Patent expert weighs in on CRISPR dispute between UC & Broad

CRISPR dispute patent

Adapted from NIH CRISPR-Cas9 image

The patent dispute on CRISPR between UC/Jennifer Doudna and The Broad/Feng Zhang has been the subject of major attention including in a recent piece on Stanford Center for Law & Biosciences Blog. There is a lot of confusion over this important CRISPR dispute so I turned to a patent expert for their take on this via an interview below.

The interview with this anonymous expert provides some helpful, informed assessments on the CRISPR patent dispute including fresh perspectives. This person has different views than those expressed in other media outlets and in some instances from those in that recent piece on the Stanford Law/Bioethics Blog.

(1) Are there elements to this patent situation that most people are overlooking?

Answer: Probably the most important thing that people are overlooking is the scope of the claimed subject material of the Doudna patent application (and also the Zhang patents in question); only Cas9 is claimed. The claims do not cover other nucleases that could work with CRISPR (such as Zhang’s recently published Cpf1, which is “smaller and easier to program than Cas9”) and thus would not exclude someone from using other nucleases. Therefore, I think that even if the Doudna patent is issued and the Zhang patents are pulled, the Doudna patent would still not exclude someone from using CRISPR/Cpf1 (or CRISPR with other non-Cas9 nucleases that may be eventually discovered or engineered). Similarly, the current Zhang patents also would not exclude someone from using CRISPR with non-Cas9 nucleases, although I assume that Zhang and the Broad Institute will now also be applying for patent coverage for Cpf1. Indeed, the Broad Institute’s press release for the discovery of Cpf1 mentioned that “The Broad Institute and MIT plan to offer non-exclusive licenses to enable commercial tool and service providers to add this enzyme to their CRISPR pipeline and services” and “We see much more to come, even beyond Cpf1 and Cas9, with other enzymes that may be repurposed for further genome editing advances.” Interestingly, this means that if the Doudna patent is issued and the Zhang CRISPR/Cas9 patents are pulled, Zhang and the Broad Institute (and their licensees) will still be able to use CRISPR/Cpf1, whereas if the Doudna patent application is rejected and the Zhang CRISPR/Cas9 patents are left intact and Zhang then patents CRISPR/Cpf1, then Zhang and the Broad Institute could potentially control the entire current CRISPR intellectual property landscape (at least until someone finds or engineers yet another CRISPR-compatible nuclease). Of course, I should also point out that all of this is assuming that there will be no settlement between the UC and the Broad Institute in the current interference proceeding; some kind of settlement could still happen. These aspects regarding the actual scope of the patent claims have been largely ignored by most media reports and were not mentioned in the Stanford Law blog. Therefore, although the outcome of the CRISPR/Cas9 patent battle is indeed important as this is the first wave of patents for this technology, it is important to keep things in perspective, and some assertions from the Stanford Law blog such as “needless to say, this is a monumental event for patent attorneys, molecular biologists, the PTO, and the world” and “the biotech patent dispute of this century” feel to me somewhat overstated.

(2) Do the recent developments chronicled on the Stanford Law Blog favor Doudna’s patent claim?

Answer: The recent developments chronicled in the Stanford Law blog simply amount to the initiation of the interference proceeding, which I think that people have been expecting for a while, perhaps ever since the Zhang patents were issued and it was known that the Doudna application was still under examination. The USPTO’s decision to actually institute the interference proceeding means that the USPTO supports the Doudna team’s position that the scope of the inventions claimed by Doudna and Zhang overlaps and that the claims are in conflict. This is a step forward for Doudna’s team because it opens up the possibility that the Doudna patent could be issued and that the Zhang patents could be pulled. However, the initiation of the interference proceeding itself does not favor or disfavor Doudna’s patent claim; indeed, the entire purpose of having the interference proceeding is to determine whether or not Doudna’s patent claims can be allowed, and as the Stanford Law blog pointed out, decisions to institute interference proceedings are largely pro forma. Basically, we still have no idea who will come out on top, except that it should be whoever can show that they invented CRISPR/Cas9 first.

Regarding the reporting in the Stanford Law blog, though, I disagree with trying to draw a difference between Zhang and Doudna with regard to the modification of eukaryotic cells; Doudna specifically recites it in the originally filed claims (see, e.g., original claim 95), not to mention discussing it in the originally filed specification. Genome editing of human cells is explicitly presented in Doudna’s Example 2, and in vivo genome editing in mice is described in Example 7. Thus, the argument that Doudna did not contemplate eukaryotic applications and the suggestion that Zhang’s claims might thus somehow be more worthy, valuable, or patentable (on this basis) is incorrect. Furthermore, it should be noted that patent claims allow the patent owner to exclude someone else from practicing the invention; they do not give the owner a right to practice what is claimed. Therefore, even if hypothetically Doudna had not explicitly contemplated eukaryotic applications (which is not the case), the broader scope of the resulting claims (which would basically cover any genomic editing by CRISPR/Cas9, not just in eukaryotic cells) might even be considered to make the Doudna patent more (and not less) valuable than the Zhang patent.

Despite disagreeing with some elements of its analysis of the actual patents, I do however think that the Stanford Law blog did a great job of describing what happens in interference proceedings, which are mysterious even to many patent practitioners.

(3) What could be the deciding factor in this patent dispute?

Answer: Because the application was filed under the first-to-invent rule, I think that in the end that it will come down to establishing the actual date of invention on both sides, i.e., at what point both sides achieved what they are claiming. If Doudna has lab data showing that she invented CRISPR/Cas9 before Zhang, then her patent should be allowed and Zhang’s patents should be pulled. If Zhang instead has lab data showing that he invented CRISPR/Cas9 first, then his patents should stand and Doudna’s application should be rejected.

However, Doudna’s team could also go after Zhang’s patents or vice versa by attacking the patentability of the claims themselves under the written description and enablement requirements, i.e., by saying that the invention is not described sufficiently to indicate that Doudna/Zhang was actually in possession of the claimed matter at the time of filing/invention or to enable one of ordinary skill in the art to actually practice the invention. The outcome would be to invalidate Zhang’s patents or reject Doudna’s application on that basis. It should be noted, though, that the enablement and written description aspects were already considered to not be an issue by the examiners for both Zhang’s patents and Doudna’s application. The patentability of the claims over the prior art could also potentially be made an issue, although similar to the enablement and written description aspects, the examiners already found Zhang’s patents and Doudna’s application to be clear of any prior art, and any prior art that one side further attempts to apply would probably also apply to the other side in terms of date, as the dates of invention and the claimed subject material seem to be so similar.

Another possibility is that if the claims of either Doudna’s application or Zhang’s patents, or both, could be amended so that the claimed material no longer overlaps, then Zhang could keep his patents and Doudna’s patent could still be issued. However, given that they are both claiming the same fundamental CRISPR/Cas9 technology, I think that this is unlikely. Still another possibility is that some other kind of settlement could be reached, although it seems as though they have already been trying to settle without success, and as the Stanford Law blog rightly notes, settlement is discouraged in interference proceedings. Thus, I predict that in the end it will probably amount to who has the earliest lab records to determine whether Doudna’s application is issued and Zhang’s patents are pulled or whether Doudna’s application is rejected and Zhang keeps his patents (unless there is a settlement of some kind).

(4) There have been some suggestions that the original UC attorneys were outplayed by those at the Broad. What do you think?

Answer: I don’t really think so. The strategies used are different, but it is difficult to say that one is better than the other. The Broad Institute attorneys opted for the fast track to ensure early granting and as a result had to limit what they were claiming (hence the limiting of the claims to eukaryotic cells) and submit relatively few claims. The UC attorneys submitted lots of claims with a broader scope and more coverage, but then had to deal with the longer conventional prosecution process. I think that one could argue, though, that in the end the pathway used wouldn’t make a much of a difference with regard to total patent coverage that could eventually be obtained, as subsequent applications with broader claims could be filed to increase coverage in the case of Zhang, or an interference proceeding can be requested if the patent prosecution process takes too long and someone else patents the invention first in the case of Doudna.

Furthermore, I think that the question of whether or not the UC’s original attorneys were outplayed by the Broad’s may not even be relevant because the outcome of this patent dispute will likely not depend on whose patents were issued first or on whose patent claims initially had a broader scope. For example, if Doudna’s attorneys had instead filed a fast-track application where the patent was granted before Zhang’s but Zhang believed that he had invented CRISPR/Cas9 before Doudna, then Zhang could have requested the interference proceeding and we would still basically be where we are today. Rather, I think that the quality of the initial attorneys and what they did will only impact the interference proceedings if written description or enablement issues come up, because the resolution of such issues would depend on how solidly the patents/patent applications were written. I think that the arguments made by the currently litigating attorneys will have a greater impact, but the greatest impact will probably come from the dates in the lab records.

(5) What about the “mysterious third party” in this patent situation?

Answer: It could be anyone. However, a couple of things to note:

(a) Whoever it is used a solo attorney (not a firm) who doesn’t seem to do much (if any) life science/biotech work, so I am guessing that (like the Yamanaka patent challenge) they are not serious and are not major players in the field.

(b) The attorney is based out of the San Francisco Bay Area, so I think that whoever it is probably is from the Bay Area, too. I think that it’s highly unlikely that someone from Boston, New York, DC, or another major city (or another country, for that matter) would turn to a local solo non-biotech practitioner who is not from their own locality for this. A look at the patent applications handled by the former firm of the attorney confirms that almost all are from Bay Area inventors. However, another possibility would be if the third party is a friend or relative of the attorney.

(c) Most importantly, the “prior art” cited in both 3rd party submissions was largely filed or published after Doudna’s effective filing date, and thus well after her presumed invention date; the only documents that predate her filing/invention date only generally describe the CRISPR/Cas system as it exists naturally in bacterial cells, which would not render obvious what Doudna did in turning it into a bioengineering tool by including the targeting RNA and activator RNA and then using it in other cell types. Therefore, none of the documents are actually applicable as prior art, and so it is really unclear why the “mysterious third party” would have thought that filing these submissions would affect the granting of Doudna’s patent. The examiner working on this thought the same thing, as she did not consider a single one of the references supplied by the third party to be applicable to reject Doudna’s application.

(d) Furthermore, some of the documents in the 3rd party submission (for example, WO2013141680, which is the main patent application cited) were even already disclosed in the information disclosure statements originally filed with Doudna’s application- so the USPTO had already even been informed about some of these documents by Doudna’s own team. Therefore, I think that the “mysterious third party” submissions were never really an issue, although the fact that someone was actively trying to prevent Doudna’s patent from being issued is indeed interesting.

19 thoughts on “Patent expert weighs in on CRISPR dispute between UC & Broad


  1. Paul-
    Well, I believe that your interviewee is an attorney- only an attorney would write a reply so dense, jargon-rich, and nuanced. Even though I’ve read a lot of stuff like this, I still thought: “kill me now…” halfway through.

    I can find the answer to this question, but it’s easier to ask. Did Doudna file under first to invent, while Zhang’s later filing was under the new rules, first to file?


    • @Jeanne,
      I personally thought the interviewee was quite well-spoken and clear. With all your patent experience, what’s your take on this CRISPR dispute?
      Elliot in his comment addressed your question.
      Paul


  2. On October 15, 2013, Zhang filed his own patent application (No. 14/054,414)—months after the first-to-file rules came into effect—but claiming a December 12, 2012 priority date under the old first-to-invent rules.

    Doudna’s original patent application (No. 13/842,859), filed on March 15, 2013—a day before the PTO’s first-to-file rules came into effect—but given a priority date of May 25, 2012

    from the Stanford Law post


  3. @ Dr. Loring:

    Both applications (Doudna’s and Zhang’s) claimed priority to back before the transition to first-to-file, so both filings were under first-to-invent. The effective filing date for Doudna was 5/25/2012, whereas that for Zhang was 12/2/2012. The transition to first-to-file was 3/16/2013.


  4. I was left confused, but Shinsakan explained that both patents fall under the “first to invent” rules because it’s the priority date, not the filing date, that determines what rules the patent falls under. I know, why call it “first to file” if the filing date isn’t so relevant?

    My take is that I’m glad that there is discussion going on in public. I’m learning from it.


  5. @ Dr. Loring:

    With “first-to-file,” the effective filing date (i.e., the earliest priority date) is everything, and the filing date is indeed relevant. With “first-to-invent,” it’s the date of invention (which is different from the filing date or priority date).

    If the current patents were filed were under “first-to-file,” then there would be no dispute because the effective filing date for Doudna is 5/25/2012, whereas that for Zhang is 12/12/2012. Doudna clearly filed first.

    However, because the patents were filed under “first-to-invent,” it may still be possible for Zhang to have invented CRISPR/Cas9 first, even though Doudna filed the patent application first. The only way to determine who was first to invent (as opposed to first to file) is to evaluate the lab records to see when the invention was actually achieved on both sides.

    Because the US moved to first-to-file in March 2013, “first-to-invent” disputes such as the CRISPR dispute will soon become a thing of the past.


  6. Note — “priority” date does refer to an earlier patent filing — in both Doudna and Zhang’s cases the ‘priority’ filing was a United States provisional patent application filing.


  7. I don’t think anybody actually “invented” CRISPR/Cas9 – it already existed in nature – so is it not the tools and uses thereof that are claimed?

    The first to file (or invent) rule only comes into effect if others make the same claims later. So do the patents make exactly the same claims?

    Also if I understand correctly, Doudna’s claims were drastically rewritten twice after the priority filing in order to “more clearly” claim the use of CRISPR in eukaryotes. But if these revisions are not upheld as being priority valid claims then Doudna may get the first patent, but not for use in eukaryotes – and that’s where the business end is.


  8. @ Gary:

    (1) Although CRISPR/Cas9 is present in nature, the CRISPR/Cas9 systems claimed by Zhang and Doudna include elements not present in nature (such as the expression vector and targeting of euakryotic DNA in Zhang or the targeting/activating RNA duplex of Doudna) Therefore, both Zhang and Doudna claim non-naturally occurring CRISPR/Cas9 systems, in addition to methods of using them.

    (2) The scope of the patent claims is the same. This seems to be at least part of why Doudna made the final set of claim amendments – to ensure that the scope of the claims overlaps sufficiently to have the interference proceeding instituted.

    (3) First, Doudna’s amendments to the claims are considered valid because Doudna had already described the use of CRISPR in eukaryotic cells in the originally filed specification and claims, so there is no new matter being introduced. Second, as the post above described, a patent gives someone the right to exclude someone from using the invention- it’s doesn’t give someone the right to use an invention. This means that even if Doudna’s claims did not specifically claim eukaryotes, they would still exclude someone else from using CRISPR/Cas9 in eukaryotes because the claims more generally cover the of CRISPR/Cas9 to edit any DNA. Regardless, the use in eukaryotes is specifically supported in Doudna’s originally filed specification and claims.


  9. @Shinsakan – thanks for those very clear explanations. So it really is a claim by claim battle, and if Doudna’s claims are already accepted as being priority-setting, it looks bad for The Broad Institute and the spin-offs based on their patents.

    People will find ways around Cas9, (such as Cpf1 and others) to give them freedom to operate, but the core market for fast gene editing tools will certainly belong to Cas9 for some time.

    It’s going to be interesting to see how this plays out on the markets if and when Caribou, Editas, et al. plan for IPOs.


  10. @Gary:

    “Priority” refers to the earliest date that the individual applications can refer back to as their effective filing date; it does not mean that Doudna has been given “priority” over Zhang. It has not yet been concluded that Doudna invented the claimed subject matter first (which is what will establish who gets the patent); that is the purpose of having the interference proceeding. I don’t think that it can yet be said that it looks either bad or good for either Doudna or Zhang. We simply do not have enough information as to who actually invented first.

    I agree that because Cas9 was the first characterized nuclease, most of the approaches thus far have used Cas9, thus forming the core of the current market. However, I think that other nucleases such as Cpf1 are more than just ways around a central Cas9-based technology. For example, Cpf1 has been suggested to be “smaller and easier to program than Cas9.” I don’t see what would prevent rapid or widespread adoption of other nucleases (such as Cpf1) as they become available. The fundamental part of the technology is the CRISPR phenomenon, and not necessarily the specific nuclease used to cut the DNA. Cas9 was the first one to be found, but that doesn’t mean that it is the only one or even the best one.


  11. @Shinsakan – in theory nothing prevents widespread adoption of other nucleases, but markets don’t always follow logic. The first mover principle will keep Cas9 as market leader for some time, even if Cpf1 is ultimately better in some way (like VHS vs. Betamax). But this is only relevant to the tools market and not the therapeutics market that will employ many different systems for development.


  12. @Gary:

    I don’t think that VHS vs. Betamax is an apt comparison here because VHS and Betamax were completely different technologies with different equipment where manufacturing companies basically had to commit to building one of two entirely different devices for one of two entirely different formats of content, leading to a “format war.” In contrast, CRISPR/Cas9 and CRISPR/Cpf1 both rely on the same CRISPR phenomenon for the same potential applications; the primary difference only lies in which nuclease is used. Cpf1 could easily be swapped for Cas9 in a kit, or both types of kits could easily be manufactured side-by-side, with both options then available to consumers even from the same manufacturer (in the same way that the same manufacturer can offer different antibodies for immunoassays that may target different epitopes but still target the same protein). A better example might be the diversity of commercially available iPS cell generation kits, where many of these kits do not use the original combination of Yamanaka factors; even the “tools market” can use many different systems.

    Furthermore, Betamax was actually introduced 2 years before VHS (1975 vs. 1977 in the US, respectively), so wouldn’t the first mover principle actually suggest that Betamax should have won over VHS? If anything, the Betamax vs. VHS example suggests that rather than or perhaps to a greater extent than the introduction date, marketing and cost to the end user might actually play a larger role in what gets widespread adoption.

    In terms of advantages, Cpf1 only requires one RNA in the complex (rather than two), which might make it easier to use than Cas9. Cpf1 is also smaller, which might make it easier to deliver into cells and tissues if a kit using recombinant protein is used. Cpf1 targets PAM sequences different from those targeted by Cas9 that might make it better at editing the human genome than Cas9. Cpf1 cuts far away from the recognition site (unlike Cas9), which might facilitate re-cutting and correct editing if the target gene becomes mutated at the cut site (Cas9 genome editing is known to not have 100% fidelity). Cpf1 cuts with staggered overhangs rather than the blunt ends of Cas9, which should make the cut sites less prone to mutation. See http://www.genengnews.com/gen-news-highlights/crispr-cpf1-may-outsnip-crispr-cas9/81251791/ for a more extensive discussion of these points.

    Based on these perceived advantages, I think that it would be easy for a manufacturer to market a CRISPR/Cpf1 kit as a research tool and gain market share, even if CRISPR/Cas9 was introduced first. It may still be possible for CRISPR/Cas9 to hold a high proportion of the market share because of its earlier introduction and greater familiarity among researchers, but I don’t think that it would necessarily be totally dominant, especially as additional nucleases become available and if strong marketing tactics are applied.


  13. @Shinsakan – a first mover is the first to secure a new market, but not necessarily the first to invent, to make or try to sell – Betamax may have existed first, but they were not able to create the market, whereas VHS did so through aggressive marketing – hence VHS was the first-mover.

    Cas9 has achieved this position for the rapid inexpensive gene editing market even before ownership has been sorted out. So despite both Cas9 and Cpf1 both using the CRISPR phenomenon, in the tools market Cpf1 will not be aiming to create a new market, but to impinge on the market of the fist-mover – and that is much more difficult as trust and experience is now associated with Cas9. Cpf1 would need to show advantages or different uses over Cas9, similar to the different antibodies or iPSC kits you mentioned. This is possible through nuclease tailoring, which Zhang has been involved in recently.

    But I’m not giving any investment advice here, I leave that to seekingalpha.com


  14. @Shinsakan and @Gary:

    From a business perspective, biotech companies trying to go after Cpf1 makes sense. After all, they can market to those same labs that can’t do molecular biology themselves.

    However, from a biological perspective and from what’s shown in the paper, only one advantage actually holds up at this point: Cpf1 is smaller.

    The two-RNA thing can be gotten around by the fusion guide, so a non-advantage for Cpf1.

    Cutting far from the recognition site wasn’t actually demonstrated to be advantageous in the paper. If Cpf1 behaves anything like Cas9, that may actually be neutral or a disadvantage due to the amount of time these nucleases remain bound to DNA.

    The staggered versus blunt really bothers me because it seemingly stems from a misunderstanding of the literature and an incorrect application of traditional in vitro plasmid cloning principles. If anything, with what is known about how mammalian cells respond to different types of DNA double strand breaks, blunt breaks are repaired much more precisely than overhangs. Overhangs in a mammalian cell context are actually preferentially chewed back and lost.

    Also, there’s a good portion of the therapeutics market that wouldn’t use either of these (as well as TALENs).


  15. @Repairman:

    Thanks for the insight! Regarding the smaller size, do you think that this advantage would also be neutralized if the kit uses gene delivery for expression of the nuclease within the cell (as recited in the patent claims of both Zhang and Doudna for Cas9) rather than trying to get the entire intact nuclease molecule into the cell and nucleus?

    @Gary:

    What you are saying makes sense, and I know how important trust and experience are for lab reagents. I am just wondering whether such trust and experience issues would really be a barrier to the adoption of different nucleases if provided by a trusted source, as the underlying principle/mechanism of CRISPR is still the same (to me, this seems to be like introducing RT and PCR kits with different varieties of polymerase), and whether VHS-style marketing could be applied to a CRISPR kit using another nuclease to increase its popularity.


  16. @Shinsakan

    It would depend on what the application is.
    If the nuclease is being used for cell culture (in a non-therapeutic manner), size advantage is lost because one can always find a relatively useful method for transfection.

    If the nuclease is being used in vivo or for a cell-based therapeutic, I could see the size mattering somewhat, but this is more based on the traditional thinking of desiring a single vector system rather than a mix for viral (particularly AAV) delivery. Cpf1 might be better in this case, although Editas and Feng’s lab spent a lot of time on the Staph aureus Cas9, which is also smaller than SpCas9 (SaCas9 may be bigger than Cpf1, I can’t remember off the top of my head).

    Bottom line: There’s a lot of claims in that Broad press release about what Cpf1 could presumably be better at, but the paper doesn’t really demonstrate any of those.

Leave a Reply