How’s this for strange? What’s likely the most important biotech patent battle of the decade is now being fought under outmoded rules that the U.S. Congress, in a rare spasm of common purpose three years ago, agreed roundly were due for a once-in-a-lifetime overhaul.
The fight is over CRISPR/Cas9, a potential Nobel-winning biotech discovery, and shorthand for a new way to edit and otherwise modify genomes. As a biologist’s research tool, it’s already invaluable. As a medicine, it could fulfill the promise of gene therapy, snipping out faulty genes that cause disease, perhaps replacing them with new, improved ones.
Who invented it and when is the subject of the fight which, like battles over other once-in-a-generation biotechnologies such as RNA interference, monoclonal antibodies, and polymerase chain reaction, could have consequences that resound for decades.
President Obama signed the America Invents Act—a reboot of the U.S. patent system—in 2011. The old system awarded patents to those who were first to invent; the new system rewards those who are first to file a patent application. Because of the grinding gears of putting laws in place, the CRISPR/Cas9 case revolves around the old paradigm—who was first to invent key aspects of the technology. Yet as we’ll see, some aspects of the new rules apply, too. Patent strategy is tricky enough as it is, but this is like playing three-dimensional chess on two different boards.
“There’s a lot of speculation about what can happen and lots of permutations,” says Chelsea Loughran, a patent attorney with Wolf, Greenfield & Sacks in Boston. (Loughran and her colleague Pat Granahan have been tracking the CRISPR/Cas9 IP situation for some time but have no ties to any parties involved.)
The scrum is all the more intense because the technology is fast to build and easy to use. Within a couple years of its invention, it has caught on across the world with scientists who want to cut out or replace genes in organisms from bacteria to mice to monkeys. Work in human cells is starting to emerge, too. New ideas and new uses seem to pop up every day.
“The time frame is so compressed here,” says Caribou Biosciences chief scientific officer Andy May. “The foundational work and the improvements are happening so much closer than in previous instances [of new biotechnologies]. Usually those improvements take many years to develop.”
In a conversation last week, May stressed to me he couldn’t talk about anything patent-related. As I wrote about last month, Caribou has teamed up with Intellia Therapeutics of Cambridge, MA, one of three for-profit startups in the race to turn CRISPR/Cas9 into the latest form of gene therapy. Intellia has exclusive license to Caribou’s suite of CRISPR/Cas9 technology for human therapeutic use.
Caribou’s IP stems from the University of California, Berkeley, where biochemist Jennifer Doudna worked on a crucial idea with others at Berkeley and beyond. They took a defense system that bacteria deploy against viruses, first discovered a quarter-century ago by Japanese researchers, and made key changes that turned the system into a gene-modification tool. They published their work in Science in 2012.
They filed a patent application in March 2013, just before the system switched from first-to-invent to first-to-file. The patent has not yet been granted.
Doudna’s co-inventor, Emmanuelle Charpentier, who has posts at institutes in Sweden and Germany, assigned her portion of the rights not to Caribou but to Crispr Therapeutics, a group founded by Versant Ventures in London. (I wrote about Charpentier’s decision and much of the background of the story for Start-Up earlier this year.)
Here’s where things get a bit twisted around. Until recently Doudna had ties to another CRISPR/Cas9 startup, Editas Medicine in Cambridge, MA. Editas launched a year ago with big backing from three blue-chip Boston-area venture firms and several of the area’s academic boldface names—plus Doudna herself—as founders. At the time of Editas’s launch, it was often reported that the company had some rights to Doudna’s work. That wasn’t so. She helped found Editas even though Caribou had exclusive rights to her foundational work.
As of Dec. 15, she was still listed as a founder. But Doudna is no longer affiliated with Editas, as first reported by MIT Technology Review earlier this month. Via e-mail, she told me she resigned from the scientific advisory board in June. She otherwise declined to comment, calling it her matter of policy not to “comment on processes and procedures while they are pending before the US PTO.”
Editas has plenty of other intellectual property in hand. It announced December 1 a first set of licenses that include work from nearly all its scientific cofounders as well as from Duke University’s Charles Gersbach, who specializes in Duchenne muscular dystrophy.
Emphasizing that there’s much more beneath the surface, Editas CEO Katrine Bosley gave me an outline of what the licenses bring to Editas. From Feng Zhang of the Broad Institute/MIT and George Church of Harvard University comes the use of CRISPR/Cas9 in eukaryotic cells; from Keith Joung of Massachusetts General Hospital, methods to make CRISPR/Cas9 perform more accurate cuts in DNA; and from David Liu of Harvard, delivery technology. Delivery—getting the right amount of CRISPR/Cas9 editing machinery into the right cells—will be no easy task. (Just ask anyone familiar with the slow progress of RNA interference-related therapeutics.)
“Together it’s a unique combination and foundational to the field,” says Bosley. “The concept of genome editing been around a long time” but creating a molecular entity that can do it is another thing entirely, she said.
Only Feng Zhang has received a patent for his work. Indeed, it was the first and remains the only CRISPR-related patent to be granted.
That patent is at odds with the work of Doudna, Charpentier, and colleagues (I’ll call them the Berkeley group for shorthand purposes). An examination of patent records shows why.
Before the U.S. patent office awarded Zhang and the Broad the patent, it rejected it. This isn’t unusual; most patents that are eventually granted go through temporary rejections, says Richard Blaylock, a patent attorney with Pillsbury Winthrop Shaw Pittman in San Diego, and a former biochemist.
But in one of its rejection notices, the PTO specifically pointed to the Berkeley group’s patent application, saying it “anticipates the claimed invention” described in the Zhang/Broad application. Zhang submitted a 15-page addendum, using dates of his own experiments and reports about the Berkeley group’s work to assert that he was first to invent the use of CRISPR/Cas9 to modify eukaryotic cells (that is, anything more advanced than bacteria or archaea).
Three months later, Zhang had his patent.
In a conversation this September, according to PTO records, the Berkeley group’s lawyers asked PTO to reconsider and open a type of appeal trial. It’s called interference and happens when two groups claim to invent the same subject matter. But this form of dispute resolution (there are others) is a legacy of the old patent system. “This kind of proceeding has gone away under the new law,” says Blaylock of Pillsbury.
The old system rewards those who were first to invent; the new system rewards those who are first to file a patent application. Because of the timing of the applications, this battle is being fought under the old rules. If indeed the PTO opens an interference hearing, each side will try to prove with lab notebooks, photographs, e-mail and other evidence that it was the first to invent this new gene-editing technology.
“So much of what happens in an interference proceeding is how good the notebook records and data are,” says Pat Granahan of Wolf, Greenfield. “I might have the best data in the world, get it to work in human cells, but I didn’t put the month, day, year on the page.”
No one I contacted for this story would tell me, on or off the record, if PTO has in fact agreed to re-open the case, or when the agency might make a decision.
So that’s one outstanding question: Will the PTO look at the Broad and Berkeley claims side by side? If so, says Granahan, it could be a “lengthy expensive process.”
“It’s not unusual for some of everyone’s claims to fall,” she says. “It’s unlikely for any party to come out of the proceeding with all the claims in their pocket.”
There’s a deeper issue at play that straddles the scientific and legal realms. Even if Zhang and colleagues say they were first to show that CRISPR/Cas9 could work beyond bacterial cells, does it matter? There is something in patent law called “obviousness.” If someone else invents a vehicle and demonstrates its usefulness on a dirt road, you might not be able to win a patent for a similar vehicle that drives on asphalt.
The beauty of CRISPR/Cas9 as a gene-modification tool is that, to change your target, you only need to change the sequence of the RNA guide that brings the enzyme—the scissors—right to the spot you want to cut. There’s no need to rebuild the scissors.
In other words: What’s the real breakthrough? The claim to invent the basic machinery? Or the claim to invent that machinery in the majority of cells that will matter for research and therapeutic purposes? Overly simple, yes, and of course the science and patent law will require far more complicated arguments. But it’s a good basic question to keep in mind as the fight proceeds.
For those who like a little intrigue in their patent scuffles, there’s also this question: Who submitted an anonymous challenge to the Berkeley group’s patent try in September, soon after Doudna cut ties with Editas? The fight might be governed by old “first to invent” rules, but the new patent rules have also expanded what amounts to a window for public comments. Encouraged by PTO, the comments can last for dozens of pages, and the commenters can remain anonymous.
In PTO-speak, it’s called a “third party relevance submission,” in essence an exhaustive compilation of prior art—examples meant to undermine an inventor’s case for a patent. Someone took pains to point out flaws in more than half of the 155 claims in the Berkeley group’s patent application. Think of it as an anonymous call—a very long call—to a detective that brings attention to alleged skeletons in someone’s closet.
These anonymous submissions are not trivial to produce, and even though new rules make them easier, challengers have to keep a close eye on the timing.
I was curious which other CRISPR-related patent applications have drawn similar challenges. The Zhang/Broad patent never attracted one, according to the PTO records. One application from a group at Lithuania’s Vilnius University did, but otherwise, all other applications I could find either hadn’t attracted a challenge yet, or haven’t had their contents made public (which of course makes a challenge quite difficult).
I would expect more challenges of all stripes to surface, but it’s hard to imagine the field getting bogged down. CRISPR/Cas9 is simply too widespread and too useful. Improvements will emerge at a furious pace, potentially rendering obsolete what’s currently state-of-the-art. Getting CRISPR/Cas9 into cells of all types, using various delivery mechanisms, is one wide-open area.
Another is making the technology more “specific”—hitting the targets it’s supposed to, and avoiding the ones it isn’t. “There still aren’t good ways to measure off-target activity on a genome-wide scale,” says Caribou CSO May. “We still suffer from the ‘looking under the lamppost’ problem,” which means researchers looking for problems can only find them by looking where the “light” is shining.
Still another is to expand CRISPR/Cas9’s limited DNA-cutting repertoire. The technology is limited because the version of Cas9 mainly in use now, borrowed from the bacterium S. pyogenes, only recognizes and snips genes with a certain nucleotide sequence. Cas9 from other bacteria recognize different sequences and might also be therapeutically useful.
Vinod Ranganathan, a researcher in the Johns Hopkins University School of Medicine’s ophthalmology department, says he’s found another way to expand that repertoire. He and colleagues published work this summer in Nature Communications that describes a different method of producing the short strands of guide RNA, giving them the flexibility of ending either with the nucleotide G (guanine) or A (adenine), instead of only with A. That, he says, gives the editing machinery access to more sites along the genome.
As the modifications, improvements, and new ideas accrue, the field’s foundational IP—both Editas’s Bosley and Intellia CEO Nessan Bermingham have described what they’ve licensed in exactly that way—isn’t likely to become obsolete. In RNA interference, Alnylam Pharmaceuticals (NASDAQ: ALNY) about ten years ago sewed up a huge portion of the so-called Tuschl patents and, despite ups and downs, has dominated the field ever since.
Go back further to monoclonal antibodies. As they rose to pre-eminence among biotherapeutics, Genentech squeezed as much value as possible from the Cabilly patents, originally filed in the early 1980s. Protein Design Labs couldn’t bring its own antibody products to market, but its Queen patents were spectacularly lucrative.
How ironic is it, then, that what’s arguably the most important biotech patent fight of the decade could be the last great fight under rules that no longer apply. Who benefits remains to be seen.
“This IP fight is going to burn up a lot of money,” Bermingham says. “It’s worthless if no product gets to the marketplace to help people.”
By posting a comment, you agree to our terms and conditions.