Urnov and Holmes had glimpsed gene correction—the ZFN created a DNA break, replaced by a piece of genetic information. They looped in Gregory, Rebar, and Miller to make a circle of five, and came up with a definitive experiment to rule out the possibility of an artefact caused by sample contamination. Urnov selected a cell that was homozygous for the gene being edited. If there was contamination, there would always be two forms of the gene. But if the only signal was the sequence being introduced, that would indicate that the native gene had been replaced—edited—by the external sequence. Over one long weekend, Urnov tested a group of edited cells. The first few were normal, unchanged, boring. The next was a heterozygote—one normal variant, one altered. This continued until—“Ode to Joy!”—he found a cell in which both copies of the gene had been changed. Urnov dashed off an email to Holmes: “A HOMOZYGOTE!!!” That was soon followed by “ANOTHER homozygote!!!”
There’s a Russian proverb that says: “If you grab the rope, don’t complain that the cart is too heavy.” It was time to open the curtain and put the finishing touches on the all-important scientific report. Urnov ran the most important experiment of his life—a wondrously low-tech experiment devised by Ed Southern in the 1970s, in which DNA fragments are ingeniously sucked out of a gel and transferred onto a nylon membrane with the absorbent assistance of a 4” stack of paper towels (the Southern blot). Holmes showed they could reverse the edited change, while the experiments in cancer cells were repeated in clinically relevant white blood cells.
Before submitting the paper to Nature, one of Sangamo’s key advisors, Sir Aaron Klug, proposed Urnov and colleagues use the term “high-efficiency gene correction” rather than modification. (The manuscript copy bearing Klug’s handwritten comments remains one of Urnov’s most prized possessions.) After two rounds of review, Nature published the report that rewrote the gene therapy playbook in April 2005.18 Sangamo had demonstrated the feasibility of correcting human genetic mutations. Moreover, the method avoided the problem of insertional mutagenesis that had marred the French gene therapy trial. “The ‘hit and run’ mechanism of ZFN action uncouples the therapeutically beneficial changes made to the genome from any need to integrate exogenous DNA, while still generating a permanently modified cell,” Urnov wrote.
When the Nature editors asked for ideas for a cover headline, Urnov suggested “genome editing.” (His father had just become the editor in chief of a Russian journal of literary criticism.) Five years after the completion of the first draft of the human genome, scientists had demonstrated the feasibility of rewriting the language of life to fix a genetic disease.
WIRED magazine’s Sam Jaffe reported on the landmark “nano-surgery” technique with a headline that hopefully earned the copy editor a bonus: “Giving Genetic Disease the Finger.”19 Jaffe quoted David Baltimore: “This doesn’t just deliver a foreign gene into the cell. It actually deletes the miscoded portion and fixes the problem.” The potential to target any gene in the genome was plain to see. Chandra’s review of the paper for Nature Biotechnology was entitled: “Magic scissors for genome surgery.”20
The next step was to move toward treating SCID patients, which required performing gene editing in stem cells. But to the team’s despair, all they found were small DNA sequence insertions and deletions, gene knockouts not precision repair. “This was, putting it mildly, not the droid we were looking for,” says Urnov. The impasse was broken in style by the company’s new chief medical officer, Dale Ando.
“I know exactly what to do,” Ando said. “And I know what gene, and what disease. We’re not going to do bubble boy disease. We’re going to do HIV.”
“Um, okay,” Urnov said
“We’re going to do CCR5 in T cells”
“Okay.”
“And we’re going to collaborate with Carl June.”
“Who’s that?”
Ando started laughing. Not a bad way to make an impression on your first day in the job.
Few areas of medical research were more urgent or competitive in the mid-1990s than HIV, which was first described as an acquired immune deficiency syndrome (AIDS) in a handful of patients in 1981. As the epidemic spread, scientists in the Bay Area observed that some people possessed a natural immunity to the virus. Meanwhile, several groups identified the protein receptor footholds—CD4 and a co-receptor called CXCR4—that enable HIV to gain entry into white blood cells.
In June 1996, five separate reports incriminating another membrane protein, CCR5 (C-C chemokine receptor-5), as a second co-receptor were rushed into print by the top three journals, all within a week of each other. If HIV was a blimp that is snagged by the Empire State Building (CD4), then CCR5 was the cable car ferrying passengers—the HIV genetic material—to the ground. Like tabloid newspapers, premier science journals can get competitive trying to be the first to publish a research breakthrough. Alas, one of those CCR5 reports21 in Cell was pushed into production so hastily that several pages ended up being printed upside down.
One of the senior authors of that report was Marc Parmentier, a Belgian physician-scientist, who had a hunch that abnormalities in CCR5 might explain the slow disease progression in some people exposed to HIV. Parmentier’s team took samples from three such individuals and found a glaring thirty-two-base gap (Δ32 or “delta 32”) in the middle of the CCR5 gene.22 The size of this deletion left little doubt that the function of the truncated protein was compromised. After testing hundreds of samples and volunteers, he found that the Δ32 variant was surprisingly common in Europeans—a carrier frequency (meaning one copy) of about 10 percent—but not a single HIV patient carried two copies of the Δ32 variant. A colleague showed that white blood cells with the Δ32 gene were resistant to HIV infection. By the time Parmentier submitted the report to Nature in July,23 another group had found the same results on a larger cohort of patients.
In the 1980s, Stephen O’Brien, a lab chief at the NIH, embarked on a search for genetic factors that influence HIV susceptibility and progression. O’Brien and geneticist Michael Dean began systematically screening candidate genes in their HIV population. After twelve years, the NIH team had examined more than one hundred candidate mutations in thousands of HIV patients without success. But the glut of CCR5 papers revealed one of the best candidates in years.24 On July 4, while O’Brien was at the cinema watching the premiere of Independence Day, his team was furiously sequencing samples. They too uncovered the Δ32 variant, but didn’t observe any Δ32 homozygotes in more than 1,300 HIV patients. About 1–2 percent of the American population is a Δ32 homozygote, but HIV patients almost never are. Without a portal into the white blood cell, HIV can land but it can’t infect.II
The geographic distribution of CCR5 Δ32 is interesting: it is most common in northern Europeans at a frequency of 5–15 percent. But as you travel farther south and east, the frequency drops—Δ32 is almost nonexistent in Africans and Asians. This pattern suggests that it must have been positively selected for a reason that has nothing to do with HIV, which didn’t cross over to humans until the early 20th century. O’Brien felt the only reasonable explanation was “a mysterious, but breathtaking, fatal infectious disease outbreak which, like AIDS, exerted a huge mortality, and from which CCR5 Δ32 carriers were resistant.” The prime candidate is the Black Death, which ravaged Europe throughout the Middle Ages. Perhaps the Δ32 variant arose in Scandinavia in response to an earlier plague.