Guest post by Professor Jake Sherkow of New York Law School, who is currently a Visiting Scholar at Stanford Law School.
Yesterday, the Federal Circuit heard oral argument in the dispute between the University of California and the Broad Institute over a set of fundamental patents covering CRISPR-Cas9, the revolutionary gene-editing technology. Lisa has been kind enough to invite me to write a few words here about the dispute, and I thought I’d take that generous opportunity to discuss two aspects of yesterday’s argument: the basics of the appeal and, given that this blog is devoted to legal scholarship about patent law, what the argument can teach us, if anything, about IP scholarship in general. I think the short answer to the second question is, Quite a lot, although perhaps not for obvious reasons.
CRISPR-Cas9, for you luddites out there, is a breakthrough technology in molecular biology that allows scientists to “edit” DNA, both in a test-tube and in cells, in a way that’s cheap, easy, flexible, and almost infinitely programmable. Like all breakthrough technologies, there were other substitute technologies before it, but they were, by contrast, expensive, difficult, clumsy, and ad-hoc. To give, perhaps, one good illustration about how CRISPR-Cas9 took the world by storm: since the technology was first elucidated in 2012, there have been over 2,000 papers describing experiments using the technology—an astonishing level of scientific acceptance around the world.
The lure of the technology is its simplicity: it needs, essentially, only two ingredients—a DNA cutting enzyme, also called a nuclease (here, Cas9); and short piece of RNA, DNA’s molecular cousin, that can be programmed—crafting its As, Cs, Ts, and Gs—to bind to the gene of interest in the cell and activate the enzyme. Jennifer Doudna and Emmanuelle Charpentier—often referred to as the UC scientists, even though Charpentier was then at University of Umeå in Sweden—were among the first to figure all this out and get it to actually work in test tubes and bacteria. Virginius Šikšnys of Vilnius University in Lithuana also had a relatively contemporaneous paper showing, for the most part, the same thing. Both groups—the UC scientists and Šikšnys—separately filed for patents in 2012.
Shortly after the UC and Šikšnys papers, a number of groups tried to work the technology in the cells of higher organisms, notably humans. If it worked, it had the potential to spell a breakthrough for therapeutics; it could be used to edit defective copies of genes in living patients. In late 2012, Feng Zhang of the Broad Institute successfully performed those experiments in human cells and filed for patents covering his iteration of the technology.
Zhang’s filings—all of which matured into issued patents—ultimately triggered a patent interference between UC and the Broad Institute, with the UC’s patent application still-to-be issued. The interference saw the filing of thousands of pages of attorney argument, almost ten motions on dispositive issues, and expert testimony from a number of other molecular biologists in the field. The PTAB heard oral argument on the interference on December 6, 2016 and issued its decision on February 15, 2017.
Interferences are famously arcane. But the predicate test to conducting an interference is simply to determine whether there really is an interference in the first place: whether each side’s patents or patent applications are obvious in light of one another. (Perhaps evocative of some exceedingly rare play in baseball, this is called the “two-way interference test.”) And in its February 2017 decision, the PTAB concluded that they did not—that Zhang’s patents disclosed an invention that was, itself, nonobvious in light of Doudna’s patent application. The PTAB’s decision was long, and is itself worth a read. But the basis for its decision rested on its interpretation of a mound of conflicting evidence. Yes, Doudna did not disclose how, precisely, to move the system to human cells. But she did provide a roadmap of sorts. Yes, there was no “guarantee of success” in getting the system to work in human cells. And prior efforts for other gene-editing systems—the ones that were clunky and expensive—had some major failures. But the solutions, here, were simple and predictable. Weighing this—and relying heavily on the notion that a person having ordinary skill in the art would not, at the time, have had a “reasonable expectation of success” that simple solutions would work—the PTAB concluded there was “no interference in fact” between Zhang’s patents and Doudna’s patent application.
UC appealed. But this set up a tough task for UC. The “no interference in fact” determination is, itself, a factual one, and can only be overturned if it was made without the basis of “substantial evidence.” The standard should be familiar to those who litigate quasi-judicial fact-finding determinations from agencies, such as Social Security appeals. And, as one can guess, it’s a high bar: typically speaking, substantial evidence is satisfied so long as the evidence considered would have been enough to persuade a reasonable fact-finder of its outcome—even if that outcome is wrong. The standard does allow for reversals, but typically only if the lower tribunal committed reversible legal error.
UC’s brief thus focused on two things: how far the “substantial evidence” standard really went and whether the PTAB’s description of the nonobviousness standard—important for the two-way interference test—was correct. Being neither an admin law nor appellate procedure expert, I was looking forward to getting educating on the meaning of “substantial evidence.” I expected a raft of cases and examples where appellate courts tossed out good, thoughtful agency determinations for nonetheless lacking substantial evidence. I was sorely disappointed. UC barely touched the standard in its briefing—perhaps for good reason.
As for UC’s brief’s discussion of nonobviousness, I thought it was a closer call. Yes, the PTAB used some language in its opinion suggesting that a lack of a reasonable expectation of success meant that follow-on inventions were necessarily obvious. And that, would be a violation of some express language in KSR v. Teleflex that such considerations should be weighed flexibly in light of the Graham interpretation of prior art and other secondary factors. But I did think the UC took some of these statements out of context—perhaps, ironically, not unlike what the PTAB did concerning Doudna’s expectations of her technology working in human cells. I wasn’t persuaded; but I thought UC did the best with the hand they were dealt.
The oral argument panel consisted of Chief Judge Prost and Judges Schall and Moore. Donald Verrilli, former Solicitor General and he of Obamacare fame, argued for UC. Raymond Nimrod of Quinn Emmanuel argued for the Broad Institute. The arguments opened with UC being immediately challenged by the Chief and Judge Moore on the substantial evidence hurdle. Verrilli, seeing only enemies down that road, then tacked his argument to the legal error in the PTAB’s obviousness determination. Verrilli relied, heavy, on the contemporaneous success of Doudna’s colleagues in getting CRISPR-Cas9 to work in human cells, and how the PTAB’s review of this did not comport with KSR. This also, argued Verrilli, improperly shifted the burden or proof to UC rather than the Broad Institute. During this, the Chief Judge seemed to express some skepticism of the PTAB’s recitation of the standard. But then, as Verrilli’s description of the science unfolded, Judge Moore began to take Verrilli to task for simply engaging in a reweighing of evidence. Verrilli’s complaints that contemporaneous success with “off the shelf” tools was not inventive science sparked a vocal disagreement from Judge Moore, who explained that such efforts were common—but they didn’t mean anyone necessarily thought they were going to work. Verrilli, trying to get his explanation heard, was shut down. “That’s how science works, Mr. Verrilli,” explained Judge Moore. “Don’t interrupt me.” After forty minutes of fun—if you’re me—the case was submitted.
Given the standard, the briefs, and the tenor of the argument, I think it’s likely that the Broad Institute will win. Judge Moore seems to think—and I share her view, if so—that the substantial evidence controls the outcome in this case. Substantial evidence is a high, high bar. And we could disagree with the PTAB’s decision—but that doesn’t mean its view of evidence was unreasonable or that it utterly ignored crucial facts. I don’t necessarily share the Chief Judge’s skepticism about the PTAB’s recitation of KSR—I frankly don’t think the PTAB muddled the standard as much as UC said in its papers. But even if so, it’s hard for me to envision the slip opinion exclaiming that, substantial evidence be damned, this poor description of KSR makes it all reversible error. Judge Schall was silent, so I have no clue what he thinks. But if you buy into heuristics, à la Peter Lee’s Yale Law Journal article, it’s easier to rely on the PTAB and write a substantial evidence opinion than to commit to a gnarly reversal of a technically tough case. Either way, I think the Broad Institute has at least two votes in its favor. And because I don’t think the Supreme Court will touch this case with a ten-foot pole, I think this will ultimately be an ending chapter in the CRISPR patent saga.
So: what does this all mean for scholars, if anything? Well, I think a lot—but not for the typical reasons that cases make good scholarly fodder. The CRISPR case, it seems, is not legally groundbreaking in any way. It’s unlikely we’ll get an opinion with a radical departure on the substantial evidence standard or even nonobviousness. I also doubt we’ll get a fractured opinion, and the case probably won’t go en banc. Like I said earlier, the Supreme Court would never touch it.
But I think the dispute is important for patent scholars because it demonstrates that patents covering even revolutionary technology can be legally mundane and still have significant real-world impact. Who wins the CRISPR patent dispute is likely to decide a great many important things—the speed of technological development, its cost, and the extent of patent thickets in the real-world. If good scholarship strives, in part, to make sense of momentous legal events—say, the constitutional affirmance of same-sex marriage, the suspension of cost-sharing reduction payments, or the end of deferred immigration violation prosecutions—then the CRISPR patent dispute should be considered in that vein. I bring this up because there is a current of criticism of younger scholars who pick up bright, shiny developments elsewhere in the world and simply apply the law to them (guilty), especially where such developments have little to say about how the law will change in response. But oftentimes, the law simply does not change. Parties negotiate in the shadow of the law, not its flames. And as scholars who, I hope, are ultimately concerned with how things work in the real world, we should lace our enthusiasm of momentous events in the law with momentous events in the world.
More practically, attending these oral arguments also showed me how important it is to see how the law gets applied—how the sausage gets made. This is something the bulk of us has from clerking or practice experience, or both. But making a regular habit of it demonstrates how important factual nuances are, even on appeal. This is all the more important for patent scholarship where factual predicates of the underlying technology can go a great deal in informing—or misinforming—your reader. This case showed me—and keeps showing me, as this blog post makes evident—that sometimes lengthy descriptions of technology are required to even understand what the underlying factual argument is. With exception of empirical legal scholarship, there’s enough patent scholarship out there that’s thin on the technical facts and devoid of instructive hypotheticals. Ultimately, I think that’s a disservice to the judiciary and the academy—just as a gross misinterpretation of the technology would have been to the judges in this case. Many scholars, in framing their work, love the So what? question. We also love the Who decides? question. We should put the How so? question in that pantheon—how would a piece of technology produce such a disagreement about the law?
Lastly, the oral arguments yesterday provided some great fodder for future work. There were two cases argued prior to the CRISPR case that presented some fascinating questions: The first, Martin v. Wilkie (Case No. 17-1747), was a Court of Appeals for Veterans Claims case that asked whether an egregiously delayed decision by the Board of Veterans Appeals constituted a due process violation, and whether such a violation would continue after remand. That is, if the legal result of the appeal is a remand, but also an acknowledgment that the remand would increase an otherwise constitutionally violative delay in adjudication, how is a court supposed to weigh those? The second was a § 1498 patent case (Hitkansut LLC v. U.S., Case No. 17-1853) with a simple question: does § 1498 provide for prospective remedies like running royalties? Prospective remedies in private infringement cases don’t have the greatest statutory grounding. But this makes running royalties in § 1498 cases especially suspect because the statute is both jurisdictional in nature and written in the past tense. These are issues that I would never have dreamed of without hearing them live. And, given that they are predicated by real cases with fact patterns that seem to come up frequently, would make—in my estimation—good vehicles for future scholarship.