult to edit a single gene accurately that researchers had no easy way to test which mutations actually mattered, and thus which ones to target when looking for a cure.
That alone would qualify as a major advance, but Crispr’s reach will almost certainly be far greater — in part because so many industries now rely on genetic engineering. Researchers have begun using Crispr to develop better biofuels and to create new enzymes for industrial markets, where they’re used in laundry detergents, water treatment and paper milling. In agriculture, companies are using Crispr to make crops more pest- and drought-resistant, without using genes spliced in from other species, like a flounder gene inside a tomato. (DuPont is collaborating with Doudna’s company, Caribou Biosciences, to grow Crispr-edited corn and wheat, which are expected to reach supermarkets within five years.) Livestock breeders can harness it to produce animals with more muscle mass and leaner meat, faster and more predictably than with ordinary crossbreeding. Food conglomerates, including Dannon, are already deploying Crispr to create strains of bacteria that produce more flavorful yogurt; other fermented foods — cheese, bread, pickles — will probably follow.
For researchers studying complicated psychiatric disorders, Crispr may be a particular boon. ‘‘The major roadblock in that whole field has been that mice are just not good models,’’ says Feng Zhang, a biologist at the Broad Institute who pioneered the use of Crispr in human cells. ‘‘They often don’t even have the same brain structures as are affected in those diseases.’’ Zhang and Robert Desimone, director of M.I.T.’s McGovern Institute for Brain Research, are among a number of researchers now hoping to use Crispr to generate primate models for illnesses like autism and schizophrenia, which are thought to involve multiple mutations in a variety of combinations.
Farther afield, researchers are considering how Crispr might be used to eliminate malarial mosquitoes, or target invasive species like Asian carp in the Great Lakes. ‘‘There’s an almost frantic feeling of discovery,’’ one scientist told me. ‘‘Crispr has made so many experiments possible — it’s like standing in a candy store and knowing that you can choose just three things. Meanwhile, there are a thousand more experiments that you wish you could try, if only you had the time.’’ One prominent scientist estimated that Crispr was now being used by nearly every genetic-engineering lab in the world.
Rodolphe Barrangou, a biologist at North Carolina State University who was on the team that first identified Crispr as part of a bacterial immune system, notes that it is nearly unprecedented for a single tool to result in such a scrum. ‘‘You’ve got Fortune 50 companies fighting with start-ups fighting with universities,’’ he says. ‘‘That almost never happens. But with Crispr, the range of potential uses is so huge — everybody wants in.’’
These days, Doudna works out of a lofty office with sweeping views of San Francisco Bay and U.C. Berkeley’s elegant marble bell tower in the foreground. On the day I arrived this spring, her husband, Jamie Cate, a biochemist specializing in protein formation and biofuels, was finishing lunch at his desk before catching a flight to Chicago, where he was slated to give a talk. Cate relocated his office several years ago so it would be next to Doudna’s; the couple has a 13-year-old son, and they try to avoid scheduling work trips that overlap — a process that has lately become more challenging. ‘‘Jennifer’s schedule is. … ’’ Cate rolled his eyes. ‘‘Working next door to her is sometimes the only way we can coordinate.’’
In person, Doudna is politely formidable. Tall and rail-thin, she has an unusual intensity; talking with her feels a bit like sitting opposite a heron as it eyeballs an unpromising fish. Among the scientists I spoke to, nearly all mentioned the uncanny nature of Doudna’s intuition; more than one described her as having ‘‘a kind of sixth sense’’ about which problems to tackle. ‘‘Everybody would love to be the one to identify the game-changing problem and solve it,’’ says Blake Weidenheft, a professor at the Montana State University who has also spent time in Doudna’s lab. ‘‘But when you’re looking around, trying to figure out what to work on, it’s not at all obvious which path is going to lead to that kind of breakthrough. That’s something Jennifer is very good at.’’
Doudna’s interest in science developed early. In the 1970s, her father, Martin, got a job teaching literature at the University of Hawaii, and the family settled in Hilo. At the time, Hilo had what a former schoolmate, Lisa Twigg-Smith, described as ‘‘a sugar-plantation culture’’ — multiethnic, blue-collar, tough — and Doudna, with her blond hair and earnest interest in school, stood out. ‘‘I was kind of a nerdy, geeky type,’’ Doudna says. ‘‘And I loved math. People teased me about it. I felt pretty much like an outcast.’’
In sixth grade, Doudna began spending time with Twigg-Smith. Both girls lived in a tract-home development where, Twigg-Smith says, ‘‘every third house was the same.’’ The development was new, but it was built into wild countryside; after school, Lisa would ride her bicycle over to Jennifer’s house, and the girls would head up the cane road, an overgrown track that led to a nearby sugar-cane field. The landscape was rugged and lush, a plane of steep lava rock overgrown with ferns and cut with flooded meadows where the grass grew in clumps.
‘‘We spent a lot of time walking and looking, just talking about whatever was the topic of the day,’’ Twigg-Smith says. ‘‘We’d follow wild-pig trails, or else we’d just look at things — at mosses that were in bloom, or mushrooms.’’ While on these walks, Doudna became interested in a species of mimosa known locally as hila-hila, or sleeping grass. The plant’s runners were thorny and tough, but its leaves were unexpectedly delicate and folded shut if touched. Doudna recalls being fascinated — not by the plant, but by the mechanism that drove it. ‘‘I’d look at that and think: Now, how does that work? There’s some chemical change that responds to touch, that allows these leaves to close. But what is it? I couldn’t imagine.’’
She was encouraged by a family friend, Don Hemmes, a biologist who specialized in fungi. She spent a summer working in his lab, collecting samples and embedding them in resins, then examining thin slices under an electron microscope. Though many teenagers would have found the work tedious, Doudna was captivated by it. ‘‘I looked forward to going in every day,’’ she says. ‘‘I would get this feeling — a sort of excitement. Like, ‘I’m going to uncover a mystery today!’?’’
From the start, Doudna told me, she has been drawn not to the larger forces in the natural world but to the hidden, microscopic mechanisms that drive it: the molecular chain that magically makes the hila-hila close. But in many ways she was an unlikely genetics pioneer. Before developing Crispr, Doudna had little experience editing genes and had never worked in human or animal research. And while she was well known in her field, her work — which focused on the structure and function of various kinds of RNA — was considered abstruse, even by science standards. ‘‘It’s a pretty rarefied interest,’’ one scientist told me. ‘‘To be good at it, you have to be very patient and meticulous.’’
At the time Doudna began studying RNA, in graduate school, the molecule was thought of primarily as a go-between: When a coil of DNA unzipped, the RNA would copy the orders encoded by a given gene, then ferry those instructions over to the cell’s protein-making machinery, which would use them as a blueprint for snapping together amino acids. But even as public attention (and research grants) focused on familiar problems like sequencing genes, the future was quietly arriving from another direction. To everyone’s surprise, RNA turned out to be both more diverse and more powerful than researchers imagined: Thousands of varieties have been identified, some with the power to turn genes on or off, or to disable invaders by jamming their genetic machinery. While DNA is a library containing all our genes, in other words, RNA is more like a fleet of dynamic librarians, constantly cataloging and patrolling the stacks.
Doudna made her name as a postdoctoral researcher by mapping the structure of a particular type of RNA known as a ribozyme: a molecule able to catalyze chemical reactions by twisting to bring different atoms in contact with one another. Researchers typically figure out how something like a ribozyme functions by laboriously tracing the individual interactions among atoms — a process that Tom Cech, who was Doudna’s postdoctoral adviser, likened to mapping a forest at night by feeling your way from tree to tree.
Having a detailed image of a ribozyme would bypass that process, but there was a reason that no such thing existed. At the time that Doudna began her work, Cech says, only a single RNA structure had been mapped in 20 years of attempts — and that one was much smaller and simpler. Mapping all the elements in something as complex as a ribozyme was widely assumed to be impossible. As Cech put it, ‘‘If we had asked the N.I.H. to fund this project, we would have been laughed out of the room.’’
The project took three years, but Doudna and her collaborators eventually succeeded. ‘‘They basically showed that this thing that people thought was absolutely impossible — imaging the 3-D structure of large RNA molecules — actually was possible,’’ Cech says. Doudna’s success started a stampede among researchers and was the beginning of the thriving field of RNA imaging. ‘‘She has an uncanny knack for picking the best experiments to answer a question,’’ Cech says. ‘‘From an early age, she had this talent for solving very daunting problems. Her whole career, in some ways, was preparation for Crispr.’’
One day this summer, I joined Doudna for a meeting at Caribou Biosciences, the biotech company she helped found in 2012. These days, much of the company’s research — like the research being done in hundreds of other labs — is focused on refining the Crispr technology and reducing its ‘‘off target’’ effects, where the Crispr mechanism cuts the gene in the wrong place.
That morning, Doudna was just back from testifying before a Senate committee that had been convened to grapple with Crispr’s implications — as well as its potential for misuse. A postdoctoral student described a process for making mice that had a mutation associated with human lung cancer, by using Crisprs packed into a virus that could be inhaled. Highly accurate mouse models of lung cancer could be made much faster, radically accelerating the pace of research. This was exciting but also alarming. As Doudna has said, a minor mistake in the design of the guide RNA could result in a virus that introduced the same cancer-causing mutations into human lungs. ‘‘It seemed incredibly scary that you might have students who were working with such a thing,’’ Doudna said in an interview with the journal Nature. ‘‘It’s important for people to appreciate what this technology can do.’’ While Crispr has opened the door to a huge number of experiments that were previously impossible, it has also enabled almost anyone to try them.
In March, for instance, researchers figured out how to use Crispr to create a ‘‘mutagenic chain reaction’’ — one that speeds up how quickly a mutated species can be produced. But because the chain-reaction mechanism is inherited, scientists note, it’s risky: Should the system escape to the wild, the mutation could end up in every plant or animal in the population.
Similar debates have surrounded even the earliest technologies for editing genes, as Doudna is keenly aware. In 2012, just months after Crispr debuted, she organized a meeting in Napa, Calif., to discuss the ethical issues this new technology raised. The concern at the time, Doudna explained, was whether someone might use Crispr to change the DNA of human embryos before the technology was genuinely ready. In April, researchers at Sun Yat-sen University in China published a paper describing how they used Crispr to edit a human embryo to repair the gene responsible for a potentially life-threatening blood disorder, though the embryos used weren’t viable and thus couldn’t be carried to term. Such experimentation is currently under a voluntary moratorium in the United States and tightly regulated in Europe. Researchers at Britain’s Francis Crick Institute recently petitioned for permission to pursue it, arguing that experiments like these could provide ‘‘fundamental insights into early human development.’’
There’s a vast difference between using Crispr to study embryonic development in the lab and using it to create a genetically modified infant who lives and grows to adulthood. (Which is again different from using Crispr to do gene therapy by fixing a mutation in cells taken out of a patient, then infusing those cells back in.) In practice, Doudna has said, it will most likely be years before we can safely edit the DNA of an embryo. But she also notes that the prospect of editing embryos so that they don’t carry disease-causing genes goes to the heart of Crispr’s potential. She has received email from young women with the BRCA breast-cancer mutation, asking whether Crispr could keep them from passing that mutation on to their children — not by selecting embryos in vitro, but by removing the mutation from the child’s genetic code altogether. ‘‘So at some point, you have to ask: What if we could rid a person’s germ line, and all their future generations, of that risk?’’ Doudna observed. ‘‘When does one risk outweigh another?’’
Debates are already simmering over whether certain qualities, including deafness, should actually be considered valuable sources of human diversity instead of disabilities. (The neurodiversity movement has made a similar case for Asperger’s, arguing that people with the diagnosis don’t need fixing: They just think differently, in ways that aren’t necessarily worse.) The other challenge will lie in how we’ll assess the inevitable risks inherent in creating the first genetically modified human. While researchers will surely work hard to mitigate those dangers — in part by testing Crispr extensively, on both human embryos and in animals, including primates — it’s not possible to foresee every potential error, genetic interaction or unintended consequence.
Should embryo-editing eventually become a realistic possibility, it will still require a leap. But it also won’t be the first time that we have created such a tool. The trick will be in deciding when the potential benefits of such a modification — a child who can’t get H.I.V., say, or is less likely to develop Alzheimer’s — outweigh the possible downsides.
Even as editing genes in human embryos remains a long way off, Doudna recognizes that it won’t always be. At the meeting in Napa, she said, ‘‘Someone at the table said, ‘There may come a time when, ethically, we can’t not do this.’?’’ She paused. ‘‘That kind of made everybody sit back and think about it differently.’’
Jennifer Kahn is a contributing writer for the magazine and a Ferris professor of journalism in the Council of the Humanities at Princeton University.