Why UC Berkeley deserves the main CRISPR patent

crispr nihSome months back a USPTO court issued a ruling that most interpreted as meaning the Broad Institute had won the so-called ‘CRISPR patent battle’ in the U.S. and that UC Berkeley, Jennifer Doudna, and Emmanuelle Charpentier had lost. Now this week Berkeley has appealed that ruling. It seems the odds are against Berkeley prevailing in its appeal, but frankly Berkeley deserves the main CRISPR patent and Broad doesn’t. Interestingly, the European Patent Office apparently agrees with this view and disagrees with the USPTO. Update: note that the Berkeley patent application itself also mentions eukaryotic use.

At the heart of the original decision that favored the Broad was an illogical argument by the USPTO court. They said that the research of Doudna and Charpentier did not make the work that the Broad later patented based on the work of Feng Zhang obvious. In my view Doudna and Charpentier’s work in fact did render Zhang’s later work a totally obvious next step.


A hypothetical scenario can help to illustrate this.

Let’s say a colleague tells me something along the lines of “Hey, I found this novel nuclease we are calling ‘DUH1’ that cuts DNA in a nifty new way in a prokaryote and in a test tube” and they publish that. Of course, after that many people are going to want to try DUH1 in eukaryotes. Duh, it’s a no-brainer, right? It’s therefore bizarre that the USPTO would think the step to try CRISPR in eukaryotes was not obvious after Doudna & colleagues groundbreaking work.

Flip it around too and imagine that the hypothetical colleague who discovered DUH1 only reported that it worked in vivo and then someone else was allowed to patent that DUH1 could be used in vitro on plasmid DNA in a tube. Does that make any sense? Someone else could patent the in vitro use of DUH1 over the inventor who discovered DUH1 first and reported how it worked in vivo? Even if was a bit of a challenge to get DUH1 to work in vitro, I don’t think that makes sense.

Back to the real CRISPR world, does the in vivo to in vitro or in vitro to in vivo or prokaryote to eukaryote “directionality” of the research flow matter for a patent? I’m not sure, but in theory it shouldn’t in this case as the next steps were obvious. How obvious?

If you read Doudna and Charpentier’s seminal Science paper, the abstract concludes with a statement for all the world suggesting the use of CRISPR-Cas9 for genomic editing in general and I took that to mean in eukaryotes too:

“Our study reveals a family of endonucleases that use dual-RNAs for site-specific DNA cleavage and highlights the potential to exploit the system for RNA-programmable genome editing.”

and the paper itself ends:

“We propose an alternative methodology based on RNA-programmed Cas9 that could offer considerable potential for gene-targeting and genome-editing applications.”

You’re telling me that these statements were meant to be restricted to only prokaryotes or DNA in a tube? Really? Nope.

Strangely the patent court apparently felt that Doudna’s public statements about it being a challenge to get CRISPR to work in eukaryotes was a big deal in rendering their decision, but again technical difficulty does not equate to an idea being non-obvious. For sure kudos to Zhang, who was technically speaking quite adept to get the CRISPR-Cas9 system to work well in eukaryotes quickly, but even if the Broad ironed out key technical kinks in getting CRISPR-Cas9 to work well inside eukaryotic cells that still doesn’t justify them having the main CRISPR patent. It’s just not conceptually or technically different enough from the earlier Doudna and Charpentier work. To me it’s not even a close call, but USPTO got it totally wrong.

Another exercise reinforces my argument. Can anyone imagine Zhang publishing his first CRISPR work (which by the way cites and heavily relies on the works of Doudna and Charpentier) if he didn’t have those earlier key papers of Doudna and Charpentier to build on moving forward? No way. Could Doudna and/or Charpentier and others have gotten CRISPR to work in eukaryotes without Zhang? Yes and almost certainly it was already inevitable before Zhang even published his key Science paper.

For all these reasons, Berkeley deserves the main patent based on simple common sense, but whether things will turn out that way longer term seems far less clear.

Some may say that no one should get to patent CRISPR, but these days that’s probably a naive perspective. For more on the history of patenting (or lack thereof) of nucleases and in particular restriction enzymes, this is an interesting read. 

7 thoughts on “Why UC Berkeley deserves the main CRISPR patent”

  1. stemcellrepairman

    The real lesson here is, as a bioscientist applying for a patent, never say anything that can be misconstrued (or just never talk to the press period without great care).

    I seem to remember the court deciding that George Church had nothing of import to say on the patent dispute and that the experience and direct observations of a rotation student in Feng’s lab at the time were also not relevant.

    I also seem to remember Dana Carroll’s analysis of Feng’s patent leading to the claims being somewhat incongruous with the data presented (specifically, if you follow Feng’s data in the patent, one reaches the conclusion that Cas9 itself is not required for modification).

    On one hand, you have the legal definition of “non-obvious”.
    On the other, you have several major genome engineers stating the eukaryotic application was immediately obvious to anyone skilled in the art as soon as the Doudna paper went up.

    This whole thing was not handled well by any party, in my opinion

  2. Reminds me of the WARF patents on human ES cells. Their claim was that it was NOT obvious to use the same methods that had been used to derive mouse ES cells from mouse blastocysts in 1981 to derive human ES cells from human blastocysts in 1998. It still makes my head spin…

  3. I am going to have to disagree with Shinsakan. I read almost all documents pertaining the case and I don’t think they suggest that Broad (or any other molecular biology lab for that matter) did not have a reasonable expectation of success in applying Doudna’s methods to eukaryotic cells.

    The purpose of the non-obviousness requirement is to guarantee that patents reward *inventive skill*, not routine optimization. The problem is that, while everyone agrees that routine work should not be granted patents, people cannot agree on what exactly routine work is. In any case, the evidence the PTAB used to support the claim of unpredictability seemed to me debatable at best. First, they seemed to have relied much more heavily on Broad’s expert witness than on Berkeley’s , for reasons that are not clear. Second, and most importantly, they cited quotes from Doudna herself saying variations of “we still cannot be sure it will work on eukaryotic cells” as proof of unpredictability. However, any scientist knows that a researcher will not say something works *unless it has already been done*. This is very different from the legal concepts of “reasonable expectation of success” and “undue experimentation”: in patent law, it’s very well established that you don’t need to prove something works to be entitled to “a reasonable expectation of success” (hence the term “reasonable”).

    The problem with this case, and many others, is that the people judging whether something requires undue experimentation do not understand the science well enough. I haven’t met any molecular scientist so far that did not think that, once Doudna et al’s papers were published, it was just a matter of time and routine skill to make it work in eukaryotes. Zhang was rewarded for being the first to do something that many others could easily replicate – and that’s exactly what patents are supposed to NOT do.

  4. Obviousness is not only about conceptual obviousness. Overcoming technical difficulty can actually result in non-obviousness even when something is conceptually obvious. For something to be obvious, it has to be not only conceptually obvious to do but also be able to be achieved without undue experimentation and with a reasonable expectation of success. Here, the finding was apparently that even though they had clearly conceptually described the use of CRISPR-Cas9 in eukaryotic cells in their patent application, at the time of invention for Berkeley (remember that this time predates the time they published their papers but is the critical time for patenting), they had not actually achieved it, and the evidence submitted by the Broad showed that undue experimentation beyond what Berkeley had done was required to get it to work in eukaryotic cells. I am not saying that I agree or disagree with this because I honestly don’t know whether or not going eukaryotic really required so much further experimentation or tweaking to take it from “routine” to “undue”, but at least based on the evidence submitted to the court, the ruling seems to make sense. Although the decision might not feel right, the decision is based on the law, not feelings. The law prevents people from patenting things they have not actually done or that would not have been easily within their reach to do.

    1. Hi Shinsakan,
      Good points, including about what was in the UC patent versus in their first paper.
      But I don’t think the court successfully hit upon non-obvious achievements in support of the Broad’s patent even from a technical perspective.
      What did Zhang do that was non-obvious?
      Was that really undo experimentation beyond what others had already done or would soon do themselves in a routine, logical progression?
      I don’t think so, but clearly others disagree.
      I also think the concept is most important here (over technical optimization) and that was Doudna/Charpentier’s without a doubt.

      1. Paul,

        Thanks! I agree, but for patents (as opposed to science) it’s not just the concept that is most important- it’s the combination of concept + execution. I also don’t know what Zhang did that was non-routine- my (very limited) understanding is that it might have had something to do with the guide strands? If I remember right, I think that Zhang did talk about what he did in the declarations made to the court. However, I think that the decision is more about Doudna/Charpentier than about Zhang- the court is saying that it would have been non-obvious for anyone, not just Zhang, to go from Doudna/Charpentier’s disclosure to eukaryotic cells (based at least in part on the evidence submitted that Doudna’s lab was having difficulty getting it to work in eukaryotic cells). So when Zhang shows up with actual data for eukaryotic cells, then that would make it non-obvious, no matter how he got there. I think that what I would want to know is whether it really was such a big deal to go from prokaryotic to eukaryotic cells, and also why Berkeley didn’t wait to file until they had at least some eukaryotic data. Maybe they knew that it was still going to take some time and that the transition to first-to-file was coming up, at which point they would lose everything if someone else filed first, so they decided to hedge their bets? Or maybe they just didn’t even think or realize (like most of us) that eukaryotic vs. prokaryotic would be an issue?

Comments are closed.