Skip to Main Content

As soon as Harvard biologist Kevin Esvelt began reading the scientific paper, he had a desperate question: Who are these guys?

The authors had used a controversial new molecular technique to try to force a certain gene to be inherited by all of a fruit fly’s offspring. Confounding a basic principle of genetics, they had succeeded. Nearly every one of the young flies carried the gene, for yellow pigmentation.


While turning a bunch of flies yellow may sound innocuous, “gene drives” — as biologists call the cellular machinery that guarantees inheritance — have enormous potential promise as well as risks. Because gene drives could rapidly propagate novel DNA through an entire population in the wild, they could be used, proponents say, to eradicate marauders such as the cane toads overrunning Australia. They might make mosquitoes resistant to the microbes that cause malaria or dengue fever, or even block the gene that makes locusts swarm, saving millions of tons of crops every year.

But gene drives might also doom important species to extinction, change the course of evolution, and perhaps be used to create bioweapons. That has caught the attention of the United Nations office that oversees the biological weapons treaty as well as of the FBI’s Weapons of Mass Destruction directorate, as STAT reported last week. And a science meeting Thursday includes an eye-catching agenda item: gene drive’s potential for “entomological warfare.” 

“Most gene drive research is being done in academic and government labs,” said Dr. Amesh Adalja, an infectious disease and biosecurity expert at the University of Pittsburgh Medical Center who will be speaking at the National Academy of Sciences meeting. “But if a lone wolf or terrorist group is working on this, the regulation [of gene drives that government officials are contemplating] wouldn’t make any difference.”


Hence Esvelt’s alarm that he had never heard of the authors of the fruit-fly study, though he had spent weeks tracking down every scientist he thought was working on gene drives to get them to agree to proceed carefully. “I was kicking myself,” Esvelt said.

No experiments using gene drives in vertebrates, let alone humans, are known to be underway. But a growing number of labs are working with the technology in about six species, estimated Esvelt, a research fellow at the Wyss Institute for Biologically Inspired Engineering in Boston. As political scientist Kenneth Oye of MIT put it, “Man, are things moving fast.”

How fast?

In July 2014, Esvelt’s group, led by Harvard biologist George Church, laid out in the journal eLife the theoretical uses of gene drives and how to construct them using the powerful new molecular technique called CRISPR. They knew of no lab doing such work.

By January, Esvelt and colleagues reported creating gene drives in yeast, getting more than 99 percent inheritance; they’re also making gene drives in roundworms, partly to work out the basic recipe and partly in hopes that gene drives might eliminate diseases caused by the creatures.

In March, the fruit-fly scientists published the paper that shocked Esvelt.

“Man, are things moving fast.”

Kenneth Oye, Massachusetts Institute of Technology

In an attempt to identify possible ramifications of gene drives, the National Academy of Sciences has assembled experts to assess current regulations and recommend whether additional oversight is needed. The committee held its first meeting in July and will hold its sixth this week.

The Academy panel is expected to deliver a report next spring on, among other questions, biological techniques to reduce the risk of unintended consequences of gene drives.

From the evidence so far, however, the science threatens to outrace efforts to ensure that gene-drive research is conducted safely. For one thing, the key molecular tools are as easily available as tea cosies on Etsy. “In terms of parts you can buy on eBay,” Esvelt said, “probably for less than $10,000 you could” construct a gene drive. 

Flouting Darwin

Two groundbreaking discoveries facilitated the surge in gene-drive experiments. One was theoretical. In 2003 evolutionary biologist Austin Burt of Imperial College London outlined an ingenious way to flout the laws of both inheritance and evolution.

For more than a century, geneticists have known that, in organisms that pair up to reproduce, most genes have a 50-50 chance of being inherited, as Gregor Mendel famously showed in the 19th century with pea plants. Thanks to Mendelian inheritance, offspring carry either mom or dad’s version of the gene for, say, digesting lactose. Similarly, evolutionary biologists have known since Charles Darwin that natural selection eliminates inherited traits that reduce organisms’ fitness, such as mosquitoes’ susceptibility to the insecticide DDT.

Burt described how gene drive can outsmart both Mendel and Darwin. In nature, he noted, some genes copy themselves into multiple places in a genome. Like a dishonest candidate stuffing the ballot box, these extra copies increase a gene’s odds of “winning” — being inherited. Other genes destroy related ones, also ensuring their transmission into the next generation.

Although this natural gene drive is unpredictable and scattershot, Burt showed that certain tricks of molecular biology could achieve the same results: causing a gene to be inherited by many more organisms through many generations than standard genetics and natural selection allow.

But because the then-available laboratory tricks were inefficient and cumbersome, his idea lay on the shelf for nearly a decade.

Things changed abruptly in 2012. In a paper that June, scientists demonstrated how it might be possible to efficiently edit genes — that is, how to snip DNA at a particular spot and insert different DNA, a sort of biological version of word processing’s “find and replace.” This system, called CRISPR-Cas9, makes gene drive feasible. The 2012 paper used DNA in a test tube, but within six months researchers had built on it to edit genes in plants and animals.

“Since the 1970s we’ve been able to genetically engineer individual organisms,” Burt told STAT. “With gene drive” made practical by CRISPR, “we could change the genetics of vast populations.”

(Visualization by Talia Bronshtein/STAT; Source: PubMed)
Drag timeline handles to filter CRISPR studies by year, and hover over or click the colored circles to see publication details.

In their July 2014 paper, Esvelt and his colleagues described how gene drive would work: CRISPR would cut a target gene in a cell or cells, probably an egg or embryo. Cells would repair the cut by incorporating replacement DNA, also carried by CRISPR, resulting in an edited genome. Unlike in traditional genetic engineering, where researchers slip a foreign gene into an organism’s genome and call it a day, for gene drive they would insert, along with the replacement gene, what is essentially a molecular Xerox machine.

The copier would reproduce the replacement gene in such a way that almost every descendant of the engineered organism would inherit two copies of the introduced gene. After enough generations (exactly how many depends on factors such as generation length and how many organisms are engineered), every descendant would carry the foreign gene.

CRISPR-aided gene drive might be harnessed to target insect-borne disease. If CRISPR replaced the gene in mosquitoes that lets them detect the odor of people, and substituted a dud, and if gene drive ensured the dud was carried by both chromosomes, then every offspring would have a double dose of the dud. Eventual result: mosquitoes that can’t smell humans, reducing their odds of biting. Modifying malaria-carrying mosquitoes in this way could reduce the spread of a disease that kills an estimated 600,000 people every year.

This possibility was purely speculative when Esvelt and his colleagues outlined it in the summer of 2014. No one had used CRISPR-aided gene drive in mosquitoes or any other insect. Then the fruit-fly paper landed.

‘We were stunned’

Graduate student Valentino Gantz of the University of California, San Diego and his advisor, biologist Ethan Bier, didn’t set out to create a gene drive. Gantz was studying which genes produce the intricate vein patterns in the wings of fruit flies. To do that, he needed to create a menagerie of mutants, including some flies with multiple mutations — painstaking, time-consuming work.

“We thought, wouldn’t it be cool if you could come up with a trick that would give you the mutants you wanted in one generation?” Bier recalled. That trick was CRISPR. In the summer of 2014, Gantz used CRISPR to edit wing-vein genes. “We immediately understood that [the edited gene] could potentially propagate in a way that would lead to gene drive,” Bier said.

They therefore took precautions. Gantz conducted the experiment behind five sets of locked doors that required fingerprint identification to open. He housed the flies in a tube inside a tube inside a box. “And we thought it would be safer if I worked alone,” Gantz said, to minimize the chance that a mutant fly the size of a pinhead would hitch a ride on a human into the wilds of southern California.

Among the changes Gantz inserted in his flies with CRISPR was a gene for yellow pigmentation. That would be easier to spot than a wing-vein pattern. Gantz started the experiments in late November 2014. On Dec. 18 he saw the first signs of success: flies born from the CRISPR-modified embryos were yellow. He had edited the embryos’ genomes.

“With gene drive, we could change the genetics of vast populations.”

Austin Burt, Imperial College London

Then came the real test, determining whether the recessive yellow gene drove out dominant genes for other colors. Gantz mated the yellow flies to unedited ones. Their progeny hatched last Dec. 28. Nearly all — 97 percent — were yellow. It was an astonishing violation of the textbook laws of inheritance, which say only one in four offspring should express the recessive yellow trait. He had constructed a gene drive.

“We were stunned,” Bier recalled. The CRISPR-based gene-drive system “completely eliminated Mendelian constraints. It will revolutionize genetics.”

At a 97 percent rate of inheritance, one lone mosquito carrying an edited gene that prevents it from spreading malaria could make an entire population of mosquitoes resistant to the disease in less than a year.

Scientists had been trying to genetically engineer mosquitoes to stop spreading malaria for nearly two decades, without much success. A leader in that effort has been biologist Anthony James of the University of California, Irvine. In 2012, James’s team developed a mosquito that, thanks to a genetic tweak, produced antibodies against the malaria microbe, destroying it before it could be transmitted to the mosquito’s next blood meal. His lab has accomplished similar feats with mosquitoes that carry the dengue virus, but the genetic engineering has been inefficient and laborious.

Once the UCSD duo had their gene-drive flies, they e-mailed James. “Holy s***! James replied. “Can we do it in mosquitoes?

To find out, the two labs are collaborating. They’re using CRISPR to edit a mosquito genome to block the insect from spreading malaria or dengue, and gene drive to make every descendant inherit the trait. The team declined to divulge results, but James called them “promising.”

“We think we can make it so it’s completely impossible for a mosquito to propagate malaria,” Bier said.

In London, Burt and fellow Imperial College London biologist Andrea Crisanti are also trying to harness gene drive to disarm malaria mosquitoes. “We’re mostly focused on disrupting reproduction,” Burt said. Using an enzyme that targets some 200 sites on the X chromosomes in mosquito eggs, the team has managed to shred X chromosomes so thoroughly “that it’s too much for the cell to repair,” Burt said. The result: eggs carry only the Y chromosome, which makes sons, and no X, which makes daughters.

So far, they have gotten 95 percent sons using an editing tool other than CRISPR. And in a paper submitted to a journal, they will report that “CRISPR works in Anopheles,” the malaria-carrying mosquito, Crisanti said. “It speeds up” genome editing and gene drive “quite a bit.”

Whether other labs are secretly developing gene drives is, of course, impossible to say. Esvelt has made it his mission to keep that from happening. “I strongly believe we have a moral obligation to let others know when our research could directly impact their lives, and gene drives are a poster child when it comes to shared impact,” he said. “Working with community guidance is arguably the only ethical way to proceed.”

Unintended consequences

Esvelt heard about the UCSD paper a few days before it was published in Science last March. He immediately Skyped Bier to ask what precautions he had taken.

When Bier told him about the tube-in-a-tube-in-a-box confinement, Esvelt asked whether it was earthquake-proof. “I’ve thrown these things against walls, and still, the caps [on the tubes] don’t come off,” Bier recalled saying. “Even in an earthquake … This is not some kind of monster that can’t be confined.”

Esvelt wasn’t so sure Bier’s precautions were adequate. The UCSD scientists’ confinement system was “problematic,” he said. It “encourages other laboratories to do the same, which will eventually lead to an accidental release.” Even if such an accident had no ecological consequences, he added, it would make scientists seem like careless DNA cowboys, perhaps leading to calls to shut down gene-drive research.

Esvelt stepped up his efforts to get labs to agree to specific biosafety steps to prevent the escape of mutant organisms carrying gene drives, including inserting a sort of biological kill switch. The resulting letter was published in the journal Science in July with 27 signatories, including researchers from the only teams that have reported using CRISPR-aided gene drives — at the Wyss and at UCSD.

Taking their own advice, Esvelt and Church, who are authors of a patent filed on one form of gene drive, reported Monday that they had constructed biological machinery able to “overwrite” a CRISPR gene drive in laboratory yeast. Although the original gene drive remains in the organisms, it’s inactive.

Such reversibility, they say, could minimize the chance of gene drives released into the wild turning into an ecological disaster. That’s crucial, especially since releasing into the wild is exactly how gene drives could accomplish what their proponents envision. 

On the afternoon that Esvelt told STAT about his shock at the unexpected UCSD paper, he’d spent the morning with scientists from Australia asking whether gene drive could rid the island of cane toads. Introduced to Australia in 1935 to control beetles that were devouring sugarcane crops, the poisonous predators (native to South America) have been spreading unchecked. Toxic to birds, the voracious toads feed on insects and crowd out native species.

Where poisoning, hunting, freezing, and whacking have failed, gene drive might succeed. One idea would be to drive a gene that makes the toads die when they’re exposed to an otherwise-harmless compound. Gene drive might similarly eliminate other invasive species, including the zebra mussels clogging the Great Lakes, the Asian carp driving native species out of America’s waterways, and the pythons that have obliterated raccoons and rabbits in Florida’s Everglades.

While that may sound ecologically worthwhile, intervening in nature seldom goes as planned. For one thing, these invasive species are also established species, said environmental biologist Todd Kuiken of the Woodrow Wilson Center. “I don’t think we have a good way to evaluate what happens if we remove a species from a system as large as” the Great Lakes, let alone Australia, he said.

Nor do scientists know what else, besides what they intend, gene drive might affect in the wild. Genes don’t necessarily stay put. Through a process called horizontal transfer, they can jump from one organism to another, even an unrelated species. A gene drive that kills cane toads could jump to a benign creature. A gene drive that prevents locusts from swarming, benefitting farmers, might jump to bees, threatening their ability to pollinate and produce honey. “We need to understand what would happen if gene drive transferred into other species,” Kuiken said.

Perhaps the greatest worry is a reiteration of Esvelt’s “who are these guys?” experience. With the rise of do-it-yourself biologists operating outside institutions, it’s anyone’s guess what they’re doing.

MIT’s Oye, for instance, raises the “most extreme scenario” of bioterrorists altering the genomes of disease-causing organisms to make them more lethal or more infectious, and using gene drives to spread that trait throughout a population.

“I’m not saying that any decent, sensible human being would want to do it,” Oye said, “but it is possible.”