Who will win the CRISPR-Cas9 appeal at the Federal Circuit?
Zachary Silbersher
The epic CRISPR-Cas9 patent dispute has reached the Federal Circuit. UC previously lost its bid to the Patent Office to eviscerate Broad’s patents through an interference proceeding. That decision is now on appeal. The briefing at the Federal Circuit is complete, and oral argument is scheduled for April 30, 2018. Who will win at the Federal Circuit—UC or Broad?
(The “Broad” parties include the Broad Institute as well as Harvard and MIT. The “UC” parties includes the University of California, University of Vienna and Emmanuelle Charpentier.)
The patent fight over CRISPR-Cas9 is about its application in eukaryotic cells, i.e., human and animal cells. The patents on that will confer exclusive rights on the potentially most groundbreaking scientific and commercial breakthroughs for years to come. CRISPR-Cas9 is gene-editing technology that has been touted as a revolutionary game-changer across multiple fields. It has potential to cure genetic disorders, such as cystic fibrosis and sickle-cell anemia, but also improve agriculture, among other benefits. To put it simply, the stakes are high.
What exactly is CRISPR-Cas9?
The CRISPR-Cas9 fight is about who will hold the patents for taking a naturally occurring system in single-celled organisms and implementing it in plant and animal cells.
CRISPR systems occur naturally only in prokaryotes. Prokaryotes are single-cell organisms that include bacteria and archaea. Prokaryotes are distinguishable from eukaryotes. The distinction between prokaryotes and eukaryotes is important for this matter. The heart of the revolutionary potential behind CRISPR-Cas9 revolution can be boiled down to taking a naturally-occurring system in prokaryotes and applying it to eukaryotes. That matters because eukaryotes include plant and animal cells, which is where the groundbreaking commercial and scientific potential resides.
“CRISPR” is itself an acronym, for “clustered regularly interspaced short palindromic repeats.” That’s a mouthful, and harder still because the term was coined before the function of the process was completely understood. Natural CRISPR systems in prokaryotes act as defense-mechanisms. When attacked by a pathogen, the CRISPR system acquires a short snippet of DNA from the pathogen. That foreign DNA snippet is then used as a form of immunity against future attacks from the same pathogen. This works by detecting and destroying similar DNA during subsequent attacks. To detect similar DNA, the DNA snippets from pathogens are stored together as a family. The family of DNA sequences show repeating sequences of nucleotides. In between those repeating sequences are the foreign DNA snippets previously acquired from foreign pathogens. The critical point here is that prokaryotes have a naturally-occurring system (CRISPR) that is very good at targeting and precisely cutting DNA.
At its most basic form, gene-editing is adding or removing genes from the genome. For that to work, gene-editing technology must be able to target and precisely cleave a DNA segment at a precise point. That is where the naturally-occurring CRISPR system in prokaryotes comes in. In order to acquire the foreign DNA snippets for immunity, the CRISPR system naturally harbors an effective mechanism for targeting and cutting precise DNA segments. Thus, the CRISPR-Cas9 revolution is all about harnessing that naturally-occurring mechanism in prokaryotes (bacteria) and applying a programmable version in eukaryotes (plant and animal cells). The patent fight between UC and Broad is about who came up with that idea first.
What is the patent fight between UC and Broad?
The pending appeal before the Federal Circuit technically arose from an interference proceeding at the Patent Office. An interference proceeding addresses what happens when two people each separately invent something at roughly the same time—who gets the patent? In most countries, the first person to file a patent application wins. That’s now the law in the U.S., but when Broad’s patents were filed, the old law still applied.
Under the old law, the first person who really “invented” the idea, or reduced it to practice, could prevail. That means, if persons A and B each have the same stroke of genius, and person A is first to file a patent application, person B could nonetheless still be awarded the patent if person B can prove that he or she actually “invented” the idea first. To do that, person B would have to provoke what was known as an “interference” proceeding. (In light of the US change to a first-to-file regime, interference proceedings will soon no longer exist.) Putting aside why the proceeding is called an “interference,” the point of an interference is for the Patent Office to determine—who had the idea first?—regardless of who made it to the door of the Patent Office first. (Had this fight between UC and Broad occurred a few years later, Broad would most assuredly have prevailed. Or, rather, UC would have sought to protect the IP through different tactics.)
Both UC and Broad filed patent applications directed to CRISPR-Cas9 inventions. (This dispute technically involves multiple patents and patent applications filed on both sides, but the particular differences between the patents and applications is not at issue at the Federal Circuit appeal.) Though both parties filed patents for CRISPR-Cas9 inventions, there was a fundamental difference between UC’s patents and Broad’s patents. Put simply, UC’s patents are genus patents, covering any environment, whereas Broad’s patents are species patents, specifically covering the eukaryote environment.
UC filed for a series of patents that identified the components of a programmable CRISPR-Cas9 system. The system had three components: (1) a Cas9 protein; (2) a “targeter” RNA that guides Cas9 to the target DNA; and (3) an “activator” RNA that activates the Cas9 protein to cleave the target DNA. Technically, there are different types of CRISPR systems. The Cas9 type refers to a specific protein. When the Cas9 protein forms a complex with the “targeter” and “activator” RNAs, they jointly act as a pair of molecular scissors that can, in theory, be programmed to precisely target specific DNA and cut it in a precise location. (UC also discovered that the two RNAs, the “targeter” and “activator”, could be fused together as one and still work effectively, but that purported invention does not appear at issue in this case.) Thus, UC’s patents were directed to using CRISPR-Cas9 very generally, either in prokaryotes, eukaryotes or outside of cells.
By contrast, Broad’s patents were filed later, but they are more narrow. They are directed to a system specifically for using CRISPR-Cas9 in eukaryotes. Broad’s patents purported to be based upon work showing the success of an engineered CRISPR-Cas9 system in eukaryotic cells. That work was published in an article, Cong 2013. Although UC’s patents included claims that actually covered use of CRISPR in eukaryotic cells, UC’s patents did not include any studies showing that it had actually accomplished that. Instead, UC published its own work, Jinek 2012, which showed results of CRISPR-Cas9 in an in vitro (test tube) environment, and purported to identify the system’s minimal and sufficient components. The Jinek 2012 article also pointed to using CRISPR-Cas9 in eukaryotic cells, but did not actually demonstrate doing so. UC did not finish its own experiments using CRISPR-Cas9 in eukaryotic cells until after submission of the Cong 2013.
Accordingly, because UC’s own work related to eukaryotic cells came after Broad’s Cong 2013 article, UC could not rely upon that work to undermine Broad’s patents—i.e., to argue that those results showed that UC actually showed CRISPR-Cas9 functioning in eukaryotic cells before Broad. Thus, UC made a tactical decision to provoke an interference using its earlier environment-neutral patents. In other words, UC used its its earlier filed patents—the patents generally directed to using CRISPR-Cas9 in any environment (prokaryotes, eukaryotes, outside of cells)—and filed an interference against Broad’s patents, which are specifically directed to using CRISPR-Cas9 in eukaryotic cells.
UC’s argument was simple: even though its earlier patents did not actually show CRISPR-Cas9 functioning in eukaryotic cells, it nevertheless was the first to disclose the minimum, sufficient components of doing so. Further, according to UC, actually carrying out the process of using the system in eukaryotic cells was a matter of routine experimentation. This argument, on its face, is stronger than it might sound. Under general principles of patent law, if part of your invention really is matter of routine experimentation, there is no need to actually carry out that experiment as a condition to getting your patent. Indeed, patents within the biological and biotech spaces are often granted and upheld in the absence of any studies or data actually showing the purported invention. It is nowhere near a hard-and-fast rule that experiments are required, but rather, the devil is in the details.
If UC’s tactic was successful, it would deprive Broad of any patents to using CRISPR-Cas9 in eukaryotic cells. Yet, to prevail, UC would effectively have to show that its work in the Jinek 2012 article rendered obvious Broad’s eukaryotic patents. Put another way, UC had to show that its in vitro work, which identified the minimal and sufficient components of a CRISPR-Cas9 system, made it a matter of routine and conventional experimentation to take UC’s discovery of those sufficient components, and implement them in eukaryotic cells.
To prove its case, UC made several arguments. UC argued that its in vitro results showing the minimum and sufficient components of a CRISPR-Cas9 system clearly motivated other scientists to try using the system in eukaryotic cells. UC argued that the community was already aware of numerous conventional and well-known techniques for introducing and expressing proteins and RNA in eukaryotic cells. Indeed, UC argued that proof of this was in the pudding—six groups successfully implemented CRISPR-Cas9 in eukaryotic cells within months of publication of Jinek 2012. And the groups used purportedly well-known techniques for doing so (which is potentially one of its strongest arguments in this appeal.) UC also argued the prior art showed other prokaryotic proteins to manipulate DNA in eukaryotes had been done successfully, thus reinforcing the view that it was routine to implement CRISPR in eukaryotes after UC’s work.
In response, Broad argued that there was nothing routine or conventional about taking the CRISPR system and successfully engineering it in the eukaryotic cells. Broad argued that even if Jinek 2012 motivated other scientists to use CRISPR-Cas9 in eukaryotic cells, there was no expectation of success. Indeed, just because there is a clear desire to do something does not mean figuring out how to do it is obvious. According to Broad, even if Jinek 2012 showed the CRISPR-Cas9 system’s minimal components, implementing it in eukaryotic cells was neither routine nor conventional. To support this, Broad pointed to the statements from the UC team itself. Members of that team publicly expressed doubt as to whether CRISPR-Cas9 system’s could be successfully used in eukaryotes. For instance, one of UC’s inventors, Dr. Jennifer Doudna stated after Jinek 2012, “[w]e weren’t sure if CRISPR/Cas9 would work in eukaryotes—plant and animal cells.” She also purportedly stated that getting CRISPR to work in human cells would be a “profound discovery.”
Broad also pointed to concerns expressed by other scientists that identified several reasons why CRISPR-Cas9 may not function well in eukaryotes. For instance, UC’s hired expert in this litigation, Dr. Dana Carroll, published an article suggesting that prior gene-editing technologies relied upon functions naturally occurring in eukaryotic cells, whereas CRISPR-Cas9 only naturally occurs in prokaryotes. He was also concerned, for instance, that CRISPR-Cas9 systems could be degraded by enzymes existing in eukaryotic cells but not prokaryotic ones.
Broad also drove home the multitude of structural differences between simple prokaryotic cells and complex eukaryotes, which Broad claimed diminished any reasonable expectation of success. The differences between prokaryotic cells and eukaryotic cells is not immaterial to this dispute. At its most basic, prokaryotic cells are simple whereas eukaryotic cells are complex. Eukaryotic cells form multicellular organisms, whereas prokaryotic cells do not. Eukaryotic cells includes nucleuses and complex structures to house DNA, and prokaryotic cells lack almost all of this structure. Eukaryotic cells and prokaryotic cells also have different environments (temperature, pH, and so forth) that may impact whether transferring one mechanism to another is easy or hard. There are many other differences. Broad’s point was that transferring a naturally-occurring mechanism in prokaryotic cells to eukaryotic cells was simply not straightforward. (That question, in essence, is really what the Federal Circuit appeal is all about.)
The Patent Office agreed with Broad. In short, the PTO found that UC’s work and its patents clearly motivated scientists to try the system in eukaryotes, but it did not show how to do it. And because of that, successfully doing it remained unpredictable. The PTO found that doubts expressed by both UC and Broad members showed that any success was not expected. The PTO stated, “contemporaneous statements cited by both parties persuade us that one of ordinary skill in the art would not have reasonably expected success before experiments in eukaryotic cells were done.” The PTO also found that the multitude of differences between prokaryotic and eukaryotic cells made transferring CRISPR-Cas9 to eukaryotes neither routine nor conventional. The PTO stated, “differences between prokaryotic and eukaryotic systems would make use of CRISPR-Cas9 in a eukaryotic system unpredictable even though it was known to work endogenously in prokaryotes.” The PTO also found that, even though multiple groups “immediately” succeeded in using CRISPR-Cas9 in eukaryotes after Jinek 2012, that only showed a motivation to try, but not a reasonable expectation that anyone would actually succeed.
What does UC have to do to win the appeal?
Among litigators, the age-old wisdom often passed around is, the best tactic for winning an appeal is not to lose the underlying case in the first place. That’s another way of saying, if you’re seeking to reverse a decision on appeal, you’re already behind the eight-ball. Unfortunately for UC, that is borne out in this case by the standard that will be applied by the Federal Circuit.
The standard on appeal for assessing the PTO’s legal decision is de novo, but the PTO’s factual determinations are subject to substantial evidence. In other words, UC does not have the benefit of the Federal Circuit reviewing this case with a fresh set of eyes. Instead, Court must give deference to the PTO’s factual findings.
That creates an uphill battle for UC because this appeal is about many factual findings. For instance, the PTO afforded considerable weight to the statements by UC’s team, including Dr. Doudna, and to its litigation expert, Dr. Carroll. They each appear to have expressed doubt that CRISPR-Cas9 could function in eukaryotes. UC argues that these statements were often taken out of context and, likewise afforded too much weight by the PTO. Even if that is debatable, those types of credibility determinations would technically be subject to considerable deference by the Federal Circuit. That means, unless they are egregiously wrong, they should not be disturbed by the appellate court. That’s the uphill battle that UC faces.
UC’s most promising argument is therefore a legal one, rather than a factual one. This is because the legal conclusions are reviewed de novo. UC argues that the PTO applied the wrong standard in determining that its own patents did not render obvious Broad’s patents. According to UC, the PTO found that neither UC’s patents or any other prior art relating to transferring mechanisms from prokaryotic cells to eukaryotic cells provided specific instructions on how to do it with CRISPR-Cas9. The standard for obviousness is technically a reasonable expectation of success, not an absolute expectation of success. If prior art had to precisely teach a purported invention in order to invalidate it, then no patent would ever be deemed obvious.
Here, therefore, UC argues that the PTO essentially demanded too much. Had it properly focused on the fact that Broad’s patents actually used conventional and routine techniques for showing CRISPR-Cas9 functioning in eukaryotic cells, then that would underscore that a reasonable rather than actual expectation of success existed.
UC’s argument definitely has traction, and UC’s chances of reversing the PTO’s decision are not negligible. Broad’s principle retort to this argument is that there was enthusiastic response to the Cong 2013, which published Broad’s success implementing CRISPR-Cas9 in eukaryotic cells. Yet, that reaction may have had more to do with the heightened interest in this technology, rather than indisputable evidence that Broad did something truly innovative by adapting UC’s Jinek 2012 findings to eukaryotes.
Therefore, despite the uphill battle faced by UC, we view the strongest fact in favor of UC to be that Broad’s success at implementing CRISPR in eukaryotic cells was purportedly based upon well-known techniques. The Federal Circuit is mindful of the importance of this case, and will undoubtedly pay close attention to whether overstepped the standard is finding that Broad’s invention was not obvious of UC’s earlier work. During oral argument on April 30, it will be interesting tell if the Court probes Broad’s lawyer for an admission that Broad’s own patents used conventional techniques, rather than something truly unconventional. This is arguably the biggest weakness in Broad’s case, and something to listen for during oral argument.
So how can we briefly summarize the arguments on both sides?
The big question presented by this appeal is whether it was “routine” or “conventional” (i.e., straightforward) to go from UC’s work (Jinek 2012), where UC identified the basic requirements for a CRISPR-Cas9 system, and using it in eukaryotic cells (Cong 2013), where Broad showed successful results. In other words, are Broad’s patents obvious in light of UC’s patents?
Both sides have clearly compelling arguments. For UC, the proof is in the pudding that it was routine step to go from Jinek 2012 to using CRISPR-Cas9 in eukaryotes. Within months, six different groups did it. More importantly, they used conventional techniques to do so, rather than groundbreaking techniques in their own right. UC’s claim is that they successfully mapped out the process, and Broad simply followed the map, and thus spoils of this “invention” should be awarded to UC. By contrast, for Broad, the proof is also in the pudding. The social, scientific and commercial potential in CRISPR-Cas9 is enormous. If you put enough brilliant minds behind anything, they will succeed. Jinek 2012 motivated everyone to try the system in eukaryotic cells, but actually doing it took work. Indeed, the UC team members admitted openly that pulling it off would be profound. If it was so obvious, then presumably UC would have done it first. But their team didn’t. And UC shouldn’t, according to Broad, get the benefit of any patents for work it didn’t do. At the end of the day, however, UC has the higher burden to meet to reverse the PTO’s interference decision.
What are the consequences of this case?
It bears emphasis that only Broad faces the prospect of losing its patents through this case (although there will have to be further determinations within the remanded interference proceeding at the Patent Office before that happens.) UC has its owns patents. Those patents technically encompass using CRISPR-Cas9 in eukaryotic cells, and those patents are not technically threatened regardless of whether UC wins or loses this interference appeal. That fact raises a couple of questions.
If UC loses this appeal, can it still license/use its own patents? The shortest answer is, Yes, but … Even if UC loses this interference fight, its own patents will survive. And that means that UC could theoretically license its own patents, which are very broad, and which technically cover eukaryotic cells. UC could technically sue any companies trying to capitalize on CRISPR-Cas9. More realistically, any startups hoping to do so will wisely seek a license to Broad’s patents (if they prevail on this appeal,) as well as to UC’s patents as well. So, UC and its follow-on startups could still cash in on the CRISPR-Cas9 revolution even if it loses this appeal.
Put another way, if UC loses, then both UC and Broad will both have patents covering CRISPR-Cas9 in eukaryotes. Both could potentially sue each other at any time. Although, for the reasons discussed below, UC’s patents will be significantly weakened compared to Broad’s patent. Broad’s patents would thus be stronger. Either way, investors could potentially demand that, to avoid litigation uncertainty, both Broad and UC take licenses to each others patents.
So what’s the “But”? If UC will keep its own patents, then what does it have to lose in this appeal? If UC loses this appeal, then that won’t technically invalidate UC’s patents. But, it will give anyone sued by UC for refusing to take a license a nearly slam-dunk argument for invalidating UC’s patents. During this proceeding, UC has argued that its own patents covered the eukaryote environment, and thus rendered obvious Broad’s patents. The PTO disagreed, and essentially held that even though UC’s own patents claimed CRISPR-Cas9 in the eukaryotic environment, the patents themselves did not disclose enough to show that a person of skill had a reasonable expectation of success in using CRISPR-Cas9 in eukaryotic cells.
That’s another way of saying that UC’s patents lack adequate written description for using CRISPR-Cas9 in the eukaryotic environment. And that’s another way of saying that UC’s patents are most likely invalid. Especially if UC tries to insist that Broad or any other companies are infringing its patents through use of CRISPR-Cas9 in eukaryotic cells, which is where all the commercial potential for CRISPR resides. Thus, if UC loses this appeal over the interference, then UC will still have its own patents, but their value will be considerably diminished. On the other hand, if UC prevails on this appeal, then the case will have to return to the Patent Office to parse exactly what UC invented and Broad did not. Broad could stand to lose all or a significant portion of its patent rights, whereas UC will hold the patents on using CRISPR-Cas9 on eukaryotic cells.
With so much to lose on both sides, and potentially so much to gain through a collaboration, it begs the question why there has not been a settlement. Some observers have remarked how vigorously both sides have pursued this battle. Presumably, given the tremendous potential of this technology, a collaboration between UC and Broad, either in the form of a patent pool or joint-licensing could still be worthwhile for all. Indeed, at least one patent pool for CRISPR-Cas9 organized by MPEG-LA has already formed.
Further, UC and Broad are not the only players in this space. Other CRISPR-Cas9 patents have been granted to other research groups already. A patent pool is likely in everyone’s interest, at some point. If the patents are too dispersed among too many parties, that may eventually yield the efficient infringement tactics that plague so many other industries in the US, thus indirectly devaluing the value of anyone’s individual patents.
What about IPRs? Can Broad or UC file IPRs against each other’s patents? They can. But the outcome of this appeal will greatly influence the result of those IPRs. If UC wins, then Broad loses its patents, and thus there is no reason to file IPRs. If Broad wins, then it could theoretically file IPRs against UC’s patents, but the arguments made during this interference proceeding will be used against it. Broad will have to argue that UC’s patents are obvious. But Broad’s patents include using CRISPR-Cas9 in the eukaryotic environment, and Broad has just paid millions of dollars to its lawyers to argue that doing that was really difficult, not routine, and definitely not obvious. In addition, Broad could not argue in an IPR that UC’s patents lack adequate written description (as discussed in the preceding paragraph,) because those types of arguments are not permitted in an IPR. They are, by contrast, permitted in an PGR (Post Grant Review), but UC’s patents are not likely eligible for Post-Grant Review because their filing date is too early. In short, if UC loses, no one will likely be able to invalidate those patents unless UC sues them first. In reality, that could significantly drive down the terms that UC can charge for licensing. If it demands too much, licensees will just forgo a license, knowing that any litigation will open up an easy shot to invalidate the patents anyway.
These types of enormous patent battles over groundbreaking technology that could literally change the world do not happen that often. Some may argue that this case is reminiscent of the telephone patent wars in the late 1800’s between Alexander Graham Bell and Elisha Gray, which also bore through an interference over who first invented certain aspects of the telephone. The result, whether just or not, and whether correct or not, minted Alexander Graham Bell in the minds of generations to come. The stakes may be equally high in this case. Listening to the Court’s questioning at oral argument on April 30 is likely to provide further insight into the outcome of this case.